Altered Fovea in AQP4-Igg–Seropositive Neuromyelitis Optica Spectrum Disorders E805

Home , GM2 (ganglioside), Luigi Novarese, Tremor (character)

Volume 7, Number 5, September 2020 Neurology.org/NN

A peer-reviewed clinical and translational neurology open access journal

ARTICLE Altered fovea in AQP4-IgG–seropositive neuromyelitis optica spectrum disorders e805

ARTICLE Treatment and outcome of aquaporin-4 antibody–positive NMOSD: A multinational pediatric study e837

ARTICLE Guillain-Barré syndrome: Th e fi rst documented COVID-19–triggered autoimmune neurologic disease: More to come with myositis in the offi ng e781

ARTICLE Extending rituximab dosing intervals in patients with MS during the COVID-19 pandemic and beyond? e825 Academy Officers Neurology® is a registered trademark of the American Academy of Neurology (registration valid in the United States). James C. Stevens, MD, FAAN, President Neurology® Neuroimmunology & Neuroinflammation (eISSN 2332-7812) is an Orly Avitzur, MD, MBA, FAAN, President Elect open access journal published online for the American Academy of Neurology, Ann H. Tilton, MD, FAAN, Vice President 201 Chicago Ave, Minneapolis, MN 55415, by Wolters Kluwer Health, Inc. at 14700 Citicorp Drive, Bldg 3, Hagerstown, MD 21742. Business offices are Carlayne E. Jackson, MD, FAAN, Secretary located at Two Commerce Square, 2001 Market St., Philadelphia, PA 19103. Production offices are located at 351 West Camden Street, Baltimore, MD Janis M. Miyasaki, MD, MEd, FRCPC, FAAN, Treasurer 21201-2436. © 2020 American Academy of Neurology. Ralph L. Sacco, MD, MS, FAAN, Past President Neurology® Neuroimmunology & Neuroinflammation is an official journal of the American Academy of Neurology. Journal website: Neurology.org/NN, AAN website: AAN.com CEO, American Academy of Neurology Copyright and Permission Information: Please go to the journal website Mary E. Post, MBA, CAE (www.neurology.org/nn) and click the Permissions tab for the relevant article. Chief Executive Officer Alternatively, send an email to [email protected]. General information about permissions can be found here: https://www.lww.com/ 201 Chicago Ave journal-permission. Minneapolis, MN 55415 Disclaimer: Opinions expressed by the authors and advertisers are not necessarily those of the American Academy of Neurology, its affiliates, or of Tel: 612-928-6100 the Publisher. The American Academy of Neurology, its affiliates, and the Publisher disclaim any liability to any party for the accuracy, completeness, efficacy, or availability of the material contained in this publication (including Editorial Office drug dosages) or for any damages arising out of the use or non-use of any of the Patricia K. Baskin, MS, Executive Editor material contained in this publication. Rachel A. Anderson, Administrative Assistant Advertising Sales Representatives: Wolters Kluwer, 333 Seventh Avenue, New York, NY 10001. Contacts: Eileen Henry, tel: 732-778-2261, fax: 973-215-2485, [email protected] and in Europe: Craig Silver, tel: +44 7855 Morgan S. Sorenson, Managing Editor 062 550 or e-mail: [email protected]. fl Careers & Events: Monique McLaughlin, Wolters Kluwer, Two Commerce Neurology® Neuroimmunology & Neuroin ammation Square, 2001 Market Street, Philadelphia, PA 19103, tel: (215)-521- 8468, fax: Neurology® Genetics 215-521-8801; [email protected]. Reprints: Meredith Edelman, Commercial Reprint Sales, Wolters Kluwer, Two Commerce Square, 2001 Market Street, Philadelphia, PA 19103, tel: Kathleen M. Pieper, Senior Managing Editor, Neurology® 215-356-2721; [email protected]; reprintsolutions@ Karen Skaja, Senior Editorial Coordinator wolterskluwer.com. Skyler M. Kane, Editorial Coordinator Special projects: US & Canada: Alan Moore, Wolters Kluwer, Two Commerce Square, 2001 Market Street, Philadelphia, PA 19103, tel: 215-521-8638, Margaret A. Rei, Editorial Coordinator [email protected]. International: Andrew Wible, Senior Manager, Rights, Licensing, and Partnerships, Wolters Kluwer, translationrights@ wolterskluwer.com. Lee Ann Kleffman, Managing Editor, Neurology® Clinical Practice

Andrea Rahkola, Production Editor Kristen Swendsrud, Production Coordinator

Sharon L. Quimby, Digital Managing Editor Kaitlyn Aman Ramm, Digital Multimedia/Graphics Coordinator Justin Daugherty, Digital Multimedia/Podcast Coordinator Madeleine Sendek, MPH, Digital Marketing Communications Coordinator

Publisher Wolters Kluwer Baltimore, MD

Publishing Staff Sarah Carrera, MBA, Executive Publisher Jessica Heise, Production Team Leader, Neurology Journals Megen Miller, Production Editor Stacy Drossner, Production Associate

Neurology® Neuroimmunology & Neuroinflammation

Editor JosepO.Dalmau,MD,PhD Vision Neurology® Neuroimmunology & Neuroinflammation will be the premier Deputy Editor peer-reviewed journal in the fields of Scott S. Zamvil, MD, PhD, FAAN neuroimmunology and neuroinflammation.

Associate Editors Marinos Dalakas, MD, FAAN Mission Neurology: Neuroimmunology & Michael Heneka, MD Neuroinﬂammation will provide neurologists Dennis Kolson, MD, PhD and clinical research scientists with Friedmann Paul, MD outstanding peer-reviewed articles, editorials,

Editorial Board and reviews to enhance patient care and Rohit Aggarwal, MD, MS Jun Li, MD, PhD facilitate basic and translational research. Sergio E. Baranzini, PhD Lin Mei, MD, PhD Susanne Benseler, MD Andrew McKeon, MD Jeffrey L. Bennett, MD, PhD, FAAN Christopher Power, MD Editorial Tel: 612-928-6400 Séverine Boillée, PhD Richard M. Ransohoff,MD Inquiries Toll-free: 800-957-3182 (US) Matthijs C. Brouwer, MD, PhD Christine Stadelmann, MD Fax: 612-454-2748 Wolfgang Brück, MD Israel Steiner, MD [email protected] David B. Clifford, MD, FAAN Lawrence Steinman, MD Russell C. Dale, MBChB, MRCP, PhD Jesper Tegnér, PhD, MSc Oscar H. Del Brutto, MD, FAAN Silvia N. Tenembaum, MD Stay facebook.com/NeurologyNN Thomas G. Forsthuber, MD Maarten J. Titulaer, MD, PhD Connected Kazuo Fujihara, MD Ari Waisman, PhD twitter.com/greenjournal Jeffrey M. Gelfand, MD, MAS, FAAN Hugh J. Willison, MBBS, PhD Joseph El Khoury, MD Gregory F. Wu, MD, PhD youtube.com/user/NeurologyJournal

Markus P. Kummer, PhD instagram.com/aanbrain

Neurology® Journals

Editor-in-Chief Classification of Evidence Review Team JoséG.Merino,MD,MPhil,FAAN Melissa J. Armstrong, MD Richard L. Barbano, MD, PhD, FAAN Deputy Editor Richard M. Dubinsky, MD, MPH, FAAN Olga Ciccarelli, MD, PhD, FRCP Jeffrey J. Fletcher, MD, MSc Gary M. Franklin, MD, MPH, FAAN Section Editors David S. Gloss II, MD, MPH&TM Biostatistics John J. Halperin, MD, FAAN Richard J. Kryscio, PhD Jason Lazarou, MSc, MD Sue Leurgans, PhD Steven R. Messé, MD, FAAN V. Shane Pankratz, PhD Pushpa Narayanaswami, MBBS, DM, FAAN Alex Rae-Grant, MD Classification of Evidence Evaluations Gary S. Gronseth, MD, FAAN

Equity, Diversity, & Inclusion RoyH.Hamilton,MD,MS,FAAN Holly E. Hinson, MD, MCR, FAAN

Podcasts Stacey L. Clardy, MD, PhD, FAAN Jeﬀrey B. Ratliﬀ, MD, Deputy Podcast Editor

Ombudsman Jonathan W. Mink, MD, PhD, FAAN

Scientiﬁc Integrity Advisor David S. Knopman, MD, FAAN

e807 Ocrelizumab shorter infusion: Primary results from the ENSEMBLE PLUS substudy in patients with MS H.-P. Hartung, on behalf of the ENSEMBLE Steering Committee members and study investigators Open Access Class of Evidence

e805 Altered fovea in AQP4-IgG–seropositive neuromyelitis optica spectrum disorders S. Motamedi, F.C. Oertel, S.K. Yadav, E.M. Kadas, M. Weise, J. Havla, Cover Image M. Ringelstein, O. Aktas, P. Albrecht, K. Ruprecht, J. Bellmann-Strobl, 3D foveal shape analysis method overview. The left panel shows inner H.G. Zimmermann, F. Paul, and A.U. Brandt limiting membrane surface smoothing and radial reconstruction using the Open Access cubic Bezier polynomial. The right panel shows rim disk (blue), slope ﬂ e809 Recent cannabis use in HIV is associated with reduced disk (red), and pit at disk (green). The lower panel shows rim height, ﬂ average pit depth, and central foveal thickness. Stylized by Kaitlyn Aman in ammatory markers in CSF and blood Ramm, Digital Multimedia/Graphics Coordinator. R.J. Ellis, S.N. Peterson, Y. Li, R. Schrier, J. Iudicello, S. Letendre, E. Morgan, B. Tang, I. Grant, and M. Cherner See page e805 Open Access

Editor’s Corner e817 Antiparanodal antibodies and IgG subclasses in acute autoimmune neuropathy e858 N2 in the time of COVID-19 L. Appeltshauser, A.-M. Brunder, A. Heinius, P. Kortv¨ ´elyessy, J. Dalmau K.-P. Wandinger, R. Junker, C. Villmann, C. Sommer, F. Leypoldt, and K. Doppler Open Access Open Access Editorial, p. e843

Editorial e815 CSF chitinase 3-like-1 association with disability of e843 Isotyping paranodal antibodies in inflammatory primary progressive MS neuropathies: One step closer to precision care F. Pérez-Miralles, D. Prefasi, A. Garc´ıa-Merino, F. Gascon-Gim´ énez, N. Medrano, J. Castillo-Villalba, L. Cubas, C. Alcalá, S. Gil-Perot´ın, C. Lleixà and L. Querol R. Gomez-Ballesteros,´ J. Maurino, E. Alvarez-Garc´ ´ıa, and Open Access B. Casanova Open Access Class of Evidence Articles e819 Painful trigeminal neuropathy associated with e781 Guillain-Barré syndrome: The first documented anti-Plexin D1 antibody COVID-19–triggered autoimmune neurologic T. Fujii, R. Yamasaki, Y. Miyachi, K. Iinuma, Y. Hashimoto, N. Isobe, disease: More to come with myositis in the offing T. Matsushita, and J.-i. Kira Open Access M.C. Dalakas Open Access e825 Extending rituximab dosing intervals in patients with MS during the COVID-19 pandemic e787 Risk of COVID-19 infection in MS and neuromyelitis optica spectrum disorders and beyond? A. Maarouf, A. Rico, C. Boutiere, M. Perriguey, S. Demortiere, M. Fan, W. Qiu, B. Bu, Y. Xu, H. Yang, D. Huang, A.Y. Lau, J. Guo, J. Pelletier, and B. Audoin, Under the aegis of OFSEP M.-N. Zhang, X. Zhang, C.-S. Yang, J. Chen, P. Zheng, Q. Liu, C. Zhang, and F.-D. Shi Open Access Open Access e827 Genetic determinants of the humoral immune e789 COVID-19-related acute necrotizing encephalopathy response in MS with brain stem involvement in a patient with C. Gasperi, T.F.M. Andlauer, A. Keating, B. Knier, A. Klein, V. Pernpeintner, P. Lichtner, R. Gold, F. Zipp, F. Then Bergh, aplastic anemia M. Stangel, H. Tumani, B. Wildemann, H. Wiendl, A. Bayas, L. Dixon, J. Varley, A. Gontsarova, D. Mallon, F. Tona, D. Muir, T. Kumpfel,¨ U.K. Zettl, R.A. Linker, U. Ziemann, M. Knop, C. Warnke, A. Luqmani, I.H. Jenkins, R. Nicholas, B. Jones, and A. Everitt M.A. Friese, F. Paul, B. Tackenberg, A. Berthele, and B. Hemmer Open Access Open Access

Continued

Copyright ª 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited. TABLE OF CONTENTS Volume 7, Number 5, September 2020 Neurology.org/NN

e829 Motor cortex transcriptome reveals microglial key e854 Regional microglial activation in the substantia nigra events in amyotrophic lateral sclerosis is linked with fatigue in MS O. Dols-Icardo, V. Montal, S. Sirisi, G. Lopez-Pernas,´ T. Singhal, S. Cicero, H. Pan, K. Carter, S. Dubey, R. Chu, B. Glanz, L. Cervera-Carles, M. Querol-Vilaseca, L. Muñoz, O. Belbin, S. Hurwitz, S. Tauhid, M.-A. Park, M. Kijewski, E. Stern, R. Bakshi, D. Alcolea, L. Molina-Porcel, J. Pegueroles, J. Turon-Sans,´ R. Blesa, D. Silbersweig, and H.L. Weiner A. Lleo,´ J. Fortea, R. Rojas-Garc´ıa, and J. Clarimon´ Open Access Open Access e856 Long-term prognostic value of longitudinal κ e831 High free light chain is a potential biomarker for measurements of blood neurofilament levels double seronegative and ocular myasthenia gravis D.A. Häring, H. Kropshofer, L. Kappos, J.A. Cohen, A. Shah, A. Wilf-Yarkoni, Y. Alkalay, T. Brenner, and A. Karni R. Meinert, D. Leppert, D. Tomic, and J. Kuhle Open Access Class of Evidence Open Access Class of Evidence e835 COVID-19 outcomes in MS: Observational study of e860 Validation of the NEOS score in Chinese patients with early experience from NYU Multiple Sclerosis anti-NMDAR encephalitis Comprehensive Care Center Y. Peng, F. Dai, L. Liu, W. Chen, H. Yan, A. Liu, X. Zhang, X. Wang, J. He, Y. Li, C. Li, L. Chen, Y. Zhao, L. Li, Q. Ma, and J. Wang E. Parrotta, I. Kister, L. Charvet, C. Sammarco, V. Saha, R.E. Charlson, J. Howard, J.M. Gutman, M. Gottesman, N. Abou-Fayssal, R. Wolintz, Open Access Class of Evidence M. Keilson, C. Fernandez-Carbonell, L.B. Krupp, and L. Zhovtis Ryerson e862 Open Access Progressive multifocal leukoencephalopathy and sarcoidosis under interleukin 7: The price of healing e837 Treatment and outcome of aquaporin-4 A. Guffroy, M. Solis, V. Gies, Y. Dieudonne, C. Kuhnert, antibody–positive NMOSD: A multinational C. Lenormand, L. Kremer, A. Molitor, R. Carapito, Y. Hansmann, pediatric study V. Poindron, T. Martin, S. Hirschi, and A.-S. Korganow Open Access R.B. Paolilo, Y. Hacohen, E. Yazbeck, T. Armangue, A. Bruijstens, C. Lechner, S.L. Apostolos-Pereira, Y. Martynenko, M. Breu, C. de Medeiros Rimkus, E. Wassmer, M. Baumann, L. Papetti, Clinical/Scientific Notes M. Capobianco, B. Kornek, K. Rostásy, J.A. da Paz, O. Ciccarelli, M. Lim, A. Saiz, R. Neuteboom, R. Marignier, C. Hemingway, e783 COVID-19 in MS: Initial observations from the D.K. Sato, and K. Deiva fi Open Access Class of Evidence Paci c Northwest J.D. Bowen, J. Brink, T.R. Brown, E.B. Lucassen, K. Smoot, e839 Effect of the sphingosine-1-phosphate receptor modulator A. Wundes, and P. Repovic ozanimod on leukocyte subtypes in relapsing MS Open Access S. Harris, J.Q. Tran, H. Southworth, C.M. Spencer, B.A.C. Cree, and e785 S.S. Zamvil Guillain-Barré syndrome related to SARS-CoV-2 Open Access infection K. Bigaut, M. Mallaret, S. Baloglu, B. Nemoz, P. Morand, F. Baicry, e845 Tailoring B cell depletion therapy in MS according to A. Godon, P. Voulleminot, L. Kremer, J.-B. Chanson, and J. de Seze memory B cell monitoring Open Access G. Novi, F. Bovis, S. Fabbri, F. Tazza, P. Gazzola, I. Maietta, D. Curro,` e791 N. Bruschi, L. Roccatagliata, G. Boffa, C. Lapucci, G. Pesce, A case of cerebral vasculitis due to M. Cellerino, C. Solaro, A. Laroni, E. Capello, G. Mancardi, neurobartonellosis M. Sormani, M. Inglese, and A. Uccelli M. Poursheykhi, F. Mithani, T. Garg, C. Cajavilca, S. Jaijakul, S. Fung, Open Access Class of Evidence R. Klucznik, and R. Gadhia Open Access e847 Infliximab treatment in pathology-confirmed neurosarcoidosis e793 The rare case of a 20-year-old male with rapidly D. Fritz, W.M.C. Timmermans, J.A.M. van Laar, P.M. van Hagen, progressive primary angiitis of the CNS with T.A.M. Siepman, D. van de Beek, and M.C. Brouwer a good outcome Open Access Class of Evidence E. Graham, T. Shoemaker, D. Stefoski, M. Kontzialis, A. Naumaan, and R.K. Garg e849 Clinically based score predicting cryptogenic NORSE at the early stage of status epilepticus Open Access A. Yanagida, N. Kanazawa, J. Kaneko, A. Kaneko, R. Iwase, H. Suga, e797 Acute disseminated encephalomyelitis after Y. Nonoda, Y. Onozawa, E. Kitamura, K. Nishiyama, and T. Iizuka SARS-CoV-2 infection Open Access G. Novi, T. Rossi, E. Pedemonte, L. Saitta, C. Rolla, L. Roccatagliata, M. Inglese, and D. Farinini e852 High-throughput investigation of molecular and cellular biomarkers in NMOSD Open Access S.S. Yandamuri, R. Jiang, A. Sharma, E. Cotzomi, C. Zografou, e799 Severe paradoxical disease activation following A.K. Ma, J.S. Alvey, L.J. Cook, T.J. Smith, M.R. Yeaman, and K.C. O’Connor, on behalf of the Guthy-Jackson Charitable alemtuzumab treatment for multiple sclerosis Foundation CIRCLES Study Group J. Brannigan, J.L. Jones, and S.R.L. Stacpoole Open Access Open Access

Copyright ª 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited. TABLE OF CONTENTS Volume 7, Number 5, September 2020 Neurology.org/NN

e801 Acute necrotizing encephalopathy and myocarditis in e823 COVID-19-associated ophthalmoparesis and a young patient with COVID-19 hypothalamic involvement A. Elkady and A.A. Rabinstein E. Pascual-Goñi, J. Fortea, A. Mart´ınez-Domeño, N. Rabella, M. Tecame, C. Gomez-Oliva,´ L. Querol, and B. Gomez-Ans´ on´ Open Access Open Access e803 COVID-19-associated acute necrotizing myelitis e833 J. Sotoca and Y. Rodr´ıguez-Alvarez´ High proportion of Guillain-Barré syndrome Open Access associated with chikungunya in Northeast Brazil A.M.B. Matos, F.M. Maia Carvalho, D.L. Malta, C.L. Rodrigues, e811 Should interferons take front stage as an essential MS A.C. Félix, C.S. Pannuti, A.D.R. Lima, D.L.A. Esposito,´ L.M.B. dos Santos, F. von Glehn, J.K.B. Colares, B.A.L. da Fonseca, disease-modifying therapy in the era of coronavirus A.C.P. de Oliveira, and C.M. Romano disease 2019? Open Access C. Maguire, T. Frohman, S.S. Zamvil, E. Frohman, and E. Melamed Open Access Views & Reviews e813 CNS inflammatory vasculopathy with antimyelin oligodendrocyte glycoprotein antibodies in COVID-19 e841 Interleukin-6 in neuromyelitis optica spectrum A.A. Pinto, L.S. Carroll, V. Nar, A. Varatharaj, and I. Galea disorder pathophysiology Open Access Kazuo Fujihara, Jeffrey L. Bennett, Jerome de Seze, Masayuki Haramura, Ingo Kleiter, Brian G. Weinshenker, e821 Increased CSF levels of IL-1β, IL-6, and ACE Delene Kang, Tabasum Mughal, and Takashi Yamamura in SARS-CoV-2–associated encephalitis Open Access M. Bodro, Y. Compta, L. Llanso,´ D. Esteller, A. Doncel-Moriano, A. Mesa, A. Rodr´ıguez, J. Sarto, E. Mart´ınez-Hernandez, A. Vlagea, N. Egri, X. Filella, M. Morales-Ruiz, J. Yague,¨ A.´ Soriano, F. Graus, and Correction F. Garc´ıa, on behalf of the “Hospital Cl´ınic Infecto-COVID-19” and “Hospital Cl´ınic Neuro-COVID-19” groups e850 Guillain-Barré syndrome related to SARS-CoV-2 Open Access infection

Copyright ª 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited. ’ Volume 7, Number 5, September 2020 Editor sCorner Neurology.org/NN

Josep Dalmau, MD, PhD, FAAN, Editor, Neurology® Neuroimmunology & Neuroinﬂammation N2 in the time of COVID-19

Neurol Neuroimmunol Neuroinﬂamm September 2020 vol. 7 no. 5 e858. doi:10.1212/NXI.0000000000000858

This issue of Neurology® Neuroimmunology & Neuroinflammation reflects the impact of the COVID-19 pandemic with 15 studies related to inflammatory or autoimmune neurologic complications in patients with this disease, including Guillain-Barré syndrome, CNS inflammatory disorders, and several studies addressing the risk of COVID-19 in patients with neuromyelitis optica spectrum disorders (NMOSD) or multiple sclerosis (MS). Among the latter, a study conducted in New York by Parotta et al.1 included 72 patients with MS (55 relapsing and 17 progressive) or related disorders (4 patients) and confirmed (37 patients) or suspected COVID- 19 illness. Most patients did not require hospitalization despite being on disease-modifying therapies; factors associated with critical illness were similar to those of the general at-risk patient population. On the other hand, most of the articles related to CNS disorders that occurred in association with COVID-19 are single cases describing acute necrotizing encephalitis or myelitis; acute disseminated encephalomyelitis; transient encephalitis with increased levels of interleukin 1 (IL1), IL6, and angiotensin-converting enzyme; and 2 cases of inflammatory vasculopathy (one of them with newly identified oligodendrocyte glycoprotein antibodies). Unrelated to the topic of the neurologic complications of COVID-19, the current issue of N2 contains multiple interesting studies, among which I selected 3 for additional comments.

In a retrospective study of pediatric patients with aquaporin 4 (AQP4) antibody-positive NMOSD performed in a center in Brazil and 13 European centers, Barbosa Paolilo et al.2 described the clinical phenotypes, treatment response, and outcome of 67 children. After a median follow-up of 4 years, 58% of patients had permanent disability, and those with non-White ethnicity had a worse expanded disability status scale (EDSS); these patients also had a shorter time from disease onset to first relapse. Overall, the most frequent phenotypes at disease onset were optic neuritis (45% bilateral), transverse myelitis, and isolated area postrema syndrome. CSF pleocytosis was present in 71% of patients. During follow-up, 297 attacks were assessed. Optic neuritis at disease onset was associated with poor visual outcome, and younger age at disease onset was associated with cognitive impairment. Several different treatments were used (azathioprine, mycophenolate mofetil, or rituximab), and all were associated with a reduction of annualized relapse rate. None of the children treated with rituximab as first-line immunotherapy had relapses. This study shows that AQP4-antibody NMOSD in children is an aggressive disease with permanent disability occurring in over 50% of patients. The study also shows that the approach to treatment varies according to investigators, centers, and countries, leading the authors to emphasize the need for international consensus on treatment protocols to minimize the heterogeneity of treatment regimens used worldwide.

In another study, Fujii et al.3 investigated whether patients with idiopathic painful trigeminal neuropathy (IPTN) had antibodies against Plexin D1 and whether the antibodies were able to bind trigeminal ganglion neurons. The rationale for this study is based in previous work of the same authors showing that about 10% of patients with neuropathic pain had antibodies against

From the ICREA-IDIBAPS Hospital Cl´ınic, University of Barcelona, Spain; and Department of Neurology, University of Pennsylvania, Philadelphia.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Plexin D1, which is a member of a family of transmembrane sample size and short clinical follow-up (1 year); however, the protein receptors that bind semaphorins as ligands.4 Plexin D1 study identified significant differences and tendencies that was found expressed in endothelial cells, immunocytes, muscle deserve confirmation in future investigations. Larger sample cells, and small dorsal root ganglia (DRG) neurons. Patients sizes and SIMOA technology, preferably in serum, are also with neuropathic pain and Plexin D1 antibodies frequently needed to determine the role of NfL levels in patients with developed burning pain and thermal hyperalgesia along with PPMS. sensory impairment, suggesting involvement of C-fiber type pain nerves and DRG neurons.4 In the current study, the au- In addition to these studies, this September issue of N2 con- thors investigated 21 consecutive patients with IPTN and 35 tains an excellent review on COVID-19 and Guillain-Barré controls with neuropathic pain. Three (14%) of the 21 patients syndrome by Dalakas6 and other interesting studies that I am had Plexin D1 antibodies; in addition to facial, perioral, and sure will catch your attention. tongue pain or numbness, these patients also had limb or trunk neuropathic pain. The presence of tongue pain was more fre- Study funding quent than in Plexin D1 antibody-negative IPTN patients. No targeted funding reported. Two of the patients received immunotherapy and one improved; the patient who did not improve had been symp- Disclosure tomatic for 15 years, whereas the one who improved had J. Dalmau holds patents for the use of Ma2, NMDAR, symptoms for only a few months. Overall, this study shows that GABABR, GABAAR, DPPX, and IgLON5 as autoantibody small subsets of patients with neuropathic pain or IPTN have tests and receives royalties from the use of these tests. He is Plexin D1 antibodies and that some of these patients can po- editor of Neurology: Neuroimmunology & Neuroinflammation. tentially improve with immunotherapy. Future studies with a Go to Neurology.org/NN for full disclosures. larger number of patients are needed to better delineate the phenotype of these autoimmune neuropathic pain syndromes. References 5 1. Parotta E, Kister I, Charvet L, et al. COVID-19 outcomes in MS: observational In another study, Pérez-Miralles et al. investigated the role of study of early experience from NYU Multiple Sclerosis Comprehensive Care CSF chitinase 3-like-1 (CHI3L1), CHI3L2, and neurofilament Center. Neurol Neuroimmunol Neuroinflamm 2020;7:e835. doi: 10.1212/NXI. 0000000000000835. light chain (NfL) in predicting the course of primary pro- 2. Barbosa Paolilo R, Hacohen Y, Yazbeck E, et al. Treatment and outcome of aquaporin- gressive MS (PPMS). Twenty-five patients with disease du- 4 antibody–positive NMOSD: a multinational pediatric study. Neurol Neuro- ≤ ≥ immunol Neuroinflamm 2020;7:e837. doi: 10.1212/NXI.0000000000000837. ration 10 years and without disease-modifying therapy for 6 3. Fujii T, Yamasaki R, Miyachi Y, et al. Painful trigeminal neuropathy associated with months were included in the study. The authors found that anti-Plexin D1 antibody. Neurol Neuroimmunol Neuroinflamm 2020;7:e819. doi: 10. 1212/NXI.0000000000000819. increasing CHI3L1 levels associated with higher EDSS scores 4. Fujii T, Yamasaki R, Iinuma K, et al. A novel autoantibody against Plexin D1 in at baseline and at month 12 and tended to be associated with a patients with neuropathic pain. Ann Neurol 2018;84:208–224. 5. Pérez-Miralles F, Prefasi D, Garcia-Merino A. CSF chitinase 3-like-1 association with higher risk of disability progression. By contrast, increasing disability of primary progressive MS. Neurol Neuroimmunol Neuroinflamm 2020;7: CHI3L2 levels tended to correlate with lower baseline EDSS e815. doi: 10.1212/NXI.0000000000000815. 6. Dalakas MC. Guillain-Barré syndrome: the first documented COVID-19–triggered scores, and no correlation was found with NfL levels. The autoimmune neurologic disease: more to come with myositis in the offing. Neurol authors acknowledge several limitations including the small Neuroimmunol Neuroinflamm 2020;7:e781. doi: 10.1212/NXI.0000000000000781.

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN EDITORIAL OPEN ACCESS Isotyping paranodal antibodies in inﬂammatory neuropathies One step closer to precision care

Cinta Lleix`a, BSc, and Luis Querol, MD, PhD Correspondence Dr. Querol Neurol Neuroimmunol Neuroinﬂamm 2020;7:e843. doi:10.1212/NXI.0000000000000843 [email protected]

The discovery of autoantibodies against paranodal proteins such as neurofascin-155 (NF155), RELATED ARTICLE contactin-1 (CNTN1), or contactin-associated protein-1 (CASPR1) in inflammatory neuropathies has led to the description of subsets of patients with specific phenotypic features.1 Antiparanodal antibodies and IgG subclasses in acute These antibodies mostly belong to the immunoglobulin G (IgG)4 subclass, but IgG1, IgG2, or – autoimmune neuropathy IgG3 autoantibodies have also been described.2 4 Different autoantibody isotypes are associated with different effector mechanisms causing nerve damage5,6; testing them may have Page e817 implications to inform therapeutic choices or to predict prognosis.

In this issue of Neurology: Neuroimmunology & Neuroinﬂammation, an article entitled Anti- paranodal antibodies and IgG subclasses in acute autoimmune neuropathy7 by Luise Appeltshauser et al. provides evidence that IgG2/3 antibodies can be found in acute onset inﬂammatory neuropathies associated with paranodal antibodies and suggests a correlation between the clinical features and therapeutic response and the IgG isotypes associated to the disease.

The authors screened for anti-CNTN1 and anti-CASPR1 antibodies in a cohort of patients with Guillain-Barré syndrome (GBS) and acute chronic inflammatory demyelinating polyradiculoneuropathy (A-CIDP), including follow-up autoantibody testing in some patients. The frequency of the detected antibodies (around 4% of GBS/A-CIDP patients) confirms data from previous reports published by this and other groups. They found 5 patients with IgG2 or IgG3 autoantibodies against CNTN1, CASPR1, or both proteins, and 1 patient with IgG4 autoantibodies against CNTN1 and suggest that IgG3 antibodies may associate with a better response to IVIg. A recent study in CIDP does not find a relationship of IVIg efficacy and terminal complement inhibition in CIDP without paranodal antibodies.8 However, the authors of this study have previously reported that complement deposition mediated by IgG3 autoantibodies targeting paranodal proteins can be modulated by IVIg.6 Thus, the mechanisms through which IVIg exert their effect in CIDP may differ depending on the subtype of CIDP and associated autoantibodies. Other mechanisms may explain the poorer response to IVIg in diseases caused by autoantibodies of the IgG4 isotype. For example, IgG4-producing plasma cells have been reported to have regulatory phenotypes (IL10+) and lower expression levels of the inhibitory immunoglobulin receptor FcGRIIb.9 Further studies should clarify the underlying mechanisms explaining this differential response to IVIg of autoimmune diseases depending on the autoantibody isotype; apart from the role of complement inhibition induced by IVIg proposed by the authors, it seems reasonable to assume that the antibody-producing cells that produce antibodies that have antagonistic functions (i.e., proinflammatory IgG1-3 vs anti-inflammatory IgG4) may respond differently to the immunomodulatory effects of IVIg.

The authors also describe an interesting patient with A-CIDP in which the autoantibody subclass switches from IgG3 in the acute phase of the disease to IgG4 in the chronic stage; the target of the autoantibody also shifts over time from CNTN1 and CASPR1 to CASPR1 alone. This simultaneous change in the antigenic target and the isotype, which should be conﬁrmed in

From the Neuromuscular Diseases Unit (C.L., L.Q.), Neurology Department, Hospital de la Santa Creu I Sant Pau, Barcelona; and Centro para la Investigacion´ Biom´edica en Red en Enfermedades Raras (C.L., L.Q.), CIBERER, Madrid, Spain.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work, provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 other patients, may be explained by 2 related facts: first, IgG4 true antigenic target and assess the diagnostic accuracy of each antibodies appear late in the immune response, after several test in all nodo/paranodal autoantibodies. rounds of affinity maturation and somatic hypermutation have occurred.10 Second, this fine tuning of the affinity may drive In conclusion, detection of antibodies against paranodal the autoantibody response toward the CASPR1 epitope. proteins, followed by IgG isotype testing in seropositive cases Whether this epitope is a immunodominant epitope driving and their longitudinal monitoring during disease course the affinity selection in all CIDP patients in which an anti- should be considered in the diagnostic workup in in- CASPR1 IgG4 response is detected, as it happens in other flammatory neuropathies to improve pathophysiologic diseases such as anti-MusK myasthenia gravis,11 and whether knowledge, diagnostic accuracy, and treatment selection. this phenomenon occurs in other CIDP or GBS patients with anti-CASRP1 antibodies, remains to be elucidated. In most of Study funding the studies regarding paranodal antibodies in CIDP, samples No targeted funding reported. were acquired in the chronic phase of the disease, and there fi are few longitudinal studies in the eld. Therefore, data on the Disclosure appearance and features of paranodal autoantibodies in the The authors report no disclosures relevant to the manuscript. acute phase of CIDP are scarce. For this reason, although Go to Neurology.org/NN for full disclosures. larger prospective studies are still needed, the association of IgG isotypes and disease progression and treatment response References described in this study could be important for optimal patient 1. Pascual-Goñi E, Mart´ın-Aguilar L, Querol L. Autoantibodies in chronic inflammatory demyelinating polyradiculoneuropathy. Curr Opin Neurol 2019;32:651–657. care in each moment of the disease. 2. Doppler K, Appeltshauser L, Villmann C, et al. Auto-antibodies to contactin- associated protein 1 (Caspr) in two patients with painful inflammatory neuropathy. The report also describes 2 patients with antibodies against Brain 2016;139, 2617–2630. 3. Stengel H, Vural A, Brunder A-M, et al. Anti-pan-neurofascin IgG3 as a marker of both CNTN1 and CASPR1 proteins in the acute phase of the fulminant autoimmune neuropathy. Neurol Neuroimmunol Neuroinflamm 2019;6: disease. There is uncertainty in the field as to whether the e603. doi:10.1212/NXI.0000000000000603. 4. Doppler K, Schuster Y, Appeltshauser L, et al. Anti-CNTN1 IgG3 induces acute immune response in patients with antibodies against the conduction block and motor deficits in a passive transfer rat model. CNTN1/CASPR1 complex is targeting an epitope arising J Neuroinflammation 2019;16:73–13. 5. Manso C, Querol L, Mekaouche M, Illa I, Devaux JJ. Contactin-1 IgG4 antibodies cause from the binding of both CNTN1 and CASPR1, or separately paranode dismantling and conduction defects. Brain 2016;139(Pt 6):1700–1712. against each of CNTN1 and CASPR1 proteins. Authors 6. Appeltshauser L, Weishaupt A, Sommer C, Doppler K. Complement deposition induced by binding of anti-contactin-1 auto-antibodies is modified by immunoglob- provide descriptive data suggesting that the latter may be ulins. Exp Neurol 2017;287:84–90. happening. Considering the changes that authors report in the 7. Appeltshauser L, Brunder A-M, Heinius A, et al. Antiparanodal antibodies and IgG fi subclasses in acute autoimmune neuropathy. Neurol Neuroimmunol Neuroinflamm antigenic speci city over time in some patients, it could well 2020;7:e817. doi:10.1212/NXI.0000000000000817. be that this uncertainty can be clarified by studying longitu- 8. Keller CW, Quast I, Dalakas MC, Lünemann JD. IVIG efficacy in CIDP patients is not associated with terminal complement inhibition. J Neuroimmunol 2019;330:23–27. dinally (as the authors did in this report) the IgG isotypes in 9. Van De Veen W, Stanic B, Yaman G, et al. IgG4 production is confined to human IL- all patients and analyzing epitope changes over time in pa- 10-producing regulatory B cells that suppress antigen-specific immune responses. fi J Allergy Clin Immunol 2013;131:1204–1212. tients previously classi ed as having antibodies only against 10. Vidarsson G, Dekkers G, Rispens T. IgG subclasses and allotypes: from structure to the CNTN1/CASPR1 complex. Because technical issues can effector functions. Front Immunol 2014;5:1–17. 11. Huijbers MG, Vink AF, Niks EH, et al. Longitudinal epitope mapping in MuSK also explain some of these discrepancies, collaborative, myasthenia gravis: implications for disease severity. J Neuroimmunol 2016;291: interlaboratory validation studies are needed to elucidate the 82–88.

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN ARTICLE OPEN ACCESS Guillain-Barré syndrome: The first documented COVID-19–triggered autoimmune neurologic disease More to come with myositis in the offing

Marinos C. Dalakas, MD Correspondence Dr. Dalakas Neurol Neuroimmunol Neuroinﬂamm 2020;7:e781. doi:10.1212/NXI.0000000000000781 [email protected]

Abstract MORE ONLINE Objective COVID-19 Resources To present the COVID-19–associated GBS, the prototypic viral-triggered autoimmune disease, For the latest articles, in the context of other emerging COVID-19–triggered autoimmunities, and discuss potential invited commentaries, and concerns with ongoing neuroimmunotherapies. blogs from physicians around the world Methods NPub.org/COVID19 Eleven GBS cases in four key COVID-19 hotspots are discussed regarding presenting symptoms, response to therapies and cross-reactivity of COVID spike proteins with nerve glycolipids. Emerging cases of COVID-19–triggered autoimmune necrotizing myositis (NAM) and encephalopathies are also reviewed in the context of viral invasion, autoimmunity and ongoing immunotherapies.

Results Collective data indicate that in this pandemic any patient presenting with an acute paralytic disease-like GBS, encephalomyelitis or myositis-even without systemic symptoms, may represent the first manifestation of COVID-19. Anosmia, ageusia, other cranial neuropathies and lymphocytopenia are red flags enhancing early diagnostic suspicion. In Miller-Fisher Syndrome, ganglioside antibodies against GD1b, instead of QG1b, were found; because the COVID-19 spike protein also binds to sialic acid-containing glycoproteins for cell-entry and anti-GD1b antibodies typically cause ataxic neuropathy, cross-reactivity between COVID-19–bearing gangliosides and peripheral nerve glycolipids was addressed. Elevated Creatine Kinase (>10,000) is reported in 10% of COVID-19–infected patients; two such patients presented with painful muscle weakness responding to IVIg indicating that COVID-19–triggered NAM is an overlooked entity. Cases of acute necrotizing brainstem encephalitis, cranial neuropathies with leptomeningeal enhancement, and tumefactive postgadolinium-enhanced demyelinating lesions are now emerging with the need to explore neuroinvasion and autoimmunity. Concerns for modifications-if any-of chronic immunotherapies with steroids, mycophenolate, azathioprine, IVIg, and anti-B-cell agents were addressed; the role of complement in innate immunity to viral responses and anti-complement therapeutics (i.e. eculizumab) were reviewed.

Conclusions Emerging data indicate that COVID-19 can trigger not only GBS but other autoimmune neurological diseases necessitating vigilance for early diagnosis and therapy initiation. Although COVID-19 infection, like most other viruses, can potentially worsen patients with pre-existing autoimmunity, there is no evidence that patients with autoimmune neurological diseases stable on common immunotherapies are facing increased risks of infection.

From the Department of Neurology, Thomas Jefferson University, Philadelphia, PA, and the Neuroimmunology Unit, National and Kapodistrian University of Athens Medical School, Greece.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

The Article Processing Charge was funded by the author. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary ACE-2 = angiotensin-converting enzyme 2; AIDP = acute inﬂammatory demyelinating polyneuropathy; AMAN = acute motor axonal neuropathy; CK = creatine kinase; CRP = C-reactive protein; GBS = Guillain-Barr´e syndrome; ICU = intensive care unit; IVIg = intravenous immunoglobulin; MERS = Middle East respiratory syndrome; MFS = Miller Fisher syndrome; NAM = necrotizing autoimmune myositis.

– Guillain-Barré syndromes (GBSs) comprise a spectrum of Italy, Spain, and now France.7 11 The cases are of special polyneuropathies characterized by acute (within 1–4 weeks) neuroimmunologic and practical interest while highlight what ascending motor weakness, mild or moderate sensory ab- is more to come. normalities, occasional cranial nerve involvement, and muscle or radicular pain. According to the degree of involvement of The very first case—and the only one from Wuhan—refers to the motor or sensory nerves, myelin sheath, axon, or cranial a woman who 4 days after returning to Shanghai from Wuhan nerve predominance, the most common subtypes are the presented with acute lower extremity weakness and areflexia acute inflammatory demyelinating polyneuropathy (AIDP), that progressed over 3 days to the arms without any systemic acute motor axonal neuropathy (AMAN), and the Miller symptoms.7 When first seen, GBS was suspected and con- Fisher syndrome (MFS) characterized by acute oph- firmed by elevated CSF protein and electromyographic thalmoplegia, gait ataxia, and areflexia.1 GBS is one of the features of demyelinating neuropathy. Her laboratory test prototypic viral-triggered autoimmune neurologic diseases as results were normal, except of lymphocytopenia and approximately 70% of the patients have a preceding by 1–3 thrombocytopenia. She was started on intravenous immu- – weeks, flu-like, viral illness.1 3 Among the infectious agents noglobulin (IVIg) but 4 days later developed fever, cough, associated with triggering sporadic GBS are viruses, including and pneumonia and tested COVID-19 positive. After 3 influenza, enteroviruses, cytomegalovirus, Epstein–Barr virus, weeks, her strength normalized and lymphocytopenia re- herpes simplex virus, hepatitis, or HIV, and bacteria, such as solved. While at the regular neurology ward, however, 2 of Campylobacter jejuni, Mycoplasma pneumoniae, and Haemo- her relatives taking care of her tested COVID-19 positive philus influenzae.1 GBS outbreaks have been also associated with pneumonia, while 2 of her neurologists and 6 nurses with viral epidemics or pandemics, including H1N1, swine flu treating her GBS were put on isolation. A/New Jersey influenza strain, arthropod-borne flaviviruses, such as the West Nile virus, chikungunya, or Zika, and with Italy has just reported 6 patients with GBS,8 1 in this issue.9 All coronaviruses, including the Middle East respiratory syn- presented with subacute onset of upper and lower extremity – drome (MERS)-CoV and SARS-CoV.1 3 weakness, distal paresthesias, and sensory deficits 3–10 days after experiencing cough, anosmia, ageusia, and sore throat. We are now more than 4 months into the peak of the COVID- On admission, all had lymphocytopenia and high C-reactive 19 pandemic with >5,000,000 infections and >330,000 deaths protein (CRP) that led to diagnosis of COVID-19. One pa- steadily increasingly worldwide, and although various re- tient had facial diplegia and sensory ataxia.8 The CSF showed spiratory and cardiac complications have been reported, we elevated protein concentration but was negative for COVID- have not yet seen COVID-19–related neuroinflammatory or 19. Electrophysiology was consistent with demyelinating or neuroautoimmune diseases as with the other viral outbreaks, axonal GBS. MRI showed enhancement of the caudal nerve including with coronaviruses MERS-CoV and SARS-CoV, roots or facial nerve. Symptoms rapidly progressed to tetra- which have 75–80% identical viral genome sequence with plegia requiring mechanical ventilation. Antiganglioside COVID-19.4 Whether this is due to COVID-19–inducing antibodies in 3/6 tested patients were negative. All received severe respiratory compromise soon after the median 4-day IVIg with variable recovery; 1 died. incubation period5 and underrecognition of neurologic events owing to overwhelming urgency to focus on life-saving efforts The third series refers to 2 men from Spain who, 3–5 days is unclear. In the most up-to-date large published series, apart after experiencing low-grade fever, malaise, anosmia, and from multifactorial acute cerebrovascular events in 5.7% of ageusia, developed MFS or polyneuritis cranialis.10 The pa- patients, the main COVID-19–related neurologic symptoms tient with MFS presented with oculomotor nerve palsies, have been hypogeusia (in 5.6%), hyposmia (5.1%), and very diplopia, perioral paresthesias, areflexia, ataxia, and elevated high creatine kinase (CK) levels, with myalgia (in 19.3%) CSF protein concentration and the other with diplopia, bi- indicating potential CNS, peripheral nervous system, and lateral abducens nerve palsy, and anosmia. Both had lym- myopathic manifestations.6 Things are however rapidly phopenia and tested positive for COVID-19; their SCF was changing as just only the last 4 weeks the first COVID-19– COVID-19 negative. The patient with MFS was positive for triggered neurologic events are reported with at least 11 cases GD1b ganglioside antibodies and improved with IVIg. In of GBS, the prototypic viral-triggered autoimmune neurologic both, the neurologic features completely resolved, except for disease, observed in the 4 main COVID-19 hotspots, Wuhan, residual anosmia and ageusia. Last, 2 cases from France, a 43-

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN year-old man and a 70-year-old woman, also presented with suggests neurotropism targeting olfactory neurons. SARS-CoV acute anosmia, followed by rapidly progressive weakness, and MERS-CoV, the 2 coronaviruses similar to COVID-19, are paresthesias, ataxia, areflexia, and multiple cranial palsies neurovirulent and can enter the brain via olfactory nerves.13 In (including III, V, VI, VII, and VIII), MRI enhancement of mice, after oronasal infection with SARS-CoV, the virus not cranial nerves, roots, and plexus, high CSF protein concen- only infects epithelial cells of the respiratory tract but also the tration but negative for COVID-19, typical features of acute olfactory receptor neurons in the neuroepithelium gaining ac- demyelinating neuropathy by electrophysiology, and clinical cess to the olfactory bulb and brainstem.13 These viruses can improvement after IVIg.11 also enter the CNS via retrograde axonal transport through other cranial nerves, such as trigeminal, which possesses noci- ceptive neuronal receptors in the nasal cavity, the sensory fibers GBS in the COVID-19 pandemic of the glossopharyngeal, and via peripheral nerves.13 The present GBS series, where oculomotor, trigeminal, and MRI- These early GBS cases, despite incomplete immunologic enhanced facial and nerves roots were concurrently affected, ffi workup considering the enormous di culties in highly stressed strengthens (but not prove) this notion. Accordingly, it will not in-hospital settings, provide early clues on what to expect in the be unexpected in the weeks and months ahead to see other months ahead regarding the common acute autoimmune neu- COVID-19–infected patients with neurologic signs related to rologic conditions, such as GBS, polyneuritis cranialis, enceph- multiple cranial nerves, brainstem, and peripheral nerves with alitis, encephalomyelitis, or myositis, which for years we have MRI enhancement in nerves and meninges. been casually referring to as postviral if seen after febrile illnesses. First and foremost, the practicing neurologists in this pan- Autoimmunity of COVID-19–GBS: demic should now be aware that a patient who presents with fi an acute paralytic disease—like GBS, encephalomyelitis, or signi cance of sialic acids present in myositis—even without fever, dyspnea, or any systemic the coronaviruses and peripheral symptoms, may represent the first manifestation of COVID- 19. This is compelling, considering that only 43% of COVID- nerve myelin 4–6 19–positive patients on admission have fever, and many of In 7/11 tested patients with COVID-19–GBS, the virus was the present patients with GBS did not have any COVID-19 not detected in the CSF, implying no direct root infection or symptoms at presentation. These early GBS cases also high- intrathecal viral replication. The improvement of several light that 2 clinical and laboratory signs, anosmia/ageusia and patients with IVIg and the presence of GD1b antibodies in 1 lymphocytopenia/thrombocytopenia, are red flags in sus- tested patient suggest a postviral-triggered immune response – pecting COVID-19 in otherwise asymptomatic patients with similar to other postviral-induced GBS cases1 3 or other post- acute neurologic events. They further confirm what we had viral autoimmune neurologic disorders.14 ThereportedGD1b feared from other viral pandemics that COVID-19 can trigger ganglioside antibodies, however, although in contrast to GQ1b neurologic autoimmunity; in contrast to the other postviral antibodies typically seen in MFS, are of very special interest. neurologic diseases, however, COVID-19 requires high degree of suspicion as a potential hidden trigger to prevent The SARS-CoV 3a protein contains oligosaccharides with inadvertent viral transmission to health care personnel and direct evidence that sialic acids play a critical role in human 7 patient relatives, as in the Wuhan case. The series also coronavirus infection.15 It has been just shown that the at- highlights that GBS peaks 5–10 days after the first COVID-19 tachment of coronaviruses to the surface of respiratory cells is symptoms, which in intensive care unit (ICU) settings helps mediated by the spike (S) viral protein, which binds not only to distinguish GBS from critical illness neuropathy that usu- to the angiotensin-converting enzyme 2 (ACE-2) receptor for ally appears later in the course of very sick ICU patients. entry16 but also to sialic acid–containing glycoproteins and gangliosides on cell surfaces.15 Such a dual receptor/ attachment is proposed to be a reason for the increased Significance of anosmia/ageusia and transmissibility of COVID-19 compared with SARS-CoV that other cranial nerve palsies binds only to ACE-2 receptor.15,16 Of relevance to GBS is that various gangliosides, most commonly those containing either The early manifestation of anosmia and ageusia not only in the a disialosyl moiety, such as GD1b, GQ1b, and GT1b, or 2 present GBS series that often occurred in conjunction with gangliosides that share epitopes with GM2, or a combination other cranial neuropathies but also in large worldwide cohorts of GM2 and GM1, GM1 and GD1b, can serve as antigens in reporting sudden loss of smell and taste early in the infection in patients with neuropathies.17 When IgM recognizes the Gal up to 60% of COVID-19 carriers5,6,12 is highly informative (pl-3) GalNAc moiety of GM1, which is found on the surface about COVID-19 neurovirulence or even possible viral entry of motor neurons, there is clinically a motor neuropathy, but if into the brain. In contrast to commonly reversible anosmia recognizes epitopes containing disialosyl groups of GDlb, when the non-neural olfactory epithelial cells are virally infec- which is present on the dorsal root ganglionic neurons, there ted, the often persistent anosmia/ageusia after COVID-19 is sensory ataxic neuropathy.17 Immunization of rabbits with

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 GDlb also causes sensory ataxic neuropathy mimicking the chronic statin use as overstated, but autoimmune, paraneo- – human disease.18 Of interest, the first described patient with plastic, or postviral.22 25 Considering that very high CK level sensory ataxic neuropathy and GDlb antibodies had also and painful muscle weakness were seen in 10% of COVID- ophthalmoplegia,19 as seen in MFS and the present series.10 19–positive patients,6 a potentially treatable autoimmune my- As COVID-19 spike interacts with the GalNAc residue of opathy has been likely overlooked. Studies with muscle biopsy GM1 and ganglioside dimers for anchoring to cell surface and antibody screening are urgently needed in such patients gangliosides,15 cross-reactivity between epitopes within the because therapy with IVIg can improve strength, reduce mor- COVID-19 spike-bearing gangliosides and signature sugar bidity, and facilitate recovery, as reported in the first 2 cases.20,21 residues of surface peripheral nerve glycolipids is a very likely COVID-19–associated myositis is also highly interesting be- possibility. Such typical molecular mimicry has been shown cause ACE-2, the SARS-CoV receptor, is reportedly expressed between peripheral nerve glycolipids and Campylobacter jejuni in skeletal muscle.26 If this is confirmed, COVID-19 may rep- – or Zika virus that also trigger GBS.1 3 Accordingly, all GBS resent the first virus directly capable of infecting muscle fibers. subtypes (AIDP, AMAN, and MFS) can be expected with None of the viruses implicated in viral-triggered myositis has COVID-19, necessitating screening for ganglioside antibodies sofarbeenshowntoinfectmuscle27,28; instead, viruses in- to assess autoimmunity. An interesting therapeutic compo- duce an immune T cell with clonal expansion of viral-specific nent in this association is the emerging data that chloroquine, T cells or macrophage-mediated, muscle fiber autoinvasion an antimalarial drug under investigation for treating COVID- with abundant proinflammatory cytokines.22,29,30 19, binds with high-affinity sialic acids and GM1 gangliosides and, in the presence of chloroquine, the SARS-CoV viral spike cannot bind gangliosides to infect the targeted cells.15 If Other emerging postinfectious benefit is confirmed and safety is established, chloroquine autoimmune neurologic disorders may be of added therapeutic value in future patients with COVID-19–triggered GBS in conjunction with IVIg. Acute disseminated encephalomyelitis, as described for the other coronaviruses, or postinfectious brainstem encephalitis, myelitis, and radiculoneuropathies will not be unexpected. An What is more to come with myositis in atypical case of acute necrotizing encephalopathy, attributed to ffi immune-mediated process and proinflammatory cytokines, the o ng and another with meningoencephalitis and leptomeningeal 31 Among the other potential COVID-19–associated autoim- enhancement have been already noted. This picture is now mune diseases, the first alarming concern is inflammatory becoming more clear with 2 reports in this issue of the Journal, myopathy, especially necrotizing autoimmune myositis documented by impressive MRIs; one, a case of acute necrotizing brainstem encephalopathy, presented with epilepsy and (NAM) because very high CK levels >10,000 with myalgia and 32 weakness are now reported in more than 10% of COVID- changes in the right mesial temporal lobe and hippocampus 19–infected patients.6 Although COVID-19–associated my- and another with altered mental status and multiple white opathy has not yet been studied but only characterized as matter tumefactive lesions with postgadolinium enhancement 6 in periventricular areas and corpus callosum suggestive of de- skeletal muscle injury or rhabdomyolysis, 2 just published 33 cases suggest an autoimmune COVID-19–triggered NAM. myelination. Not underestimating the overwhelming burden ff One, an 88-year-old man from New York presented with acute of acute COVID-19 to medical sta ,vigilanceforthesedis- orders is needed along with screening for autoimmune en- bilateral thigh weakness and inability to get up from the toilet, 14 without fever or other systemic symptoms, and very high CK cephalitis autoantibodies because these cases can potentially level (13,581 U/L).20 He was found COVID-19 positive and respond to early initiation of immunotherapy, especially with ff given hydroxychloroquine, and a week later, his painful weak- IVIg, whenever indicated, which may additionally o er various ff ness improved with CK reduction. The other, a 60-year-old potentially protective antibodies and anticytokine e ects. man from Wuhan had a 6-day history of fever, cough, and COVID-19–positive pneumonia with normal strength and CK; fi 7 days later, although systemically had improved, his CRP Residual neurologic de cits in doubled and developed painful muscle weakness with very high COVID-19–recovered patients: need CK (11,842 U/L).21 He was given IVIg and his strength im- for a systematic study proved while became COVID-19 negative. No neurologic data are yet available about the discharged Myopathic symptoms in a severe systemic viral disease are patients who survived the catastrophic effects of the virus. We multifactorial, but an acute onset of severe muscle weakness know that several people have permanently lost smell and taste, with increased inflammatory markers and very high CK levels which is a form of disability affecting quality of life, depriving in the thousands, as described above, is consistent with auto- tasting pleasures, and the ability to detect danger signals, like immune inflammatory myopathy within the spectrum of NAM, smelling gas or fire. Many discharged patients require assistance similar to what was first reported with HIV early in that because of muscle weakness and gait unsteadiness, which is epidemic.22,23 The most common causes of NAM are not arguably multifactorial; some patients may have had critical

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN illness neuropathy and deconditioning with significant muscle on suppressing the innate immunity that facilitates infection atrophy worsened by comorbidities; others may have neuro- of macrophages and viral spread. Similar data from New toxicity or myocytotoxicity from antiviral therapies, like first York City area show that the incidence of hospitalization – described with antiretrovirals or chloroquine34 36; still others among patients with immune-mediated inflammatory dis- may have the residual effects from an unrecognized primary eases on therapy with steroids and biologic agents was myopathy, neuropathy, or myelopathy due to postviral auto- consistent with that among patients with COVID-19 in the immunity. A study exploring the patients’ current causes of general population, concluding that ongoing use of biologics residual muscle weakness and sensory deficits is urgently is not associated with worse COVID-19 outcomes.40 There needed using EMG, muscle or nerve biopsies, autoantibody is also new evidence suggesting possible beneficial effect of screening, and CSF or imaging studies to determine immediate anticomplement therapies.41 Complement is an integral or long-term recovery prospects, identify potential reversibility, component of the innate immune response to viruses and an and accelerate return to normalcy. instigator of proinflammatory responses with evidence that activation of C3 exacerbates SARS-CoV–associated acute respiratory distress syndrome and C3-C5 complement deposits Patients with autoimmune neurologic are abundant in the lung biopsies from patients with COVID- 19.41 It was proposed that complement inhibition may alleviate diseases: how is COVID-19 changing the inflammatory complications of COVID-19 infection lead- ongoing immunotherapies and the ing to ongoing trials with anti-C3 and anti-C5 agents.41,42 On role of complement this basis, eculizumab, an anti-C5 monoclonal antibody approved for neuromyelitis optica spectrum disorder and myas- Patients with autoimmune neuropathies, especially chronic thenia gravis with some benefits in patients with GBS, may not inflammatory demyelinating polyneuropathy and multifocal be withheld, if indicated, as, like IVIg, may theoretically have motor neuropathy, but also with other autoimmunities like added benefit. myasthenia gravis, MS, and inflammatory myopathies, have been justifiably concerned as to whether their disease status adds an additional risk placing them into an immunosup- Study funding pressed or immunocompromised category. There is no current No targeted funding reported. evidence that any of the aforementioned autoimmune disorder itself makes them more susceptible to COVID-19, but some Disclosure immunosuppressive or even immunomodulating therapies M. Dalakas is an associate editor for Neurology: Neuro- they are receiving may have this potential, although there are immunology & Neuroinflammation. Disclosures available: no validated data. Most autoimmune neuromuscular Neurology.org/NN. patients are maintained on steroids, mycophenolate, or azathioprine while most chronic inflammatory de- Publication history myelinating polyneuropathy receive monthly IVIg. The Received by Neurology: Neuroimmunology & Neuroinflammation same applies to patients with MS where it seems safe to start May 1, 2020. Accepted in final form May 4, 2020. or continue treatment with the standard disease-modifying drugs. If clinically stable and not lymphopenic, there are no References 1. Dalakas MC. Pathogenesis of immune-mediated neuropathies Biochimica et. Bio- compelling or data-driven reasons to change anything in phyiscal Acta 2015;1852:658–666. these patients and disturb clinical stability. For patients on 2. Wakerley BR, Yuki N. Infectious and noninfectious triggers in Guillain-Barre syn- monthly IVIg, there may be even a theoretical advantage that drome. Expert Rev Clin Immunol 2013;9:627–639. 3. Cao-Lormeau VM, Blake A, Mons S, et al. Guillain-Barre syndrome outbreak asso- IVIg offers additional protection due to natural autoanti- ciated with Zika virus infection in French Polynesia: a case-control study. Lancet bodies; if IVIg is not infused as home infusion, switching to 2016;387:1531–1539. 4. Zhu N, Zhang D, Wang W, et al. A novel coronavirus from patients with pneumonia in self-administered subcutaneous IgG might be an option to China, 2019. N Engl J Med 2020;382:727–733. diminish exposure, as has been proven effective.37 For 5. Guan WJ, Ni ZY, Hu Y, et al. Clinical characteristics of coronavirus disease 2019 in China. N Engl J Med 2020;382:1708–1720. patients on rituximab, the infusion intervals can be pro- 6. Mao L, Jin H,Wang M, et al. Neurologic manifestations of hospitalized patients with longed to more than 6 months because both B-cell reduction coronavirus disease 2019 in Wuhan, China. JAMA Neurol 2020. 10.1001/jamaneurol. fi 38 2020.1127. and clinical bene t can persist longer. New emerging data 7. Zhao H, Shen D, Zhou H, Liu J, Chen S. Guillain-Barré syndrome associated with provide credence and reassurance regarding these issues, SARS-CoV-2 infection: causality or coincidence? Lancet Neurol 2020;19: 383–384. especially on immunomodulating drugs. In a large number of 8. Toscano G, Palmerini F, Ravaglia S, et al. Guillain–Barré syndrome associated with patients from Wuhan, published in this issue of the Journal,39 SARS-CoV-2. New Engl J Med 2020. 10.1056/NEJMc2009191. 9. Alberti P, Beretta S, Piatti M et al. Guillain-Barré syndrome related to COVID-19 it was shown that altered immunity induced by disease- infection. Neurol Neuroimmunol Neuroinflamm 2020;7:e741. doi: 10.1212/ modifying drugs in patients with MS or neuromyelitis optica NXI.0000000000000741. ffi 10. Gutiérrez-Ortiz C, Méndez A, Rodrigo-Rey S, et al. Miller Fisher Syndrome and spectrum disorder appears insu cient to enhance suscepti- polyneuritis cranialis in COVID-19. Neurology Epub 2020 Apr 17. doi: 10.1212/ bility to COVID-19 infection. The results are important but WNL.0000000000009619. 11. Bigaut K, Mallaret M, Baloglu S, et al. Guillain-Barré syndrome related to SARS-CoV- not unexpected considering that most of these therapies 2 infection. Neurol Neuroimmunol Neuroinflamm 2020;7:e785. doi: 10.1212/ target the adaptive immune response with little, if any, effect NXI.0000000000000785.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 5 12. Yan CH, Faraji F, Ptajapati D, DeConde AS. Association of chemosensory dysfunc- 29. Illa I, Nath A, Dalakas MC. Immunocytochemical and virological characteristics of tion and covid-19 in patients presenting with influenza-like symptoms. Int Forum HIV-associated inflammatory myopathies: similarities with seronegative polymyositis. Allergy Rhinology Epub 2020 Apr 12. doi: 10.1002/alr.22579. Ann Neurol 1991;29:474–481. 13. Desforges M, Coupanec A, Dubeau P, et al. Human coronaviruses and other respiratory viruses: 30. Dalakas MC, Rakocevic G, Shatunov A, et al. Inclusion body myositis with human underestimated opportunistic pathogens of the central nervous system. Viruses 2019;12:14. immunodeficiency virus infection: four cases with clonal expansion of viral-specific 14. Armangue T, Spatola M, Vlagea A, et al. Frequency, symptoms, risk factors, and T cells. Ann Neurol 2007;61:466–475. outcomes of autoimmune encephalitis after herpes simplex encephalitis: a prospective 31. Moriguchi T, Harii N, Goto J, et al. A first Case of Meningitis/Encephalitis associated observational study and retrospective analysis. Lancet Neurol 2018;17:760–772. with SARS-Coronavirus-2. Int J Infect Dis 2020;94:55–58. 15. Fantini J, Di Scala C, Chahinian H, Yahi N. Structural and molecular modeling studies 32. Dixon L, Varley J, Gontsarova A, et al. COVID-19-related acute necrotizing en- reveal a new mechanism of action of chloroquine and hydroxychloroquine against SARS- cephalopathy with brain stem involvement in a patient with aplastic anemia. Neurol CoV-2 infection. Int J Antimicrob Agents 2020. 10.1016/j.ijantimicag.2020.105960. Neuroimmunol Neuroinflamm 2020;7:e789. doi: 10.1212/NXI.0000000000000789. 16. Vaduganathan M, Vardeny O, Michel T, et al. Renin–angiotensin–aldosterone system 33. Brun G, Hak JF, Coze S, et al. COVID-19–White matter and globus pallidum lesions: inhibitors in patients with covid-19. N Engl J Med 2020;382:1653–1659. Demyelination or small-vessel vasculitis? Neurol Neuroimmunol Neuroinflamm 17. Dalakas MC. Quarles RH. Autoimmune ataxic Neuropathies (sensory gangliono- 2020;7:e777. doi: 10.1212/NXI.0000000000000777. pathies): are glycolipids the responsible autoantigens? Ann Neurol 1996;39:419–422. 34. Dalakas MC, Illa I, Pezeshkpour GH, Laukaitis JP, Cohen B, Griffin J. Mitochondrial 18. Kusunoki S, Shimizu J, Chiba A, et al. Experimental sensory neuropathy induced by myopathy caused by long-term zidovudine therapy. N Engl J Med 1990;332: sensitization with ganglioside GD1b. Ann Neurol 1996;39:424–431. 1098–1105. 19. Daune GC, Farrer RG, Dalakas MC, Quarles RH. Sensory neuropathy associated with 35. Dalakas MC. Peripheral neuropathy and antiretroviral drugs. J Peripher Nerv Syst monoclonal IgM to GD1b ganglioside. Ann Neurol 1992;31:683–685. 2001;6:14–20. 20. Suwanwongse K, Shabarek N. Rhabdomyolysis as a presentation of 2019 novel 36. Stein M, Bell MJ, Ang LC. Hydroxychloroquine neuromyotoxicity. J Rheumatol 2000; coronavirus disease. Cureus 2020;12:e7561. 27:2927–2931. 21. Jin M, Tong Q. Rhabdomyolysis as potential late complication associated with 2019 37. van Schaik IN, Bril V, van Geloven N, et al; For the PATH Study Group. Sub- novel coronavirus disease. Emerg Infect Dis 2020;26. cutaneous immunoglobulin for maintenance treatment in chronic inflammatory de- 22. Dalakas MC. Inflammatory muscle diseases. N Engl J Med 2015;372:1734–1747. myelinating polyneuropathy (PATH): a randomised, double-blind, placebo- 23. Dalakas MC, Pezeshkpour GH, Gravell M, Sever JL. Polymyositis associated with controlled, phase 3 trial. Lancet Neurol 2017;17:35–46. AIDS retrovirus. JAMA 1986;256:2381–2383. 38. Dalakas MC. Inhibition of B cell functions: implications for neurology. Neurology 24. Dalakas MC. Are autoantibodies pathogenic in necrotizing myopathy? Nat Rev 2008;70:2252–2260. Rheumatol 2018;14:251–252. 39. Fan M, Qiu W, Bu B, et al. Risk of COVID-19 infection in MS and neuromyelitis 25. Watanabe Y, Uruha A, Suzuki S, et al. Clinical features and prognosis in anti-SRP and anti- optica spectrum disorders. Neurol Neuroimmunol Neuroinflamm 2020;7:e787. doi: HMGCR necrotizing myopathy. J Neurol Neurosurg Psychiatry 2016;87:1038–1044. 10.1212/NXI.0000000000000787. 26. Cabello-Verrugio C, Morales M, Rivera J, et al. Renin-angiotensin system: an old 40. Habernan R, Axelrad J, Chen A, et al. Covid-19 in immune-mediated inflammatory player with novel functions in skeletal muscle. Med Res Rev 2015;35:437–463. diseases—case series from New York. 10.1056/NEJMc2009567. 27. Leff RL, Love LA, Miller FW, et al. Viruses in the idiopathic inflammatory myopathies: 41. Risitano AM, Mastellos DC, Huber-Lang M, et al. Complement as a target in COVID- absence of candidate viral genomes in muscle. Lancet 1992;339:1192–1195. 19? Nat Rev Immunol 2020. 10.1038/s41577-020-0320-7. 28. Leon-Monzon M, Lamperth L, Dalakas MC. Search for HIV proviral DNA and 42. Eculizumab (Soliris) in Covid-19 Infected Patients (SOLID-C19). Available at: amplified sequences in the muscle biopsies of patients with HIV polymyositis. Muscle https://clinicaltrials.gov/ct2/show/NCT04288713. NLM identifier: NCT04288713. Nerve 1993;16:408–413. Accessed April 27, 2020.

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN ARTICLE OPEN ACCESS Risk of COVID-19 infection in MS and neuromyelitis optica spectrum disorders

Moli Fan, MD, PhD,* Wei Qiu, MD, PhD,* Bitao Bu, MD, PhD,* Yan Xu, MD, PhD,* Huan Yang, MD, Correspondence Dehui Huang, MD, Alexander Y. Lau, MD, Jun Guo, MD, PhD, Mei-Ni Zhang, MD, Xinghu Zhang, MD, PhD, Dr. Shi [email protected] Chun-Sheng Yang, MD, PhD, Jingshan Chen, MD, Pei Zheng, MD, Qiang Liu, MD, PhD, Chao Zhang, MD, PhD, and Fu-Dong Shi, MD, PhD

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e787. doi:10.1212/NXI.0000000000000787

Abstract MORE ONLINE Objective COVID-19 Resources Disease-modifying drugs (DMDs) may alter the immune status and thus increase the sus- For the latest articles, ceptibility to coronavirus disease 2019 (COVID-19) in patients with MS or neuromyelitis invited commentaries, and optica spectrum disorders (NMOSD). However, evidence supporting this notion is currently blogs from physicians lacking. In this study, we conducted a survey on the risk of COVID-19 in patients with MS and around the world NMOSD. NPub.org/COVID19

Methods The survey was conducted through the Chinese Medical Network for Neuroinﬂammation. Patients in 10 MS centers from 8 cities including Wuhan were included. Information about MS and NMOSD disease duration and the usage of DMDs were collected. Data of suspected cases of COVID-19 were obtained from hospital visits, questionnaires, and patient self-reporting. Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection was conﬁrmed through clinical evaluation by a panel of experts in conjunction with chest CT and viral RNA detection.

Results Eight hundred eighty-two of 1,804 (48.89%) patients with MS and 2,129 of 3,060 (69.58%) patients with NMOSD were receiving DMDs. There were no alterations in the patients’ DMD regimen during January 15, 2020, to March 15, 2020, the 3-month period. None of the patients with MS treated with DMDs had COVID-19. However, 2 patients with relapsing NMOSD were diagnosed with COVID-19-related pneumonia. After treatment, both patients recovered from pneumonia and neither patient experienced new attacks due to predisposing SARS-CoV-2 infection in the following 2 months.

Conclusions No increased risk of COVID-19 infection was observed in patients with MS or NMOSD, irrespective of whether these patients received DMDs. A battery of stringent preventive measures adopted by neurologists to reduce COVID-19 infection in these patients may have contributed to low risk of COVID-19 infection.

*These authors contributed equally to this work.

From the Department of Neurology (M.F., H.Y., J.C., P. Zheng., C. Zhang, F.-D.S.), Tianjin Medical University General Hospital; China National Clinical Research Center for Neurological Diseases (X. Zhang, F.-D-.S.), Jing-Jin Center for Neuroinfalmmation Beijing Tiantan Hospital, Capital Medical University; Department of Neurology (W.Q.), the Third Affiliated Hospital of Sun Yat-sen University, Guangzhou; Department of Neurology (B.B.), Tongji Hospital Affiliated to Tongji Medical College, Huazhong University of Science and Technology, Wuhan; Department of Neurology (Y.X.), Peking Union Medical College Hospital, Beijing; Xiangya Hospital of Central South University (H. Yang), Changsha; Department of Neurology (D. Huang), General Hospital of Chinese People’s Liberation Army, Beijing; Division of Neurology (A.Y.L.), Department of Medicine and Therapeutics, Prince of Wales Hospital, the Chinese University of Hong Kong; Department of Neurology (J.G.), Tangdu Hospital, Air Force Military Medical University, Xi’an; Department of Neurology (M. Zhang), the First Affiliated Hospital of Shanxi Medical University, Taiyuan, China.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

The Article Processing Charge was funded by the authors. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary CMNN = Chinese Medical Network for Neuroinﬂammation; COVID-19 = coronavirus disease 2019; DMD = disease- modifying drug; NMOSD = neuromyelitis optica spectrum disorder; SARS-CoV-2 = severe acute respiratory syndrome coronavirus 2.

The number of the novel coronavirus disease 2019 (COVID- with MS and NMOSD managed in 10 centers from 8 cities 19) cases and mortality grow rapidly around the world. Aged including Wuhan were enrolled in this study. Information individuals with preexisting medical conditions are at in- about disease duration and usage of disease-modifying drugs creased risk for severe acute respiratory syndrome coronavirus (DMDs) were collected. Data of suspected cases of COVID- 2 (SARS-CoV-2) infection, with disease and mortality oc- 19 were obtained from hospital visits, online questionnaires – curring predominantly in this group.1 3 It is thus conceivable (supplementary file, links.lww.com/NXI/A266), and patient that patients with MS or neuromyelitis optica spectrum dis- self-reporting via emails or phone calls. Patients who did not orders (NMOSDs) who are receiving disease-modifying respond to questionnaires were excluded from this study. treatments have increased susceptibility to COVID-19 in- Confirmation of SARS-CoV-2 infection was determined fection and disease. However, evidence supporting this notion through clinical evaluation by a panel of experts in conjunc- is currently lacking. tion with a set of tests including chest CT and viral RNA detection by real-time fluorescent PCR from nasopharyngeal swab samples. Methods Data availability We have conducted a survey through the Chinese Medical All data are published in this article. Network for Neuroinflammation (CMNN). The CMNN consists of neurologists specialized in managing patients with MS and NMOSD from 47 hospitals across China (figure 1). Results MS and NMOSD were diagnosed according to the 2017 McDonald criteria and 2015 International Panel for Neuro- We conducted a survey for COVID-19 infection during myelitis Optica Diagnosis criteria, respectively. All patients January 15 to March 15. One thousand eight hundred four of

Figure 1 Distribution map of hospitals of the Chinese Medical Network for Neuroinflammation (CMNN)

The cities marked red were surveyed for the usage of disease-modifying drugs. The CMNN was founded on February 25, 2018. It is a branch of the China National Clinical Research Center for Neurological Diseases. Until now, members of the CMNN include 47 hospitals from 21 provinces and munici- palities in mainland China. The location and number of hospitals are illustrated. CMNN’s mission includes consultation for government policy-making, introduction of disease-modifying drugs into the National Reimbursement Drug List, fostering research collaborations, and educating neuroimmunologists.

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN 1,836 (98.26%) patients with MS and 3,060 of 3,128 Discussion (97.82%) patients with NMOSD from 10 MS/NMOSD centers across China, including Tongji Hospital from Natalizumab, rituximab, and fingolimod usage is associated Wuhan, have responded (figure 2). The mean disease du- with reactivation of John Cunningham virus and progressive 5 ration was 4.64 years for patients with MS and 4.94 years for multifocal leukoencephalopathy in patients with MS. Side patients with NMOSD. For MS, 882 (48.89%) patients were effects for newer therapies such as eculizumab, inebilizu- receiving DMDs, and the remaining patients received non- mab, satralizumab, and tocilizumab in NMOSD are unclear, DMDs such as corticosteroids, traditional Chinese medi- partly because of the relatively short duration and small cines, or other drugs. For NMOSD, 2,129 (69.58%) patients numbers of patients’ exposure to these medications. Be- were receiving DMDs. There were no alterations in the cause these DMDs interfere with multiple arms of the im- patients’ DMD regimen during this 3-month period (table, mune system, altered immune functions in these patients figures 3 and 4). None of the patients with MS treated are expected. Despite this anticipated risk, we did not ob- with DMDs were diagnosed with COVID-19. However, 2 serve an escalated rate of COVID-19 infection, even at the patients with relapsing NMOSD, a 53-year-old man from epicenter of the outbreak in Wuhan. SARS-CoV-2 has Wuhan and a 49-year-old man from Beijing, were diagnosed spread to all the provinces of China from the Wuhan epi- with COVID-19-related pneumonia, and SARS-CoV-2 in- center since January 2020. The overall incidence of fection was confirmed by viral RNA detection. Both patients COVID-19 infection in China is estimated at 6/105 subjects had received oral methylprednisolone as maintenance ther- comparable with the estimated incidence of NMOSD in apy to prevent relapses, and neither had been treated with China. These results from China are consistent with other DMDs after diagnosis. After treatment according to observations in some MS centers from Korea, Japan, and the Chinese protocols for COVID-19,4 both patients re- Singapore where no COVID-19 cases have been reported for covered from pneumonia and neither patient experienced patients with MS or NMOSD treated with DMDs (per- new attacks due to predisposing SARS-CoV-2 infection in sonal communications). As with the general public, the following 2 months. Neurologists from centers who have patients with MS and NMOSD do not have immunity to not contributed to these data have reported no COVID-19- SARS-CoV-2, and altered immunity induced by DMDs in infected patients with MS or NMOSD diagnosed in their patients with MS or NMOSD, if any, appears insufficient to clinic. enhance the susceptibility to infection.

Figure 2 Distribution map of COVID-19 cases across China and cities and hospitals surveyed (red) for this study

COVID-19 = coronavirus disease 2019.

Table DMDs used by MS and NMOSD patients and confirmed COVID-19 cases in 10 centers from 8 cities in China

Cities of MS/NMOSD centers surveyed

Characteristics Wuhan Beijing Guangzhou Changsha Hong Kong Tianjin Xi’an Taiyuan Total

Population (million) 14.2 21.5 14.9 8.2 7.5 15.6 10.0 4.2 96.1

COVID-19 confirmed 50,004 456 351 242 162 136 120 20 51,491

No. of patients with MS 78 796a 350 152 160 90 34 144 1804

Disease duration, mean, y 6.63 1.9 7 4.55 11.45 4.61 5.96 5.2 4.64

DMDs, n (%) 51 (65.40) 324 (42.13) 146 (41.71) 56 (36.84) 138 (86.25) 56 (62.22) 27 (79.41) 84 (58.33) 882 (48.89)

Interferon beta 9 (11.50) 51 (6.63) 5 (1.43) 14 (9.21) 34 (21.26)b 0 1 (2.94) 45 (31.25) 159 (8.81)

Teriflunomide 30 (38.50) 226 (29.39) 91 (26.00) 42 (27.63) 20 (12.50) 30 (33.33) 9 (26.47) 27 (18.75) 475 (26.33)

Fingolimod 0 27 (3.51) 2 (0.57) 0 31 (19.38) 1 (1.25) 2 (5.88) 0 63 (3.49)

Rituximab 6 (7.70) 12 (1.56) 34 (9.71) 0 4 (2.50) 25 (31.25) 15 (44.12) 12 (8.33) 108 (5.99)

Dimethyl fumarate 0 1 (0.13) 6 (1.71) 0 39 (24.38) 0 0 0 46 (2.55)

Cladribine 0 0 0 0 6 (3.75) 0 0 0 6 (0.33)

Alemtuzumab 0 0 0 0 4 (2.50) 0 0 0 4 (0.22)

COVID-19 confirmed 000 00 0000

No. of patients with NMOSD 588 1,079c 450 453 30 155 158 147 3,060

Disease duration, mean, y 5.87 3.8 6 3.75 12.69 5.52 5.5 7.2 4.94

DMDs, n (%) 420 (71.43) 697 (64.6) 375 (83.33) 268 (59.16) 22 (73.33) 100 (64.52) 150 (94.94) 97 (65.99) 2,129 (69.58)

Methylprednisoloned 36 (6.12) 30 (2.78) 412 (91.56) 113 (24.94) 4 13.33) 55 (35.48) 8 (5.06) 137 (93.20) 795 (25.98)

Azathioprinee 18 (3.06) 104 (9.64) 180 (40.00) 65 (14.40) 12 (40.00) 2 (1.29) 7 (4.43) 17 (11.56) 405 (13.24)

Mycophenolate mofetil 24 (4.08) 439 (40.69) 146 (32.44) 173 (38.15) 6 (20.00) 6 (3.87) 6 (3.80) 32 (21.77) 832 (27.19)

Tacrolimus 360 (61.22) 17 (1.58) 14 (3.11) 12 (2.60) 0 0 0 0 403 (13.17)

Rituximab 12 (2.04) 93 (8.62) 29 (6.44) 18 (3.94) 3 (10.00) 55 (35.48) 137 (86.71) 34 (23.135) 381 (12.45)

Tocilizumab 0 10 (0.93) 6 (1.33) 0 0 35 (22.58) 0 11 (7.48) 62 (2.03)

Cyclophosphamide 0 34 (3.15) 0 0 0 2 (1.29) 0 3 (2.04) 39 (1.27)

COVID-19 confirmed 110 00 0002

Abbreviations: COVID-19 = coronavirus disease 2019; DMD = disease-modifying drug; NMOSD = neuromyelitis optica spectrum disorder. a Of all patients with MS from centers in Beijing, 154 patients were from Beijing Tiantan Hospital, 206 were from the General Hospital of Chinese People’s Liberation Army, and 436 from Peking Union Medical College Hospital. b For patients receiving interferon beta treatment in Hong Kong, 25 received Rebif, 5 received Avonex, and 4 received Betaferon. In other centers, patients received Betaferon treatment. c Of all patients with NMOSD from centers in Beijing, 280 patients were from Beijing Tiantan Hospital, 101 were from the General Hospital of Chinese People’s Liberation Army, and 698 from Peking Union Medical College Hospital. d Methylprednisolone was used as monotherapy. e A proportion of patients who received treatment with immunosuppressants used concomitant methylprednisolone. Figure 3 Proportions of patients with MS in 10 centers from 8 cities surveyed who received disease-modifying drugs and methylprednisolone

COVID-19 = coronavirus disease 2019; NMOSD = neuromyelitis optica spectrum disorder.

From the start of the SARS-CoV-2 epidemic in December 2019 on DMDs differs from those in other countries. Specifically, and during its progression in China, neuroimmunologists natalizumab, cladribine, and alemtuzumab are not available have paid special attention to their patients with MS and to the patients from mainland China. Nevertheless, the NMOSD receiving DMDs who are at increased risk for collection of DMDs used by patients from Hong Kong, infection. Meticulous preventive protocols have been Korea, Japan, and Singapore are similar to those used in adopted, which have undoubtedly minimized the exposure the United States and EU, implying the generality of our of these at-risk patients, and thereby have contributed to the conclusions. absence or low rates of SARS-CoV-2 infection in our patients. These measures include but are not limited to the Despite these limitations, our data originated from a large following: scaling up viral test capacities and swift isolation number of patients that include individuals from the of infected patients, online patient consultations to reduce hardest hit city of Wuhan and its surrounding regions, ar- visits to hospitals, and coordinating patients with local care guing against a significantly increased risk of COVID-19 givers to perform infusions and routine monitoring. infection and disease in patients with MS and NMOSD Moreover, new patients or those with a suspected new MS treated with DMDs. However, a variety of stringent or NMOSD attack were segregated in special hospital units measures that have been taken to protect these patients with single rooms, which is not a common practice in must have contributed to relative low risk of COVID-19 Chinese inpatient facilities. Portions of the Department of infection in our cohort. Because COVID-19 has already Neurology have further been reconfigured to accommodate infected 3,387 medical professionals and has caused 28 patients with suspected infected. Visits from friends and deaths among them in China, extraordinary steps in man- relatives have been strictly controlled. aging neurologic patients to protect both patients and medical professionals from COVID-19 infection are criti- This study has several limitations. Self-reporting and cally important. Because COVID-19 is becoming a serious questionnaires may have missed some patients, especially global health concern6 and its spread is still intensifying in those who had minor symptoms. In rare cases, patients many parts of the world, a full-scale preparedness plan, with refused to disclose this information to avoid mandatory unconditional support for manpower and resources from quarantine in designated facilities. One aspect of the DMDs federal and local governments, is required to better manage available for patients from mainland China and durations neurologic patients during the pandemic.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 5 Figure 4 Proportions of patients with NMOSD in 10 centers from 8 cities surveyed who received disease-modifying drugs and methylprednisolone

COVID-19 = coronavirus disease 2019; NMOSD = neuromyelitis optica spectrum disorder.

Dedication University, Hohhot); Jiawei Wang, MD (Beijing Tongren Hospital, Capital Medical University, Beijing); Lihua Wang, This article is dedicated to our colleagues Bin Zhao, Guang- MD (The 2nd Aﬃliated Hospital of Harbin Medical xun Yan, Yuanyuang Qu, Shasha Han, Mengxi Li, Lianghui University, Harbin); Tao Jin, MD (First Hospital of Jilin Yang, Lei Su, Jing Xu, Mengya Xing, Zhe Zhang, Jinmei Wang, University, Changchun); Manxia Wang, MD (The Second Wei-Na Jin from the Departments of Neurology, Tianjin Hospital of Lanzhou University, Lanzhou); Xiangjun Chen, Medical University General Hospital and Beijing Tiantan MD (Huashan Hospital Fudan University, Shanghai). The Hospital, and to those medical professionals across China, authors also thank D.-C. Tian, K.-B. Shi, and members who went to Wuhan, and Departments of Infectious Disease of Jing-Jin Neuroimmunology team for extensive support; caring for patients with COVID-19. Drs. Luc Van Kear and Samuel X. Shi for English editing.

Study funding Acknowledgment This study was supported by the Advanced Innovation Center The authors thank the following neurologist colleagues for for Human Brain Protection (Neuroimmune 01), and the sharing information on the risk of COVID-19 infections in National Science Foundation of China (91642205, 81830038, their clinics: Kazuo Fujihara, MD (Department of Multiple and 81601019). Sclerosis Therapeutics, Fukushima Medical University School of Medicine, Japan); Ho Jin Kim, MD, PhD Disclosure (Division of Clinical Research, Research Institute & The authors report no disclosures relevant to the manuscript. Hospital of National Cancer Center, Korea); Kevin Tan, None of the funding organizations were involved in the design BM BS, MRCP (UK) (Department of Neurology, National and conduct of the study; collection, management, analysis, and Neuroscience Institute, Singapore); Weibin Liu, MD (The interpretation of the data; preparation, review, or approval of First Affiliated Hospital of Sun Yat-sen University, Guangz- the manuscript; or the decision to submit the manuscript for hou); Ying Fu, MD (The First Affiliated Hospital of Fujian publication. Go to Neurology.org/NN for full disclosures. Medical University, Fuzhou); Xiaokun Qi, MD (The Navy General Hospital, Beijing); Guangzhi Liu, MD (Beijing Publication history Anzhen Hospital, Capital Medical University, Beijing); Zeyu Received by Neurology: Neuroimmunology & Neuroinflammation Li, MD (The Affiliated Hospital of Inner Mongolia Medical April 20, 2020. Accepted in final form April 27, 2020.

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Appendix Authors Appendix (continued)

Name Location Contribution Name Location Contribution

Moli Fan, Department of Neurology, Data collection, statistical Jingshan Department of Neurology, Data collection, statistical MD, PhD Tianjin Medical University analysis, data Chen, MD Tianjin Medical University analysis, data General Hospital, China interpretation, and General Hospital, China interpretation and administrative, technical, or drafting of the manuscript, material support and administrative, technical, or material Wei Qiu, Department of Neurology, Data collection, statistical support MD, PhD the Third Affiliated Hospital analysis, and data of Sun Yat-sen University, interpretation Pei Zheng, Department of Neurology, Data collection, statistical Guangzhou, China MD Tianjin Medical University analysis, data General Hospital, China interpretation and Bitao Bu, Department of Neurology, Data collection, statistical drafting of the manuscript, MD, PhD Tongji Hospital affiliated to analysis, and data and administrative, Tongji Medical College, interpretation technical, or material Huazhong University of support Science and Technology, Wuhan, China Qiang Liu, Department of Neurology, Data collection, statistical MD, PhD Tianjin Medical University analysis, and data Yan Xu, Department of Neurology, Data collection, statistical General Hospital, China interpretation and MD, PhD Peking Union Medical College analysis, and data drafting of the Hospital, Beijing, China interpretation manuscript

Huan Department of Neurology, Data collection, Chao Department of Neurology, Data collection, statistical Yang, MD, Xiangya Hospital of Central statisticalanalysis, and data Zhang, Tianjin Medical University analysis, data PhD South University, Changsha, interpretation MD, PhD General Hospital, China interpretation, drafting China. of the manuscript, and administrative, Dehui Department of Neurology, Data collection, statistical technical, or material Huang, General Hospital of Chinese analysis, and data support MD People’s Liberation Army, interpretation Beijing, China Fu-Dong Department of Neurology, Concept and design, Shi, MD, Tianjin Medical University obtained funding, Alexander Division of Neurology, Data collection, statistical PhD General Hospital, data collection, statistical Y. Lau, MD Department of Medicine and analysis, and data China; China National analysis, data Therapeutics, Prince of Wales interpretation Clinical Research Center interpretation, drafting Hospital, The Chinese for Neurological of the manuscript, University of Hong Kong, China Diseases, Jing-Jin Center supervision of this for Neuroinfalmmation, study, and administrative, Jun Guo, Department of Neurology, Data collection, statistical Beijing Tiantan Hospital, technical, or material MD, PhD Tangdu Hospital, Air Force analysis, and data Capital Medical support Military Medical University, interpretation University, Xi’an, China China

Mei-Ni Department of Neurology, Data collection, statistical Zhang, MD The First Affiliated Hospital analysis, and data of Shanxi Medical interpretation University, Taiyuan, China References 1. Huang CL, Wang YM, Li XW, et al. Clinical features of patients infected with 2019 Xinghu China National Clinical Data collection, statistical novel coronavirus in Wuhan, China. Lancet 2020;395:497–506. Zhang, MD Research Center for analysis, and data 2. Li Q, Guan XH, Wu P, et al. Early transmission dynamics in Wuhan, China, of novel Neurological Diseases, interpretation coronavirus-infected pneumonia. N Engl J Med 2020;382:1199–1207. Department of Neurology, 3. WangDW,HuB,HuC,etal.Clinical characteristics of 138 hospitalized patients Beijing Tiantan Hospital, Capital with 2019 novel coronavirus-infected pneumonia in Wuhan, China. JAMA 2020; Medical University, China 323:1061–1069. 4. Wang T, Zhu F, Cao Z, An Y, Gao Y, Jiang B. Comorbidities and multi-organ injuries Chun- Department of Neurology, Data collection, statistical in the treatment of COVID-19. Lancet 2020;395:e52. Sheng Tianjin Medical University analysis, and data 5. Gustavo L, Peter A, Joachim B, et al. Infection risks among patients with multiple Yang, MD, General Hospital, China interpretation and sclerosis treated with ﬁngolimod, natalizumab, rituximab, and injectable therapies. PhD administrative, technical, or JAMA Neurol 2019;77:184–191. material support 6. Wang C, Peter WH, Frederick GH, et al. A novel coronavirus outbreak of global health concern. Lancet 2020;395:470–473.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 7 ARTICLE OPEN ACCESS COVID-19-related acute necrotizing encephalopathy with brain stem involvement in a patient with aplastic anemia

Luke Dixon, FRCR, James Varley, PhD, Anastassia Gontsarova, FRCR, Dermot Mallon, PhD, Correspondence Francesca Tona, MD, David Muir, MRCPath, Asad Luqmani, FRCPath, Ieuan Harri Jenkins, MD, Dr. Dixon [email protected] Richard Nicholas, PhD, Brynmor Jones, FRCR, and Alex Everitt, PhD

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e789. doi:10.1212/NXI.0000000000000789

Abstract MORE ONLINE Objective COVID-19 Resources To describe a novel case of coronavirus disease 2019 (COVID-19)-associated acute necrotizing For the latest articles, encephalopathy (ANE) in a patient with aplastic anemia where there was early brain stem- invited commentaries, and predominant involvement. blogs from physicians around the world Methods NPub.org/COVID19 Evaluation of cause, clinical symptoms, and treatment response.

Results A 59-year-old woman with a background of transfusion-dependent aplastic anemia presented with seizures and reduced level of consciousness 10 days after the onset of subjective fever, cough, and headache. Nasopharyngeal swab testing for severe acute respiratory syndrome coronavirus (SARS-CoV-2) was positive, and CT during admission demonstrated diﬀuse swelling of the brain stem. She required intubation and mechanical ventilation for airway protection, given her reduced level of consciousness. The patient’s condition deteriorated, and MRI on day 6 demonstrated worsening brain stem swelling with symmetrical hemorrhagic lesions in the brain stem, amygdalae, putamina, and thalamic nuclei. Appearances were consistent with hemorrhagic ANE with early brain stem involvement. The patient showed no response to steroid therapy and died on the eighth day of admission.

Conclusions COVID-19 may be associated with an acute severe encephalopathy and, in this case, was considered most likely to represent an immune-mediated phenomenon. As the pandemic continues, we anticipate that the spectrum of neurologic presentation will broaden. It will be important to delineate the full clinical range of emergent COVID-19-related neurologic disease.

From the Department of Imaging (L.D., A.G., D. Mallon, F.T., B.J.), Imperial College Healthcare NHS Trust; Department of Neurosciences (J.V., I.H.J., A.E.), Imperial College Healthcare NHS Trust; Northwest London Pathology (D. Muir); Department of Hematology (A.L.), Imperial College Healthcare NHS Trust; Centre for Neuroinflammation and Neurodegeneration (R.N.), Faculty of Medicine, Imperial College London; and Department of Visual Neuroscience (R.N.), UCL Institute of Ophthalmology, London, United Kingdom.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

The Article Processing Charge was funded by the Imperial College London University. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary ANE = acute necrotizing encephalopathy; COVID-19 = coronavirus disease 2019; GCS = Glasgow Coma Score; GTCS = generalized tonic-clonic seizure; SARS-CoV-2 = severe acute respiratory syndrome coronavirus 2.

Coronavirus disease 2019 (COVID-19) is caused by severe no focal deficits. Chest radiography showed right basal con- acute respiratory syndrome coronavirus 2 (SARS-CoV-2) solidation, and a CT scan of the head showed early swelling of and, was first detected in the human population in late De- the brain stem (figure 1). On admission, her blood cell count cember 2019. As of April 30, 2020, over 3 million cases have showed unchanged anemia and thrombocytopenia (platelet been reported worldwide and over 230,000 people have died count 29 × 109/L) which was consistent with the patient’s from the infection. The typical presentation of SARS-CoV-2 history of aplastic anemia (table). Unlike previous blood with fever and respiratory symptoms is well recognized, al- counts, there was new lymphopenia (lymphocytes 0.3 from though there is comparatively little reported on the neuro- 1.4 2 days earlier). Nasopharyngeal swab RT-PCR testing for logic sequelae. Despite the paucity of reported cases, there is SARS-CoV-2 returned positive, thus confirming a diagnosis of increasing evidence that patients with severe COVID-19 COVID-19. The patient was started on levetiracetam and IV often develop neurologic manifestations such as impaired ceftriaxone, aciclovir, amoxicillin, and clarithromycin. In light 1 consciousness. There have been reported cases of COVID- of the severe thrombocytopenia and risk of hemorrhage, hu- 19-associated encephalopathy and a single case of pre- man leukocyte antigen-matched platelets (because of known sumptive COVID-19-related acute necrotizing hemorrhagic platelet HLA antibodies) were transfused intermittently to encephalopathy associated with symmetrical hemorrhagic, maintain the platelet count >50 × 109/L. necrotic lesions in both the thalamic nuclei and amygdalae.2 Here, we report a further case of possible COVID-19-related Twelve hours after admission, the patient’sGCSfellto5 necrotizing hemorrhagic encephalopathy associated with early (E1, V1, and M3), with associated development of an ex- brain stem involvement. tensor left plantar response and an unreactive left pupil. Although the patient’s respiratory symptoms remained relatively mild, she underwent endotracheal intubation for Case report airway protection and was transferred to the intensive care A 59-year-old woman presented to the emergency department unit for mechanical ventilation. Repeat head CT showed with recurrent fleeting episodes of vacant staring and speech increased hypodensity and swelling of the brain stem, and arrest associated with flexion of both shoulders and a brief a new area of cortical and subcortical hypodensity in the left witnessed generalized tonic-clonic seizure (GTCS), fol- occipital lobe initially suggested an acute posterior circula- lowed by postictal reduced consciousness. Shortly after ar- tion infarct. A subsequent computed tomography angiogram rival in the emergency department, the patient vomited and excluded an acute vascular occlusion but showed worsening had a further GTCS. She had a history of aplastic anemia brain stem swelling with subtle intrinsic pontine hemorrhage treated with intermittent red blood cell and platelet trans- and new symmetrical hypodensities in the deep gray matter fi fusions. The patient had no significant paroxysmal nocturnal and amygdalae ( gure 1). Appearances were suggestive of hemoglobinuria clone. She had received immunosuppressive a rapidly evolving encephalopathy with severe involvement therapy in the past, but not recently. She had returned from of the brain stem. On the fifth day of admission, lumbar a trip to Afghanistan 3 weeks before presentation and de- puncture was performed immediately after platelet trans- veloped transient abdominal pain and diarrhea. Ten days fusion. CSF opening pressure was 28 cm water, and CSF before her neurologic presentation, she developed a persis- analysis showed increased protein concentration (2.3 g/L) 3 tent cough, sore throat, shivering, and headache, with sub- and a normal white cell count of 4/mm . Subsequent standard sequent shortness of breath and myalgia. Three days before CSF virology PCR (herpes simplex virus 1 and 2, varicella presentation, she had a routine telephone hematology clinic zoster virus, adenovirus, cytomegalovirus, Epstein-Barr virus, consultation and the following day underwent a blood test enterovirus, parechovrius, and human herpesvirus 6), CSF (table, day-2) which showed a stable platelet count (33 × PCR for SARS-CoV-2, and CSF culture were negative. 109/L), anemia (hemoglobin 103 g/L), and leukopenia (total white cell count 3.8 × 109/L, lymphocyte count 1.4 × On the sixth day of admission, an MRI of the head demon- 109/L, neutrophil count 1.8 × 109/L). strated extensive, relatively symmetrical changes throughout the supratentorial and infratentorial compartments. There was Physical examination revealed reduced consciousness with diffuse swelling and hemorrhage in the brain stem and both a Glasgow Coma Score (GCS) of 11 of 15, body temperature of amygdalae. Extensive abnormal signal and microhemorrhage 36.9°C, blood pressure of 130/63 mm Hg, pulse of 82 beats per were found in a symmetrical distribution within the dorsolat- minute, respiratory rate of 22 breaths per minute, and oxygen eral putamina, ventrolateral thalamic nuclei, subinsular regions, saturation of 97% on ambient air. Neurologic assessment found splenium of the corpus callosum, cingulate gyri, and subcortical

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Table Clinical laboratory results

Day of COVID-19 symptoms 71011121416

Day of admission −20 1 3 5 7

Hemoglobin (g/L, 114–150) 103 105 98 79 71 62

WBC (109/L, 4.2–11.2) 3.8 3.2 5.1 2.9 2.4 3.3

Neutrophils (109/L, 2.0–7.1) 1.8 2.7 4.3 2.5 1.6 2.5

Lymphocytes (109/L, 1.1–3.6) 1.4 0.3 0.5 0.3 0.6 0.6

Monocytes (109/L, 0.3–0.9) 0.6 0.2 0.3 0.1 0.2 0.2

Platelets (109/L, 135–400) 33 29 66 (post platelets) 51 47 65

Sodium (mmol/L, 133–146) 139 135 140 150 149 144

Potassium (mmol/L, 3.5–5.3) 3.6 3.9 3.2 3.8 3.2 3.6

Blood urea nitrogen (mmol/L, 2.5–7.8) 2.5 4.3 4.9 8.5 7.1 6.6

Creatinine (umol/L, 55–110) 104 101 105 116 116 113

Corrected calcium (mmol/L, 2.2–2.6) 2.16 2.07 2.17 1.98 2.1

ALT (unit/L, 0–34) 11 17 13 11 9 13

Total bilirubin (umol/L, 0–21) 88 129 6 5

CRP (mg/L, 0.0–5.0) 6.8 8.1 8.0 105.8 91.8 144.9

Ferritin (ug/L, 20–300) 98 253 314 544

D-dimer (ng/mL <500) 2033 1,203 1997 2044

Creatine kinase (unit/L) 54 55

Troponin (ng/L) 19 24

Abbreviations: ALT = alanine transaminase; COVID-19 = Coronavirus disease 2019; CRP = C-reactive protein; WBC = white blood cell count.

perirolandic regions (figure 2). These regions demonstrated secondary to the severe cerebral and pontine swelling. severe swelling and restricted diffusion with peripheral en- Overall imaging features supported a diffuse hemorrhagic hancement (figure 3). There was partial effacement of the acute necrotizing encephalopathy (ANE) with involvement ventricles, temporal uncal herniation, effacement of the of the brain stem. Neurologic examination after withdrawal basal cisterns, and moderate cerebellar tonsillar herniation of sedation revealed intact corneal reflexes and normal pupillary

Figure 1 CT of the head findings over time

Axial CT head images on different dates. From left to right, premorbid previous CT performed in 2016, 2020 day 0 admission CT, and day 1 follow-up CT. Early admission CT demonstrates subtle new swelling of the brain stem, and the follow-up CT 1 day later shows progression of the swelling with new central hemorrhagic foci (closed arrow) and symmetrical hypodensities in both amygdalae (chevrons). On day 1 of the follow-up CT, there was also hypodensity in both thalami and dorsolateral putamina (not shown).

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 Figure 2 T2-weighted and susceptibility-weighted MRI head at day 6

Serial T2-weighted (A) and susceptibility-weighted (B) axial MRI images of the brain demonstrating abnormal swelling and T2-weighted signal (A) with intrinsic hemorrhage (B) in the subcortical perirolandic regions (diamond arrow heads), dorsolateral putamina, ventrolateral thalamic nuclei and subinsular regions (open arrow heads), amygdalae (chevrons) and, pons (closed arrow heads). Ab- normal signal is also shown in the splenium of the corpus callosum and cingulate gyri (not labeled).

responses to light. Doll’s eye response was reduced. She coughed Discussion on suction and initiated breathing but required pressure support mechanical ventilation. She displayed no response to verbal To our knowledge, we report the second case of presumptive command or painful stimuli. However, based on the severity of COVID-19-related hemorrhagic ANE. Similar to the previously the MRI ﬁndings, the patient was deemed to have a very poor reported case, our patient had relatively symmetrical hemor- neurologic prognosis. She showed no sign of neurologic im- rhagic lesions in the amygdalae and thalamic nuclei although, by provement after high dose dexamethasone and on the 10th day contrast, there was also extensive involvement of the pons and of admission died after the withdrawal of ventilatory support. medulla and, to a lesser extent, the striatum and subcortical Her family declined a postmortem study. perirolandic regions.2 In our case, the patient’saplasticanemia

Figure 3 Diffusion weighted and contrast-enhanced T1-weighted MRI head at day 6

Serial diffusion-weighted imaging (A) and contrast-enhanced T1-weighted (B) axial MRI images of the brain demonstrating abnormal restricted diffusion (A) and peripheral enhancement (B) in the same areas as the abnormal T2-weighted signal. Namely, the subcortical perirolandic regions (diamond arrow heads), dorsolateral putamina, ventrolateral thalamic nuclei and subinsular regions (open arrow heads), amygdalae (chevrons), and pons (closed arrow heads). Re- stricted diffusion is also shown in the splenium of the corpus callosum and cingulate gyri (not labeled).

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN with severe thrombocytopenia likely contributed to the hemorrhagic component of the encephalopathy. However, outside Appendix Authors of bone marrow transplantation and active immunosuppressant Name Location Contributions therapy, there is no reported predisposition to encephalopathy in patients with aplastic anemia.3,4 Our patient was not lym- Luke Dixon Department Data acquisition, of Imaging, Imperial drafting/revising phopenic before COVID-19, and therefore, it is unlikely that the College Healthcare the manuscript, aplastic anemia caused an impaired immune response. NHS Trust, London, analysis or UK. interpretation of the data, and ANE is a rare encephalopathy most frequently encountered study concept or design in children and characterized by multiple, symmetrical lesions in the thalami, striatum, cerebral white matter, and James Varley Department of Data acquisition, 5,6 Neurosciences, drafting/revising brain stem. ANE has previously been linked to several Imperial College the manuscript, infective agents including influenza-A, herpes simplex virus, Healthcare NHS and analysis or fl 6,7 Trust, London, UK. interpretation of in uenza-B, mycoplasma, and human herpes virus-6. As the data in this case, ANE is often rapidly progressive with seizures, Anastassia Gontsarova Department of Data acquisition reduced consciousness, and vomiting, usually occurring Imaging, Imperial and analysis or 5,7 12–72 hours after symptom onset of the viral infection. As College Healthcare interpretation of NHS Trust, London, the data in our case, CSF analysis frequently shows an elevated protein UK. concentration but a normal white blood cell count.5,7 The pre- cise etiology and pathophysiology of ANE remains unclear. In Dermot Mallon Department of Drafting/revising Imaging, Imperial the manuscript reported cases, the suspected causative pathogen is rarely College Healthcare detected in the CSF by PCR assay.5,7 In an autopsy of a patient NHS Trust, London, UK. with ANE secondary to H1N1 influenza, a notable absence of perivascular or meningeal inflammation was found.6 There- Francesca Tona Department of Drafting/revising Imaging, Imperial the manuscript fore, ANE is not believed to be a productofdirectinfection College Healthcare but the result of an immune-mediated process involving proin- NHS Trust, London, flammatory cytokines.8,9 A possible hyperinflammatory response UK. in COVID-19 is supported by the recent observation that David Muir Northwest London Drafting/revising COVID-19-related acute respiratory distress syndrome may be Pathology, UK the manuscript driven by a macrophage activation syndrome or cytokine Asad Luqmani Department of Data acquisition 10 Hematology, and analysis or storm. Unfortunately, in our case, testing for interleukins in the Imperial College interpretation of serum or CSF was not possible. Despite this, an immune- Healthcare NHS the data mediated phenomenon, as opposed to a neurotropic effect, is Trust, London, UK. suggested because CSF RT-PCR testing for SARS-CoV-2 was Ieuan Harri Jenkins Department of Data acquisition, negative and no other potential causative agent was identified. Neurosciences, drafting/revising Imperial College the manuscript, Healthcare NHS and analysis or Further exploration of potential COVID-19-related CNS pa- Trust, London, UK. interpretation of the data thology is needed and requires national and international collaboration to collect large, organized data sets. The possibility Richard Nicholas Centre for Data acquisition Neuroinflammation and analysis or of an immune-mediated process and its therapeutic implica- and interpretation of tions also warrants greater study. Neurodegeneration, the data Faculty of Medicine, Imperial College Study funding London, London, UK. & Department of No targeted funding reported. Visual Neuroscience, UCL Institute of Ophthalmology, Disclosure London, UK. L. Dixon, J. Varley, A. Gontsarova, D. Mallon, F Tona, D. Muir, A. Luqmani, I. Jenkins, R. Nicholas, B. Jones, and A. Brynmor Jones Department of Data acquisition, Imaging, Imperial drafting/revising Everitt reports no disclosures. Go to Neurology.org/NN for College Healthcare the manuscript, full disclosures. NHS Trust, London, analysis or UK. interpretation of the data, and Publication history study concept or design Received by Neurology: Neuroimmunology & Neuroinflammation April 14, 2020. Accepted in final form May 5, 2020. Continued

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 5 3. Mashima K, Yano S, Yokoyama H, et al. Epstein-barr virus-associated lymphoproli- Appendix (continued) ferative disorder with encephalitis following anti-thymocyte globulin for aplastic anemia resolved with rituximab therapy: a case report and literature review. Intern – Name Location Contributions Med 2017;56:701 706. 4. Tam DYS, Cheng FWT, Chan PKS, et al. Intact survival of refractory CMV Alex Daniel Everitt Department of Data acquisition, limbic encephalitis in a patient with severe aplastic anemia after unrelated bone – Neurosciences, drafting/revising marrow transplantation. J Pediatr Hematol Oncol 2012;34:472 474. Imperial College the manuscript, 5. Wong AM, Simon EM, Zimmerman RA, Wang HS, Toh CH, Ng SH. Acute necro- Healthcare NHS analysis or tizing encephalopathy of childhood: correlation of MR findings and clinical outcome. Trust, London, UK. interpretation of Am J Neuroradiol 2006;27:1919–1923. the data, study 6. Offiah C, Hall E. Acute necrotizing encephalopathy associated with novel influenza concept or H1N1 (pdm09) infection: MRI and correlation with brain necropsy. J Pediatr design, and Study Neuroradiol 2013;2:319–324. Supervision 7. Mizuguchi M. Acute necrotizing encephalopathy of childhood: a novel form of acute encephalopathy prevalent in Japan and Taiwan. Brain Dev 1997;19:81–92. 8. Sugaya N. Influenza-associated encephalopathy in Japan: pathogenesis and treatment. References Pediatr Int 2000;42:215–218. 9. Ito Y, Ichiyama T, Kimura H, et al. Detection of influenza virus RNA by reverse 1. Mao L, Jin H, Wang M, et al. Neurologic manifestations of hospitalized patients with coro- fl fl navirus disease 2019 in Wuhan, China. JAMA Neurol Epub 2020 April 10. doi: 10.1001/ transcription-PCR and proin ammatory cytokines in in uenza-virus-associated en- – jamaneurol.2020.1127. cephalopathy. J Med Virol 1999;58:420 425. ’ 2. Poyiadji N, Shahin G, Noujaim D, Stone M, Patel S, Griffith B. COVID-19–associated 10. McGonagleD,SharifK,ORegan A, Bridgewood C. The role of cytokines including acute hemorrhagic necrotizing encephalopathy: CT and MRI features. Radiology interleukin-6 in COVID-19 induced pneumonia and macrophage activation Epub 2020 Mar 31. doi: 10.1148/radiol.2020201187. syndrome-like disease. Autoimmun Rev 2020:102537.

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN ARTICLE OPEN ACCESS CLASS OF EVIDENCE Ocrelizumab shorter infusion Primary results from the ENSEMBLE PLUS substudy in patients with MS

Hans-Peter Hartung, MD, FRCP, FAAN, on behalf of the ENSEMBLE Steering Committee members and study Correspondence investigators Dr. Hartung hans-peter.hartung@ Neurol Neuroimmunol Neuroinﬂamm 2020;7:e807. doi:10.1212/NXI.0000000000000807 uni-duesseldorf.de

Abstract MORE ONLINE Objective Class of Evidence To assess the safety of ocrelizumab (OCR) shorter duration infusion in patients with MS. Criteria for rating therapeutic and diagnostic Methods studies ENSEMBLE PLUS is a randomized, double-blind substudy to the single-arm ENSEMBLE study NPub.org/coe (NCT03085810). In ENSEMBLE, patients with early stage relapsing-remitting MS received OCR 600 mg initially as two 300 mg IV infusions 2 weeks apart and subsequently as a single 3.5-hour 600 mg infusion every 24 weeks for 192 weeks. In ENSEMBLE PLUS, OCR 600 mg administered over the approved 3.5-hour infusion time (conventional duration) is compared with a 2-hour infusion (shorter duration). The primary end point was the proportion of patients with infusion- related reactions (IRRs) after the ﬁrst randomized dose (assessed during and up to 24 hours postinfusion).

Results From November 1, 2018, to September 27, 2019, 580 patients were randomized 1:1 to the conventional or shorter infusion group. After the ﬁrst randomized dose, 67 of 291 patients (23.1%) in the conventional and 71 of 289 patients (24.6%) intheshorterinfusiongroupexperiencedIRRs. Most IRRs were mild or moderate in both groups; one patient in each group experienced a severe IRR, and in both groups, 98.6% (136 of 138) of all IRRs resolved without sequelae. No IRRs were serious, life-threatening, or fatal. No IRR-related discontinuation occurred. During the ﬁrst randomized dose, 14 of 291 (4.8%) and 25 of 289 (8.7%) patients in the conventional and shorter infusion groups, respectively, had IRRs leading to infusion slowing/interruption.

Conclusion The frequency and severity of IRRs were similar between conventional and shorter OCR infusions. Shortening the infusion time to 2 hours reduces the total infusion site stay time and lessens the overall patient and site staﬀ burden.

Classification of evidence This interventional study provides Class I evidence that the frequency and severity of IRRs were similar at the ﬁrst randomized dose using OCR (600 mg) infusions of conventional and shorter duration in patients with relapsing-remitting MS.

Clinical trial identifier number NCT03085810.

From the Department of Neurology, UKD, Center of Neurology and Neuropsychiatry and LVR-Klinikum, Medical Faculty, Heinrich-Heine University Dusseldorf,¨ Germany.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

The article processing charge was funded by the F. Hoffmann-La Roche Ltd.

ENSEMBLE Steering Committee members and coinvestigators are listed in the appendix 2 at the end of article. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary AE = adverse event; iCCOD = interim clinical cutoﬀ; IRR = infusion-related reaction; ITT = intent-to-treat; OCR = ocrelizumab; RRMS = relapsing-remitting MS.

Ocrelizumab (OCR) is a humanized anti-CD20 monoclonal sodium chloride infusion over the remaining 1.5 hours in the antibody approved uniquely for both relapsing and primary shorter infusion group, every 24 weeks for the remainder of progressive MS.1,2 The current OCR infusion schedule, in- the study duration (figure 1B). Blood samples were only cluding mandatory premedication 1-hour preinfusion and collected at the first OCR infusion postrandomization and 30 1-hour postinfusion observations, requires an on-site stay of minutes after the completion of the shorter and conventional 5.5–6 hours. Shortened infusion times can minimize the treat- infusion, representing the peak concentration (Cmax) ment burden for patients, reduce the time required at the in- of OCR. fusion site, and lead to decreased workloads for site staff, – without compromising patient safety.3 5 Here,wedescribethe Patients, site personnel, and sponsor employees remained primary findings from the ENSEMBLE PLUS study evaluating blinded during the study. Infusions were preloaded and placed the safety, including infusion-related reactions (IRRs), of into standardized infusion cover bags on an infusion rack; the a shorter vs conventional infusion of OCR in patients with early infusion administration pump was covered and operated only relapsing-remitting MS (RRMS). by an unblinded infusion nurse.

Standard protocol approvals, registrations, Methods and patient consents The relevant institutional review boards/ethics committees Trial design and patients approved the trial protocols (NCT03085810). All patients The ENSEMBLE PLUS substudy is a prospective, multicenter, provided written informed consent. The Steering Committee randomized, double-blind phase IIIb study designed to evaluate and study investigators gathered the data, and the sponsor the safety of a shorter duration infusion of OCR in patients with performed the data analyses. The author and Steering Com- early stage RRMS enrolled in the main ENSEMBLE study. In mittee agreed to submit the manuscript for publication. ENSEMBLE, treatment-na¨ıve patients (age 18–55 years) with aconfirmed diagnosis of RRMS,6 disease duration ≤3years,one Study objectives or more relapses/signs of MRI activity in the previous The primary research question provides Class I evidence to 12 months, and an Expanded Disability Status Scale score of determine if the frequency and severity of IRRs after the first 0–3.5 (inclusive) received OCR 600 mg infusions every randomized dose differed using OCR (600 mg) infusions of 24 weeks for 192 weeks (up to 8 doses), with mandatory pre- conventional and shorter duration in patients with RRMS. The medication. Patients with a previous serious OCR-related IRR primary end point was the proportion of patients with IRRs were excluded from the substudy. The target enrollment was during or within 24 hours after the first randomized dose using 700 patients in the ENSEMBLE PLUS substudy, which included shorter vs conventional duration OCR infusion groups 150 patients already enrolled in the main ENSEMBLE study (IRRs assessed during and 24 hours postinfusion). Secondary plus 550 newly enrolled patients. end points include the severity and symptoms of IRRs, IRRs leading to treatment discontinuation, the proportion of In all patients, the first dose of OCR was administered, per patients with IRRs overall, and the overall safety. label, as an initial dose of two 300 mg infusions, separated by 14 days (figure 1A). Randomization was performed with the Safety reporting use of an interactive web response system in permuted blocks IRRs were classified as occurring during infusion or within (block size = 4). Randomization to either the conventional or 24 hours after the end of the infusion (collected by follow-up shorter infusion group occurred at week 24 for newly enrolled telephone contact). IRR events occurring in a patient at both patients. For patients already enrolled in the main ENSEM- time points (during and postinfusion) were reported as 2 sep- BLE study, randomization occurred at their next scheduled arate IRRs per infusion. Safety was assessed through the mon- infusion (week 48, 72, 96, or 120). Patients eligible to take itoring and recording of adverse events (AEs) and serious AEs. part in this substudy were randomized (1:1) into conven- AEs were defined as all AEs including IRRs and serious MS tional 3.5-hour and shorter 2-hour infusion groups stratified relapses, but excluding nonserious MS relapses. AEs were by region (United States, Canada, and Australia vs the rest of reported from the first randomized dose onward up to the the world) and dose at which the patient is randomized. interim clinical cutoff date (iCCOD). Patients received 600 mg OCR in 500 mL 0.9% sodium chloride infused over approximately 3.5 hours in the con- Statistical methods ventional infusion group (with a mimic switch infusion at The proportion of patients with IRRs that occurred during or approximately 2 hours) or 2 hours, followed by a 100 mL 0.9% within 24 hours after the first randomized dose of OCR were

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Figure 1 ENSEMBLE PLUS (A) study design and (B) infusion schedule

The ENSEMBLE PLUS primary end point is the proportion of patients with IRRs after the first randomized dose (frequency and severity assessed during and 24 hours postinfusion). aRandomization of new patients at week 24. AH = antihistamine; IRR = infusion-related reaction; IVMP = methylprednisolone; PK = pharmacokinetic assessment. summarized as point estimates of the between-treatment Data availability difference and associated symmetric 2-sided 95% CIs. All The authors confirm that the data supporting the findings of summaries of IRRs are based on the intent-to-treat (ITT) this study are available within the article and from the cor- population; overall AEs are based on the safety population. responding author on reasonable request. Analyses are based on patients who had completed the 24-hour evaluation period after the first randomized dose by Qualified researchers may request access to individual patient- the prespecified iCCOD of September 27, 2019. A descriptive level data through the clinical study data request platform analysis (mean and median in patients with/without IRR) was (https://vivli.org/). Further details on the Roche criteria for used to describe any association between the Cmax of OCR vs eligible studies are available here (https://vivli.org/members/ IRR maximum intensity. ourmembers/). For further details on Roche Global Policy on

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 Figure 2 Patient disposition and analysis of population

One patient disclosed that they were pregnant after randomization but before receiving any study treatment. Per protocol, treatment is withheld from patients who become pregnant during the study. There was also one withdrawal from the conventional infusion group because of an adverse event (depressive symptom) that was considered unrelated to the study treatment but because of concurrent illness of depression. A discontinuation visit had not been scheduled or undertaken for the patient at the time of CCOD; hence, this patient could not be included in any of the tables which display discontinuation. Other: accidental unblinding. CCOD = clinical cutoff date.

the Sharing of Clinical Study Information and how to request patients with prefirst randomized dose IRRs (n = 71/291 access to related clinical study documents, see roche.com/ [24.4%] in the conventional and n = 78/289 [27.0%] in the research_and_development/who_we_are_how_we_work/ shorter infusion groups). clinical_trials/our_commitment_to_data_sharing.htm. Shorter infusion time summary All patients received at least one randomized OCR infusion, Results except for one patient in the conventional infusion group. In the conventional infusion group, 236 of 291 (81.1%), 54 of 291 Patient disposition and analysis population (18.6%), and zero patients received one, 2, and 3 randomized A total of 586 patients were enrolled in the ongoing EN- doses, respectively; this was 233 of 289 (80.6%), 55 of 289 SEMBLE PLUS study by the prespecified iCCOD (183 from (19.0%), and 1 of 289 (0.3%) patients in the shorter infusion the main ENSEMBLE study and 403 newly enrolled group. Overall, the median (range) infusion time was 215 patients) across 21 countries. Of the 586 patients enrolled, (195–350) and 120 (109–255) minutes in the conventional 580 patients were randomized (1:1), stratified by region and and shorter infusion groups, respectively. dose at which the patient is randomized, to the conventional infusion group (N = 291) or shorter infusion group (N = Infusion-related reactions 289, figure 2). Two patients (0.7%) were withdrawn from The incidence of IRRs at the first randomized dose (primary the shorter infusion group. All patients received the full end point) in patients was comparable between the conven- 600 mg dose in each group. Baseline demographics and tional (n = 67/291 [23.1%]) and shorter (n = 71/289 [24.6%]) disease characteristics were well-balanced across conven- infusion groups (stratified difference in proportions [95% CI]: tional and shorter infusion groups; most patients were fe- 2.0% [−4.7% to 8.7%], table). Of patients experiencing IRRs, male (181 of 291 [62.2%]/186 of 289 [64.4%]), with a mean the onset of IRR symptoms occurred during infusion for n = 27 age (SD) of 34.0 (8.5)/34.2 (8.8) years and a mean (SD) of 67 (40.3%) and n = 40 of 71 (56.3%) patients and within duration since relapsing MS diagnosis of 1.1 (0.6)/(0.7) 24 hours postinfusion for n = 48/67 (71.6%) and n = 40/71 years. There was a slight imbalance in the proportion of (56.3%) patients, in conventional and shorter infusion groups,

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Table Summary of (1) primary end point (proportion of patients with IRRs after the first randomized dose) and severity, (2) IRRs at the first randomized dose leading to intervention in OCR infusion, (3) symptoms and management of IRRs, and (4) AEs

Conventional infusion (N = 291) Shorter infusion (N = 289)

No. (%) of patients with an infusion 290 (99.7) 289 (100)

(1) No. (%) of patients with any IRR (primary end point) 67 (23.1) 71 (24.6)

Unstratified difference (95% CI) 1.5 (−5.5 to 8.4)

Stratified difference (95% CI)a 2.0 (−4.7 to 8.7)

Mild (grade 1) 46 (15.9) 47 (16.3)

Moderate (grade 2) 21 (7.2) 23 (8.0)

Severe (grade 3) 0 1 (0.4)

(2) No. (%) of patients with any IRR leading to intervention in OCR infusion 14 (4.8) 25 (8.7)

Infusion discontinued 00

Infusion temporarily interrupted 10 (71.4) 13 (52.0)

Infusion slowed down 4 (28.6) 12 (48.0)

(3) No. (%) of patients with any IRRb,c 67 (23.1) 71 (24.6)

Throat irritation 12 (17.9) 22 (31.0)

Fatigue 17 (25.4) 17 (23.9)

Headache 21 (31.3) 13 (18.3)

No. (%) of patients with any symptomatic treatment for any IRRd 26 (38.8) 25 (35.2)

Paracetamol 9 (34.6) 3 (12.0)

Diphenhydramine hydrochloride 4 (15.4) 6 (24.0)

Chlorpheniramine 4 (15.4) 3 (12.0)

(4) No. (%) of patients with at least one AEe 125 (43.4) 120 (41.2)

Total no. of AEs 224 228

Total no. of deaths 00

Total no. (%) of patients with at least one:

AE with fatal outcome 00

Serious AE 3 (1.0) 3 (1.0)

Serious AE leading to withdrawal from OCR treatment 1 (0.3) 0

Serious AE leading to OCR temporary delay 1 (0.3) 0

AE leading to withdrawal from OCR treatment 1 (0.3) 0

AE leading to OCR temporary delay/dose interruption 2 (0.7) 4 (1.4)

IRRs leading to withdrawal from OCR treatment at the first randomized dose 00

IRRs leading to withdrawal from OCR treatment at any randomized dose 00

Abbreviations: AE = adverse event; IRR = infusion-related reaction; ITT = intent-to-treat; OCR = ocrelizumab. All patients, ITT population. a The stratified estimated difference between the proportions in the 2 randomized groups is the weighted average of the proportion difference across strata (region and dose at which the patient is randomized) based on Cochran-Mantel-Haenszel weights. b Most frequent symptoms, i.e., in ≥10% of patients with IRRs. c Percentages of patients with any symptoms are based on the number of patients with any IRR. d Percentages of patients with any symptomatic treatments are based on the number of patients with any symptomatic treatment for any IRR. e Summaries of safety data were performed using the safety population, which included all randomized patients who received any dose or a part of a dose of ocrelizumab (N = 291 shorter infusion group and N = 288 conventional infusion group). One patient in the conventional infusion group who did not receive treatment was excluded from the safety population. Two patients in the shorter infusion group received the wrong treatment they were randomized to, sothe N for ITT differs from the N for safety population.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 5 respectively. Most IRRs were mild (grade 1; n = 46/67 [68.7%] Study funding and n = 47/71 [66.2%]) or moderate (grade 2; n = 21/67 This research was funded by F. Hoffmann-La Roche Ltd, [31.3%] and n = 23/71 [32.4%]) in the conventional and Basel, Switzerland. This manuscript was submitted by the shorter infusion groups, respectively. One severe IRR (grade 3) author on behalf of the ENSEMBLE PLUS principle inves- occurred in the shorter (fatigue at first randomized dose) and in tigators. The author had full editorial control of the manu- the conventional (laryngeal inflammation at second randomized script and provided their final approval of all content. The dose) infusion groups. No IRRs were life-threatening, serious, first author wrote the first draft of the manuscript with as- or fatal, and >98% (136/138) of all IRRs resolved without sistance from an independent medical writer funded by F. sequelae in both groups. The most frequent symptoms associ- Hoffmann-La Roche Ltd. Medical editorial assistance was ated with IRRs in both groups were throat irritation, fatigue, and provided by Terence Smith of Articulate Science and was headache, with the most common treatments being para- funded by F. Hoffmann-La Roche Ltd. cetamol, diphenhydramine hydrochloride, and chlorpheniramine (table). Overall, n = 14 of 291 (4.8%) and n = 25 of 289 Disclosure (8.7%) patients in the conventional and shorter infusion groups H-P Hartung has received honoraria for serving on steering had IRRs leading to temporary infusion interruption/slowing, and data monitoring committees from Bayer Healthcare, respectively (table). There was no correlation between peak Biogen, Celgene BMS, GeNeuro, MedImmune, Merck, serum OCR concentration and observed IRRs. Novartis, Roche, Sanofi Genzyme, Teva, TG Therapeutics, and Viela Bio with approval by the Rector of Heinrich-Heine Adverse events University Düsseldorf. Go to Neurology.org/NN for full fi Overall, the AE pro le between conventional and shorter disclosures. infusion groups was balanced; the most common AEs in either infusion group were IRRs (table). One patient (0.3%) Publication history from the conventional infusion group withdrew because of an Received by Neurology: Neuroimmunology & Neuroinflammation AE (depressive symptom). Serious AEs occurred in n = 3 April 20, 2020. Accepted in final form May 19, 2020. (1.0%) patients in both groups (conventional infusion: events, n = 1 typhoid fever and intraductal papilloma of breast, n = 2 depressive symptoms; shorter infusion: events, n = 1 urinary tract infection, peripheral edema, and hypotension). Appendix 1 Author Name Location Contribution

H-P Department of Neurology, Design and conceptualized Discussion Hartung UKD, Center of Neurology and study, reviewed the data, Neuropsychiatry and LVR- and drafted and revised the This primary analysis of the ENSEMBLE PLUS study shows Klinikum, manuscript content. Heinrich-Heine University that the frequency, severity, and symptoms of IRRs were similar Dusseldorf,¨ Dusseldorf,¨ between conventional and shorter OCR infusion periods. Germany During the first randomized dose, there was a moderately higher incidence of IRRs leading to infusion slowing/ interruption in the shorter (25 of 289 patients; 8.7%) vs conventional (14 of 291 patients; 4.8%) infusion group. Overall, – AEs were consistent with the known safety profile of OCR,7 9 and no new safety signals were observed with a shorter infusion time. The safety profile of OCR remains unchanged. Short- Appendix 2 Coinvestigators ening the OCR infusion time to 2 hours may reduce the total Investigator Country Role Contribution site stay time and lessen the burden on patients and site staff, which is of particular importance in light of the current coro- Hans-Peter Germany Head of Steering Design and Hartung (Head Committee and conceptualized navirus disease 2019 pandemic. of Steering Site Investigator study, reviewed Committee) (Neurologist) the data, and drafted and revised the Acknowledgment manuscript The author thanks all patients, their families, and the content. As a member of the investigators who participated in this trial (including the Steering ENSEMBLE PLUS study Steering Committee, which pro- Committee, participated in vided study oversight). The author also thanks the in- study oversight dependent data monitoring committee for performing data and led and coordinated analysis and safety monitoring, and is grateful to Regine communication Buffels and Jad Abdul Samad (of F. Hoffmann-La Roche Ltd) among sites for additional critical review of this manuscript and technical within their country. advice.

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Appendix 2 (continued) Appendix 2 (continued)

Investigator Country Role Contribution Investigator Country Role Contribution

Thomas Berger Austria Steering As a member of Bruno Brochet France Steering As a member of Committee the Steering Committee the Steering member and Committee, member and Committee, Site Investigator participated in Site Investigator participated in (Neurologist) study oversight (Neurologist) study oversight and led and and led and coordinated coordinated communication communication among sites among sites within within their their country. country. Carlos Nos Spain Steering As a member of Timothy USA Steering As a member of Committee the Steering Vollmer Committee the Steering member and Committee, member and Committee, Site Investigator participated in Site Investigator participated in (Neurologist) study oversight (Neurologist) study oversight and led and and led and coordinated coordinated communication communication among sites among sites within their within their country. country. Francesco Patti Italy Steering As a member of Robert Bermel USA Steering As a member of Committee the Steering Committee the Steering member and Committee, member and Committee, Site Investigator participated in Site Investigator participated in (Neurologist) study oversight (Neurologist) study oversight and led and and led and coordinated coordinated communication communication among sites among sites within within their their country. country.

Amy Perrin Ross USA Steering As a member of Rana Turkey Steering As a member of Committee the Steering Karabudak Committee the Steering member (MS Committee, member and Committee, Nurse Specialist) participated in Site Investigator participated in study oversight. (Neurologist) study oversight and led and Mark Freedman Canada Steering As a member of coordinated Committee the Steering communication member and Committee, among sites Site Investigator participated in within their (Neurologist) study oversight country. and led and coordinated William Carroll Australia Steering As a member of communication Committee the Steering among sites member and Committee, within their Site Investigator participated in country. (Neurologist) study oversight and led and Ludo Belgium Steering As a member of coordinated Vanopdenbosch Committee the Steering communication member and Committee, among sites Site Investigator participated in within their (Neurologist) study oversight country. and led and coordinated Trygve Holmoy Norway Steering As a member of communication Committee the Steering among sites member and Committee, within their Site Investigator participated in country. (Neurologist) study oversight and led and Joep Killestein The Steering As a member of coordinated Netherlands Committee the Steering communication member and Committee, among sites Site Investigator participated in within their (Neurologist) study oversight country. and led and coordinated Jens Wuerfel Switzerland Steering As a member of communication (MRI, not Committee the Steering among sites involved in member (MRI Committee, within their ENSEMBLE Plus) Specialist) participated in country. study oversight.

Continued

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 7 3. Pritchard CH, Greenwald MW, Kremer JM, et al. Safety of infusing rituximab at Appendix 2 (continued) a more rapid rate in patients with rheumatoid arthritis: results from the RATE-RA study. BMC Musculoskelet Disord 2014;15:177. 4. Sehn LH, Donaldson J, Filewich A, et al. Rapid infusion rituximab in combination Investigator Country Role Contribution with corticosteroid-containing chemotherapy or as maintenance therapy is well tolerated and can safely be delivered in the community setting. Blood 2007;109: Ralph Benedict USA Steering As a member of 4171–4173. (Cognition, not Committee the Steering 5. Tuthill M, Crook T, Corbet T, King J, Webb A. Rapid infusion of rituximab over 60 involved in member and Committee, min. Eur J Haematol 2009;82:322–325. ENSEMBLE Plus) Site Investigator participated in 6. Polman CH, Reingold SC, Banwell B, et al. Diagnostic criteria for multiple (Cognition study oversight. sclerosis: 2010 revisions to the McDonald criteria. Ann Neurol 2011;69: Specialist) 292–302. 7. Mayer L, Kappos L, Racke MK, et al. Ocrelizumab infusion experience in patients with relapsing and primary progressive multiple sclerosis: results from the phase 3 randomized OPERA I, OPERA II, and ORATORIO studies. Mult Scler Relat Disord 2019;30:236–243. References 8. Hauser SL, Bar-Or A, Comi G, et al. Ocrelizumab versus interferon beta-1a in re- 1. Genentech. Ocrevus (package insert). South San Francisco: Genentech, Inc.; 2019. lapsing multiple sclerosis. N Engl J Med 2017;376:221–234. 2. Roche. Ocrevus (summary of product characteristics). Welwyn Garden City: Roche 9. Montalban X, Hauser SL, Kappos L, et al. Ocrelizumab versus placebo in primary Products Limited; 2019. progressive multiple sclerosis. N Engl J Med 2017;376:209–220.

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN ARTICLE OPEN ACCESS Altered fovea in AQP4-IgG–seropositive neuromyelitis optica spectrum disorders

Seyedamirhosein Motamedi, MSc, Frederike C. Oertel, MD, Sunil K. Yadav, PhD, Ella M. Kadas, PhD, Correspondence Margit Weise, MSc, Joachim Havla, MD, Marius Ringelstein, MD, Orhan Aktas, MD, Philipp Albrecht, MD, Dr. Brandt [email protected] Klemens Ruprecht, MD, Judith Bellmann-Strobl, MD, Hanna G. Zimmermann, PhD, Friedemann Paul, MD, and Alexander U. Brandt, MD

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e805. doi:10.1212/NXI.0000000000000805 Abstract Objective To investigate disease-speciﬁc foveal shape changes in patients with neuromyelitis optica spectrum disorders (NMOSDs) using foveal morphometry.

Methods This cross-sectional study included macular spectral domain optical coherence tomography scans of 52 eyes from 28 patients with aquaporin-4 immunoglobulin G (AQP4-IgG)-seropositive NMOSD, 116 eyes from 60 patients with MS, and 123 eyes from 62 healthy controls (HCs), retrospectively, and an independent conﬁrmatory cohort comprised 33/33 patients with NMOSD/MS. The fovea was characterized using 3D foveal morphometry. We included peripapillary retinal nerve ﬁber layer (pRNFL) thickness and combined macular ganglion cell and inner plexiform layer (GCIPL) volume to account for optic neuritis (ON)-related neuroaxonal damage.

Results Group comparison showed significant differences compared with HC in the majority of foveal shape parameters in NMOSD, but not MS. Pit flat disk area, average pit flat disk diameter, inner rim volume, and major slope disk length, as selected parameters, showed differences between NMOSD and MS (p value = 0.017, 0.002, 0.005, and 0.033, respectively). This effect was independent of ON. Area under the curve was between 0.7 and 0.8 (receiver operating characteristic curve) for discriminating between NMOSD and MS. Pit flat disk area and average pit flat disk diameter changes independent of ON were confirmed in an independent cohort.

Conclusions Foveal morphometry reveals a wider and ﬂatter fovea in NMOSD in comparison to MS and HC. Comparison to MS and accounting for ON suggest this eﬀect to be at least in part independent of ON. This supports a primary retinopathy in AQP4-IgG–seropositive NMOSD.

From the Experimental and Clinical Research Center (S.M., F.C.O., J.B.-S., H.G.Z., F.P., A.U.B.), Max-Delbruck¨ Center for Molecular Medicine and Charité - Universitätsmedizin Berlin, corporate member of Freie Universität Berlin, Humboldt-Universität zu Berlin, and Berlin Institute of Health; NeuroCure Clinical Research Center (S.M., F.C.O., S.K.Y., E.M.K., J.B.-S., H.G.Z., F.P., A.U.B.), Charité – Universitätsmedizin Berlin, corporate member of Freie Universität Berlin, Humboldt-Universität zu Berlin, and Berlin Institute of Health, Germany; Division of Neuroinflammation and Glial Biology (F.C.O.), University of California, San Francisco; Nocturne GmbH (S.K.Y., E.M.K.), Berlin; Department of Neurology (M.W., M.R., O.A., P.A.), Medical Faculty, Heinrich Heine University, Dusseldorf;¨ Institute of Clinical Neuroimmunology (J.H.), LMU Hospital, Ludwig-Maximilians University, Munich; Department of Neurology (M.R.), Center for Neurology and Neuropsychiatry, LVR-Klinikum Dusseldorf;¨ Department of Neurology (K.R., F.P.), Charité - Universitätsmedizin Berlin, corporate member of Freie Universität Berlin, Humboldt-Universität zu Berlin, and Berlin Institute of Health, Germany; and Department of Neurology (A.U.B.), University of California, Irvine.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary AQP4-IgG = aquaporin-4 immunoglobulin G; ART = automatic real time; AUC = area under the curve; B =estimate;FT =foveal thickness; GCIPL = combined macular ganglion cell and inner plexiform layer; HC = healthy control; ILM = inner limiting membrane; INL =innernuclearlayer;MOG = myelin oligodendrocyte glycoprotein; NMOSD = neuromyelitis optica spectrum disorder; OCT = optical coherence tomography; ON = optic neuritis; ON− = eyes without a history of ON; ON+ = eyes with a history of ON; pRNFL = peripapillary retinal nerve ﬁber layer; ROC = receiver operating characteristic; SE =standarderrorofB.

Aquaporin-4 immunoglobulin G (AQP4-IgG)-seropositive refractive error >6 diopters), which can affect the retina.19 neuromyelitis optica spectrum disorder (NMOSD) is an in- Eyes with an episode of ON within the last 6 months before flammatory astrocytopathy defined by pathogenic serum im- the OCT examinations were excluded. Of 46 patients – munoglobulin G antibodies against aquaporin-4.1 3 enclosed in the study, we included 28 patients with NMOSD in the analysis after applying the inclusion and exclusion cri- Optic neuritis (ON) is a hallmark of NMOSD and leads to teria (table 1). We additionally included 60 patients with severe neuroaxonal damage in optic nerve and retina associ- relapsing-remitting MS according to the 2010 revised – ated with oftentimes severe vision loss.4 8 Retinal optical McDonald criteria,20 from 2 cohort studies about MS and coherence tomography (OCT) can be used to measure this clinically isolated syndrome and 62 HCs, both groups age and – damage9 12: Peripapillary retinal nerve fiber layer (pRNFL) sex matched to the NMOSD cohort, in this study (table 1). and combined macular ganglion cell and inner plexiform layer Data from 17 patients with AQP4-IgG–seropositive NMOSD (GCIPL) typically become thinner, whereas inner nuclear (61%) were already included in a previous study by Oertel – layer (INL) becomes thicker as a result of ON.6,13 15 et al.17 High-contrast visual acuity was measured using Early Treatment in Diabetes Retinopathy Study charts at a 4-m Recently, a foveal thickness (FT) reduction has been reported distance with an Optec 6500 P system (Stereo Optical, Chi- in eyes never experiencing an ON in patients with AQP4- cago, IL), with best correction and under photopic conditions. IgG–seropositive NMOSD,16,17 suggesting either subclinical optic nerve inflammation or primary retinal astrocytopathy in Aconfirmatory cohort consisting of macular OCTs from 58 NMOSD.8 This change in FT appeared to be driven by eyes of 33 patients with AQP4-IgG–seropositive NMOSD a change in foveal shape, with a normally V-shaped fovea (eyes with a history of ON [ON+]: 27; 33 women; age: 49.2 ± appearing more widened and U-shaped with flattened disk in 15.4 years) and 62 eyes of 33 patients with MS (ON+: 12; 32 eyes of patients with AQP4-IgG–seropositive NMOSD.17 women; age: 49.7 ± 14.7 years) from longitudinal prospective observational cohort studies at the Department of Neurology, Because FT is a weak measure for foveal shape, we developed Universitätsklinikum Düsseldorf at Heinrich Heine University, a 3D foveal morphometry method, which we previously de- Düsseldorf, Germany, was included in this study, following the scribed and validated in detail.18 Here, we use this approach to same inclusion and exclusion criteria. MS and NMOSD groups investigate the foveal shape in patients with AQP4- were well matched in this cohort for age (p =0.812)andsex(p IgG–seropositive NMOSD. We compare findings against = 1), but not for the proportion of eyes with ON (p =0.001). measurements in patients with MS, which also presents with The NMOSD group is well matched to the Berlin cohort for ON, and against healthy controls (HCs). Our goal was to age (p =0.113),sex(p = 0.214), and ON+ (p =0.507). investigate whether foveal changes are characteristic to AQP4- IgG–seropositive NMOSD and not simply caused by ON. Ethics statement The study was approved by the local ethics committee at Charité—Universitätsmedizin Berlin (EA1/131/09, EA1/163/ Methods 12, and EA1/182/10). The confirmatory OCT data were collected under approval from the local ethics committee at Study population Heinrich Heine University Düsseldorf (4389R). The study was In this analysis, we retrospectively included data from an conducted according to the Declaration of Helsinki in its cur- ongoing observational cohort study in patients with NMOSD rently applicable version and the applicable German and Eu- at the NeuroCure Clinical Research Center at Charité— ropean laws. All the participants gave written informed consent. Universitätsmedizin Berlin, Germany, acquired from August 2013 to November 2016. Inclusion criteria were a minimum age Optical coherence tomography of 18 years and fulfilling the diagnostic criteria for AQP4- All retinal OCT images (exploratory and confirmatory co- IgG–seropositive NMOSD according to the 2015 International hort) were taken using Spectralis spectral domain OCT Consensus Diagnostic Criteria.7 AQP4-IgG–seropositivity devices from Heidelberg Engineering (Heidelberg, Ger- was tested using a cell-based assay (Euroimmun, Lübeck, many), with activated eye tracker and automatic real-time Germany). Exclusion criteria were any other neurologic or (ART) averaging. The pRNFL thickness was calculated using ophthalmologic disorder (e.g., glaucoma, diabetes, and standard ring scans around the optic nerve head (12°, single B

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Table 1 Demographic description of NMOSD, MS, and HC cohorts

NMOSD MS HC

No. of patients (N) 28 60 62

No. of eyes (N) 52 116 123

Sex (female) (N [%])a 26 (93%) 55 (92%) 56 (90%)

Age (y) (mean ± SD)b 43.6 ± 11.5 39.0 ± 10.9 41.7 ± 13.5

Patients with a history of ON (N [%]) 16 (57%) 29 (48%) —

Eyes with a history of ON (N [%])c 20 (38%) 32 (28%) —

Number of ONs per eye (median [range]) 1.5 (1–8) 1 (1–2) —

VA for ON eyes (logMAR) (mean ± SD)d 0.36 ± 0.69 −0.09 ± 0.10

EDSS score (median [range]) 3(0–6.5) 2 (0–4.5) —

Disease duration (y) (mean ± SD) 6.8 ± 4.8 7.9 ± 8.4 — pRNFL (μm) (mean ± SD) 81.2 ± 22.1 90.8 ± 15.3 97.6 ± 8.8

GCIPL (mm3) (mean ± SD) 1.69 ± 0.29 1.83 ± 0.24 1.94 ± 0.14

INL (mm3) (mean ± SD) 0.94 ± 0.09 0.96 ± 0.06 0.95 ± 0.06

FT (μm) (mean ± SD) 262.9 ± 14.9 272.1 ± 20.3 272.3 ± 23.1

Abbreviations: EDSS = Expanded Disability Status Scale; FT = foveal thickness; GCIPL = combined macular ganglion cell and inner plexiform layer volume; HC = healthy controls; INL = inner nuclear layer volume; logMAR = logarithm of the minimum angle of resolution; NMOSD = neuromyelitis optica spectrum disorders; ON = optic neuritis; pRNFL = peripapillary retinal nerve fiber layer thickness; VA = high-contrast visual acuity. a Sex match: p value = 1. b Age match: p value: HC vs MS = 0.382, HC vs NMOSD = 0.437, MS vs NMOSD = 0.056. c ON match: p value = 0.199. d VA measurements for 25 (78%) ON eyes of patients with MS and 17 (85%) ON eyes of patients with NMOSD were available. scan with 1536 A scans, 16 ≤ ART ≤100). The volume of the points with the maximum slopes in the parafoveal area; and pit GCIPL and INL was calculated in a 6-mm diameter around flat disk, which characterizes the flatness of the foveal pit. Each the fovea, and FT was measured in a 1-mm-diameter area surface is described by 4 parameters—the length in the around the fovea, based on macular volume scans (25° × 30°, dominant direction (major axis): major length; the length in 61 B scans with 768 A scans per each B scan, ART = 15). the second dominant direction (perpendicular to the major Intraretinal layer segmentation was performed and corrected axis): minor length; the area; and the average diameter. In if needed using Heidelberg Eye Explorer (HEYEX version addition, the distance between the fovea (the minimum point) 1.9.10.0) by an experienced grader. All OCT scans were and the center of rim disk: average pit depth; the distance quality controlled according to the OSCAR-IB criteria,19,21 between the fovea and the reference plane: central foveal and OCT data are reported in accordance with the Advised thickness; the average height of the rim points: average rim Protocol for OCT Study Terminology and Elements height; the volume between the ILM surface and the reference (APOSTEL) recommendations.22 Four eyes from the ex- plane: rim volume; the volume between the ILM surface and ploratory cohort were excluded from the study because of rim disk: pit volume; the volume between the ILM surface and inadequate OCT scan quality. the reference plane within 1-mm-diameter cylinder centered at the fovea: inner rim volume; and the average slope at the Foveal morphometry maximum slope points: average maximum pit slope are mea- All macular volume OCTs were analyzed using a 3D foveal sured by this method to characterize the 3D foveal shape. 18 morphometry described previously in detail. In brief, first, Figure 1 gives an overview of the method and the defined the 3D macular scan is flattened based on the segmentation of parameters. Test-retest reliability was excellent in all foveal the Bruch membrane as the reference plane, and then the morphometry parameters with intraclass correlation coef- inner limiting membrane (ILM) surface is smoothed and ficients >0.9 in all cases (table e-1, links.lww.com/NXI/A270). reconstructed radially using a cubic Bezier polynomial model. Based on the reconstructed ILM surface, several parameters Statistical analysis are defined to describe the foveal and parafoveal shape. Three Sex and ON differences between groups were tested using χ2 surfaces are defined in this foveal shape analysis method: rim tests, and age differences were tested using 2-sample Wil- disk, which connects the points on the surface with the max- coxon tests. Performance measurements were based on the imum height (rim points); slope disk, which connects the area under the curve (AUC) for receiver operating

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 Figure 1 Three-dimensional foveal shape analysis method overview

(A) ILM surface smoothing and radial reconstruction using the cubic Bezier polynomial. (B) Rim height, average pit depth, and central foveal thickness. (C) Rim disk (blue), slope disk (red), and pit flat disk (green) and major and minor axes on each surface. (D) Rim volume, pit volume, and inner rim volume. APD = average pit depth, CFT = central foveal thickness; ILM = inner limiting membrane; major = major axis; minor = minor axis.

characteristic (ROC) curves. All linear regression analyses Results were performed using linear mixed-effect models including intereye within-patient correlations, age, and sex as random Foveal shape changes in NMOSD and MS effects. The marginal and conditional coefficients of de- First, we analyzed foveal shape in patients with NMOSD and termination of the linear models were calculated with pseudo MS and compared results with measurements in HCs. Foveal R-squared. Stepwise logistic regression analysis for model se- shape was altered in patients with NMOSD, but only mildly lection was performed by the Akaike Information Criteria with affected in patients with MS, both in comparison to HCs (table both backward and forward modes of stepwise search based 2). Foveal parameters stratified by history of ON are included on generalized linear models. In this exploratory study, we in supplemental data (table e-2, links.lww.com/NXI/A271). corrected p values for multiple testing using the Benjamini- Hochberg procedure. In addition, the identified parameter In contrast, both patients with MS and NMOSD showed differences were tested in a second independent cohort neuroaxonal damage typically occurring after ON: pRNFL and obtained with the same scanning protocol at a different center. GCIPL were lower in patients with NMOSD in comparison to One-sided p values were reported for the confirmatory cohort, HCs (pRNFL: standard error of B [SE] = −17.7 [3.0] μm, 3 not corrected for multiple testing. Sample size for the confir- p < 0.001, GCIPL: B [SE] = −0.27 [0.04] mm , p < 0.001), but matory cohort was calculated for 1-sided 2-sample t-test with also in patients with MS in comparison to HCs (pRNFL: B 90% power and a significance level of 0.05. To adjust this [SE] = −7.0 (2.0) μm, p < 0.001, GCIPL: B [SE] = −0.12 [0.03] 3 estimate for eye-based statistics, we added 60% sample size to mm , p = 0.001). account for intereye within-patient effects and additional covariates. All statistical analysis were performed in R version Parameter selection 3.5.023 with packages stats, lme4, lmerTest, MuMIn, ROCR, Next, we selected parameters with the highest potential to be ggplot2, plotROC, pwr, multcomp, and ggpubr packages. The abnormal in NMOSD: We therefore analyzed parameter p values less than 0.05 were considered significant. performance in discriminating between eyes from patients with NMOSD and MS, regardless of ON status (table 3 and Data availability figure 2A). This was followed by a linear regression analysis All data are available on reasonable request from the corre- against diagnosis, history of ON, and their interaction effect to sponding author. derive effect sizes and group differences accounting for ON.

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Table 2 Foveal shape analysis parameters results (mean ± SD) and linear regression analysis for NMOSD and MS vs HC

HC vs NMOSD HC vs MS HC (mean ± MS (mean ± NMOSD (mean ± SD) SD) SD) B (SE) p B (SE) p

Average pit depth (mm) 0.117 ± 0.021 0.111 ± 0.018 0.101 ± 0.026 −0.017 <0.001 −0.006 (0.003) 0.078 (0.005)

Central foveal thickness (mm) 0.231 ± 0.015 0.229 ± 0.017 0.228 ± 0.017 −0.004 0.330 −0.002 (0.003) 0.535 (0.004)

Average rim height (mm) 0.348 ± 0.014 0.340 ± 0.016 0.328 ± 0.018 −0.021 <0.001 −0.009 (0.002) <0.001 (0.003)

Average rim disk diameter (mm) 2.184 ± 0.115 2.152 ± 0.110 2.132 ± 0.130 −0.056 0.037 −0.034 (0.020) 0.082 (0.027)

Rim disk area (mm2) 3.717 ± 0.387 3.606 ± 0.371 3.545 ± 0.436 −0.184 0.041 −0.116 (0.066) 0.079 (0.090)

Major rim disk length (mm) 0.630 ± 0.066 0.613 ± 0.063 0.600 ± 0.075 −0.032 0.039 −0.018 (0.011) 0.112 (0.015)

Minor rim disk length (mm) 0.619 ± 0.065 0.599 ± 0.062 0.590 ± 0.072 −0.030 0.042 −0.021 (0.011) 0.055 (0.015)

Average slope disk diameter (mm) 0.663 ± 0.119 0.653 ± 0.152 0.771 ± 0.136 0.104 (0.028) <0.001 −0.012 (0.025) 0.626

Slope disk area (mm2) 0.361 ± 0.131 0.358 ± 0.180 0.486 ± 0.164 0.120 (0.032) <0.001 −0.005 (0.028) 0.858

− Major slope disk length (mm) 0.068 ± 0.025 0.068 ± 0.034 0.090 ± 0.029 0.021 (0.006) <0.001 −4.8e 4 (0.005) 0.930

Minor slope disk length (mm) 0.053 ± 0.019 0.052 ± 0.026 0.072 ± 0.026 0.019 (0.005) <0.001 −0.001 (0.004) 0.772

Average pit flat disk diameter 0.215 ± 0.030 0.211 ± 0.039 0.257 ± 0.052 0.042 (0.008) <0.001 −0.005 (0.006) 0.440 (mm)

Pit flat disk area (mm2) 0.037 ± 0.010 0.036 ± 0.015 0.054 ± 0.025 0.017 (0.003) <0.001 −0.001 (0.002) 0.691

Major pit flat disk length (mm) 0.0067 ± 0.0065 ± 0.0098 ± 0.0048 0.0032 <0.001 −0.0001 0.725 0.0018 0.0027 (0.0007) (0.0004)

Minor pit flat disk length (mm) 0.0058 ± 0.0056 ± 0.0083 ± 0.0036 0.0026 <0.001 −0.0002 0.628 0.0016 0.0023 (0.0005) (0.0004)

Rim volume (mm3) 1.045 ± 0.153 0.983 ± 0.133 0.910 ± 0.170 −0.141 <0.001 −0.061 (0.025) 0.013 (0.036)

Inner rim volume (mm3) 0.104 ± 0.018 0.103 ± 0.019 0.088 ± 0.015 −0.017 <0.001 −0.001 (0.003) 0.753 (0.004)

Pit volume (mm3) 0.252 ± 0.043 0.246 ± 0.051 0.259 ± 0.044 0.005 (0.010) 0.606 −0.007 (0.008) 0.402

Average maximum pit slope 12.16 ± 3.38 11.10 ± 2.41 9.86 ± 3.11 −2.42 (0.74) 0.001 −1.03 (0.52) 0.047 (degrees)

Abbreviations: B = estimate; HC = healthy controls; MS = patients with MS; NMOSD = patients with neuromyelitis optica spectrum disorders; SE = standard error of B. Significant p values are marked in bold.

Table 3 shows the results of this selection process, ordered by further investigate this, we repeated the AUC analysis, but this AUC. The best parameter selected from the ROC analysis was time separately for the eyes without a history of ON (ON−)and pit flat disk area (AUC = 0.798, figure 2A). To derive a final set the ON+ (table e-2, links.lww.com/NXI/A271). Indeed, foveal of relevant parameters, we computed a stepwise logistic re- shape was altered also in the ON−. Here, the best foveal shape gression model to predict NMOSD vs MS, including only the parameter to distinguish ON− from patients with NMOSD and parameters with AUC ≥0.7. This selected 4 parameters: pit patients with MS was minor pit flat disk length (AUC = 0.804, flat disk area, average pit flat disk diameter, inner rim volume, figure 2B). In ON+, the best-performing parameter to discrim- and major slope disk length (figure 2, D–G). inate between patients with NMOSD and patients with MS was major pit flat disk length (AUC = 0.817, figure 2C). Of note, Association with ON and neuroaxonal damage NMOSD ON− also showed signs of mild neuroaxonal damage A crucial question is whether these parameters react to ON- compared with HC (pRNFL: B [SE] = −5.7 [2.4] μm, p = 0.017; related damage or are indeed at least partially independent. To GCIPL: B [SE] = −0.12 [0.04] mm3, p = 0.001).

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 5 Table 3 AUC and linear regression analysis results for NMOSD vs MS, sorted in ascending order of AUC

Linear regression with interaction effects of diagnosis and ON history

MS vs NMOSD ON2 vs ON+ NMOSD × ON NMOSD vs MS 2 2 (AUC) B (SE) p B(SE) p B (SE) p RMarg RCond

Pit flat disk area (mm2) 0.798 0.011 (0.005) 0.021 0.001 (0.002) 0.544 0.015 (0.004) 0.001 0.176 0.843

Average pit flat disk diameter 0.796 0.033 (0.011) 0.002 0.004 (0.005) 0.423 0.029 (0.009) 0.002 0.200 0.876 (mm)

Minor pit flat disk length (mm) 0.796 0.0018 0.011 0.0002 0.645 0.0020 0.007 0.172 0.827 (0.0007) (0.0003) (0.0007)

Major pit flat disk length (mm) 0.790 0.0019 0.038 0.0003 0.456 0.0031 <0.001 0.177 0.855 (0.0009) (0.0004) (0.0008)

Inner rim volume (mm3) 0.755 −0.012 0.006 −0.002 0.170 −0.007 0.006 0.144 0.940 (0.004) (0.001) (0.003)

Minor slope disk length (mm) 0.744 0.017 (0.006) 0.013 0.001 (0.002) 0.554 0.008 (0.004) 0.051 0.111 0.941

Slope disk area (mm2) 0.739 0.102 (0.042) 0.022 0.010 (0.011) 0.371 0.053 (0.022) 0.022 0.096 0.958

Average slope disk diameter 0.739 0.094 (0.035) 0.010 0.006 (0.009) 0.509 0.050 (0.019) 0.010 0.116 0.956 (mm)

Major slope disk length (mm) 0.733 0.017 (0.008) 0.040 0.002 (0.002) 0.265 0.010 (0.004) 0.019 0.083 0.963

Average rim height (mm) 0.664 −0.008 0.022 −0.007 0.001 −0.009 0.022 0.107 0.882 (0.003) (0.002) (0.004)

Rim volume (mm3) 0.627 −0.053 0.124 −0.061 <0.001 −0.044 0.164 0.081 0.878 (0.032) (0.016) (0.032)

Average pit depth (mm) 0.602 −0.007 0.147 −0.007 <0.001 −0.009 0.025 0.069 0.931 (0.005) (0.002) (0.004)

Pit volume (mm3) 0.594 0.012 (0.011) 0.369 −0.006 0.239 0.002 (0.008) 0.845 0.013 0.917 (0.004)

Average maximum pit slope 0.593 −0.74 (0.59) 0.210 −0.73 (0.26) 0.008 −1.24 (0.51) 0.019 0.073 0.908 (degrees)

Major rim disk length (mm) 0.556 −0.010 0.671 −0.021 0.002 −0.005 0.697 0.025 0.902 (0.015) (0.006) (0.013)

Average rim disk diameter 0.550 −0.014 0.657 −0.039 0.002 −0.010 0.657 0.027 0.891 (mm) (0.026) (0.012) (0.023)

Rim disk area (mm2) 0.549 −0.043 0.628 −0.128 0.002 −0.038 0.628 0.027 0.892 (0.088) (0.039) (0.077)

Minor rim disk length (mm) 0.545 −0.004 0.761 −0.022 0.002 −0.008 0.756 0.029 0.878 (0.015) (0.007) (0.013)

− Central foveal thickness (mm) 0.543 −0.001 0.891 1.5e 4 0.891 −0.001 0.891 0.002 0.946 (0.004) (0.001) (0.002)

Abbreviations: AUC = area under the curve; B = estimate; MS = patients with MS; NMOSD = patients with neuromyelitis optica spectrum disorders; NMOSD × ON = interaction effect of diagnosis and ON; ON = optic neuritis; ON− = eyes without a history of ON; ON+ = eyes with a history of ON; SE = standard error of B; 2 2 RCond = conditional R-squared; RMarg = marginal R-squared. Significant p values and AUC ≥0.7 are marked in bold.

Foveal changes may also be driven not by ON per se, but by considered sensitive parameters for ON severity. Here, it was the amount of neuroaxonal damage after ON. Here, MS may confirmed that the observed group differences are unlikely to be a problematic control group because the amount of ON- be explained by ON severity differences alone. related retinal damage is typically lesser than in NMOSD.10,24 To investigate whether differences in neuroaxonal damage are See supplemental data for more results on regression analysis indeed able to explain the observed foveal differences, we corrected for the neuroaxonal damage (table e-3, links.lww. repeated the linear regression model analyses for the pre- com/NXI/A272). Figure 2, H–I shows rim volume and av- viously selected 4 parameters, but this time corrected addi- erage pit depth as example foveal shape parameters signifi- tionally for GCIPL and INL (table 4), which can be cantly dependent on ON status but not on diagnosis. Figure 2,

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Figure 2 ROC curves, exploratory data analysis for selected parameters, and sample fovea

ROC curves for best-performing foveal shape and standard OCT parameters discriminating between (A) NMOSD vs MS, (B) NMOSD ON− vs MS ON−, and (C) NMOSD ON + vs MS ON+. Box and dot plots for (D) pit flat disk area, (E) average pit flat disk diameter, (F) inner rim volume, and (G) major slope disk length, the selected 4 foveal shape parameters. (H) Rim volume and (I) average pit depth, as example foveal shape parameters affected by ON but not diagnosis. A sample central (foveal) B scan of (J) MS ON− and (K) NMOSD ON−, chosen from the median of the selected pit flat disk parameters in each group, demonstrating the difference in foveal pit (pit flat disk), between NMOSD and MS in eyes without a history of ON. AUC = area under the curve; FT = foveal thickness; pRNFL = peripapillary retinal nerve fiber layer thickness; INL = inner nuclear layer volume; HC = healthy controls; MS = patients with MS; NMOSD = patients with neuromyelitis optica spectrum disorders; ON = optic neuritis; ON− = eyes without a history of ON; ON+ = eyes with a history of ON; ROC = receiver operating characteristic.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 7 Table 4 Linear regression results for the selected foveal shape analysis parameters corrected for GCIPL and INL

Linear regression with interaction effects of diagnosis and ON history corrected for GCIPL

MS vs NMOSD ON2 vs ON+ GCIPL (mm3) NMOSD × ON

2 2 B (SE) p B(SE) p B (SE) p B (SE) p RMarg RCond

− Pit flat disk area (mm2) 0.010 0.031 −3.2e 4 0.886 −0.016 0.010 0.012 0.010 0.212 0.858 (0.005) (0.002) (0.006) (0.004)

− Average pit flat disk diameter 0.030 0.006 −8.3e 5 0.986 −0.038 0.006 0.023 0.019 0.230 0.889 (mm) (0.011) (0.005) (0.013) (0.009)

− Inner rim volume (mm3) −0.011 0.013 −7.7e 4 0.554 0.010 0.013 −0.006 0.038 0.159 0.945 (0.004) (0.001) (0.004) (0.003)

− Major slope disk length (mm) 0.017 0.051 0.002 (0.002) 0.360 5.5e 5 0.992 0.010 0.033 0.082 0.962 (0.008) (0.006) (0.004)

Linear regression with interaction effects of diagnosis and ON history corrected for INL

MS vs NMOSD ON2 vs ON+ INL (mm3) NMOSD × ON

2 2 B (SE) p B (SE) p B (SE) p B (SE) p RMarg RCond

Pit flat disk area (mm2) 0.012 0.017 0.001 0.771 0.041 0.145 0.013 0.014 0.193 0.851 (0.005) (0.002) (0.024) (0.004)

Average pit flat disk diameter 0.035 0.005 0.003 0.482 0.075 0.200 0.027 0.013 0.208 0.882 (mm) (0.011) (0.005) (0.054) (0.010)

Inner rim volume (mm3) −0.012 0.009 −0.002 0.212 0.003 0.860 −0.008 0.009 0.143 0.939 (0.004) (0.001) (0.018) (0.003)

Major slope disk length (mm) 0.016 0.061 0.002 0.312 −0.014 0.623 0.010 0.022 0.083 0.962 (0.008) (0.002) (0.028) (0.004)

Abbreviations: B = estimate; GCIPL = combined macular ganglion cell and inner plexiform layer volume; INL = inner nuclear layer volume; MS = patients with MS; NMOSD = patients with neuromyelitis optica spectrum disorders; NMOSD × ON = interaction effect of diagnosis and ON; ON = optic neuritis; ON− = eyes 2 2 without a history of ON; ON+ = eyes with a history of ON; SE = standard error of B; RCond = conditional R-squared; RMarg = marginal R-squared. Significant p values are marked in bold.

J–K shows sample central B scans (crossing the fovea) of Discussion ON− from patients with NMOSD and MS, chosen from the median of the selected pit flat disk parameters in each group. Using a novel foveal morphometry approach, we here show that foveal shape is altered in patients with AQP4- Parameter confirmation IgG–seropositive NMOSD. Our results further support that Finally, we tested whether the parameters identified in the ex- these changes cannot be explained by neuroaxonal damage ploratory analysis could be confirmed in an independent cohort resulting from ON alone. of patients with NMOSD and MS measured with the same device and protocol at an independent center. Based on dif- Foveal morphometry described a flatter and wider fovea in ferences in the selected parameters, we determined the mini- AQP4-IgG–seropositive NMOSD both in comparison to MS mum sample size for a confirmatory cohort as n = 38, 29, 35, and HC (figure 2, J–K). This is characterized by increased pit and 59 eyes per group, based on measurements for pit flat disk flat disk area, increased average pit flat disk diameter, reduced area, average pit flat disk diameter, inner rim volume, and major inner rim volume, and increased major slope disk length. Al- slope disk length, respectively. In this confirmatory cohort, pit though neuroaxonal damage from ON altered the foveal shape flat disk area and average pit flat disk diameter were confirmed as well, we observed robust changes in these parameters also in to be significantly different in NMOSD in comparison to MS (B eyes never experiencing an ON and when correcting for ON or [SE] = 0.007 [0.004] mm2, p =0.035andB[SE]=0.018 neuroaxonal damage in the statistical models in all eyes. [0.010] mm, p = 0.039, respectively), neither dependent on ON (p = 0.254 and 0.184) nor on NMOSD-specificON(p =0.293 The foveal shape changes in AQP4-IgG–seropositive NMOSD and 0.382). Differences in inner rim volume were not significant reported in our study are supported by previous studies, which in the confirmatory cohort (diagnosis: p = 0.125; ON: p = investigated thickness or volume changes as indirect evidence 0.080; NMOSD-specificON:p = 0.056). Major slope disk for foveal shape changes. Jeong et al.16 and Oertel et al.17 length only showed a significant association with NMOSD- showed a significant reduction in FT in eyes of patients with specific ON (diagnosis: p = 0.155; ON history: p = 0.370; AQP4-IgG–seropositive NMOSD independent of ON in NMOSD-specific ON: B [SE] = 0.012 [0.007] mm, p =0.046). comparison to HCs.

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN A pathophysiologic explanation for the observed changes could be observed changes may be caused or affected by covert ON and the presence of a primary retinopathy in AQP4-IgG–seropositive neuroaxonal damage of the optic nerve. NMOSD, mediated by AQP4-IgG. The principal glial cell of the retina is the Müller cell, expresses AQP4, and is enriched around The main limitation of our study was the low sample size for AQP4- the fovea.25 Müller cell bodies reside in the INL, but process IgG–seropositive NMOSD, especially for patients without a history stretch through the whole thickness of the retina, linking retinal of ON, which is unfortunately common in studies investigating neurons and photoreceptors with blood vessels. Importantly, an- NMOSD. Another limitation of this study is that the diagnostic imal studies have shown complement-independent AQP4 loss in value is unclear, as we only used scans from 1 OCT device using Müller cells in rats induced by AQP4-IgG, which is in line with an a single scanning protocol. It is unclear how scans from different AQP4-IgG–mediated primary retinopathy in NMOSD.26,27 OCT devices and scanning protocols can be compared. AQP4-IgG–mediated primary retinal astrocytopathy has been also suggested in human by AQP4-IgG–seropositive NMOSD Foveal morphometry may potentially be useful for differential autopsy cases.28 AQP4 is expressed in Müller cell end feet— diagnosis of AQP4-IgG–seropositive NMOSD. Typically, analogous to astrocytic end feet—at the blood-retina barrier.29 pRNFL and GCIPL as well as other OCT parameters associ- For AQP4-IgG circulating in serum to reach its antigen, the blood- ated with neuroaxonal damage are mostly nonspecifictothe retina barrier presumably needs to be disrupted. Blood-retina or underlying ON etiology. In contrast, many foveal morphom- -brain barrier disruptions are typically associated with an acute etry parameters showed significant differences between inflammatory event and then conceptionally linked to an acute patients with NMOSD and MS in this study. Parameter se- attack involving complement.30 It is unclear whether or to which lection resulted in 4 promising parameters describing foveal extent blood-retina/brain barrier disruptions occur in NMOSD differences: pit flat disk area, average pit flat disk diameter, and other diseases that do not lead to full attack cascades. A recent major slope disk length, and inner rim volume. Only the first 2 analysis of the NMOSD momentum trial data31 revealed that parameters could be confirmed in an independent cohort. The elevated glial fibrillary acidic protein levels in serum were associ- reason may be the different frequency of ON in the confir- ated with an increased attack risk, independently suggesting that matory cohort between patients with MS and NMOSD, which there is indeed subclinical astrocyte damage outside attacks (An- exemplifies the need for additional confirmation, especially in nual European Committee for Treatment and Research in Mul- eyes that are inconspicuous in regard to neuroaxonal damage tiple Sclerosis [ECTRIMS] 2019, P1609).32 A recent study could from ON. Future work should further investigate this by further show that in rats, blood-brain barrier breakdown is not comparing patients with myelin oligodendrocyte glycoprotein necessary for NMOSD pathology, but that NMOSD-like disease (MOG)-IgG–seropositive disease against patients with MOG/ can be caused by AQP4-IgG circulating in CSF.33 We recently AQP4-IgG double-negative NMOSD, as well as clarify effects reported progressive GCIPL loss without ON in a longitudinal of scan protocols and foveal variability in healthy persons. study investigating an overlapping cohort.34 Further evidence for a primary retinopathy comes from Tian et al.35 reporting inner Study funding retinal layer thinning independent from ON. Significant changes Supported by the Einstein Foundation Berlin (Einstein Junior in vascularization of the fovea were also shown in patients with Scholarship to S.M.), the German Federal Ministry of Eco- AQP4-IgG–seropositive NMOSD in comparison to HCs using nomic Affairs and Energy (BMWI EXIST 03EFEBE079 to – OCT angiography.36 38 A.U.B. and E.M.K.), German Research Foundation (DFG Exc. 257 to F.P. and A.U.B.), German Federal Ministry of Educa- Alternatively, foveal changes could be caused by subclinical tion and Research (BMBF Neu2 ADVISIMS to F.P. and ON. Occasionally, studies have reported neuroaxonal damage A.U.B. as well as part of the “German Competence Network also in eyes without prior ON in AQP4-IgG–seropositive Multiple Sclerosis” (KKNMS), project NationNMO, NMOSD, which could be interpreted as evidence as such.39 01GI1602B to O.A.), and Novartis (research grant to H.G.Z.). However, earlier studies had cohort heterogeneity due to incomplete antibody characterization or inclusion of patients Disclosure with antibody-negative NMOSD. Furthermore, ON in E.M. Kadas, S.K. Yadav, A.U. Brandt, S. Motamedi, and F. NMOSD often occurs near the chiasm, and neuroaxonal Paul are named as coinventors on the patent application for damage can be caused by chiasmal crossover from an affected the foveal shape analysis method used by this manuscript eye to the fellow eye,40 which might not be clinically apparent. (“Method for estimating shape parameters of the fovea by In this study, we found evidence of neuroaxonal damage also optical coherence tomography”, International Publication in eyes reported to have never experienced ON, which is in Number: “WO 2019/016319 A1”). E.M. Kadas, S.K. Yadav, F. agreement with our recent findings in a multicenter study.34 It Paul, and A.U. Brandt are cofounders and hold shares in is possible that our data set also included fellow eyes that were technology start-up Nocturne GmbH, which has commercial affected from cross-chiasmal affects during a contralateral ON. interest in OCT applications in neurology. E.M. Kadas and The number of patients with NMOSD only experiencing S.K. Yadav are now employees of Nocturne GmbH. H.G. transverse myelitis and no ON was too small, which is why we Zimmermann received a research grant from Novartis. All refrained from analyzing eyes of these patients separately. In other authors report no relevant disclosures. Go to Neurol- consequence, we cannot fully determine to which extent the ogy.org/NN for full disclosures.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 9 Publication history References fl 1. Jarius S, Paul F, Franciotta D, et al. Mechanisms of disease: aquaporin-4 antibodies in Received by Neurology: Neuroimmunology & Neuroin ammation – fi neuromyelitis optica. Nat Clin Pract Neurol 2008;4:202 214. January 30, 2020. Accepted in nal form April 16, 2020. 2. Paul F, Jarius S, Aktas O, et al. Antibody to aquaporin 4 in the diagnosis of neuromyelitis optica. PLoS Med 2007;4:e133. 3. Zekeridou A, Lennon VA. Aquaporin-4 autoimmunity. Neurol Neuroimmunol Neuroinflamm 2015;2:e110. doi: 10.1212/NXI.0000000000000110. 4. Jarius S, Wildemann B, Paul F. Neuromyelitis optica: clinical features, immunopatho- – Appendix Authors genesis and treatment: neuromyelitis optica. Clin Exp Immunol 2014;176:149 164. 5. Schmidt F, Zimmermann H, Mikolajczak J, et al. Severe structural and functional Name Location Contribution visual system damage leads to profound loss of vision-related quality of life in patients with neuromyelitis optica spectrum disorders. Mult Scler Relat Disord – Seyedamirhosein Charité—Universitätsmedizin Collected the data; conducted 2017;11:45 50. Motamedi, MSc Berlin, Germany the statistical analysis; 6. Schneider E, Zimmermann H, Oberwahrenbrock T, et al. Optical coherence to- contributed to development of mography reveals distinct patterns of retinal damage in neuromyelitis optica and the foveal shape analysis multiple sclerosis. PLoS One 2013;8:e66151. method; and drafted the 7. Wingerchuk DM, Banwell B, Bennett JL, et al. International consensus diagnostic manuscript for intellectual criteria for neuromyelitis optica spectrum disorders. Neurology 2015;85:177–189. content 8. Yamamura T, Nakashima I. Foveal thinning in neuromyelitis optica: a sign of retinal astrocytopathy? Neurol Neuroimmunol Neuroinflamm 2017;4:e347. doi: 10.1212/ Frederike C. Charité—Universitätsmedizin Collected the data; performed Oertel, MD Berlin, Germany OCT quality control and NXI.0000000000000347. segmentation; contributed to 9. Oertel FC, Zimmermann H, Paul F, Brandt AU. Optical coherence tomography in data interpretation; and neuromyelitis optica spectrum disorders: potential advantages for individualized revised the manuscript for monitoring of progression and therapy. EPMA J 2018;9:21–33. intellectual content 10. Bennett JL, de Seze J, Lana-Peixoto M, et al. Neuromyelitis optica and multiple sclerosis: seeing differences through optical coherence tomography. Mult Scler Sunil K. Yadav, Charité—Universitätsmedizin Developed the foveal shape Houndmills Basingstoke Engl 2015;21:678–688. PhD Berlin, Germany analysis method and revised 11. Oertel FC, Zimmermann H, Mikolajczak J, et al. Contribution of blood vessels to the manuscript for retinal nerve fiber layer thickness in NMOSD. Neurol Neuroimmunol Neuroinflamm intellectual content 2017;4:e338. doi: 10.1212/NXI.0000000000000338. Ella M. Kadas, PhD Charité—Universitätsmedizin Contributed to development 12. Oberwahrenbrock T, Traber GL, Lukas S, et al. Multicenter reliability of semi- Berlin, Germany of the foveal shape analysis automatic retinal layer segmentation using OCT. Neurol Neuroimmunol Neuro- method and revised the inflamm 2018;5:e449. doi: 10.1212/NXI.0000000000000449. manuscript for intellectual 13. Kaufhold F, Zimmermann H, Schneider E, et al. Optic neuritis is associated with inner content nuclear layer thickening and microcystic macular edema independently of multiple sclerosis. PLoS One 2013;8:e71145. Margit Weise, Heinrich Heine University, Major role in acquisition of 14. Syc SB, Saidha S, Newsome SD, et al. Optical coherence tomography segmentation MSc Dusseldorf,¨ Germany the confirmatory data and reveals ganglion cell layer pathology after optic neuritis. Brain 2012;135:521–533. revised the manuscript for 15. Pache F, Zimmermann H, Mikolajczak J, et al. MOG-IgG in NMO and related intellectual content disorders: a multicenter study of 50 patients. Part 4: afferent visual system damage Joachim Havla, Ludwig-Maximilians Contributed to data after optic neuritis in MOG-IgG-seropositive versus AQP4-IgG-seropositive patients. fl MD University, Munich, Germany interpretation and revised the J Neuroin ammation 2016;13:282. manuscript for intellectual 16. Jeong IH, Kim HJ, Kim NH, Jeong KS, Park CY. Subclinical primary retinal pathology content in neuromyelitis optica spectrum disorder. J Neurol 2016;263:1343–1348. 17. Oertel FC, Kuchling J, Zimmermann H, et al. Microstructural visual system changes in Marius Heinrich Heine University, Major role in acquisition of AQP4-antibody–seropositive NMOSD. Neurol Neuroimmunol Neuroinflamm 2017; Ringelstein, MD Dusseldorf,¨ Germany the confirmatory data and 4:e334. doi: 10.1212/NXI.0000000000000334. revised the manuscript for 18. Yadav SK, Motamedi S, Oberwahrenbrock T, et al. CuBe: parametric modeling of 3D intellectual content foveal shape using cubic Bézier. Biomed Opt Express 2017;8:4181. 19. Tewarie P, Balk L, Costello F, et al. The OSCAR-IB consensus criteria for retinal OCT Orhan Aktas, MD Heinrich Heine University, Major role in acquisition of Dusseldorf,¨ Germany the confirmatory data and quality assessment. PLoS One 2012;7:e34823. revised the manuscript for 20. Polman CH, Reingold SC, Banwell B, et al. Diagnostic criteria for multiple sclerosis: intellectual content 2010 revisions to the McDonald criteria. Ann Neurol 2011;69:292–302. 21. Schippling S, Balk LJ, Costello F, et al. Quality control for retinal OCT in multiple Philipp Albrecht, Heinrich Heine University, Major role in acquisition of sclerosis: validation of the OSCAR-IB criteria. Mult Scler Houndmills Basingstoke MD Dusseldorf,¨ Germany the confirmatory data and Engl 2015;21:163–170. revised the manuscript for 22. Cruz-Herranz A, Balk LJ, Oberwahrenbrock T, et al. The APOSTEL recom- intellectual content mendations for reporting quantitative optical coherence tomography studies. Neu- rology 2016;86:2303–2309. Klemens Charité—Universitätsmedizin Contributed to study 23. R Core Team. R: A Language and Environment for Statistical Computing [online]. Ruprecht, MD Berlin, Germany management and revised the manuscript for intellectual Vienna: R Foundation for Statistical Computing; 2018. Available at: www.R-project.org/. content Accessed June 13, 2019. 24. Ratchford JN, Quigg ME, Conger A, et al. Optical coherence tomography helps differ- Judith Bellmann- Charité—Universitätsmedizin Contributed to study entiate neuromyelitis optica and MS optic neuropathies. Neurology 2009;73:302–308. Strobl, MD Berlin, Germany management and revised the 25. Bringmann A, Pannicke T, Grosche J, et al. Müller cells in the healthy and diseased manuscript for intellectual retina. Prog Retin Eye Res 2006;25:397–424. content 26. Felix CM, Levin MH, Verkman AS. Complement-independent retinal pathology produced by intravitreal injection of neuromyelitis optica immunoglobulin G. — Hanna G. Charité Universitätsmedizin Contributed to data J Neuroinflammation 2016;13:275. Zimmermann, Berlin, Germany interpretation and revised the 27. Zeka B, Hastermann M, Kaufmann N, et al. Aquaporin 4-specific T cells and NMO- PhD manuscript for intellectual content IgG cause primary retinal damage in experimental NMO/SD. Acta Neuropathol Commun 2016;4:82. Friedemann Paul, Charité—Universitätsmedizin Contributed to study 28. Hokari M, Yokoseki A, Arakawa M, et al. Clinicopathological features in anterior MD Berlin, Germany management; contributed to visual pathway in neuromyelitis optica. Ann Neurol 2016;79:605–624. data interpretation; and 29. Goodyear MJ, Crewther SG, Junghans BM. A role for aquaporin-4 in fluid regulation revised the manuscript for in the inner retina. Vis Neurosci 2009;26:159–165. intellectual content 30. Takeshita Y, Obermeier B, Cotleur AC, et al. Effects of neuromyelitis optica-IgG at the blood-brain barrier in vitro. Neurol Neuroimmunol Neuroinflamm 2017;4:e311. doi: — Alexander U. Charité Universitätsmedizin Designed and conceptualized 10.1212/NXI.0000000000000311. Brandt, MD Berlin, Germany the study; supervised the 31. Cree BAC, Bennett JL, Kim HJ, et al. Inebilizumab for the treatment of neuromyelitis statistical analysis; and drafted the manuscript for optica spectrum disorder (N-MOmentum): a double-blind, randomised placebo- – intellectual content controlled phase 2/3 trial. Lancet Lond Engl 2019;394:1352 1363. 32. ECTRIMS 2019—late breaking news abstracts. Mult Scler J 2019;25:890–938.

10 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN 33. Hillebrand S, Schanda K, Nigritinou M, et al. Circulating AQP4-specific auto- 37. Kwapong WR, Peng C, He Z, Zhuang X, Shen M, Lu F. Altered macular microvas- antibodies alone can induce neuromyelitis optica spectrum disorder in the rat. Acta culature in neuromyelitis optica spectrum disorders. Am J Ophthalmol 2018;192: – Neuropathol (Berl) 2019;137:467 485. 47–55. 34. Oertel FC, Havla J, Roca-Fernández A, et al. Retinal ganglion cell loss in neuro- 38. Green AJ, Cree BAC. Distinctive retinal nerve fibre layer and vascular changes in myelitis optica: a longitudinal study. J Neurol Neurosurg Psychiatry 2018;89: neuromyelitis optica following optic neuritis. J Neurol Neurosurg Psychiatry 2009;80: 1259–1265. 1002–1005. 35. Tian DC, Su L, Fan M, et al. Bidirectional degeneration in the visual pathway in 39. Ringelstein M, Harmel J, Zimmermann H, et al. Longitudinal optic neuritis-unrelated neuromyelitis optica spectrum disorder (NMOSD). Mult Scler J 2018;24: 1585–1593. visual evoked potential changes in NMO spectrum disorders. Neurology 2020;94: – 36. Huang Y, Zhou L, ZhangBao J, et al. Peripapillary and parafoveal vascular network e407 e418. ff assessment by optical coherence tomography angiography in aquaporin-4 antibody- 40. Ramanathan S, Prelog K, Barnes EH, et al. Radiological di erentiation of optic neuritis positive neuromyelitis optica spectrum disorders. Br J Ophthalmol 2019;103: with myelin oligodendrocyte glycoprotein antibodies, aquaporin-4 antibodies, and 789–796. multiple sclerosis. Mult Scler Houndmills Basingstoke Engl 2016;22:470–482.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 11 ARTICLE OPEN ACCESS Recent cannabis use in HIV is associated with reduced inﬂammatory markers in CSF and blood

Ronald J. Ellis, MD, PhD, Scott N. Peterson, PhD, Yueling Li, MD, Rachel Schrier, PhD, Jenny Iudicello, PhD, Correspondence Scott Letendre, MD, Erin Morgan, PhD, Bin Tang, PhD, Igor Grant, MD, and Mariana Cherner, PhD Dr. Ellis [email protected] Neurol Neuroimmunol Neuroinflamm 2020;7:e809. doi:10.1212/NXI.0000000000000809 Abstract Objective To determine whether cannabis may reduce HIV-related persistent inflammation, we evaluated the relationship of cannabis use in people with HIV (PWH) to inflammatory cytokines in CSF and blood plasma.

Methods We measured a panel of proinﬂammatory cytokines (interleukin [IL]-16, C-reactive protein [CRP], IL-6, interferon gamma-induced protein [IP]-10, soluble CD14, and soluble tumor necrosis factor receptor type II [sTNFRII]) in CSF and blood plasma in PWH and HIV− individuals who did or did not use cannabis at various levels of exposure. Participants in this observational cohort were recruited from community sources and underwent lumbar puncture and phlebotomy. Cannabis use parameters were characterized by self-report based on a semistructured timeline follow-back interview. Cytokines were measured using commercially available immunoassays. Data were analyzed using factor analysis.

Results Participants were 35 PWH and 21 HIV− individuals, mean (SD) age 45.4 (14.5) years, 41 cannabis ever users, and 15 never users. PWH and HIV− were not different in recency, cumulative months, grams, or density of cannabis use. A factor analysis using CSF biomarkers yielded a factor loading on CRP, IL-16, and sTNFRII that was significantly associated with recency of cannabis use (more recent use associated with lower factor 1 values, reflecting less inflammation; r = 0.331 [95% CI 0.0175, 0.586]). In particular, more recent cannabis use was related to lower IL-16 levels (r = 0.549 [0.282, 0.737]). Plasma biomarkers yielded a factor loading on sTNFRII and IP-10 that was associated with more recent cannabis use (more recent use related to less inflammation; r = 0.374 [0.0660, 0.617]).

Conclusions Recent cannabis use was associated with lower levels of inflammatory biomarkers, both in CSF and blood, but in different patterns. These results are consistent with compartmentalization of immune effects of cannabis. The principal active components of cannabis are highly lipid soluble and sequestered in brain tissue; thus, our findings are consistent with specific anti- neuroinflammatory effects that may benefit HIV neurologic dysfunction.

From the Departments of Neurosciences and Psychiatry (R.J.E.), University of California, San Diego; Sanford Burnham Prebys Medical Discovery Institute (S.N.P.); LECOM health - Millcreek Community Hospital (Y.L.), Erie, PA; Department of Pathology (R.S.), Department of Psychiatry (J.I., E.M., B.T., I.G., M.C.), and Departments of Medicine and Psychiatry (S.L.), University of California, San Diego.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary ART = antiretroviral therapy; CB2R = CB2 receptor; CRP = C-reactive protein; HAND = HIV-associated neurologic dysfunction; IL = interleukin; IP = interferon gamma-induced protein; PWH = people with HIV; RBC = red blood cell; THC = delta-9-tetrahydrocannabinol; TLFBI = timeline follow-back interview; TNFRII = tumor necrosis factor receptor type II.

Despite antiretroviral therapy (ART), HIV-associated neuro- Clinical assessment logic dysfunction (HAND) persists in approximately one-third All participants underwent a comprehensive neuromedical as- of people with HIV (PWH). Cannabis use is highly prevalent sessment, blood draw, and lumbar puncture. Parameters of among PWH compared with people who are HIV−.1,2 Previous lifetime cannabis use were estimated using a semistructured research has demonstrated that microglia are chronically acti- timeline follow-back interview (TLFBI) as previously de- vated in PWH, reflecting neuroinflammation, and may con- scribed.13 The TLFBI uses a calendar method to evaluate daily – tribute to HAND.3 5 Although numerous studies have patterns and frequency of substance use over a specified period. evaluated cannabis effects on systemic inflammation and im- It has high retest reliability, convergent and discriminant val- mune activation,6,7 no previous studies have reported on can- idity with other measures, agreement with collateral inform- nabis use in relation to CSF inflammatory markers. It is known ants’ reports of patients’ substance use, and agreement with that expression of interleukin 16 (IL-16) is increased in results from patients’ urine assays.14 Cannabis recency was microglial nodules in HIV encephalitis. Microglia express CB2 assessed as self-reported months since last cannabis use. Other receptor (CB2R), a cannabinoid receptor that regulates their cannabis variables assessed included self-reported frequency, activation.8,9 In other neuroinflammatory diseases, neutralizing density, and cumulative dose. IL-16 reduces demyelination and axonal damage. Cannabis downregulates IL-16 in myeloid cells including microglia.10 Laboratory evaluations Given that cannabis may have protective anti-inflammatory HIV disease was diagnosed by ELISA with Western blot con- effects, is neuroprotective in neurodegenerative disorders firmation. Among HIV+ participants, HIV viral load in plasma (Alzheimer disease and Parkinson disease), and is lipophilic and was measured using reverse transcriptase PCR (Amplicor; concentrated 3 to 10 times higher in the brain than in blood, we Roche Diagnostics, Indianapolis, IN) and deemed undetectable sought to examine possible links between cannabis use and IL- at a lower limit of quantitation of 50 copies/mL. CD4 was 16 as a marker of neuroinflammation in HIV. Cannabis bi- measured by flow cytometry, and nadir CD4 was assessed by ological effects may differ by sex.11,12 Given the potential neu- self-report. Biomarkers measured in plasma and CSF by im- roprotective effects of cannabis and its widespread use in PWH, munoassay were (IL)-16 (Merck Sharp & Dohme, Rockville, cannabis could be an intervention to protect against HAND. MD, kit #K151RED; quantitation range CSF and plasma 6.18–2,530 pg/mL), C-reactive protein (CRP) (MSD, We sought to evaluate possible links between cannabis use and K151STD; quantitation range CSF and plasma 12.5–195,000 biomarkers of neuroinflammation in HIV. Figure 1 illustrates the proposed model of cannabis modulation of HIV-induced microglial activation. We hypothesized that inflammatory bio- Figure 1 Conceptual model markers would be reduced in HIV+ individuals who more recently used cannabis.

Methods Participants This was a cross-sectional prospective observational cohort study of HIV+ and HIV− individuals recruited from community sources. Inclusion criteria were availability of data on cannabis use. Exclusions were diagnoses of other substance use disorders within the past year, positive urine toxicology for substances of abuse other than cannabis, and presence of an active, major psychiatric or neurologic condition such as schizophrenia or epilepsy. HIV activates resting microglia, which produce cytokines reflecting neuroinflammation. Cannabis (especially THC) acts as a ligand for the endocanna- Standard protocol approvals, registrations, binoid receptor CB2, reducing neuroinflammation. CB2R = cannabinoid and patient consents receptor type 2; CRP = C-reactive protein; IL = interleukin; IP = interferon gamma-induced protein; THC = delta-9-tetrahydrocannabinol; TNFRII = tumor All participants signed an institutional review board– necrosis factor receptor type II. approved written consent.

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Tukey Honest Significant Difference tests. Factor analysis was Table 1 Participant demographic and clinical performed in CSF and plasma separately to identify underlying characteristics by HIV serostatus factors that explain correlations among cytokines measured as fi HIV2 HIV+ p described above. The rst 3 factors were chosen based on the criterion of eigenvalues greater than 1. Each factor includes the N 21 35 cytokines with loading (correlation, r) value at least 0.30. We Age, y (mean ± SD) 42.7 (14.2) 47 (14.7) 0.28 used multivariable linear regression models to test the in- ff Education, y (mean ± SD) 13.4 (2.8) 13.7 (2.6) 0.68 teraction e ect between HIV status and cannabis recency on CSF factors 1–3. In the absence of the interaction effect, ad- Female sex (N, %) 7 (33%) 3 (9%) 0.02 ditive effects were tested. In addition, simple linear models were Non-Hispanic white (N, %) 10 (47%) 10 (37%) 0.15 used to regress age, current CD4, and cannabis recency on CSF factor 1, and then the relationship of CSF factor 1 with cannabis CD4 nadir (median, IQR) — 278 (280, 535) — recency was adjusted for age using multivariable regression. Current CD4 (median, IQR) — 699 (494, 935) — Finally, the correlation between plasma factor 2 and cannabis recency was assessed with Pearson correlation. Sensitivity Groups were similar except with respect to sex, with more women in the HIV− group. analyses were conducted on the subset of participants who were virally suppressed. All analyses were conducted using JMP Pro version 14.0.0 (JMP, Version 12.0.1.; SAS Institute Inc., pg/mL), IL-6 (MSD K15049D; quantitation range CSF Cary, NC, 2018). 0.183–749 pg/mL; plasma 3.2–10,000 pg/mL), interferon gamma-induced protein (IP)-10 (MSD, K15047D; quantita- Data availability tion range CSF IP-10 and plasma 3.2–10,000 pg/mL), soluble Data will be made available on request. (s) CD14 (R&D, Minneapolis, MN, kit #DC140; quantitation range CSF and plasma 250–16,000 pg/mL), and soluble tumor necrosis factor receptor type II (TNFRII) (MSD F21ZS-3; Results quantitation range CSF and plasma 12.2–50,000 pg/mL). Participants were 35 PWH and 21 HIV− individuals, mean Statistical analysis (SD) age 45.4 (14.5) years, 41 cannabis ever users, and 15 HIV and cannabis group differences on background charac- never users. PLWH had a mean CD4 of 699; 94% took ART, teristics (i.e., demographics and neuropsychiatric and neuro- and 86% were virologically suppressed. Demographics and medical characteristics) were examined using analysis of clinical characteristics by HIV serostatus are summarized in variance, Wilcoxon/Kruskal-Wallis tests, and χ2 statistics as table 1. Recency of cannabis use was recorded as a continuous appropriate. Except for CSF IL-16 and TNFRII and plasma variable (median 10 months; interquartile range [IQR] 1–61 – soluble CD14, log10 transformation of all of the other cytokine months; range 0.03 414 months). Among users, the route of measures improved their fit to a normal distribution, and the administration was smoked, and participants did not know the transformed values were used in analyses. To determine the specific formulation (delta-9-tetrahydrocannabinol [THC] vs combined effects of HIV and cannabis on CSF and plasma IL- cannabidiol content); the median grams per day over the last 16 levels, multivariable linear regression analyses modeled IL- 30 days was 0.5 (IQR 0.125, 0.75). The percent with a lifetime 16 as a function of HIV, cannabis use, and their interaction. history of any substance abuse was 41%, with the majority of Follow-up comparisons examined pairwise group differences in this being alcohol (38.6%). Only 8% had a lifetime history of IL-16 levels with correction for multiple comparisons using cannabis abuse or dependence. Cannabis ever users did not

Table 2 Plasma vs CSF correlations for each of the biomarkers

Variable By variable Correlation 95% CI

Log10 plasma IL-16 CSF IL-16 (pg/mL) 0.239 −0.0473, 0.455

Log10 (plasma CRP) Log10 (CSF CRP) 0.808 0.692, 0.883

Log10 (plasma IL-6) Log10 (CSF IL-6) −0.069 −0.326, 0.198

Log10 (plasma IP-10) Log10 (CSF IP-10) 0.287 0.0258, 0.511

Plasma sCD14 Log10 (CSF sCD14) 0.198 −0.0685, 0.438

Log10 (plasma sTNFRII) CSF sTNFRII 0.311 0.0529, 0.531

Abbreviations: CRP = C-reactive protein; IL = interleukin; IP-10 = interferon gamma-induced protein 10; sCD14 = soluble CD14. Correlations were significant (higher plasma associated with higher CSF) for CRP, IP-10, and sTNFRII, but not for the other 3 biomarkers.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 Table 3 Factor loadings of the CSF and plasma factors

CRP IL-16 IL-6 IP-10 sTNFRII sCD14

CSF factor 1 0.955 0.398 0.332

CSF factor 2 0.453 0.333 0.432 0.694

CSF factor 3 0.707 0.487

Plasma factor 1 0.811 0.425 0.746

Plasma factor 2 0.346 0.954

Plasma factor 3 0.961

Abbreviations: CRP = C-reactive protein; IL = interleukin; IP-10 = interferon gamma-induced protein 10; sCD14 = soluble CD14.

differ from never users on age (45.3 ± 12.8 vs 45.7 ± 18.9 years; use was no longer significantly correlated with CSF CRP (p = p = 0.928), female sex (20% vs 17.1%), or non-Hispanic white 0.247). In a multivariate model with HIV and cannabis recency ethnicity (46.7% vs 39.0%). PWH and HIV− were not different as independent variables, HIV × cannabis recency interaction in recency, cumulative months, grams, or density of use. was not significant, indicating that the cannabis recency effect Months since last use was unrelated to current and nadir CD4 was not different in PWH than in HIV− participants. Viral loads (p = 0.15 and 0.43, respectively). Older participants had more were detectable in 5 participants. In a multivariable model in- months since last cannabis use than younger. Months since last cluding CSF factor 1 and viral load detectability, the latter was use was unrelated to race/ethnicity (p =0.37).Menand nonsignificant (p = 0.631). CSF factor 1 was not associated women did not differ with respect to months since last cannabis with cumulative duration, quantity, or density of cannabis use. use (67.8 ± 127 vs 70.2 ± 111). Four individuals tested positive Higher CSF factor 1 was significantly associated with older age on a urine drug screen for cannabis. (r = 0.414 [0.169, 0.610]) and higher current CD4 (r = 0.338 [0.00514, 0.603]. In a multivariable regression, after adjusting Although in most cases, PWH had higher levels of biomarkers of for age, the relationship between CSF factor 1 and cannabis inflammation than HIV− individuals, these differences reached recency was no longer significant (p = 0.227). CD4 did not significance only for IP-10 (for CSF, 753 ± 430 vs 493 ± 435; and reach significance in a multivariable model (p = 0.167). CSF for plasma, 588 ± 349 vs 367 ± 215). Table 2 shows correlations factors 2 and 3 were not related to cannabis use. between plasma and CSF levels of individual biomarkers. No- tably, plasma-CSF correlations were generally modest or nonsignificant, except for CRP (r = 0.808 [0.692, 0.883]). Figure 2 More recent cannabis use correlated with lower Factor analysis CSF factor 1 (indicating reduced CRP, IL-16, and Separate factor analyses were performed for CSF and plasma sTNFRII) biomarkers to reduce relatively large number of individual biomarkers into a fewer number of dimensions for analysis. Table 3 shows the factor loadings for each of the resulting factors. Levels of all factors were nonsignificantly higher in PWH than in HIV−.

Cannabis recency and CSF factors More recent cannabis use was significantly associated with lower CSF factor 1 values (r = 0.331; 95% CI 0.0175, 0.586; figure 2) and with lower plasma factor 2 values (r = 0.374; 95% CI 0.0660, 0.617; figure 3A). Among 15 individuals who had never used cannabis, CSF factor 2 levels were higher than the 41 who had ever used cannabis (0.138 ± 0.819 vs −0.051 ± 0.765). Evaluating the individual components of CSF factor 1, more recent cannabis use was related to lower CSF CRP and IL-16 levels (r = 0.336 [95% CI 0.0223, 0.589] and r = 0.549 A factor analysis was performed on the 6 CSF markers CRP, IL-16, sTNFRII, [0.282, 0.737], but not sTNFRII (r = 0.080 [−0.241, 0.386] sCD14, IP-10, and IL-6. Only factor 1 was associated with recency of cannabis (figure 3, A–C). Older age was associated with higher CSF use. The Pearson correlation coefficient and its 95% CI are shown. Note logarithmic time scale. N = 41 ever used cannabis. CRP = C-reactive protein; IL = CRP (r = 0.504 [0.278, 0.677]), but not IL-16 (r = 0.0633 interleukin; IP = interferon gamma-induced protein; PWH = people with HIV. [−0.203, 0.321]). After accounting for age, recency of cannabis

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Figure 3 (A–C) More recent cannabis use correlated with 2 of the components of CSF factor 1, lower CSF CRP, and IL-16, but not with TNFRII levels

Pearson correlation coefficients and their 95% CIs are shown. CRP = C-reactive protein; IL = interleukin; TNFRII = tumor necrosis factor receptor type II.

Cannabis recency and plasma factors To evaluate for possible confounding by blood plasma contam- More recent cannabis use correlated with lower plasma ination of CSF, we used the following approach. First, we cal- factor 2 (indicating reduced sTNFRII and IP-10; r = 0.374 culated the correlations between CSF biomarkers and CSF red [0.0660, 0.617]; figure 4). Higher plasma factor 1 levels blood cell (RBC) count, a marker of blood contamination. Blood were associated with older age r = 0.381 (0.132, 0.586), but contamination would be particularly significant in the case where age was unrelated to the other 2 factors. Age did not in- plasma biomarker levels were much higher than those for CSF. fluence the relationship between factor 1 and recency of For example, plasma CRP levels were, on average, 1000-fold cannabis use. Lower nadir CD4 was related to higher plasma higher than CSF levels. Although log10 CSF and log10 plasma factor 2 (r = −0.425 [−0.664, −0.107]). In a multivariable CRP levels were significantly correlated (r = 0.808 [0.692, model containing both cannabis use recency and nadir CD4, 0.883]), CSF CRP levels were not associated with CSF RBC (r = neither variable reached significance. No other de- −0.00473 [−0.267, 0.258]), indicating that the correlation be- mographic or clinical variables were significantly related to tween CSF and blood was not attributable to blood contamina- plasma factor 2. tion. With the exception of IL-6 (r = 0.443 [0.204, 0.632]), CSF biomarker levels were not correlated with CSF RBC counts (ps> Table 4 shows Spearman correlations between CSF and 0.10). None of the CSF factor scores was significantly associated fi plasma factors and their 95% CIs. With the exception of CSF with CSF RBC. Despite the signi cant correlation of log10 CSF factor 1 and plasma factor 1, the factors were independent IL-6 with CSF RBC, log10 CSF and plasma IL-16 levels were from one another. uncorrelated (r = 0.218 [−0.0473, 0.455]).

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 5 Figure 4 More recent cannabis use correlated with lower plasma factor 2 (A), indicating reduced sTNFRII (B) and IP-10 (C)

Pearson correlation coefficients and their 95% CIs are shown for the plasma factor 2 and for its individual components, sTNFRII and IP-10. IP = interferon gamma-induced protein.

Discussion a receptor for TNF-α and indicates a regulatory response to increased inflammation. Elevated CSF levels of sTNFRII were sig- We found that recent cannabis use was associated with reduced nificantly correlated with worse cognition. Thus, the 3 levels of a factor loading on IL-16, CRP, and sTNFRII, reflecting components of the factor we found to be associated with cannabis reduced neuroinflammation. IL-16, originally called lymphocyte recency are involved in the inflammatory response in the CNS. chemoattractant factor,15 has been shown to recruit and activate many cells expressing the CD4 molecule, including monocytes, Although CB1 receptors are the principal type found in the CNS eosinophils, and dendritic cells.16 It stimulates the production of and account for the psychoactive effects of ligands such as THC, proinflammatory cytokines by human monocytes.17 IL-16 is un- CB2RsalsoareexpressedintheCNSofhumansandmany usual among the cytokines because it engages the CD4 molecule as animal species.7,21,22 In humans, the bulk of CB2R expression is its receptor instead of a unique IL-16R. Thus, although IL-16 can by microglia and THP-1 cells (a spontaneously immortalized block HIV cell entry, it nevertheless remains a proinflammatory monocyte-like cell line),23 as well as astrocytes,24 consistent with cytokine. In particular, because activated microglia express IL-16,18 aroleininflammation. Both in vitro studies and animal models our findings suggest reduced microglial activation with recent ex- show that CB2R is anti-inflammatory and may be part of the posure to cannabis. In addition, IL-16 is an endogenous ligand for general neuroprotective action of the endocannabinoid system the CD4 molecule (a receptor for HIV) and is known for its by decreasing glial reactivity.25 Both natural and synthetic can- chemotactic and anti–HIV-1 activities.19 CRP is a well-known nabinoids have been demonstrated to be neuroprotective after acute phase reactant. Although CRP expression is largely hepatic, it various types of CNS insults, such as stroke.26 is also expressed by multiple cells of the CNS, particularly during neuroinflammation, and CNS-specific expression of CRP asso- Although numerous studies have evaluated cannabis effects on ciates with more severe disease.20 sTNFRII (also known as p75) is systemic inflammation and immune activation in HIV,6 no

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Table 4 Spearman correlations between CSF and plasma factors and their 95% CIs

Variable By variable Correlation Lower 95% Upper 95%

Plasma factor 1 CSF factor 1 0.788 0.663 0.870

Plasma factor 1 CSF factor 2 −0.111 −0.364 0.156

Plasma factor 1 CSF factor 3 0.0828 −0.184 0.338

Plasma factor 2 CSF factor 1 0.139 −0.129 0.388

Plasma factor 2 CSF factor 2 0.158 −0.110 0.404

Plasma factor 2 CSF factor 3 0.218 −0.0477 0.455

Plasma factor 3 CSF factor 1 0.150 −0.117 0.397

Plasma factor 3 CSF factor 2 0.136 −0.132 0.385

Plasma factor 3 CSF factor 3 0.109 −0.159 0.361

With the exception of CSF factor 1 and plasma factor 1, the factors were independent. previous studies have reported on cannabis use in relation to with well-characterized histories of substance use, we found CSF soluble markers of inflammation in HIV. Also, it is often that exclusive use of cannabis was associated with a faster decay difficult to distinguish between heavy and recent cannabis use of HIV DNA during suppressive ART.33 Thus, cannabis use because heavy users are more likely to have used recently. For may reduce the HIV reservoir.34 example, in 1 study, heavy cannabis users had fewer activated immune cells (CD4+ and CD8+ T cells expressing Class II Our study is limited by several considerations. This was a small human leukocyte antigen and CD38 [HLA-DR+CD38+]) pilot study and subject to selection bias and omitted variable in peripheral blood compared with non–cannabis-using bias. Self-reported cannabis use is subject to recall and other individuals. Heavy cannabis users also had fewer IL-23 and biases, particularly in clinical settings, where use may have phorbol 12-myristate 13-acetate-induced tumor necrosis negative legal consequences. In this study, research participants factor-α–producing antigen-presenting cells27 in blood. were explicitly informed that all research data would be held in Complementing these findings, we show that soluble markers confidence. We did not characterize the composition of can- of inflammation in blood—specifically sTNFRII, IP-10—also nabis used in this study with respect to cannabidiol (CBD) or are lower in those who used cannabis. These findings are THC, and we did not measure THC and CBD in CSF and consistent with another study showing reduced plasma IP-10 blood. This is important because in previous reports, CBD and in cannabis users.28 In another study, cannabis users had lower THC showed varying effects on anti-inflammatory pathways, TNF-α levels than nonusers.29 including nuclear factor kappa-light-chain-enhancer of activated B cells and interferon-dependent pathways,35 and CBD The specific cytokines associated with cannabis use differed for accounting for the bulk of anti-inflammatory effects. CSF and plasma. This suggests compartmentalization of immune responses between the CNS and systemic compart- Futurestudiesshouldseektoreplicatethesefindings in a larger ments. We found a high correlation between CSF and plasma sample, directly measuring the response of inflammatory mark- CRP, but not between CSF and plasma levels of the other ers to cannabis administration in a controlled setting, and markers. Previously published evidence indicates that blood- characterize the time course of changes. Vaping and smoking borne CRP can indeed cross the blood-brain barrier.30 In ad- may have different effects. CNS inflammatory markers should dition, cannabinoids may achieve greater concentrations in be evaluated with respect to neuropsychological performance CNS due to their lipophilicity and longer duration of action due and imaging measures of neuroinflammation such as to sequestration in lipid-rich tissues. [18F]N-(2-(2-fluoroethoxy)benzyl)-N-(4-phenoxypyridin- 3-yl)acetamide. THC, CBD, and endocannabinoids should be Our results have substantial implications for HIV infection. measured in CSF and blood. In the future, cannabinoids may Because immune activation increases HIV target cells and HIV be studied as treatments for neuroinflammation in HIV. RNA, reduction of immune activation might lower viral load. This is in line with a previous study reporting that cannabis use Acknowledgment was associated with lower plasma HIV RNA among recently The Translational Methamphetamine AIDS Research Center infected PWH33 and with reports of reduced HIV replication (TMARC) is supported by Center award P50DA026306 from and cellular infection rate in the presence of cannabinoids in the National Institute on Drug Abuse (NIDA) and is affiliated vitro.31,32 In addition, in a previous longitudinal study of PWH with the University of California, San Diego (UCSD), the

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 7 Sanford-Burnham Medical Discovery Institute (SBMDI), and the University of California, Irvine (UCI). The TMARC comprises Appendix (continued) – Administrative Coordinating Core (ACC) Executive Unit: Name Location Contribution Director – Igor Grant, MD; Co-Directors – Ronald J. Ellis, MD, PhD, Scott L. Letendre, MD, and Cristian L. Achim, MD, PhD; Yueling Li, University of California, Interpreted the data and revised – MD San Diego the manuscript for intellectual Center Manager Mariana Cherner, PhD; Associate Center content Managers – Erin E. Morgan, PhD, and Jared Young, PhD; Data Rachel University of California, Interpreted the data and revised Management and Information Systems (DMIS) Unit: Ian S. Schrier, San Diego the manuscript for intellectual Abramson, PhD (Unit Chief) and Clint Cushman, BA (Unit PhD content – Manager); ACC Statistics Unit: Florin Vaida, PhD (Unit Jenny University of California, Interpreted the data and revised Chief),IanS.Abramson,PhD,BinTang,PhD,andAnyaUmlauf, Iudicello, San Diego the manuscript for intellectual MS; ACC – Participant Unit: J. Hampton Atkinson, MD (Unit PhD content Chief) and Jennifer Marquie-Beck, MPH (Unit Manager); Scott University of California, Major role in the acquisition of Behavioral Assessment and Medical (BAM) Core – Neuro- Letendre, San Diego data MD medical and Laboratory Unit (NLU): Scott L. Letendre, MD (Core Co-Director/NLU Chief) and Ronald J. Ellis, MD, PhD; Erin University of California, Interpreted the data and revised Morgan, San Diego the manuscript for intellectual BAM Core – Neuropsychiatric Unit (NPU): Robert K. Heaton, PhD content PhD (Core Co-Director/NPU Chief), J. Hampton Atkinson, Bin Tang, University of California, Performed statistical analyses MD, Thomas D. Marcotte, PhD, Erin E. Morgan, PhD, and PhD San Diego Matthew Dawson (NPU Manager); Neuroimaging (NI) Core: Igor University of California, Interpreted the data and revised GregoryG.Brown,PhD(CoreDirector), Thomas T. Liu, PhD, Grant, MD San Diego the manuscript for intellectual Miriam Scadeng, PhD, Christine Fennema-Notestine, PhD, Sarah content

L. Archibald, MA, John R. Hesselink, MD, Mary Jane Meloy, Mariana University of California, Interpreted the data and revised PhD, and Craig E.L. Stark, PhD; Neuroscience and Animal Cherner, San Diego the manuscript for intellectual Models (NAM) Core: Cristian L. Achim, MD, PhD (Core PhD content Director), Marcus Kaul, PhD, and Virawudh Soontornniyomkij, MD; Pilot and Developmental (PAD) Core: Mariana Cherner, PhD (Core Director) and Stuart A. Lipton, MD, PhD; Project 1: References 1. Anthony JC, Lopez-Quintero C, Alshaarawy O. Cannabis epidemiology: a selective Arpi Minassian, PhD (Project Director), William Perry, PhD, review. Curr Pharm Des 2017;22:6340–6352. Mark A. Geyer, PhD, and Jared W. Young, PhD; Project 2: 2. Pacek LR, Towe SL, Hobkirk AL, Nash D, Goodwin RD. Frequency of cannabis use and medical cannabis use among persons living with HIV in the United States: Amanda B. Grethe, PhD (Project Director) and Susan F. Tapert, findings from a nationally representative sample. AIDS Educ Prev 2018;30: PhD; Project 3: Erin E. Morgan, PhD (Project Director) and Igor 169–181. 3. Wu B, Huang Y, Braun AL, et al. Glutaminase-containing microvesicles from HIV-1- Grant, MD; Project 4: Samuel Barnes, PhD (Project Director); infected macrophages and immune-activated microglia induce neurotoxicity. Mol Project 5: Marcus Kaul, PhD(ProjectDirector). Neurodegener 2015;10:61. 4. Chen NC, Partridge AT, Sell C, Torres C, Martin-Garcia J. Fate of microglia during HIV-1 infection: from activation to senescence? Glia 2017;65:431–446. Study funding 5. Periyasamy P, Thangaraj A, Guo ML, Hu G, Callen S, Buch S. Epigenetic promoter No targeted funding reported. DNA methylation of miR-124 promotes HIV-1 tat-mediated microglial activation via MECP2-STAT3 Axis. J Neurosci 2018;38:5367–5383. 6. Klein TW. Cannabinoid-based drugs as anti-inflammatory therapeutics. Nat Rev Disclosure Immunol 2005;5:400. 7. Maresz K, Carrier EJ, Ponomarev ED, Hillard CJ, Dittel BN. Modulation of the R.J. Ellis, S.N. Peterson, Y. Li, R. Schrier, J. Iudicello, S. cannabinoid CB2 receptor in microglial cells in response to inflammatory stimuli. Letendre, E. Morgan, B. Tang, I. Grant, and M. Cherner report J Neurochem 2005;95:437–445. ff no disclosures. Go to Neurology.org/NN for full disclosures. 8. Carlisle SJ, Marciano-Cabral F, Staab A, Ludwick C, Cabral GA. Di erential expression of the CB2 cannabinoid receptor by rodent macrophages and macrophage- like cells in relation to cell activation. Int Immunopharmacol 2002;2:69–82. Publication history 9. Guida F, Luongo L, Boccella S, et al. Palmitoylethanolamide induces microglia fl changes associated with increased migration and phagocytic activity: involvement of Received by Neurology: Neuroimmunology & Neuroin ammation the CB2 receptor. Sci Rep 2017;7:375. December 22, 2019. Accepted in final form May 15, 2020. 10. Nair MP, Figueroa G, Casteleiro G, Munoz K, Agudelo M. Alcohol versus cannabinoids: a review of their opposite neuro-immunomodulatory effects and future therapeutic potentials. J Alcohol Drug Depend 2015;3:184. 11. Fattore L, Fratta W. How important are sex differences in cannabinoid action? Br J Appendix Authors Pharmacol 2010;160:544–548. 12. Struik D, Sanna F, Fattore L. The modulating role of sex and anabolic-androgenic Name Location Contribution steroid hormones in cannabinoid sensitivity. Front Behav Neurosci 2018;12:249. 13. Sobell LC, Sobell MB. Timeline Followback User’s Guide: A Calendar Method for Ronald J. University of California, Designed and conceptualized the Assessing Alcohol and Drug Use. Toronto: Addiction Research Foundation; 1996. Ellis, MD, San Diego study; analyzed the data; and 14. Fals-Stewart W, O’Farrell TJ, Freitas TT, McFarlin SK, Rutigliano P. The timeline PhD drafted the manuscript for followback reports of psychoactive substance use by drug-abusing patients: psycho- intellectual content metric properties. J Consult Clin Psychol 2000;68:134–144. 15. Cruikshank W, Center DM. Modulation of lymphocyte migration by human lym- Scott N. Sanford Burnham Interpreted the data and revised phokines. II. Purification of a lymphotactic factor (LCF). J Immunol 1982;128: Peterson, Prebys Medical the manuscript for intellectual 2569–2574. PhD Discovery Institute content 16. Cruikshank WW, Kornfeld H, Center DM. Interleukin-16. J Leukoc Biol 2000;67: 757–766.

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN 17. Mathy NL, Scheuer W, Lanzendorfer M, et al. Interleukin-16 stimulates the expres- 27. Manuzak JA, Gott TM, Kirkwood JS, et al. Heavy cannabis use associated with sion and production of pro-inflammatory cytokines by human monocytes. Immu- reduction in activated and inflammatory immune cell frequencies in antiretroviral nology 2000;100:63–69. therapy-treated human immunodeficiency virus-infected individuals. Clin Infect Dis 18. Liebrich M, Guo LH, Schluesener HJ, et al. Expression of interleukin-16 by tumor- 2018;66:1872–1882. associated macrophages/activated microglia in high-grade astrocytic brain tumors. 28. Rizzo MD, Crawford RB, Henriquez JE, et al. HIV-infected cannabis users have lower Arch Immunol Ther Exp (Warsz) 2007;55:41–47. circulating CD16+ monocytes and IFN-gamma-inducible protein 10 levels compared 19. Zhao ML, Si Q, Lee SC. IL-16 expression in lymphocytes and microglia in HIV-1 with nonusing HIV patients. AIDS 2018;32:419–429. encephalitis. Neuropathol Appl Neurobiol 2004;30:233–242. 29. Keen L II, Turner AD. Differential effects of self-reported lifetime marijuana use on 20. Wright T, McCrory M, Morgan T, et al. Central nervous system-specific expression of interleukin-1 alpha and tumor necrosis factor in African American adults. J Behav Med C-reactive protein exacerbates experimental autoimmune encephalomyelitis. 2015;38:527–534. (BA3P.202). J Immunol 2014;192(1 suppl):44.8. 30. Rajs G, Finzi-Yeheskel Z, Rajs A, Mayer M. C-reactive protein concentrations in 21. Benito C, Romero JP, Tolon RM, et al. Cannabinoid CB1 and CB2 receptors and fatty cerebral spinal fluid in gram-positive and gram-negative bacterial meningitis. Clin acid amide hydrolase are specific markers of plaque cell subtypes in human multiple Chem 2002;48:591–592. sclerosis. J Neurosci 2007;27:2396–2402. 31. Costantino CM, Gupta A, Yewdall AW, Dale BM, Devi LA, Chen BK. Cannabinoid 22. Van Sickle MD, Duncan M, Kingsley PJ, et al. Identification and functional charac- receptor 2-mediated attenuation of CXCR4-tropic HIV infection in primary CD4+ terization of brainstem cannabinoid CB2 receptors. Science 2005;310:329–332. T cells. PLoS One 2012;7:e33961. 23. Klegeris A, Bissonnette CJ, McGeer PL. Reduction of human monocytic cell neu- 32. Ramirez SH, Reichenbach NL, Fan S, et al. Attenuation of HIV-1 replication in rotoxicity and cytokine secretion by ligands of the cannabinoid-type CB2 receptor. Br macrophages by cannabinoid receptor 2 agonists. J Leukoc Biol 2013;93: J Pharmacol 2003;139:775–786. 801–810. 24. Sheng WS, Hu S, Min X, Cabral GA, Lokensgard JR, Peterson PK. Synthetic can- 33. Chaillon A, Nakazawa M, Anderson C, et al. Effect of cannabis use on human im- nabinoid WIN55,212-2 inhibits generation of inflammatory mediators by IL-1beta- munodeficiency virus DNA during suppressive antiretroviral therapy. Clin Infect Dis stimulated human astrocytes. Glia 2005;49:211–219. 2020;70:140–143. 25. Martin-Moreno AM, Reigada D, Ramirez BG, et al. Cannabidiol and other cannabi- 34. Eisenstein TK, Meissler JJ. Effects of cannabinoids on T-cell function and resistance to noids reduce microglial activation in vitro and in vivo: relevance to Alzheimer’s infection. J Neuroimmune Pharmacol 2015;10:204–216. disease. Mol Pharmacol 2011;79:964–973. 35. Juknat A, Pietr M, Kozela E, et al. Differential transcriptional profiles mediated by 26. Fernandez-Ruiz J, Moreno-Martet M, Rodriguez-Cueto C, et al. Prospects for can- exposure to the cannabinoids cannabidiol and Delta9-tetrahydrocannabinol in BV-2 nabinoid therapies in basal ganglia disorders. Br J Pharmacol 2011;163:1365–1378. microglial cells. Br J Pharmacol 2012;165:2512–2528.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 9 ARTICLE OPEN ACCESS Antiparanodal antibodies and IgG subclasses in acute autoimmune neuropathy

Luise Appeltshauser, MD, Anna-Michelle Brunder, Annika Heinius, Peter Kortv¨ ´elyessy, MD, Correspondence Klaus-Peter Wandinger, MD, Ralf Junker, MD, Carmen Villmann, PhD, Claudia Sommer, MD, Dr. Appeltshauser [email protected] Frank Leypoldt, MD, and Kathrin Doppler, MD

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e817. doi:10.1212/NXI.0000000000000817

Abstract RELATED ARTICLE Objective Editorial Isotyping paranodal To determine whether IgG subclasses of antiparanodal autoantibodies are related to disease fl course and treatment response in acute- to subacute-onset neuropathies, we retrospectively antibodies in in ammatory screened 161 baseline serum/CSF samples and 66 follow-up serum/CSF samples. neuropathies: One step closer to precision care Methods Page e843 We used ELISA and immunofluorescence assays to detect antiparanodal IgG and their subclasses and titers in serum/CSF of patients with Guillain-Barré syndrome (GBS), recurrent GBS (R-GBS), Miller-Fisher syndrome, and acute- to subacute-onset chronic inflammatory demyelinating polyradiculoneuropathy (A-CIDP). We evaluated clinical data retrospectively.

Results We detected antiparanodal autoantibodies with a prevalence of 4.3% (7/161), more often in A-CIDP (4/23, 17.4%) compared with GBS (3/114, 2.6%). Longitudinal subclass analysis in the patients with GBS revealed IgG2/3 autoantibodies against Caspr-1 and against anti– contactin-1/Caspr-1, which disappeared at remission. At disease onset, patients with A-CIDP had IgG2/3 anti–Caspr-1 and anti–contactin-1/Caspr-1 or IgG4 anti–contactin-1 antibodies, IgG3 being associated with good response to IV immunoglobulins (IVIg). In the chronic phase of disease, IgG subclass of one patient with A-CIDP switched from IgG3 to IgG4.

Conclusion Our data (1) conﬁrm and extend previous observations that antiparanodal IgG2/3 but not IgG4 antibodies can occur in acute-onset neuropathies manifesting as monophasic GBS, (2) suggest association of IgG3 to a favorable response to IVIg, and (3) lend support to the hypothesis that in some patients, an IgG subclass switch from IgG3 to IgG4 may be the correlate of a secondary progressive or relapsing course following a GBS-like onset.

From the Department of Neurology (L.A., A.-M.B., C.S., K.D.), University Hospital of Wurzburg;¨ Neuroimmunology Section (A.H., K.-P.W., R.J., F.L.), Institute of Clinical Chemistry, University Hospital of Schleswig-Holstein Campus Kiel; Department of Neurology (P.K.), University Hospital of Magdeburg; and Institute for Clinical Neurobiology (C.V.), University Hospital of Wurzburg,¨ Germany.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary A-CIDP = acute-onset chronic inflammatory demyelinating polyradiculoneuropathy; Caspr-1 = contactin-1-associated protein-1; CBA = cell-based assay; CIDP = chronic inflammatory demyelinating polyradiculoneuropathy; GBS = Guillain-Barré syndrome; HEK = human embryonic kidney; HLA = human leukocyte antigen; ICU =intensivecareunit;IVIg =IV immunoglobulin; MFS = Miller-Fisher syndrome; NF =neurofascin;OD = optical density; R-GBS = recurrent GBS.

Autoantibodies against the paranodal antigens contactin-1, et al.18 and 18 patients as Miller-Fisher syndrome (MFS). In 23 contactin-associated protein-1 (Caspr-1), and neurofascin-155 patients, the initial diagnosis was GBS, but was later reverted to (NF155) have been described as biomarkers for a new entity of CIDP because of a disease progression >2 months (diagnostic – inflammatory neuropathies classified as paranodopathies.1 3 In certainty according to the EFNS criteria19:definite CIDP in 10 the chronic phase of disease, IgG4-seropositive patients do not patients, probable CIDP in 3 patients, possible CIDP in 1 – respond to IV immunoglobulins (IVIg), but to rituximab.4 7 patient, and EFNS electrodiagnostic criteria not fulfilled in Noninflammatory IgG4 autoantibodies are pathogenic, possibly 9/23 patients). Eighteen of 23 patients with CIDP fulfilled the by inhibition of the interaction between contactin-1/Caspr-1 criteria for acute-onset CIDP (A-CIDP) proposed in previous – – and NF155 and by NF155 depletion.8 11 Autoantibodies of the publications,20 22 and 5/23 patients with CIDP showed a sub- IgG3 subclass have been described (1) in monophasic disease, acute-onset (peak ≤90 days). In the following, all patients with (2) at the subacute onset, (3) in patients with anti–pan- acute- to subacute-onset CIDP are referred to as part of the neurofascin autoantibodies and severe course of disease, and A-CIDP cohort. Follow-up sera and CSF samples were avail- (4) most recently in chronic inflammatory demyelinating poly- able in 66 patients, including 3 seropositive patients. We in- radiculoneuropathy (CIDP) with clinical features indistinguish- cluded sera of 40 healthy controls recruited in former studies.6 able from seronegative patients but with a good response to Sera of all patients had already been tested for anti-NF155 – IVIg.6,7,11 13 Proinflammatory IgG3 antibodies lead to comple- autoantibodies in a previous study.12 ment deposition in vitro14 and in vivo, resulting in reversible conduction failure in Lewis rats intraneurally injected with Binding assays on murine teased fibers anti–contactin-1 IgG315 and may therefore play a role in the We used binding assays on murine teased fibers at a serum di- acute onset of paranodopathies. However, data on antiparanodal lution of 1:100 and 1:500 to screen all patients’ and controls’ sera 23 autoantibodies in Guillain-Barré syndrome (GBS) or the acute for antiparanodal autoantibodies as previously described. Tit- onset of CIDP are scarce because previous studies mostly fo- ers of seropositive samples were measured with a dilution series TM cused on CIDP and patients were recruited during the chronic using Cy3 -conjugated goat anti-human IgG 1:300 (Jackson, phase of disease. IgG subclass distribution and the associated West Grove). In positive sera, we performed double immuno- clinical phenotype have never been investigated longitudinally. fluorescence with sera diluted 1:50 and 1:100 and rabbit anti– We therefore aimed at determining the prevalence and IgG Caspr-1 diluted 1:1,000 (Abcam, Cambridge, United Kingdom), subclass of paranodal autoantibodies in a cohort of patients with using MFP488-conjugated goat anti-human IgG diluted 1:500 TM acute to subacute inflammatory neuropathies including follow- (MoBiTec, Göttingen, Germany) and Cy3 -conjugated don- up of seropositive patients. We hypothesize that IgG subclass key anti-rabbit IgG diluted 1:300 (Jackson) as secondary anti- and titer are related to the course of disease and therapeutic bodies. For subclass analysis, we used FITC-conjugated response in acute-onset paranodopathy. secondary antibodies diluted 1:100, namely mouse anti-human IgG1, IgG4 (Abcam), mouse anti-human IgG2 (Merck, Darm- stadt, Germany), and sheep anti-human IgG3 (Rockland Methods Immunochemicals, Inc., Pottstown, PA). A healthy control serum and serum of patients with anti–Caspr-1IgG3andIgG4 Patients and controls antibodies from a previous study7 stained on separate slides One hundred sixty-one patients with suspected GBS and served as negative and positive controls for double immunoflu- subacute inflammatory neuropathy (peak of symptoms ≤90 orescence and subclass analysis. We tested CSF of seropositive days) who had undergone diagnostic lumbar puncture at the patients of the first assessment and follow-up on teased fibers in University Hospitals of Kiel and Magdeburg between 2001 and a dilution of 1:20 and 1:50 as described above. CSF of five 2016 were included into the study. We assessed clinical data patients who were seronegative served as negative controls. retrospectively by analysis of discharge letters and documented laboratory, electrophysiologic, and MRI examinations. Table ELISA e-1, links.lww.com/NXI/A274, summarizes demographic data. We performed ELISA for detection of anti–contactin-1 in The diagnosis of GBS was confirmed in n = 114 patients with duplets with sera of all patients and 40 healthy controls as pre- diagnostic certainty according to the Brighton criteria16,17: level viously described.6,12,23,24 For anti–Caspr-1 ELISA, we coated 1 in 73 patients, level 2 in 29 patients, level 3 in 6 patients, and Chinese hamster ovary cell line–derived recombinant human level 4 in 6 patients. We classified 6 patients as recurrent GBS Caspr-1 protein (Research and Diagnostic Systems, Inc., Min- (R-GBS) according to the criteria adapted from Kuitwaard neapolis, MN) at a dilution of 2 μg/mL on MaxiSorb 96-well

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN plates (Thermo Scientific Fisher, Minneapolis, MN) and tested anonymized manner on request from any qualified in- baseline/follow-up sera of seropositive patients and 40 healthy vestigator for purposes of replicating procedures and results. controls or rabbit anti–Caspr-1 (Lifespan Biosciences, Seattle, WA) in duplets, using corresponding horseradish peroxidase– conjugated secondary antibodies6 andconjugatedanti-rabbit Results IgG (DakoCytomation, Glostrup, Denmark). For each sample, Antiparanodal autoantibodies of different IgG we subtracted corresponding values of uncoated duplet wells to subclasses in 4.3% of the patients reduce background signals. We set the threshold for both assays We applied immunofluorescence binding assays on murine at 5 SDs above the mean of healthy controls. Antibody titers teased fibers as a screening tool for detection of anti- were determined by serial dilutions. Subclass-specificsecondary 12 paranodal autoantibodies. We observed serum binding to antibodies served for subclass analysis as previously described. the paranodal junctional region colocalizing with the com- We tested CSF of seropositive patients in a dilution of 1:20 in mercial anti–Caspr-1 antibody (figure 1A) in seven patients anti–contactin-1 ELISA and anti–Caspr-1 ELISA (exception: of the total cohort (7/161; prevalence 4.3%), in none of the patient 4 baseline CSF in both ELISAs and patient 3 in anti- patients of the R-GBS (0/6) and MFS (0/18) subcohort, nor Caspr-1 ELISA due to lack of material) as described above, with in the healthy controls (0/40), but in 3/114 (2.6%) of the CSFof5seronegativepatientsasnegativecontrols. GBS subcohort and 4/23 (17,4%) of the A-CIDP subcohort. Preabsorption experiments Thus, paranodal autoantibodies occurred more frequently To determine the specific paranodal target of the autoanti- in A-CIDP compared with GBS (OR = 7.79, 95% CI – bodies, we serially preincubated seropositive sera and serum 1.61 37.95). Two of the sera with paranodal binding (1 GBS fi – from a healthy control with contactin-1 (CNTN1)– and and 1 A-CIDP) had been identi ed as anti NF155 sero- 12 fi Caspr-1 (CNTNAP1)–transfected human embryonic kidney positive in a recent study ; therefore, experimental ndings (HEK) 293 cells as previously described7 and afterward and clinical associations of these patients are not reported fi performed binding assays on murine teased fibers as de- here. Target antigens of the other ve seropositive patients fi scribed above. were determined using cell-based assay (CBA; gure 1B) and ELISA (see below) and confirmed by preabsorption Cell-based assay experiments (figure 1C). In the GBS subcohort, the two For specification, we tested sera of patients positive in teased seropositive patients showed anti–Caspr-1 (patient 1) and fibers assay or ELISA and sera of five controls on HEK293 cells anti–contactin-1/Caspr-1 (patient 2) antibodies. Seroposi- transfected with plasmids of human Caspr-1 (CNTNAP1)and tive A-CIDP patients carried antibodies against either con- rat contactin-1 (CNTN1) as previously described7,23 and used tactin-1 (patient 3), contactin-1/Caspr-1 (patient 4) or anti- polyclonal rabbit anti–contactin-1 IgG (Abcam) and mono- Caspr-1 (patient 5). At the onset of disease, the two patients clonal mouse anti–Caspr-1 IgG1 (Santa Cruz Biotechnology, with GBS showed IgG2 and IgG3 antibodies (table 1). Dallas, TX) for double immunofluorescence in positive During the acute phase at onset, sera of the three patients patients with corresponding Cy3TM-conjugated donkey later classified as A-CIDP carried IgG2/4 (patient 3) and anti-rabbit IgG (Abcam), donkey anti-mouse IgG, and IgG2/3 (patients 4 and 5), as detected by binding assays and MFP488-conjugated goat anti-human IgG (MoBiTec) sec- ELISA (figure 2, A and B). Autoantibody titers are shown in ondary antibodies. Sera of patients with anti–Caspr-1 and table 1. The results of the binding assays on teased fibers, anti–contactin-1 antibodies from previous studies6,7 served ELISA, and CBA showed concordance regarding the specific as positive controls and sera of healthy controls as negative target antigen of the antibodies and the predominating IgG controls in the double immunofluorescence assays. subclass. There were only minor differences of the antibody titer and coexisting subclasses in ELISA and teased fibers Statistical analysis assay. In case of double positivity for anti-contactin-1/ We performed statistical analysis (descriptive statistics, Caspr-1(patients2and4),preabsorptionassaysdetected calculation of OR, and 95% CI) with SPSS 25.0 (IBM, only the dominating antibody (table 1). Armonk, NY). Subclass switch from IgG3 to IgG4 at follow-up Standard protocol approvals, registrations, in one patient and patient consents In 66/161 patients, follow-up serum and CSF samples were The Ethics Committees of the University of Würzburg and available, including three seropositive patients (1 GBS and 2 the University of Kiel approved to conduct this study. Written A-CIDP). None of the follow-up sera of initially seronega- informed consent was obtained from all participants in the tive patients or controls showed positive results in the teased study. fibers assay (data not shown) nor in the contactin-1 ELISA (figure 3A). Serum of the patient with GBS acquired in Data availability remission 55 months after the first assessment, when the Data not published within this article are available at the patient did not show any remaining neurologic deficits, was University Hospital of Würzburg or will be shared in an seronegative in all experimental assays (figure 3, B and C).

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 Figure 1 Detection of antiparanodal autoantibodies by binding assays on teased fibers, CBA, and preabsorption assay

(A) Colocalization overlay photomicrographs of double immunofluorescence assays with commercial autoantibody displayed in red and human serum in green. Colocalization at the paranodes (arrows) appears yellow and indicates antiparanodal autoantibodies in patients 1–5(A.b–A.f), but not in the negative control serum (A.a). (B) The specific target of the autoantibodies is illustrated by colocalization overlay photomicrographs on transfected HEK293 cells (arrowheads) with commercial antibody in red, serum in green, and nucleus staining in blue. Colocalization appears yellow (merge) and confirms anti– contactin-1 in patients 2 (B.c), 3 (B.d), and 4 (B.e), but not in patients 1 (B.b), 5 (B.f), and the negative control (B.a). Sera of patients 1, 2, 4, and 5 (B.h, B.i, B.k, and B.l) colocalize on Caspr-1 (CNTNAP1)–transfected HEK293 cells as proof of anti–Caspr-1 autoantibodies, whereas sera of the negative control (B.g) and patient 3 (B.j) do not show anti–Caspr-1 positivity. (C) Single immunofluorescence on murine teased fibers after HEK293 contactin-1 and Caspr-1 preabsorption shows that antiparanodal autoantibodies disappear after contactin-1 preabsorption in patient 3 (C.d) and after Caspr-1 preabsorption in patient 4 (C.f), proving specificity to the respective target antigen. Scale bar = 10 μm. CBA = cell-based assay; HEK = human embryonic kidney.

Follow-up serum of one patient with A-CIDP two months subclasses between the acute and chronic stage of the disease after onset, i.e., still in the acute phase of disease (patient 5), (figure 2C). showed IgG3 autoantibodies against Caspr-1, with no difference in antibody titer and subclass compared with onset Intrathecal antiparanodal autoantibodies in (figure 3, B and C). A follow-up-sample during the chronic patients with high antibody titer and blood- stage (28 months after the onset) revealed a reduction of the brain barrier dysfunction titer (1:50 in teased fibers,negativeinELISA,table1)of ELISA with CSF samples of the seropositive patients and five anti–Caspr-1 IgG autoantibodies with subclasses not mea- disease controls showed elevated optical density (OD) at surable due to the low titer (data not shown). The patient a dilution of 1:20 only in anti–contactin-1 patient 3 with now presented with only very mild sensorimotor impair- a very high serum titer (see appendix e-1, links.lww.com/ ment and ameliorated nerve conduction studies. Follow-up NXI/A275) and anti–Caspr-1 patient 5 with signs of se- serum and CSF of the other patient with A-CIDP (patient 4) verely disrupted blood-brain barrier (high CSF protein, MRI who was initially IgG3–anti-contactin-1/Caspr-1 double cauda equina enhancement, see appendix e-1, links.lww. positive were acquired 120 months after the acute onset, com/NXI/A275), but equally low OD values for the other when the patient had developed relapsing-remitting senso- seronegative and seropositive patients (see figure e-1, links. rimotor CIDP. At follow-up, we could no longer detect lww.com/NXI/A273). Binding assays on murine teased anti–contactin-1, but detected and confirmed anti–Caspr-1 fibers showed weak binding of CSF to the paranodes in these autoantibodies via anti–Caspr-1 ELISA, CBA, and pre- two patients, but not in the other seropositive patients or absorption assay (table 1 and figure 3, B and C). Subclass controls. None of the seropositive patients showed signs of analysis now revealed reactivity against IgG2/4, but not intrathecal autoantibody production, with CSF IgG index IgG3 subclass anymore, giving evidence of a switch of IgG and antiparanodal autoantibody specific ASI (antibody

Table 1 Summary of the autoantibody status of seropositive patients 1–5 at baseline and follow-up

CSF Serum

Teased fibers and ELISA Preabsorption CBA

Paranodal binding/OD Titer IgG subclass

Patient 1 (baseline, no follow-up)

Teased fibers Negative Positive 1:100 IgG3

Contactin-1 0.202 0.465 n.d. n.d. Negative Negative

Caspr-1 0.072 2.961 1:200 n.m. Positive Positive

Patient 2 (baseline)

Teased fibers Negative Positive 1:200 n.d.

Contactin-1 0.247 3.538 1:200 IgG2 Negative Positive

Caspr-1 0.129 3.965 1:1,000 IgG2 Positive Positive

Patient 2 (follow-up)

Teased fibers n.d. Negative n.d. n.d.

Contactin-1 n.d. 0.039 n.d. n.d. n.d. Negative

Caspr-1 n.d. 0.025 n.d. n.d. n.d. Negative

Patient 3 (baseline, no follow-up)

Teased fibers Negative Positive 1:30,000 IgG4

Contactin-1 0.516 4.492 1:40,000 IgG4/2 Positive Positive

Caspr-1 n.d. 0.073 n.d. n.d. Negative Negative

Patient 4 (baseline)

Teased fibers Negative Positive 1:500 IgG3

Contactin-1 0.267 2.294 1:200 IgG3 Negative Positive

Caspr-1 n.d.* 3.011 1:500 IgG3/2 Positive Positive

Patient 4 (follow-up)

Teased fibers Negative Positive 1:500 IgG4

Contactin-1 0.227 0.206 n.d. n.d. Negative Negative

Caspr-1 0.138 3.618 1:500 IgG2/4 Positive Positive

Continued 5 6 erlg:Nuomuooy&Nuonlmain|Vlm ,Nme etme 00Neurology.org/NN 2020 September | 5 Number 7, Volume | Neuroinflammation & Neuroimmunology Neurology:

Table 1 Summary of the autoantibody status of seropositive patients 1–5 at baseline and follow-up (continued)

CSF Serum

Teased fibers and ELISA Preabsorption CBA

Paranodal binding/OD Titer IgG subclass

Patient 5 (baseline)

Teased fibers Positive Positive 1:100 IgG3

Contactin-1 0.219 0.260 n.d. n.d. Negative Negative

Caspr-1 0.136 n.d.* n.d.* n.d.* Positive Positive

Patient 5 (first follow-up)

Teased fibers Positive Positive 1:100 IgG3

Contactin-1 0.236 0.243 n.d. n.d. Negative Negative

Caspr-1 0.364 3.802 1:200 IgG3/2 Positive Positive

Patient 5 (second follow-up)

Teased fibers n.d. Positive 1:50 n.m.

Contactin-1 n.d. 0.168 n.d. n.d. n.d. Negative

Caspr-1 n.d. 0.013 n.d. n.d. n.d. Negative

Abbreviations: A-CIDP = acute-onset chronic inflammatory demyelinating polyradiculoneuropathy; Caspr-1 = contactin-1–associated protein-1; CBA = cell-based assay; GBS = Guillain-Barr´e syndrome; n.d. = not done; n.d.* = not done due to lack of material; n.m. = not measurable; OD = optical density. Results of teased fibers assays are displayed in the first row for each patient; contactin-1 and Caspr-1-specific results are displayed in lines 2 and 3. ODs above the threshold and results considered positive are highlighted in bold letters. There is a 100% concordance regarding target specificity in the teased fibers assay, ELISA, and CBA, but minor differences in IgG titer and coexisting subclass between the teased fibers assay and ELISA. In case of double positivity, preabsorption assays only detect the target antigen with the higher antibody titer (see ELISA titer). Figure 2 Serum subclass analysis at baseline and at follow-up with IgG subclass switch

(A) Results of anti–contactin-1 and anti–Caspr-1 ELISA with baseline sera and subclass-specific IgG1-4 autoantibodies (grayscale) are shown as % of the OD of the total IgG (using anti-human IgG autoantibody) on y-axis. In patient 1, ELISA IgG subclasses were not measurable. ELISA revealed mainly IgG2 subclass in patient 2 (GBS), IgG2/IgG4 in patient 3 (A-CIDP), and IgG2/3 in patients 4 and 5 (A-CIDP). (B) Acute phase teased fibers subclass analysis with baselinesera showed binding to the paranodal regions (arrows) only when using IgG3-specific secondary antibody in patients 1 (GBS) and 4 (A-CIDP), visualized by photomicrographs of single immunofluorescence staining on teased fibers. (C) Anti–Caspr-1 IgG subclass ELISA with follow-up serum of patient 4 now revealed IgG 2/4 subclass. Photomicrographs show immunofluorescence staining of murine teased fibers with subclass-specific IgG3 and IgG4 autoantibodies at follow-up. Paranodal binding disappeared using IgG3 antibody, but occurred using IgG4 subclass-specific antibody, proving subclass switch in patient 4. Scale bar = 10 μm. A-CIDP = acute-onset chronic inflammatory demyelinating polyradiculoneuropathy; GBS = Guillain-Barr´e syndrome. speciﬁcity index) being in a normal range (see appendix e-1, three patients with A-CIDP presented with moderate to severe links.lww.com/NXI/A275). paresis and multimodal sensory impairment at disease onset. During the course of disease, all three patients with A-CIDP Clinical features of seropositive patients reported neuropathic pain requiring pregabalin/gabapentin Figure 4 schematically summarizes the clinical course related or even opioid treatment. Two of three patients with to the experimental results. Appendix e-1, links.lww.com/ A-CIDP presented sensory ataxia. Nerve conduction studies NXI/A275, summarizes clinical data of seropositive patients showed demyelinating features in all seropositive patients 1–5 and provides a detailed description of the clinical disease with conduction block in 3/5 patients. Sural nerve biopsy course in individuals. The two seropositive patients with was performed in one patient (patient 5) six months after the GBS were both middle-aged women presenting with severe onset and showed axonal loss without signs of demyelination. motor, but only mild sensory involvement. Cranial nerve CSF analysis revealed cytoalbuminologic dissociation and involvement, autonomic symptoms, and neuropathic pain blood-brain barrier dysfunction in all seropositive patients. were present in 1/2 patients with GBS only. Both patients The patient with subacute-onset CIDP, highly elevated CSF required treatment on the intensive care unit (ICU). The protein, and detection of anti–Caspr-1 in both serum and

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 7 Figure 3 Antiparanodal IgG autoantibodies at follow-up assessment

(A) Contactin-1 IgG ELISA of all patients at first assessment, of 66/161 patients at follow-up, and of 40 healthy controls. Optical density (OD) at 450 nm is displayed on the y-axis. The threshold for positive results is set at 5 SDs above the mean of controls (0.615). Patients 2–4 show positive results at first assessment. The other patients and controls show values below the threshold. Sera of patients 2, 4, and 5 are negative in anti–contactin-1 IgG ELISA at follow- up, as well as 63 other follow-up sera. Patients 1 and 3 were lost to follow-up. (B) Overlay photomicrographs of teased fibers double immunofluorescence and (C) contactin-1– and Caspr-1 (CNTNAP1)–transfected HEK293 cells show loss of anti–contactin-1/Caspr-1 autoantibodies in patient 2 (GBS) at follow-up (B.a, C.a, C.d), but colocalization of commercial anti–Caspr-1 antibody and sera at paranodes in patients 4 and 5 with A-CIDP (B.b, B.c, arrows). Both sera bind to Caspr-1 (CNTNAP1)–transfected HEK293 cells (C.e, C.f, arrowheads), but not to contactin-1–transfected cells (C.b, C.c). Scale bar = 10 μm. A-CIDP = acute-onset chronic inflammatory demyelinating polyradiculoneuropathy; GBS = Guillain-Barr´e syndrome; HEK = human embryonic kidney.

CSF (patient 5) developed action tremor during the course Discussion of disease. We detected antiparanodal autoantibodies in 4.3% of patients with Antibody status is concordant with disease acute to subacute immune-mediated neuropathy, more frequently activity and therapy response in CIDP with acute to subacute onset than in GBS. We found One of two patients with GBS (patient 2) received IVIg, IgG2/3 in monophasic GBS, which supports the notion of pre- which did not lead to amelioration. This patient was IgG2 > vious studies that IgG3 is the predominant subclass in monophasic IgG3 seropositive (patient 2, figure 2A). Both patients with disease.7 By analyzing sera from different time points, we could GBS responded well to plasma exchange. The two patients show that the antibody status was concordant with the disease with A-CIDP and IgG3 autoantibodies (patients 4 and 5) severity. Therefore, our data support pathogenicity of autoanti- showed good response to IVIg treatment in the acute phase of bodies and suggest that autoantibody status and titer are valid disease. Ongoing IVIg treatment led to further improvement indicators for disease activity, as proposed in previous studies on – of symptoms and was stopped after 25 cycles in patient 5 paranodopathy and further IgG4-related neurologic diseases.25 27 because of a remission of symptoms. In the other patient with A-CIDP and anti–contactin-1/Caspr-1 IgG3 at disease onset In one patient, we could demonstrate a subclass switch from (patient 4), IVIg showed a good therapeutic effect in the acute IgG3 in acute disease to IgG4 in chronic disease, a finding that phase and led to remission not requiring further therapy. In we had hypothesized in a previous study.6 In other autoimmune the chronic phase, IVIg was initiated again and stopped after diseases, e.g., in membranous glomerulonephritis, subclass 26 cycles because of a loss of the therapeutic effect. At that switch of anti-PLA2R autoantibodies from IgG1 to IgG4 also stage, the autoantibody subclass had already switched from associates with progression to chronic disease.28 According to IgG3 to IgG4 (for detailed treatment response in the course of a linear model of autoantibody response, unspecificIgMre- disease, see appendix e-1, links.lww.com/NXI/A275). The sponse is followed by IgG isotype, produced by antigen-experi- patient with A-CIDP and high-titer anti–contactin-1 IgG4 enced B cells. These IgG are of the IgG3 subclass and autoantibodies (patient 3) had to be treated on ICU because sequentially switch to IgG1, 2 and 4.29,30 Concurrently, auto- of the severity of symptoms, developed renal dysfunction, did antibodies strongly gain antigen affinity, with IgG4 showing the not respond to IVIg treatment, and did not ameliorate during highest affinity to its target.31,32 Disease progression in para- the course of disease. nodopathy might therefore depend on IgG subclass properties

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Figure 4 Scheme of the autoantibody status in the course of disease in seropositive patients

The course of disease is shown on the x-axis by pseudo-logarithmic display of the time course. Time point 0 is set at baseline assessment of serum/CSF. The color code indicates severity of symptoms. Results of serum assessment are displayed in black. Patients 1 and 3 were lost to follow-up.

and affinities. Gain of specificity and affinity at switch from IgG3 validated in larger studies, these findings might have a direct to IgG4 might explain different binding characteristics of anti- impact on the treatment regime in affected patients. Testing of paranodal autoantibodies in the patient whose IgG3 autoanti- autoantibody subclass during the course of disease may be bodies reacted both against contactin-1/Caspr-1 epitopes, but useful as an indicator of a positive response to IVIg in case of whose IgG4 autoantibodies were specifictoCaspr-1.Here,we IgG3, whereas IgG4-related paranodopathy shows very good report IgG2/4 (1) in patient 4 in the chronic phase after subclass response to rituximab.4,6,7,11 switch and (2) in another patient at disease onset, showing nonameliorating disease with high disability similar to previous The trigger of the autoantibody production and the patho- reports on anti–contactin-1 IgG4-related disease.11,13,33 So far, mechanism behind possible IgG switching are still unknown. studies have only reported antiparanodal IgG4 autoantibodies in Three of five seropositive patients in this study had diabetes patients with chronic disease.7,11,34,35 We therefore suggest that mellitus type 2 as a comorbidity diagnosed years before the high-affinity IgG4 autoantibody binding leads to chronic disease, onset of disease. Previous studies report disruption of the blood- with functional impairment possibly caused by disturbance of the brain barrier and the nodal architecture in patients with diabetes antigen interaction, paranodal architecture, and paranodal mellitus, thereby possibly exposing paranodal antigens to the – complex formation as demonstrated previously.6,8 11 immune response.36,37 A recent study has identified human leukocyte antigen (HLA)-DRB15 as a risk factor for CIDP Patients of our study with predominant IgG3 subclass showed associated with anti-NF155 antibodies, and another study good response to IVIg, whereas patients with IgG2 and IgG4 reported different HLA antibody patterns with subclass switch initially did not respond to IVIg or lost responsiveness in the after renal transplantation.38,39 Whether interindividual differ- disease course. Several studies described good responses to ences in antibody production and switching during the course of IVIg in IgG3-associated paranodopathy, but poor responses in disease depend on (1) HLA alleles, (2) antigen exposure, or (3) IgG4-associated disease.5,6,11,13,34 IgG3 can initiate comple- further immunologic mechanisms has to be investigated in ment activation, opsonization, antigen cross-linking, and in- further studies. ternalization, whereas IgG2 shows little C1q binding and IgG4 completely lacks these Fc-mediated effector functions.31 In an Regarding characteristic clinical features of paranodopathy, we in vivo study on anti–contactin-1 pathogenicity, we suggested report neuropathic pain in anti–Caspr-1–related disease as complement-mediated functional impairment at the paranodes previously described in patients with higher autoantibody tit- as the pathophysiologic correlate in IgG3-mediated para- ers.7 Different titers might explain discrepancies between our nodopathy earlier and could show in vitro that IVIg inhibits results and a recent study that did not report pain in anti–Caspr- complement deposition and activation dose-dependently.14,15 1–positive patients.7,11 In our patients, further features of IgG4- Our data therefore support the hypothesis that therapy re- related chronic paranodopathy such as tremor, sensory ataxia, sponsiveness might depend on IgG subclass and subclass- and nephropathy5,6,13,40 were only present in the course of related pathomechanism of antiparanodal autoantibodies. If disease, suggesting that they evolve due to chronic autoantibody

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 9 exposure. The pathomechanistic correlate of action tremor, Disclosure which recently has also been described in anti–Caspr-1 IgG3- L. Appeltshauser, K-P. Wandinger, R. Junker, C. Sommer, F. related paranodopathy,11 is still under research, but studies Leypoldt, and K. Doppler work for an academic institution suggest cerebellar origin.5,6,12 In this study, we detected in- offering commercial antibody diagnostics. A.M. Brunder, A. trathecal autoantibodies in a patient with tremor, as previously Heinius, P. Körtvelyéssy, and C. Villmann report no dis- reported in a case of anti–pan-neurofascin–positive para- closures relevant to the manuscript. C. Sommer has served on nodopathy.12 The antiparanodal autoantibodies most probably scientific advisory boards for Algiax, Alnylam, Air Liquide, diffuse into the intrathecal compartment due to a blood-brain Akcea, Astellas, Bayer, Grifols, Takeda, and UCB. She reports barrier dysfunction, as indicatedbynormalIgGandASIindices being a member of speakers’ bureau and receiving speaker in our patients. Our CSF data suggest that detection of anti- honoraria from Akcea, Alnylam, Novartis, Pfizer, Sanofi- paranodal antibodies in CSF shows little sensitivity and depends Aventis, and Teva. C. Sommer served as a journal editor, on either high titer or strong leakage of the blood-brain barrier. associate editor, or editorial advisory board member for the We therefore encourage further studies on intrathecal anti- European Journal of Neurology, PLoS ONE, and PAIN Reports. paranodal autoantibodies targeting sensitivity, pathogenicity, F. Leypoldt reports speaker honoraria from Bayer, Roche, and association to clinical symptoms. Novartis, and Fresenius, travel funding from Merck, Grifols, and Bayer, and serving on advisory boards for Roche, Biogen, Low prevalence of autoantibodies against paranodal proteins in and Alexion. Go to Neurology.org/NN for full disclosures. acute inflammatory neuropathy, retrospective assessment, and the low number of seropositive patients available for follow-up limit Publication history our study. Clinical characteristics at acute onset and subclass- Received by Neurology: Neuroimmunology & Neuroinflammation related seropositivity as a possible risk factor for chronic pro- March 2, 2020. Accepted in final form May 19, 2020. gression when detected at the onset should be investigated in large multicentric studies, considering low overall prevalence of GBS and A-CIDP as well as further possible confounders in analysis of Appendix Authors paranodopathy-specific clinical symptoms, e.g., coinciding painful Name Location Contribution

diabetic neuropathy. The question of sequential subclass switching Luise University Hospital of Major role in the acquisition, in paranodopathy needs to be addressed in larger prospective Appeltshauser, Wurzburg¨ analysis, and interpretation studies. Nevertheless, our data (1) propose autoantibody subclass MD of the data and drafting and revising the manuscript for screening to be considered in patients with GBS and A-CIDP, as intellectual content results might have direct implications on diagnosis and therapeutic Anna-Michelle University Hospital of Major role in the acquisition regime during the course of disease, and (2) pave the way for Brunder Wurzburg¨ and analysis of the data further clinical and pathomechanistic studies on antiparanodal Annika Heinius University Hospital of Major role in the acquisition autoantibodies in acute to subacute immune-mediated neuropathy. Schleswig-Holstein, and analysis of the data Campus Kiel

Acknowledgment Peter University Hospital of Revising the manuscript for The authors thank Barbara Reuter and Antonia Kohl for their Kortv¨ élyessy, Magdeburg intellectual content and MD contribution of vital superb technical assistance. They thank Maximilian Frömmer for reagents/tools/patients provision and processing of clinical data and PD Dr. Jakob Matschke for provision of slides and images of supplementary Klaus-Peter University Hospital of Revising the manuscript for Wandinger, MD Schleswig-Holstein, intellectual content and morphologic data. The study was supported by a grant of the Campus Kiel contribution of vital Interdisciplinary Center of Clinical Research of the Medical reagents/tools/patients Faculty of Würzburg to KD and CV and a grant of the German Ralf Junker, MD University Hospital of Revising the manuscript for Research Foundation to KD (DFG, DO-2219/1-1). LA is Schleswig-Holstein, intellectual content and Campus Kiel contribution of vital supported by a research fellowship as a clinician scientist of the reagents/tools/patients Interdisciplinary Center of Clinical Research of the Medical Carmen University Hospital of Revising the manuscript for Faculty of Würzburg. FL receives funding from the German Villmann, PhD Wurzburg¨ intellectual content and Ministry of Education and Research (BMBF, 01GM1908A) and contribution of vital the German Research Foundation (DFG, LE3064/2-1). CS reagents/tools/patients receives funding from the German Research Foundation (DFG Claudia University Hospital of Conceptualization of the Sommer, MD Wurzburg¨ study; revising the 328/9-1; DFG UE 171/4-1; DFG SO 328/10-1; DFG, SFB manuscript for intellectual 1158; and DFG SO 328 13/1), from the German Ministry of content; and major role in Education and Research (BMBF CMT-Net PI), and from the interpretation of data European Union (Horizon 2020). Frank Leypoldt, University Hospital of Drafting and revising the MD Schleswig-Holstein, manuscript for intellectual Campus Kiel content; design and Study funding conceptualization of the This study was supported by the Open Access Publication study; and major role in interpretation of data Fund of the University of Würzburg.

10 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN 18. Kuitwaard K, van Koningsveld R, Ruts L, Jacobs BC, van Doorn PA. Recurrent – Appendix (continued) Guillain-Barre syndrome. J Neurol Neurosurg Psychiatry 2009;80:56 59. 19. Van den Bergh PY, Hadden RD, Bouche P, et al. European Federation of Neurological Societies/Peripheral Nerve Society guideline on management of chronic in- Name Location Contribution flammatory demyelinating polyradiculoneuropathy: report of a joint task force of the European Federation of Neurological Societies and the Peripheral Nerve Society— Kathrin University Hospital of Drafting and revising the first revision. Eur J Neurol 2010;17:356–363. Doppler, MD Wurzburg¨ manuscript for intellectual 20. Ruts L, Drenthen J, Jacobs BC, van Doorn PA. Distinguishing acute-onset CIDP from content; design and fluctuating Guillain-Barre syndrome: a prospective study. Neurology 2010;74: conceptualization of the 1680–1686. study; major role in analysis 21. McCombe PA, Pollard JD, McLeod JG. Chronic inflammatory demyelinating poly- and interpretation of data; radiculoneuropathy. A clinical and electrophysiological study of 92 cases. Brain 1987; and study supervision 110:1617–1630. 22. Alessandro L, Pastor Rueda JM, Wilken M, et al. Differences between acute-onset chronic inflammatory demyelinating polyneuropathy and acute inflammatory demyelinating polyneuropathy in adult patients. J Peripher Nerv Syst 2018;23: References 154–158. 1. Fehmi J, Scherer SS, Willison HJ, Rinaldi S. Nodes, paranodes and neuropathies. 23. Doppler K, Appeltshauser L, Kramer HH, et al. Contactin-1 and neurofascin-155/- J Neurol Neurosurg Psychiatry 2018;89:61–71. 186 are not targets of auto-antibodies in multifocal motor neuropathy. PLoS One 2. Uncini A, Vallat JM. Autoimmune nodo-paranodopathies of peripheral nerve: the 2015;10:e0134274. concept is gaining ground. J Neurol Neurosurg Psychiatry 2018;89:627–635. 24. Ng JK, Malotka J, Kawakami N, et al. Neurofascin as a target for autoantibodies in 3. Querol L, Illa I. Paranodal and other autoantibodies in chronic inflammatory neu- peripheral neuropathies. Neurology 2012;79:2241–2248. ropathies. Curr Opin Neurol 2015;28:474–479. 25. Fujita A, Ogata H, Yamasaki R, Matsushita T, Kira JI. Parallel fluctuation of anti-neurofascin 4. Querol L, Rojas-Garcia R, Diaz-Manera J, et al. Rituximab in treatment-resistant 155 antibody levels with clinico-electrophysiological findings in patients with chronic in- CIDP with antibodies against paranodal proteins. Neurol Neuroimmunol Neuro- flammatory demyelinating polyradiculoneuropathy. J Neurol Sci 2018;384:107–112. inflamm 2015;2:e149. doi: 10.1212/NXI.0000000000000149. 26. Niks EH, van Leeuwen Y, Leite MI, et al. Clinical fluctuations in MuSK myasthenia 5. Querol L, Nogales-Gadea G, Rojas-Garcia R, et al. Neurofascin IgG4 antibodies in gravis are related to antigen-specific IgG4 instead of IgG1. J Neuroimmunol 2008; CIDP associate with disabling tremor and poor response to IVIg. Neurology 2014;82: 195:151–156. 879–886. 27. Huijbers MG, Vink AF, Niks EH, et al. Longitudinal epitope mapping in MuSK 6. Doppler K, Appeltshauser L, Wilhelmi K, et al. Destruction of paranodal architecture myasthenia gravis: implications for disease severity. J Neuroimmunol 2016;291: in inflammatory neuropathy with anti-contactin-1 autoantibodies. J Neurol Neuro- 82–88. surg Psychiatry 2015;86:720–728. 28. Huang CC, Lehman A, Albawardi A, et al. IgG subclass staining in renal biopsies with 7. Doppler K, Appeltshauser L, Villmann C, et al. Auto-antibodies to contactin- membranous glomerulonephritis indicates subclass switch during disease progression. associated protein 1 (Caspr) in two patients with painful inflammatory neuropathy. Mod Pathol 2013;26:799–805. Brain 2016;139:2617–2630. 29. Collins AM, Jackson KJ. A temporal model of human IgE and IgG antibody function. 8. Manso C, Querol L, Mekaouche M, Illa I, Devaux JJ. Contactin-1 IgG4 antibodies Front Immunol 2013;4:235. cause paranode dismantling and conduction defects. Brain 2016;139:1700–1712. 30. Valenzuela NM, Schaub S. The Biology of IgG subclasses and their clinical relevance 9. Manso C, Querol L, Lleixa C, et al. Anti-Neurofascin-155 IgG4 antibodies prevent to transplantation. Transplantation 2018;102:S7–S13. paranodal complex formation in vivo. J Clin Invest 2019;129:2222–2236. 31. Vidarsson G, Dekkers G, Rispens T. IgG subclasses and allotypes: from structure to 10. Labasque M, Hivert B, Nogales-Gadea G, Querol L, Illa I, Faivre-Sarrailh C. Specific effector functions. Front Immunol 2014;5:520–532. contactin N-glycans are implicated in neurofascin binding and autoimmune targeting 32. Bruhns P, Iannascoli B, England P, et al. Specificity and affinity of human Fcgamma receptors in peripheral neuropathies. J Biol Chem 2014;289:7907–7918. and their polymorphic variants for human IgG subclasses. Blood 2009;113:3716–3725. 11. Cortese A, Lombardi R, Briani C, et al. Antibodies to neurofascin, contactin-1, and 33. Querol L, Nogales-Gadea G, Rojas-Garcia R, et al. Antibodies to contactin-1 in contactin-associated protein 1 in CIDP: clinical relevance of IgG isotype. Neurol chronic inflammatory demyelinating polyneuropathy. Ann Neurol 2013;73:370–380. Neuroimmunol Neuroinflamm 2019;7:e639. doi: 10.1212/NXI.0000000000000639. 34. Devaux JJ, Miura Y, Fukami Y, et al. Neurofascin-155 IgG4 in chronic inflammatory 12. Stengel H, Vural A, Brunder AM, et al. Anti-pan-neurofascin IgG3 as a marker of demyelinating polyneuropathy. Neurology 2016;86:800–807. fulminant autoimmune neuropathy. Neurol Neuroimmunol Neuroinflamm 2019;6: 35. Ogata H, Yamasaki R, Hiwatashi A, et al. Characterization of IgG4 anti-neurofascin e603. doi: 10.1212/NXI.0000000000000603. 155 antibody-positive polyneuropathy. Ann Clin Transl Neurol 2015;2:960–971. 13. Miura Y, Devaux JJ, Fukami Y, et al. Contactin 1 IgG4 associates to chronic inflammatory 36. Doppler K, Frank F, Koschker AC, Reiners K, Sommer C. Nodes of Ranvier in skin demyelinating polyneuropathy with sensory ataxia. Brain 2015;138:1484–1491. biopsies of patients with diabetes mellitus. J Peripher Nerv Syst 2017;22:182–190. 14. Appeltshauser L, Weishaupt A, Sommer C, Doppler K. Complement deposition 37. Banks W. The blood-brain barrier interface in diabetes mellitus: dysfunctions, induced by binding of anti-contactin-1 auto-antibodies is modified by immunoglo- mechanisms, and approaches to treatment. Curr Pharm Des 2020;26:1438–1447. bulins. Exp Neurol 2017;287:84–90. 38. Arnold ML, Ntokou IS, Doxiadis II, Spriewald BM, Boletis JN, Iniotaki AG. Donor- 15. Doppler K, Schuster Y, Appeltshauser L, et al. Anti-CNTN1 IgG3 induces acute specific HLA antibodies: evaluating the risk for graft loss in renal transplant recipients conduction block and motor deficits in a passive transfer rat model. J Neuro- with isotype switch from complement fixing IgG1/IgG3 to noncomplement fixing inflammation 2019;16:73. IgG2/IgG4 anti-HLA alloantibodies. Transpl Int 2014;27:253–261. 16. Sejvar JJ, Kohl KS, Gidudu J, et al. Guillain-Barre syndrome and Fisher syndrome: case 39. Martinez-Martinez L, Lleixa MC, Boera-Carnicero G, et al. Anti-NF155 chronic in- definitions and guidelines for collection, analysis, and presentation of immunization flammatory demyelinating polyradiculoneuropathy strongly associates to HLA- safety data. Vaccine 2011;29:599–612. DRB15. J Neuroinflammation 2017;14:224. 17. Fokke C, van den Berg B, Drenthen J, Walgaard C, van Doorn PA, Jacobs BC. 40. Hashimoto Y, Ogata H, Yamasaki R, et al. Chronic inflammatory demyelinating Diagnosis of Guillain-Barre syndrome and validation of Brighton criteria. Brain 2014; polyneuropathy with concurrent membranous nephropathy: an anti-paranode and 137:33–43. podocyte protein antibody study and literature survey. Front Neurol 2018;9:997.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 11 ARTICLE OPEN ACCESS CLASS OF EVIDENCE CSF chitinase 3-like-1 association with disability of primary progressive MS

Francisco Pérez-Miralles, MD, PhD, Daniel Prefasi, MD, PhD, Antonio Garc´ıa-Merino, MD, PhD, Correspondence ´ Francisco Gascon-Gim´ énez, MD, Nicolás Medrano, MD, Jessica Castillo-Villalba, MD, Laura Cubas, MD, Dr. Perez-Miralles [email protected] Carmen Alcalá, MD, PhD, Sara Gil-Perot´ın, MD, Roc´ıoGomez-Ballesteros,´ BSci, MBA, Jorge Maurino, MD, Esther Alvarez-Garc´ ´ıa, DVM, PhD, and Bonaventura Casanova, MD, PhD

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e815. doi:10.1212/NXI.0000000000000815

Abstract MORE ONLINE Objective Class of Evidence To assess the role of CSF chitinase 3-like-1 (CHI3L1), chitinase 3-like-2 (CHI3L2), and Criteria for rating neuroﬁlament light chain (NfL) in predicting the course of primary progressive MS (PPMS). therapeutic and diagnostic studies Methods NPub.org/coe We analyzed CSF CHI3L1, CHI3L2, and NfL levels in 25 patients with PPMS with disease duration ≤10 years and no disease-modifying therapy for ≥6 months from the prospective Understanding Primary Progressive Multiple Sclerosis cohort study. CSF samples taken at disease diagnosis were analyzed using commercial ELISAs and following the manufacturer’s instructions. Data on Expanded Disability Status Scale (EDSS) scores, disability progression, and cognitive function according to the Brief Repeatable Neuropsychological Battery were also assessed throughout the 1-year study follow-up.

Results Increasing CHI3L1 levels correlated with higher EDSS scores at baseline (ρ = 0.490, 95% CI 0.118–0.742, p = 0.013) and month 12 (ρ = 0.455, 95% CI 0.063–0.725, p = 0.026) and tended to be associated with a higher risk of disability progression according to EDSS scores (OR = 1.008, 95% CI 0.999–1.017, p = 0.089). Increasing CHI3L2 levels also tended to correlate with lower baseline EDSS scores (ρ = −0.366, 95% CI -0.676–0.054, p = 0.086). There was no correlation with regard to NfL levels.

Conclusions This analysis supports the association between CSF CHI3L1 levels and neurologic disability according to EDSS scores in patients with PPMS. Other chitinase-like proteins such as CHI3L2 may also be involved.

Classification of evidence This study provides Class II evidence that CSF CHI3L1 is associated with neurologic disability in patients with PPMS.

From the Neuroimmunology Unit-National Referral Center for Demyelinating Diseases (F.P.-M., C.A., B.C.), Hospital Universitari i Polit`ecnic La Fe, Valencia, Spain; Medical De- partment (D.P., N.M., R.G.-B., J.M.), Roche Farma S.A., Madrid, Spain; Neurology Department (A.G.-M.), Hospital Universitario Puerta de Hierro, Madrid, Spain; Neurology Department (F.G.-G.), Hospital Cl´ınico Universitario de Valencia, Valencia, Spain; Multiple Sclerosis and Neuroregeneration Research Group (J.C.-V., L.C., S.G.-P.), Institut d’Investigacio´ Sanit`aria La Fe, Valencia, Spain; and Dynamic Science S.L. (E.A.-G.),´ Madrid, Spain.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

The Article Processing Charge was funded by Roche. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary 9-HPT = 9-Hole Peg Test; CHI3L1 = chitinase 3-like-1; CHI3L2 = chitinase 3-like-2; CIS = clinically isolated syndrome; EDSS = Expanded Disability Status Scale; NfL = neuroﬁlament light chain; PPMS = primary progressive MS; T25-fw = Timed 25-Foot Walk; UPPMS = Understanding Primary Progressive Multiple Sclerosis.

Several studies have explored CSF biomarkers reflecting in- Data on Expanded Disability Status Scale (EDSS) scores, dis- flammatory and neurodegenerative processes underlying MS, ability progression, and cognitive function were collected suggesting the prognostic role of chitinase 3-like-1 (CHI3L1), throughout the 1-year study follow-up. Disability progression chitinase 3-like-2 (CHI3L2), and neurofilament light chain was assessed and confirmed ≥3 months later according to the – (NfL) in conversion to definite MS.1 4 Reported data have also EDSS score (≥1-point increase when the baseline score was pointed at measuring CHI3L1 and NfL to anticipate conver- ≤5.0 or ≥0.5-point increase when the baseline score was ≥5.5), sion to progressive disease,5 as well as supporting the biomarker Timed 25-Foot Walk (T25-fw; ≥20% increase from baseline), use of CHI3L1 to monitor disease activity in secondary pro- 9-Hole Peg Test (9-HPT; ≥20% increase from baseline), and gressive MS6 and response to interferon-beta in relapsing- a composite of previously mentioned variables. The Rao Brief remitting patients.7 These proteins may also be involved in the Repeatable Neuropsychological Battery was used to assess pathogenesis of the different MS forms, with a progressive in- cognitive function.9 crease in CHI3L1 levels with disease stages from clinically isolated syndrome (CIS), and an opposite variation of CHI3L2 Statistical analyses expression in relapsing-remitting and progressive forms.1 CHI3L1, CHI3L2, and NfL levels were correlated with EDSS However, they have not been fully characterized in patients scores at enrollment (baseline) and month 12 using Spear- with primary progressive MS (PPMS), and their role as pre- man rank correlations. Associations of these levels with 12- dictive biomarkers of disease course remains unknown. month disability progression and cognitive assessments were analyzed using Spearman rank correlations and logistic Therefore, we approached the analysis of CSF CHI3L1, regressions. Indeterminate/missing data were not considered CHI3L2, and NfL levels to assess their role in predicting in the analyses, which were performed by Dynamic Science disease course in patients with PPMS. S.L. (Madrid, Spain) with the Statistical Package for the Social Sciences (SPSS) version 22.0 (SPSS Inc, Chicago) and a significance level of 0.05. Methods Standard protocol approvals, registrations, Study design and participants and patient consents This pilot CSF analysis included a sample of patients with PPMS The study was conducted according to Good Pharmacoepi- from the prospective Understanding Primary Progressive Multiple demiology Practices, the Declaration of Helsinki, and national Sclerosis (UPPMS) cohort study and provides Class II evidence. regulations. The ethics committee of Hospital Universitario 12 de Octubre approved it, and all patients gave their written Patients were consecutively recruited from neurologic informed consent. departments at 11 Spanish hospitals between January and July 2017. Key eligibility criteria included age ≥18 years, PPMS Data availability 8 diagnosis according to the 2010 McDonald criteria, disease All data are reported within the article and available anony- ≤ duration 10 years from neurologic symptom onset, and no mized by request from qualified investigators. disease-modifying therapy within the past 6 months. Assessments Results When CSF samples taken at MS diagnosis were available, an aliquot was sent to the Laboratory of Neuroimmunology “Dr. Twenty-five of the 55 patients enrolled in the UPPMS study Francisco Coret”—Institut d’Investigació Sanitària La Fe had CSF samples taken at disease diagnosis (table). (Valencia, Spain). Commercially available ELISAs for CHI3L1 (Quantikine ELISA kit, R&D Systems Inc., Minne- Increasing CHI3L1 levels correlated with higher EDSS scores at apolis), CHI3L2 (Human CHI3L2 ELISA kit, Cusabio baseline (Spearman ρ = 0.490, 95% CI 0.118–0.742, p = 0.013) Technology LLC, Houston), and NfL (NF-light neurofila- and month 12 (Spearman ρ = 0.455, 95% CI 0.063–0.725, p = ment ELISA kit, UmanDiagnostics AB, Umea, Sweden) were 0.026) (figure). Conversely, increasing CHI3L2 levels tended to used according to the manufacturer’s instructions. Mean correlate with lower baseline EDSS scores (Spearman ρ = intra-assay variations were 6%, 8.9%, and 1.7% for CHI3L1, −0.366, 95% CI −0.676 to 0.054, p = 0.086) (figure). There was CHI3L2, and NfL, respectively; all interassay variation coef- no significant Spearman rank correlation between NfL levels and ficients were <15%. EDSS scores at baseline or month 12 (figure).

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN increasing CSF CHI3L1 expression was associated with Table Patient characteristics (N = 25) a higher and earlier risk of developing a sustained EDSS score ≥ 3 Characteristic Value 3 within 5 years after CIS. The CHI3L1 is a member of the glycoside hydrolase 18 chitinase family, whose increase Age (y), median (IQR) throughout MS stages seems to derive from the extension of At MS diagnosis 52.7 (50.6–58.7) diffuse brain inflammation associated with neurologic impairment rather than acute inflammation during relapses1 and At baseline 56.3 (53.1–61.9) the detrimental role of astrocyte activation in disease Sex, n (%) pathogenesis.3

Male 13 (52.0) In addition, there was a trend toward ﬁnding lower CHI3L2 Female 12 (48.0) levels at disease diagnosis to be related with higher EDSS EDSS score, median (IQR) scores at study baseline and vice versa, suggesting a potential ﬁ At MS diagnosis 3.5 (2.5–4.5)a predictive role in PPMS that deserves further con rmation. CHI3L2 is another member of the glycoside hydrolase 18 At baseline 4.0 (3.8–6.0) chitinase family that was reported to predict MS development 10 At month 12 4.3 (3.6–6.0)b in patients with CIS. Contrasting to CH3L1, CHI3L2 levels are lower in progressive vs relapsing MS, which might result CSF samples taken at MS diagnosis, median 1 (IQR) from their distinct temporal pattern of expression. However, we found no evidence on its role in predicting EDSS scores – CHI3L1 level (ng/mL) 199.1 (140.8 266.8) either in this patient population or in patients with PPMS. CHI3L2 level (ng/mL) 48.0 (24.6–64.2)a

NfL level (pg/mL) 470.7 (314.5–601.8)b Furthermore, our analyses provide no evidence to suggest the predictive value of NfL in patients with PPMS. As NfL is Time since MS diagnosis (y), median (IQR) 3.4 (1.2–6.4) a cytoskeletal polypeptide of the axon whose release denotes

Abbreviations: CHI3L1 = chitinase 3-like-1; CHI3L2 = chitinase 3-like-2; EDSS = axonal impairment, the lacking association may result from Expanded Disability Status Scale; IQR = interquartile range; NfL = neuro- the underlying progressive neurodegenerative process rather filament light chain. 5 a Missing data, n = 2. than axonal damage outbreaks. b Missing data, n = 1. We acknowledge that the analysis has limitations that may affect its statistical power, including the limited study follow- Three (12.0%) patients experienced disability progression up, sample size, and patients with disability progression. according to EDSS scores at month 12, 1 (4.0%) according to However, we identified significant differences and tendencies T25-fw, none according to 9-HPT, and 4 (16.0%) according deserving confirmation. Despite the absence of significant to a composite of EDSS scores, T25-fw, or 9-HPT. Higher differences or tendencies according to NfL levels, additional CHI3L1 levels tended to be associated with a higher risk of data on larger sample sizes, longer follow-ups, and using disability progression according to this composite (OR = SIMOA technology, preferably in serum, are needed to con- 1.008, 95% CI 0.999–1.018, p = 0.092) and EDSS scores (OR firm its role in patients with PPMS. In addition, the obser- = 1.008, 95% CI 0.999–1.017, p = 0.089), although there was vational, prospective, and multicenter nature of the study no association regarding CHI3L2 (OR = 1.014, 95% CI supports the generalizability of analysis findings to clinical 0.978–1.051, p = 0.450) and NfL levels (OR = 0.996, 95% CI practice. 0.988–1.004, p = 0.327). There was no association with other disability progression measurements or cognitive assessments This analysis supports the association between CSF CHI3L1 (data not shown). levels and neurologic disability according to EDSS scores in patients with PPMS. Other chitinase-like proteins such as CHI3L2 may also be involved. These findings warrant further Discussion confirmatory assessment in larger samples of patients with PPMS. This pilot CSF analysis showed that CHI3L1 levels at PPMS diagnosis were related to EDSS scores at study follow-up. Acknowledgment Higher CHI3L1 levels also tended to be associated with The authors are extremely grateful to all the patients and their a higher risk of disability progression according to EDSS families for making the UPPMS study possible. Medical scores, although the relatively short study follow-up and the writing support was provided by Esther Alvarez-Garc´´ ıa reduced sample size limited its power to detect significant (Dynamic Science S.L.) during the preparation of this article differences. These results are in line with a previous report (funded by Roche Farma S.A), who contributed to drafting, showing that CSF CHI3L1 correlated with EDSS scores editing, and word processing the manuscript as well as through years 1–4 after the CIS episode.4 In addition, responding to the editor and reviewers’ comments.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 Figure Correlations between biomarker levels and Expanded Disability Status Scale scores

This figure shows the results of the Spearman rank correlations of chitinase 3-like-1 (A), chitinase 3-like-2 (B), and neurofilament light chain (C) levels with Expanded Disability Status Scale scores.

Study funding Disclosure The authors disclosed receipt of the following financial F. Pérez-Miralles was part of the steering committee of the support for the research, authorship, and publication of this UPPMS study and has received compensation for serving on article: This work was supported by Roche Farma S.A. scientific advisory boards or speaking honoraria from Almirall, (ML39253). Biogen Idec, Genzyme, Merck Serono, Mylan, Novartis, Roche,

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Sanofi-Aventis, and Teva, outside the submitted work. D. Prefasi is an employee of Roche Farma S.A. A. Garc´ıa-Merino has re- Appendix (continued) ceived consultant and/or lecture fees from Merck, Teva, Biogen, Name Location Contribution Novartis, Roche, and Sanofi.F.Gascón-Giménez reports no disclosures. N. Medrano is an employee of Roche Farma S.A. J. Carmen Hospital Universitari i Design and ı Alcala,´ MD, Politècnic La Fe, conceptualization of the Castillo-Villalba, L. Cubas, C. Alcalá, and S. Gil-Perotńreportno PhD Valencia, Spain study; major role in the disclosures. R. Gómez-Ballesteros and J. Maurino are employees acquisition of data; analysis ´ ı and interpretation of the of Roche Farma S.A. E. Alvarez-GarcáisanemployeeofDy- data; and drafting and namic Science S.L. B. Casanova reports no disclosures. Go to revising the manuscript

Neurology.org/NN for full disclosures. Sara Gil- Instituto de Design and Perot´ın, MD Investigacion´ Sanitaria conceptualization of the La Fe, Valencia, Spain study; major role in the Publication history acquisition of data; analysis Received by Neurology: Neuroimmunology & Neuroinﬂammation and interpretation of the January 29, 2020. Accepted in ﬁnal form May 1, 2020. data; and drafting and revising the manuscript

Roc´ıoGomez-´ Roche Farma S.A., Design and Ballesteros, Madrid, Spain conceptualization of the Appendix Authors BSci, MBA study; analysis and interpretation of the data; Name Location Contribution and drafting and revising the manuscript Francisco Hospital Universitari i Design and P´erez- Polit`ecnic La Fe, conceptualization of the Jorge Maurino, Roche Farma S.A., Design and Miralles, MD, Valencia, Spain study; major role in the MD Madrid, Spain conceptualization of the PhD acquisition of data; analysis study; analysis and and interpretation of the interpretation of the data; data; and drafting and and drafting and revising revising the manuscript the manuscript

Daniel Prefasi, Roche Farma S.A., Design and Esther Dynamic Science S.L. Drafting and revising the MD, PhD Madrid, Spain conceptualization of the Alvarez-´ manuscript; responding to study; analysis and Garc´ıa, DVM, the reviewer comments; interpretation of the data; PhD technical editing; and drafting and revising copyediting; and word the manuscript processing

Antonio Hospital Universitario Design and Bonaventura Hospital Universitari i Design and ı Garc´a- Puerta de Hierro, conceptualization of the Casanova, MD, Polit`ecnic La Fe, conceptualization of the Merino, MD, Madrid, Spain study; major role in the PhD Valencia, Spain study; major role in the PhD acquisition of data; acquisition of data; analysis and interpretation analysis and interpretation of the data; and drafting of the data; and drafting and revising the and revising the manuscript 4manuscript

Francisco Hospital Cl´ınico Design and Gascon-´ Universitario de conceptualization of the Giménez, MD Valencia, Valencia, study; major role in the References Spain acquisition of data; analysis 1. Hinsinger G, Galeotti N, Nabholz N, et al. Chitinase 3-like proteins as diagnostic and and interpretation of the prognostic biomarkers of multiple sclerosis. Mult Scler 2015;21:1251–1261. data; and drafting and 2. Arrambide G, Espejo C, Eixarch H, et al. Neurofilament light chain level is a weak risk revising the manuscript factor for the development of MS. Neurology 2016;87:1076–1084. 3. Canto E, Tintore M, Villar LM, et al. Chitinase 3-like 1: prognostic biomarker in Nicolás Roche Farma S.A., Design and clinically isolated syndromes. Brain 2015;138:918–931. Medrano, MD Madrid, Spain conceptualization of the 4. Comabella M, Fernandez M, Martin R, et al. Cerebrospinal fluid chitinase 3-like 1 study; analysis and interpretation of the data; levels are associated with conversion to multiple sclerosis. Brain 2010;133: – and drafting and revising 1082 1093. fl the manuscript 5. Gil-Perotin S, Castillo-Villalba J, Cubas-Nunez L, et al. Combined cerebrospinal uid neurofilament light chain protein and chitinase-3 like-1 levels in defining disease Jessica Instituto de Design and course and prognosis in multiple sclerosis. Front Neurol 2019;10:1008. Castillo- Investigacion´ Sanitaria conceptualization of the 6. Burman J, Raininko R, Blennow K, Zetterberg H, Axelsson M, Malmestrom C. YKL- Villalba, MD La Fe, Valencia, Spain study; major role in the 40 is a CSF biomarker of intrathecal inflammation in secondary progressive multiple acquisition of data; analysis sclerosis. J Neuroimmunol 2016;292:52–57. and interpretation of the 7. Matute-Blanch C, Rio J, Villar LM, et al. Chitinase 3-like 1 is associated with the data; and drafting and response to interferon-beta treatment in multiple sclerosis. J Neuroimmunol 2017; revising the manuscript 303:62–65. 8. Polman CH, Reingold SC, Banwell B, et al. Diagnostic criteria for multiple sclerosis: Laura Cubas, Instituto de Design and 2010 revisions to the McDonald criteria. Ann Neurol 2011;69:292–302. MD Investigacion´ Sanitaria conceptualization of the 9. Boringa JB, Lazeron RH, Reuling IE, et al. The brief repeatable battery of neuro- La Fe, Valencia, Spain study; major role in the acquisition of data; analysis psychological tests: normative values allow application in multiple sclerosis clinical – and interpretation of the practice. Mult Scler 2001;7:263 267. fl data; and drafting and 10. Mollgaard M, Degn M, Sellebjerg F, Frederiksen JL, Modvig S. Cerebrospinal uid revising the manuscript chitinase-3-like 2 and chitotriosidase are potential prognostic biomarkers in early multiple sclerosis. Eur J Neurol 2016;23:898–905.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 5 ARTICLE OPEN ACCESS Painful trigeminal neuropathy associated with anti-Plexin D1 antibody

Takayuki Fujii, MD, PhD, Ryo Yamasaki, MD, PhD, Yukino Miyachi, MSc, Kyoko Iinuma, MSc, Yu Hashimoto, MD, Correspondence Noriko Isobe, MD, PhD, Takuya Matsushita, MD, PhD, and Jun-ichi Kira, MD, PhD Dr. Kira [email protected] Neurol Neuroimmunol Neuroinﬂamm 2020;7:e819. doi:10.1212/NXI.0000000000000819 Abstract Objective To determine whether anti-Plexin D1 antibody (Plexin D1-immunoglobulin G [IgG]), which is associated with limb and trunk neuropathic pain (NP) and binds to pain-conducting small unmyelinated dorsal root ganglion (DRG) neurons, exists in patients with idiopathic painful trigeminal neuropathy (IPTN) and whether Plexin D1-IgG binds to trigeminal ganglion (TG) neurons.

Methods We enrolled 21 consecutive patients with IPTN and 35 age- and sex-matched controls without NP (25 healthy persons and 10 with neurodegenerative diseases). We measured serum Plexin D1-IgG using a mouse DRG tissue–based indirect immunoﬂuorescence assay (IFA) and by Western blotting (WB) using a recombinant human Plexin D1 (rhPlexin D1) accompanied by immunoadsorption tests with rhPlexin D1. The reactivity of Plexin D1-IgG toward mouse TG, brain, heart, and kidney was assessed by tissue-based IFAs.

Results Serum Plexin D1-IgG was detected more frequently in IPTN than in controls by both IFA and WB (14.3% vs 0%, p = 0.048). Three Plexin D1-IgG–positive patients also had limb or trunk NP and commonly showed tongue pain. In tissue-based IFAs, IgG from 2 Plexin D1- IgG–positive patients immunostained small TG neurons, which was prevented by preincubation with rhPlexin D1. Moreover, Plexin D1-IgG immunostaining mostly colocalized with isolectin B4-positive pain-conducting unmyelinated TG neurons. IFAs of other tissues with the same IgG revealed weak immunoreactivity only in endothelial cells, which was prevented by preincubation with rhPlexin D1.

Conclusions Plexin D1-IgG, which binds to pain-conducting small TG neurons in addition to DRG neurons, can be present in IPTN as well as limb and trunk NP.

From the Department of Neurological Therapeutics (T.F., N.I.), Neurological Institute, Graduate School of Medical Sciences, Kyushu University, Higashi-ku, Fukuoka; Department of Neurology (R.Y., Y.M., K.I., Y.H., T.M., J.K.), Neurological Institute, Graduate School of Medical Sciences, Kyushu University, Higashi-ku, Fukuoka; Translational Neuroscience Center (J.K.), Graduate School of Medicine, and School of Pharmacy at Fukuoka, International University of Health and Welfare, Ookawa, Fukuoka; and Department of Neurology (J.K.), Brain and Nerve Center, Fukuoka Central Hospital, International University of Health and Welfare, Chuou-ku, Japan.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary DRG = dorsal root ganglion; HC = healthy control; IB4 = isolectin B4; ICHD-3 = International Classification of Headache Disorders 3rd edition; IFA = immunofluorescence assay; IgG = immunoglobulin G; IPTN = idiopathic painful trigeminal neuropathy; IVIg = IV immunoglobulin; NCS = nerve conduction study; NFH = neurofilament heavy chain; NP = neuropathic pain; PE = plasma exchange; rhPlexin D1 = recombinant human Plexin D1; TG = trigeminal ganglion; TN = trigeminal nerve; WB = Western blotting.

Painful trigeminal neuropathy (PTN) presents with facial assays (IFAs) and Western blotting (WB) using recombinant pain that coincides with the distribution of the trigeminal human Plexin D1 (rhPlexin D1) accompanied by immu- nerves (TNs). PTN develops in a variety of underlying con- noadsorption tests with rhPlexin D1 (R&D Systems, Minne- ditions, but its pathomechanism is frequently undetermined. apolis).2 Before testing, patients’ sera were preabsorbed with The International Classification of Headache Disorders 3rd mouse liver powder (Rockland, Gilbertsville).3 edition (ICHD-3) defines such cases with unknown mecha- – nism as idiopathic PTN (IPTN).1 Mouse tissue based IFAs IFAs were conducted using patient IgG and 4-μmparaffinsec- ff fi We recently reported anti-Plexin D1 antibody (Plexin D1- tions processed from 10% bu ered formalin- xed adult male – 2 immunoglobulin G [IgG]) in the sera of 10% of patients with C57BL/6 mouse tissues (10 12 weeks old). We also per- limb and trunk neuropathic pain (NP).2 Plexin D1-IgG binds to formed double immunostaining of TG neurons with patient IgG – and exerts cytotoxic effects against isolectin B4 (IB4)-positive and Alexa Fluor 594 conjugated anti-IB4 (Thermo Fisher Sci- fi pain-conducting small unmyelinated dorsal root ganglion (DRG) enti c, Waltham, 1:1,000) and with patient IgG and anti- fi neurons.2 NP was improved in all Plexin D1-IgG–positive cases neuro lament heavy chain (NFH) (Covance, Princeton, 1:500). 2 treated with plasma exchange. NP occasionally manifests facial Data availability pain; therefore, we assessed whether Plexin D1-IgG exists in Any data not published within the article will be shared in patients with IPTN and determined whether Plexin D1-IgG anonymized form by request from any qualified investigator. binds to trigeminal ganglion (TG) neurons. Results Methods Detection of Plexin D1-IgG in IPTN Patients Serum Plexin D1-IgG detected by both IFA and WB was more We enrolled 21 consecutive patients with IPTN who attended frequent in patients with IPTN than in controls (3/21 [14.3%] fi our clinic between 2008 and 2019, and we reviewed their vs 0/35 [0%], p =0.048)(gure 1 and table). The overall medical records. An IPTN diagnosis was based on the estab- coincidence rate of positive WB and IFA results was 98.2% lished criteria1: unilateral or bilateral facial pain colocalizing (55/56, 1 control had an immunoreactive IgG band on WB but with one or both TNs, clinically evident positive (hyperalgesia negative immunoreactivity to mouse DRG). Three patients – and allodynia) and/or negative (hypesthesia and hypoalgesia) who were Plexin D1-IgG positive also had limb or trunk NP signs of TN dysfunction, no identified cause, and not better and commonly showed tongue pain, which was more frequent accounted for by another ICHD-3 diagnosis. Patients with compared with patients with IPTN who were Plexin D1- – some underlying diseases were not excluded unless the IgG negative (100% vs 11.1%, p = 0.03) (table). Otherwise, no ff mechanism causing PTN was established. As controls, 35 age- di erence was found in any parameter examined between the 2 and sex-matched subjects without NP were used (25 healthy groups. Here, we present 3 IPTN cases with Plexin D1-IgG. persons and 10 with neurodegenerative diseases). Case 1 A 64-year-old woman had a 15-year history of persistent pain Standard protocol approvals, registrations, and numbness in her tongue and perioral area, which gradually and patient consents spread to all 3 divisions of bilateral TNs and both upper limbs. This study was approved by the Ethical Committee of Kyushu Neurologic examination revealed a decreased sensation to University (#29-40 and #30-164). All patients and controls pinprick and light touch in all TN divisions and the upper provided written informed consent. Animal experiments were limbs. A nerve conduction study (NCS) showed decreased performed according to the protocols approved by the In- sensory action potentials in her upper limbs. Bilateral blink stitutional Animal Care and Use Committee at Kyushu Uni- reflexes showed delayed R1 and R2 responses. IV immuno- versity (A19-109). globulin (IVIg) administration did not improve her symptoms. Plexin D1-IgG detection Case 2 For all participants, serum Plexin D1-IgG was measured by Two months before admission, a 46-year-old woman with both mouse DRG tissue–based indirect immunofluorescence atopic myelitis developed persistent pain and numbness of

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Figure 1 Detection of Plexin D1-IgG by IFA using mouse DRG and WB with rhPlexin D1

(A) IFA with mouse DRG for case 1. IgG from case 1 (green) bound to small DRG neurons (upper images). The immunostaining (green) of small DRG neurons by IgG from case 1 was prevented by preincubation with rhPlexin D1 (2 μg/mL) (lower images). Nuclei are counterstained with 49,6- diamidino-2-phenylindole (DAPI) (blue). (B) IFA with mouse DRG for case 2. IgG from case 2 (green) bound to small DRG neurons (upper images). The immunostaining (green) of small DRG neurons by IgG from case 2 was prevented by preincubation with rhPlexin D1 (2 μg/mL) (lower images). Nuclei are counterstained with DAPI (blue). (C) WB analysis using rhPlexin D1 and IgG from cases 1, 2, and 3, 3 patients with IPTN without Plexin D1-IgG, 3 healthy controls, 3 patients with neurodegenerative diseases, and a commercial anti-human Plexin D1 antibody. WB of rhPlexin D1 with IgG samples from cases 1, 2, and 3 showed a common immunoreactive band between 150 and 250 kDa, similar to the commercial anti-human Plexin D1 antibody (R&D Systems, Minne- apolis). All other samples were non- reactive. Scale bars: (A and B) = 50 μm. DRG = dorsal root ganglion; IFA = immunofluorescence assay; IPTN = idiopathic painful trigeminal neuropathy; ND = neurodegenerative diseases; Plexin D1-IgG = anti- Plexin D1 antibody; rhPlexin D1 = recombinant human Plexin D1; WB = Western blotting. the tongue, face, and a girdle-like sensation around mid- Case 3 trunk. Neurologically, she had hypesthesia in the tongue, An 81-year-old woman with a history of hepatitis B virus– face, and mid-trunk on admission. Spinal cord MRI related liver cancer presented with tingling in her tongue and revealed T2 hyperintense lesions at the level of Th8, the soles of her feet, which she had experienced over the whereas cranial MRI showed no relevant lesions. Her pain preceding 12 years. Neurologic examination revealed a de- in the tongue, face, and mid-trunk was relieved by a com- creased sensation to pinprick in her tongue and feet soles but bined administration of corticosteroids, azathioprine, and preserved vibratory sensation and normal deep tendon analgesics. reﬂexes. NCS showed normal ﬁndings. Therefore, her

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 Table Clinical features of patients with IPTN with and without Plexin D1-IgG

Total patients with Patients with IPTN with Patients with IPTN without IPTN (N = 21) Plexin D1-IgG (n = 3) Plexin D1-IgG (n = 18) puncorra pcorra

Female, n (%) 14/21 (66.7) 3/3 (100.0) 11/18 (61.1) NS NS

Age at the time of serum 62.8 ± 14.8 64.0 ± 18.0 62.6 ± 14.8 NS NS sampling (mean ± SD), y

Age at onset (mean ± SD), y 57.9 ± 15.3 54.7 ± 12.5 58.4 ± 16.0 NS NS

Disease duration at the time of 4.9 ± 5.4 9.3 ± 8.1 4.2 ± 4.8 NS NS serum sampling (mean ± SD), y

Localization of pain, n (%)

V1 11/21 (52.4) 0/3 (0.0) 11/18 (61.1) NS NS

V2 18/21 (85.7) 2/3 (66.7) 16/18 (88.9) NS NS

V3 18/21 (85.7) 3/3 (100.0) 15/18 (83.3) NS NS

Tongue 5/21 (23.8) 3/3 (100.0) 2/18 (11.1) 0.0075 0.03

Laterality of pain, n (%)

Unilateral 10/21 (47.6) 0/3 (0.0) 10/18 (55.6) NS NS

Bilateral 11/21 (52.4) 3/3 (100.0) 8/18 (44.4) NS NS

Pain characters, n (%)

Tingling 18/21 (85.7) 3/3 (100.0) 15/18 (83.3) NS NS

Burning 1/21 (4.8) 0/3 (0.0) 1/18 (5.6) NS NS

Aching 2/21 (9.5) 0/3 (0.0) 2/18 (11.1) NS NS

Abnormalities in clinical sensory testing, n (%)

Hypesthesia 20/21 (95.2) 3/3 (100.0) 17/18 (94.4) NS NS

Hyperesthesia/allodynia 2/21 (9.5) 1/3 (33.3) 1/18 (5.6) NS NS

History of allergy, n (%) 6/21 (28.6) 2/3 (66.7) 4/18 (22.2) NS NS

Abnormalities of the blink 6/9 (66.7) 1/1 (100.0) 5/8 (62.5) NS NS reflex, n (%)

Abbreviation: IPTN = idiopathic painful trigeminal neuropathy. The Mann-Whitney U test was used to compare continuous variables, and the χ2 test or Fisher exact probability test (when criteria for the χ2 test were not uncorr fulfilled) was used to compare categorical variables between each group. Statistical significance was set at p < 0.05. Uncorrected p values (p ) were corr corrected by multiplying them by the number of comparisons in the same categories (Bonferroni-Dunn correction) to calculate corrected p values (p ). a Patients with IPTN with Plexin D1-IgG vs patients with IPTN without Plexin D1-IgG.

diagnosis was small fiber neuropathy of unknown etiology. neurons in addition to DRG.2 IFAs of mouse brain, heart, and She received an analgesic agent only without immunotherapy kidney tissues with patient IgG demonstrated weak immu- to avoid hepatitis B virus reactivation. noreactivity only in endothelial cells, which was prevented by preincubation with rhPlexin D1 (figure 2E). Tissue-based IFA IgG from cases 1 and 2 bound to small TG neurons in IFAs (figure 2, A and B). Preincubation with rhPlexin D1 prevented Discussion the staining of small TG neurons by the patients’ IgG (figure 2, A and B), indicating that Plexin D1-IgG bound to small TG The main findings of the present study are as follows: (1) neurons. Dual labeling revealed that IB4 immunostaining, a small fraction of patients with IPTN had Plexin D1-IgG and a marker for pain-conducting unmyelinated neurons, mostly (2) Plexin D1-IgG bound directly to small TG neurons, es- colocalized with patient IgG binding (figure 2C). Conversely, pecially IB4-positive pain-conducting unmyelinated neurons. patient IgG did not colocalize with NFH, a marker for myelinated neurons (figure 2D). These findings indicated that Plexin D1 mRNA is present at low levels in various normal Plexin D1-IgG binds to pain-conducting unmyelinated TG adult tissues.4 Recent studies found Plexin D1 mRNA and

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN dysfunction in NP should also be carefully monitored. Tran- Figure 2 IFA using mouse tissues and IgG from patients scriptome analyses of TG and DRG neurons revealed that with IPTN with Plexin D1-IgG Plexin D1 mRNA is signiﬁcantly enriched in DRG compared with TG neurons.5 Therefore, limb and trunk pain may occur more frequently than facial pain in Plexin D1-IgG–positive cases.

IVIg was ineffective for pain relief in case 1, whereas a combined administration of corticosteroids and azathioprine improved pain in case 2. The difference in disease duration (2 months vs 15 years) may explain the inconsistent effects of immunotherapy because long disease duration is associated with neural damage in autoantibody-mediated neuropathy.7 Alternatively, Plexin D1-IgG–mediated disease may be resistant to IVIg because some subclasses of IgG autoantibodies are reported to show resistance to IVIg.8

A study limitation is that we did not establish a cell-based assay for Plexin D1-IgG because transfection of various cell types, including human embryonic kidney 293 cells, with plasmids encoding Plexin D1 induced apoptosis.9 Establish- ment of a reliable cell-based assay for Plexin D1-IgG requires further study.

In conclusion, we have demonstrated the presence of Plexin D1-IgG in a small fraction of patients with IPTN and Plexin D1-IgG reactivity toward small TG neurons, which indicates that Plexin D1-IgG–related NP could manifest as facial pain in addition to limb and trunk pain.

Acknowledgment The authors thank Jeremy Allen, PhD, from the Edanz Group (edanzediting.com/ac) for editing a draft of this manuscript. (A) IFA with mouse TG for case 1. IgG from case 1 (green) bound to small TG neurons (left image). The immunostaining (green) of small TG neurons by IgG from case 1 was prevented by preincubation with rhPlexin D1 (2 μg/mL) Study funding (right image). Nuclei are counterstained with 49,6-diamidino-2-phenylindole (DAPI) (blue). (B) IFA with mouse TG for case 2. IgG from case 2 (green) bound This study was funded in part by grants from the Japan So- to small TG neurons (left image). The immunostaining (green) of small TG neurons by IgG from case 2 was prevented by preincubation with rhPlexin ciety for the Promotion of Science (JSPS) KAKENHI (Grant D1 (2 μg/mL) (right image). Nuclei are counterstained with DAPI (blue). (C) Nos. 19H01045 and 19K17037). Immunostaining for IB4 (red), a marker of unmyelinated C-fiber type TG neurons, mostly colocalized with IgG (green) in a representative patient with IPTN (case 1). (D) Immunostaining for NFH (red), a marker of myelinated Aβ Disclosure and Aδ fiber type TG neurons, did not colocalize with IgG binding (green) in case 1. (E) IFAs with the mouse brain, heart, and kidney using IgG from T. Fujii is supported by grants from JSPS KAKENHI (Grant a representative patient with IPTN (case 1). IgG from case 1 (green) re- No. 19K17037), Mitsubishi Tanabe Pharma, Osoegawa strictedly bound to endothelial cells (white arrows) in the cerebral cortex (upper image), heart (middle image), and kidney (lower image). The weak Neurology Clinic, Bayer Yakuhin, Ltd., and the Japan Blood immunostaining (green) of endothelial cells by IgG from case 1 was prevented by preincubation with rhPlexin D1 (2 μg/mL). Nuclei are counter- Products Organization and received speaker honoraria pay- stained with DAPI (blue). Scale bars: (A–E) = 50 μm. IB4 = isolectin B4; IFA = ments from the Takeda Pharmaceutical Company and Eisai. immunofluorescence assay; IPTN = idiopathic painful trigeminal neuropathy; NFH = neurofilament heavy chain; Plexin D1-IgG = anti-Plexin D1 anti- R. Yamasaki is supported by a grant from JSPS KAKENHI body; rhPlexin D1 = recombinant human Plexin D1; TG = trigeminal (Grant No. 19K07963). Y. Miyachi, K. Iinuma, and Y. ganglion. Hashimoto report no disclosures. N. Isobe is supported by grants from Mitsubishi Tanabe Pharma, Osoegawa Neurol- ogy Clinic, Bayer Yakuhin, Ltd., and the Japan Blood Products protein in normal adult human and mouse DRG2,5 but not in Organization. T. Matsushita received speaker honoraria pay- the normal adult human brain and heart.6 In our study, Plexin ments from Mitsubishi Tanabe Pharma, the Takeda Phar- D1-IgG from patients with IPTN immunoreacted with DRG maceutical Company, and Biogen Japan. J.-i. Kira is supported neurons but only weakly immunostained endothelial cells in by grants from JSPS KAKENHI (Grant No. 19H01045) and the brain, heart, and kidney, which is largely consistent with Health and Labour Sciences Research Grants on Intractable previous ﬁndings. NP is attributable to preferential binding of Diseases [H29-Nanchitou (Nan)-Ippan-043] and received Plexin D1-IgG to DRG neurons, although endothelial cell consultancy fees, speaking fees, and/or honoraria from

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 5 Novartis Pharma, Mitsubishi Tanabe Pharma, Boehringer Ingelheim, Teijin Pharma, Takeda Pharmaceutical Company, Appendix (continued) fi Otsuka Pharmaceutical, Astellas Pharma, P zer Japan, and Name Location Contribution Eisai. Go to Neurology.org/NN for full disclosures. Noriko Isobe, Kyushu Major role in data acquisition and MD, PhD University, clinical data of patients Publication history Fukuoka, Japan Received by Neurology: Neuroimmunology & Neuroinflammation March 24, 2020. Accepted in final form May 14, 2020. Takuya Kyushu Major role in data acquisition and Matsushita, University, clinical data of patients MD, PhD Fukuoka, Japan

Jun-ichi Kira, Kyushu Designed and conceptualized the MD, PhD University, study; interpreted the data; and Appendix Authors Fukuoka, Japan revised the manuscript for intellectual content Name Location Contribution

Takayuki Kyushu Designed and conceptualized the Fujii, MD, PhD University, study; analyzed the data; and References Fukuoka, Japan drafted the manuscript for 1. Headache Classification Committee of the International Headache Society (IHS) The intellectual content International Classification of Headache Disorders, 3rd edition. Cephalalgia 2018;38:1–211. 2. Fujii T, Yamasaki R, Iiunuma K, et al. A novel autoantibody against Plexin D1 in Ryo Kyushu Analyzed and interpreted the data patients with neuropathic pain. Ann Neurol 2018;84:208–224. Yamasaki, University, 3. Prineas JW, Parratt JDE. Multiple sclerosis: serum anti-CNS autoantibodies. Mult MD, PhD Fukuoka, Japan Scler 2018;24:610–622. 4. Rehman M, Gurrapu S, Cagnoni G, et al. PlexinD1 is a novel transcriptional target and Yukino Kyushu Major role in data acquisition, IFA, effector of notch signaling in cancer cells. PLoS One 2016;11:e0164660. Miyachi, MSc University, and WB 5. MegatS,RayPR,Tavares-FerreiraD,etal.Differences between dorsal root and trigeminal Fukuoka, Japan ganglion nociceptors in mice revealed by translational profiling. J Neurosci 2019;39:6829–6847. 6. Roodink I, Verrijp K, Raats J, et al. Plexin D1 is ubiquitously expressed on tumor Kyoko Kyushu Major role in data acquisition and vessels and tumor cells in solid malignancies. BMC Cancer 2009;9:297. Iinuma, MSc University, WB 7. Ogata H, Yamasaki R, Hiwatashi A, et al. Characterization of IgG4 anti-neurofascin Fukuoka, Japan 155 antibody-positive polyneuropathy. Ann Clin Transl Neurol 2015;2:960–971. 8. Devaux JJ, Miura Y, Fukami Y, et al. Neurofascin-155 IgG4 in chronic inflammatory Yu Kyushu Major role in data acquisition and demyelinating polyneuropathy. Neurology 2016;86:800–807. Hashimoto, University, clinical data of patients 9. Luchino J, Hocine M, Amoureux MC, et al. Semaphorin 3E suppresses tumor cell MD Fukuoka, Japan death triggered by the Plexin D1 dependence receptor in metastatic breast cancers. Cancer Cell 2013;24:673–685.

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN ARTICLE OPEN ACCESS Extending rituximab dosing intervals in patients with MS during the COVID-19 pandemic and beyond?

Adil Maarouf, MD, PhD,* Audrey Rico, MD, PhD,* Clemence Boutiere, MD, Marine Perriguey, MD, Correspondence Sarah Demortiere, MD, Jean Pelletier, MD, PhD,† and Bertrand Audoin, MD, PhD,† Under the aegis of OFSEP Dr. Audoin [email protected] Neurol Neuroimmunol Neuroinﬂamm 2020;7:e825. doi:10.1212/NXI.0000000000000825

MORE ONLINE Abstract COVID-19 Resources Objective For the latest articles, To evaluate disease activity in patients with relapsing-remitting MS (RRMS) receiving ritux- invited commentaries, and imab with an extended dosing interval. blogs from physicians around the world Methods NPub.org/COVID19 In the context of COVID-19 pandemic, this was an interim analysis of an ongoing prospective observational study of patients who were stable on rituximab for at least 6 months and who had a planned extended dosing interval of 24 months. Only data for patients with active RRMS before rituximab were analyzed.

Results Among 177 patients receiving rituximab, 33 had RRMS and MRI activity before rituximab and at least 8 months of follow-up after the last infusion. The mean (SD) age was 40 (14) years, 25 were females, the mean disease duration was 10 (6.8) years, the mean annual relapse rate (ARR) before rituximab was 1.7 (1.3), and the median Expanded Disability Status Scale (EDSS) score before rituximab was 4.5 (1–7). Before extended dosing, when rituximab was infused every 6 months, the mean (SD) ARR decreased to 0.04 (0.1) (p < 0.0001) and the EDSS score to 4 (0–7) (p = 0.04). At the time of this analysis, the median follow-up since the last infusion was 11 (8–31) months. No patient showed relapse or disability progression. In total, 30 patients had at least 1 MRI performed since the last infusion (median time between the last MRI and the last infusion 10 [8–31] months). No MRI showed activity. The CD19+ cell proportion was >1% for 10 of 25 patients at the last count (median time 8 [6–25] months).

Conclusions An extended dosing interval for rituximab for patients with stable MS during the COVID-19 pandemic may be associated with a low risk of disease activity.

*These authors contributed equally to this work as co‐first authors.

†These authors contributed equally to this work as co‐last authors.

From the Service de Neurologie, Poleˆ de Neurosciences Cliniques, APHM, Hopitalˆ de la Timone, Aix Marseille Univ, France.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary ARR = annual relapse rate; CEL = contrast-enhancing lesion; EDSS = Expanded Disability Status Scale; nT2L = new T2 lesion; OFSEP = Observatoire Français de la Scl´erose en Plaques; RRMS = relapsing-remitting MS.

In the emergency context of the COVID-19 pandemic, infusion 6 months ago. This decision was based on the absence maintaining anti-CD20 therapy is problematic because of the of standardized administration scheme for rituximab in RRMS as well-known risk of severe infectious diseases developing in demonstrated by the heterogeneity of dosing intervals reported patients under this therapy.1 The wait-and-see option, in- in the literature3,4,6,7 along with our experience with patients volving a survey of the potential increase in the incidence of stopping rituximab for various reasons and to limit the potential severe COVID-19 infection in patients receiving anti-CD20 infectious side effects related to hypogammaglobulinemia.8 therapy before changing recommendations, is unsafe and Particularly, the 24-month interval was chosen according to ethically questionable. One careful option would be to delay arecentstudyfinding a potential slight waning of the rituximab reinfusion during the pandemic to limit immunodeficiency effectat24monthsafterthelastinfusion.6 during this period.2 We limited this interim analysis to data concerning patients Anti-CD20 therapies are usually administered every 6 with RRMS showing disease activity confirmed by MRI per- months, but their efficacy may be more prolonged in MS. In formed during the year before rituximab initiation (new T2 pivotal studies of rituximab in relapsing-remitting MS lesion [nT2L] or contrast-enhancing lesion [CEL]) and with (RRMS),3,4 efficacy was maintained for 12 months. Recently, the last clinical follow-up at least 8 months after the last rit- Juto et al.5 did not find any return of disease activity in patients uximab infusion. interrupting rituximab for different reasons. However, most patients switched to another treatment after rituximab Lymphocyte count withdrawal. Flow cytometry immunophenotyping was used to count CD19+ lymphocytes. At least 5,000 lymphocytes were ana- All these studies suggest that extending the delay between 2 lyzed by Navios flow cytometry (Beckman Coulter, Miami, + infusions to 12 months could be possible in MS. However, FL). The analysis was stopped when a minimum of 20 CD19 this possibility cannot exclude a potential return of disease events were detected. The maximum number of lymphocytes activity after 12 months, especially in patients with highly analyzed was 200,000. Lymphocyte counting was planned active RRMS. This issue must be addressed before system- every 6 months. atically considering postponing anti-CD20 reinfusion during the COVID-19 pandemic. Standard protocol approvals, registrations, and patient consents On March 15, 2020, at the beginning of the COVID-19 epi- This study was conducted within the framework of the na- demic in France, an emergency meeting was organized in the tional French registry designated as OFSEP (Observatoire tertiary MS center of Marseille to develop local recom- Français de la Sclérose en Plaques; ClinicalTrials.gov no. mendations for treatment management during this period. NCT02889965). Patients enrolled in our OFSEP center For suggesting an anti-CD20 therapy strategy, we decided to provided written consent for participation. OFSEP received perform an interim analysis of the data from a larger ongoing approval from the Comité Consultatif sur le Traitement de ’ monocentric prospective observational study of patients with l Information en matière de Recherche dans le domaine de la ’ MS receiving rituximab off-label with extended dosing. For Santé and Commission nationale de l informatique et des this interim analysis, only data for patients with active RRMS libertés for storing clinical, biological, and imaging data for just before rituximab were analyzed because of the highest risk research purposes. This study was covered by this general of return of disease activity in these patients. approval and did not require any additional procedure according to French laws. Methods Statistical analysis JMP 14.1.0 (SAS Institute Inc., Cary, NC) was used for sta- Protocol and participants tistical analyses. Changes in the mean relapse rate and Ex- In 2018, our department initiated change in clinical practice panded Disability Status Scale (EDSS) score were assessed by concerning the dosing interval used for off-label rituximab in the Wilcoxon signed-rank test. The proportion of patients RRMS. All neurologists (A.M., A.R., C.B., S.D., J.P., and B.A.) with MRI activity was compared by the Fisher exact test. have extended the interval between 2 infusions to 24 months, Because of multiple comparisons for the primary outcome maintaining clinical visits every 6 months and MRI moni- (annual relapse rate [ARR] before, during, and after every toring at least annually. Extending dosing was used for only 6-month infusion of rituximab), the p value was considered patients showing no disease activity since the last rituximab significant at 0.01 (Bonferroni correction).

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN For secondary outcomes (EDSS score and MRI lesions), p included in the analysis. All patients, except 1 who previously values are presented with effect size assessed by r rank cor- received natalizumab were positive for JC virus and required relation score for paired quantitative values (according to r = highly effective therapy. The patient who previously received sqrt(t2)/sqrt(t2 + df) where t is the t-ratio of the test and df =n natalizumab showed MS disease activity despite this therapy. − 1, n is the number of observations) and with Cohen w for Up to June 2019, some planned 6-month rituximab infusions qualitative parameters (according to w = 2/N, N is the total were administered in only some patients because we needed sample size). An effect size 0.1–0.3 was considered small, time to inform and convince patients about the rationale of 0.3–0.5 medium, 0.5–0.8 large, and >0.8 very large. this new administration scheme.

Disease activity during rituximab treatment Results before extended dosing In total, 29 of the 33 patients received rituximab every 6 Study population months before starting the standardized extended dosing The flow of participants is shown in figure 1, and the char- protocol (median [range] 6 [2–9] cycles) (table and figure 2). acteristics of the population are shown in the table. Since the The induction consisted of 1,000 mg infused twice at a 2-week onset of the new administration scheme, only 1 patient interval. Maintenance treatment consisted of a single infusion showed disease activity after the first infusion and did not of 1,000 mg. During treatment with 6-month interval dosing, receive the extended dosing. This patient did not show de- the mean (SD) ARR decreased from 1.7 (1.3) to 0.04 (0.1) (p pleted CD19+ cells after the second infusion and was positive < 0.001, r = 0.79) and the median (range) EDSS score from for anti-rituximab antibodies. In total, 33 patients were 4.5 (1–7) to 4 (0–7) (p = 0.04, r = 0.37). Five relapses

Figure 1 Flow of participants in the study

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 Table Demographic, clinical, and MRI characteristics of patients with RRMS (n = 33)

Characteristics Active RRMS p Value Effect size

Sex (F/M) 25/8

Age, y, mean (SD) 40 (14)

Disease duration, y, mean (SD) 10 (6.8)

Annual relapse rate during the year before rituximab, mean (SD) 1.7 (1.3)

Annual relapse rate during the 6-mo interval administration scheme, mean (SD) 0.04 (0.1) <0.001a r = 0.79

Annual relapse rate after the last rituximab infusion, mean (SD) 0 (0) <0.001b r = 0.80

EDSS score before rituximab, median (range) 4.5 (1–7)

EDSS score at the last rituximab infusion, median (range) 4(0–7) 0.04c r = 0.37

EDSS score at the last follow-up, median (range) 4(0–7) 0.01d r = 0.42

MRI before rituximab onset

% Patients with new T2 lesion(s) 100

% Patients with gadolinium-enhancing lesion(s) 75

MRI during the 6-mo interval administration scheme

% Patients with new T2 lesion(s) 15 <0.001e w = 0.74

% Patients with gadolinium-enhancing lesion(s) 6 <0.001e w = 0.50

MRI after the last rituximab infusion

% Patients with new T2 lesion(s) 0 <0.001f w=1

% Patients with gadolinium-enhancing lesion(s) 0 <0.001f w = 0.52

Treatment before rituximab, mean washout period, d (range)

Fingolimod (n = 12) 26 (15–42)

Mitoxantrone (n = 3) 105 (30–150)

Natalizumab (n = 1) 30

Cyclophosphamide (n = 1) 270

Dimethyl fumarate (n = 1) 42

Teriflunomide (n = 1) 15

Glatiramer acetate (n = 4) 0(0–0)

Abbreviations: EDSS = Expanded Disability Status Scale; RRMS = relapsing-remitting MS. Effect size was estimated with r rank correlation score for paired quantitative values and with Cohen w for qualitative parameters. An effect size 0.1–0.3 was considered small, 0.3–0.5 medium, 0.5–0.8 large, and >0.8 very large. a Wilcoxon signed-rank test between the annual relapse rate during the year before rituximab and the annual relapse rate during the 6-month interval administration scheme. b Wilcoxon signed-rank test between the annual relapse rate during the year before rituximab and the annual relapse rate after the last rituximab. c Wilcoxon signed-rank test between the EDSS score before rituximab and the EDSS score at the last rituximab infusion. d Wilcoxon signed-rank test between the EDSS score before rituximab and the EDSS score at the last follow-up. e Fisher exact test comparing the proportion of patients with new T2 lesion/gadolinium-enhancing lesion before rituximab with that during the 6-month interval administration scheme. f Fisher exact test comparing the proportion of patients with new T2 lesion/gadolinium-enhancing lesion before rituximab with that after the last rituximab.

occurred in 4 patients at 4, 5, 5, 10, and 15 months after month (nT2L and CEL), 3 months (nT2L and CEL), 3 rituximab onset. All patients underwent MRI. The mean months (nT2L no CEL), 3 months (nT2L no CEL), and 8 (range) number of MRI sessions per patient was 3 (1–7). The months (nT2L no CEL) after rituximab initiation. mean (SD) time between rituximab onset and the ﬁrst MRI was 5.5 (3.5) months, between the ﬁrst and second MRI was 9 Disease activity during extended dosing (3.6) months, and between the second and third MRI was 11 At the time of this interim analysis (between March 18 and 25, (5) months. MRI activity was found in only 5 patients at 1 2020), the median (range) interval between the last follow-up

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Figure 2 Timeline framework for the patients’ cohort

Each row on the y-axis represents a patient. On the x-axis, 0 indicates the last rituximab infusion. Diamond shape (¤) represents a relapse. Less-than sign (<) represents the onset of rituximab. More-than sign (>) represents the last clinical follow-up. Performed MRI examinations are represented by a T2 axial image. Only active MRIs are surrounded by a crown (a new T2 lesion and/or a gadolinium-enhanced lesion). Numbers represent CD19 proportion of total lymphocytes. Overall, the median interval between the last clinical follow-up and the last rituximab infusion was 11 months (range 8–31 months). The median interval between the last MRI and the last rituximab infusion was 10 months (range 8–31 months). and the last rituximab infusion was 11 (8–31) months (table administered. Moreover, of particular note, there was no and figure 2). Only 1 patient reached the 24-month time point switch to any other treatment. (31 months). This patient refused new reinfusion, arguing neurologic stability and that rituximab worsened his psoriasis. Maintenance of the efficacy of rituximab during a relatively During the follow-up, no patients showed relapse or disability long period is not fully understood. Of note, B-cell counts progression. During this period, 30 patients had at least 1 MRI during the extended dosing period showed significant re- and 22 of these had at least 1 MRI at least 6 months after the emergence of B cells in more than half of tested patients. last rituximab treatment. The median (range) interval be- Nevertheless, none of these patients experienced disease ac- tween the last MRI and the last rituximab infusion was 10 tivity, which suggests that the effect of rituximab in MS is (8–31) months. No MRI showed activity (nT2L or CEL). maintained after B-cell repopulation. This situation contrasts + CD19 B-lymphocyte count was measured in 25 patients. The with other pathologies such as AQP4 antibody disease, which – 9,10 mean (range) CD19 proportion was 1.7% (0 6.7) of tends to relapse when B cells repopulate. total lymphocytes at the last count performed after the last – rituximab infusion (median [range] interval 8 [6 25] This study is not without limitations. First, the sample size was months). CD19 proportion was >1% in 10 of 25 patients and small, which limits the robustness of the findings. However, >0.1% in 13 of 25. the population selected was homogeneous, including only patients with highly active RRMS, which facilitates general- ization to patients with lower disease activity. Second, the Discussion number of patients with a long follow-up (≥12 months) after These results suggest that the option to delay rituximab the last rituximab infusion was low (n = 14), inherent to the during the COVID-19 pandemic in RRMS could be consid- interim aspect of the study. However, no patient in this group ered. We reveal that patients with RRMS with a high level of showed relapse, which suggests that the potential for fast activity before rituximab initiation did not demonstrate any return to disease activity after rituximab withdrawal is un- return of disease activity after rituximab withdrawal during likely. Third, the relatively short follow-up prevents any de- a period of 8–31 months. Importantly, no disease activity was finitive conclusion about the potential effect of low dosing on found whatever the number of rituximab cycles previously medium-term disability progression suggested by 1 previous

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 5 study.11 Finally, lack of exhaustive assessment of CD19+- lymphocyte counts due to organization failure and memory Appendix (continued)

B-cell monitoring limits the interpretation of the present Name Location Contribution ﬁndings. Marine Aix Marseille University, Major role in the Perriguey, APHM, Hopitalˆ de la acquisition of data In this emergency context of the COVID-19 pandemic, lack of MD Timone, Poleˆ de knowledge of the potential consequences of anti-CD20 Neurosciences Cliniques

therapy on prognosis with this infection warrants careful Sarah Aix Marseille University, Major role in the consideration by neurologists. The present findings suggest Demortiere, APHM, Hopitalˆ de la acquisition of data MD Timone, Poleˆ de that extended interval dosing for stable patients with MS re- Neurosciences Cliniques ceiving rituximab during the COVID-19 pandemic may be Jean Aix Marseille University, Designed and associated with a low risk of relapse or MRI activity. A ran- Pelletier, APHM, Hopitalˆ de la conceptualized the study; domized clinical trial of extended interval dosing is required. MD, PhD Timone, Poleˆ de analyzed the data; and Neurosciences Cliniques drafted the manuscript for intellectual content Study funding No targeted funding reported. Bertrand Aix Marseille University, Designed and Audoin, MD, APHM, Hopitalˆ de la conceptualized the study; PhD Timone, Poleˆ de analyzed the data; and Disclosure Neurosciences Cliniques drafted the manuscript for intellectual content The authors report no disclosures relevant to the manuscript. Go to Neurology.org/NN for full disclosures. References Publication history 1. Luna G, Alping P, Burman J, et al. Infection risks among patients with multiple fi Received by Neurology: Neuroimmunology & Neuroinflammation sclerosis treated with ngolimod, natalizumab, rituximab, and injectable therapies. fi JAMA Neurol 2020;77:184. April 30, 2020. Accepted in nal form May 27, 2020. 2. Brownlee W, Bourdette D, Broadley S, Killestein J, Ciccarelli O. Treating multiple sclerosis and neuromyelitis optica spectrum disorder during the COVID-19 pandemic. Neurology 2020;94:949–952. doi: 10.1212/WNL.0000000000009507. 3. Bar-Or A, Calabresi PAJ, Arnold D, et al. Rituximab in relapsing-remitting multiple Appendix Authors sclerosis: a 72-week, open-label, phase I trial. Ann Neurol 2008;63:395–400. 4. Hauser SL, Waubant E, Arnold DL, et al. B-cell depletion with rituximab in relapsing- Name Location Contribution remitting multiple sclerosis. N Engl J Med 2008;358:676–688. 5. Juto A, Fink K, Al Nimer F, Piehl F. Interrupting rituximab treatment in relapsing- Adil Aix Marseille University, Designed and remitting multiple sclerosis; no evidence of rebound disease activity. Mult Scler Relat Maarouf, APHM, Hopitalˆ de la conceptualized the study; Disord 2020;37:101468. MD, PhD Timone, Poleˆ de analyzed the data; and 6. de Flon P, Gunnarsson M, Laurell K, et al. Reduced inflammation in relapsing-remitting Neurosciences Cliniques drafted the manuscript for multiple sclerosis after therapy switch to rituximab. Neurology 2016;87:141–147. intellectual content 7. Salzer J, Svenningsson R, Alping P, et al. Rituximab in multiple sclerosis. Neurology 2016;87:2074–2081. Audrey Rico, Aix Marseille University, Designed and 8. Barmettler S, Ong MS, Farmer JR, Choi H, Walter J. Association of immunoglobulin MD, PhD APHM, Hopitalˆ de la conceptualized the study; levels, infectious risk, and mortality with rituximab and hypogammaglobulinemia. Timone, Poleˆ de analyzed the data; and JAMA Netw Open 2018;1:e184169. Neurosciences Cliniques drafted the manuscript for 9. Kim S, Huh S, Lee S, Joung A, Kim H. A 5-year follow-up of rituximab treatment in intellectual content patients with neuromyelitis optica spectrum disorder. JAMA Neurol 2013;70:1110–1117. 10. Durozard P, Rico A, Boutiere C, et al. Comparison of the response to rituximab Clemence Aix Marseille University, Interpreted the data and between myelin oligodendrocyte glycoprotein and aquaporin-4 antibody diseases. Boutiere, APHM, Hopitalˆ de la revised the manuscript for Ann Neurol 2020;87:256–266. MD Timone, Poleˆ de intellectual content 11. Kletzl H, Gibiansky E, Petry C, et al. Pharmacokinetics, pharmacodynamics and Neurosciences Cliniques exposure-response analyses of ocrelizumab in patients with multiple sclerosis (N4.001). Neurology 2019;92:N4.001.

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN ARTICLE OPEN ACCESS Genetic determinants of the humoral immune response in MS

Christiane Gasperi, MD, Till F.M. Andlauer, PhD, Ana Keating, MSc, Benjamin Knier, MD, Ana Klein, MD, Correspondence Verena Pernpeintner, MD, Peter Lichtner, PhD, Ralf Gold, MD, Frauke Zipp, MD, Florian Then Bergh, MD, Dr. Hemmer [email protected] Martin Stangel, MD, Hayrettin Tumani, MD, Brigitte Wildemann, MD, Heinz Wiendl, MD, Antonios Bayas, MD, Tania Kümpfel, MD, Uwe K. Zettl, MD, Ralf A. Linker, MD, Ulf Ziemann, MD, Matthias Knop, MD, Clemens Warnke, MD, Manuel A. Friese, MD, Friedemann Paul, MD, Bjorn¨ Tackenberg, MD, Achim Berthele, MD, and Bernhard Hemmer, MD

Neurol Neuroimmunol Neuroinflamm 2020;7:e827. doi:10.1212/NXI.0000000000000827 Abstract Objective In this observational study, we investigated the impact of genetic factors at the immunoglobulin heavy chain constant locus on chromosome 14 and the major histocompatibility complex region on intrathecal immunoglobulin G, A, and M levels as well as on B cells and plasmablasts in the CSF and blood of patients with multiple sclerosis (MS). Methods Using regression analyses, we tested genetic variants on chromosome 14 and imputed human leukocyte antigen (HLA) alleles for associations with intrathecal immunoglobulins in 1,279 patients with MS or clinically isolated syndrome and with blood and CSF B cells and plasmablasts in 301 and 348 patients, respectively. Results The minor alleles of variants on chromosome 14 were associated with higher intrathecal immu- − noglobulin G levels (β = 0.58 [0.47 to 0.68], lowest adjusted p =2.32×1023), and lower − intrathecal immunoglobulin M (β = −0.56 [−0.67 to −0.46], p =2.06×10 24)andA(β = −0.42 − [−0.54 to −0.31], p = 7.48 × 10 11) levels. Alleles from the HLA-B*07:02-DRB1*15:01-DQA1*01: − 02-DQB1*06:02 haplotype were associated with higher (lowest p =2.14×10 7) and HLA-B*44: − 02 with lower (β = −0.35 [−0.54 to −0.17], p =1.38×10 2) immunoglobulin G levels. Of interest, different HLA alleles were associated with lower intrathecal immunoglobulin M (HLA-C*02:02, − β = −0.45 [−0.61 to −0.28], p =1.01×105) and higher immunoglobulin A levels (HLA-DQA1*01:03-DQB1*06:03-DRB1*13:01 haplotype, β = 0.40 [0.21 to 0.60], p =4.46× − 10 3). The impact of HLA alleles on intrathecal immunoglobulin G and M levels could mostly be explained by associations with CSF B cells and plasmablasts. Conclusion Although some HLA alleles seem to primarily drive the extent of humoral immune responses in the CNS by increasing CSF B cells and plasmablasts, genetic variants at the immunoglobulin heavy chain constant locus might regulate intrathecal immunoglobulins levels via different mechanisms.

From the Department of Neurology (C.G., T.F.M.A., A. Keating, B.K., A. Klein, V.P., A. Berthele, B.H.), Klinikum rechts der Isar, School of Medicine, Technical University of Munich; Institute of Human Genetics (P.L.), Helmholtz Zentrum München, Neuherberg; Department of Neurology (R.G.), St. Josef Hospital, Ruhr-University Bochum; Department of Neurology, Focus Program Translational Neurosciences (FTN) and Research Center for Immunotherapy (FZI), Rhine-Main Neuroscience Network (rmn2) (F.Z.), University Medical Center of the Johannes Gutenberg University Mainz; Department of Neurology and Translational Center for Regenerative Medicine (F.T.B.), University of Leipzig; Clinical Neuroimmunology and Neurochemistry (M.S.), Department of Neurology, Hannover Medical School, Hannover; Department of Neurology (H.T.), University of Ulm; Clinic of Neurology Dietenbronn (H.T.), Schwendi; Department of Neurology (B.W.), University Hospital Heidelberg; Department of Neurology (H.W.), University of Münster; Department of Neurology (A. Bayas), University Hospital Augsburg; Institute of Clinical Neuroimmunology (T.K.), University Hospital and Biomedical Center, Ludwig-Maximilians University Munich; Department of Neurology (U.K.Z.), Neuroimmunological Section, University of Rostock; Department of Neurology (R.A.L.), University Hospital Erlangen; Department of Neurology (R.A.L.), University of Regensburg; Department of Neurology & Stroke and Hertie-Institute for Clinical Brain Research (U.Z.), Eberhard-Karls-Universität Tübingen; Max Planck Institute of Psychiatry (M.K.), Munich; Department of Neurology (C.W.), Medical Faculty, Heinrich Heine University, Düsseldorf; Department of Neurology (C.W.), University Hospital Cologne; Institute of Neuroimmunology and Multiple Sclerosis (M.A.F), University Medical Centre Hamburg-Eppendorf, Hamburg; NeuroCure Clinical Research Center (F.P.), Charité—Universitätsmedizin Berlin, Corporate Member of Freie Universität Berlin, Humboldt-Universität zu Berlin; Berlin Institute of Health and Experimental and Clinical Research Center (F.P.), Max Delbrück Center for Molecular Medicine and Charité—Universitätsmedizin Berlin; and Center of Neuroimmunology (B.T.), Philipps-University Marburg; and Munich Cluster for Systems Neurology (SyNergy) (B.H.), Germany.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary CIS = clinically isolated syndrome; GWAS = genome-wide association study; HLA = human leukocyte antigen; IgA = immunoglobulin A; IgG = immunoglobulin G; IGHC = immunoglobulin heavy chain constant; IGHG = immunoglobulin heavy constant gamma; IgM = immunoglobulin M; LD = linkage disequilibrium; MHC = major histocompatibility complex; SNP = single nucleotide polymorphism.

An elevated immunoglobulin G (IgG) index is seen in 70% of Standard protocol approvals, registrations, patients with multiple sclerosis (MS), whereas intrathecal syn- and patient consents thesis of immunoglobulin M (IgM) and immunoglobulin A We obtained written informed consent from all patients (IgA) occurs less frequently (20% and 9%, respectively).1 Al- according to the Declaration of Helsinki and collected sam- though the amount of intrathecal production of Igs varies strongly ples with ethical approval at the recruitment sites. The ethic between patients, it remains relatively stable over the disease committee at the Technical University of Munich approved – course—even under disease-modifying treatment.2 4 Agenetic the study. contribution to intrathecal Ig synthesis, therefore, likely exists. CSF protein analysis We could previously demonstrate in a genome-wide associa- CSF analysis was performed at each center independently. If tion study (GWAS) that genetic variants located at the im- CSF data from more than 1 time point were available, we only munoglobulin heavy chain constant (IGHC) locus on considered the first CSF sampling data. CSF and serum chromosome 14 are associated with the IgG index in patients concentrations for albumin and the 3 Ig classes IgG, IgM, and with MS or clinically isolated syndrome (CIS).5 This finding IgA were measured in parallel by standard turbidimetric or was replicated in a large multicenter GWAS.6 nephelometric assays, depending on the center. We calculated CSF/serum quotients (QIgG, QIgM, QIgA, and Qalb) as well In addition, the GWAS by Goris et al.6 showed that the as IgG, IgM, and IgA indices as QIgG/Qalb, QIgM/Qalb, and haplotype rs9271640*A-rs6457617*G that correlates with the QIgA/Qalb, respectively. human leukocyte antigen (HLA) allele HLA-DRB1*15: 01—the strongest known MS risk allele7,8—was also associ- Flow cytometric analysis fl ated with higher IgG indices. We performed ow cytometric analysis of CSF and blood immune cells for 348 and 301 treatment-naive patients, re- 14 The aim of the present study was to further investigate the spectively, as described previously. Flow cytometry data were influence of genetic variants and HLA alleles on intrathecal available only for patients treated at the Klinikum rechts der immunoglobulin synthesis in a large cohort of patients with Isar. We used the following antibodies for staining of B cells and MS or CIS. Based on our previous findings, we aimed at plasmablasts: CD45 (clone HI30, BD Biosciences), CD19 a more detailed characterization of the association of the (clone J3.119, Beckman Coulter), and CD138 (clone B-A38, fl IGHC locus and not only intrathecal IgG but also IgM and Beckman Coulter) and analyzed the stained cells using a ow IgA levels. To further elucidate the mechanisms by which the cytometer (CyAn ADP, Beckman Coulter). We then gated the genetic variants alter the intrathecal immune response, we cells on CD45 to select all leukocytes and subsequently on + + + + analyzed possible associations with serum concentrations of CD19 (CD45 CD19 B cells) and CD138 (CD45 CD19 + IgG, IgA, and IgM, as well as with the proportion of B cells and CD138 plasmablasts). We determined cell numbers using plasmablasts in CSF and blood. FlowJo v10 (FlowJo LLC) and calculated percentages of B cells and plasmablasts of all CD45+ cells. Methods Genotyping and quality control The previously described variants at the IGHC locus associated Cohorts with the IgG index are not well represented on most of the We analyzed DNA samples of 1,279 patients with MS or CIS available whole-genome genotyping microarrays. We therefore including all patients with available DNA samples and CSF data genotyped 16 single nucleotide polymorphisms (SNPs) at the at the Klinikum rechts der Isar of the Technical University IGHC locus on a MassARRAY system using MALDI-TOF of Munich as well as patients recruited by the German MS mass spectrometry with iPLEX Gold chemistry (Agena) and competence network.9 Diagnosis was based on standard di- called genotypes with Typer Analyzer 4.0.22.67. For genotyp- – agnostic criteria.10 13 Of 2,559 patients with available DNA, we ing, we collected DNA samples from venous blood and stored excluded all patients with missing data on sex, age, or date of the samples at −80°C. During quality control, we excluded lumbar puncture. In all patients, lumbar puncture had been variants with a minor allele frequency <1%, a Hardy-Weinberg performed as part of the diagnostic workup. We performed equilibrium test p value <0.0001, or a call rate <98%, leaving 13 quality control on available genetic data as described below and variants for further analysis. Genome-wide genotyping SNP excluded patients without available genome-wide chip data. data had been acquired using 2 different microarrays (Illumina

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN OmniExpress v1.0, v1.1, and v1.2 and Illumina 660-Quad) in using 2 different microarray types. As visual inspection of the different batches at the Max Planck Institute of Psychiatry in multidimensional scaling components showed no distinction Munich, Germany, the Helmholtz Zentrum Munich in Neu- between these data sets, we did not include the genotyping herberg, Germany, and the Wellcome Trust Sanger Institute in chip as a covariate. We determined homoscedasticity of re- Cambridge, United Kingdom. Genotype calling had been gression residuals using the Breusch-Pagan test; for models performed with GenomeStudio Genotyping Module v2.0 or showing evidence for heteroscedasticity, we used robust with Illuminus.15 We conducted quality control of the genotype sandwich error estimators (R package sandwich). We tested data using PLINK v1.90b6.916,17 as described previously.8 We the normality of residuals by a Shapiro-Wilk test and cor- excluded individuals with a genotyping rate <98%, cryptic re- rected p values for multiple testing using the Bonferroni latedness >1/8, and any genetic outliers with a distance in the procedure for the number of LD groups with r2 > 0.7 (n = 85 first 2 multidimensional scaling ancestry components of the independent tests, 4 LD groups of the chromosome 14 var- identity-by-state matrix of >5 SDs. We further excluded indi- iants, and 81 HLA allele LD groups). To determine the viduals with deviation of autosomal heterozygosity >4 SDs phenotypic variance explained by the analyzed genetic var- from the mean and individuals with heterozygosity on the X iants (R2), we conducted linear regression analyses using the chromosome of < −0.2. residuals from a null model including all covariates as the dependent variable and the respective genetic variant as the HLA imputation independent variable. We performed HLA allele imputation using SNP2HLA v1.0.3 (Beagle v3.04) and the Type 1 Diabetes Genetics Consortium We carried all variants and HLA alleles significantly associated – imputation panel, as previously described.18 20 After quality with any of the Ig indices forward to secondary exploratory control, we selected 98 HLA alleles with 4-digit resolution, an analyses on their associations with rank-transformed Ig serum allele frequency of ≥1%, and a Beagle imputation r2 ≥ 0.3 for concentrations and proportions of CSF and blood B cells further analysis. We also analyzed the following haplotypes (CD45+ CD19+ cells) and plasmablasts (CD45+ CD19+ determined using Beagle phasing results: HLA-A*03:01- CD138+ cells). IgM indices and serum IgM concentrations C*07:02-B*07:02-DRB1*15:01-DQA1*01:02-DQB1*06:02, did not follow a normal distribution after inverse rank trans- HLA-DQA1*01:03-DQB1*06:03-DRB1*13:01, HLA-A*02: formation. We therefore validated associations with these 01-B*44:02-C*05:01-DRB1*04:01, and HLA-A*02:01-B*27: traits using permutation analyses (100,000 permutations). 02-C*02:02-DRB1*16:01. We performed causal mediation analyses including nonparametric bootstrap for estimation of CIs and p values using Linkage disequilibrium of the variants on the R package mediation21 with the same covariates as de- chromosome 14 and HLA alleles scribed above with 10,000 simulations. We performed all Of 16 genotyped variants, 13 variants at the IGHC locus passed statistical analyses using R v3.5.1.22 quality control. Using a linkage disequilibrium (LD) threshold of r2 > 0.7, we defined 4 LD groups: A: rs10136766, rs1071803, Data availability fi rs111608686, rs1134590, rs11621145, rs12884389, rs12897751, The data that support the ndings of this study are available rs2725142, rs2753571, and rs34398108; B: rs1059216; C: from the corresponding author on reasonable request. rs61984162; and D: rs8009156. For the 4-digit HLA alleles, 7LDgroupswith2memberseachand5LDgroupswith3 alleles could be identified. We calculated LD using PLINK Results 16,17 v1.90b6.9. Study cohort, CSF, and flow cytometry data Table e-1, links.lww.com/NXI/A277, shows the demographic Statistical analyses data and aggregated CSF and flow cytometry parameters for As the primary analysis, we investigated associations of the all 1,279 patients. Four hundred twenty of these samples were genotyped variants on chromosome 14 and the imputed HLA included in one or both of the mentioned previous genetic alleles with Ig indices (transformed by inverse rank normali- studies on IgG indices.5,6 We observed significant correlations zation) by linear regression. For the IGHC variants, we either between all 3 Ig indices (Spearman ρ = 0.26, 95% CI chose a dominant or an additive allelic model for the re- [0.20–0.31] for IgG and IgM indices; ρ = 0.18 [0.12–0.24] for gression analyses, following visual inspection of the data, and IgG and IgA indices; and ρ = 0.47 [0.42–0.52] for IgM and included sex, age at lumbar puncture, and the sequencing IgA indices). plate as covariates. For the imputed HLA alleles, we analyzed dosage data (2× the probability for being homozygous for the Genetic factors at the IGHC locus associated allele + 1× the probability for being heterozygous) using the with Ig indices same covariates except the sequencing plate. To correct for The minor alleles of all 10 variants from LD group A on population stratification, we added up to 5 first multidimen- chromosome 14 were associated with higher IgG indices sional scaling components as covariates if they were associ- (table 1 and figure 1A). All 10 variants were also significantly ated either with the dependent variable or the investigated associated with IgA and IgM indices. However, these associ- variant or HLA allele. The DNA samples had been genotyped ations had a reversed sign, i.e., the minor alleles were

Table 1 Association of variants on chromosome 14 with the Ig indices

Regression on IgG indices Regression on IgM indices Regression on IgA indices

Variant (LD group) EA AF β (95% CI) Adjusted p N β (95% CI) Adjusted p N β (95% CI) Adjusted p N

− − − rs10136766 (A) A 25.5 0.57 (0.46 to 0.68) 1.02 × 10 21 1,229 −0.55 (−0.65 to −0.44) 6.35 × 10 22 1,151 −0.41 (−0.53 to −0.30) 3.57 × 10 10 1,143

− − − rs1071803 (A) T 25.8 0.57 (0.46 to 0.68) 1.34 × 10 21 1,227 −0.57 (−0.67 to −0.46) 1.51 × 10 23 1,150 −0.41 (−0.52 to −0.29) 7.79 × 10 10 1,142

− − − rs11160868 (A) T 25.7 0.58 (0.47 to 0.68) 2.32 × 10 23 1,262 −0.56 (−0.66 to −0.45) 2.61 × 10 23 1,189 −0.39 (−0.50 to −0.27) 2.69 × 10 09 1,181

− − − rs1134590 (A) C 21.8 0.55 (0.43 to 0.66) 2.89 × 10 19 1,262 −0.47 (−0.58 to −0.36) 1.37 × 10 15 1,185 −0.37 (−0.49 to −0.26) 3.34 × 10 08 1,177

− − − rs11621145 (A) G 27.1 0.53 (0.42 to 0.64) 8.39 × 10 19 1,227 −0.53 (−0.64 to −0.42) 2.22 × 10 20 1,149 −0.40 (−0.52 to −0.29) 1.05 × 10 09 1,141

− − − rs12884389 (A) C 28.7 0.51 (0.40 to 0.62) 1.72 × 10 17 1,229 −0.48 (−0.59 to −0.37) 2.51 × 10 16 1,152 −0.42 (−0.54 to −0.31) 7.48 × 10 11 1,144

− − − rs12897751 (A) G 25.6 0.57 (0.46 to 0.68) 4.92 × 10 23 1,278 −0.56 (−0.67 to −0.46) 2.06 × 10 24 1,200 −0.40 (−0.51 to −0.29) 4.05 × 10 10 1,192

− − − rs2725142 (A) G 29.0 0.51 (0.40 to 0.62) 1.13 × 10 17 1,233 −0.48 (−0.59 to −0.37) 1.46 × 10 16 1,156 −0.40 (−0.51 to −0.28) 1.07 × 10 09 1,148

− − − rs2753571 (A) A 29.1 0.51 (0.40 to 0.62) 3.65 × 10 18 1,231 −0.46 (−0.57 to −0.36) 3.11 × 10 15 1,154 −0.39 (−0.50 to −0.27) 4.50 × 10 09 1,146

− − − rs34398108 (A) A 27.9 0.52 (0.42 to 0.63) 2.40 × 10 19 1,229 −0.54 (−0.64 to −0.43) 2.71 × 10 21 1,153 −0.40 (−0.51 to −0.28) 1.19 × 10 09 1,145

− − − rs1059216 (B) C6.2−0.10 (−0.27 to 0.07) 1.00 × 10 00 1,270 −0.17 (−0.33 to 0.00) 1.00 × 10 00 1,193 −0.12 (−0.30 to 0.05) 1.00 × 10 00 1,185

− − − rs61984162 (C) A 2.8 0.20 (−0.04 to 0.45) 1.00 × 10 00 1,276 −0.20 (−0.44 to 0.04) 1.00 × 10 00 1,198 −0.15 (−0.40 to 0.11) 1.00 × 10 00 1,190

− − − rs8009156 (D) T 44.8 −0.21 (−0.29 to −0.13) 5.80 × 10 06 1,256 0.19 (0.11 to 0.26) 1.07 × 10 04 1,178 0.14 (0.06 to 0.22) 5.18 × 10 02 1,170

Abbreviations: AF = allele frequency; EA = effect allele; Ig = immunoglobulin; IgA = immunoglobulin A; IgG = immunoglobulin G; IgM = immunoglobulin M; LD = linkage disequilibrium; N = number of patients. Linear regression on Ig indices transformed by inverse rank normalization and variants on chromosome 14. p Values were adjusted by Bonferroni correction for multiple testing for n = 85 tests. IgM index were validated using permutation analyses (data not Figure 1 Ig indices in carriers and noncarriers of the effect shown). alleles of variants on chromosome 14 To fine map the association of variants at the IGHC locus with Ig indices, we performed stepwise conditional regression analyses, adjusting, in each step, for the variants with the most robust support for association. For IgG indices, we identified rs12897751 as the top-associated variant (explaining 7.6% of IgG index variance) and observed no evidence for a second causal effect at this locus. SNP rs12897751 also showed the most robust association with IgM indices (explaining 8.2% of the variance), and we observed weak evidence for a possible second causal effect: SNP rs34398109 was associated at nominal significance when conditioning for rs12897751 (β = −0.24 [−0.47 to −0.01], p = 0.037). For IgA indices, rs12884389 was the top-associated SNP, explaining 4.2% of the variance, and there was weak evidence for a second effect because rs11621145 was associated at nominal significance in the conditional analysis (β = −0.19 [−0.37 to −0.01], p = 0.035).

Association of HLA alleles with Ig indices Of 98 analyzed 4-digit HLA alleles, 9 showed an association with at least 1 of the 3 Ig indices after correction for multiple testing (table 2). HLA-DRB1*15:01, HLA-DQB1*06:02, HLA-DQA1*01:02, and HLA-B*07:02 were all signiﬁcantly associated with higher IgG indices (ﬁgure 2A for HLA- DQB1*06:02). We performed haplotype level analyses on the extended HLA-A*03:01-C*07:02-B*07:02-DRB1*15:01- DQA1*01:02-DQB1*06:02 haplotype with stepwise addition of the single HLA alleles of this haplotype. The most robust support for association could be observed for the HLA-DRB1*15:01-DQA1*01:02-DQB1*06:02 haplotype that explained 2.6% of the variance of IgG indices. When analyzed separately, HLA-DQB1*06:02 showed the most robust association. In addition, HLA-B*44:02 showed an association with lower IgG indices, explaining 1.0% of the IgG index variance. This association was independent of the HLA-DRB1*15:01-DQA1*01:02-DQB1*06:02 haplotype in a conditional analysis (β = −0.34 [−0.52 to −0.16], − unadjusted p = 2.64 × 10 04).

Only HLA-C*02:02 was associated with IgM indices after

Box plots showing IgG (A), IgM (B), and IgA (C) indices transformed by inverse correction for multiple testing. This allele was associated with rank normalization for noncarriers, heterozygotes, and homozygotes of the lower IgM indices, explaining 1.8% of the variance. All HLA effect allele of the most significantly associated variants on chromosome 14. fi Bars represent the median, boxes the interquartile ranges, the vertical lines alleles signi cantly associated with IgG indices were also as- the data range without outliers, and dots the outliers. sociated with IgM indices at a nominal significance level (table 2 and figure 2B), but these associations were not significant after correction for multiple testing. associated with lower IgM (figure 1B) and IgA (figure 1C) indices (table 1). Visual evaluation of the rank-transformed Ig HLA-DRB1*13:01, HLA-DQB1*06:03, and HLA-DQA1*01: indices by genotype for these variants was consistent with 03 were associated with higher IgA indices. All 3 HLA alleles a dominant model of inheritance. The minor allele rs8009156*T are part of the HLA-DQA1*01:03-DQB1*06:03-DRB1*13:01 was associated with lower IgG indices and higher IgM and IgA haplotype, and the association of the haplotype was stronger indices, but the association with IgA indices was not significant than the association of the single alleles, explaining 1.6% of the after correction for multiple testing. Associations for the IgA index variance (table 2).

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 5 Table 2 Associations of HLA alleles with the Ig indices

Regression on IgG indices Regression on IgM indices Regression on IgA indices (no. of patients = 1,279) (no. of patients = 1,200) (no. of patients = 1,192)

HLA allele AF β (95% CI) Adjusted p β (95% CI) Adjusted p β (95% CI) Adjusted p

− − − B*07:02 17.8 0.18 (0.08 to 0.28) 2.92 × 10 02 0.12 (0.02 to 0.22) 1.00 × 10 00 0.09 (−0.02 to 0.19) 1.00 × 10 00

− − − DQA1*01:02 33.0 0.17 (0.09 to 0.25) 7.61 × 10 03 0.11 (0.02 to 0.19) 1.00 × 10 00 0.05 (−0.04 to 0.13) 1.00 × 10 00

− − − DQB1*06:02 26.5 0.27 (0.18 to 0.36) 2.41 × 10 07 0.15 (0.06 to 0.24) 1.05 × 10 01 0.09 (0.00 to 0.19) 1.00 × 10 00

− − − DRB1*15:01 27.3 0.24 (0.16 to 0.33) 5.92 × 10 06 0.13 (0.05 to 0.22) 2.57 × 10 01 0.09 (0.00 to 0.18) 1.00 × 10 00

− − − DQA1*01:03 6.0 0.10 (−0.07 to 0.26) 1.00 × 10 00 0.15 (−0.02 to 0.31) 1.00 × 10 00 0.37 (0.19 to 0.56) 6.71 × 10 03

− − − DQB1*06:03 5.5 0.06 (−0.10 to 0.22) 1.00 × 10 00 0.17 (0.00 to 0.33) 1.00 × 10 00 0.34 (0.15 to 0.53) 3.40 × 10 02

− − − DRB1*13:01 5.4 0.10 (−0.07 to 0.27) 1.00 × 10 00 0.18 (0.01 to 0.35) 1.00 × 10 00 0.39 (0.19 to 0.59) 8.59 × 10 03

− − − B*44:02 4.5 −0.35 (−0.54 to −0.17) 1.38 × 10 02 −0.25 (−0.45 to −0.05) 7.44 × 10 01 −0.03 (−0.25 to 0.18) 1.00 × 10 00

− − − C*02:02 4.3 −0.19 (−0.38 to −0.01) 1.00 × 10 00 −0.45 (−0.61 to −0.28) 1.01 × 10 05 −0.22 (−0.41 to −0.03) 1.00 × 10 00

− − − Haplotype1 26.0 0.27 (0.18 to 0.36) 2.14 × 10 07 0.15 (0.06 to 0.24) 9.07 × 10 02 0.10 (0.01 to 0.19) 1.00 × 10 00

− − − Haplotype2 5.5 0.11 (−0.06 to 0.28) 1.00 × 10 00 0.18 (0.01 to 0.35) 1.00 × 10 00 0.40 (0.21 to 0.60) 4.46 × 10 03

Abbreviations: AF = allele frequency; Haplotype1 = HLA-DRB1*15:01-DQA1*01:02-DQB1*06:02; Haplotype2 = HLA-DQA1*01:03-DQB1*06:03-DRB1*13:01; HLA = human leukocyte antigen; Ig = immunoglobulin; IgA = immunoglobulin A; IgG = immunoglobulin G; IgM = immunoglobulin M; LD = linkage disequilibrium. Linear regression on Ig indices transformed by inverse rank normalization and 4-digit imputed HLA alleles. p Values were adjusted by Bonferroni correction for multiple testing for n = 85 tests.

Association of genetic variants and HLA alleles this was the HLA allele with the strongest association with the with serum Ig concentrations IgG indices). HLA-C*02:02 was associated with lower CSF To further investigate the mechanisms by which the genetic B cell and plasmablast proportions (table 3). HLA-DRB1*13: factors might have an effect on Ig indices, we performed sec- 01, HLA-DQA1*01:03, HLA-DQB1*06:03, and HLA-B*44:02 ondary regression analyses on blood Ig concentrations for were not associated with CSF B cells or plasmablasts. There variants and HLA alleles associated with Ig indices. Except for was no significant association of any HLA allele associated with rs8009156, the minor alleles of all other variants on chromo- the Ig indices with percentages of peripheral blood B cells or some 14 associated with IgG indices were also associated with plasmablasts (data not shown). higher IgG serum levels, but the effect sizes of these associations were smaller compared with the analyses on IgG indices To investigate whether the association of the HLA alleles (β = 0.15 - 0.28, data not shown). There was no association of from the HLA-DRB1*15:01-DQA1*01:02-DQB1*06:02 any of the analyzed variants at the IGHC locus with serum haplotype and HLA-C*02:02 with CSF B cell and plasma- IgM concentrations, and only rs12897751 was associated blast proportions fully explains the associations with the Ig with lower serum IgA concentrations. There was no asso- indices, we performed causal mediation analyses. These ciation of any HLA allele with serum Ig levels after correction analyses could only be performed in a smaller proportion of for multiple testing (data not shown). the patients with available flow cytometry data and showed nominally significant results. We observed full mediation of Association of genetic variants and HLA alleles the effect of the HLA alleles from the HLA-DRB1*15:01- with blood and CSF B cells and plasmablasts DQA1*01:02-DQB1*06:02 haplotype on IgG indices by Of the analyzed variants on chromosome 14, only rs2725142 increased CSF B cell or plasmablast proportions (shown and rs2753571 showed an association with proportions of CSF for HLA-DQB1*06:02 in figure e-1, links.lww.com/NXI/ B cells (β = −0.29 [−0.52 to −0.07], adjusted p = 0.022 for both A277) and full mediation of the effect of HLA-C*02:02 on variants), but were associated with a lower proportion of CSF IgM indices by decreased CSF plasmablast proportions B cells. None of the other analyzed variants on chromosome 14 (figure e-1, links.lww.com/NXI/A277). were significantly associated with the proportions of B cells and plasmablasts in the CSF or peripheral blood (data not shown). Epistasis between genetic factors at the IGHC HLA-DRB1*15:01 was associated with higher CSF B cell locus and HLA alleles on Ig indices and plasmablast proportions and HLA-DQB1*06:02 and We tested for epistatic interactions between the variants at the HLA-DQA1*01:02 with higher CSF plasmablast pro- IGHC locus and HLA alleles. No significant interactions were portions (table 3 and figure 2 shown for HLA-DQB*06:02 as found between the top-associated IGHC variant, rs12897751,

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Figure 2 Ig indices and CSF B cell and plasmablast proportions for carriers and noncarriers of HLA-DQB1*06:02

Box plots showing the rank-transformed IgG indices (A), IgM indices (B), CSF B cell (C), and plasmablast (D) proportions (all transformed by inverse rank normalization) for noncarriers, heterozygotes, and homozygotes of HLA-DQB1*06:02. Bars represent the median, boxes the interquartile ranges, the vertical lines the data range without outliers, and dots the outliers. and alleles from the HLA-B*07:02-DRB1*15:01-DQA1*01: We confirmed and fine mapped a previously reported associ- 02-DQB1*06:02 haplotype on IgG indices or between ation between genetic variants at the IGHC locus and in- rs12897751 and HLA-C*02:02 on IgM levels. However, we trathecal IgG indices.5,6 The effect alleles of a highly correlated observed evidence for epistatic interaction between group of 10 SNPs were significantly associated with higher IgG rs12897751 and HLA-B*44:02 on IgG indices (interaction indices. SNP rs12897751—an intronic variant in the immu- −03 term β = −0.58 [−0.94 to −0.22], p = 6.12 × 10 ,adjusted noglobulin heavy constant gamma 3 (IGHG3)gene—showed for 4 independent test, figure e-1C, links.lww.com/NXI/ the most robust support for association. Another variant A277) and between rs12884389 and HLA alleles from the strongly associated with IgG indices, but not independently of HLA-DQA1*01:03-DQB1*06:03-DRB1*13:01 haplotype on rs12897751, was the missense variant rs1071803. Variant IgA indices (interaction term for the haplotype and rs1071803 (in high LD with rs12897751, r2 = 0.95) defines the β − − − rs12884389 = 0.61 [ 0.97 to 0.26], adjusted p = 3.22 × IgG1 allotype G1m17 by altering the amino acid sequence of −03 fi 10 , gure e-1D, links.lww.com/NXI/A277). the CH1 domain of IgG1. The G1m17 allotype is part of the Gm21*;17,1;.. and the Gm21*;17,1,2;.. haplotypes, which are prevalent haplotypes in Caucasian and Mongoloid pop- Discussion ulations.23 The functional consequences of allotypes are poorly In this study, we report associations of genetic factors in 2 understood. Allotypes have been shown to correlate with IgG regions—the IGHC locus on chromosome 14 and the major plasma concentrations24,25; they alter the IgG half-life26 and histocompatibility complex (MHC) region on chromosome might influence the distribution of antibodies to specific tissues, 6—with IgG, IgA, and IgM indices in patients with MS or CIS. affect class switching, or alter secondary messenger RNA

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 7 Table 3 Association of HLA alleles with CSF B cells and plasmablasts

Regression on CSF B cell proportions Regression on CSF plasmablast proportions (no. of patients = 348) (no. of patients = 348)

HLA allele AF β (95% CI) Adjusted p β (95% CI) Adjusted p

− − B*07:02 17.5 0.13 (−0.06 to 0.31) 8.90 × 10 01 0.15 (−0.03 to 0.33) 5.35 × 10 01

− − DQA1*01:02 32.3 0.19 (0.03 to 0.36) 1.17 × 10 01 0.23 (0.07 to 0.39) 3.11 × 10 02

− − DQB1*06:02 25.5 0.23 (0.05 to 0.40) 5.50 × 10 02 0.23 (0.06 to 0.41) 3.74 × 10 02

− − DRB1*15:01 26.7 0.23 (0.06 to 0.40) 4.45 × 10 02 0.25 (0.08 to 0.42) 1.91 × 10 02

− − DQA1*01:03 6.0 −0.12 (−0.43 to 0.20) 1.00 × 10 00 −0.09 (−0.40 to 0.22) 1.00 × 10 00

− − DQB1*06:03 5.5 −0.12 (−0.43 to 0.19) 1.00 × 10 00 −0.04 (−0.34 to 0.27) 1.00 × 10 00

− − DRB1*13:01 5.8 −0.13 (−0.45 to 0.19) 1.00 × 10 00 −0.07 (−0.39 to 0.25) 1.00 × 10 00

− − B*44:02 3.2 0.07 (−0.35 to 0.49) 1.00 × 10 00 0.25 (−0.16 to 0.66) 1.00 × 10 00

− − C*02:02 4.3 −0.57 (−0.99 to −0.14) 4.52 × 10 02 −0.57 (−0.91 to −0.23) 5.35 × 10 03

Abbreviations: AF = allele frequency; HLA = human leukocyte antigen; Ig = immunoglobulin; IgA = immunoglobulin A; IgG = immunoglobulin G; IgM = immunoglobulin M; LD = linkage disequilibrium. Linear regression on CSF B cell and plasmablast proportions transformed by inverse rank normalization and 4-digit imputed HLA alleles. p Values were adjusted by Bonferroni correction for multiple testing for n = 5 tests (only HLA alleles associated with at least one of the Ig indices were tested).

structures and affect the transcription rate.27,28 Although Eight HLA alleles were associated with Ig indices in this based on conditional analyses, rs1071803 does not appear to study. Alleles that are part of the extended HLA-DRB1*15: be the causal variant for the association of the IGHC locus 01-DQA1*01:02-DQB1*06:02 haplotype were significantly and the IgG index, it is however possible that other genetic associated with higher IgG indices and nominally associ- variants causing Ig allotypy, not genotyped in the present ated with higher IgM indices. The HLA-DRB1*15:01- study but in LD with the investigated variants, are causal for DQA1*01:02-DQB1*06:02 haplotype showed the most this relationship. Alternatively, as rs12897751 has been robust support of associationandHLA-DQB1*06:02was shown to be associated with higher protein expression levels the most significantly associated single allele. Causal me- of IGHG1, IGHG2, and IGHG3 (GTEx v829), an influence diation analysis showed that the effect of these HLA alleles of this or another correlated variant at this locus on gene on the IgG index is completely explained by their associa- expression might affect IgG levels. tion with higher CSF B cell and plasmablast proportions. ThesameistruefortheassociationofHLA-C*02:02with Most of the genetic variants on chromosome 14 associated lower IgM indices and lower CSF B lymphocyte pro- with Ig indices were also associated with serum IgG levels. portions. HLA-B*44:02 was independently associated These associations were, however, much weaker than the with lower IgG indices, and we observed some evidence for associations with Ig indices. Variants at the IGHC locus might epistasis between HLA-B*44:02 and the top-associated thus not only affect the amount of intrathecal Igs but also have IGHC variant rs12897751. The HLA-DQA1*01:03- some effect on serum IgG concentrations. Analysis of flow DQB1*06:03-DRB1*13:01 haplotype and the individual cytometry data from CSF cells showed an association of 2 HLA alleles that are part of this haplotype, on the other SNPs on chromosome 14 with CSF B cells. As the effect hand, were associated with higher IgA indices, but did not alleles of these variants were, however, associated with lower show any association with CSF B cells or plasmablasts. We CSF B cell proportions, it is unlikely that this explains their did however observe evidence for epistatic interactions association with higher IgG indices. between the HLA-DQA1*01:03-DQB1*06:03-DRB1*13: 01 haplotype and IGHC variants. Different HLA alleles thus Of interest, the minor alleles of all variants on chromosome appear to have a differential effect on the intrathecal pro- 14 that were associated with higher IgG indices were, at the duction of the Ig classes IgG, IgM, and IgA, probably due to same time, associated with lower IgM and IgA indices. This different underlying mechanisms. Because of the design of is especially striking, as higher IgG indices were correlated this study, we cannot conclude whether the observed effects with higher IgA and IgM indices. How the variants at the are specific for MS or possibly shared by other inflammatory IGHC locus influence intrathecal IgM and IgA is, therefore, neurologic diseases. unclear; possible scenarios are an increased class switching to IgG-producing cells with resulting lower concentrations We describe 2 genetic regions—the IGHC locus on chro- of IgA and IgM or feedback mechanisms resulting in reduced mosome 14 and the MHC region on chromosome 6—that synthesis of IgM and IgA. were associated with the amount of intrathecal IgG, IgM,

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN and/or IgA in patients with MS or CIS. Our findings suggest Disclosure differential mechanisms by which the 2 genetic regions in- Dr. C. Gasperi received funding from the Deutsche For- fluence intrathecal Ig synthesis or concentration (figure e-2, schungsgemeinschaft (DFG, German Research Founda- links.lww.com/NXI/A277). HLA alleles in LD with a known tion).Dr.T.F.M.AndlauerandA.Keatingreportno MS risk allele, HLA-DRB1*15:01, appeared to influence the disclosures relevant to the manuscript. Dr. B. Knier received proportion of intrathecal B cells and plasmablasts and thereby funding from the Federal Ministry for Education and Re- increase the intrathecal synthesis of immunoglobulins, espe- search (grant from the German Competence Network of cially of IgG. The same was true for HLA-C*02:02 associated MS) and intramural funding from the Technical University with lower CSF B lymphocyte proportions and IgM indices. of Munich. He received a research grant from Novartis The DQA1*01:03-DQB1*06:03-DRB1*13:01 haplotype was outside the submitted work. Dr. A. Klein, Dr. V. Pern- associated with higher IgA indices, but not with higher CSF peintner,andDr.P.Lichtnerreport no disclosures relevant B lymphocyte proportions. We cannot conclude from the to the manuscript. Dr. R. Gold has received speaker’sand fl present study how these alleles in uence CSF IgA levels. board honoraria from Baxter, Bayer Schering, Biogen Idec, CLB Behring, Genzyme, Merck Serono, Novartis, Stendhal, Genetic variants at the IGHC locus on chromosome 14 were Talecris, and Teva Pharmaceutical Industries and grant associated with higher IgG and lower IgA and IgM indices. support (via his department) from Bayer Schering, Biogen fl These variants did not in uence the composition of Idec, Genzyme, Merck Serono, Novartis, and Teva Phar- B lymphocytes in the CSF. We therefore believe that they maceutical Industries. Dr. F. Zipp has recently received re- fl might in uence the amount of intrathecal Ig via other search grants and/or consultation funds from DFG, BMBF, mechanisms, such as an altered CSF immunoglobulin — PMSA,MPG,Genzyme,MerckSerono,RocheNovartis, homeostasis probably as a result of changes of the protein Sanofi Aventis, Celgene, ONO, and Octapharma. Dr. F. structure caused by the variation in the amino acid sequence Then Bergh has received funding from the Deutsche For- associated with the Gm21* haplotypes. Understanding the schungsgemeinschaft; received through his institution, re- mechanisms by which the IGHC locus influences intrathecal search support for investigator-initiated studies from Ig levels may have implications for the design of future Actelion and Novartis; served on scientific advisory boards therapeutic antibodies to ensure a better enrichment and for Novartis, Sanofi/Genzyme, and Roche; received support persistence in the CNS compartment. to attend a scientific meeting from Biogen; and received Acknowledgment personal honoraria for speaking from Bayer Schering, Bio- gen, Roche, and Sanofi/Genzyme. Dr. M. Stangel has re- The authors thank all contributors of the study, especially the fi study nurses, for their motivated collaboration and recruitment ceived honoraria for scienti c lectures or consultancy from efforts, all patients and relatives for their participation and Bayer HealthCare, Biogen, Baxter/Baxalta, CSL Behring, fi support, and the data monitoring and administrative personnel Euroimmun, Grifols, Merck Serono, Novartis, Roche, Sano of the German Competence Network of Multiple Sclerosis and Aventis, and Teva Pharmaceutical Industries. His institution the Klinikum rechts der Isar. received research support from Bayer HealthCare, Biogen Idec, Genzyme, Merck Serono, Novartis, and Teva Phar- Study funding maceutical Industries. Dr. H. Tumani received speaker The German National Multiple Sclerosis Cohort is sup- honoraria from Bayer, Biogen, Fresenius, Genzyme, Merck, ported by grants from the German Federal Ministry for Novartis, Roche, Siemens, and Teva; he or his institution has Education and Research (grants 01GI0914 [Bochum], received research support from Fresenius, Genzyme, Merck, 01GI1601B [Marburg], and 01GI1601D [Munich]) and the and Novartis, none related to this work. Dr. B. Wildemann German Research Foundation (DFG) in the framework of has received research grants and/or honoraria from Merck fi the Transregional Collaborative Research Center SFB CRC Serono, Biogen, Teva, Novartis, Sano Genzyme, and Bayer TR-128 (Drs Hemmer, Wiendl, Zipp, and Gold). Bernhard HealthCare and research grants from the Dietmar Hopp Hemmer received funding for the study by the European Foundation, the Klaus Tschira Foundation, the Bundesmi- Union’s Horizon 2020 Research and Innovation Program nisterium für Bildung und Forschung (BMBF), and the (grant MultipleMS, EU RIA 733161) and the Deutsche Deutsche Forschungsgemeinschaft (DFG). All conflicts are Forschungsgemeinschaft (DFG, German Research Foun- not relevant to the topic of the study. Dr. H. Wiendl receives dation) under Germany’s Excellence Strategy within the honoraria for acting as a member of scientific advisory framework of the Munich Cluster for Systems Neurology boards and as consultant for Biogen, Evgen, MedDay (EXC 2145 SyNergy—ID 390857198). Bernhard Hemmer, Pharmaceuticals, Merck Serono, Novartis, Roche Pharma Till Andlauer, Christiane Gasperi, Tania Kümpfel, Ulf Zie- AG, and Sanofi Genzyme, as well as speaker honoraria mann, and Antonios Bayas are associated with DIFUTURE and travel support from Alexion, Biogen, Cognomed, F. (Data Integration for Future Medicine, BMBF 01ZZ1804 Hoffmann-La Roche Ltd., Gemeinnützige Hertie-Stiftung, [A-I]). The biobank of the Department of Neurology as part Merck Serono, Novartis, Roche Pharma AG, Sanofi Gen- of the Joint Biobank Munich in the framework of the Ger- zyme, Teva, and WebMD Global. Prof. Wiendl is acting as a man Biobank Node supported the study. paid consultant for AbbVie, Actelion, Biogen, IGES, Johnson

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 9 & Johnson, Novartis, Roche, Sanofi Genzyme, and the Swiss HealthCare, Biogen, CSL Behring, Gilead, Grifols, Merck Multiple Sclerosis Society. His research is funded by the German Serono, Novartis, Octapharma, Roche, Sanofi Genzyme, Teva, Ministry for Education and Research (BMBF), Deutsche For- and UCB Pharma. His university received unrestricted research schungsgemeinschaft (DFG), Else Kröner Fresenius Founda- grants from Biogen Idec, Novartis, Teva, Bayer HealthCare, tion, Fresenius Foundation, Hertie Foundation, NRW Ministry CSL Behring, Grifols, Octapharma, Sanofi Genzyme, and UCB of Education and Research, Interdisciplinary Center for Clinical Pharma. Dr. A. Berthele reports compensations for clinical Studies (IZKF) Muenster and RE Children’s Foundation, Bio- trials received by his institution from Alexion Pharmaceuticals, gen GmbH, GlaxoSmithKline GmbH, Roche Pharma AG, and Biogen, Novartis Pharmaceuticals, Roche, Sanofi Genzyme, Sanofi Genzyme. Dr. A. Bayas has received personal compen- and Teva Pharmaceuticals and personal fees and nonfinancial sation from Merck, Biogen, Bayer Vital, Novartis, Teva, Roche, support from Bayer HealthCare, Biogen, Merck Serono, Celgene, and Sanofi/Genzyme and grants for congress trips and Mylan, Novartis Pharmaceuticals, Roche, and Sanofi Genzyme, participation from Biogen, Teva, Novartis, Sanofi/Genzyme, all outside the work presented. Dr. B. Hemmer has served on Celgene, and Merck. All conflicts are not relevant to the topic of scientific advisory boards for Novartis; he has served as DMSC the study. Dr. T. Kümpfel has received travel expenses and member for AllergyCare, Polpharma, and TG Therapeutics; speaker honoraria from Bayer HealthCare, Teva Pharma, Merck, he or his institution has received speaker honoraria from Novartis Pharma, Sanofi Aventis/Genzyme, CLB Behring, Desitin; his institution received research grants from Roche Pharma, and Biogen as well as grant support from Bayer Regeneron for MS research; he holds part of 2 patents: one Schering AG, Novartis, and Chugai Pharma. All conflicts are not for the detection of antibodies against KIR4.1 in a sub- relevant to the topic of the study. Dr. U.K. Zettl has received population of patients with MS and one for genetic deter- research grants and/or speaker honorary from Almirall, Aventis, minants of neutralizing antibodies to interferon. All conflicts Bayer, Biogen, Merck Serono, Novartis, Roche, Teva, and are not relevant to the topic of the study. Go to Neurology. Bundesministerium für Bildung und Forschung (BMBF). All org/NN for full disclosures. conflicts are not relevant to the topic of the study. Dr. R.A. Linker received research support and/or personal compen- Publication history sation for activities with Bayer HealthCare, Biogen, Received by Neurology: Neuroimmunology & Neuroinflammation Genzyme/Sanofi, Merck, Novartis Pharma, Roche, and Teva February 28, 2020. Accepted in final form May 26, 2020. Pharma. Dr. U. Ziemann has received grants from European Research Council, German Research Foundation, German Ministry of Education and Research, Biogen Idec GmbH, Servier, and Janssen Pharmaceuticals NV, all not related to Appendix Authors this work, and consulting honoraria from Biogen Idec Name Location Contribution GmbH, Bayer Vital GmbH, Bristol-Myers Squibb GmbH, fi Christiane Technical University Munich, Designed and P zer, CorTec GmbH, and Medtronic GmbH, all not related Gasperi, MD Munich, Germany conceptualized the to this work. Dr. M. Knop received honoraria for serving on study; performed fi statistical analyses; scienti c advisory boards and as a speaker from Merck and drafted the Serono, Novartis, Genzyme, Biogen, Pfizer/BMS, and manuscript

Roche. Dr. C. Warnke received speaker honoraria (in- Till F.M. Technical University Munich, Acquired and stitutional only) and/or research funding from Biogen, Andlauer, PhD Munich, Germany analyzed the data and revised the Novartis,andRoche.Dr.M.A.Friesehasreceivedspeaker manuscript for honoraria from Biogen, Novartis, and EMD. All conflicts are intellectual content not relevant to this study. Dr. F. Paul served on the steering Ana Keating, Technical University Munich, Acquired and committee for Novartis OCTIMS study and MedImmune; MSc Munich, Germany analyzed the data and received speaker honoraria and travel funding from Bayer, revised the manuscript for Novartis, Biogen Idec, Teva, Sanofi Aventis/Genzyme, intellectual content Merck Serono, Alexion, Chugai, and MedImmune; is an Benjamin Technical University Munich, Acquired and academic editor for PLoS One; is an associate editor for Knier, MD Munich, Germany analyzed the data and Neurology®: Neuroimmunology & Neuroinflammation;has revised the fi manuscript for consulted for Sano Genzyme, Biogen Idec, and MedI- intellectual content mmune; and received research support from Bayer, Novar- fi Ana Klein, MD Technical University Munich, Acquired and tis, Biogen Idec, Teva, Sano Aventis/Genzyme, Alexion, Munich, Germany analyzed the data and Merck Serono, German Research Council, Werth Stiftung of revised the manuscript for the City of Cologne, German Ministry of Education and intellectual content Research (BMBF Competence Network Multiple Sclerosis), Verena Technical University Munich, Acquired and Guthy-Jackson Charitable Foundation, National Multiple Pernpeintner, Munich, Germany analyzed the data and Sclerosis Society, and Arthur Arnstein Stiftung Berlin. Dr. B. MD revised the Tackenberg received personal speaker honoraria and con- manuscript for intellectual content sultancy fees as a speaker and advisor from Alexion, Bayer

10 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Appendix (continued) Appendix (continued)

Name Location Contribution Name Location Contribution

Peter Helmholtz Zentrum München, Acquired and analyzed Matthias Max Planck Institute of Acquired and Lichtner, PhD Neuherberg, Germany the data and revised Knop, MD Psychiatry, Munich, Germany analyzed the data and the manuscript for revised the intellectual content manuscript for intellectual content Ralf Gold, MD Ruhr-University Bochum, Acquired and Bochum, Germany analyzed the data and Clemens Medical Faculty, Heinrich Acquired and revised the Warnke, MD Heine University Düsseldorf, analyzed the data and manuscript for Düsseldorf, Germany revised the intellectual content manuscript for intellectual content Frauke Zipp, Johannes Gutenberg Acquired and MD University Mainz, Mainz, analyzed the data and Manuel A. University Medical Centre Acquired and Germany revised the Friese, MD Hamburg-Eppendorf, analyzed the data and manuscript for Hamburg, Germany revised the intellectual content manuscript for intellectual content Florian Then University of Leipzig, Leipzig, Acquired and Bergh, MD Germany analyzed the data and Friedemann Charité—Universitätsmedizin Acquired and revised the Paul, MD Berlin, Berlin, Germany analyzed the data and manuscript for revised the intellectual content manuscript for intellectual content Martin Hannover Medical School, Acquired and Stangel, MD Hannover, Germany analyzed the data and Bjorn¨ Philipps-University Marburg, Acquired and revised the Tackenberg, Marburg, Germany analyzed the data and manuscript for MD revised the intellectual content manuscript for intellectual content Hayrettin Clinic of Neurology Acquired and Tumani, MD Dietenbronn, Schwendi, analyzed the data and Achim Technical University Munich, Acquired and Germany revised the Berthele, MD Munich, Germany analyzed the data and manuscript for revised the intellectual content manuscript for intellectual content Brigitte University Hospital Acquired and Wildemann, Heidelberg, Heidelberg, analyzed the data and Bernhard Technical University Munich, Designed and MD Germany revised the Hemmer, MD Munich, Germany conceptualized the manuscript for study; supervised the intellectual content research; and revised the manuscript for Heinz Wiendl, University of Münster, Acquired and intellectual content MD Münster, Germany analyzed the data and revised the manuscript for intellectual content References Antonios University Hospital Augsburg, Acquired and 1. Reiber H, Ungefehr S, Jacobi C. The intrathecal, polyspecific and oligoclonal immune Bayas, MD Augsburg, Germany analyzed the data and response in multiple sclerosis. Mult Scler 1998;4:111–117. revised the 2. Bonnan M. Intrathecal IgG synthesis: a resistant and valuable target for future mul- manuscript for tiple sclerosis treatments. Mult Scler Int 2015;2015:296184. intellectual content 3. Stauch C, Reiber H, Rauchenzauner M, et al. Intrathecal IgM synthesis in pediatric MS is not a negative prognostic marker of disease progression: quantitative versus Tania Ludwig-Maximilians Acquired and qualitative IgM analysis. Mult Scler 2011;17:327–334. Kümpfel, MD University Munich, Munich, analyzed the data and 4. Walsh MJ, Tourtellotte WW. Temporal invariance and clonal uniformity of brain and Germany revised the cerebrospinal IgG, IgA, and IgM in multiple sclerosis. J Exp Med 1986;163:41–53. manuscript for 5. Buck D, Albrecht E, Aslam M, et al. Genetic variants in the immunoglobulin heavy intellectual content chain locus are associated with the IgG index in multiple sclerosis. Ann Neurol 2013; 73:86–94. Uwe K. Zettl, University of Rostock, Acquired and 6. Goris A, Pauwels I, Gustavsen MW, et al. Genetic variants are major determinants of MD Rostock, Germany analyzed the data and CSF antibody levels in multiple sclerosis. Brain 2015;138:632–643. revised the 7. Moutsianas L, Jostins L, Beecham AH, et al. Class II HLA interactions modulate manuscript for genetic risk for multiple sclerosis. Nat Genet 2015;47:1107–1113. intellectual content 8. Andlauer TF, Buck D, Antony G, et al. Novel multiple sclerosis susceptibility loci implicated in epigenetic regulation. Sci Adv 2016;2:e1501678. Ralf A. Linker, University of Regensburg, Acquired and 9. von Bismarck O, Dankowski T, Ambrosius B, et al. Treatment choices and neuro- MD Regensburg, Germany analyzed the data and psychological symptoms of a large cohort of early MS. Neurol Neuroimmunol revised the Neuroinflamm 2018;5:e446. 10.1212/NXI.0000000000000446. manuscript for 10. Poser CM, Paty DW, Scheinberg L, et al. New diagnostic criteria for multiple sclerosis: intellectual content guidelines for research protocols. Ann Neurol 1983;13:227–231. 11. McDonald WI, Compston A, Edan G, et al. Recommended diagnostic criteria for Ulf Ziemann, Eberhard-Karls-Universität Acquired and multiple sclerosis: guidelines from the International Panel on the Diagnosis of Mul- MD Tübingen, Tübingen, Germany analyzed the data and tiple Sclerosis. Ann Neurol 2001;50:121–127. revised the 12. Polman CH, Reingold SC, Edan G, et al. Diagnostic criteria for multiple sclerosis: manuscript for 2005 revisions to the “McDonald Criteria”. Ann Neurol 2005;58:840–846. intellectual content 13. Polman CH, Reingold SC, Banwell B, et al. Diagnostic criteria for multiple sclerosis: 2010 revisions to the McDonald criteria. Ann Neurol 2011;69:292–302.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 11 14. Cepok S, Rosche B, Grummel V, et al. Short-lived plasma blasts are the main B cell 23. Lefranc MP, Lefranc G. Human Gm, Km, and Am allotypes and their molecular effector subset during the course of multiple sclerosis. Brain 2005;128:1667–1676. characterization: a remarkable demonstration of polymorphism. In: Christiansen F, 15. Teo YY, Inouye M, Small KS, et al. A genotype calling algorithm for the Illumina Tait B, ed. Immunogenetics Methods in Molecular Biology (Methods and Protocols). BeadArray platform. Bioinformatics 2007;23:2741–2746. Totowa: Humana Press, 2012. 16. Purcell S, Neale B, Todd-Brown K, et al. PLINK: a tool set for whole-genome association 24. Seppala IJ, Sarvas H, Makela O. Low concentrations of Gm allotypic subsets G3 mg and population-based linkage analyses. Am J Hum Genet 2007;81:559–575. and G1 mf in homozygotes and heterozygotes. J Immunol 1993;151:2529–2537. 17. Chang CC, Chow CC, Tellier LC, Vattikuti S, Purcell SM, Lee JJ. Second-generation 25. Morell A, Skvaril F, Steinberg AG, Van Loghem E, Terry WD. Correlations between PLINK: rising to the challenge of larger and richer datasets. Gigascience 2015;4:7. the concentrations of the four sub-classes of IgG and Gm Allotypes in normal human 18. International HIV Controllers Study, Pereyra F, Jia X, McLaren PJ, et al. The major sera. J Immunol 1972;108:195–206. genetic determinants of HIV-1 control affect HLA class I peptide presentation. Sci- 26. Einarsdottir H, Ji Y, Visser R, et al. H435-containing immunoglobulin G3 allotypes are ence 2010;330:1551–1557. transported efficiently across the human placenta: implications for alloantibody- 19. Jia X, Han B, Onengut-Gumuscu S, et al. Imputing amino acid polymorphisms in mediated diseases of the newborn. Transfusion 2014;54:665–671. human leukocyte antigens. PLoS One 2013;8:e64683. 27. Pan Q, Petit-Frere C, Hammarstrom L. An allotype-associated polymorphism in the 20. Browning BL, Browning SR. A unified approach to genotype imputation and gamma3 promoter determines the germ-line gamma3 transcriptional rate but does haplotype-phase inference for large data sets of trios and unrelated individuals. Am J not influence switching and subsequent IgG3 production. Eur J Immunol 2000;30: Hum Genet 2009;84:210–223. 2388–2393. 21. Tingley D, Yamamoto T, Hirose K, Keele L, Imai K. Mediation: R package for causal 28. Vidarsson G, Dekkers G, Rispens T. IgG subclasses and allotypes: from structure to mediation analysis. J Stat Sotw 2014;59:1–38. effector functions. Front Immunol 2014;5:520. 22. R Core Team. R: A Language and Environment for Statistical Computing. Vienna, 29. Mele M, Ferreira PG, Reverter F, et al. Human genomics. The human transcriptome Austria: R Core Team; 2018. Available at: https://www.R-project.org/. across tissues and individuals. Science 2015;348:660–665.

12 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN ARTICLE OPEN ACCESS Motor cortex transcriptome reveals microglial key events in amyotrophic lateral sclerosis

Oriol Dols-Icardo, PhD, V´ıctor Montal, MSc, Sonia` Sirisi, PhD, Gema Lopez-Pernas,´ MSc, Correspondence ´ Laura Cervera-Carles, PhD, Marta Querol-Vilaseca, MSc, Laia Muñoz, BSc, Olivia Belbin, PhD, Dr. Clarimon [email protected] ´ Daniel Alcolea, MD, PhD, Laura Molina-Porcel, MD, PhD, Jordi Pegueroles, MSc, Janina Turon-Sans, MD, or Dr. Dols-Icardo Rafael Blesa, MD, PhD, Alberto Lleo,´ MD, PhD, Juan Fortea, MD, PhD, Ricard Rojas-Garc´ıa, MD, PhD, and [email protected] Jordi Clarimon,´ PhD

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e829. doi:10.1212/NXI.0000000000000829 Abstract Objective To identify transcriptomic changes, neuropathologic correlates, and cellular subpopulations in the motor cortex of sporadic amyotrophic lateral sclerosis (ALS).

Methods We performed massive RNA sequencing of the motor cortex of patients with ALS (n = 11) and healthy controls (HCs; n = 8) and analyzed gene expression alterations, diﬀerential isoform usage, and gene coexpression networks. Furthermore, we used cell type deconvolution algorithms with human single-nucleus RNA sequencing data as reference to identify perturbations in cell type composition associated with ALS. We performed immunohistochemical techniques to evaluate neuropathologic changes in this brain region.

Results We report extensive RNA expression alterations at gene and isoform levels, characterized by the enrichment of neuroinﬂammatory and synaptic-related pathways. The assembly of gene coexpression modules conﬁrmed the involvement of these 2 major transcriptomic changes, which also showed opposite directions related to the disease. Cell type deconvolution revealed an overrepresentation of microglial cells in ALS compared with HC. Notably, microgliosis was driven by a subcellular population presenting a gene expression signature overlapping with the recently described disease-associated microglia (DAM). Using immunohistochemistry, we further evidenced that this microglial subpopulation is overrepresented in ALS and that the density of pTDP43 aggregates negatively correlates with the proportion of microglial cells.

Conclusions DAM has a central role in microglia-related neuroinﬂammatory changes in the motor cortex of patients with ALS, and these alterations are coupled with a reduced expression of postsynaptic transcripts.

From the Memory Unit (O.D.-I., V.M., S.S., G.L.-P., L.C.-C., M.Q.-V., L.M., O.B., D.A., J.P., R.B., A.L., J.F., J.C.), Neurology Department and Sant Pau Biomedical Research Institute, Hospital de la Santa Creu i Sant Pau, Universitat Autonoma` de Barcelona; Network Center for Biomedical Research in Neurodegenerative Diseases (CIBERNED) (O.D.-I., V.M., S.S., G.L.-P., L.C.- C., M.Q.-V., L.M., O.B., D.A., J.P., R.B., A.L., J.F., J.C.), Madrid; Neurological Tissue Bank of the Biobanc-Hospital Cl´ınic-IDIBAPS (L.M.-P.), Barcelona; Alzheimer’s Disease and Other Cognitive Disorders Unit, Neurology Department (L.M.-P.), Hospital Cl´ınic, Institut d’Investigacions Biom`ediques August Pi i Sunyer, University of Barcelona; Network Center for Biomedical Research in Rare Diseases (CIBERER) (J.T.-S., R.R.-G.), Madrid; and Neuromuscular Disorders Unit (J.T.-S., R.R.-G.), Department of Neurology, Hospital de la Santa Creu i Sant Pau, Universitat Autonoma` de Barcelona, Spain.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary ALS = amyotrophic lateral sclerosis; DAM = disease-associated microglia; FFPE = formalin-ﬁxed paraﬃn-embedded; GO = gene ontology; HC = healthy control; IHC = immunohistochemistry; KEGG = Kyoto Encyclopedia of Genes and Genomes; MuSiC = Multi-Subject Single Cell deconvolution; NTB = Neurological Tissue Bank; qPCR = quantitative real-time PCR; RIN = RNA integrity; RNAseq = RNA sequencing; ROI = region of interest; ROSMAP = Religious Orders Study and Memory and Aging Project; snRNAseq = single-nucleus RNAseq; TREM2 = triggering receptor expressed on myeloid cells 2; WGCNA = weighted gene coexpression network analyses.

Amyotrophic lateral sclerosis (ALS) is a neurodegenerative The brain motor cortex is affected at the most early stages of disease neuropathologically characterized by the aberrant the disease and is one of the most vulnerable regions in ALS. cytoplasmic aggregation and phosphorylation of the 43-kDa However, studies of this critical region both at the transcrip- transactive response DNA-binding protein (pTDP43) in tional and immunohistochemical level are lacking. In the the majority of ALS cases at postmortem evaluation, which present work, we aim to characterize the motor cortex of is known to occur in a sequential manner starting in the sporadic ALS cases through total RNAseq analyses. We also motor cortex.1 Although the causes that lead to nongenetic use cell type deconvolution using human snRNAseq data as forms of ALS (sporadic ALS) have yet to be fully de- reference and immunohistochemical analyses to resolve termined, the presence of neuroinflammation is a consistent a signature of neuropathologic changes associated with ALS in feature in the CNS of affected patients.2 Microglia are this particular brain region. macrophage-like innate immune cells of the CNS that, as a result of disease conditions, change their gene expression profile to adapt and acquire a reactive state.3 Recently, a study using single-cell RNA sequencing from the CNS of Methods Alzheimer disease and ALS mouse models reported that Human samples aspecific microglial subpopulation, known as disease- The study included human motor cortex (Brodmann area 4) associated microglia (DAM), was the cell type responsible samples provided by the Neurological Tissue Bank (NTB) of for microgliosis and that triggering receptor expressed on the Biobanc-Hospital Cl´ınic-IDIBAPS. None of the brain myeloid cells 2 (TREM2) is a major driver of DAM acti- tissues presented any infarcts in the motor cortex. Diagnosis vation.4 In support of this discovery, some chitinase proof ALS complied with the El Escorial criteria during life,14 and teins, which are known markers of microglial activation,5,6 none of the patients had a family history of ALS or dementia, have been consistently shown to be increased in the CSF of nor did show any sign of cognitive impairment. All patients patients with ALS.7,8 presented pTDP43 inclusions in the motor cortex at postmortem examination. None of the samples carried the The pivotal role of TDP43 and other genes known to cause C9orf72 hexanucleotide repeat expansion or mutations in the ALS (such as FUS, hnRNPA2B1, hnRNPA1, TAF15 or TBK1 gene, the most common genetic causes related to adult- TIA1) in the metabolism of RNA has implicated RNA onset ALS in Spain.15,16 dyshomeostasis as a crucial event in the pathophysiology of the disease.9 An unbiased resource to obtain a compre- RNA extraction and sequencing hensive signature of gene and isoform expression changes Using a mortar and liquid nitrogen, the tissue (60 mg) was associated with the whole tissue response to disease con- grinded to powder and transferred to a solution of 600 μLof ditions is total RNA sequencing (RNAseq) in bulk tissue. TRIzol reagent (Thermo Fisher Scientific, Waltham, MA). Furthermore, the advent of novel sequencing tools such as We used standard recommendations and procedures to ex- single-nucleus RNAseq (snRNAseq) has opened a new tract RNA with TRIzol, the RNeasy Mini Kit (Qiagen, Hilden, window to better explore the transcriptome at a cell level Germany), and the Rnase-Free Dnase Set (Qiagen). Qubit and has unraveled a complex and huge variety of human was used to measure RNA concentration, whereas RNA in- cell types with unique expression profiles in the human tegrity (RIN) was verified on an Agilent 2100 Bioanalyzer – brain.10 12 These transcriptomic signatures can be applied (Agilent Technologies, Santa Clara, CA). Only samples with to perform cell type deconvolution of bulk RNAseq data a quantity threshold of 5 μg and RIN ≥6.5 were used for total using novel deconvolution methods. These methods take RNAseq. The final study group included 11 ALS cases and 8 into account cross-subject and cross-cell variability of gene healthy controls (HCs) (table 1). Paired-end sequencing li- expression profiles, without relying on preselected mark- braries were prepared using the TruSeq Stranded Total RNA ers, thereby providing more realistic cell type proportion Library Preparation kit (Illumina, San Diego, CA) and se- estimations and yielding crucial information about the quenced on the Illumina HiSeq 2500 platform by the Centro cellular heterogeneity that is associated with a pathologic Nacional de Análisis Genómicos (CNAG; Barcelona, Spain) condition.13 with 101-bp paired-end reads to achieve at least 100 million

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Table 1 Demographic, clinical, and quality control data

Sample ID Group Age at death, y Sex Age at onset, y Duration, mo Site of onset RIN PMI, h

AF5214 ALS 64 Male 62 33 Limb 8.2 7

AF5215 ALS 50 Male 46 58 Limb 8.8 13

AF5216 ALS 60 Female 54 72 Limb 8.3 6

AF5218 ALS 53 Female 50 32 Limb 8.4 9

AF5220 ALS 82 Male 81 1 Bulbar 7.7 16

AF5222 ALS 54 Female 53 18 Limb 9 7

AF5224 ALS 70 Female 68 24 Limb 8.4 14

AF5227 ALS 78 Male 76 24 Limb 8.1 12

AF5228 ALS 77 Female 75 31 Bulbar 7 16

AF5229 ALS 57 Female 54 24 Limb 6.8 5

AF5232 ALS 52 Male 51 12 Limb 7.4 18

AF5219 HC 64 Male ———8.2 10

AF5221 HC 78 Male ———7.9 6

AF5226 HC 81 Female ———7.5 23

AF5231 HC 83 Female ———7.3 7

AF5234 HC 58 Male ———7.5 4

AF5235 HC 76 Male ———7.5 11

AF5236 HC 83 Female ———7.7 7

AF5237 HC 68 Female ———6.5 13

Abbreviations: ALS = amyotrophic lateral sclerosis; HC = healthy control; PMI = postmortem interval; RIN = RNA integrity number. paired-end reads for each sample (ranging from 103.447.000 Validation of transcriptome changes by to 138.990.000 paired-end reads). quantitative real-time PCR We performed quantitative real-time PCR (qPCR) on the Data processing same brain-derived RNA samples used for total RNAseq (11 fi We aligned FASTQ les to the Grch38.p12 genome assembly sALS cases and 8 controls). A total of 500 ng of RNA was used 17 18 using STAR v2.6.1a. We then used GATK Best Practices to generate cDNA with the RevertAid First Strand cDNA for variant calling and FeatureCounts (within the Subread Synthesis Kit (Thermo Fisher Scientific), as per the manu- 19 v.1.6.2 package ) to assign fragments to each gene feature facturer’s instructions. All qPCRs were conducted using Fast fi included in the Grch38v94 gene transfer format le. For dif- SYBR Green Master Mix (Thermo Fisher Scientific) and run ferential isoform usage, we used the STAR v.2.6.1 package in on an ABI Prism 7900HT Fast Real-Time PCR System quantMode TranscriptomeSAM mode and used Salmon to (Applied Biosystems, Foster City, CA). All primer pairs used 20 ff obtain isoform expression values. For di erential gene ex- are listed in table e-1, links.lww.com/NXI/A280. For relative ff pression and di erential isoform usage analyses, we used quantification, we applied the DDCt method, using RPL13 as 21 DESeq2 v1.24, which applies the Benjamini-Hochberg for p an endogenous control to normalize the data. We decided to value adjustment. We used a false discovery rate threshold of use RPL13 as a housekeeping gene as it showed the smallest 5%. To perform gene ontology (GO) and Kyoto Encyclo- coefficient of variation (<3.5%) compared with the 2 other pedia of Genes and Genomes (KEGG) pathway enrichment putative housekeeping genes that we assessed (RPL0 and 22 analyses, we used Metascape 3.0 (metascape.org), applying GAPDH). default parameters (minimum overlap = 3, p value cutoff = 0.01, and minimum enrichment = 1.5). For the synaptic en- Cell type deconvolution richment analyses, we used the web server SynGO (syngo- We used the recently developed Multi-Subject Single Cell portal.org/), which provides an expert-curated resource for deconvolution (MuSiC)13 method to perform cell type synaptic function and gene enrichment analysis.23 deconvolution of our bulk RNAseq data. As a reference, we

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 used the most comprehensive human snRNAseq data set peroxidase secondary antibody (1:200) for 1 hour at room available. This data set includes the RNAseq from 70,634 cells temperature. Primary antibodies were Iba1 rabbit poly- in the frontal cortex (Brodmann area 10) of 24 Alzheimer clonal (1:500; Wako Chemicals, Osaka, Japan) and ps409/ disease cases and 24 HCs from the Religious Orders Study 410-TDP rabbit polyclonal (1:2000; Cosmo Bio, Tokyo, and Memory and Aging Project (ROSMAP),12 grouped into Japan). Following peroxidase development with 3,39-dia- 8 major cellular populations (excitatory neurons, inhibitory minobenzidine tetrahydrochloride (Dako liquid DAB + neurons, microglia, astrocytes, oligodendrocytes, oligoden- substrate Chromogen System; Dako Denmark, Glostrup, drocyte precursor cells, endothelial cells, and pericytes). To Denmark), sections were stained with hematoxylin (EnVi- check consistency and validate our results, we used 2 in- sion FLEX Haematoxylin; Dako Denmark) and dehydrated dependent snRNAseq repositories as a reference to estimate using increasing concentrations of ethanol and then major cell type proportions. To this aim, we downloaded the mounted with DPX mounting medium (PanReac Appli- medial temporal gyrus snRNAseq data derived from the Allen Chem, ITW Reagents, Chicago, IL). For immunofluores- Brain Atlas, which includes 15,928 cells from 8 human tissue cence, we used 3 ALS and 3 control motor cortex sections, donors. We selected this data set as it contains double the which were incubated with primary antibodies overnight at number of nuclei than the primary visual cortex and the an- 4°C. We used Iba1 rabbit polyclonal antibody as a marker terior cingulate cortex, also contained in the database.10 The for the broad microglial population (1:500; Wako Chem- third data set comprised 10,319 cells derived from the frontal icals) and human HLA-DP, DQ, DR antigen (MHC class II cortex of 4 individuals.11 Cell type deconvolution of cellular markers), mouse clone CR3/43 (1:20; Agilent, subpopulations was performed using the ROSMAP data as M077501–2), as a marker for activated cells. MHCII marker reference. genes are specifically expressed in the Mic1 gene signature.12 Thus, we decided to combine these primary anti- Weighted gene coexpression network analyses bodies to visualize Mic1 cells as performed in the study of We accounted for genes with more than 10 counts across all Mathys et al.12 as well as to estimate the proportion of cells individuals. DESeq2 was then used to normalize the input of this particular subpopulation. Then, we incubated sec- data to construct coexpression networks using the weighted tions for 1 hour at room temperature with secondary anti- 24 gene coexpression network analyses (WGCNA) package. bodies (1:1,000), which were conjugated with Alexa Fluor We then constructed the signed weighted correlation net- 488 and Alexa Fluor 555 (Invitrogen, Carlsbad, CA). Cell work, which takes into account both negative and positive nuclei visualized with Hoechst 33258 (Life Technologies). correlations, using the manual function in WGCNA, applying Sections were mounted with Shandon Immu-Mount the biweight midcorrelation, selecting a power of 7 (the (Thermo Fisher Scientific). We acquired all the images lowest possible power term where topology fits a scale free with a Leica TCS SP5 Confocal Laser Scanning Microscope, network), and run in a single block analysis. We defined all with a 63× objective. modules by the hybrid treecutting option with deepsplit parameter = 2, applying a minimal module size of 30 genes and Quantification analysis merging modules with a cutHeight ≤0.25. Module eigengenes We obtained full-section immunohistochemistry (IHC) (first principal component as a summary of each module) images with Pannoramic MIDI II (3DHistech). Blinded to summarized the modules and correlations performed using the neuropathologic diagnoses, we delimited 6 gray matter spearman correlation for continuous traits and Pearson cor- areas of the motor cortex from each case, and for each gray relation for binary traits (bicor and corPvalueStudent, in- matter area, we randomly generated 6 regions of interest cluded in the WGCNA package). We used Cytoscape 3.7.2 to (ROIs)withthesamesizeusingthemulticrop.mscript.We visualize networks. used the Iba1.m script to quantify densities for Iba1 staining in each ROI, which represents a total of 36 images per in- Immunohistochemistry dividual. This algorithm binarizes the random images and We used formalin-fixed paraffin-embedded (FFPE) motor compute densities of protein expression to quantify the cortex tissue sections of 5 μm in 11 patients with ALS and 7 number of immunoreactive objects. Both in-house algo- controls. One control sample (AF5237) was excluded from rithms can be freely accessed at github.com/Memo- the analysis due to lack of availability of FFPE motor cortex ryUnitSantPau. We used MATLAB R2017b software to sections. We dewaxed the FFPE sections by placing them in develop the algorithm (The MathWorks Inc., Natick, MA). xylene and decreasing concentrations of ethanol, followed For pTDP43 quantification,2different researchers blinded by hydration 30 minutes in distilled water. We then treated to neuropathologic diagnoses quantified immunoreactive tissue sections with a solution containing 10% methanol, 3% positive objects manually. We evaluated immunofluores- H2O2, and PBS for 10 minutes to inhibit endogenous per- cence using FIJI imaging software. We estimated an auto- oxidase and boiled in citrate 1x solution to mediate antigen mated threshold to create binary images. Following the noise retrieval. Once cooled, we blocked all sections with 5% removal with the despeckle and the remove outliers’ filters, bovine serum albumin for 1 hour at room temperature and we determined the number of Iba1+ microglial cells incubated them overnight with the corresponding primary expressing MHC class II markers using the analyze particles antibody at 4°C and then with the anti-rabbit horse radish function.

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Statistical analysis measures in these 10 genes (table e-3, links.lww.com/NXI/ We used the Shapiro-Wilk test to test for deviation from A280). GO and KEGG pathway analyses revealed an enriched a standard distribution. For correlation analyses, we de- involvement of the inflammatory response in ALS (figure termined the Pearson or Spearman correlation coefficients 1C). The assessment of changes at the isoform expression using the ggscatter function within the ggpubr package in R. level resulted in 167 upregulated and 40 downregulated iso- We performed mean comparison analyses through the Stu- forms, of which 181 were from unique genes (table e-4, links. dent t test or Student test with Welch correction (using t.test lww.com/NXI/A280). Enrichment analyses revealed the in- function in R), depending on data distribution. Statistical volvement of postsynaptic density and immune response as significance for all tests was set at 5% (α = 0.05), and all dominant pathways (figure e-1, links.lww.com/NXI/A279). statistical tests were 2 sided. We used R version 3.6.2 to To further characterize and confirm the synaptic involvement, perform all statistical analyses. we used the SynGO database, which provides an expert- curated resource for synaptic function and gene enrichment Standard protocol approvals, registrations, analysis.23 This analysis confirmed the enrichment of post- and patient consents synaptic components in our list of differentially expressed The Sant Pau Hospital Ethics Committee approved the study isoforms (table e-5, links.lww.com/NXI/A280). Among the with all experiments with human tissue performed in accor- list of altered genes identified by the differential isoform ex- dance with the Declaration of Helsinki. The NTB of the pression and gene expression models, 32 of them overlapped Biobanc-Hospital Cl´ınic-IDIBAPS supplied all tissue follow- between both approaches (table e-6, links.lww.com/NXI/ ing approval from their Scientific Advisory Committee. A280). The involvement of synaptic-related changes was confirmed from the differential isoform usage analysis (figure Data availability statement e-2, links.lww.com/NXI/A279), whereas the enrichment of fi We have deposited all raw sequencing data (FASTQ les) at inflammatory markers was evidenced in the gene model ap- the European Genome-phenome Archive (EGA), which is proach (figure e-3, links.lww.com/NXI/A279). hosted by the EBI and the CRG under accession number EGAS00001004286. Cell type deconvolution To characterize the variability of cell types and compare cellular proportions between groups, we applied the recently Results developed MuSiC algorithm. As reference, we used data from Demographics, clinical features, and the first and most comprehensive human snRNAseq study sample composition performed in brain tissue of individuals with a neurodegen- We performed RNAseq analysis of postmortem motor cor- erative disease, which includes 24 patients with Alzheimer 12 tex samples from 11 ALS cases (6 females) and 8 individuals disease and 24 HCs. Cell type deconvolution revealed an without neurologic disease (4 females). Postmortem interval overrepresentation of microglial cells in the ALS motor cortex and RIN did not differ between groups (p = 0.68 and p = compared with HC (p = 0.025, fold change = 1.65) and 0.089, respectively). The mean age at death was lower for a reduction in excitatory neurons (p = 0.101, fold change = ALS cases compared with HC (63.4 years; range: 54–82 -0.85) (figure 2). We confirmed microglial upregulation using 10,11 years in ALS and 73.9 years; range: 58–83 years in HC, p = 2 additional independent reference data sets. In both 0.04). Patients had an average disease onset of 60.9 years cases, we found the same pattern of downregulation of ex- (range: 46–81 years) and a disease duration of 29.9 months citatory neurons (figures e-4 and e-5, links.lww.com/NXI/ (range: 1–72 months). Nine of the 11 cases (81.8%) pre- A279). To further gain insight into which microglial sub- sented a limb disease onset (table 1). We used RNAseq data population might be driving this effect, we estimated the to rule out the presence of other ALS-disease causing proportion of each subpopulation identified in the ROSMAP mutations, confirming the presumably sporadic nature of our study. Our results showed that a specific microglial sub- ALS group of cases. population (Mic1, as named in the study performed by Mathys and collaborators and recognized as the human Gene expression alterations and differential DAM12) is disproportionately presented in the ALS motor − isoform usage cortex (p = 6.6 × 10 3, fold change = 4.33) (figure 2). Among Differential gene expression between patients with ALS and the 77 marker genes, which characterize the Mic1 subcluster HCs disclosed a total of 108 upregulated and 16 down- as DAM (Mic1 unique genes or those overlapping with regulated genes (adjusted p < 0.05) (figure 1A and table e-2, DAM),12 22 of them (28.6%) are genes differentially links.lww.com/NXI/A280). We validated 9 of the 10 genes expressed in our bulk RNAseq analysis (unadjusted p < 0.05; selected by qPCR, as we did not confirm the CHI3L2 gene (table e-7, links.lww.com/NXI/A280). Of interest, the pro- overexpression in patients with ALS (p = 0.085) (figure 1B). portion of Mic1 cells showed a positive correlation with the − Correlation analyses between counts derived from total expression of TREM2 (p = 1.6 × 10 3; R = 0.67; figure e-6, RNAseq data and qPCR expression values indicated a high links.lww.com/NXI/A279), a receptor required for the tran- and significant concordance between both expression sition from the homeostatic microglia to DAM.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 5 Figure 1 Differential regulation of gene expression in ALS and HC

(A) Volcano plot displaying differentially expressed genes between the ALS and the HC motor cortex. The vertical axis (y-axis) corresponds to the −log10 adjusted p value, and the horizontal axis (x-axis) represents the log2 fold change value obtained when comparing ALS with HC gene expression. Significantly differential expressed genes are depicted with blue circles (adjusted p value <0.05), whereas gray circles display the nonsignificant genes. Gene names are shown for the 10 genes selected for qPCR validation. (B) Box plot showing the relative RNA expressionvaluesinALSandHCobtained through qPCR for the 10 genes selected for validation. Dashed line corresponds to reference relative gene expression. *p < 0.05; **p < 0.01. (C) Top 5 gene ontology and KEGG pathway enrichment analyses obtained from the list of genes showing a significant differential expression. ALS = amyotrophic lateral sclerosis; HC = healthy control.

Gene coexpression network analyses inflammation and synaptic alterations are interconnected in To identify gene clusters with varying coexpression patterns ALS pathologic processes. Furthermore, the proportion of that could behave differently between patients and healthy Mic1 cells showed a strong direct correlation with the 2 in- individuals, and elucidate possible biological mechanisms flammatory gene coexpression modules associated with ALS − driving pathologic processes, we performed a gene coex- (MEblack; p = 2.7 × 10 8, R = 0.92; and MEyellow; p = 1.4 × − pression network analyses using the WGCNA package. The 10 3, R = 0.68) and a negative correlation with the synaptic − analysis disclosed a total of 8 modules. Among them, 3 module (MEpink; p = 2.4 × 10 4, R = −0.75) (figure 3). modules associated with disease status (MEblack, p = 0.003; R Overall, our results from differential gene and isoform ex- = 0.64; MEyellow, p = 0.025; R = 0.57; and MEpink, p = pression, gene coexpression modules, and cell type decon- 0.011; R = −0.51; figure 3). Moreover, MEblack and MEyel- volution comparisons strongly indicate that microglia-related low were enriched for inflammatory responses, whereas inflammatory changes, mainly driven by the Mic1 sub- MEpink had an overrepresentation of genes involved in population (also known as DAM), are central in ALS path- synaptic and neuronal functions (table e-8, links.lww.com/ ophysiology and that these inflammatory processes are closely NXI/A280). The SynGO database confirmed the post- related to the synaptic disturbances that are present in the synaptic signature of this enrichment in MEpink (table e-9, motor cortex. links.lww.com/NXI/A280). Of interest, the 2 inflammatory gene coexpression modules (MEblack and MEyellow) Immunohistochemistry showed a strong negative correlation with the postsynaptic To further investigate the inflammatory response in post- − − module (MEpink) (p =8×10 5; R = −0.78 and p =3×10 4; mortem brain tissue, we performed IHC analyses using Iba1, R = −0.74, respectively) (figure 3), suggesting that both a protein whose expression is primarily restricted to homeostatic

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Figure 2 Cell type composition in the ALS and HC motor cortex

Box plot showing major cell type and Mic1 proportions for HC and ALS estimated by MuSiC using the ROSMAP human single-nucleus RNAseq data. *p < 0.05; **p < 0.01. ALS = amyotrophic lateral sclerosis; HC = healthy control; MuSiC = Multi-Subject Single Cell deconvolution; ROSMAP = Religious Orders Study and Memory and Aging Project. microglia, to examine the broad population of microglial cells, compared with HCs (p = 0.017) (figure e-9A, links.lww.com/ and MHC class II, a marker responsible for antigen recognition NXI/A279). Of interest, a high degree of variability on the and the activation of the adaptive immune system, which is number of pTDP43 aggregates was noted in ALS and was not expressed in the Mic1 cell subpopulation. Positive immunos- explained by age at death, age at onset, or disease duration. taining was increased in the ALS motor cortex for Iba1; how- The burden of pTDP43 inclusions inversely correlated with ever, differences between patients and controls did not reach the proportion of microglial cells in our group of patients with statistical significance (p = 0.106; figure e-7, links.lww.com/ ALS (p = 0.0026; R = −0.81) (figure e-9B, links.lww.com/ NXI/A279). Notwithstanding, co-immunofluorescence of Iba1 NXI/A279), and the same pattern was found with the and MHC class II markers revealed an increase in the number of microglial subpopulation Mic1 (p = 0.019; R = −0.69). microglial cells expressing MHCII markers in ALS compared with HC (70.1% vs 10%, respectively; p = 0.021) (figure 4 and figure e-8, links.lww.com/NXI/A279), strengthening our Discussion results and suggesting that Mic1 is driving neuroinflammation in the ALS motor cortex. The motor cortex is one of the major vulnerable and early affected regions in ALS and represents a target region to We finally assessed the frequency of pTDP43 immunoreac- disentangle key pathologic processes in this neurodegenerative structures in the motor cortex. As expected, patients with tive disorder. To date, few studies have performed a whole – ALS showed a higher density of pTDP43 inclusions transcriptomic assessment of the CNS in ALS.25 27 Through

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 7 Figure 3 WGCNA visualization and correlation map constructed with WGCNA module eigengenes and cell type proportions

Network diagrams showing the 50 most connected genes for each of the 3 modules associated with ALS (MEblack, MEyellow, and MEpink) are depicted on the top of the figure. Node darkness and size are proportional to the number of connections within the module. Correlation matrix showing the correlation coefficients between cell type proportions is depicted for the 6 major cell types and the Mic1 (DAM) cell subtype. The 3 significant modules (MEblack, MEyellow, and MEpink), each of one sharing unique groups of coexpressed genes that are differentially expressed between patients with ALS and controls, are also included in the correlogram. Only significant correlations are depicted with a colored circle (blue for direct and red for inverse correlations). The size of the circle is proportional to the correlation significance. The plot indicates the high correlation between the 3 modules and how these modules are related to the proportion of specific cell types. ALS = amyotrophic lateral sclerosis; DAM = disease-associated microglia; HC = healthy control; WGGNA = weighted gene coexpression correlation network analyses.

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Figure 4 Immunohistochemistry of Mic1 markers

Immunohistochemistry with anti-IBA1 (red) and anti-MHC class II (green) antibodies in the ALS and HC motor cortex. Scale bar corresponds to 50 μm. ALS = amyotrophic lateral sclerosis; HC = healthy control. an unbiased transcriptomic analysis using high-throughput related to glial activation and inflammation.27 Our results RNA sequencing, we have elucidated gene and isoform ex- clearly reinforce the idea of these inflammatory-related pression alterations, gene coexpression networks, and cell changes in the human motor cortex and emphasize the role type proportions associated with ALS. For the first time in the of DAM as a key factor in this process. First, differential gene ALS field, we have performed cell type deconvolution using and isoform expression data point toward an inflammatory human brain single-nucleus RNA sequencing data from 3 response as a major event that is strongly intensified in the independent sources as reference, thus providing highly reli- ALS motor cortex. Second, the assembly of weighted gene able results that have been reinforced through our immuno- coexpression networks resulted in 2 significant modules as- histochemical analyses. sociated with ALS (MEblack and MEyellow), both highly enriched with genes related to inflammatory functions. Third, We report a list of 124 genes differentially expressed in the cell type deconvolution of the bulk RNAseq data demon- ALS motor cortex, which reflect the RNA expression changes strated an increased proportion of microglial cells compared in this brain region. Among them, chitinase-related genes with the motor cortex of healthy cases. In this context, studies CHI3L1 and CHI3L2 presented 2 of the most prominent in mice have recently evidenced DAM as the microglial sub- shifts in gene expression, whereas CHIT1 showed a nominal population with a more prominent role in ALS.4 Our results significance. Of interest, these chitinases are neuro- show that Mic1, the human microglial subpopulation that inflammatory biomarkers, which have been shown to be harbors the majority of markers found in the previously de- consistently increased in the CSF of patients with ALS scribed DAM transcriptomic signature,12 is the main micro- compared with neurologically healthy individuals.7,8 These glial subpopulation that drives microgliosis in ALS. In fact, results indicate that among the unbiased signature of RNA almost a third of Mic1 (DAM) marker genes are deregulated alterations provided herein, some of them might lead to the in the motor cortex of our group of patients with ALS. Of discovery of novel promising biomarkers. note, Mic1 proportion highly correlated with the expression of TREM2, which is required to enhance the proinflammatory An exacerbated innate immune response with microgliosis has stage of DAM.4 been recently described in the ALS motor cortex.28 A recent study that has investigated the whole transcriptome of the Our immunohistochemical analyses did not show a relevant ALS motor cortex suggested that among ALS cases, a sub- increased density of Iba1 (a widely used marker of microglial group of them is characterized by a molecular signature processes) in ALS-related brain tissue, nor was the expression

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 9 of its coding gene (AIF1) altered in our RNASeq data set. Overall, our study strongly suggests that DAM plays a key role Previous studies have frequently provided contradictory in driving neuroinflammatory changes and synapse loss in the results when using this marker to assess microgliosis.29 This ALS motor cortex. The identification of specific microglial somewhat unexpected result could be explained by the fact populations with well-defined transcriptional signatures will that some marker genes might downregulate on microglial contribute to disentangle new mechanisms and novel thera- activation. Also, whereas cell type deconvolution is performed peutic targets to fight against this devastating disorder. using a complete catalogue of gene expression profiles obtained from brain-derived snRNAseq, immunohistochem- Acknowledgment istry only uses a single marker and does not reflect the com- The authors are indebted to the IDIBAPS Biobank for sample plexity and heterogeneity of this cell type, making it an and data procurement. underpowered method to detect differences in the proportion of cell subpopulations. That being said, we did validate the Study funding increased proportion of Mic1 cells in the ALS motor cortex The authors are indebted to the “Fundación Española para el through co-immunofluorescence of 2 markers (Iba1 and Fomento de la Investigación de la Esclerosis Lateral MHC class II) previously shown to characterize this micro- Amiotrófica (FUNDELA)” for funding the present study. O. glial subpopulation.12 This finding further establishes Mic1 Dols-Icardo is a recipient of a grant by The Association for (the human disease-associated microglia) as the microglial Frontotemporal Degeneration (Clinical Research Post- subpopulation that drives microgliosis in the ALS motor doctoral Fellowship, AFTD 2019–2021). This work was cortex. supported by research grants from Institute of Health Carlos III (ISCIII), Spain PI18/00326 to J.C., PI18/00435 to D.A., We observed a high degree of variability in the density of and INT19/00016 to D.A., and by the Department of Health pTDP43 inclusions across patients, which could not be Generalitat de Catalunya PERIS program SLT006/17/125 to explained by any of the demographic or clinical features D.A. This work was also supported in part by Generalitat de available in this study. Notably, we found a striking inverse Catalunya (2017 SGR 00547) to the “Grup de Recerca en correlation between the proportion of microglial cells and the Demències: Sant Pau.” amount of pTDP43 aggregates. These results are in line with recent in vivo and in vitro studies, suggesting that although Disclosure microgliosis arises as a phagocytic response to pTDP43 The authors report no disclosures relevant to the manuscript. aggregates, at some point, these cells lose their ability to clear Go to Neurology.org/NN for full disclosures. these neuropathologic insults and are downregulated.30,31 Whether biofluid levels of microglial markers, such as Publication history TREM2, could be used as a proxy of pTDP43 density in the Received by Neurology: Neuroimmunology & Neuroinflammation fi motor cortex is an avenue worth pursuing and would be February 21, 2020. Accepted in nal form May 15, 2020. a valuable addition to the biomarker arsenal for use in designing clinical trials and assessing therapeutic efficacy.

32 Synaptic dysfunction is an early pathogenic event in ALS. Appendix Authors Our data point toward an underrepresentation of post- Name Location Contribution synaptic markers and a decrease of excitatory neurons in patients with ALS. Together, these results are consistent Oriol Dols- Sant Pau Biomedical Major role in the design Icardo, PhD Research Institute, Hospital and conceptualization of with upper motor neuron degeneration occurring in this de la Santa Creu i Sant Pau, the study; drafted the brain region. A limitation of our approach is the lack of an Barcelona, Spain manuscript; and analyzed available motor cortex snRNAseq data set, precluding any the data firm and more detailed conclusion related to the alteration of V´ıctor Sant Pau Biomedical Analyzed the data and Montal, MSc Research Institute, Hospital participated in manuscript Betz cells, a unique class of motor neurons expressed in this de la Santa Creu i Sant Pau, drafting specific brain area. Recent studies in other neurodegenera- Barcelona, Spain tive diseases have suggested that microglia are a key and early Sonia` Sirisi, Sant Pau Biomedical Analyzed the data and mediator of synapse loss through phagocytosis induced by PhD Research Institute, Hospital participated in manuscript the complement cascade.33,34 Our gene expression data de la Santa Creu i Sant Pau, drafting Barcelona, Spain show an increased expression of some key complement cascade-related genes and the most significant enriched Gema Sant Pau Biomedical Analyzed the data and Lopez-´ Research Institute, Hospital performed a part of the pathway corresponds to the complement and coagulation Pernas, MSc de la Santa Creu i Sant Pau, experiments cascades, reinforcing this hypothesis. Furthermore, our Barcelona, Spain fl results indicate that the 2 neuroin ammatory-related mod- Laura Sant Pau Biomedical Participated in manuscript ules of gene expression (MEblack and MEyellow) and the Cervera- Research Institute, Hospital drafting and data Carles, PhD de la Santa Creu i Sant Pau, interpretation proportion of Mic1 cells inversely correlate with the pres- Barcelona, Spain ence of synaptic markers.

10 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN 5. Steinacker P, Verde F, Fang L, et al. Chitotriosidase (CHIT1) is increased in microglia Appendix (continued) and macrophages in spinal cord of amyotrophic lateral sclerosis and cerebrospinal fluid levels correlate with disease severity and progression. J Neurol Neurosurg Psychiatry 2018;89:239–247. Name Location Contribution 6. Thompson AG, Gray E, Thézénas ML, et al. Cerebrospinal fluid macrophage biomarkers in amyotrophic lateral sclerosis. Ann Neurol 2018;83:258–268. Marta Sant Pau Biomedical Participated in manuscript 7. Illán-Gala I, Alcolea D, Montal V, et al. CSF sAPPβ, YKL-40, and NfL along the ALS- Querol- Research Institute, Hospital drafting and data FTD spectrum. Neurology 2018;91:e1619–e1628. Vilaseca, de la Santa Creu i Sant Pau, interpretation 8. Thompson AG, Gray E, Bampton A, Raciborska D, Talbot K, Turner MR. CSF MSc Barcelona, Spain chitinase proteins in amyotrophic lateral sclerosis. J Neurol Neurosurg Psychiatry 2019;90:1215–1220. Laia Muñoz, Sant Pau Biomedical Contributed to sample 9. Ling SC, Polymenidou M, Cleveland DW. Converging mechanisms in ALS and FTD: BSc Research Institute, Hospital preparation and disrupted RNA and protein homeostasis. Neuron 2013;79:416–438. de la Santa Creu i Sant Pau, participated in 10. Hodge RD, Bakken TE, Miller JA, et al. Conserved cell types with divergent features in Barcelona, Spain experimental procedures human versus mouse cortex. Nature 2019;573:61–68. 11. Lake BB, Chen S. Sos BC Integrative single-cell analysis of transcriptional and epi- Olivia Sant Pau Biomedical Participated in the revision genetic states in the human adult brain. Nat Biotechnol 2018;36:70–80. Belbin, PhD Research Institute, Hospital and editing of the 12. Mathys H, Davila-Velderrain J, Peng Z, et al. Single-cell transcriptomic analysis of de la Santa Creu i Sant Pau, manuscript and data Alzheimer’s disease. Nature 2019;570:332–337. Barcelona, Spain interpretation 13. WangX,ParkJ,SusztakK,ZhangNR,LiM.Bulktissuecelltypedeconvo- lution with multi-subject single-cell expression reference. Nat Commun 2019; Daniel Sant Pau Biomedical Participated in data 10:380. Alcolea, Research Institute, Hospital collection, analysis, and 14. Brooks BR, Miller RG, Swash M, Munsat TL; World Federation of Neurology Re- MD, PhD de la Santa Creu i Sant Pau, data interpretation search Group on Motor Neuron Diseases. El Escorial revisited: revised criteria for the Barcelona, Spain diagnosis of amyotrophic lateral sclerosis. Amyotroph Lateral Scler Other Motor Neuron Disord 2000;1:293–299. Laura Neurological Tissue Bank Contributed to sample 15. Garc´ıa-Redondo A, Dols-Icardo O, Rojas-Garc´ıa R, et al. Analysis of the C9orf72 gene Molina- of the Biobanc-Hospital preparation and in patients with amyotrophic lateral sclerosis in Spain and different populations Porcel, MD, Cl´ınic-IDIBAPS, Barcelona, participated in worldwide. Hum Mutat 2013;34:79–82. PhD Spain experimental procedures 16. Dols-Icardo O, Garc´ıa-Redondo A, Rojas-Garc´ıa R, et al. Analysis of known amyotrophic lateral sclerosis and frontotemporal dementia genes reveals a substantial ge- Jordi Sant Pau Biomedical Participated in data netic burden in patients manifesting both diseases not carrying the C9orf72 expansion Pegueroles, Research Institute, Hospital analysis mutation. J Neurol Neurosurg Psychiatry 2018;89:162–168. MSc de la Santa Creu i Sant Pau, 17. Dobin A, Davis CA, Schlesinger F, et al. STAR: ultrafast universal RNA-seq aligner. Barcelona, Spain Bioinformatics 2013;29:15–21. 18. McKenna A, Hanna M, Banks E, et al. The Genome Analysis Toolkit: a MapReduce Janina Sant Pau Biomedical Major role in the framework for analyzing next-generation DNA sequencing data. Genome Res 2010; Turon-Sans,´ Research Institute, Hospital acquisition of samples and 20:1297–1303. MD de la Santa Creu i Sant Pau, data and reviewed and 19. Liao Y, Smyth GK, Shi W. The R package Rsubread is easier, faster, cheaper and better Barcelona, Spain edited the manuscript for alignment and quantification of RNA sequencing reads. Nucleic Acids Res 2019; 47:e47. Rafael Sant Pau Biomedical Major role in the 20. PatroR,DuggalG,LoveMI,IrizarryRA,KingsfordC.Salmonprovidesfastand Blesa, MD, Research Institute, Hospital acquisition of samples and bias-aware quantification of transcript expression. Nat Methods 2017;14: PhD de la Santa Creu i Sant Pau, data and reviewed and 417–419. Barcelona, Spain edited the manuscript 21. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 2014;15:550. Alberto Sant Pau Biomedical Major role in the 22. Zhou Y, Zhou B, Pache L, et al. Metascape provides a biologist-oriented resource for Lleo,´ MD, Research Institute, Hospital acquisition of samples and the analysis of systems-level datasets. Nat Commun 2019;10:1523. PhD de la Santa Creu i Sant Pau, data and reviewed and 23. Koopmans F, van Nierop P, Andres-Alonso M, et al. SynGO: an evidence-based, Barcelona, Spain edited the manuscript expert-curated knowledge base for the synapse. Neuron 2019;103:217–234.e4. 24. Langfelder P, Horvath S. WGCNA: an R package for weighted correlation network Juan Fortea, Sant Pau Biomedical Major role in the analysis. BMC Bioinformatics 2008;9:559. MD, PhD Research Institute, Hospital acquisition of samples and 25. Prudencio M, Belzil VV, Batra R, et al. Distinct brain transcriptome profiles in de la Santa Creu i Sant Pau, data and reviewed and C9orf72-associated and sporadic ALS. Nat Neurosci 2015;18:1175–1182. Barcelona, Spain edited the manuscript 26. D’Erchia AM, Gallo A, Manzari C, et al. Massive transcriptome sequencing of human spinal cord tissues provides new insights into motor neuron degeneration in ALS. Sci Ricard Sant Pau Biomedical Major role in the Rep 2017;7:10046. Rojas- Research Institute, Hospital acquisition of samples and 27. Tam OH, Rozhkov NV, Shaw R, et al. Postmortem cortex samples identify distinct Garc´ıa, MD, de la Santa Creu i Sant Pau, data and reviewed and molecular subtypes of ALS: retrotransposon activation, oxidative stress, and activated PhD Barcelona, Spain edited the manuscript glia. Cell Rep 2019;29:1164–1177.e5. 28. Jara JH, Genç B, Stanford MJ, et al. Evidence for an early innate immune response in Jordi Sant Pau Biomedical Major role in funding the motor cortex of ALS. J Neuroinflammation 2017;14:129. Clarimon,´ Research Institute, Hospital acquisition; design and 29. Hopperton KE, Mohammad D, Trépanier MO, Giuliano V, Bazinet RP. Markers of PhD de la Santa Creu i Sant Pau, conceptualization of the microglia in post-mortem brain samples from patients with Alzheimer’s disease: Barcelona, Spain study; drafted the a systematic review. Mol Psychiatry 2018;23:177–198. manuscript; and analyzed 30. Spiller KJ, Restrepo CR, Khan T, et al. Microglia-mediated recovery from ALS- the data relevant motor neuron degeneration in a mouse model of TDP-43 proteinopathy. Nat Neurosci 2018;21:329–340. 31. SvahnAJ,DonEK,BadrockAP,etal.Nucleo-cytoplasmic transport of TDP- 43 studied in real time: impaired microglia function leads to axonal spreading References of TDP-43 in degenerating motor neurons. Acta Neuropathol 2018;136: 1. Brettschneider J, Del Tredici K, Toledo JB, et al. Stages of pTDP-43 pathology in 445–459. amyotrophic lateral sclerosis. Ann Neurol 2013;74:20–38. 32. Henstridge CM, Pickett E, Spires-Jones TL. Synaptic pathology: a shared mechanism 2. Liu J, Wang F. Role of neuroinflammation in amyotrophic lateral sclerosis: cellular in neurological disease. Ageing Res Rev 2016;28:72–84. mechanisms and therapeutic implications. Front Immunol 2017;8:1005. 33. Hong S, Beja-Glasser VF, Nfonoyim BM, et al. Complement and microglia mediate 3. Prinz M, Jung S, Priller J. Microglia biology: one century of evolving concepts. Cell early synapse loss in Alzheimer mouse models. Science 2016;352:712–716. 2019;179:292–311. 34. Dejanovic B, Huntley MA, De Mazière A, et al. Changes in the synaptic proteome in 4. Keren-Shaul H, Spinrad A, Weiner A, et al. A unique microglia type Associated with tauopathy and rescue of tau-induced synapse loss by C1q antibodies. Neuron 2018; restricting development of Alzheimer’s disease. Cell 2017;169:1276–1290. 100:1322–1336.e7.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 11 ARTICLE OPEN ACCESS CLASS OF EVIDENCE High κ free light chain is a potential biomarker for double seronegative and ocular myasthenia gravis

Adi Wilf-Yarkoni, MD, Yifat Alkalay, PhD, Talma Brenner, PhD, and Arnon Karni, MD, PhD Correspondence Dr. Karni Neurol Neuroimmunol Neuroinﬂamm 2020;7:e831. doi:10.1212/NXI.0000000000000831 [email protected]

Abstract MORE ONLINE Objective Class of Evidence To investigate the hypothesis that free light chain (FLC) sera levels could serve as a biomarker Criteria for rating for myasthenia gravis (MG), especially for the subgroups of seronegative MG and ocular MG. therapeutic and diagnostic studies Methods NPub.org/coe Sera from 73 patients with MG (20 seronegative for antiacetylcholine receptor [AChR] and anti–muscle-speciﬁc kinase and 53 positive for anti-AChR, which were clinically divided into 24 patients with ocular type, 45 with generalized type, and 4 with unequivocal clinical manifestation) and 49 healthy controls were studied for κ FLC and λ FLC levels with the Freelite human FLC kits.

Results The κ but not the λ levels of FLC were significantly increased in the patients with MG, including those with double seronegative MG and ocular MG, compared with the healthy controls. The specificity for double seronegative MG and ocular MG were both 98.0% when κ FLC was ≥25.0 mg/L. Increased κ FLC levels were not affected by the patient’s sex, age at MG onset, the presence of thymic pathology, or different treatments.

Conclusions Elevated serum κ FLC may serve as a biomarker for MG in suspected patients who are double seronegative and in those with only ocular manifestations when serology is inconclusive.

Classification of evidence This study provides Class III evidence that high κ FLC levels distinguished patients with MG, including those who were double seronegative, from healthy controls.

From the Neuroimmunology and Multiple Sclerosis Unit of the Neurology Division (A.W.-Y., A.K.), Tel Aviv Sourasky Medical Center; Clinical Immunology Laboratory (Y.A.), Tel Aviv Sourasky Medical Center; Laboratory of Neuroimmunology (T.B.), Department of Neurology, the Agnes Ginges Center for Human Neurogenetics, Hadassah-Hebrew University Medical Center, Jerusalem; Sackler Faculty of Medicine (A.K.), Tel Aviv University; and Sagol School of Neuroscience (A.K.), Tel Aviv University, Israel.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

The Article Processing Charge was funded by the Medical Research Infrastructure of TASMC. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary AChR = acetylcholine receptor; ANA = antinuclear antibody; ARR = absolute risk reduction; ENA = extracted nuclear antibody; FLC = free light chain; G-MG = generalized myasthenia gravis; HC = healthy control; IVIg = IV immunoglobulin; LRP4 = lipoprotein receptor–related protein 4; MG = myasthenia gravis; MuSK = muscle-speciﬁc kinase; NNT = number needed to treat; O-MG = ocular myasthenia gravis; ROC = receiver operating characteristic; SN-MG = serum-negative myasthenia gravis; SP-MG = serum-positive myasthenia gravis; TPO = thyroid peroxidase.

Myasthenia gravis (MG) is an antibody-mediated autoimmune Study design disease affecting the postsynaptic neuromuscular junctions of This is a case-control prospective study that compares the – striated skeletal muscles.1 3 The clinical manifestation includes levels of FLCs in the sera of patients with MG and of HCs. muscle weakness, which can be localized to ocular muscles (ocular MG [O-MG]) or distributed in extraocular muscles Patients and controls (generalized MG [G-MG]).2 The diagnosis of MG is confirmed One hundred twenty-eight potential donors who include 79 by the combination of symptoms, electrical physiologic studies consecutive patients with MG and 49 healthy volunteers were demonstrating neuromuscular junction dysfunction, and a pos- screened. Sixteen of the patients were excluded (because of the fl itive test for specificantibodies.4,5 Antibodies against acetyl- exclusion criteria that are listed below and in the ow diagram choline receptors (AChRs),6 muscle-specifickinase(MuSK),7 [supplementary data, links.lww.com/NXI/A281]). To increase and lipoprotein receptor–related protein 4 (LRP4)3,8 can be the number of subjects with SN-MG, we invited an additional found in about 90% of patients with MG (seropositive [SP]), 10 patients who were known to have SN-MG to participate in and about 10% remain with undetected specificautoantibody the study. Overall, blood samples were drawn between 2017 (seronegative [SN]).1 The diagnosis of MG may be obscure in and 2019 from 73 patients with MG who were referred to the SN patients. The failure in finding a specificantibodyforMG Neuroimmunology Unit at the Tel Aviv Sourasky Medical leaves a degree of insecurity in the diagnosis of SN-MG, and it is Center, Tel Aviv, Israel, and from 49 healthy individuals who recommended that serologic tests be repeated several months served as controls (HCs) (table). The diagnosis of MG was fi following negative test results.1 AbiomarkerforMGinthese de ned by clinical and supportive features of neurophysiology fi patients may therefore add confidence in the diagnosis of MG. tests of single- ber EMG and/or serology of AChR antibodies or anti-MuSK antibodies. The patients underwent a chest CT The production of antibodies is generally also accompanied by scan or a chest MRI scan, and those with radiologic evidence of the synthesis of immunoglobulin light chains. The circulating thymus enlargement or a suspected thymoma underwent levels of light chains may be increased in conditions of excess thymectomy. immunoglobulin production, as in antibody-mediated diseases and in renal failure.9,10 Recent studies have demonstrated that The distribution of patients according to the Myasthenia fi this overproduction of light chains has a biological and im- Gravis Foundation of America clinical classi cation was Class munologic role.9 An increase in free light chain (FLC) pro- I = 24 patients, Class IIA = 20 patients, Class IIB = 4 patients, – duction has been reported in several autoimmune diseases.11 17 Class IIIA = 9 patients, Class IIIB = 6 patients, Class IVA = 4 patients, and Class IVB = 2 patients. The classification of 4 To the best of our knowledge, only 1 study has evaluated FLC patients was equivocal. Class I, which is ocular MG (O-MG), levels in MG, and it reported an elevation of both κ FLC and λ was diagnosed in patients with pure ocular involvement last- FLC in their 34 study patients.18 We hypothesized that κ FLC ing over 2 years because 90% of them will be presumed to 19 and λ FLC levels may be biomarkers for MG, especially for remain with ocular MG throughout their illness. The disease SN-MG for which diagnosis can be difficult. Therefore, we duration among these patients was between 2 and 3 years in 2 studied the κ FLC and λ FLC levels in patients with MG, patients, between 3 and 5 years in 6 patients, and above 5 including those with SN-MG, and in healthy controls (HCs). years in 16 patients. We also analyzed the results according to various clinical fl forms of the disease in a large number of patients with MG. The diagnosis of patients with SN-MG was based on uc- tuating muscular weakness in a typical distribution for MG with supportive positive diagnostic tests of single-fiber EMG Methods and/or repetitive stimulation EMG and/or positive edrophonium tests in addition to ruling out other relevant Standard protocol approvals, registrations, medical conditions that may present with the same type of and patient consent weakness (such as thyroid ophthalmopathy, congenital The study protocol was approved by the Tel Aviv Sourasky myasthenia, Lambert-Eaton myasthenic syndrome, myositis, Medical Center Institutional Review Board for human experi- mitochondrial syndromes, amyotrophic lateral sclerosis, and ments (Helsinki Committee, No. 0702-15). All the participants CNS demyelinating diseases) by the clinical presentation signed written informed consent. and course, relevant blood tests, EMG, and brain MRI in

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Table Demographic and clinical characteristics of the patients

N Male Female Age at onset, y

Ocular MGa 24 16 8 62.4 ± 16.7

Generalized MGa 45 25 20 56.1 ± 16.8

Early-onset MG (age ≤50 y) 21 11 10 38.1 ± 10.8

Late-onset MG 52 31 21 66.8 ± 10.8

Seropositive for AChR 53 34 19 57.8 ± 17.2

Double seronegative (anti-AChR and anti-MuSK) 20 8 12 60.7 ± 16.5

Thymus pathology 12 5 7 40.2 ± 14.2

Total MG 73 42 31 58.6 ± 16.9

Healthy controls 49 21 28 43.4 ± 12.9b

Abbreviations: AChR = acetylcholine receptor; MG = myasthenia gravis; MuSK = muscle-specific kinase. a The clinical manifestation was unequivocal for 4 patients. b Age at sampling.

patients who had negative tests for anti-AChR and anti- Primary research questions MuSK antibodies. Patients were treated according to the Can κ FLC and λ FLC levels be biomarkers for patients with accepted guidelines. MG and for patients with double SN-MG.

The exclusion criteria were a diagnosis of plasma cell dys- Statistical analysis crasia disorders, a diagnosis or suspicion of another chronic Data for group comparisons are presented as mean ± SD and inflammatory condition known to be associated with in- with 95% confidence intervals (CIs). Statistical analysis and the creased serum levels of FLC (such as rheumatoid arthritis, graphical representation of the data were performed with systemic sclerosis, Sjögren syndrome, and systemic lupus GraphPad Prism (version 8.2.1; GraphPad Software, San Diego, erythematosus),10,14 a glomerular filtration rate of <60 mL/ CA), SPSS (version 22; SPSS Inc., Chicago, IL). Comparisons min/1.73 m2, and treatment with plasmapheresis for the past between 2 groups were performed using the Student unpaired t 3 months. We recruited healthy individuals to serve as the test with Welch’s correction when the variances were unequal. control group to test whether FLC levels are high in the sera The Welch analysis of variance test was applied to compare of patients with MG. The data collection was planned before multiple group differences for unequal variance, and the Dun- the tests for FLCs were performed. nett T3 multiple comparisons test was used to compare within- group differences. A p value of <0.05 was considered statistically Laboratory testing significant. A multivariate analysis was performed to evaluate Blood serum samples were separated by standard centrifu- possible confounding effects. A receiver operating characteristic gation, divided into aliquots, and stored frozen until analysis. (ROC) curve analysis for calculating the area under the ROC The samples were then assayed in a blinded fashion. FLC curve was used to determine the diagnostic accuracy of FLC serum levels were assessed with the Freelite Human κ and λ values as diagnostic markers for MG. The number needed to Free Kits (The Binding Site Ltd., Birmingham, United treat (NNT) and absolute risk reduction (ARR) were added Kingdom) on a SPAPLUS instrument (The Binding Site when an estimation of cutoff levels of FLC was analyzed. The Ltd.). Anti-AChR Ab was tested by ELISA (EUROIMMUN Pearson correlation coefficient was applied to test correlations AG, Luebeck, Germany). Anti-MuSK Ab was assessed by between anti-AChR and FLC levels. RIA with the MuSK kit (RSR Ltd., Cardiff,UnitedKing- dom). Antinuclear antibody (ANA), antimitochondrial, and Data availability antismooth muscle were tested by immunofluorescence Anonymized data will be shared by request from any qualified (AESKU GROUP, Wendelsheim, Germany). Extracted investigator. nuclear antibodies (ENAs) including antibodies against SSA, SSB, RNP SM, JO1, and SCL70, as well as antibodies against histone, dsDNA, thyroglobulin, and thyroid peroxidase Results (TPO), were tested by ELISA (Alegria by ORGENTEC Diagnostika GmbH, Mainz, Germany). Rheumatoid factor The 122 participants in this study included 73 patients with was assessed by a nephelometric BNtmII System by Siemens MG (42 men and 31 women), with a mean age of (mean ± SD) Medical Solutions, Malvern, PA. 59.1 ± 14.2 years, and 49 HCs (21 men and 28 women) with

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 a mean age of 43.4 ± 12.8 years. The patients’ clinical mani- between the levels of anti-AChR and κ FLC (r = −0.059) or λ festations, serologic results, sex, and age at MG onset are FLC (r = −0.044). summarized in table. There were 53 patients with detectable anti-AChR antibodies in the serum and 20 patients with neg- FLC levels in patients with MG according to ative results for anti-AChR and anti-MuSK. clinical characteristics We compared the FLC levels in MG subgroups according to FLC levels in patients with MG and according to clinical manifestations, age at MG onset, and thymus pathology. serologic status Figure 2, A and B shows the κ and λ FLC levels in the O-MG, The κ FLC serum levels were higher in the patients with MG G-MG, and HC groups. The κ FLC but not the λ FLC levels compared with the HCs (23.7 ± 8.3 mg/L, CI = 21.8–25.68 differed significantly among the 3 groups (p < 0.0001). A mg/L and 15.9 ± 5.0 mg/L, CI = 14.5–17.3 mg/L, respectively, multiple comparisons test revealed an increase in κ FLC in the p <0.0001,figure 1A). However, there were no significant O-MG group (n = 24, 23.0 ± 6.2 mg/L, CI = 20.0–25.0 mg/L) differences in their λ FLC serum levels (15.9 ± 5.7 mg/L, CI = compared with the HC group (15.9 ± 5.0 mg/L, CI = 14.5–17.3 14.6–17.2 mg/L and 14.5 ± 4.6 mg/L, CI = 13.2–15.8 mg/L, mg/L, figure 2A). There was no difference between the O-MG respectively, p =N.S.,figure 1B). Moreover, there was a sig- and G-MG groups (23.8 ± 9.2, CI = 21.1–26.5 mg/L). nificant increase in the κ/λ ratio in the patients with MG compared with the HCs (1.5 ± 0.41, CI = 1.4–1.6 and 1.1 ± 0.2, Given that the diagnosis of ocular MG may often be chal- CI = 1.0–1.2, respectively, p <0.0001,figure 1C). These in- lenging, there is a need for a biomarker for ocular MG. The creased κ FLC levels in the patients with MG compared with ROC analysis for κ FLC in patients with ocular MG had an area the HCs were confirmed after correction for age and sex in under the curve of 0.81, 95% CI = 0.71–0.91 (figure 2C). For a multivariant analysis (F =16.34,p <0.001). example, the calculated specificity of κ FLC ≥25 mg/L of patients with O-MG compared with the HCs was 98.0%, with One of the main objectives of this study was to test the FLC a sensitivity of 37.5%, NNT = 2.7, and ARR = −0.375. There serum levels in patients who tested negative for antibodies was an increase in the κ/λ ratio between O-MG (1.5 ± 0.5, CI = against both AChR and MuSK (double SN-MG). The κ but 1.3–1.7) and G-MG (1.5 ± 0.4, CI = 1.4–1.6) groups and the not the λ FLC levels differed significantly among the 3 groups HCs (1.1 ± 0.2, CI = 1.0–1.2). (p < 0.0001). A multiple comparisons test revealed a significant increase in κ FLC in the patients with double SN-MG (n = 20, Early-onset MG was defined as onset at ≤50 years of age, and 26.8 ± 12.0 mg/L, CI = 21.5–32.0 mg/L) compared with the late-onset MG was defined as onset >50 years of age.5 There HCs (p < 0.002, figure 1D). There was no significant difference was also no difference in the κ FLC and λ FLC levels between in κ FLC levels between double SN patients and SP patients for patients with early-onset MG (n = 21, 20.9 ± 6.5 mg/L, CI = anti-AChR (n = 53, 22.5 ± 6.2 mg/L, CI = 20.8–24.2 mg/L, p = 18.1–23.6 mg/L and 14.7 ± 4.5 mg/L, CI = 12.7–16.6 mg/L, N.S.). The κ/λ ratio, however, differed significantly among the respectively) and patients with late-onset MG (n = 52, 24.8 ± 3groups(p < 0.0001). There was a significant increase in the 8.8 mg/L, CI = 22.4–27.2 mg/L and 16.4 ± 6.2 mg/L, CI = κ/λ ratio in the patients with SN-MG (1.6 ± 0.5, CI = 1.4–1.9) 14.8–18.1 mg/L, respectively) (figure 2, C and D). Twelve compared with the HCs (p < 0.001, figure 1F). None of the patients with MG had an abnormal thymus, including 2 study patients were found to have anti-MuSK antibodies. patients with thymoma and 10 patients with thymus hyper- plasia. An analysis of those patients revealed no difference in the The difficulties in diagnosing MG in patients who have neg- κ FLC and λ FLC levels between patients with thymic pa- ative serology tests for myasthenia-specific autoantibodies thology (23.8 ± 8.2 mg/L, CI = 19.2–28.4 and 16.8 ± 4.9 mg/L, establish a clear-cut need for a biomarker(s) for this patient CI = 14.0–19.5 mg/L, respectively) and those patients without subgroup. A ROC analysis for κ FLC in patients with double (22.6 ± 6.6 mg/L, CI = 20.7–24.4 and 15.6 ± 4.7 mg/L, CI = SN-MG has an area under the curve of 0.79, CI = 0.65–0.91, 14.3–16.9 mg/L, respectively, figure 2, E and F). Twenty-eight p = 0.0002 (figure 1G). For example, the calculated specificity of the patients were being regularly treated by IV immuno- of κ FLC ≥25 mg/L of patients with double SN-MG com- globulin (IVIg) at a dose of 0.4 g/kg every 3–6weeks.Ablood pared with HCs was 98.0%, with a sensitivity of 45.0%, sample from these patients was always taken just before the NNT = 2.2, and ARR = −0.45. next IVIg treatment. There was no difference in the κ FLC and λ FLC levels between the patients with IVIg-treated MG (22.7 The sera of all patients with double SN-MG were tested for ± 7.8 mg/L, CI = 19.8–25.6 and 16.4 ± 5.3 mg/L, CI = ANA, ENA (that includes antibodies against SSA, SSB, RNP, 14.4–18.3 mg/L, respectively) and the patients with non-IVIg– SM, JO1, and SCL70), antibodies against histone, dsDNA, treated MG (23.0 ± 6.1 mg/L, CI = 21.0–25.1 and 15.7 ± 4.2 thyroglobulin, TPO, mitochondrial, smooth muscle, and mg/L, CI = 14.3–17.1 mg/L, respectively). There were no rheumatoid factor. We found that 1 patient was positive for differences in κ FLC and λ FLC levels between patients who ANA (homogenous pattern), anti-dsDNA, and anti-histone (κ were treated with pyridostigmine alone (n = 22, 22.4 ± 6.6 mg/ FLC = 15.5 mg/L and λ FLC = 11.4 mg/L). Another patient L, CI = 19.6–25.2 and 15.5 ± 3.9 mg/L, CI = 14.0–17.1 mg/L, with double SN-MG was positive for TPO (κ FLC = 16.9 mg/ respectively) and those who were also treated with prednisone Landλ FLC = 11.7 mg/L). We did not find any correlation and/or azathioprine but not IVIg (n = 22, 22.6 ± 3.3 mg/L,

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Figure 1 FLC levels in patients with MG and according to serologic status

Free light chain (FLC) levels in the sera of patients with seropositive myasthenia gravis (SP-MG) and patients with double seronegative MG (SN-MG). Serum κ (A) and λ (B) FLC levels and the κ/λ ratio (C) in patients with MG and healthy controls (HCs). Serum κ (D) and λ (E) FLC levels and the κ/λ ratio (F) in patients with double SN-MG, in patients with SP-MG for antiacetylcholine receptor antibody, and in HCs. A receiver operating characteristic curve for κ FLC in patients with double SN-MG (G). Significant p values for the t test and Dunnett T3 multiple comparisons test are displayed: *p < 0.01; and **p < 0.0001.

CI = 21.2–23.9 mg/L and 14.3 ± 4.4 mg/L, CI = 12.5–16.2 become an area of growing interest. A number of studies mg/L, respectively) and those who were treated with IVIg. One have demonstrated that FLCs are not a waste of immu- patient was treated with rituximab. noglobulin synthesis but rather active molecules with immunologic properties. FLCs have been shown to Consistently high κ FLC levels participate in several key processes of immune responses. To assess the reliability of FLC measurements, serum was They are necessary to adjust polymorphonuclear function collected once more from 28 patients with MG at an average and stimulation,20 they cause mast cell degranulation that interval of 6 months. The statistical analysis with a paired t test releases proinflammatory mediators,21 and they may have revealed no significant difference between the 2 samples (22.3 direct toxicity in the kidney.9 Overproduction of FLC may ± 6.8 mg/L, CI = 19.8–24.8 mg/L vs 21.4 ± 7.4 mg/L, CI = occur after an excess of antibody production by B cells, 18.6–24.2 mg/L). The mean ± SD of percentage of change usually as a result of chronic immune stimulation. There- between the sample pairs was 16.0% ± 11.9%, CI = 11.6–20.4. fore, the measurement of FLC has been proposed as a biomarker of B-cell activity.10 Discussion Little is known about FLC in MG. The main result of our Thesearchforthebiologicalroleandpossibleclinicalap- current study is the finding that patients with MG have plication of FLC in various autoimmune diseases has higher levels of serum κ FLC but not of λ FLC compared

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 5 Figure 2 FLC levels according to MG subgroups

Serum κ (A) and λ (B) free light chain (FLC) levels in patients with pure ocular myasthenia gravis (O-MG) and patients with generalized myasthenia gravis (G- MG). A receiver operating characteristic curve for κ FLC in patients with O-MG (C). Serum κ (D) and λ (E) FLC levels in early-onset myasthenia gravis (EO-MG, age ≤50 years) and late-onset myasthenia gravis (LO-MG, age >50 years). Serum κ (F) and λ (G) FLC levels in MG patients with thymus pathology and normal thymus. HC = healthy control.

with HCs. Furthermore, and even more important, we ob- thesepatientswerelow,andthereforedidnotcontributeto served that the levels of κ FLC are high in patients with the overall increase in κ FLC levels in this patient group, we double SN-MG compared with HCs. This is a heretofore did not exclude them from the analysis. unreported immunologic finding of double SN-MG. The differential diagnosis of patients with suspected SN-MG is It is important to emphasize that being a biomarker does not according to the patient’sspecific clinical expression, and the make κ FLC a diagnostic test for MG, as the antibodies against diagnostic investigation is by verification of the clinical postsynaptic antigen of the neuromuscular junction, but at course and manifestations, blood tests, electromyographic a high value and under the above-mentioned clinical con- tests, and imaging studies. For patients who were not di- ditions, it supports the diagnosis of MG. This finding suggests agnosed with another disease, a high level of κ FLC and of overactivity of plasma cells in this patient subgroup. It is the κ/λ ratio may serve as a supportive biomarker for MG, conceivable, therefore, that there is a postsynaptic-directed when there is no comorbidity with a known medical con- antibody production that has not yet been identified, or that dition that is associated with high FLC levels. Also, none of the current methods are not sensitive enough to detect it in our patients with double SN-MG were diagnosed with any this subgroup of patients with double SN-MG.22 other diseases known to have high FLC levels, 2 of our patients with double SN-MG tested positive to autoanti- Clinical presentation, age at MG onset, and the presence of bodies (1 had positive ANA, anti-dsDNA, and anti-histone, thymus pathology did not affect the κ FLC levels that were high and the other had positive TPO). Because the FLC levels in in both O-MG and G-MG patients as well as among the early-

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN onset and late-onset MG patients. O-MG is often part of the Our study has several limitations. The number of SN differential diagnosis of disorders of the brainstem, ocular patients is relatively small because only about 10% of motor nerves, and eye muscles, and the confirmatory tests of patients with MG are SN. We intentionally recruited a higher edrophonium, serum AChR antibodies, and electromyographic proportion of SN patients who comprised 27.4% of the evaluations may fail to support the diagnosis of O-MG.23 Our studied MG group to be able to test the hypothesis of in- results suggest that high levels of κ FLC can serve as a bio- creased FLC levels in this patient subgroup. Indeed, we marker to support the diagnosis of O-MG in these cases. found significantly increased levels of κ FLC in patients with double SN-MG. A larger group of SN subjects would have Twenty-eight of our patients with MG were treated with IVIg, allowed stronger support for these results. Anti-LRP4 anti- but we did not find any effect of this therapy on their FLC levels body and anticlustered AChR antibody were not evaluated, when their blood sample was drawn just before the next sched- and anti-MuSK was not detected in our patients. Therefore, uled treatment. This observation can be explained by the fact that we could not study the levels of FLC in patients who are although IVIg solutions contain FLC, their half-life is only a few positive to one or more of these autoantibodies or in patients hours. There were also no differences in κ FLC levels in these who are also negative for anti-LRP4 and anticlustered AChR patients and the patients who were treated with pyridostigmine antibodies. Because these tests are not available in many alone or with immunosuppressive agents without IVIg. Our centers, we believe that the relatively simple test for de- results, therefore, suggest no treatment effect on κ FLC levels. termining FLC levels can support the diagnosis of MG in such situations. Our control group comprised healthy indi- Our results of elevated κ FLC levels in patients with MG are in viduals and not patients with other diseases that can mimic line with recent observations in 34 serum samples from 17 MG; therefore, our conclusions are limited to cases that AChR-positive and 13 MuSK-positive patients.18 However, those mimickers were ruled out. those authors found a significant increase in free λ chain levels as well. A review of their data reveals that the patients’ free λ chain Elevation in κ FLC levels is not specific for MG, and it can be levels were higher than those of the controls but were within the found in hematologic malignancies and in other autoimmune normal range, whereas the κ/λ ratio was significantly higher in diseases. Therefore, κ FLC can serve as a biomarker only in both MG groups, which indicates substantial increases in free κ patients whose clinical manifestations suggest the possibility of FLC. The explanation for increased κ FLC levels in autoim- MG and who do not have any other disorder known to be mune diseases is not clear. This finding could possibly be associated with increased κ FLC levels. In conditions in which explained by the dominance of κ FLC in humans because the κ the only diagnostic support for MG is a high κ FLC level, and chain is rearranged first during IgG production and is quanti- after ruling out other disorders that are associated with high κ tatively more common.16,24 Furthermore, λ FLC is usually di- FLC levels, it would be appropriate to consider performing meric in form, whereas κ FLC is generally monomeric but can hematologic and immunologic tests for these disorder markers exist as noncovalently linked dimers.25 In addition, the 2 forms to rule them out. Because we did not assess the association of FLC probably serve as different modulators of the immu- between FLC levels with disease activity, severity, and response nologic response because λ FLCsaremoreeffective silencers of to therapy, we cannot comment on the use of FLC as a bio- autoreactive B cells,26 which may explain why the κ FLC level is marker for MG activity, as reported in other autoimmune more elevated in classic autoimmune diseases, such as MG. diseases,9,10,14,25 or for the possibility of its being a marker of response to therapy. Further investigation is needed to study Ahighκ FLC level was also shown to be associated and to have these questions. prognostic value in several autoimmune conditions. Specifi- cally, serum κ chains and κ/λ ratios were significantly higher in In conclusion, our data support the use of κ FLC as a bio- systemic sclerosis and correlated with disease activity.11 marker for MG, especially in patients with double SN-MG Moreover, κ chains were significantly and strongly correlated and in patients suspected of having O-MG for which a high κ with disease activity in Sjögren syndrome, and patients with FLC level seems to be specific for MG. systemic lupus erythematosus demonstrated elevated concentrations of serum FLC (mainly κ chains), with a strong inverse Acknowledgment correlation having been found between the κ chain levels and The authors thank Ms. Esther Eshkol, the institutional the C3 levels.11,12 CSF κ FLC levels have a high sensitivity and medical copyeditor, for editing the manuscript for non- specificity for the diagnosis of MS, and they can predict con- intellectual content. – version from a clinically isolated syndrome to MS.15 17 Study funding The diagnosis of MG may be challenging, especially in patients The study was funded in part by The Binding Site Group Ltd. with no detectable antibodies in their serum. To the best of our and by Medison Pharmaceuticals Ltd. knowledge, this is the first description of elevated κ FLC levels in patients with double SN-MG. The measurement of κ FLC is Disclosure reliable, inexpensive, and rapid and therefore readily available The authors report no disclosures relevant to the manuscript. for supporting the possible diagnosis of these patients. Go to Neurology.org/NN for full disclosures.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 7 Publication history 8. Zisimopoulou P, Evangelakou P, Tzartos J, et al. A comprehensive analysis of the fl epidemiology and clinical characteristics of anti-LRP4 in myasthenia gravis. Received by Neurology: Neuroimmunology & Neuroin ammation J Autoimmun 2014;52:139–145. February 29, 2020. Accepted in final form May 29, 2020. 9. Esparvarinha M, Nickho H, Mohammadi H, Aghebati-Maleki L, Abdolalizadeh J, Majidi J. The role of free kappa and lambda light chains in the pathogenesis and treatment of inflammatory diseases. Biomed Pharmacother 2017;91:632–644. 10. Napodano C, Pocino K, Rigante D, et al. Free light chains and autoimmunity. Appendix Authors Autoimmun Rev 2019;18:484–492. 11. Bosello S, Basile U, De Lorenzis E, et al. Free light chains of immunoglobulins in Name Location Contribution patients with systemic sclerosis: correlations with lung involvement and inflammatory milieu. J Clin Pathol 2018;71:620–625. Adi Wilf- Rabin Medical Center Analysis or interpretation of the 12. Chiche L, Cournac JM, Mancini J, et al. Normalization of serum-free light chains in Yarkoni, Israel, Petach Tikva, Israel data and drafting or revising patients with systemic lupus erythematosus upon rituximab treatment and correlation MD the manuscript for intellectual with biological disease activity. Clin Rheumatol 2011;30:685–689. content 13. Gottenberg JE, Aucouturier F, Goetz J, et al. Serum immunoglobulin free light chain assessment in rheumatoid arthritis and primary Sjogren’s syndrome. Ann Rheum Dis Yifat Tel Aviv Sourasky Medical Major role in the acquisition of 2007;66:23–27. Alkalay, Center, Israel data 14. Gulli F, Napodano C, Marino M, et al. Serum immunoglobulin free light chain levels PhD in systemic autoimmune rheumatic diseases. Clin Exp Immunol 2020;199:163–171. 15. Hassan-Smith G, Durant L, Tsentemeidou A, et al. High sensitivity and specificity of Talma Hadassah-Hebrew Major role in the acquisition of elevated cerebrospinal fluid kappa free light chains in suspected multiple sclerosis. Brenner, University Medical Center, data J Neuroimmunol 2014;276:175–179. PhD Jerusalem, Israel 16. Ramsden DB. Multiple sclerosis: assay of free immunoglobulin light chains. Ann Clin Biochem 2017;54:5–13. Arnon Tel Aviv Sourasky Medical Design or conceptualization of 17. Senel M, Mojib-Yezdani F, Braisch U, et al. CSF free light chains as a marker of Karni, Center, Israel the study and drafting or intrathecal immunoglobulin synthesis in multiple sclerosis: a blood-CSF barrier re- MD, PhD revising the manuscript for lated evaluation in a large cohort. Front Immunol 2019;10:641. intellectual content 18. Basile U, Marino M, Napodano C, et al. Serological immunoglobulin-free light chain profile in myasthenia gravis patients. J Immunol Res 2018;2018:9646209. 19. Kerty E, Elsais A, Argov Z, Evoli A, Gilhus NE. EFNS/ENS guidelines for the treatment of ocular myasthenia. Eur J Neurol 2014;21:687–693. References 20. Braber S, Thio M, Blokhuis BR, et al. An association between neutrophils and im- 1. Gilhus NE. Myasthenia gravis. N Engl J Med 2017;376:e25. munoglobulin free light chains in the pathogenesis of chronic obstructive pulmonary 2. Gilhus NE, Verschuuren JJ. Myasthenia gravis: subgroup classification and therapeutic disease. Am J Respir Crit Care Med 2012;185:817–824. strategies. Lancet Neurol 2015;14:1023–1036. 21. Redegeld FA, van der Heijden MW, Kool M, et al. Immunoglobulin-free light chains 3. Berrih-Aknin S, Le Panse R. Myasthenia gravis: a comprehensive review of immune elicit immediate hypersensitivity-like responses. Nat Med 2002;8:694–701. dysregulation and etiological mechanisms. J Autoimmun 2014;52:90–100. 22. Devic P, Petiot P, Simonet T, et al. Antibodies to clustered acetylcholine receptor: 4. Zisimopoulou P, Brenner T, Trakas N, Tzartos SJ. Serological diagnostics in myas- expanding the phenotype. Eur J Neurol 2014;21:130–134. thenia gravis based on novel assays and recently identified antigens. Autoimmun Rev 23. Kusner LL, Puwanant A, Kaminski HJ. Ocular myasthenia: diagnosis, treatment, and 2013;12:924–930. pathogenesis. Neurologist 2006;12:231–239. 5. Berrih-Aknin S, Frenkian-Cuvelier M, Eymard B. Diagnostic and clinical classification 24. van der Heijden M, Kraneveld A, Redegeld F. Free immunoglobulin light chains as target of autoimmune myasthenia gravis. J Autoimmun 2014;48–49:143–148. in the treatment of chronic inflammatory diseases. Eur J Pharmacol 2006;533:319–326. 6. Vincent A, Newsom-Davis J. Acetylcholine receptor antibody as a diagnostic test for 25. Boivin D, Provencal M, Gendron S, et al. Purification and characterization of a stim- myasthenia gravis: results in 153 validated cases and 2967 diagnostic assays. J Neurol ulator of plasmin generation from the antiangiogenic agent Neovastat: identification Neurosurg Psychiatry 1985;48:1246–1252. as immunoglobulin kappa light chain. Arch Biochem Biophys 2004;431:197–206. 7. Evoli A, Tonali PA, Padua L, et al. Clinical correlates with anti-MuSK antibodies in 26. Wardemann H, Hammersen J, Nussenzweig MC. Human autoantibody silencing by generalized seronegative myasthenia gravis. Brain 2003;126:2304–2311. immunoglobulin light chains. J Exp Med 2004;200:191–199.

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN ARTICLE OPEN ACCESS COVID-19 outcomes in MS Observational study of early experience from NYU Multiple Sclerosis Comprehensive Care Center

Erica Parrotta, DO, Ilya Kister, MD, Leigh Charvet, PhD, Carrie Sammarco, DNP, Valerie Saha, NP, Correspondence Robert Erik Charlson, MD, Jonathan Howard, MD, Josef Maxwell Gutman, MD, Malcolm Gottesman, MD, Dr. Zhovtis Ryerson lana.zhovtisryerson@ Nada Abou-Fayssal, MD, Robyn Wolintz, MD, Marshall Keilson, MD, Cristina Fernandez-Carbonell, MD, nyulangone.org Lauren B. Krupp, MD, and Lana Zhovtis Ryerson, MD

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e835. doi:10.1212/NXI.0000000000000835

Abstract MORE ONLINE

Objective COVID-19 Resources To report outcomes on patients with multiple sclerosis (MS) and related disorders with For the latest articles, coronavirus disease 2019 (COVID-19) illness. invited commentaries, and blogs from physicians Methods around the world From March 16 to April 30, 2020, patients with MS or related disorders at NYU Langone MS NPub.org/COVID19 Comprehensive Care Center were identiﬁed with laboratory-conﬁrmed or suspected COVID- 19. The diagnosis was established using a standardized questionnaire or by review of in-patient hospital records.

Results We identiﬁed 76 patients (55 with relapsing MS, of which 9 had pediatric onset; 17 with progressive MS; and 4 with related disorders). Thirty-seven underwent PCR testing and were conﬁrmed positive. Of the entire group, 64 (84%) patients were on disease-modifying therapy (DMT) including anti-CD20 therapies (n = 34, 44.7%) and sphingosine-1-phosphate receptor modulators (n = 10, 13.5%). The most common COVID-19 symptoms were fever and cough, but 21.1% of patients had neurologic symptom recrudescence preceding or coinciding with the infection. A total of 18 (23.7%) were hospitalized; 8 (10.5%) had COVID-19 critical illness or related death. Features more common among those hospitalized or with critical illness or death were older age, presence of comorbidities, progressive disease, and a nonambulatory status. No DMT class was associated with an increased risk of hospitalization or fatal outcome.

Conclusions Most patients with MS with COVID-19 do not require hospitalization despite being on DMTs. Factors associated with critical illness were similar to the general at-risk patient population. DMT use did not emerge as a predictor of poor COVID-19 outcome in this preliminary sample.

From the NYU Langone Multiple Sclerosis Comprehensive Care Centers (E.P., I.K., L.C., C.S., V.S., R.E.C., J.H., J.M.G., M.G., N.A.-F., R.W., M.K., L.B.K., L.Z.R.), New York, NY; and Cohen’s Children Medical Center Northwell Health (C.F.-C.), Lake Success, NY.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary CAD = coronary artery disease; COVID-19 = coronavirus disease 2019; DMT = disease-modifying therapy; ICU = intensive care unit.

At present, we do not know whether multiple sclerosis Data availability (MS) or disease-modifying therapies (DMTs) for MS in- Anonymized data can be made available on request for research crease the risk of acquiring coronavirus disease 2019 purposes by submitting a request to the corresponding author. No (COVID-19) or worsen the course (hospitalizations, in- deidentified patient data or study-related documents will be shared. tensive care unit [ICU], and death). DMT medications have immunosuppressive effects that could hamper mounting an effective immune response to the infection.1 Results On the other hand, immunosuppression could offer pro- A total of 76 patients met the inclusion criteria, 72 (93%) tection by downregulating hyperinflammation and the cy- with MS, and 4 (7%) with related disorders (neuromyelitis tokine storm associated with COVID-19.2 optica spectrum disorder, chronic relapsing inflammatory optic neuropathy, neurosarcoidosis, and myelin oligodendrocyte New York City emerged as the epicenter of the COVID-19 glycoprotein-immunoglobulin G–associated disorder). The pandemic in the United States in March 2020. Given the average age of the full sample was 44.9 ± 15.2 years (range widespread prevalence of COVID-19 in our community, 13–71 years), and 61.8% were female. The disease duration was clinicians at the NYU Multiple Sclerosis Comprehensive Care 15.2 ± 10.7 years (range 1–52 years). Racial breakdown was 50 Center (MSCCC) received numerous reports of Severe Acute (65.8%) white, 21 (27.6%) black, 3 (3.9%) Asian, 1 (1.3%) Respiratory Syndrome Coronavirus 2 (SARS-COV2) infection Pacific Islander, and 1 (1.3%) other. Hispanic ethnicity was from patients and began systematically collecting symptom reported by 15 (19.7%). Of the 72 patients with MS, 55 data and the clinical course. This timely, real-world observa- (76.4%) had a relapsing-remitting subtype and 17 (23.6%) tional study on outcomes of COVID-19 in actively treated primary or secondary progressive subtypes. Sixty-five patients patients with MS help inform clinicians as they counsel patients (84%) were on DMT. One patient resided in a nursing facility. with MS and guide treatment decisions during the pandemic. He had an uncomplicated course.

Methods Common symptoms reported included fever (68.4%), cough (68.4%), fatigue (38.2%), shortness of breath (31.6%), and For this observational study, demographic and clinical fea- myalgias/arthralgias (26.3%). Other frequent symptoms in- tures were collected on patients currently followed at cluded anosmia (22%), ageusia (19.7%), and headache MSCCC and its 4 affiliated sites in the greater New York area (21.1%). A subset reported neurologic symptom recrudescence (2 in Long Island and 2 in Brooklyn) with a history of (21.1%) suggestive of relapse. In some cases, neurologic COVID-19 from March 16 through April 30, 2020. All symptoms preceded viral symptoms by several days. patients who contacted the center with infectious symptoms or were seen during routine teleneurology visits were queried Of the 84% of patients on DMTs, 44.7% were treated with anti- regarding COVID-19 exposure using a standardized in- CD20 therapies (rituximab n = 18; ocrelizumab n = 16), 13.5% strument. Inclusion criteria were any patient with MS or re- on sphingosine-1-phosphate (S1P) modulators (fingolimod n lated disorders who was diagnosed with COVID-19 by = 8; siponimod n = 2), 7.9% (n = 6) on glatiramer acetate, 5.3% a health care provider (based on symptoms, course, radio- (n = 4) each on natalizumab and dimethyl fumarate, and 3.9% graphic findings consistent with CDC COVID-19 criteria,3 (n = 3) on beta-interferons. Table 1 summarizes outcomes by and/or positive SARS-COV2 PCR when available). For DMT class. There were no observed differences between DMT hospitalized patients, in-patient records were reviewed. NYU use among those who were and were not hospitalized or be- School of Medicine Institutional Review Board approval was tween those specifically treated with anti-CD20 therapies (43% obtained for the study. vs 50%, OR = 0.76 [CI 0.26–2.18]).

Descriptive statistics were used to summarize the de- As shown in table 2, of the full sample, 18 (23.7%) were hospi- mographic and clinical characteristics of patient. Continuous talized. The hospitalized vs nonhospitalized patients were more variables were described in terms of means and SDs, and likely to be older (CI 1.22–17.11, p = 0.03), have progressive MS categorical variables were summarized as counts and per- subtype (OR 4.11 [CI 1.21–13.97], p = 0.04), required ambula- centages. No imputation was made for missing data. Diﬀer- tory assistance or were nonambulatory (OR 4.27 [CI 1.38–13.27], ences between hospitalized vs nonhospitalized patients were p = 0.01), and have comorbid obesity (OR 6.25 [CI 1.90–20.50], compared by the t test, χ2 test, or Fisher exact test as indicated, p = 0.003). Coronary artery disease (CAD) was observed in n = 3 with CIs set at 95%. hospitalized patients and no cases in nonhospitalized patients.

Table 1 Disease-modifying therapy specific outcomes

Demographics Anti-CD20 S1P inhibitors Glatiramer acetate Natalizumab Dimethyl fumarate Interferon IVIG None

n 34 10 6 4 4 3 3 12

Age, years (SD) 38.72 (SD 14.9) 44.90 (SD 11.2) 53.3 (SD 14.8) 37.8 (SD 17.9) 52.8 (SD 7.2) 57.6 (SD 16.7) 57 (SD 12.8) 49.8 (SD 15.8)

Female n (%) 20 (62.5%) 6 (60%) 4 (66.7%) 2 (50%) 2 (50%) 1 (33.3%) 3 (100%) 8 (66.7%)

Ambulation status, n (%)

Ambulatory 27 (79.4%) 9 (90%) 2 (33.3%) 4 (100%) 3 (75%) 3 (100%) — 8 (66.7%)

Ambulatory with assistance 4 (11.8%) 1 (10%) 3 (50%) —— —2 (66.7%) 1 (8.3%)

Nonambulatory 3 (8.8%) — 1 (16.7%) — 1 (25%) — 1 (33.3%) 3 (25%)

Hospitalization

Not hospitalized 25 (73.5%) 9 (90%) 5 (83.3%) 3 (75%) 2 (50%) 3 (100%) 3 (100%) 8 (66.7%)

Hospitalized 9 (26.5%) 1 (10%) 1 (16.7%) 1 (25%) 2 (50%) ——4 (33.3%)

Comorbidities associated with COVID-19

Obesity 9 (26.5%) 3 (30%) 2 (33.3%) 2 (50%) 1 (25%) — 1 (33.3%) 6 (50%)

Hypertension 4 (11.8%) 3 (30%) 1 (16.7%) 1 (25%) 2 (50%) — 2 (66.7%) 4 (33.3%)

Diabetes 1 (2.9%) 1 (10%) 1 (16.7%) 1 (25%) 1 (25%) — 1 (33.3%) 2 (16.7%)

CAD 1 (2.9%) — 1 (16.7%) 1 (25%) ————

COVID-19 outcomes

Death 2 (5.8%) — 1 (16.7%) 1 (25%) ———2 (16.7%)

Ongoing treatment/still recovering 8 (23.5%) 3 (30%) 1 (16.7%) —— —1 (33.3%) 2 (16.7%)

Recovered 24 (70.6%) 7 (70%) 4 (66.7%) 3 (75%) 4 (100%) 3 (100%) 2 (66.7%) 8 (66.7%)

Abbreviations: Anti-CD20 = ocrelizumab and rituximab; COVID-19 = coronavirus disease 2019; Interferons = interferon beta-1a and interferon beta-1b; IVIG = IV immunoglobulins; S1P inhibitors = fingolimod and siponimod. One 58-year-old woman with a history of hypertension on leflunomide had a mild course. 3 Table 2 Demographics of hospitalized vs nonhospitalized patients with COVID-19

Demographics Hospitalized (n = 18) Nonhospitalized (n = 58)

Age, years (range, SD) 52.0 (17–71, 14.6) 42.7 (13–71, 14.8)

Sex (% female) 10 (55.6%) 37 (63.8%)

Race, n (% white) 12 (66.7%) 38 (65.5%)

Ethnicity, n (% Hispanic) 6 (33.3%) 9 (15.5%)

Clinical diagnosis

RRMS 8 (44.4%) 47 (81.0%)

Progressive MS 7 (38.8%) 10 (17.2%)

Other demyelinating syndrome 3 (16.7%) 1 (1.7%)

Disease duration, years (range, SD) 17.2 (5–36, 10.5) 14.6 (1–52, 10.8)

Ambulation status, n (%)

Ambulatory 9 (50%) 47 (81%)

Ambulatory with assistance 5 (27.8%) 6 (10.3%)

Nonambulatory 4 (22.2%) 5 (8.6%)

Comorbidities associated with COVID-19

Obesity 11 (61%) 12 (20.7%)

Hypertension 5 (27.8%) 12 (20.6%)

CAD 3 (16.6%) —

Diabetes 2 (11.1%) 6 (10.3%)

Venous thromboembolism 3 (16.6%) 1 (1.7%)

COVID-19 outcomes

Death, n (%) 5 (27.7%) 1 (1.7%)

Ongoing treatment/still recovering 4 (22.2%) 10 (17.2%)

Recovered, n (%) 9 (50.0%) 47 (81.0%)

Abbreviations: CAD = coronary artery disease; COVID-19 = coronavirus disease 2019; DMT = disease-modifying therapy.

History of venous thromboembolism was recorded in n = 3 relapsing-remitting MS. Noted comorbidities were obesity hospitalized and n = 1 nonhospitalized patient. (n = 3), type I diabetes (n = 1), or both (n = 1). Two of the 9 were hospitalized requiring supplemental oxygen. None Eight patients (10.5%) had critical illness deﬁned by ICU admis- required invasive ventilation. Eight patients were either fully sion (n = 5) and/or death (n = 6) (table 3). Compared with the recovered or recovering at censoring date. One remains entire MS sample, the critically ill patients were older (mean of 57.7 hospitalized. ± 10.5 years, range 42–71 years), more likely to have a progressive subtype (50.0%), and or required assistance for ambulation/ Discussion nonambulatory (62.5% requiring assistance or nonambulatory). Following the pattern observed in those who were hospitalized, the Patients with MS and related disorders often seek guidance critically ill group had had high rates of comorbid obesity (62.5%), regarding the impact of the disease and medication on their CAD (25%), and venous thromboembolism (37.5%). There were risk of COVID-19. This observational study—although sub- no reports of stroke. One patient had hypercoagulability resulting ject to sampling and ascertainment biases—provides some in multiple venous thromboembolisms and ultimately died. insights regarding COVID-19 disease course in actively treated patients with MS and related disorders. Nine patients with pediatric-onset MS were identiﬁed as the center has a large pediatric MS sample (table 4). Ages The rate of hospitalization in our patients (24%) and mor- ranged from 13 to 26 years; 8 were female, and all had tality (7.9%) are in line with the data of another published MS

Table 3 Critically ill and deceased patient outcomes

Diagnosis and Age/ Race/ disease duration Ambulation sex ethnicity (y) DMT status Comorbidities COVID course and outcome

42/ Black/non- RRMS—18 y Rituximab Nonambulatory Chronic anticoagulation for VTE, Hodgkin Several weeks of fevers, cough, and dyspnea. A family member in the home M Hispanic lymphoma, and ITB was also sick with similar symptoms. Died at home.

46/ Black/non- SPMS—26 y None Nonambulatory Diabetes, morbid obesity, hypertension, chronic Hospitalized with dyspnea, fever, and cough and received high-flow oxygen. M Hispanic anticoagulation for VTE, and hyperlipidemia Discharged to a rehabilitation facility on supplemental oxygen

50/F White/non- RRMS—13 y None Ambulatory Hypertension, obesity, and hypothyroid Hospitalized with respiratory failure following treatment with 5 d of IVMP for Hispanic MS exacerbation. Received ventilator and ECMO support. Deceased.

60/F Black/non- RRMS—19 y Natalizumab Ambulatory CAD with cardiac stents, hypertension, and obesity Hospitalized with fatigue, cough, fever, and respiratory failure. Found to have Hispanic DVT and PE. Had cardiac arrest. Deceased.

62/ White/ NMOSD—7 y Rituximab Ambulatory Obesity Hospitalized with dyspnea and fever and on high-flow oxygen. Remains M Hispanic hospitalized.

65/F White/non- SPMS—31 y None Nonambulatory ITB and neurogenic bladder with indwelling Foley Hospitalized with fever, dyspnea, and code status DNR/DNI. Deceased Hispanic

66/ White/non- SPMS—33 y Ocrelizumab Nonambulatory Remote history of testicular and prostate cancer Hospitalized with respiratory failure and received ventilator support. M Hispanic and ITB Deceased.

71/ Black/non- SPMS—30 y Glatiramer Ambulatory Chronic anticoagulation for VTE obesity Hospitalized with dyspnea and fever and had cardiac arrest. Deceased. M Hispanic acetate with assistance

Abbreviations: CAD = coronary artery disease; COVID-19 = coronavirus disease 2019; DMT = disease-modifying therapy; DNR/DNI = do not resituate/do not intubate; DVT = deep venous thrombosis; ECMO = extracorporeal membrane oxygenation; ITB = intrathecal baclofen pump; IVMP = IV methylprednisolone; PE = pulmonary embolus; RRMS = relapsing-remitting MS; SPMS = secondary progressive MS; VTE = venous thrombosis. Of the 8 patients, 7 tested positive for COVID-19. The patient who died at home was not tested for COVID as he did not receive any hospital care before expiration. Both he and a family member he resided with had symptoms of fever, cough, and dyspnea for several weeks prior. The family member has since recovered. 5 of COVID-19–infected patients on anti-CD20 therapies Table 4 Demographics of patients with pediatric-onset (44.7%) compared with our MSCCC population in which MS with COVID-19 33.1% of patients take anti-CD20 therapies. This observation Full sample could be an artifact of sampling bias or might represent an (n = 9) increased susceptibility to COVID-19 infection as anti-CD-20

Demographics therapies increase the risk of non-COVID-19 infections in general.1 It is worth noting that although we had a larger-than- Age, years (range, SD) 19.44 (13–26, 4.12) expected portion of patients on anti-CD20 therapies, there is no evidence to suggest worse outcomes. Sex (% female) 7 (77.7%)

Race, n (% white) 6 (66.6%) As a note of caution in the care of patients with MS, a subgroup of individuals reported worsening of preexisting neu- Ethnicity, n (% Hispanic) 5 (55.5%) rologic symptoms before or at onset of COVID-19 symptoms. Clinical history One such patient was treated with several days of high-dose

MS subtype, n (% RRMS) 9 (100%) steroids after which they presented with respiratory and cir- culatory failure and ultimately died (table 3). Infections are – Disease duration, years (range, SD) 6.11 (2 12, 3.55) a well-known cause of pseudoexacerbations, and in areas with DMT, n (%) high prevalence of COVID-19, testing for SARS-COV2

Glatiramer 1 (11.1%) should be considered in patients with acute worsening of preexisting symptoms before steroid treatment is initiated. Ocrelizumab 2 (22.2%)

Rituximab 4 (44.4%) Our study has several limitations. This was a convenience sample and not randomly selected, nor was the entire practice Natalizumab 1 (11.1%) systematically surveyed. Patients were identified during routine None 1 (11.1%) teleneurology visits, if they notified the office, or if — Ambulation status, n (% fully ambulatory) 9% (100) hospitalized likely leading to an overrepresentation of more symptomatic individuals. Patients on higher potency infusible COVID-19 testing, comorbidities, and hospitalization DMTs were more likely to have frequent follow-up and may have been more readily captured. In contrast, patients in nursing Tested for COVID-19, n (%) 5 (55.5%) homes, who would be expected to be most severely affected by Positive, n (% tested) 4 (80%) COVID-19, visit specialized centerslessfrequentlyandcouldbe underrepresented in our cohort. Although we used a systematic Diabetes, type I 2 (22.2%) questionnaire to collect relevant data, we could not verify these Obesity 3 (33.3%) data independently unless the patient was seen by an NYU ffi Asthma 1 (11.1%) Langone-a liated physician or hospital. Another problem was lack of access to COVID-19 testing in our area. Less than half of Hospitalized, n (%) 2 (22.2%) our patients (48.7%) underwent SARS-COV2 PCR testing. As

Abbreviation: DMT = disease-modifying therapy. shown in table 5, subgroup analysis showed that this group was not different from the overall sample with respect to demographic or MS-related features. patient survey4 and for the general population of patients with COVID-19.5 Similar risk factors were identified in- Our early experience with COVID-19 at NYU Langone cluding older age, male sex, and high number of comor- MSCCC could inform clinicians taking care of patients with MS bidities such as obesity, diabetes, hypertension, and during the pandemic. Our findings suggest that individuals with – CAD.6 8 MS-specific features associated with more severe MS who experience COVID-19 have similar disease course, COVID-19 included nonambulatory status and pro- outcomes, and risk factors for complications as the general gressive disease course. Given the small sample size, we population. Rigorous, population-based studies are needed to could not determine whether these patients were at higher confirm our preliminary findings and better define the risk of risk, given their advanced age and other comorbidities, or COVID-19 infection with respect to individual DMTs. Future whether worse disability in and of itself represents an ad- studiesshouldassesstheroleofbaselinelymphocytecountsand ditional risk factor. immunoglobulin levels with respect to viral susceptibility and course as well as examine the frequency of serologies to We did not observe an association between DMT class and COVID-19 according to DMT class. As we and others collect COVID-19 outcome in univariate comparisons; however, our more data and contribute to larger MS registries, we expect sample size is small, and these preliminary findings should be more answers will be forthcoming to further guide patient interpreted with caution. There is a relatively high proportion management.

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Table 5 Demographics and outcomes of all patients vs COVID-19 PCR confirmed positive

All patients (n = 76) COVID PCR positive (n = 37)

Demographics

Age, years (range, SD) 44.9 (13–71, 15.2) 47.5 (13–71, 15.15)

Sex, F/M (% female) 47/29 (61.8%) 23/14 (62.2%)

Race, n (%)

White 50 (65.8%) 24 (64.9%)

Black 21 (27.6%) 12 (32.4%)

Asian 3 (3.9%) —

Pacific Islander 1 (1.3%) —

Other 1 (1.3%) 1 (2.7%)

Ethnicity, n (%)

Non-Hispanic 59 (77.6%) 28 (75.7%)

Hispanic 15 (19.7%) 8 (21.6%)

Other 2 (2.6%) 1 (2.7%)

Smoking status, n (%)

Never 52 (68.4%) 25 (67.5%)

Former 20 (26.3%) 10 (27.02)

Current 2 (2.6%) 1 (2.7%)

Unknown 2 (2.6%) 1 (2.7%)

Clinical

MS, n (%)

RRMS 55 (72.3%) 25 (67.6%)

SPMS 15 (19.7%) 10 (27.0%)

PPMS 2 (2.6%) —

Other demyelinating syndrome, n (%) 4 (5.2%) 2 (5.4%)

Disease duration, years (range, SD) 15.2 (1–52, 10.7) 16.9 (2–52, 11.46)

DMT, n (%)

Anti-CD20 34 (44.7%) 14 (37.8%)

S1P Inhibitors 10 (13.2%) 4 (10.8%)

Glatiramer acetate 6 (7.9%) 4 (10.8%)

Natalizumab 4 (5.3%) 3 (8.1%)

Dimethyl fumarate 4 (5.3%) 2 (5.4%)

Leflunomide 1 (1.3%) 1 (2.7%)

Interferons 3 (3.9%) 1 (2.7%)

IVIG 3 (3.9%) 1 (2.7%)

None 12 (15.8%) 7 (18.9%)

Ambulation status, n (%)

Ambulatory 56 (73.7%) 24 (64.9%)

Ambulation with assistance 11 (14.4%) 7 (18.9%)

Continued

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 7 Table 5 Demographics and outcomes of all patients vs COVID-19 PCR confirmed positive (continued)

All patients (n = 76) COVID PCR positive (n = 37)

Nonambulatory 9 (11.8%) 6 (16.2%)

Comorbidities associated with COVID-19

Hypertension 17 (22.4%) 11 (29.7%)

Obesity 23 (30.3%) 11 (29.7%)

Diabetes 8 (10.5%) 5 (13.5%)

VTE 4 (5.3%) 3 (8.1%)

CAD 3 (3.9%) 3 (8.1%)

History of cancer 4 (5.3%) 2 (5.4%)

Baclofen pump 3 (3.9%) 2 (5.4%)

Indwelling Foley 3 (3.9%) 2 (5.4%)

None 24 (31.6%) 12 (32.4%)

COVID-19 outcomes

Death, n (%) 6 (7.9%) 5 (13.5%)

Ongoing treatment/still recovering 14 (18.4%) 7 (18.9%)

Recovered, n (%) 56 (73.7%) 25 (67.6%)

Abbreviations: Anti-CD20 = ocrelizumab and rituximab; CAD = coronary artery disease; COVID-19 = coronavirus disease 2019; DMT = disease-modifying therapy; Interferons = interferon beta-1a and interferon beta-1b; IVIG = IV immunoglobulins; S1P inhibitors = fingolimod and siponimod; VTE = venous thromboembolism.

Study funding No targeted funding reported. Appendix Authors

Name Location Contribution Disclosure E. Parrotta reports no disclosures. I. Kister served on advisory Erica NYU Langone Organization and Parrotta, DO Comprehensive Care acquisition of data; boards for Biogen and Genentech and received consulting Center, New York, NY analyzed and interpreted fees from Roche and research support for investigator- data; drafted results; and revised the manuscript for initiated grants from Sanoﬁ Genzyme, Biogen, EMD Serono, intellectual content National MS Society, and the Guthy-Jackson Charitable Ilya Kister, NYU Langone Interpreted the data; Foundation. L. Charvet reports no disclosures. C. Sammarco MD Comprehensive Care drafted the discussion; receives consulting compensation from Biogen and Gen- Center, New York, NY and revised the manuscript for intellectual entech. V. Saha, R.E. Charlson, J. Howard, J.M. Gutman, M. content Gottesman, N. Abou-Fayssal, R. Wolintz, M. Keilson, and C. Leigh NYU Langone Analyzed the data; Fernandez-Carbonell report no disclosures. L.B. Krupp has Charvet, PhD Comprehensive Care performed statistical received advisory board/consulting fees, travel and meal Center, New York, NY analysis; and revised the ﬁ manuscript for intellectual allowances, and/or research funding from Sano Aventis, content Biogen, Novartis, Gerson Lehrman, EMD Serono, Allergan Inc., and Tesaro Inc. She is also a noncompensated consultant Carrie NYU Langone Major role in acquisition of Sammarco, Comprehensive Care data and/or advisory board member with Novartis and Celgene DNP Center, New York, NY and receives royalties for use of the Fatigue Severity Scale by Valerie Saha, NYU Langone Major role in acquisition of various biopharmaceutical entities. L. Zhovtis Ryerson re- NP Comprehensive Care data ceived personal compensation for advisory board participa- Center, New York, NY tion for Biogen and received research support from Biogen. Robert Erik NYU Langone Acquisition of data Go to Neurology.org/NN for full disclosures. Charlson, Comprehensive Care MD Center, New York, NY

Publication history Jonathan NYU Langone Acquisition of data ﬂ Howard, MD Comprehensive Care Received by Neurology: Neuroimmunology & Neuroin ammation Center, New York, NY May 11, 2020. Accepted in ﬁnal form June 12, 2020.

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Appendix (continued) Appendix (continued)

Name Location Contribution Name Location Contribution

Josef NYU Langone Acquisition of data Lauren B. NYU Langone Interpreted the data and Maxwell Comprehensive Care Krupp, MD Comprehensive Care drafted and revised the Gutman, MD Center, Huntington, NY Center, New York, NY manuscript for intellectual content Malcolm NYU Langone MS Acquisition of data Gottesman, Comprehensive Care Lana Zhovtis NYU Langone Designed and MD Center, Mineola, NY Ryerson, MD Comprehensive Care conceptualized the study; Center, New York, NY analyzed the data; and Nada Abou- NYU Langone Acquisition of data drafted and revised the Fayssal, MD Comprehensive Care manuscript for intellectual Center, Brooklyn, NY content

Robyn NYU Langone Acquisition of data Wolintz, MD Comprehensive Care Center, Brooklyn, NY 2. Mehta P, McAuley DF, Brown M, Sanchez E, Tattersall RS, Manson JJ. COVID-19: consider cytokine storm syndromes and immunosuppression. The Lancet 2020;395: Marshall NYU Langone Acquisition of data 1033–1034. Keilson, MD Comprehensive Care 3. Available at: cdc.gov/coronavirus/2019-nCoV/hcp/index.html. Accessed May 4, 2020. Center, Brooklyn, NY 4. Sormani MP. An Italian programme for COVID-19 infection in multiple sclerosis. Lancet Neurol 2020;19:481–482. ’ Cristina Cohen s Children Medical Acquisition of data 5. Zhou F, Yu T, Du R, et al. Clinical course and risk factors for mortality of adult Fernandez- Center Northwell Health, inpatients with COVID-19 in Wuhan, China: a retrospective cohort study. Lancet Carbonell, Lake Success, NY 2020;395:1054–1062. MD 6. Lighter J, Phillips M, Hochman S, et al. Obesity in patients younger than 60 years is a risk factor for Covid-19 hospital admission. Clin Infect Dis Epub 2020 Apr 9. 7. Petrilli CM, Jones SA, Yang J, et al. Factors associated with hospitalization and critical illness among 4,103 patients with COVID-19 disease in New York City. References BMJ 2020;369:m1966. 1. Luna G, Alping P, Burman J, et al. Infection risks among patients with multiple 8. Richardson S, Hirsch JS, Narasimhan M, et al. Presenting characteristics, comorbid- sclerosis treated with ﬁngolimod, natalizumab, rituximab, and injectable therapies. ities, and outcomes among 5700 patients hospitalized with COVID-19 in the New JAMA Neurol 2020;77:184–191. York City area. JAMA 2020;323:2052–2059.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 9 ARTICLE OPEN ACCESS CLASS OF EVIDENCE Treatment and outcome of aquaporin-4 antibody–positive NMOSD A multinational pediatric study

Renata Barbosa Paolilo, MD,* Yael Hacohen, MD, PhD,* Elise Yazbeck, MD, Thais Armangue, MD, PhD, Correspondence Arlette Bruijstens, MD, Christian Lechner, MD, Samira Luisa Apostolos-Pereira, MD, PhD, Dr. Paolilo [email protected] Yana Martynenko, MD, Markus Breu, MD, Carolina de Medeiros Rimkus, MD, PhD, Evangeline Wassmer, MD, Matthias Baumann, MD, Laura Papetti, PhD, Marco Capobianco, MD, Barbara Kornek, MD, PhD, Kevin Rost´asy, MD, Jos´e Albino da Paz, MD, PhD, Olga Ciccarelli, MD, PhD, Ming Lim, MD, PhD, Albert Saiz, MD, PhD, Rinze Neuteboom, MD, PhD, Romain Marignier, MD, PhD, Cheryl Hemingway, MD, PhD, Douglas Kazutoshi Sato, MD, PhD, and Kumaran Deiva, MD, PhD

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e837. doi:10.1212/NXI.0000000000000837 Abstract Objective To describe the clinical phenotypes, treatment response, and outcome of children with antibodies against aquaporin-4 (AQP4-Ab) neuromyelitis optica spectrum disorder (NMOSD).

Methods Retrospective, multicenter, and multinational study of patients with AQP4-Ab NMOSD aged <18 years at disease onset from a center in Brazil and 13 European centers. Data on demographics, clinical ﬁndings, and laboratory results were analyzed; calculation of annualized relapse rates (ARRs) pre- and on-treatment with disease-modifying therapies (DMTs) and of ORs for predictors of poor outcome was performed.

Results A total of 67 children were identified. At last follow-up (median 4 years, interquartile range 2–10 years), 37/67(57.8%) were found to have permanent disability. A more severe disease course was seen in the non-White ethnicity with both a shorter time to first relapse (p = 0.049) and a worse Expanded Disability Status Scale score at last follow-up (p = 0.008). The median ARR on treatment was 0.18 on azathioprine (n = 39, range 0–4), 0 on mycophenolate mofetil (n = 18, range 0–3), and 0 on rituximab (n = 29, range 0–2). No patient treated with rituximab as first-line therapy relapsed. Optic neuritis at onset was associated with a poor visual outcome below 20/200 (OR 8.669, 95% CI 1.764–42.616, p = 0.008), and a younger age at onset was associated with cognitive impairment (OR 0.786, 95% CI 0.644–0.959, p = 0.018).

*These authors have contributed equally to this manuscript.

From the Department of Neurology (R.B.P., S.L.A.-P., J.A.d.P.) and Department of Radiology (INRAD) (C.d.M.R.), Hospital das Cl´ınicas, Faculty of Medicine, University of São Paulo (HCFMUSP), Brazil; Queen Square MS Centre (Y.H., O.C.), UCL Institute of Neurology, Faculty of Brain Sciences, University College London; Department of Paediatric Neurology (Y.H., C.H.), Great Ormond Street Hospital for Children, London, United Kingdom; Assistance Publue-Hopitauxˆ de Paris (E.Y., K.D.), University Hospitals Paris South, Bicêtre Hospital, Le Kremlin Bicêtre, France; Neuroimmunology Program (T.A.) and Neurology Service (A.S.), Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Hospital Cl´ınic, and Pediatric Neuroimmunology Unit (T.A.), Sant Joan de Deu (SJD) Children’s Hospital, Universitat de Barcelona, Spain; Department of Neurology (A.B., R.N.), Erasmus University Medical Center, Rotterdam, the Netherlands; Division of Paediatric Neurology (C.L., M.B.), Department of Pediatrics I, Medical University of Innsbruck, Austria; Medical Center of Physical Therapy and Pain Medicine INNOVO (Y.M.), Lviv, Ukraine; Division of Pediatric Neurology (M.B.), Department of Pediatrics and Adolescent Medicine, and Department of Neurology (B.K.), Medical University of Vienna, Austria; Pediatric Neurology (E.W.), Birmingham Women and Children’s Hospital, United Kingdom; Department of Neuroscience (L.P.), Pediatric Multiple Sclerosis Center, Bambino Gesu` Children Hospital, IRCCS, Rome, Italy; Department of Neurology (M.C.) and Regional Multiple Sclerosis Centre (M.C.), University Hospital San Luigi Gonzaga, Orbassano, Italy; Department of Pediatric Neurology (K.R.), Vestische Kinder und Jugendklinik, Witten/Herdecke University, Datteln, Germany; Children’s Neurosciences (M.L.), Evelina London Children’s Hospital at Guy’s and St Thomas’ NHS Foundation Trust, King’s Health Partners Academic Health Science Centre; Faculty of Life Sciences and Medicine (M.L.), King’s College Hospital, London, United Kingdom; Neurology Department (R.M.), Hopitalˆ Neurologique Pierre Wertheimer, Hospices Civils de Lyon; French Reference Network of Rare Inflammatory Brain and Spinal Diseases (R.M., K.D.), Le Kremlin Bicêtre, France; School of Medicine (D.K.S.), Brain Institute of Rio Grande do Sul (Brains), Pontifical Catholic University of Rio Grande Do Sul (PUCRS), Porto Alegre, Brazil; and Inserm UMR 1184 (K.D.), Immunology of Viral Infections and Autoimmune Diseases, CEA, IDMIT, Le Kremlin Bicêtre, France.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

The Article Processing Charge was funded by Assistance Publique-Hopitauxˆ de Paris, Hopitauxˆ Universitaires Paris Saclay. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary ADS = acquired demyelinating syndrome; AQP4-Ab = antibodies against aquaporin-4; ARR = annualized relapse rate; AZA = azathioprine; CRF = case reporting form; DMT =disease-modifyingtherapy;EDSS = Expanded Disability Status Scale; IQR = interquartile range; MMF = mycophenolate mofetil; MOG-Ab = myelin oligodendrocyte glycoprotein antibody; NMOSD = neuromyelitis optica spectrum disorder; ON = optic neuritis; TM = transverse myelitis.

Conclusions AQP4-Ab NMOSD in children is an aggressive disease with permanent disabilities observed in over half the cohort. All DMTs were associated with a reduction of ARR. First-line rituximab prevented further clinical relapses. International consensus on treatment protocols for children is required to reduce heterogeneity of treatment regimens used worldwide.

Classification of evidence This study provides Class IV evidence that for children with AQP4-Ab NMOSD, all DMTs, particularly ﬁrst-line rituximab, reduced the ARR and prevented further clinical relapses.

Antibodies against aquaporin-4 (AQP4-Ab) were first described to 4.5% (3/64)13 of children presenting with a first pre- in 2004 in patients with neuromyelitis optica (NMO)1 allowing sentation of acquired demyelinating syndrome (ADS) and 8/ the expansion of the phenotype.2 The most recent criteria for 102 (7.8%) of children with relapsing syndromes.14 Children the diagnosis of NMO spectrum disorder (NMOSD) stratify are reported to have a less severe disease course and may take patients by the presence/absence of AQP4-Ab.3 AQP4-Ab se- longer to reach disability than adults.15 Children are at a ropositivity is associated with relapsing disease.4,5 This led to the higher risk of visual impairment compared with adults but are use of B-cell targeting therapies, which clearly reduce relapse less likely to acquire motor deficits.16 Previous pediatric rates.6 This reduction is not seen when therapies known to be publications highlighted that AQP4-Ab NMOSD in Europe is effective for MS7 are used in NMOSD. rare,9 whereas the prevalence in South America8 is higher.

The clinical features and MRI abnormalities in children with With the rarity of pediatric presentation, treatment is derived from – AQP4-Ab NMOSD are similar to the adult phenotype.8 11 adult guidelines and can be inﬂuenced by medication availability The prevalence of AQP4-Ab was reported in 0.7% (2/279)12 and cost. Current available treatments used, such as azathioprine

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN (AZA), mycophenolate mofetil (MMF), and rituximab, have not hemispheric syndromes (associated with encephalopathy or received regulatory approval for NMOSD. In this retrospective, without encephalopathy). multicenter, and multinational study, patients’ demographics, first attack features, paraclinical characteristics, and disease course are Brain and spinal cord MRI were performed in 61 and 50 described to ultimately evaluate responses to different treatment patients, respectively, at disease onset. All patients had un- strategies in children with AQP4-Ab NMOSD. dergone brain and spinal cord imaging according to local MRI protocols (not routinely including orbits). Gadolinium- enhanced imaging was performed in 54 patients. Lumbar Methods puncture for CSF analysis was performed in 57/67 (85.0%) patients. Participants In this multicenter, multinational study, we collected de- The acute treatment for each of these patients at presentation mographic, clinical, and radiologic data of 67 patients from a and subsequent episodes of relapses was decided by the single center in Brazil (São Paulo, n = 20) and from 13 centers treating pediatricians, on the basis of protocols influenced by in 7 countries as part of the EU Paediatric Demyelinating their regional and/or national reference center for CNS de- Disease Consortium (United Kingdom [n = 18], France [n = myelination, guided by severity and persistence of symptoms. 11], Spain [n = 6], Germany/Austria [n = 5], the Disease-modifying therapies (DMTs) referred to all forms of Netherlands/Belgium [n = 4], Italy [n = 2], and Ukraine [n = maintenance immunomodulation or immunosuppression 1]). This consortium was initiated to study children with therapies. ADS, as part of the European Reference Network for Rare Immunodeficiency, Autoinflammatory and Autoimmune Annualized relapse rates (ARRs) on retrospective data were Disease. calculated as number of relapses per year and only included patients with at least 6 months of follow-up after initiation of We retrospectively identified participants who were treatment. If time to treatment was less than 6 months, the recruited into the respective centers or national de- pretreatment ARR was calculated over a 6-month period. The myelination programs and fulfilled the following inclusion outcomes, as measured by a range of difficulties the patients criteria: (1) NMOSD, fulfilling the 2015 International Panel were experiencing (cognitive, visual, and motor), at last for NMO diagnosis criteria,3 (2) AQP4-Abs detected at follow-up were retrieved from the patient’s medical records to onset or at the time of a clinical relapse, using live cell-based represent the most contemporary assessment of disability. If assays in the local laboratories, and (3) age <18 years at first unavailable, this assessment was obtained directly from the presentation. patient’s primary treating physician. Patients were considered Standard protocol approvals, registrations, to have motor disability if they were a wheelchair user, needed and patient consents crutches (or a cane) to walk, or could walk for less than 500 m Patients included in this study had been enrolled in national without support. Cognitive performance was estimated by programs with respective review board/ethical committee looking at their school performance at last follow-up: patients ffi approvals (Brazil [Hospital das Cl´ınicas, Faculty of Medicine, were considered to have cognitive di culties if there was fi University of São Paulo, São Paulo], France [Hopitalˆ Bicetre,ˆ signi cant school support, grade repetition, or needing of Paris], the Netherlands [Medisch Ethische Toetsingscom- special education schools. Patients were considered to have missie Erasmus Medical Centre, Rotterdam], United King- visual sequelae if the logMAR scale was <0.3 (Snellen <20/ fi dom [West Midlands-South Birmingham Research Ethics 40) or persistent visual eld defects. The Expanded Disability fi Committee], Germany and Austria [University of Innsbruck Status Scale (EDSS) scores were documented at rst attack fi Ethics Committee], and Spain [Hospital Clinic and by Sant nadir and at nal follow-up (at point of disease stability at least Joan de Déu Children’s Hospital] or provided verbal and/or 3 months from acute or relapsing events). written informed consent to the respective referring physician (Italy, Ukraine). All data were deidentified. All CRFs were initially reviewed by the respective national leads and subsequently analyzed by 3 investigators (R.B.P., Procedure Y.H., and K.D.). Clinical data already collected as part of national demyelination programs were deidentified and entered by each AQP4-Ab testing participating investigator onto a unified case reporting form Within 1 month of an acute event (either onset or relapse), (CRF), detailing selected demographics, clinical findings and clinically symptomatic children underwent testing for serum laboratory results (AQP4-Abs, CSF cell count, protein, and AQP4-Ab, using live cell-based assays in the respective ref- oligoclonal bands), first and subsequent attack characteristics, erence laboratories of the following referring countries and treatment information. Demyelinating phenotype at on- (France,18 Brazil,19 United Kingdom,20 Spain,21 the set and relapses were clinically determined according to Netherlands,22,23 in Germany/Austria in the Clinical De- established criteria18 as being optic neuritis (ON), transverse partment of Neurology, Medical University of Innsbruck, myelitis (TM), brainstem and/or cerebellar and/or Innsbruck, Austria and in Italy and Ukraine anti-AQP4

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 antibodies testing were performed with commercial IFA Family history of autoimmune diseases was reported in 9 [Euroimmun]). Routine assessments of antibody testing in (13.7%) patients, personal history of inflammatory diseases in the CSF were not performed. 7 (10.4%), and 8 (11.9%) had a preceding infection. Patients’ demographic, clinical, and paraclinical features and EDSS Statistical analysis scores, stratified to ethnicities, are summarized in table 1. The We tested the hypothesis that patients treated with AZA, groups were similar except for shorter time to first relapse (η2 MMF, and/or rituximab experienced a reduction in ARR = 0.071, p = 0.049) and higher EDSS score at last follow-up providing Class IV evidence for therapeutic decision. (η2 = 0.056, p = 0.008) in the non-White population.

Descriptive data of demographic, clinical, radiologic, and se- First attack features rologic characteristics were reported for children with AQP4- The most frequent phenotype at onset was isolated ON (n = 20, Ab NMOSD. To compare between the different ethnicities, 29%) followed by isolated TM (n = 15, 22%) and isolated area parametric or nonparametric statistical tests (Mann-Whitney postrema syndrome (n = 11, 16.4%). Of the children presenting U and Kruskal-Wallis tests) were used for continuous distri- with ON, 9/20 (45%) had bilateral ON. Of the 15 children butions, as appropriate given normality, and χ2 or Fisher exact presenting with isolated TM, 10 (66.7%) had longitudinally tests for nominal data. In particular, comparisons between extensive TM. Six (9%) patients had simultaneous ON and TM. White and non-White ethnicities and stratified EDSS score at Six(9%)patientspresentedwithADEM,1(1.5%)patientwith last follow-up were performed. Logarithm transformation was isolated diencephalic syndrome, 3 (4.5%) with isolated brain- performed for nonparametric data. Effect sizes were calculated stem syndrome, and 5 (7.5%) patients with multifocal syn- by OR, Cohen d, and eta-squared (η2) measures. dromes involving the spinal cord, optic nerve, and brain (brainstem, area postrema, or diencephalic syndromes). In addition, parametric or nonparametric statistical tests were used for continuous distributions, as appropriate, and χ2 or At onset, 44/61 (72.1%) patients had abnormal brain MRI, Fisher exact tests for nominal data to compare patients who 32/50 (64%) patients had abnormal spinal cord MRI, and 16/ were treated with AZA, MMF, and rituximab as first-line 54 (29.6%) had contrast-enhancing lesions. Abnormal CSF treatment. A paired 2-tailed t test was used to compare ARRs was reported in 37/57 (64.9%) children, with oligoclonal before and during treatment. bands detected in 7/57 (12.2%). CSF pleocytosis was present in 35/50 (70%) examinations, with a median of 10 cells Univariate logistic regression was used to evaluate age at (interquartile range [IQR] 2.5–38.5 cells). Elevated CSF onset, sex, ethnicity, demyelinating phenotype at onset, ab- protein was found in 13/27 (48%) examinations with a me- normal MRI at onset, time to treatment, number of relapses dian of 0.030 g/dL (IQR 0.002–0.850 g/dL). Myelin oligo- before treatment, and follow-up duration as potential pre- dendrocyte glycoprotein antibodies (MOG-Abs) were dictors of motor (wheelchair or walking aid), visual (blind- negative in the 60 patients who were tested. ness), and cognitive outcomes at last follow-up. Multivariable analyses were modeled for the same outcomes, including as Disease course and outcome covariates those variables significant at the (p < 0.05) level in A total of 297 attacks were reported in the cohort. The median the univariate analysis. Results were presented as ORs and time to first relapse was 4 months (range 1.7–100.5 months). 95% CIs. A 2-sided p < 0.05 was considered significant. Data At last follow-up (median disease duration of 4 years, IQR were analyzed using SPSS 20 software and GraphPad Prism 5. 2–10 years), 58 (86.6%) patients had a second clinical event with a median ARR of 1.05 relapses per year and median Data availability (IQR) EDSS score = 2.0 (1.0–3.5). Moderate disability with Anonymized data will be shared by request from qualified EDSS score ≥3 was reported in 29/67 (43.2%). Visual im- investigators. pairment was seen in 32 (47.8%) patients, of which 20 were registered blind (visual acuity <20/200; logMAR >1), motor deficits were found in 14 (21.2%, of which 5 were wheelchair Results users and 2 used a walker), and cognitive impairment was Patients detected in 17 (25.4%) patients. A total of 67 children with AQP4-Ab NMOSD were included in this study (table 1). Mean age at onset was 10.2 ± 3.6 years A more severe disease course was seen in the non-White (range 2–16 years). The female/male ratio was 4.1:1. Of the ethnicity with worse EDSS score at last follow-up (p = 0.008) fi 67 children, 29 (43.3%) were White, 14 (20.9%) Black, 13 and a shorter time to rst relapse (p = 0.049). (19.4%) Brazilian mixed ethnicity, and 11 (16.4%) other ethnicities. Of the 67 children, 29 (43.3%) were White, 14 Response to immunotherapy (20.9%) Black, 13 (19.4%) Brazilian mixed ethnicity and DMTs were given in 63/67 (95%) children. A total of 41 11 (16.4%) other ethnicities (Asian, Indians, Nepali, Afro- patients were treated with 1 DMT, 12 with 2, 7 with 3 DMTs, Caribbean, Omani). Non-White ethnicity was more common 2 with 4 DMTs, and 1 with 5 sequential DMTs. Patients were in both European (25/47) and Brazilian (13/20) populations. commenced on DMTs at a median of 6 months from

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Table 1 Clinical/paraclinical features and outcome in children with AQP4-Ab NMOSD stratified to the different ethnicities

White (N = 29) Non-White (N = 38) All (N = 67)

Age at onset, y, mean (SD) 10.6 (3.5) 9.9 (3.7) 10.2 (3.6)

Female sex, n (%) 25/29 (86.2) 29/38 (76.3) 54/67 (80)

EDSS score at nadir, median (IQR) N = 17; 3.0 (2–4.5) N = 30; 4.7 (3.3–7.5) N = 47; 4.0 (3–6.5)

Baseline attack

Area postrema syndrome (all), n (%) 8/29 (27.6) 6/38 (15.8) 14/67 (21)

Optic neuritis presentation (all), n (%) 15/29 (51.7) 14/38 (36.8) 29/67 (43)

Transverse myelitis presentation (all), n (%) 10/29 (34.5) 14/38 (36.8) 24/67 (36)

Brainstem/cerebral involvement (all), n (%) 9/29 (31) 17/38 (44.7) 26/67 (39)

Abnormal brain MRI at onset, n (%) 20/27 (74.1) 24/34 (70.6) 44/61 (72)

Abnormal spinal cord MRI at onset, n (%) 14/21 (66.7) 17/29 (58.6) 31/50 (62)

Intrathecal OCB, n (%) 3/22 (13.6) 4/23 (17.4) 7/47 (15)

Time to first relapse, mo, median (IQR) 5.5 (3–14.2) 4 (2–5) 4 (2–10)

No. of attacks, median (IQR) 3(2–5.7) 2 (2–5) 3 (2–5)

No. of attacks in the first 2 y, mean (SD) 2.2 (1.3) 2.1 (0.9) 2.2 (1.1)

Time to treatment, mo, median (IQR) 6(2–17) 4.5 (1–10.2) 5 (1–13.5)

FU, y, median (IQR) 6(4–10.7) 3.5 (2–6.2) 4 (2–10)

Visual sequelae, n (%) 14/29 (48.3) 18/35 (51.4) 32/64 (50)

Motor sequelae, n (%) 5/29 (17.2) 9/37 (24.3) 14/66 (21)

Cognitive sequelae, n (%) 5/28 (17.9) 12/38 (31.6) 17/66 (26)

EDSS score at last FU, median (IQR) 2.0 (0–3) 2.5 (1–4) 2.0 (1–3.5)

Abbreviations: EDSS, Expanded Disability Status Scale; FU, follow-up; IQR = interquartile range; OCB = oligoclonal band. symptom onset (IQR 2–13.5 months) and a median of 2 treatment was associated with a mean reduction in the ARR of clinical attacks (IQR 1–3 clinical attacks). Of the 4 untreated 1.10 (mean ARR pretreatment 1.69 vs 0.59 on treatment, mean patients, at a median follow-up of 5 years (range 2–10 years), difference 1.10, 95% CI 0.54–1.66, t = 4.01, Cohen d = 0.63, p < only 1 patient relapsed despite persistent antibody positivity. 0.001). Twenty-five of 39 (64.1%) patients relapsed on treatment. Of these, treatment was escalated in 16 patients to MMF A total number of 139 relapses on treatment were observed in (n = 6), rituximab (n = 8), cyclosporine (n = 1), and glatiramer the cohort. Relapses occurred on all treatments, except the 2 acetate (n = 1). Of the 4 who were treated with AZA as second patients treated with ofatumumab and the 1 patient who line, 3 previously failed treatment with methotrexate, glatiramer underwent hematopoietic stem cell transplantation (HSCT). acetate, and cyclophosphamide, and in 1, treatment was de- The clinical course and disease activity in patients who un- escalated after being relapse free on rituximab. derwent therapy with maintenance treatment are illustrated in figure 1. Most frequently used treatments were AZA (total Twenty-nine children were treated with rituximab, 14 treatment duration 209.5 patient years, relapses = 92; 0.43 (48.3%) first line, 9 (31.0%) second line, 5 (17.2%) third line, relapses/treated year), rituximab (total treatment duration 88 and 1 (3.4%) fourth line. The treatment with rituximab was years, relapses = 13; 0.15 relapses/treated year), and MMF associated with a mean reduction in the ARR of 2.36 (mean (total treatment duration 46.2 years, relapses = 23; 0.49 ARR pretreatment 2.50 vs 0.14 on treatment, mean difference relapse/treated year). Patient characteristic stratified to first- 2.36, 95% CI 1.57–3.15, t = 6.13, Cohen d = 1.14, p < 0.001). line treatments is summarized in table 2 and table e-4, links. All 14 patients treated with rituximab as first line did not have lww.com/NXI/A287. further relapses. Of these, 1 patient had prolonged oral steroids (for 1 year) before commencing rituximab, and a second Thirty-nine children were treated with AZA, 35 (89.7%) as patient had additional intravenous immunoglobulin for the first-line therapy, and 4 as second-line treatment. AZA first 6 months. Seven of 29 patients (24.1%) treated with

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 5 Figure 1 Disease course in relation to respective therapies

Patient relapsed on all treatments with a total of 139 relapses on treatment reported in the cohort. Forty-three (69.3%) remained relapse free on treatment; 10/35 (28.6%) treated with azathioprine (AZA), 10/17 (58.8%) treated with mycophenolate mofetil (MMF); and 23/29 (79.3%) treated with rituximab. One patient who relapsed AZA and cyclosporine underwent hematopoietic stem cell transplantation (HSCT) and has not relapsed (follow-up 2 years).

rituximab relapsed on treatment, and these patients were were treated with ofatumumab following infusion reaction previously treated with AZA (n = 5) and MMF (n = 2). In the with rituximab and have not relapsed. Three patients treated 4 children in whom the CD19 count was measured acutely, 1 with cyclophosphamide, 2 with methotrexate and 1 with relapsed with absent B cells and 3 during B cell reconstitution. cyclosporine; all continued to relapse with worsening ARR. Two patients were treated with glatiramer acetate for a Eighteen children were treated with MMF, 10 (55.6%) first presume diagnosis of MS, and both relapsed on treatment. line, 6 (33.3%) second line, 1 (5.6%) third line, and 1 fifth line ARRs before and after treatment initiation are shown in (5.6%). MMF treatment was associated with a mean re- figure 2. duction in the ARR of 0.32 (mean ARR pretreatment 1.04 vs 0.72 on treatment, mean difference 0.32, 95% CI −0.57 to Adverse events 1.22, t = 0.77, Cohen d = 0.18, p = 0.452). Ten patients Of the 37 patients treated with AZA, 5 (13.5%) developed remained relapse free on treatment. Of the 8 who relapsed on lymphopenia (<0.5 × 109/L) without infectious complica- treatment, 5 were changed to rituximab. tions, 3 developed infections (viral meningitis, pneumonia, and varicella), and 3 did not tolerate the treatment in view of Tocilizumab was used in 2 patients as third-line DMT, and gastrointestinal symptoms or raised liver function test. One both relapsed and satralizumab in 2 patients (who were child developed lymphopenia (<0.5 × 109/L) on MMF. In- enrolled in the open-label SAkuraSky trial as an add-on fusion reactions with rituximab occurred in 3 of 29 (10.3%) treatment to MMF), who have not relapsed. Two patients children, of which 2 were switched to ofatumumab due to the

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Table 2 Patients characteristic stratified to first-line treatment

Azathioprine Mycophenolate Rituximab Other treatment/no (n = 35) mofetil (n = 10) (n = 14) treatment (n = 8)

Age when starting treatment, y, 11 (4–17) 9 (3–17) 12 (5–15) 11 (9–15) median (range)

Non-White ethnicity, n (%) 21/35 (60) 4/10 (40) 8/14 (57,1) 3/8 (37.5)

Country of origin, n (%)

Brazil (n = 20) 17/20 (85) 1/20 (5) 0 2/20 (10)

United Kingdom (n = 18) 7/18 (38.9) 5/18 (27.8) 3/18 (16.7) 3/18 (16.7)

France (n = 10) 2/10 (20) 1/10 (10) 6/10 (60) 1/10 (10)

Spain (n = 6) 2/6 (33.3) 1/6 (16.7) 3/6 (50) 0

Germany/Austria (n = 5) 4/5 (80) 0 1/5 (20) 0

Belgium/the Netherlands/Italy/Ukraine (n = 8) 3/8 (37.5) 2/8 (25) 1/8 (12.5) 2/8 (25)

Demyelinating phenotype at onset, n (%)

Optic neuritis (n = 29) 16/29 (55.1) 7/29 (24.1) 4/29 (13.7) 2/29 (6.8)

Transverse myelitis (n = 24) 13/24 (54.1) 2/24 (8.3) 7/24 (29.1) 2/24 (8.3)

All area postrema syndrome (14) 7/14 (50) 1/14 (7.1) 4/14 (28.5) 2/14 (14.2)

All brainstem/encephalic (n = 26) 13/26 (50) 3/26 (11.5) 6/26 (23) 4/26 (15.3)

Time to treatment, mo, median (range) 6(0–84) 6 (0–16) 1 (0–9) —

No. of attacks before treatment, median (range), 2(1–6) 2 (2–4) 2 (1–9) — n(%)

Median duration on treatment, y, median 5.5 (0.5–11) 1.5 (0.8–6) 2 (1–11) — (range)

ARR, median (range) 0.18 (0–4) 0 (0–2) 0 —

No. of patients who relapsed on treatment, n (%) 22/35 (62.8) 3/10 (30) 0 —

No. of patients who changed treatment, n (%) 14/35 (40) 2/10 (20) 2/14 (14.2) — additional one stopped

Abbreviation: ARR = annualized relapse rate.

side effects. One child developed persisted neutropenia on did not identify any predictors for motor disabilities (wheel- rituximab and was changed to AZA. chair or walking aid). ON at onset was associated with worse visual outcome below 20/200 (p = 0.008, OR 8.669, 95% CI Predictors of poor outcome 1.764–42.616), and younger age at onset was associated with ORs to estimate the effects of parameters at onset on the worse cognitive impairment (p = 0.018, OR 0.786, 95% CI occurrence of disability (EDSS score ≥3) are illustrated in 0.644–0.959). table 2. Children with an EDSS score ≥3 (n = 29, 43.2%) were younger, presented with ON, and had a higher EDSS score at nadir. Patients with a better outcome at last follow-up (EDSS score <3.0) were more commonly girls and presented with Discussion cerebral syndrome (table 3). Logistic regression looking at AQP4-Ab NMOSD is now well recognized in children. Nat- age at onset, ethnicity, demyelinating phenotype at onset, ural history studies suggest an attack-related stepwise accu- abnormal MRI at onset, time to treatment, and number of mulation of disabilities. Therefore, attack prevention relapses before treatment were evaluated as predictors of strategies are used as maintenance treatment. In this multi- motor, visual, and cognitive outcomes (table e-1, links.lww. national cohort of children with AQP4-Ab NMOSD, the com/NXI/A284, table e-2, links.lww.com/NXI/A285, and phenotype observed was overall similar to that reported table e-3, links.lww.com/NXI/A286). Multivariable analysis previously in children or adults.3 The time to first relapse was

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 7 Figure 2 The efficacy of various disease-modifying therapies in children with AQP4-Ab NMOSD

Annualized relapse rates (ARRs) before and after treatment initiation with the most frequently used medications. (A) Azathioprine (AZA). (B) Mycophenolate mofetil (MMF). (C) Rituximab. (D) After escalation from treatment with AZA or MMF to rituximab.

short, and the ARR pretreatment was high, indicating a high We detected cognitive impairment in 25.4% of the cohort. risk of relapses and permanent disability in children with Abnormal brain MRI was not a risk factor, but younger age at AQP4-Ab NMOSD. onset and brainstem/cerebral attacks were associated with

Table 3 Univariable analysis and OR for parameters associated with the EDSS score (EDSS ≥3) at final follow-up

EDSS score <3 (n = 38) EDSS score ≥3 (n = 29) OR (95% CI) p Value

Sex (F:M) 34:4 20:9 0.261 (0.071–0.960) 0.043

Age at presentation, y, mean (SD) 11.1 (3.1) 9.1 (3.9) 0.847 (0.732–0.980) 0.026

Ethnicity (White:non-White) 19:19 10:19 1.900 (0.702–5.141) 0.206

Optic neuritis presentation, n (%) 12/38 (31.6) 17/29 (58.6) 3.069 (1.121–8.402) 0.029

Transverse myelitis presentation, n (%) 12/38 (31.6) 12/29 (41.4) 1.529 (0.559–4.186) 0.408

Brainstem/cerebral presentation, n (%) 19/38 (50) 7/29 (24.1) 0.318 (0.110–0.920) 0.035

Abnormal brain MRI, n (%) 25/35 (71.4) 19/26 (73.1) 1.096 (0.349–3.379) 0.887

Intrathecal OCBs, n (%) 7/28 (25) 0/17 NA >0.999

EDSS score at nadir, mean (SD) N = 26; 4.0 (2.3) N = 21; 5.4 (2.2) 1.303 (1.004–1.690) 0.046

Time to treatment, mo, median (IQR) 6(1.5–16) 4.5 (1–10.7) 1.003 (0.967–1.041) 0.859

No. of relapses before treatment, median (IQR) 1(0.2–2) 1 (0–2) 1.061 (0.749–1.503) 0.739

FU, y, median (IQR) 4(3–6) 5 (2–12) 1.1109 (1.009–1.218) 0.032

Abbreviations: CI = CI for OR calculated for nominal variables; EDSS = Expanded Disability Status Scale; FU = follow-up; IQR = interquartile range; OCB= oligoclonal band.

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN cognitive deficits. Cognitive impairment had not been sys- likely reflects the more inflammatory disease seen in the tematically addressed in pediatric NMOSD, and our results younger patients.31 Of interest, in comparison to a cohort of highlight its importance. Although no objective cognitive children with MOG-Ab–associated disease,32 despite the measure was used in this study, we selected measures that are higher ARR before treatment, the ARR on treatment was consistent across the diverse educational systems that reflect lower across all treatments in the AQP4-Ab NMOSD group. clear difficulties and thus would be a reasonable surrogate for cognitive performance. ON at onset was a risk for visual im- Our cohort was not optimal for a direct evaluation of an pairment and blindness in keeping with previous report that individual treatment or a sequence of treatment effect, which young children do not recover well following ON.16 Of in- is better suited to a study design where the sequence of terest, time to starting treatment and number of relapses DMTs, the lag phase of efficacy or washout period of specific before starting treatment were not associated with worst therapies are prospectively controlled. One particular treat- outcome. As previously reported,24 non-White ethnicity was ment that deserves specific attention would be the cumulative associated with a more aggressive disease course but was not use of corticosteroids, often used in conjunction with DMTs predictive of moderate disabilities at last follow-up. and at low doses, but also during acute attacks. Time to initiation of acute treatment, the choice of treatment, and Patients were treated according to each center experience, and treatment escalation are major predictors of long-term out- – the choices of treatment were additionally influenced by the come in AQP4-Ab NMOSD.33 35 As many of the children access to treatments in each country and their related costs. presented initially to district general hospital, timing, dura- Almost all of the Brazilian children were commenced on AZA, tion, and sequence of acute treatment were not included in and none were given rituximab as first-line treatment. In the analysis. contrast, the European children were frequently commenced on rituximab and MMF (table 3). A previous pediatric co- A major limitation of this study is that not all patients were hort10 (only 20/58 patients were treated with DMTs) systematically managed, with possible biases in treatment reported residual disability in 43/48 (90%), with 26/48 initiation and/or escalation. Nevertheless, these real-world (54%) having visual impairment and 21/48 (44%) with motor clinical data from multiple countries allowed us to make im- deficits. Comparing to this cohort, the visual but not the portant observations about the treatment responsiveness of motor outcome is similar, which might be related to children with AQP4-Ab NMOSD. The important question treatment. our study raises is whether rituximab should be first-line treatment for children. Moreover, should children who are A key observation from our study was that all treatments already stable on AZA and MMF be changed to the more knowntoreducerelapsesinadults(AZA,MMF,andritux- efficacious therapies with a higher risk profile? Furthermore, imab) were associated with a reduction in the ARR. More- with new monoclonal antibodies (satralizumab,36 inebilizu- over, rituximab treatment resulted in the lowest ARR, and it mab,37 and eculizumab38) likely to become available, treat- is remarkable that all 14 patients who started on rituximab as ment algorithms that accommodate risk stratification of first line did not have any further relapses. Treatment esca- poorer prognosis (such as the patients age and severity of the lation from AZA and MMF to rituximab was beneficial in 12/ attack) and grouping of treatments with similar efficacy within 13, with only 1 child maintaining the same ARR (figure 2). the sequence and choice of treatments in pediatric AQP4-Ab The benefit of early treatment with rituximab was previously NMOSD are urgently required. demonstrated in a pediatric cohort with a range of CNS disorders including NMOSD.25 This apparent superiority Acknowledgment could be partially explained by the early treatment effect of The authors thank all collaborators who worked in this study in rituximab. In addition, in our cohort, patients were also the collection of data: Austria and Germany: M. Reindl, Clinical initiated with rituximab sooner compared with AZA/MMF Department of Neurology, Medical University of Innsbruck, (1 month vs 6–10 months) after the attack. CD19 count was Innsbruck, Austria. Brazil: Dr. Dagoberto Callegaro (Depart- not performed in all 7 patients who relapsed with rituximab. ment of Neurology, Hospital das Cl´ınicas, Faculty of Medicine, Monitoring B-cell repopulation and redosing rituximab São Paulo, Brazil). Laboratory of Brain Institute of Rio Grande might prevent relapses as demonstrated in a pediatric do Sul (Brains), Pontifical Catholic University of Rio Grande Do study.26 Three patients treated with rituximab changed Sul (PUCRS), Porto Alegre, Brazil. France: Dr. Maurey, therapy due to adverse effect, and none had severe infection. Assistance Publique-Hopitauxˆ de Paris, University Hospitals Prospective follow-up was not performed to check on Paris South, Bicetreˆ Hospital, Pediatric Neurology Department, hypogammaglobulinemia.27 Le Kremlin Bicetre,ˆ Dr. Pascale Chrétien, University Hospitals Paris South, Bicetreˆ Hospital, Immunology Department, Le Although we did not perform a direct comparison, the treat- Kremlin Bicetre,ˆ Dr. Aubart, Assistance Publique-Hopitauxˆ de ment effect observed in this pediatric cohort appeared greater Paris, Necker Hospital, Paris, Dr. Cheuret, Hopitalˆ des Enfants, – than reported in adults with AQP4-Ab NMOSD.6,28 30 The Pediatric Neurology Department, Toulouse, Dr. Lepine, Centre greater reduction in the ARR following initiation of DMTs is Hopitalo-Universitaire de Marseille, Marseille, Pr Laugel, CHU also reported in children with MS compared with adults and de Strasbourg - Hopitalˆ de Hautepierre, Pediatric Neurology

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 9 Department, Strasbourg. Spain: Dr. Maria Sepulveda (Neuro- Serono; and was awarded educational grants to organize immunology Program, Hospital Clinic-IDIBAPS, University of meetings by Novartis, Biogen Idec, Merck Serono, and Bayer. Barcelona, Spain), Dr. Victor Soto-Insuga (Pediatric Neurology A. Saiz reports no disclosures. R. Neuteboom received con- Section, Pediatric Service, Fundacion Jimenez D´ıaz, Madrid, sultation fees from Novartis and Zogenix. R. Marignier serves Spain), Maria C. Miranda-Herrero (Pediatric Neurology on the scientific advisory board for Novartis and MedImmune Section, Pediatric Service, Hospital Gregorio Marañon, Madrid, and received speaker honoraria and travel funding from Spain), Montserrat Garcia-Puig (Pediatric Neurology Section, Novartis, Biogen, Teva, Sanofi-Aventis/Genzyme, and Merck. Pediatric Service, Hospital Parc Taul´ı, Sabadell, Spain). The C. Hemingway received consulting and educational grants Netherlands: Prof. Rogier Hintzen, MD, PhD, was involved in from Novartis; served on the Global Paediatric MS Steering this study before his sudden death. Dörte Hamann and Kyra committee and received educational grants from Biogen; and Gelderman from the Department of Immunopathology and received consultation fees from Alexion. D.K. Sato received Blood Coagulation, Sanquin Diagnostic Services, Amsterdam, research support from CAPES/Brasil, CNPq/Brasil, the Netherlands. AQP4 antibodies were routinely assayed in a FAPERGS/Brasil, and Japan Society of Science and re- cell-based assay using a fluorescence-activated cell sorter (FACS, ceived speaker honoraria from Biogen, Novartis, Genzyme, LSRII, and DIVA software, Becton Dickinson, San Jose). United Merck Sereno, Quest/Athena Diagnosis, Shire, and Teva. K. Kingdom: UK Childhood Inflammatory Disorders (UK-CID) Deiva received speaking honoraria at meetings and for par- Study Group: Antonia Clarke, MD (Department of Paediatric ticipation in advisory boards from Novartis, Sanofi, and Neurology, St George’s Hospital, London), Katherine Forrest, Servier. Go to Neurology.org/NN for full disclosures. MD (Department of Paediatric Neurology, Southampton Children’s Hospital, Southampton), Andrex Lux, MD (De- Publication history partment of Paediatric Neurology, University Hospitals Bristol, Received by Neurology: Neuroimmunology & Neuroinflammation Bristol), Saleel Chandratre, MD (Department of Paediatric March 2, 2020. Accepted in final form April 28, 2020. Neurology, John Radcliffe Hospital, Oxford), Kshitij Mankad, MD, and Kling Chong, MD (Paediatric Neuroradiology, Great Ormond Street Hospital, London). Laboratory: Paddy Water, PhD, Mark Woodhall, PhD, and Jackie Palace, MD, PhD Appendix Authors (Oxford Autoimmune Neurology Group and National Health Name Location Contribution

Service National Specialised Commissioning Group for Neuro- Renata Department of Neurology, Analyzed data and myelitis Optica, United Kingdom). Barbosa Hospital das Cl´ınicas, drafted the manuscript Paolilo, MD Faculty of Medicine, for intellectual content University of São Paulo Study funding (HCFMUSP), São Paulo, This study was supported in part by the Marato TV3 Foun- Brazil dation (20141830 to T.A. and A.S.), Red Española de Escle- Yael Hacohen, Queen Square MS Centre, Designed and rosis Múltiple (REEM) (RD16/0015/0002; RD16/0015/ MD, PhD UCL Institute of Neurology, conceptualized the study; Faculty of Brain Sciences, analyzed data; and 0003 to A.S.), and Torrons Vicens Foundation (PFNR0144 University College London. drafted the manuscript to TA). Department of Paediatric for intellectual content Neurology, Great Ormond Street Hospital for Disclosure Children, London, United R.B. Paolilo received support for participating in scientific Kingdom meetings from Roche and Merck. Y. Hacohen and E. Yazbeck Elise Yazbeck, Assistance Publique- Revised the manuscript MD Hopitauxˆ de Paris, for intellectual content report no disclosures. T. Armangue received speaking hono- University Hospitals Paris raria from Novartis and travel expenses for participating in South, Bicêtre Hospital, Le ˆ scientific meetings from Roche. A. Bruijstens and C. Lechner Kremlin Bicetre, France report no disclosures. S.L. Apostolos-Pereira served as board Thais Neuroimmunology Revised the manuscript advisory for Novartis and Roche and received support for Armangue, Program, Institut for intellectual content MD, PhD d’Investigacions participating in scientific meetings from Genzyme, Roche, and Biomèdiques August Pi i Novartis. Y. Martynenko, M. Breu, C. de Medeiros Rimkus, E. Sunyer (IDIBAPS), Hospital Cl´ınic, Universitat de Wassmer, M. Baumann, and L. Papetti report no disclosures. Barcelona, Barcelona, M. Capobianco received personal honoraria for speaking at Spain. Pediatric Neuroimmunology Unit, meeting or participating in advisory boards from Biogen, Sant Joan de Deu (SJD) Merck, Novartis, Roche, Sanofi, and Teva. B. Kornek, K. Children’s Hospital, Universitat de Barcelona, Rostásy, and J.A. da Paz report no disclosures. O. Ciccarelli Barcelona, Spain receives research funding from the UK MS Society, Rosetrees Arlette Department of Neurology, Revised the manuscript Trust, and NIHR UCLH BRC and serves as a consultant for Bruijstens, Erasmus University for intellectual content Biogen, Novartis, Roche, and GE Healthcare; she is on the MD Medical Center, Rotterdam, the editorial board of Neurology. M. Lim received consultation Netherlands fees from CSL Behring; received travel grants from Merck

10 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Appendix (continued) Appendix (continued)

Name Location Contribution Name Location Contribution

Christian Division of Paediatric Revised the manuscript Olga Queen Square MS Centre, Interpreted the data and Lechner, MD Neurology, Department of for intellectual content Ciccarelli, MD, UCL Institute of Neurology, revised the manuscript Pediatrics I, Medical PhD Faculty of Brain Sciences, for intellectual content University of Innsbruck, University College London, Innsbruck, Austria Ming Lim, MD, Children’s Neurosciences, Interpreted the data and Samira Luisa Department of Neurology, Revised the manuscript PhD Evelina London Children’s revised the manuscript Apostolos- Hospital das Cl´ınicas, for intellectual content Hospital at Guy’s and St for intellectual content Pereira, MD, Faculty of Medicine, Thomas’ NHS Foundation PhD University of São Paulo Trust, King’s Health (HCFMUSP), São Paulo, Partners Academic Health Brazil Science Centre, London, United Kingdom; Faculty of Yana Medical Center of Physical Revised the manuscript Life Sciences and Martynenko, Therapy and Pain for intellectual content Medicine, King’s College MD Medicine INNOVO, Lviv, Hospital, London, United Ukraine Kingdom

Markus Breu, Division of Pediatric Revised the manuscript Albert Saiz, Neurology Service, Revised the manuscript MD Neurology, Department of for intellectual content MD, PhD Hospital Clinic, and Institut for intellectual content Pediatrics and Adolescent d’Investigacions Medicine, Medical Biom`ediques August Pi i University of Vienna, Sunyer (IDIBAPS), Vienna, Austria University of Barcelona, Barcelona, Spain Carolina de Department of Radiology Revised the manuscript Medeiros (INRAD), Hospital das for intellectual content Rinze Department of Neurology, Revised the manuscript Rimkus, MD, Cl´ınicas, Faculty of Neuteboom, Erasmus University for intellectual content PhD Medicine, University of MD, PhD Medical Center, São Paulo (HCFMUSP), São Rotterdam, the Paulo, Brazil Netherlands

Evangeline Pediatric Neurology, Revised the manuscript Romain Neurology Department, Interpreted the data and Wassmer, MD Birmingham Women and for intellectual content Marignier, Hopitalˆ Neurologique revised the manuscript Children’s Hospital, MD, PhD Pierre Wertheimer, for intellectual content Birmingham, United Hospices Civils de Lyon, Kingdom Lyon; French Reference Network of Rare Matthias Division of Paediatric Revised the manuscript Inflammatory Brain and Baumann, MD Neurology, Department of for intellectual content Spinal Diseases, Le Pediatrics I, Medical Kremlin Bicêtre, France University of Innsbruck, Innsbruck, Austria Cheryl Department of Paediatric Interpreted the data and Hemingway, Neurology, Great Ormond revised the manuscript Laura Papetti, Department of Revised the manuscript MD, PhD Street Hospital for for intellectual content PhD Neuroscience, Pediatric for intellectual content Children, London, United Multiple Sclerosis Center, Kingdom Bambino Gesu` Children Hospital, IRCCS, Rome, Douglas School of Medicine, Brain Interpreted the data and Italy Kazutoshi Institute of Rio Grande do drafted the manuscript Sato, MD, PhD Sul (Brains), Pontifical for intellectual content Marco Department of Neurology Revised the manuscript Catholic University of Rio Capobianco, and Regional Multiple for intellectual content Grande Do Sul (PUCRS), MD Sclerosis Centre, Porto Alegre, Brazil University Hospital San Luigi Gonzaga, Orbassano, Kumaran Assistance Publique- Designed and Italy Deiva, MD, Hopitauxˆ de Paris, conceptualized the study; PhD University Hospitals Paris analyzed data; and Barbara Department of Neurology, Revised the manuscript South, Bicêtre Hospital, Le drafted the manuscript Kornek, MD, Medical University of for intellectual content Kremlin Bicêtre, France. for intellectual content PhD Vienna, Vienna, Austria French Reference Network of Rare Inflammatory Kevin Rostasy,´ Department of Pediatric Revised the manuscript Brain and Spinal Diseases, MD Neurology, Vestische for intellectual content Le Kremlin Bicêtre, France. Kinder und Jugendklinik, Inserm UMR 1184, Witten/Herdecke Immunology of Viral University, Datteln, Infections and Germany Autoimmune Diseases, CEA, IDMIT, Le Kremlin José Albino da Department of Neurology, Revised the manuscript Bicêtre, France Paz, MD, PhD Hospital das Cl´ınicas, Faculty for intellectual content of Medicine, University of São Paulo (HCFMUSP), São Paulo, Brazil

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 11 References 20. Waters PJ, Pittock SJ, Bennett JL, Jarius S, Weinshenker BG, Wingerchuk DM. – 1. Lennon VA, Wingerchuk DM, Kryzer TJ, et al. A serum autoantibody marker of neu- Evaluation of aquaporin-4 antibody assays. Clin Exp Neuroimmunol 2014;5:290 303. romyelitis optica: distinction from multiple sclerosis. Lancet 2004;364:2106–2112. 21. Höftberger R, Sepulveda M, Armangue T. Antibodies to MOG and AQP4 in adults 2. Sato DK, Nakashima I, Takahashi T, et al. Aquaporin-4 antibody-positive cases be- with neuromyelitis optica and suspected limited forms of the disease. Mult Scler 2015; – yond current diagnostic criteria for NMO spectrum disorders. Neurology 2013;80: 21:1086 1094. 2210–2216. 22. van Pelt ED, Wong YY, Ketelslegers IA. Neuromyelitis optica spectrum disorders: 3. Wingerchuk DM, Banwell B, Bennett JL, et al. International consensus diagnostic comparison of clinical and magnetic resonance imaging characteristics of AQP4-IgG – criteria for neuromyelitis optica spectrum disorders. Neurology 2015;85:177–189. versus MOG-IgG seropositive cases in the Netherlands. Eur J Neurol 2015;23:580 587. 4. Rostasy K, Reindl M. Role of autoantibodies in acquired inflammatory demyelinating 23. Ketelslegers IA, Modderman PW, Vennegoor A. Antibodies against aquaporin-4 in diseases of the central nervous system in children. Neuropediatrics 2013;44:297–301. neuromyelitis optica: distinction between recurrent and monophasic patients. Mult – 5. Tillema JM, McKeon A. The spectrum of neuromyelitis optica (NMO) in childhood. Scler 2011;17:1527 1530. ff J Child Neurol 2012;27:1437–1447. 24. Kim SH, Mealy MA, Levy M, et al. Racial di erences in neuromyelitis optica spectrum – 6. Mealy MA, Wingerchuk DM, Palace J, Greenberg BM, Levy M. Comparison of disorder. Neurology 2018;91:e2089 e2099. ff relapse and treatment failure rates among patients with neuromyelitis optica: multi- 25. Dale RC, Brilot F, Du y LV, et al. Utility and safety of rituximab in pediatric auto- fl – center study of treatment efficacy. JAMA Neurol 2014;71:324–330. immune and in ammatory CNS disease. Neurology 2014;83:142 150. 7. Palace J, Leite MI, Nairne A, Vincent A. Interferon beta treatment in neuromyelitis 26. Nosadini M, Alper G, Riney CJ, et al. Rituximab monitoring and redosing in pediatric fl optica: increase in relapses and aquaporin 4 antibody titers. Arch Neurol 2010;67: neuromyelitis optica spectrum disorder. Neurol Neuroimmunol Neuroin amm 2016; 1016–1017. 3:e188. 8. Banwell B, Tenembaum S, Lennon VA, et al. Neuromyelitis optica-IgG in childhood 27. Marcinno A, Marnetto F, Valentino P, et al. Rituximab-induced hypo- inflammatory demyelinating CNS disorders. Neurology 2008;70:344–352. gammaglobulinemia in patients with neuromyelitis optica spectrum disorders. Neurol fl 9. Huppke P, Bluthner M, Bauer O, et al. Neuromyelitis optica and NMO-IgG in Neuroimmunol Neuroin amm 2018;5:e498. ffi European pediatric patients. Neurology 2010;75:1740–1744. 28. Costanzi C, Matiello M, Lucchinetti CF, et al. Azathioprine: tolerability, e cacy, and fi – 10. McKeon A, Lennon VA, Lotze T, et al. CNS aquaporin-4 autoimmunity in children. predictors of bene t in neuromyelitis optica. Neurology 2011;77:659 666. Neurology 2008;71:93–100. 29. Huh SY, Kim SH, Hyun JW, et al. Mycophenolate mofetil in the treatment of neu- – 11. Chitnis T, Ness J, Krupp L, et al. Clinical features of neuromyelitis optica in children: romyelitis optica spectrum disorder. JAMA Neurol 2014;71:1372 1378. US Network of Pediatric MS Centers report. Neurology 2016;86:245–252. 30. Kim SH, Jeong IH, Hyun JW, et al. Treatment outcomes with rituximab in 100 fl 12. Banwell B, Bar-Or A, Arnold DL, et al. Clinical, environmental, and genetic deter- patients with neuromyelitis optica: in uence of FCGR3A polymorphisms on the – minants of multiple sclerosis in children with acute demyelination: a prospective therapeutic response to rituximab. JAMA Neurol 2015;72:989 995. national cohort study. Lancet Neurol 2011;10:436–445. 31. Gorman MP, Healy BC, Polgar-Turcsanyi M, Chitnis T. Increased relapse rate in 13. Hacohen Y, Absoud M, Woodhall M, et al. Autoantibody biomarkers in childhood- pediatric-onset compared with adult-onset multiple sclerosis. Arch Neurol 2009;66: – acquired demyelinating syndromes: results from a national surveillance cohort. 54 59. J Neurol Neurosurg Psychiatry 2014;85:456–461. 32. Hacohen Y, Wong YY, Lechner C, et al. Disease course and treatment responses in 14. Hacohen Y, Mankad K, Chong WK, et al. Diagnostic algorithm for relapsing acquired children with relapsing myelin oligodendrocyte glycoprotein antibody-associated – demyelinating syndromes in children. Neurology 2017;89:269–278. disease. JAMA Neurol 2018;75:478 487. 15. Collongues N, Marignier R, Zephir H, et al. Long-term follow-up of neuromyelitis 33. Kleiter I, Gahlen A, Borisow N, et al. Neuromyelitis optica: evaluation of 871 attacks – optica with a pediatric onset. Neurology 2010;75:1084–1088. and 1,153 treatment courses. Ann Neurol 2016;79:206 216. 16. Kitley J, Leite MI, Nakashima I, et al. Prognostic factors and disease course in 34. Kleiter I, Gahlen A, Borisow N, et al. Apheresis therapies for NMOSD attacks: a retrospective fl aquaporin-4 antibody-positive patients with neuromyelitis optica spectrum disorder study of 207 therapeutic interventions. Neurol Neuroimmunol Neuroin amm 2018;5:e504. from the United Kingdom and Japan. Brain 2012;135:1834–1849. 35. Bonnan M, Valentino R, Debeugny S, et al. Short delay to initiate plasma exchange is 17. Krupp LB, Tardieu M, Amato MP, et al. International Pediatric Multiple Sclerosis the strongest predictor of outcome in severe attacks of NMO spectrum disorders. – Study Group criteria for pediatric multiple sclerosis and immune-mediated central J Neurol Neurosurg Psychiatry 2018;89:346 351. nervous system demyelinating disorders: revisions to the 2007 definitions. Mult Scler 36. Yamamura T, Kleiter I, Fujihara K, et al. Trial of satralizumab in neuromyelitis optica – 2013;19:1261–1267. spectrum disorder. N Engl J Med 2019;381:2114 2124. 18. Marignier R, Bernard-Valnet R, Giraudon P. Aquaporin-4 antibody-negative neuromyelitis 37. Cree BAC, Bennett JL, Kim HJ, et al. Inebilizumab for the treatment of neuromyelitis optica: distinct assay sensitivity-dependent entity. Neurology 2013;80:2194–2200. optica spectrum disorder (N-MOmentum): a double-blind, randomised placebo- – 19. Sato DK, Callegaro D, Lana- Peixoto MA, et al. Distinction between MOG antibody- controlled phase 2/3 trial. Lancet 2019;394:1352 1363. positive and AQP4 antibody-positive NMO spectrum disorders. Neurology 2014;82: 38. Pittock SJ, Berthele A, Fujihara K, et al. Eculizumab in aquaporin-4-positive neuro- – 474–481. myelitis optica spectrum disorder. N Engl J Med 2019;381:614 625.

12 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN ARTICLE OPEN ACCESS Eﬀect of the sphingosine-1-phosphate receptor modulator ozanimod on leukocyte subtypes in relapsing MS

Sarah Harris, PhD, Jonathan Q. Tran, PharmD, Harry Southworth, PhD, MSc, Collin M. Spencer, BS, Correspondence Bruce A.C. Cree, MD, PhD, MAS, and Scott S. Zamvil, MD, PhD Dr. Harris [email protected] Neurol Neuroimmunol Neuroinﬂamm 2020;7:e839. doi:10.1212/NXI.0000000000000839 Abstract Objective To better understand ozanimod’s mechanism of action (MOA), we conducted exploratory analyses from a phase 1 study to characterize ozanimod’seﬀect on circulating leukocyte subsets in patients with relapsing multiple sclerosis.

Methods An open-label pharmacodynamic study randomized patients to oral ozanimod hydrochloride (HCl) 0.5 (n = 13) or 1 mg/d (n = 11) for ;12 weeks (including 7-day dose escalation). Circulating leukocyte subsets were quantiﬁed using ﬂow cytometry (days 28, 56, and 85) and epigenetic cell counting (days 2, 5, 28, 56, and 85) and compared with baseline (day 1) using descriptive statistics.

Results Ozanimod caused dose-dependent reductions in absolute lymphocyte counts. Observed by both methodologies, circulating CD19+ B- and CD3+ T-cell counts were reduced by >50% with ozanimod HCl 0.5 mg and >75% with 1 mg at day 85. Based on flow cytometry, ozanimod HCl 1 mg showed greater decreases in CD4+ than CD8+ T cells, greater decreases in both CD4+ and CD8+ central memory vs effector memory T cells, and reductions in mean CD4+ and CD8+ naive T cells by ≥90% at day 85. In the flow cytometry analysis, changes in monocytes, natural killer, and natural killer T cells were minimal. Using epigenetic cell counting, greater reductions for Th17 than T regulatory cells were determined.

Conclusion Ozanimod induced dose-dependent reductions in circulating B- and T-cell counts and differential effects on naive and memory CD4+ and CD8+ T cells and CD19+ B cells. Data characterized with both a novel epigenetic cell-counting method and flow cytometry support ozanimod’s MOA.

Clinical trial registration: clinicaltrials.gov NCT02797015.

From Bristol-Myers Squibb Company (S.H., J.Q.T.), Princeton, NJ; Data Clarity Consulting Ltd. (H.S.), Stockport, England, United Kingdom; and Department of Neurology, Weill Institute for Neurosciences (C.M.S., B.A.C.C., S.S.Z.), and Program in Immunology (C.M.S., S.S.Z.), University of California San Francisco (UCSF).

Dr. Tran is currently with Mirati Therapeutics, Inc., San Diego, CA.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

The Article Processing Charge was funded by Bristol-Myers Squibb Company. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary ALC = absolute lymphocyte count; HCl = hydrochloride; NK = natural killer; MOA = mechanism of action; RMS = relapsing MS; S1P = sphingosine-1-phosphate; SLO = secondary lymphoid organ; TEMRA = terminally diﬀerentiated eﬀector T cells expressing CD45RA; VZV = varicella-zoster virus.

Ozanimod, a sphingosine-1-phosphate (S1P) receptor 1 and Patients 5 modulator, is approved in the United States for the treat- Adults aged 18–55 years with RMS, as diagnosed by the re- ment of adults with relapsing forms of multiple sclerosis (MS) vised 2010 McDonald criteria13 and exhibiting a relapsing and in Europe for the treatment of adults with relapsing- clinical course and a history of brain MRI lesions consistent remitting MS. Ozanimod was effective and well tolerated in with RMS, were enrolled. Eligible participants had no history phase 21,2 and phase 3 clinical trials of relapsing MS (RMS).3,4 of relapse with onset from 30 days before screening until The mechanism by which ozanimod exerts therapeutic effects randomization, were clinically stable during this period in MS is unknown but may involve reduced lymphocyte mi- without systemic corticosteroid or adrenocorticotropic hor- gration into the CNS.5 mone treatment, and had documentation of positive varicella- zoster virus (VZV) immunoglobulin G (IgG) antibody status By reducing lymphocyte egress from secondary lymphoid or complete VZV vaccination at least 30 days before study organs (SLOs), S1P receptor modulators decrease the entry. In addition, they were required to have an Expanded peripheral blood absolute lymphocyte count (ALC).5 The Disability Status Scale score of 0–6 and be generally healthy chemokine receptor CCR7 directs lymphocytes into SLOs, aside from RMS. Key exclusion criteria included active in- and data suggest that CCR7+ lymphocyte subpopulations fection or history of chronic infections or immunodeficiency, are responsive to S1P modulators.6 Studies of fingolimod, recent live vaccination, previous lymphocyte-depleting or the first approved S1P receptor modulator and a modula- immunosuppressant therapy, and ALC <1.000 × 109/L or – tor of receptors 1, 3, 4, and 5,7 9 indicated differential white blood cell count <3.500 × 109/L. effects on specific T- and B-cell subtypes. The differential effects of fingolimod on lymphocyte subtypes are being Standard protocol approvals, registrations, – evaluated as possible predictors of clinical response.6,8 11 and patient consent Clinical trials of ozanimod reported expected decreases in The phase 1 study was approved by an institutional review ALCs3,4 and differential effects on specific lymphocyte board and was designed and monitored in compliance with subtypes in healthy volunteers, with CCR7+ Tcells the principles of Good Clinical Practice as required by regu- (CD4+,CD8+, and central memory T cells) preferentially latory authorities and in accordance with the Declaration of decreased.12 Helsinki. All participants provided written informed consent. This study is registered on ClinicalTrials.gov (identifier: To improve the understanding of the mechanism of action NCT02797015). (MOA) of ozanimod in patients with RMS, exploratory analyses from a phase 1 study were conducted to characterize Data availability the phenotype of circulating leukocyte subsets in patients with Celgene, a Bristol-Myers Squibb Company, is committed to RMS treated with ozanimod using both flow cytometry and responsible and transparent sharing of clinical trial data with epigenetic cell-counting methodologies. patients, health care practitioners, and independent researchers for the purpose of improving scientific and medical knowledge as well as fostering innovative treatment approaches. Data re- Methods quests may be submitted to Celgene, a Bristol-Myers Squibb Company, at vivli.org/ourmember/celgene/ and must include Study design a description of the research proposal. A phase 1 randomized (1:1), open-label, multiple-dose, parallel-group pharmacodynamic study of ozanimod hydro- Flow cytometry analysis chloride (HCl) 0.5 or 1 mg/d (equivalent to ozanimod 0.46 As a prespecified pharmacodynamic analysis, the ALC was or 0.92 mg, respectively) was conducted in participants with evaluated on days 1, 5, 8, 28, 56, and 85 (end of treatment). As an RMS at 6 study centers in the United States. Participants exploratory analysis, a flow cytometry panel was used to char- were randomized to receive ozanimod HCl 0.5 or 1 mg/d for acterize circulating leukocyte subsets at baseline (day 1) and days approximately 12 weeks, which included an initial 7-day dose 28, 56, and 85. Analyzed subsets included CD19+ Bcells,CD3+ escalation consisting of ozanimod HCl 0.25 mg/d (equiva- T cells, monocytes, natural killer (NK) cells, and natural killer T lent to ozanimod 0.23 mg) on days 1–4 and 0.5 mg/d on days (NKT) cells, as well as the following T-cell subtypes: CD4+ and 5–7. All participants who completed the study were eligible CD8+ central and effector memory T cells, CD4+ and CD8+ to enter an open-label extension study (DAYBREAK; naive T cells, and CD8+ terminally differentiated effector T cells NCT02576717). expressing CD45RA (TEMRA).

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Figure 1 ALC during ozanimod treatment in patients with RMS

Error bars represent ±1 SD. aLLN established by the reference laboratory for the study. bDose escalation consisted of ozanimod HCl 0.25 mg/d on days 1–4, 0.5 mg/d on days 5–7, and then assigned dose (0.5 or 1 mg/d) for approximately 12 weeks. ALC = absolute lymphocyte count; HCl = hydrochloride; LLN = lower limit of normal; RMS = relapsing MS.

Heparinized whole blood samples were shipped overnight at Epigenetic cell-counting analysis ambient temperature from clinical sites to the University of Epigenetic cell counting was performed by Epiontis/Precision California San Francisco, where peripheral blood mononuclear for Medicine, as previously described,14 using bisulfite- cells were isolated using a standard Ficoll/density gradient converted DNA from frozen whole blood samples as sub- centrifugation protocol. Cells were stained for surface antigens strate for quantitative PCR assays for selected cell type–specific with mouse anti-human monoclonal antibodies (CD3, CD4, demethylated loci (table e-1, links.lww.com/NXI/A283). CD8, CD19, CD20, CD14, CD56, CD16, CD45RA, CD45RO, Briefly, 75 μL of EDTA-anticoagulated peripheral blood was CCR7, CCR4, CCR6, and CXCR3; BioLegend, San Diego, CA) supplemented with 67 μLoflysisbuffer comprising 54.25 μLof and LIVE/DEAD Fixable Aqua viability dye (Thermo Fisher Lysis/Binding Buffer (Invitrogen), 9 μL of proteinase K Scientific), then fixed and permeabilized using the True-Nuclear (Sigma; 30 mg/mL), and 3.75 μL of GAP[GC] plasmid (to a Buffer Kit (BioLegend). A single lot of each antibody was used final concentration of 20,000 copies per μL of whole blood) throughout the study, and each sample was tested once. Ac- and lysed for 15 minutes at 56°C. For conversion, 270 μLof quisition was performed on a Becton Dickinson LSRFortessa 70% ammonium bisulfite solution was added, and samples were cell analyzer with FACSDiva software (version 8.0). All data incubated at 80°C for 55 minutes. Bisulfite-converted DNA was were analyzed using FlowJo version 10.3. A single investigator, directly purified using the MyONE Silane genomic DNA kit who was blinded to the treatment arm, conducted all analyses. (Invitrogen) according to the manufacturer’s instructions.

Figure 2 Circulating levels of B cells (A) and T cells (B) during ozanimod treatment

Circulating levels of CD19+ B cells and CD3+ T cells were assessed as a percentage of baseline in patients with relapsing MS treated with ozanimod HCl 0.5 or 1 mg/d, using flow cytometry. HCl = hydrochloride.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 Table 1 Circulating levels of leukocyte subtypes over time with ozanimod HCl 0.5 or 1 mg/d expressed as a percentage of baseline cell count, as assessed using flow cytometry

Cell type Ozanimod HCl 0.5 mg Ozanimod HCl 1 mg

Geometric mean (95% CI) % of baseline Day 28 (n = 11) Day 56 (n = 12) Day 85 (n = 12) Day 28 (n = 8) Day 56 (n = 6) Day 85 (n = 9)

CD19+ B cells 46.3 (23.8, 90.1) 63.9 (42.8, 95.3) 42.2 (25.8, 68.9) 34.8 (23.5, 51.4) 18.0 (11.5, 28.2) 17.6 (11.2, 27.6)

CD3+ T cells 44.0 (26.4, 73.4) 48.9 (30.6, 78.1) 39.0 (24.9, 61.2) 31.4 (23.8, 41.3) 20.4 (13.5, 30.7) 18.5 (10.5, 32.5)

CD4+ T cells 39.6 (22.7, 69.0) 44.9 (26.3, 76.6) 33.4 (19.7, 56.5) 26.3 (16.8, 41.1) 10.2 (5.9, 17.8) 8.7 (4.2, 18.1)

CD8+ T cells 46.5 (26.4, 81.9) 47.0 (28.6, 77.2) 45.4 (29.8, 69.3) 41.6 (31.9, 54.3) 36.2 (21.7, 60.4) 31.9 (17.5, 58.3)

CD4+ central memory T cells 43.6 (25.3, 75.1) 44.5 (26.0, 76.1) 37.3 (23.7, 58.6) 57.6 (16.2, 205.0) 22.5 (6.9, 73.6) 17.3 (5.3, 56.6)

CD4+ effector memory T cells 35.0 (18.9, 64.7) 43.9 (23.2, 82.9) 31.3 (19.6, 49.9) 64.5 (20.2, 206.7) 54.8 (17.3, 173.8) 36.1 (10.3, 126.2)

CD8+ central memory T cells 41.3 (21.0, 81.2) 41.9 (19.6, 89.8) 45.7 (28.9, 72.2) 43.8 (22.2, 86.4) 10.1 (4.0, 25.5) 28.5 (13.2, 61.8)

CD8+ effector memory T cells 56.7 (32.3, 99.7) 62.4 (30.9, 125.8) 69.1 (51.9, 92.0) 85.0 (54.0, 133.7) 70.4 (35.1, 141.1) 81.8 (50.6, 132.4)

CD4+ naive T cells 33.6 (19.8, 57.1) 36.0 (19.9, 65.3) 26.0 (13.7, 49.5) 37.1 (10.1, 136.9) 12.6 (2.9, 54.7) 10.1 (2.5, 39.9)

CD8+ naive T cells 30.8 (18.5, 51.2) 34.6 (19.4, 61.5) 25.3 (13.9, 46.0) 16.3 (10.6, 25.0) 5.4 (3.5, 8.3) 6.7 (2.9, 15.0)

CD8+ TEMRA 53.6 77.4 (40.1, 149.2) 63.9 95.6 (70.4, 130.0) 105.2 62.2 (20.0, 193.7) (26.4, 108.9) (38.2, 106.9) (62.5, 177.1)

CD14+ monocytes 114.0 211.0 165.6 222.4 172.7 194.9 (38.7, 335.5) (108.6, 410.1) (73.9, 371.2) (113.3, 436.8) (83.8, 356.3) (150.0, 253.3)

CD3+ CD56+ NKT 76.7 105.1 87.5 79.8 (42.9, 148.3) 82.7 (58.1, 117.6) 65.0 (25.6, 165.2) (38.7, 151.8) (61.0, 181.1) (50.9, 150.4)

2 CD16 CD56+ NK 72.1 87.3 (58.4, 130.5) 63.3 128.5 (77.9, 211.9) 101.5 91.4 (57.0, 146.5) (43.4, 119.6) (33.2, 120.7) (58.3, 176.6)

2 CD16 CD56hi NK 27.1 (8.4, 83.4) 55.8 (27.7, 111.3) 24.4 (10.9, 53.3) 11.5 (3.2, 36.4) 33.8 (21.8, 51.9) 16.5 (4.4, 56.3)

2 CD16+ CD56 NK 108.9 138.9 151.7 418.8 344.0 303.5 (51.6, 229.7) (69.9, 276.2) (62.4, 369.0) (164.1, 1069.2) (186.5, 634.6) (158.3, 581.8)

CD16+ CD56+ NK 80.5 154.2 114.4 91.6 (40.1, 209.2) 159.1 94.7 (43.4, 206.8) (33.4, 193.9) (96.4, 246.5) (70.9, 184.5) (117.5, 215.4)

CD16+CD56hi NK 30.9 (7.6, 117.4) 72.8 (37.2, 141.2) 25.9 (9.8, 66.4) 20.5 (5.8, 66.6) 62.0 (36.4, 105.1) 20.3 (5.0, 74.8)

Abbreviations: HCl = hydrochloride; n = number of observed values; NK = natural killer cells; NKT = natural killer T cells; TEMRA = terminally differentiated effector T cells expressing CD45RA. Where the data contained values of 0, 1 was added to all values before log transformation and subtracted when converting back to the original scale. Percentage of baseline for CD4+ and CD8+ cells is based on % of CD3+ T cells.

Bisulﬁte conversion rates were analyzed previously and are ALC. Circulating leukocyte subset counts were compared with provided in the manufacturer’s manual with values above 98%. baseline (day 1) using descriptive statistics. The geometric Thermal cycling was performed as follows: 1 × 95°C for 35 mean, used to report a percentage of the baseline cell count, is minutes, followed by 50 × 95°C for 15 seconds, and 61°C for 1 the mean of the logged values, then transformed back to the minute in 10 μL using the Roche LightCycler 480 Probes original scale; the 95% CIs were likewise computed on the log Master. For the calculation of cell numbers from autosomal scale and then transformed back. All analyses were performed genes, a 2:1 allele-to-cell ratio was assumed.14 using R version 3.6.1 (2019-07-05) (R Core Team 2018). Statistical analysis This was a clinical pharmacology study without formal hy- Results pothesis testing. A sample size of approximately 12 subjects per dose group was deemed adequate to characterize the phar- Patients macodynamics of ozanimod. ALC, ﬂow cytometry, and epi- A total of 24 participants were randomized to ozanimod HCl genetic cell-counting data were summarized using the number 0.5 mg (n = 13) or 1 mg (n = 11); the mean (SD) age was 38.8 of observed (nonmissing) observations, without imputation; (8.4) years. Participants were predominantly women (70.8%) however, patients with a missing baseline (day 1) ALC value and White (75.0%; 5 participants [20.8%] were Black and 1 used the ALC value from the screening visit as their baseline [4.2%] was Asian). All participants completed treatment,

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Figure 3 Flow cytometry analysis of circulating levels of central vs effector memory T cells

Circulating levels of (A) CD4+ central memory T cells, (B) CD4+ effector memory T cells, (C) CD8+ central memory T cells, and (D) CD8+ effector memory T cells during treatment with ozanimod HCl 0.5 or 1 mg/d, as assessed using flow cytometry. HCl = hydrochloride. except 1 participant in the ozanimod HCl 0.5 mg group who ozanimod treatment were primarily due to decreases in cir- withdrew from the study after day 8 (figure e-1, links.lww.com/ culating CD19+ B cells and CD3+ T cells (figure 2, A and B; NXI/A282). The first participant enrolled on July 21, 2016, and table 1). There were minimal to no decreases in monocytes, the last participant completed the study on October 20, 2017. NK, and NKT cells (table 1).

Course of ALC Further analysis of specific T-cell subtypes revealed greater + + The mean (SD) ALC at baseline was 1.754 (0.489) × 109/L in decreases in CD4 T-helper cells than CD8 cytotoxic T cells the ozanimod HCl 0.5 mg group and 2.118 (0.961) × 109/L in in the ozanimod HCl 1 mg group (table 1), as well as greater + + the ozanimod HCl 1 mg group. After dose escalation, ozani- decreases in both CD4 and CD8 central memory T cells vs mod was associated with dose-dependent reductions in ALCs effector memory T cells (figure 3 and table 1). By the end of + (figure 1). At the end of treatment (day 85), the mean (SD) treatment, ozanimod HCl 1 mg reduced mean CD4 and + ALC was 1.055 (0.594) × 109/L (42.2% reduction from CD8 naive T cells by ≥ 90%; ozanimod did not reduce + baseline) in the ozanimod HCl 0.5 mg group and 0.536 circulating CD8 TEMRA (table 1). (0.262) × 109/L (73.3% reduction from baseline) in the ozanimod HCl 1 mg group. Epigenetic cell-counting analysis In the epigenetic cell-counting analysis, total circulating leukocytes at day 85 were reduced to 90% (95% CI: 78%, 104%) Effect of ozanimod on circulating and 73% (95% CI: 55%, 97%) of baseline in the ozanimod leukocyte subsets HCl 0.5 and 1 mg groups, respectively. Results for specific Flow cytometry analysis leukocyte subsets were consistent with the flow cytometry Flow cytometry analysis of circulating leukocyte subsets in- results in showing dose-dependent decreases in total circu- dicated that the dose-dependent decreases in ALCs with lating B cells and T cells with ozanimod treatment, as well as

Table 2 Circulating levels of leukocyte subtypes over time with ozanimod HCl 0.5 or 1 mg/d expressed as a percentage of baseline cell count, as assessed using epigenetic cell counting

Ozanimod HCl 0.5 mg Ozanimod HCl 1 mg

Cell type Day 2 Day 5 Day 28 Day 56 Day 85 Day 2 Day 5 Day 28 Day 56 Day 85

CD19+ B cells, n 11 13 12 12 10 11 10 11 10 10

Gmean (95% CI) 84.7 (75.8, 94.5) 76.6 (66.4, 88.3) 45.1 (34.7, 58.6) 40.8 (32.8, 50.7) 38.2 (25.9, 56.4) 89.4 (76.8, 104.2) 79.5 (69.6, 90.8) 24.6 (20.4, 29.7) 19.3 (15.4, 24.1) 17.0 (14.2, 20.2)

CD3+ T cells, n 68 88843444

Gmean (95% CI) 84.0 (74.8, 94.5) 74.5 (61.4, 90.5) 50.5 (39.3, 64.8) 32.2 (20.7, 49.9) 34.6 (20.1, 59.5) 96.2 (79.8, 116.1) 90.6 (82.3, 99.9) 19.4 (11.1, 33.8) 12.1 (5.6, 26.0) 10.6 (6.2, 18.3)

CD4+ T cells, n 10 13 12 12 11 11 10 11 11 11

Gmean (95% CI) 86.1 (76.2, 97.4) 84.3 (73.2, 97.1) 63.4 (54.7, 73.4) 44.1 (36.3, 53.5) 46.1 (32.5, 65.5) 97.4 (84.9, 111.9) 96.6 (84.2, 110.7) 26.7 (21.5, 33.2) 21.4 (16.9, 27.0) 22.5 (18.8, 27.1)

CD8+ T cells, n 11 13 12 12 11 11 10 10 10 10

Gmean (95% CI) 89.9 (76.9, 105.2) 79.0 (68.5, 91.1) 53.1 (45.5, 62.1) 43.4 (34.1, 55.3) 39.3 (29.1, 53.0) 91.5 (79.3, 105.6) 88.5 (79.5, 98.5) 32.2 (23.3, 44.6) 28.8 (20.3, 40.9) 24.5 (18.2, 33.0)

CD8+ naive 55 44377777 T cells, n

Gmean (95% CI) 79.8 (60.6, 105.0) 92.7 (75.7, 113.5) 45.2 (26.1, 78.5) 37.5 (25.3, 55.8) 38.4 (34.3, 42.9) 78.3 (64.1, 95.7) 78.9 (68.3, 91.3) 18.5 (12.9, 26.7) 16.2 (7.9, 33.1) 12.6 (7.3, 21.6)

Th17 cells, n 11 13 12 12 11 11 10 11 11 11

Gmean (95% CI) 88.2 (80.3, 96.9) 89. 0 (78.7, 100.6) 51.6 (43.5, 61.2) 38.1 (28.2, 51.5) 38.0 (25.9, 55.8) 94.7 (81.9, 109.5) 87.1 (76.6, 98.9) 20.2 (15.4, 26.5) 13.7 (9.9, 18.9) 11.5 (8.6, 15.4)

Treg cells, n 68 66743332

Gmean (95% CI) 98.2 (79.0, 122.0) 74.9 (61.3, 91.6) 62.3 (50.2, 77.3) 54.3 (41.7, 70.9) 51.7 (36.0, 74.4) 103.9 (91.5, 118.0) 86.9 (61.8, 122.3) 45.7 (18.7, 111.4) 35.3 (21.9, 57.1) 34.6 (14.5, 82.6)

PD-1+ cells, n 55 44377777

Gmean (95% CI) 86.8 (65.5, 115.1) 98.2 (70.1, 137.6) 57.7 (36.8, 90.3) 62.9 (49.5, 79.8) 76.2 (45.4, 128.0) 78.3 (64.8, 94.7) 81.0 (71.2, 92.3) 35.3 (26.0, 47.9) 35.2 (25.8, 47.9) 34.2 (26.5, 44.3)

Abbreviations: Gmean = geometric mean; HCl = hydrochloride; n = number of observed values; PD-1+ = programmed cell death 1. Figure 4 Epigenetic cell counting: circulating levels of Th17 vs T regulatory cells

Circulating levels of (A) Th17 cells and (B) Treg cells during treatment with ozanimod HCl 0.5 or 1 mg/d, as assessed using epigenetic cell counting. HCl= hydrochloride. decreases in CD4+ T-helper and CD8+ cytotoxic T cells and fingolimod in patients with MS and with siponimod15 and naive CD8+ T cells (table 2). This analysis also demonstrated ponesimod16,17 in healthy volunteers. greater reductions in Th17 cells than T regulatory (Treg) cells or programmed cell death 1 (PD-1)+ cells (figure 4 and Our flow cytometry analysis further demonstrated that ozani- table 2). mod had a minimal effect on circulating levels of other leukocyte subtypes, including NK and NKT cells and monocytes, suggesting a differential effect of treatment on specificleukocyte Discussion subpopulations. This has potential clinical implications, given the role of these subtypes in the innate immune response necessary – These data represent the first reported analyses of the effect of for immunosurveillance against infections and tumors.18 20 One ozanimod on leukocyte subsets in patients with RMS. In ad- of the concerns surrounding the use of immunomodulatory dition to flow cytometry, a novel epigenetic cell-counting therapies is the potential for increased risks of infection and method was used, and both methodologies support the MOA malignancy. However, phase 3 clinical trials of ozanimod showed of ozanimod and differential effects on peripheral blood leu- infection rates that were comparable with patients treated with kocyte subsets. Specifically, this phase 1 study confirmed that IM interferon beta-1a, and infrequent serious infections, no se- treatment with ozanimod produced dose-dependent reduc- rious opportunistic infections, and low (<1%) rates of malig- tions in ALCs over 12 weeks in patients with RMS. Results nancy among patients with RMS receiving ozanimod.3,4 from the exploratory analyses demonstrated that this reduction was due primarily to decreases in T and B cells. In the flow Our flow cytometry results were further supported by results cytometry analysis, the greatest T-cell reductions were seen in from a second analysis based on epigenetic cell counting, a CD4+ naive and central memory subsets, whereas CD8+ ef- recently developed method of immune cell quantification. fector memory T cells and TEMRA were not reduced from This method provides relative and absolute immune cell baseline (table 1). This profile is consistent with the expected counts based on PCR amplification and quantification of cell ff – fi 14,21,22 e ects of S1P1 receptor modulation, given the link between type speci c DNA methylation markers. In a previous – chemokine receptor CCR7 expression and S1P1 receptor analysis, results from epigenetic cell counting were found to mediated egress of lymphocytes from lymph nodes; CCR7 is correlate highly (Spearman rank correlation coefficients of expressed on B cells and both naive and central memory T cells, 0.96–0.98, p < 0.001) with flow cytometry results in both but not on effector memory T cells or TEMRA.9 The differ- healthy controls and HIV+ individuals for all cell types ential effects on CCR7+ lymphocyte subsets observed in the studied.14 Although correlations could not be meaningfully current study are in agreement with those reported previously estimated in our analysis due to small sample sizes and with ozanimod in healthy volunteers.12 Moreover, this profile missing data, results from the epigenetic cell-counting analysis mirrors patterns of lymphocyte effects reported with were generally consistent with flow cytometry results for the

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 7 leukocyte subtypes analyzed by both methods, specifically manuscript), and Lorraine Sweeney (proofreading), of Peloton substantial decreases in B and T cells, including CD4+ and Advantage, LLC, an OPEN Health company, and funded by CD8+ T cells and CD8+ naive T cells (tables 1 and 2). Epi- Bristol-Myers Squibb Company. genetic cell-counting results also demonstrated that ozanimod is associated with greater reductions in proinflammatory Study funding Th17 cells23 compared with Treg cells. This has potential This study was sponsored by Celgene Corporation. clinical implications in that Th17/Treg balance is believed to – be an important factor in MS disease activity.24 27 In- Disclosure terestingly, a small study assessing circulating levels of Th17 S. Harris and J.Q. Tran are employees of Bristol-Myers Squibb and Treg after fingolimod initiation for RMS reported variable Company. H. Southworth received compensation from Aptus effects on Th17 levels, in which 11 of 21 (52.4%) patients Clinical, BresMed Health Solutions, Celgene, CytomX, F2G, showed an increased proportion of Th17 cells after 4 weeks of GlaxoSmithKline, and Gossamer Bio. C.M. Spencer reports no treatment, and increased levels of Th17 cells were associated disclosures. B.A.C. Cree received personal compensation from with MS relapses after fingolimod initiation.28 Akili, Alexion, Atara, Biogen, EMD Serono, Novartis, and TG Therapeutics. S.S. Zamvil served on data safety monitoring Finally, these analyses confirmed the dose-dependent effects of boards of Opexa, BioMS, Teva, and Eli Lilly; is a member of the ozanimod on circulating levels of leukocyte subtypes, in which clinical advisory board of the Myelin Repair Foundation; re- ozanimod HCl 1 mg clearly produced greater reductions in T ceived speaker honoraria from Biogen and Teva; is a deputy and B cells than ozanimod HCl 0.5 mg. This dose dependence editor of Neurology® Neuroimmunology & Neuroinflammation; can be seen in both flow cytometry and epigenetic cell-counting consulted for Biogen, Teva, EMD Serono, Genzyme, Novartis, results. In addition, these cell counts showed continued de- and Roche; serves on the speaker’s bureau of Advanced Health creases over time in the ozanimod HCl 1 mg group, whereas a Media and Biogen; and received research support from Biogen, time-dependent effect was less evident with the ozanimod HCl Teva, the NIH, the NMSS, the Weill Institute, and Alexander 0.5 mg group. These dose-related effects on cell counts are M. and June L. Maisin Foundation. Go to Neurology.org/NN consistent with the dose-related clinical findings from the phase for full disclosures. 3 RMS studies, in which ozanimod HCl 1 mg achieved a nu- merically lower annualized relapse rate and fewer brain MRI Publication history lesions compared with ozanimod HCl 0.5 mg.3,4 Received by Neurology: Neuroimmunology & Neuroinflammation March 13, 2020. Accepted in final form June 17, 2020. Although providing valuable information for the more detailed understanding of the MOA of ozanimod and avenues for further research, the analyses reported here were explor- Appendix Authors

atory in nature. The phase 1 study forming the basis of these Name Location Contribution analyses enrolled a small number of patients (N = 24), and Sarah Harris, Bristol-Myers Squibb Interpretation of the data even smaller numbers of samples were available for analysis PhD Company, Princeton, NJ and revised the for some of the subtypes. Owing to differences in the data sets, manuscript for intellectual no formal statistical comparisons correlating results from flow content cytometry vs epigenetic cell counting were made. Neverthe- Jonathan Q. Bristol-Myers Squibb Design of the clinical study, Tran, Company, Princeton, NJ, at interpretation of the data, less, consistency in results between the 2 methods suggests PharmD the time this study was and revised the the utility of the newer epigenetic cell-counting methodology conducted manuscript for intellectual and provides greater confidence in the overall cell type– content specificeffects of ozanimod treatment. This observation Harry Data Clarity Consulting Analysis of the data and suggests that a cross-validation study of epigenetic cell- Southworth, Ltd, Stockport, England, revised the manuscript for PhD, MSc United Kingdom intellectual content counting and flow cytometry methods could be pursued. Collin M. Department of Neurology, Design of the flow Spencer, BS Weill Institute for cytometry study, These analyses support that, in line with its MOA, ozanimod Neurosciences, and acquisition and analysis of produced dose-dependent reductions in B- and T-cell counts Program in Immunology the data, and participated in peripheral blood of patients with RMS. A differential effect UCSF, San Francisco, CA in writing the manuscript of ozanimod on specific leukocyte subtypes in patients with Bruce A. C. Department of Neurology, Design of the flow Cree, MD, Weill Institute for cytometry study, RMS was observed. PhD, MAS Neurosciences, UCSF, acquisition of the data, San Francisco, CA interpretation of the data, and participated in writing Acknowledgment the manuscript The authors thank Epiontis/Precision for Medicine for conducting the epigenetic cell counting. Under guidance from Scott S. Department of Neurology, Design of the flow Zamvil, MD, Weill Institute for cytometry study, the authors, support for this manuscript was provided by Marci PhD Neurosciences, and interpretation of the data, Mikesell (help with drafting the manuscript), Jennifer Program in Immunology and participated in writing UCSF, San Francisco, CA the manuscript Fernandez, PhD, and Jessica Herr, PharmD (editing the

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN References 14. Baron U, Werner J, Schildknecht K, et al. Epigenetic immune cell counting in human blood samples for immunodiagnostics. Sci Transl Med 2018;10:eaan3508. 1. Cohen JA, Arnold DL, Comi G, et al. Safety and efficacy of the selective sphingosine 15. Gergely P, Nuesslein-Hildesheim B, Guerini D, et al. The selective sphingosine 1-phosphate receptor modulator ozanimod in relapsing multiple sclerosis (RADIANCE): 1-phosphate receptor modulator BAF312 redirects lymphocyte distribution and has a randomised, placebo-controlled, phase 2 trial. Lancet Neurol 2016;15:373–381. species-specificeffects on heart rate. Br J Pharmacol 2012;167:1035–1047. 2. Cohen JA, Comi G, Arnold DL, et al. Efficacy and safety of ozanimod in multiple 16. D’Ambrosio D, Steinmann J, Brossard P, Dingemanse J. Differential effects of sclerosis: dose-blinded extension of a randomized phase II study. Mult Scler 2019;25: ponesimod, a selective S1P1 receptor modulator, on blood-circulating human T cell 1255–1262. subpopulations. Immunopharmacol Immunotoxicol 2015;37:103–109. 3. Cohen JA, Comi G, Selmaj KW, et al. Safety and efficacy of ozanimod versus in- 17. Jurcevic S, Juif PE, Hamid C, Greenlaw R, D’Ambrosio D, Dingemanse J. Effects of terferon beta-1a in relapsing multiple sclerosis (RADIANCE): a multicentre, rand- multiple-dose ponesimod, a selective S1P1 receptor modulator, on lymphocyte omised, 24-month, phase 3 trial. Lancet Neurol 2019;18:1021–1033. subsets in healthy humans. Drug Des Devel Ther 2017;11:123–131. 4. Comi G, Kappos L, Selmaj KW, et al. Safety and efficacy of ozanimod versus in- 18. Kaur G, Trowsdale J, Fugger L. Natural killer cells and their receptors in multiple terferon beta-1a in relapsing multiple sclerosis (SUNBEAM): a multicentre, rando- sclerosis. Brain 2013;136:2657–2676. mised, minimum 12-month, phase 3 trial. Lancet Neurol 2019;18:1009–1020. 19. Wolf BJ, Choi JE, Exley MA. Novel approaches to exploiting invariant NKT cells in 5. Scott FL, Clemons B, Brooks J, et al. Ozanimod (RPC1063) is a potent sphingosine- cancer immunotherapy. Front Immunol 2018;9:384. 1-phosphate receptor-1 (S1P1) and receptor-5 (S1P5) agonist with autoimmune 20. Abel AM, Yang C, Thakar MS, Malarkannan S. Natural killer cells: development, disease-modifying activity. Br J Pharmacol 2016;173:1778–1792. maturation, and clinical utilization. Front Immunol 2018;9:1869. 6. Mehling M, Brinkmann V, Antel J, et al. FTY720 therapy exerts differential effects on 21. Wieczorek G, Asemissen A, Model F, et al. Quantitative DNA methylation analysis of T cell subsets in multiple sclerosis. Neurology 2008;71:1261–1267. FOXP3 as a new method for counting regulatory T cells in peripheral blood and solid 7. Mandala S, Hajdu R, Bergstrom J, et al. Alteration of lymphocyte trafficking by tissue. Cancer Res 2009;69:599–608. sphingosine-1-phosphate receptor agonists. Science 2002;296:346–349. 22. Singh A, Yamamoto M, Ruan J, et al. Th17/Treg ratio derived using DNA methyl- 8. Muls N, Dang HA, Sindic CJ, van Pesch V. Fingolimod increases CD39-expressing ation analysis is associated with the late phase asthmatic response. Allergy Asthma regulatory T cells in multiple sclerosis patients. PLoS One 2014;9:e113025. Clin Immunol 2014;10:32. 9. Teniente-Serra A, Hervas JV, Quirant-Sanchez B, et al. Baseline differences in 23. Gonzalez H, Pacheco R. T-cell-mediated regulation of neuroinflammation involved in minor lymphocyte subpopulations may predict response to fingolimod in relapsing- neurodegenerative diseases. J Neuroinflammation 2014;11:201. remitting multiple sclerosis patients. CNS Neurosci Ther 2016;22:584–592. 24. Shen H, Bonner JA, Shi LZ. Metabolic checkpoints in neurodegenerative T helper 17 10. Quirant-Sanchez B, Hervas-Garcia JV, Teniente-Serra A, et al. Predicting therapeutic (TH17) and neuroregenerative regulatory T (Treg) cells as new therapeutic targets response to fingolimod treatment in multiple sclerosis patients. CNS Neurosci Ther for multiple sclerosis. Neural Regen Res 2020;15:267–269. 2018;24:1175–1184. 25. Rostami A, Ciric B. Role of Th17 cells in the pathogenesis of CNS inflammatory 11. Kaufmann M, Haase R, Proschmann U, Ziemssen T, Akgun K. Real world lab data: demyelination. J Neurol Sci 2013;333:76–87. patterns of lymphocyte counts in fingolimod treated patients. Front Immunol 2018;9: 26. Brucklacher-Waldert V, Stuerner K, Kolster M, Wolthausen J, Tolosa E. Phenotypical and 2669. functional characterization of T helper 17 cells in multiple sclerosis. Brain 2009;132:3329–3341. 12. Tran JQ, Hartung JP, Peach RJ, et al. Results from the first-in-human study with 27. Liu J, Mori M, Sugimoto K, et al. Peripheral blood helper T cell profiles and their ozanimod, a novel, selective sphingosine-1-phosphate receptor modulator. J Clin clinical relevance in MOG-IgG-associated and AQP4-IgG-associated disorders and Pharmacol 2017;57:988–996. MS. J Neurol Neurosurg Psychiatry 2020;91:132–139. 13. Polman CH, Reingold SC, Banwell B, et al. Diagnostic criteria for multiple sclerosis: 28. Sato DK, Nakashima I, Bar-Or A, et al. Changes in Th17 and regulatory T cells after 2010 revisions to the McDonald criteria. Ann Neurol 2011;69:292–302. fingolimod initiation to treat multiple sclerosis. J Neuroimmunol 2014;268:95–98.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 9 ARTICLE OPEN ACCESS CLASS OF EVIDENCE Tailoring B cell depletion therapy in MS according to memory B cell monitoring

Giovanni Novi, MD, Francesca Bovis, PhD, Sabrina Fabbri, MD, Francesco Tazza, MD, Paola Gazzola, MD, Correspondence Ilaria Maietta, Daniela Curro,` MD, Nicolo` Bruschi, MD, Luca Roccatagliata, MD, PhD, Giacomo Boffa, MD, Dr. Novi [email protected] Caterina Lapucci, MD, Giampaola Pesce, PhD, Maria Cellerino, MD, Claudio Solaro, MD, Alice Laroni, MD, PhD, Elisabetta Capello, MD, Gianluigi Mancardi, MD, Mariapia Sormani, PhD, Matilde Inglese, MD, PhD,* and Antonio Uccelli, MD*

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e845. doi:10.1212/NXI.0000000000000845

Abstract MORE ONLINE Objective Class of Evidence We wanted to evaluate efficacy on inflammatory parameters of rituximab (RTX)-personalized Criteria for rating reinfusion scheme using a memory B cell–based treatment regimen. therapeutic and diagnostic studies Methods NPub.org/coe This is a prospective, uncontrolled, open-label study including patients with MS treated with RTX in 2 Italian MS units. All patients were treated with RTX induction, followed by maintenance infusion at the dosage of 375 mg/m2, according to memory B cell repopulation (0.05% of peripheral-blood mononuclear cells [PBMCs] for the first 2 years, 0.1% of PBMC for the third year). MS activity was assessed as clinical or MRI activity.

Results One hundred two patients were included in the analysis. Mean follow-up was 2.40 years (range 0.57–7.15 years). The annualized relapse rate (ARR) was 0.67 in the year before RTX start and decreased to 0.01 in the 3 years after RTX initiation (global ARR). The proportion of patient with MS activity (i.e., relapse or MRI activity) was 63.16% in the year before RTX start and decreased to 8.7% (0–6 months), 1.3% (6–12 months), 0% (12–24 months), and 0% (24–36 months). Annualized RTX infusion rates were 1.67 (95% conﬁdence interval [CI]: 1.43–1.94), 0.76 (95% CI: 0.58–0.98), and 0.78 (95% CI: 0.52–1.12) for the ﬁrst 3 years after RTX initiation, respectively. Patients were reinfused with a mean infusion interval of 367 days (range 181–839 days).

Conclusion The results of this study show that the memory B cell–based RTX reinfusion protocol is able to reduce the mean number of RTX reinfusions with persistent reduction of disease activity.

Classification of evidence This study provides Class IV evidence that for patients with MS, a memory B cell–based RTX reinfusion protocol can reduce the mean number of RTX reinfusions with persistent reduction of disease activity.

*Both last coauthors equally contributed to the manuscript.

From the Department of Neuroscience (G.N.), Ospedale Policlinico San Martino—IRCCS; Department of Health Sciences (DISSAL) (F.B., I.M., L.R., M.S.), University of Genova, Italy; Ospedale A. Micone (S.F., P.G.), Genova; Department of Neuroscience (F.T., N.B., G.B., C.L., M.C., A.L., E.C., G.M.), Rehabilitation, Ophthalmology, Genetics, Maternal and Child Health (DINOGMI), University of Genova; Ospedale Policlinico San Martino—IRCCS (F.T., N.B., L.R., A.L., E.C., M.S., M.I., A.U.), Genova, Italy; Ospedale San Paolo (D.C.), Savona; Autoimmunity Laboratory DiMI (G.P.), University of Genova, Italy; Monsignor Luigi Novarese Rehabilitation Center (C.S.), Moncrivello, Vercelli; Istituti Clinici Scientifici Maugeri (G.M.), IRCCS, Pavia; and Department of Neuroscience (M.I., A.U.), Rehabilitation, Ophthalmology, Genetics, Maternal and Child Health (DINOGMI) and Center of Excellence for Biomedical Research (CEBR), University of Genova, Italy.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary ARIR = annualized reinfusion rate; ARR = annualized relapse rate; CI = conﬁdence interval; EDSS = expanded disability status scale; mAb = monoclonal antibody; OCR = ocrelizumab; PBMC = peripheral blood mononuclear cell; PP = primary progressive; RR = relapsing-remitting; RTX = rituximab; SE = standard error; SP = secondary progressive.

The MS therapeutic field has been recently widened by the Standard protocol approvals, registrations, approval of ocrelizumab (OCR) treatment as the first anti- and patient consents CD20-depleting monoclonal antibody (mAb).1 Rituximab The local ethic committee approved treatment regimen and (RTX), a first-generation anti-CD20 mAb, has also been data collection, and patients signed written informed consent adopted as an off-label treatment in MS,2,3 and it is currently before treatment initiation. used as standard of care therapy in some European countries.4 The standard treatment regimen of anti-CD20 mAbs usually Patients consists of an “induction” phase, followed by regular fixed Patients were treated with RTX with two 1-g infusions 15 days maintenance reinfusions (usually every 6 months). However, apart as loading doses. Patients were then followed up quar- despite being a more practical approach in the daily practice, terly with memory B cell evaluation (assessed as CD19+ and the fixed doses regimen could represent an overtreatment CD27+ cells). MRI assessment was performed within 6 because B cells could be still depleted before each subsequent months of RTX initiation, followed by additional scans at the retreatment dose, as B cell immune reconstitution after B cell end of each treatment year. depletion ranges from 27 to 125 weeks with a median of 72 weeks.5 In addition, no data support the fact that resurgence Treatment 2 (and/or normalization) of CD19+ B cells is strictly associated Patients were reinfused with 375 mg/m RTX when the fi with an inflammatory activity (i.e., clinical relapse or MRI percentage of memory B cells exceeded the prede ned rein- ff activity). fusion cuto : 0.05% of peripheral blood mononuclear cells (PBMC) for the first 2 years and 0.1% of PBMC for the third Furthermore, a subgroup of B cells called memory B cells year with subsequent doubling for each year of treatment ff (characterized by CD19 and CD27 co-expressions) have been (maximum cuto at the 7th year of treatment of 1.6% of recently implied as a putative target of many MS-approved PBMC). A year-by-year increase in the threshold for reinfu- treatments (including CD20-depleting mAbs).6 Peripherical sion was adopted to further reduce the number of RTX blood memory B cell dosage has been extensively adopted in reinfusions with each year of treatment. neuromyelitis optica to tailor RTX redosing with consistent – results.7 10 Statistical analysis The Annualized relapse rate (ARR), defined as the total Consequently, evaluating peripheral blood memory B cells number of relapses divided by the total number of patient resurgence to tailor RTX retreatment in MS might optimize years, pre- and post-RTX start, and the annualized reinfusion rate (ARIR) after RTX initiation were compared by mixed- RTX redosing, reducing the number of infusions, possibly ff maintaining consistent efficacy on MRI and relapse activity, e ect negative binomial models accounting for the repeated and potentially reducing risks of adverse events. To test our measures analysis, with p-values adjusted for multiple testing hypothesis, we conducted a pilot study in 2 MS centers in Italy by the Bonferroni correction. SAS 9.3 (Institute Inc., Cary, to assess efficacy on inflammatory parameters (i.e., MRI ac- NC) and R software (version 3.5.0) were used for the tivity and clinical relapses) of memory B cells–tailored RTX computation. redosing in patients with MS. Data availability Raw data are available on appropriate request. Methods We designed a proof-of-concept, uncontrolled, single-arm, Results open-label, prospective study where we enrolled patients with One hundred two patients were enrolled in the study: 34 MS who were referred to our clinic and were treated, with an patients (33.33%) had a relapsing-remitting (RR) phenotype, off-label indication, with RTX, since 2012. Database was 29 (28.43%) a primary progressive (PP) phenotype, and 39 locked in November 2019. (38.24%) a secondary progressive (SP) phenotype. Eighty- two percent of RRMS individuals and 52% of progressive The primary research question was to evaluate efficacy on patients (regardless of a primary or secondary phenotype) inflammatory parameters of RTX-personalized reinfusion displayed disease activity at MRI carried out in the year before scheme using a memory B cell–based treatment regimen. RTX start. Complete demographic analysis is reported in

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN table. At database lock, the patient mean follow-up was 2.40 in the third year, with a 3-year global ARR of 0.01 (95% CI: years (range 0.57–7.15 years). The annualized relapse rate 0.002–0.03) (p < 0.0001) (figure 1). was 0.43 (95% confidence interval [CI]: 0.35–0.53) in the 2 years before RTX initiation and increased to 0.67 (95% CI: Of 60 patients (63.16%) with MRI evidence of activity (either 0.52–0.84) in the year before RTX initiation. As expected, Gd+ enhancing lesions or new enlarging T2/fluid attenuated ARR was dramatically reduced on RTX initiation to 0.01 inversion recovery lesions) in the year before RTX initiation, (95% CI: 0.001–0.04) in the first year, 0.01 (95% CI: 28 patients (26.32%) had a RR phenotype, whereas the 0.001–0.06) in the second year, and 0.00 (95% CI: 0.00–0.10) remaining 32 (33.68%) had a progressive phenotype.

Table Baseline characteristics for 102 patients with MS treated with RTX, grouped by MS subtype

All RR PP SP

N=102 N=34 N=29 N=39

Age at MS onset, y, mean (SD) 30.90 (10.25) 28.71 (7.38) 39.29 (9.98) 26.58 (9.02)

Gender (female), n (%) 67 (65.69) 21 (61.76) 18 (62.07) 28 (71.79)

MS duration, y, mean (SD) 11.29 (8.61) 7.68 (6.97) 6.63 (5.47) 17.92 (7.71)

Comorbidity, n (%) 37 (36.27) 15 (44.12) 8 (27.59) 14 (35.90)

Autoimmune comorbidity, n (%) 9 (8.82) 3 (8.82) 2 (6.90) 4 (10.26)

EDSS 6 mo pre-RTX, median (range) 5(0–8.5) 2.5 (0–7) 5 (1.5–6.5) 6 (1.5–8.5)

EDSS at RTX start, median (range) 5(1–8.5) 3 (1.0–7.5) 5 (1.5–6.5) 6.5 (3–8.5)

No. relapse 2 y pre-RTX, median (range) 1(0–4) 1.5 (0–4) 0 (0–3) 0 (0–2)

No. relapse 1 y pre-RTX, median (range) 0(0–3) 1 (0–3) 0 (0–2) 0 (0–2)

Presence of active MRI 1 y pre-RTX, n (%) 60/95 (63.16) 28 (82.35) 12/23 (52.17) 20/38 (52.63)

Presence of spinal lesions 1 y pre-RTX, n (%) 37/90 (41.11) 23 (67.65) 7/18 (38.89) 7/38 (18.42)

Na¨ıve patients, n (%) 18 (17.65) 7 (20.59) 9 (31.03) 2 (5.13)

Previous treatments, median (range) 2(0–7) 2 (1–5) 1 (1–4) 3 (1–7)

Last DMT before RTX, n (%) 84 (82.35) 27 (79.41) 20 (68.97) 37 (94.87)

Cyclophosphamide 6 (7.14) 1 (3.70) 2 (10.00) 3 (8.11)

Daclizumab 1 (1.19) 1 (3.70) 0 (0.00) 0 (0.00)

Dimethyl fumarate 5 (5.95) 2 (7.41) 3 (15.00) 0 (0.00)

Fingolimod 30 (35.71) 10 (37.04) 3 (15.00) 17 (49.95)

Glatiramer acetate 8 (9.52) 3 (11.11) 1 (5.00) 4 (10.81)

Interferons 11 (13.10) 4 (14.81) 3 (15.00) 4 (10.81)

Natalizumab 11 (13.10) 3 (11.11) 2 (10.00) 6 (16.22)

Teriflunomide 2 (2.38) 1 (3.70) 1 (5.00) 0 (0.00)

Other 10 (11.90) 2 (7.41) 5 (25.00) 3 (8.11)

Reasons for changing to RTX, n (%) 84 (82.35) 27 (79.41) 20 (68.97) 37 (94.87)

Inefficacy 54 (64.29) 19 (70.37) 15 (75.00) 20 (54.05)

Intolerance/adverse event 11 (13.10) 2 (7.41) 4 (20.00) 5 (13.51)

JCV+ 8 (9.52) 2 (7.41) 1 (5.00) 5 (13.51)

Pregnancy 1 (1.19) 1 (3.70) 0 (0.00) 7 (18.92)

Other 10 (11.90) 3 (11.11) 0 (0.00) 0 (0.00)

Abbreviations: DMT = disease-modifying treatment; EDSS = expanded disability status scale; JCV+ = positive for John Cunningham virus; PP = primary progressive; RR = relapsing remitting; RTX = rituximab; SP = secondary progressive.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 Figure 1 Trend of annualized relapse rate in the 2 years preceding rituximab treatment and in the 3 years after rituximab initiation

Such a proportion declined to 8.70% (8/92 patients) in the 0.62–1.20) in the second year, and 0.78 (95% CI: 0.52–1.12) first 6 months after therapy initiation (4 RR patients and 4 in the third year (figure 2). progressive patients) and then to 1.28% (1/78 PP patient) in the subsequent 6 months (till month 12), 0% (0/59 patients) The median number of memory B cells before reinfusions was in the second year, and 0% in the third year (0/32 patients). 2 cells/mm3 (range 1–53 cells/mm3) in the first year, 2 cells/ To date, none of the patients with a follow-up longer than 3 mm3 (range 1–21 cells/mm3) in the second year, and 3 cells/ years experienced relapses or disease activity during the sub- mm3 (range 1–18 cells/mm3) in the third year of treatment. sequent years of follow-up. The median (range) baseline expanded disability status scale Nine of 102 patients (8.82%) (4 RR, 4 PP, and 1 SP) had (EDSS) score was 3.0 (1.0–7.5) in patients with RRMS, 6.5 inflammatory activity (defined as relapse and/or MRI ac- (3.0–8.5) in patients with SPMS, and 5.0 (1.5–6.5) in patients tivity) during the first 6 months after therapy initiation, with PPMS. The median time to progression was 1.65 years probably as a consequence of a carryover activity from (0.38–7.15). previous disease-modifying treatments (mainly natalizumab or fingolimod) and because of delayed onset of RTX max- The proportion of patients with a 6-month confirmed EDSS imum efficacy. progression after 3 years from RTX start was 8.82% (standard error [SE] = 0.02) in the RRMS group, 17.95% (SE = 0.05) in Interestingly, ARIR was consistently reduced year by year: the SPMS group, and 48.28% (SE = 0.09) in the PPMS group. during the first year, ARIR was 1.67 (95% CI: 1.43–1.94), 0.76 (95% CI: 0.58–0.98) in the second year, and 0.78 (95% CI: Patients were reinfused with a mean infusion interval of 367 0.52–1.12) in the third year (p < 0.0001). The results are days (range 181–839 days). consistent with those obtained analyzing only the 41 patients who have completed at least 3 years of follow-up: ARIR was During the follow-up, 14/102 (13.72%) experimented an 1.75 (95% CI: 1.38–2.19) in the first year, 0.87 (95% CI: infusion-related reaction (IRR), all occurring during the first 3

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN classical schemes with a significant saving of drug and sub- Figure 2 Trend of annualized rituximab reinfusion rate in sequent economical advantage possibly arising from the re- the 3 years after rituximab initiation for the whole duction of the yearly infused drug and from the decrement of population enrolled (blue) and for the 41 patients the use of healthcare system resources. who have completed at least 3 years of follow-up (red) For patients with progressive MS (either SP or PP), our findings suggest that our regimen could reduce the chance of disability accrual through a reduction in the inflammatory activity. However, similar to other approved CD20- depleting therapies (i.e., OCR), it is reasonable to specu- late that this treatment regimen will not be able to completely abolish disability progression driven by ongoing neurodegeneration.

The main limit of our study is represented by the lack of a control group that is required for disability progression analysis and is warranted for the comparison of composite outcomes, such as no evidence of disease activity. In addition, a control group is needed to compare the adverse event’s occurrence (including hypogammaglobulinemia), although a lower intensity regimen of B cell depletion might warrant a safer risk proﬁle. To date, many patients completed at least 2 years of treatment without emergence of MS activity, and therefore, it is reasonable that a reinfusion cutoﬀ of 0.05% memory B cell is safe and potentially applicable to everyday clinical practice.

Finally, we hope that our data might prompt the conduction infusions, and 75/102 (73.53%) patients developed AE and of randomized clinical trial to assess the efficacy and safety of 19 patients developed a serious AE (see supplementary ap- tailored reinfusion schemes for CD20-depleting mAbs. pendix 1, links.lww.com/NXI/A290). At the 2-year follow-up, among patients with known immunoglobulin levels (n = 46), Study funding hypogammaglobulinemia (immunoglobulin G levels lower No targeted funding reported. than 5.6 g/L) developed in 2 patients, whereas for 2 patients, it was pre-existent and one patient developed hypo- Disclosure gammaglobulinemia after 6 years of treatment. One patient G. Novi received speaker honoraria from Merck, Novartis, died during the follow-up because of an undifferentiated and Roche. F. Bovis reports no disclosures relevant to the mediastinal tumor lesion. manuscript. S. Fabbri reports no disclosures relevant to the manuscript. F. Tazza reports no disclosures relevant to the Discussion manuscript. P. Gazzola reports no disclosures relevant to the manuscript. I. Maietta reports no disclosures relevant to the In this pilot, uncontrolled study, we show that the RTX manuscript. D. Curròreports no disclosures relevant to the reinfusion protocol based on memory B cells might be feasible manuscript. N. Bruschi reports no disclosures relevant to the and able to reduce ARIR and drug dosage, preserving efficacy manuscript. L. Roccatagliata reports no disclosures relevant to on inflammatory parameters (i.e., relapses and MRI activity). the manuscript. G. Boffa reports no disclosures relevant to the manuscript. C. Lapucci reports no disclosures relevant to the This proposed scheme moves toward a personalized medicine manuscript. G. Pesce reports no disclosures relevant to the approach in MS, a paradigm shift that is much needed in the manuscript. M. Cellerino reports no disclosures relevant to field, especially considering emerging data on long-term safety the manuscript. C. Solaro reports no disclosures relevant to of fixed doses reinfusion schemes for CD20-depleting anti- the manuscript. A. Laroni received honoraria for speaking by bodies.11 Our protocol might confirm the putative role of Biogen Idec, Novartis, and Teva; consulting fees by Merck memory B cells in MS pathogenesis, addressing relevant Serono, Sanofi-Genzyme, and Novartis; and funding for travel questions on the mechanism of action of some disease- from Teva, Merck Serono, Biogen, and Novartis. E. Capello modifying treatments in MS. Notably, we show that, with our reports no disclosures relevant to the manuscript. G. Luigi reinfusion scheme, RTX reinfusions could be performed less Mancardi reports no disclosures relevant to the manuscript. than once per year at the second year of treatment, repre- M. Pia Sormani received consulting fees from Biogen Idec, senting a 50% reduction in reinfusion rates, compared with Merck Serono, Teva, Genzyme, Roche, Novartis, GeNeuro,

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 5 and Medday. M. Inglese received honoraria or consultation fees from Roche, Biogen, Merck-Serono, Novartis, Genzyme, Appendix (continued)

and research grants from NIH, NMSS, FISM, and Teva Name Location Contribution Neuroscience. A. Uccelli received honoraria or consultation fees from Biogen, Roche, Teva, Merck-Serono, Genzyme, and Maria University of Major role in acquisition of data Cellerino, MD Genova, Genova, Novartis. Go to Neurology.org/NN for full disclosures. Italy

Claudio Antero Micone Major role in acquisition of data Publication history Solaro, MD Hospital, Genova Received by Neurology: Neuroimmunology & Neuroinﬂammation January 20, 2020. Accepted in ﬁnal form June 5, 2020. Alice Laroni, University of Major role in acquisition of data MD, PhD Genova, Genova, Italy

Elisabetta University of Major role in acquisition of data Appendix Authors Capello, MD Genova, Genova, Italy Name Location Contribution Gianluigi University of Major role in acquisition of data Mancardi, MD, Genova, Genova, Giovanni Novi, University of Major role in the acquisition of PhD Italy MD Genova, Genova, data, interpreted the data, design Italy and conceptualized study, Mariapia University of Analyzed the data and revised the analyzed the data, and drafted the Sormani, PhD Genova, Genova, manuscript for intellectual content manuscript Italy Francesca University of Analyzed the data and revised the Matilde University of Analyzed the data and revised the Bovis, PhD Genova, Genova, manuscript for intellectual content Inglese, MD, Genova, Genova, manuscript for intellectual content Italy PhD Italy Sabrina Antero Micone Major role in acquisition of data Antonio University of Analyzed the data and revised the Fabbri, MD Hospital, Genova Uccelli, MD Genova, Genova, manuscript for intellectual content Italy Francesco University of Major role in acquisition of data Tazza, MD Genova, Genova, Italy

Paola Gazzola, Antero Micone Major role in acquisition of data References MD Hospital, Genova 1. Hauser SL, Bar-Or A, Comi G, et al. Ocrelizumab versus interferon beta-1a in relapsing multiple sclerosis. New Engl J Med 2017;376:221–234. Ilaria Maietta University of Analyzed the data 2. Salzer J, Svenningsson R, Alping P, et al. Rituximab in multiple sclerosis A retro- Genova, Genova, spective observational study on safety and efficacy. Neurology 2016;87:2074–2081. Italy 3. Zecca C, Bovis F, Novi G, et al. Treatment of multiple sclerosis with rituximab: a multicentric Italian–Swiss experience. Mult Scler J 2019. Epub 2019 Oct 1. Daniela Curro,` University of Major role in acquisition of data 4. Berntsson SG, Kristoffersson A, Boström I, Feresiadou A, Burman J, Landtblom AM. MD Genova, Genova, Rapidly increasing off-label use of rituximab in multiple sclerosis in Sweden—outlier Italy or predecessor? Acta Neurol Scand 2018;138:327–331. 5. SmPC Ocrelizumab [Internet]. 2019. Available at: ema.europa.eu/en/documents/ Nicolo` University of Major role in acquisition of data product-information/ocrevus-epar-product-information_en.pdf. Accessed May 28, Bruschi, MD Genova, Genova, 2020. Italy 6. Baker D, Marta M, Pryce G, Giovannoni G, Schmierer K. Memory B cells are major targets for effective immunotherapy in relapsing multiple sclerosis. EBioMedicine Luca University of Major role in acquisition of data 2017. Roccatagliata, Genova, Genova, 7. Kim SH, Jeong IH, Hyun JW, et al. Treatment outcomes with rituximab in 100 MD, PhD Italy patients with neuromyelitis optica: influence of FCGR3A polymorphisms on the therapeutic response to rituximab. JAMA Neurol 2015;72:989–995. Giacomo University of Major role in acquisition of data 8. Novi G, Bovis F, Capobianco M, et al. Efficacy of different rituximab therapeutic Boffa, MD Genova, Genova, strategies in patients with neuromyelitis optica spectrum disorders. Mult Scler Relat Italy Disord 2019;36:101430 9. Kim SH, Kim W, Li XF, Jung IJ, Kim HJ. Repeated treatment with rituximab based Caterina University of Major role in acquisition of data on the assessment of peripheral circulating memory B cells in patients with re- Lapucci, MD Genova, Genova, lapsing neuromyelitis optica over 2 years. Arch Neurol 2011;68:1412–1420. Italy 10. Kim SH, Hyun JW, Kim HJ. Individualized B cell-targeting therapy for neuromyelitis optica spectrum disorder. Neurochem Int 2019;130:104347. Giampaola University of Major role in acquisition of data 11. Serum immunoglobulin levels and risk of serious infections in the pivotal Phase III Pesce, PhD Genova, Genova, trials of ocrelizumab in multiple sclerosis and their open-label extensions [Internet]. Italy 2019. Available at: medically.roche.com/en/search/pdfviewer.67fb1aea-a4a2-41f3- 91c4-20c8696215d2.html?cid=slpsxx1909nexxectrims2019. Accessed May 28, 2020.

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN ARTICLE OPEN ACCESS CLASS OF EVIDENCE Inﬂiximab treatment in pathology-conﬁrmed neurosarcoidosis

Daan Fritz, MD,* Wilhelmina M.C. Timmermans, MD,* Jan A.M. van Laar, MD, PhD, Correspondence P. Martin van Hagen, MD, PhD, Theodora A.M. Siepman, MD, Diederik van de Beek, MD, PhD, and Dr. Brouwer [email protected] Matthijs C. Brouwer, MD, PhD

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e847. doi:10.1212/NXI.0000000000000847

Abstract MORE ONLINE Objective Class of Evidence To assess the efficacy and risks of treatment with infliximab (anti–tumor necrosis factor alpha) Criteria for rating in pathology-confirmed neurosarcoidosis. therapeutic and diagnostic studies Methods NPub.org/coe In a retrospective study in 2 tertiary referral centers in the Netherlands, we analyzed clinical characteristics, complications, and outcome of patients with neurosarcoidosis treated with infliximab.

Results Twenty-eight patients were identified with a mean age of 42 years. Neurosarcoidosis presented with a cerebral parenchymal localization in 16 (59%), pituitary gland/hypothalamic sarcoidosis in 15 (54%), peripheral nerve involvement in 12 (43%), and chronic meningitis in 11 patients (41%). Initial treatment response after the start of infliximab was complete remission in 6 (21%) and improvement in 14 (50%), whereas 7 patients had stable disease (25%), and 1 (4%) deteriorated and died. At the end of follow-up, with a median of 32 months, 5 patients (18%) had died, and 2 (40%) were using infliximab at the time of death. Tapering or discontinuation of corticosteroids without a relapse was achieved in 19 of 28 patients (68%). In patients with decreasing dosing or discontinuation of infliximab, a relapse occurred in 5 of 19 patients (26%). Complications of infliximab were reported in 10 of 28 patients (36%) and mainly consisted of infections in 8 (29%).

Conclusion Infliximab is an effective treatment in neurosarcoidosis leading to remission or improvement in 70%. The mortality rate in infliximab-treated patients was substantial, indicating the severity of disease and treatment-associated complications.

Classification of evidence This study provides Class IV evidence that in people with pathology-confirmed neurosarcoidosis, infliximab is beneficial.

*These authors contributed equally.

From the Amsterdam UMC (D.F., D.B., M.C.B.), University of Amsterdam, Department of Neurology, Amsterdam Neuroscience, Amsterdam; Erasmus MC (W.M.C.T., J.A.M.L., P.M.H.), Department of Internal Medicine; Erasmus MC (W.M.C.T., J.A.M.L., P.M.H.), Department of Immunology; and Erasmus MC (T.A.M.S.), Department of Neurology, Rotterdam, the Netherlands.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary AMC = Academic Medical Center; EMC = Erasmus MC; IQR = interquartile range; mRS = modiﬁed Rankin Scale; TNF- α = tumor necrosis factor alpha.

Sarcoidosis is a multisystem disorder and is characterized by therapy consisted of corticosteroid therapy, and second-line the presence of granulomas that can affect every organ sys- therapy consisted of methotrexate, azathioprine, mycopheno- tem.1 The prevalence of sarcoidosis in the population is es- late mofetil, hydroxychloroquine, and cyclosporine.3 All patients timated to be between 5 and 50 per 100.000, with the highest were treated with an infliximab dosage of 5 mg/kg. In the EMC, prevalence in Northern Europe.1 Approximately 5% of pa- patients were treated at weeks 0, 2, and 6 during the induction tients with sarcoidosis have neurosarcoidosis, in which gran- phase, followed by an infusion once every 4–8 weeks based on 2 ulomas involve the nervous system. Neurosarcoidosis is a the clinical features and seriousness by their treating physician. severe form of sarcoidosis in which one-third of patients re- In the AMC, patients did not undergo an induction phase and main stable, deteriorate, or die despite immunosuppressive immediately received infusions once every 4–8weeksbasedon 2 treatment. No clinical trials have been performed in patients the clinical features and seriousness. All patients were initially with neurosarcoidosis, and treatment choices are mainly treated with Remicade and were switched in 2016 to the bio- based on evidence from non-neurologic sarcoidosis. similar Remsima. All infections, infusion reactions, and laboratory abnormalities that occurred during the use of infliximab fl Treatment with in iximab, a tumor necrosis factor alpha (TNF- were reported. The response to treatment in each case was α ) blocker, has emerged as a promising option in the past years in scored as remission or improvement on therapy (a favorable fi patients with neurosarcoidosis refractory to rst- or second-line treatment response), stable disease (e.g., unchanged compared therapy.3 TNF-α is a pivotal proinflammatory cytokine and plays – with clinical situation before treatment), and deterioration. The a central role in the formation and maintenance of granulomas.4 6 response rate was assessed about 3 months after the start of Hence, therapeutic efficacy of TNF-α antagonists has recently – infliximab depending on time of follow-up. Clinical outcome been reported in refractory systemic and neurosarcoidosis.3,7 18 was graded into functional disability at the last recorded hospital To substantiate this potential beneficial effect of TNF-α blockers, contact. The functional disability in each case was scored, using we analyzed the use of infliximab in patients with biopsy-proven the modified Rankin Scale (mRS), as asymptomatic, complaints neurosarcoidosis and evaluated the treatment response and safety without functional disability, complaints with slight functional in a large multicenter tertiary center cohort. disability (e.g., able to look after own affairs, but unable to perform all previous activities), complaints with moderate disability (e.g., neurologic deficits mildly interfering in everyday Methods life, such as inability to cycle due to motor dysfunction, but able A retrospective study was performed with inclusion of all to walk unassisted), complaints with moderate-to-severe func- patients with biopsy-proven sarcoidosis and neurologic in- tional disability (e.g., neurologic deficits interfering everyday life, volvement who were treated with infliximab before June 1, resulting in failure to return to job or school, requirement of 2017, at the Academic Medical Center (AMC) in Amsterdam special equipment such as crutches or a wheelchair, or assistance and the Erasmus MC (EMC) in Rotterdam, 2 tertiary referral with everyday activities), complaints with severe disability (e.g., centers for (neuro)sarcoidosis in the Netherlands. Ethical requires constant nursing care and attention, bedridden, and/or approval is not required in the Netherlands for a retrospective incontinent), and death. An mRS score of asymptomatic or study with anonymized patient data such as our study. Pa- complaints without functional disability was considered a fa- tients were identified by their treating physician, and data vorable outcome. were collected in a database. The diagnosis of neurosarcoidosis was based on the Zajicek et al. criteria, later Statistical analysis was performed to compare differences modified by Tavee and Stern.19,20 A positive histology for between groups using the Fisher exact test for dichotomous sarcoidosis was defined as the presence of histologic features variables and the Mann-Whitney U test for ordinal and consistent with sarcoidosis defined as noncaseating granulo- continuous variables. A p value <0.05 was considered 6 mas with epithelioid cells and macrophages. significant.

For all patients, a case record form was created containing This study was designed to assess the efficacy and risks of baseline characteristics, disease course and immune-modulating treatment with infliximab (anti–TNF-α) in pathology-confirmed medication used at baseline, clinical characteristics and results of neurosarcoidosis. This study provides Class IV evidence. ancillary investigations at baseline, infliximab treatment and treatment response, disease course and clinical outcome up to Data availability the last time of follow-up, and adverse events. Baseline was Anonymized data not published in the article are available on defined as the initiation of infliximab treatment. First-line request by any qualified investigator.

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Results values returned to normal levels in 9 (64%) and improved in 2 of 14 patients (14%), although the change was not statistically sig- Clinical characteristics nificant (figure 1). MRI of the brain and/or spinal cord was re- A total of 28 patients were included, 11 in the AMC and 17 in the peated in 21 patients at follow-up, and radiologic abnormalities EMC. Baseline characteristics, clinical manifestations, and an- attributed to neurosarcoidosis decreased in 15 (71%), remained cillary investigations at the start of infliximab are described in stable in 4 patients (19%), and increased in 1 patient (7%). When table 1 and were comparable in both centers. The median time of looking at a favorable treatment response, we did not find an follow-up was 32 months (interquartile range [IQR] 17–54 association of time between the diagnosis neurosarcoidosis and the months). The included patients had a mean age at baseline of 42 start of infliximab (p = 0.165). Also, a favorable outcome of 0 or 1 years (SD 10.3), and 16 (57%) were male. Of these patients, 16 on the mRS was not associated (p = 0.80). Furthermore, the use of (57%) were Caucasian, 6 (21%) of African descent, and 6 (21%) second-line therapy at the start of infliximab was not associated had other ethnic backgrounds. Neurologic involvement at the with a favorable treatment response (13/15 [87%] with second- start of infliximab consisted of parenchymal involvement in 16 line treatment vs 7/13 [54%] without second-line treatment, p = patients (59%), pituitary/hypothalamic involvement in 15 0.10). The use of second-line therapy is not associated with a (54%), peripheral nerve involvement in 12 (43%), chronic favorable outcome on the mRS (7/15 [47%] with second-line meningitis in 11 (41%), cranial nerve palsy in 7 (25%), hydro- treatment vs 5/13 [38%] without second-line treatment, p = 0.72). cephalus in 6 (22%), spinal cord involvement in 5 (18%), and muscle involvement in 1 patient (4%). The main neurologic In patients with a favorable treatment response, the infliximab symptoms were paresis in 12 of 28 patients (43%), sensory dosage was decreased in 2 of 20 patients (10%) and was stopped symptoms in 12 (43%), impaired gait in 11 (39%), and headache in 11 (55%), which led to a relapse in 4 of 13 patients (31%). In in 10 (35%). Overall, 27 of 28 patients (96%) had multiple patients with an unfavorable treatment response, infliximab was symptoms. The majority of patients had systemic involvement stopped in 4 of 8 patients (50%) and led to a relapse in 1 of 4 including lymph node (72%), intrapulmonary (27%), and patients (25%). In patients in whom treatment regimen was not ophthalmologic (22%) localizations of sarcoidosis. The median altered, a relapse occurred in 1 of 11 patients (9%), which was mRS score at baseline was 3 (range 0–5). According to the attributed to the stop of methotrexate because of liver toxicity. Zajicek et al. criteria, 2 patients were diagnosed with definite and 26 patients with probable neurosarcoidosis, all with pathology Overall, a relapse was not associated with the length of infliximab use confirmation of noncaseating granulomas. Treatment used at the before discontinuation (p = 0.31). In patients with a relapse after the start of infliximab consisted of corticosteroids in 11 patients treatment regimen was altered, infliximab was restarted, or dosage (39%), prednisolone and methotrexate in 6 (21%), prednisolone was increased in all, leading to a favorable treatment response in 4 and azathioprine in 4 (14%), methotrexate and prednisone with patients and stable disease in 1 patient. Autoantibodies against mycophenolate mofetil in 2 (7%), and prednisone with infliximab were tested in 5 patients who experienced a relapse and hydroxychloroquine in 1 patient (4%). Three patients were not were positive in 1 patient. This patient was switched to adalimumab, receiving immunosuppressive medication at the start of inflix- which led to improvement. At the last time of follow-up, 6 of 28 imab, but had received this previously. Before the start of patients (21%) were asymptomatic, 6 (21%) had no significant infliximab, all patients had been treated with corticosteroids, 14 disability, 5 (18%) had slight disability, 3 (11%) had moderate dis- (50%) with methotrexate, 13 (46%) with azathioprine, 5 (18%) ability, 3 patient (11%) had moderately severe disability, and 5 (18%) with mycophenolate mofetil, 2 (7%) with cyclophosphamide, died (figure 2). Overall, the change in the mRS was not significant (p and 1 (4%) with hydroxychloroquine or cyclosporine. Overall, = 0.12). Of the patients who died, cause of death was sepsis in 2 24 of 28 patients (86%) had previously been treated with a patients, and brain stem hemorrhage, malignancy, and an unknown second-line treatment (table 1). The reasons to start infliximab cause of death each occurred in 1 patient. Of these patients, only the were a relapse when tapering corticosteroids despite second-line 2 patients with infectious complications used infliximab at the time of treatment in 16 of 28 patients (57%), serious side effects of first- death, and both were on concomitant corticosteroid therapy. and/or second-line treatment in 8 (29%), chronic progression Complications attributed to infliximab treatment occurred in 10 of despite first- and/or second-line treatment in 3 (11%), and a 28 patients (36%) and consisted of infections in 8 patients (29%) relapse after tapering corticosteroids in 1 patient (4%). and elevated liver tests and an allergic reaction both in 1 patient (4%). No adverse effects could be contributed to the switch to a biosimilar. Therapeutic outcome Two patients reported an increase of pain and sensory symptoms The starting dose of infliximab of 5 mg per kilogram was given without any signs of a relapse or disease activity. once every 4 weeks in 5 patients (18%), once every 6 weeks in 12 patients (43%), and once every 8 weeks in 11 patients (39%). The median total number of infusions was 17 (range 5–50), and Discussion infliximab treatment had a median duration of 23 months (IQR 12–38 months). The initial treatment response consisted of re- We report a favorable long-term outcome in patients with missionin6patients(21%),improvement in 14 (50%), stable pathology-confirmed neurosarcoidosis treated with infliximab. disease in 7 (25%), and death in 1 patient (4%) (table 2). Treat- However, we observe a relative high mortality at follow-up and ment responses were comparable in both centers. The sIL2R a high occurrence of infectious complications. This further

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 Table 1 Baseline characteristics and disease course at baseline

Characteristic n/N (%) Characteristic n/N (%)

Age at baseline (IQR), ya 42 (33–49) Zajicek criteria

Male sex 16 (57) Definite neurosarcoidosis 2/28 (7)

Ethnicity Probable neurosarcoidosis 26/28 (93)

Caucasian 16 (57) mRS score at baseline 3 (0–5)c

African descent 6 (21) Asymptomatic 1/28 (4)

Otherb 6 (21) No significant disability 4/28 (14)

Neurologic involvement Slight disability 6/28 (21)

Parenchymal 16/27 (59) Moderate disability 6/28 (21)

Neuroendocrine 15/27 (56) Moderate-severe disability 5/28 (18)

Peripheral nerve 11/27 (41) Severe disability 6/28 (21)

Polyneuropathy 8/11 (73) Ancillary investigations

Radiculopathy 2/11 (18) Serum ACE >70 U/L 3/17 (18)

Plexopathy 1/11 (9) Serum sIL2r >3,000 U/mL 11/15 (73)

Chronic aseptic meningitis 11/27 (41) Serum CRP >10 mg/L 7/21 (33)

Cranial nerve palsy 7/27 (26) Serum ESR >20 mm/h 8/19 (42)

Hydrocephalus 6/27 (22) CSF leukocytes > 5 × 106 cells/L 6/6 (100)

Spinal cord 5/27 (19) CSF protein > 0.6 g/L 6/6 (100)

Muscle 1/27 (4) Chest CT suggestive 3/4 (75)

Neurologic symptoms 18F-FDG PET-CT suggestive 1/6 (16)

Paresis 12/28 (43) Brain MRI suggestive 21/22 (95)

Sensory symptoms 12/28 (43) Spinal cord MRI suggestive 2/7 (29)

Impaired gait 11/28 (39) Immunosuppressant use at baseline

Headache 10/28 (35) Corticosteroids 11/28 (39)

Cognitive/psychiatric 8/28 (29) Corticosteroids + methotrexate 6/28 (21)

Visual symptoms 7/28 (25) Corticosteroids + azathioprine 4/28 (14)

Impaired coordination 5/28 (18) Methotrexate 2/28 (7)

Speech impairment 4/28 (15) Corticosteroids + hydroxychloroquine 1/28 (4)

Facial palsy 4/28 (15) Corticosteroids + MMF 2/28 (7)

Vertigo 3/28 (11) None 3/28 (11)

Seizure 2/28 (7) Prior immunosuppressant use

Hearing loss 1/28 (4) Corticosteroids 28/28 (100)

Multiple symptoms 27/28 (96) Methotrexate 14/28 (50)

Systemic involvement Azathioprine 13/28 (46)

Lymph node 18/25 (72) Mycophenolate mofetil 5/28 (18)

Lungs 7/26 (27) Cyclophosphamide 2/28 (7)

Eye 6/27 (22) Hydroxychloroquine 1/28 (4)

ENT 3/28 (11) Cyclosporine 1/28 (4)

Skin 2/27 (7) No. of prior immunosuppressants

Continued

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Table 1 Baseline characteristics and disease course at baseline (continued)

Characteristic n/N (%) Characteristic n/N (%)

Joints 1/28 (4) One type 4/28 (14)

Endocrinologic 5/28 (18) Two types 16/28 (57)

Three types 6/28 (21)

Five types 2/28 (7)

Abbreviations: ACE = angiotensin-converting enzyme; ESR = erytrocyte sedimentation rate; IQR = interquartile range; MMF = mycophenolate mofetil; mRS = modified Rankin Scale; sIL2r = soluble interleukin 2 receptor. a Median (IQR). b Other ethnicities: 2 (7%) North African, 2 (7%) Hindustan, 1 (4%) Asian, and 1 (4%) Hispanic. c Median (range). establishes TNF-α antagonists as a suitable third-line agent in In our cohort, remission or clinical improvement was ach- patients with insufficient response to first- and second-line ieved in 71% of patients, which is in line with 2 previously treatment, but follow-up remains essential to taper when pos- published studies. To date, the evidence for the use of sible and evaluate for infectious complications. infliximab in neurosarcoidosis consists of case reports and

Table 2 Treatment and outcome

Characteristic n/N (%) Characteristic n/N (%)

Months between sarcoidosis and IFXa 27 (18–42) Result of change/stop IFX

Months between neurosarcoidosis and IFXa 19 (9–41) None 9/17 (53)

Months between IFX and last follow-upa 32 (17–54) Relapse 5/17 (29)

Infliximab duration in moa 23 (12–38) Unknown 3/17 (18)

Total number of infusionsb 17 (5–50) 1st- or 2nd-line treatment change

Treatment response Taper of 1st-line treatment 10/23 (43)

Remission 6/28 (21) Stop of 1st-line treatment 13/23 (57)

Improvement 14/28 (50) Stop of 2nd-line treatment 5/14 (36)

Stable disease 7/28 (25) Complications of IFX treatment 10/28 (36)

Deterioration 1/28 (4) Infectionsd 8/28 (29)

Change of neurologic imaging Elevated liver tests 1/28 (4)

Improvement 15/21 (71) Allergic reaction 1/28 (4)

Stable 4/21 (19) Modified Rankin Scale score 2 (0–6)

Other abnormalities 2/21 (10) Asymptomatic 6/28 (21)

IFX dosage decrease or stop 17/28 (61) No significant disability 6/28 (21)

Good treatment response 8/17 (47) Slight disability 5/28 (18)

Insufficient treatment response 1/17 (6) Moderate disability 3/28 (11)

Major side effects 4/17 (24) Moderate-severe disability 3/28 (11)

Stable symptoms, no disease activity 2/17 (12) Moderate disability 0/28 (0)

Otherc 2/17 (12) Death 5/28 (18)

Autoantibodies 1/5 (20) a Median (interquartile range). b Median (range). c Other: 1 patient developed autoantibodies against infliximab, and 1 patient decided to stop. d Infections consisted of pneumonia in 4 patients, urinary tract infection in combination with pneumonia in 1 patient, and urinary tract infection in 1 patient, all necessitating hospital admission.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 5 Figure 1 Serum soluble interleukin 2 receptor and angiotensin-converting enzyme values before and after infliximab

The dotted line indicates the upper limit of the normal value. p Values were calculated using the Mann-Whitney U test. (A) Serum sIL2r values before and after the start of infliximab. (B) Serum ACE values before and after the start of infliximab. ACE = angiotensin-converting enzyme; sIL2r = soluble interleukin 2 receptor.

– series and 2 retrospective multicenter cohort studies.10 18 A manifestations, it seems reasonable to start infliximab when summary of published articles describing 4 or more patients first-line treatment fails and a quick treatment response is treated with infliximab can be found in table e-1, links.lww. required or when second-line treatment fails. com/NXI/A289. The 2 retrospective multicenter cohort studies describe 18 and 66 patients with probable or definite Importantly, we report a mortality rate of 18%, which is higher neurosarcoidosis, respectively, describing improvement or compared with previously published meta-analysis of patients remission in 89% and 77% of patients.13,14 Of interest, despite with neurosarcoidosis (5%) and the aforementioned retro- a shorter median time of follow-up of 20 and 18 months, spective studies on TNF-α antagonists in neurosarcoidosis respectively, these articles describe higher relapse rates of 50% (5% and 0%, respectively).2,13,14 In a large retrospective study and 56% compared with the overall relapse rate of 21% in our evaluating prognostic factors in neurosarcoidosis with a me- cohort. A correlation between the duration of neuro- dian follow-up of 8 years, 21 of 234 patients (9%) died, 3 of sarcoidosis and a favorable treatment response was previously which due to sepsis.21 Our high mortality rate may be fi described, but this nding could not be reproduced in our explained by the severely affected population, which is se- 14 cohort. However, given the severity of the neurologic lected for infliximab treatment. Of the patients who died, 2 of 5 patients were using infliximab infusions combined with first- and second-line immunosuppressants at the time of death, Figure 2 Modified Rankin Scale score change of follow-up compared with baseline and both died due to infectious complications due to their severely immunocompromised state. The occurrence of infections is an important adverse effect, occurring in 29% of the patients. The infection rate is comparable to other studies describing infliximab use in neurosarcoidosis, which varied between 10 and 39%, although in these studies, none of the patients died due to infectious complications.13,14 The risk of infectious complications is considered to be higher when patients are treated in combination with corticosteroids or other immunosuppressive drugs.22,23 Of interest, as shown by figure 2, the high mortality rate is flanked by an increase of favorable outcome as measured by the mRS at baseline compared with follow-up. The combined data suggest that infliximab should be reserved to patients with severe disease 0: asymptomatic, 1: complaints without functional disability, 2: complaints and that tapering corticosteroids remains essential when signs with slight functional disability, 3: complaints with moderate disability, 4: of disease activity have diminished. In our cohort, starting complaints with moderate-to-severe functional disability, 5: complaints with severe disability, and 6: death. mRS = modified Rankin Scale. infliximab was followed by tapering of concomitant first-line treatment in 43% of patients and eventually complete

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN discontinuation of corticosteroids in 57% of patients. This is involvement. However, in none of the patients, peripheral nerve in line with another study describing discontinuation of ste- involvement was the primary indication for infliximab start. roids after the start of infliximab in 40% of the patients and maintenance of prednisone on 5 mg per day or less in 27% of In conclusion, anti–TNF-α antagonists can be effective in the patients.14 Prompt tapering of steroids is essential to neurosarcoidosis; however, physicians should be aware of pos- prevent adverse effects and reduce the risk of infections in sible side effects as seen in about a third of patients in our cohort, severely immunocompromised patients.24 Furthermore, with a risk of serious infectious complications. Future ran- combination therapy with another steroid-sparing agent may domized controlled trials and prospective cohort studies to as- be associated with a favorable treatment response to inflix- sess the safety of infliximabarewarrantedandshouldshedlight imab.14 This could not be reproduced in our cohort, but over whether mono- or combination therapy is preferred and considering the small sample size, a possible association might when and how to taper and stop infliximab during follow-up. be missed. Study funding Currently, no international guidelines exist on how to proceed M.C. Brouwer is supported by a grant from the Netherlands after a favorable treatment response following infliximab treat- Organization for Health Research and Development ment. In patients in whom infliximab treatment was stopped or (ZonMw; NWO-Vidi grant 2017 [917.17.308]). D. van de dosage was decreased in our cohort, 26% had a relapse of dis- Beek is supported by grants from the Netherlands Organi- ease activity compared with 56% in an earlier published study.14 zation for Health Research and Development (ZonMw; Relapse can occur after long-term infliximab treatment in pa- NWO-Vici grant 2019 [918.19.627]), the European Research tients with clinical and radiologic remission.14 It remains un- Council (ERC Starting Grant 281156), and an Innovation known at what moment in the disease course infliximab should grant by the board of directors of the Amsterdam UMC, be tapered and stopped. In case of a relapse, dosage can be Amsterdam, the Netherlands. No potential conflict of interest increased, and/or intervals between infusions can be shortened relevant for this article exists. quickly. In our cohort, when patients were reintroduced to infliximab, they were likely to show a good treatment response Disclosure again, similar to what have been reported in inflammatory bowel The authors report no disclosures. Go to Neurology.org/NN disease.25 Although rare, when a relapse occurs during inflix- for full disclosures. imab treatment or when reintroduction fails, it is important that this can be due to autoantibody formation.26,27 In a study Publication history assessing the safety and efficacy of TNF-α antagonists in re- Received by Neurology: Neuroimmunology & Neuroinflammation fractory sarcoidosis, 3 of 132 (2%) developed autoantibodies.28 March 30, 2020. Accepted in final form June 16, 2020. In other inflammatory disorders including Crohn disease, concomitant methotrexate is advised as this has been shown to be efficient in reducing immunogenicity, although evidence in Appendix Authors patients with sarcoidosis is lacking.29 Name Location Contribution

Our study has several limitations. First, both the retrospective Daan Fritz, MD Amsterdam UMC, Substantial contribution to and multicenter approach of our study resulted in heteroge- Amsterdam, The conception and design, Netherlands acquisition of data, analysis neous assessment of disease activity, treatment strategies, and and interpretation of data, outcome, as well as missing data in some patients. Furthermore, drafted the manuscript, and treatment strategies differed between the 2 centers. In the EMC, final approval of the version to be published an induction phase was used when they started patients on infliximab, whereas the AMC initiates treatment once every 4–8 Wilhelmina Erasmus MC, Substantial contribution to M.C. Rotterdam, The conception and design, weeks without an initial induction phase. Despite these differ- Timmermans, Netherlands acquisition of data, analysis ences, baseline characteristics and treatment response did not MD and interpretation of data, ff drafted the manuscript, and di er between the 2 centers. However, sIL2R measurements final approval of the version were only performed in the EMC cohort. These limitations are to be published inherent to the study design. In addition, a majority of patients Jan A.M. van Erasmus MC, Substantial contribution to were treated with first- and/or second-line therapy possibly Laar, MD, PhD Rotterdam, The conception and design, Netherlands acquisition of data, analysis contributing to treatment responses and the occurrence of side and interpretation of data, effects. Last, we included only patients treated in our tertiary revised the manuscript for important intellectual referral centers, introducing a selection bias. We feel that pa- content, and final approval of tients with neurosarcoidosis treated with infliximab must be the version to be published treated at specialized centers, which is the norm in the Neth- Martin van Erasmus MC, Interpreted the data, revised erlands. Furthermore, both tertiary referral centers are special- Hagen, MD, Rotterdam, The the manuscript, and final ized centers in neuromuscular diseases, which may add to the PhD Netherlands approval of the version to be published relative high proportions of patients with peripheral nerve Continued Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 7 10. Hostettler KE, Studler U, Tamm M, Brutsche MH. Long-term treatment with fl – Appendix (continued) in iximab in patients with sarcoidosis. Respiration 2012;83:218 224. 11. Santos E, Shaunak S, Renowden S, Scolding NJ. Treatment of refractory neurosarcoidosis with Infliximab. J Neurol Neurosurg Psychiatry 2010;81:241–246. Name Location Contribution 12. Riancho-Zarrabeitia L, Delgado-Alvarado M, Riancho J, et al. Anti-TNF-alpha therapy in the management of severe neurosarcoidosis: a report of five cases from a single Theodora A.M. Erasmus MC, Interpreted the data, revised centre and literature review. Clin Exp Rheumatol 2014;32:275–284. Siepman, MD Rotterdam, The the manuscript, and final 13. Cohen Aubart F, Bouvry D, Galanaud D, et al. Long-term outcomes of refractory Netherlands approval of the version to be neurosarcoidosis treated with infliximab. J Neurol 2017;264:891–897. published 14. Gelfand JM, Bradshaw MJ, Stern BJ, et al. Infliximab for the treatment of CNS sarcoidosis: a multi-institutional series. Neurology 2017;89:2092–2100. Diederik van de Amsterdam UMC, Substantial contribution to 15. Sodhi M, Pearson K, White ES, Culver DA. Infliximab therapy rescues cyclophosphamide Beek, MD, PhD Amsterdam, The conception and design, failure in severe central nervous system sarcoidosis. Respir Med 2009;103:268–273. Netherlands acquisition of data, analysis 16. Panselinas E, Rodgers JK, Judson MA. Clinical outcomes in sarcoidosis after cessation and interpretation of data, of infliximab treatment. Respirology 2009;14:522–528. revised the manuscript for 17. Moravan M, Segal BM. Treatment of CNS sarcoidosis with infliximab and myco- important intellectual phenolate mofetil. Neurology 2009;72:337–340. content, and final approval of 18. Russell E, Luk F, Manocha S, Ho T, O’Connor C, Hussain H. Long term follow-up of the version to be published infliximab efficacy in pulmonary and extra-pulmonary sarcoidosis refractory to conventional therapy. Semin Arthritis Rheum 2013;43:119–124. Matthijs C. Amsterdam UMC, Substantial contribution to 19. Tavee JO, Stern BJ. Neurosarcoidosis. Continuum (Minneap Minn) 2014;20: Brouwer, MD, Amsterdam, The conception and design, 545–559. PhD Netherlands acquisition of data, analysis 20. Zajicek JP, Scolding NJ, Foster O, et al. Central nervous system sarcoidosis— and interpretation of data, diagnosis and management. QJM 1999;92:103–117. revised the manuscript for 21. Joubert B, Chapelon-Abric C, Biard L, et al. Association of prognostic factors and important intellectual immunosuppressive treatment with long-term outcomes in neurosarcoidosis. JAMA content, and final approval of Neurol 2017;74:1336–1344. the version to be published 22. Garcia-Vidal C, Rodriguez-Fernandez S, Teijon S, et al. Risk factors for opportunistic infections in infliximab-treated patients: the importance of screening in prevention. Eur J Clin Microbiol Infect Dis 2009;28:331–337. 23. Deepak P, Stobaugh DJ, Ehrenpreis ED. Infectious complications of TNF-alpha References inhibitor monotherapy versus combination therapy with immunomodulators in in- 1. Valeyre D, Prasse A, Nunes H, Uzunhan Y, Brillet PY, Muller-Quernheim J. Sar- flammatory bowel disease: analysis of the Food and Drug Administration Adverse coidosis. Lancet 2014;383:1155–1167. Event Reporting System. J Gastrointestin Liver Dis 2013;22:269–276. 2. Fritz D, van de Beek D, Brouwer MC. Clinical features, treatment and outcome in 24. Rizzato G, Riboldi A, Imbimbo B, Torresin A, Milani S. The long-term efficacy and neurosarcoidosis: systematic review and meta-analysis. BMC Neurol 2016;16:220. safety of two different corticosteroids in chronic sarcoidosis. Respir Med 1997;91: 3. Fritz D, Voortman M, van de Beek D, Drent M, Brouwer MC. Many faces of neuro- 449–460. sarcoidosis: from chronic meningitis to myelopathy. Curr Opin Pulm Med 2017;23:439–446. 25. Casanova MJ, Chaparro M, Garcia-Sanchez V, et al. Evolution after anti-TNF dis- 4. Baughman RP, Iannuzzi M. Tumour necrosis factor in sarcoidosis and its potential for continuation in patients with inflammatory bowel disease: a multicenter long-term targeted therapy. BioDrugs 2003;17:425–431. follow-up study. Am J Gastroenterol 2017;112:120–131. 5. Baughman RP, Strohofer SA, Buchsbaum J, Lower EE. Release of tumor necrosis factor 26. Baert F, Noman M, Vermeire S, et al. Influence of immunogenicity on the long-term by alveolar macrophages of patients with sarcoidosis. J Lab Clin Med 1990;115:36–42. efficacy of infliximab in Crohn’s disease. N Engl J Med 2003;348:601–608. 6. Timmermans WM, van Laar JA, van Hagen MP, van Zelm MC. Immunopathogenesis of 27. Ainsworth MA, Bendtzen K, Brynskov J. Tumor necrosis factor-alpha binding ca- granulomas in chronic autoinflammatory diseases. Clin Transl Immunol 2016;5:e118. pacity and anti-infliximab antibodies measured by fluid-phase radioimmunoassays as 7. Baughman RP, Drent M, Kavuru M, et al. Infliximab therapy in patients with chronic predictors of clinical efficacy of infliximab in Crohn’s disease. Am J Gastroenterol sarcoidosis and pulmonary involvement. Am J Respir Crit Care Med 2006;174:795–802. 2008;103:944–948. 8. Rossman MD, Newman LS, Baughman RP, et al. A double-blinded, randomized, 28. Jamilloux Y, Cohen-Aubart F, Chapelon-Abric C, et al. Efficacy and safety of tumor placebo-controlled trial of infliximab in subjects with active pulmonary sarcoidosis. necrosis factor antagonists in refractory sarcoidosis: a multicenter study of 132 pa- Sarcoidosis Vasc Diffuse Lung Dis 2006;23:201–208. tients. Semin Arthritis Rheum 2017;47:288–294. 9. Adler BL, Wang CJ, Bui TL, Schilperoort HM, Armstrong AW. Anti-tumor necrosis 29. Garces S, Demengeot J, Benito-Garcia E. The immunogenicity of anti-TNF therapy in factor agents in sarcoidosis: a systematic review of efficacy and safety. Semin Arthritis immune-mediated inflammatory diseases: a systematic review of the literature with a Rheum 2019;48:1093–1104. meta-analysis. Ann Rheum Dis 2013;72:1947–1955.

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN ARTICLE OPEN ACCESS Clinically based score predicting cryptogenic NORSE at the early stage of status epilepticus

Atsuko Yanagida, MD, Naomi Kanazawa, BS, Juntaro Kaneko, MD, Atsushi Kaneko, MD, Ryoko Iwase, MD, Correspondence Hiroki Suga, MD, Yutaka Nonoda, MD, PhD, Yuya Onozawa, PhD, Eiji Kitamura, MD, PhD, Dr. Iizuka [email protected] Kazutoshi Nishiyama, MD, PhD, and Takahiro Iizuka, MD

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e849. doi:10.1212/NXI.0000000000000849 Abstract Objective To determine whether a clinically based score predicts cryptogenic new-onset refractory status epilepticus (C-NORSE) at the early stage of status epilepticus (SE) with prominent motor symptoms (SE-M) of unclear etiology.

Methods The score (range 0–6) included 6 clinical features: highly refractoriness to antiseizure drugs, previously healthy individual, presence of prodromal fever, absence of prodromal psychobehavioral or memory alterations, absence of dyskinesias, and symmetric brain MRI abnormalities (the ﬁrst 2 mandatory). We retrospectively assessed the usefulness of a high scale score (≥5) in predicting C-NORSE in 83 patients with SE-M of unclear etiology, who underwent testing for neuronal surface antibodies (NS-Abs) between January 2007, and December 2019.

Results Thirty-one (37.3%) patients had a high score. Patients with a high score had more frequent prodromal fever (28/31 vs 24/52), mechanical ventilatory support (31/31 vs 36/52), and symmetric MRI abnormalities (26/31 vs 12/52), had less frequent involuntary movements (2/ 31 vs 30/52), and had absent prodromal psychobehavioral alterations (0/31 vs 27/52), CSF oligoclonal band detection (0/27 vs 11/38), tumor association (0/31 vs 13/52), or NS-Abs (0/ 31 vs 29/52) than those with a low score (<5). Thirty-three patients (median age, 27 years; 18 [54.5%] female) were ﬁnally regarded as C-NORSE. The sensitivity and speciﬁcity of a high score for predicting C-NORSE were 93.9% (95% CI 0.87–0.94) and 100% (95% CI 0.95–1.00), respectively.

Conclusions Patients with a high score in the indicated scale are more likely to have C-NORSE, making it a useful diagnostic tool at the early stage of SE-M before antibody test results become available.

From the Department of Neurology (A.Y., N.K., J.K., A.K., R.I., H.S., E.K., K.N., T.I.) and Department of Pediatrics (Y.N.), Kitasato University School of Medicine; and Department of Clinical Laboratory (Y.O.), Kitasato University Hospital, Sagamihara, Japan.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary AE = autoimmune encephalitis; AMPAR = α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor; AQP4 = aquaporin-4; ASD = antiseizure drug; CBA = cell-based assay; C-NORSE = cryptogenic NORSE; DWI =diﬀusion-weighted image; FC = febrile convulsion; FIRES = febrile infection-related epilepsy syndrome; FLAIR = ﬂuid-attenuated inversion recovery; GABAaR = γ-aminobutyric acid A receptor; GABAbR = γ-aminobutyric acid B receptor; IgG = immunoglobulin G; IL-6 = interleukin-6; ILAE = International League Against Epilepsy; LGI1 = leucine-rich glioma-inactivated 1; MOG = myelin oligodendrocyte glycoprotein; NCSE = nonconvulsive SE; NMDAR = NMDA receptor; NORSE = new-onset refractory status epilepticus; NS-Abs = neuronal surface antibodies; OCB = oligoclonal band; PMH = past medical history; SE = status epilepticus; SE-M = SE with prominent motor symptoms; WBC = white blood cell.

New-onset refractory status epilepticus (NORSE) is a severe 7 patients who were admitted before January 1, 2007, archived neurologic emergency condition characterized by refractory serum/CSF samples obtained at onset of disease were used status epilepticus (SE) without readily identifiable cause in for antibody assays. otherwise healthy individuals.1,2 The term NORSE is now defined as a clinical presentation, not a specific diagnosis.3 Then, we selected 129 patients who fulfilled the 2015 ILAE When the cause remains unknown despite the extensive criteria for SE.6 Of those, 46 patients with nonconvulsive SE – workup, it is called cryptogenic NORSE (C-NORSE).2 4 (NCSE) were excluded because the scale score was originally developed to estimate antibody status in patients with con- According to the consensus definition, NORSE includes pa- vulsive SE. In this study, we included all patients who de- 3 tients with viral, paraneoplastic, or autoimmune etiologies ; veloped SE-M regardless of refractoriness to conventional ff however, it is crucial in clinical practice to di erentiate antiseizure drug (ASD) treatment. We assessed the sensitivity C-NORSE from secondary NORSE with neuronal surface and specificity of the high scale score (≥5) in 83 patients with antibodies (NS-Abs) or classical paraneoplastic antineuronal SE-M of unclear etiology during the early stage. antibodies because treatment strategy and outcome could be ff 5 di erent. A large cohort study reported that a half of 130 NS-Abs were measured at the laboratory of Josep Dalmau patients with NORSE remained cryptogenic, but 37% were (University of Barcelona) using both a rat brain immunohis- immune mediated; among those, the most common etiology 7–13 2 tochemistry and cell-based assay (CBA) ; they included was anti-NMDA receptor (NMDAR) encephalitis. antibodies against the NMDAR, α-amino-3-hydroxy-5-methyl- 4-isoxazolepropionic acid receptor (AMPAR), γ-aminobutyric Although antibody tests are important to determine the eti- acid B receptor (GABAbR), γ-aminobutyric acid A receptor ology, in an emergency condition, it is often difficult to get the (GABAaR), metabotropic glutamate receptor 5, dipeptidyl antibody test results in appropriate time. Therefore, we pre- peptidase-like protein 6, contactin-associated protein-like 2, viously developed a clinically based score (range 0–6) based leucine-rich glioma-inactivated 1 (LGI1), and neurexin 3. Both on 6 clinical features to predict C-NORSE at the early stage of convulsive SE, which is currently classified into SE with serum and CSF were examined in all patients except 4 (only prominent motor symptoms (SE-M) according to the 2015 CSF [n = 2] or serum [n = 2] was available). In addition to NS- International League Against Epilepsy (ILAE) criteria for SE.6 Abs, myelin oligodendrocyte glycoprotein (MOG) and However, the scale score has not been validated yet.5 aquaporin-4 (AQP4) antibodies were examined with CBA in patients with overlapping encephalitis and demyelinating syn- 14 Here we report the sensitivity and specificity of the high scale drome. Antibodies against classical paraneoplastic in- score (≥5) in predicting C-NORSE at the early stage of SE-M tracellular antigens (CV2/CRMP5, Ma2, Ri, Yo, Hu, GAD65, of unclear etiology (before NS-Ab test results are known). and amphiphysin) were measured in serum at Kitasato Uni- versity with EUROLINE (Euroimmun AG) in patients when associated tumor was suspected or those with NORSE criteria. Methods Criteria for C-NORSE Patients selection and antibody assays Although C-NORSE is not a specific diagnosis, patients were (study profile) classified into C-NORSE as a subgroup of cryptogenic epi- We first reviewed the clinical information of 180 patients with leptic syndrome in this study if those fulfilled the following 3 seizures of unclear etiology on admission or early stage of criteria: (1) new-onset refractory SE in previously healthy seizures, in whom NS-Abs were examined to investigate po- individual, (2) refractoriness to conventional ASD treatment, tential immune-mediated etiologies between January 1, 2007, and (3) no etiology identified throughout the course of the and December 31, 2019 (figure 1). These patients were ad- disease. If the etiology of SE was identified, patients were mitted to Kitasato University Hospital or other associated diagnosed with etiology-based specific diagnosis (e.g., anti- hospitals between January 1, 1999, and December 31, 2019; in NMDAR encephalitis and anti-LGI1 encephalitis). SE was

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Figure 1 Study profile

The sensitivity and specificity of the clinically based scale score indicated in the text were assessed among 83 patients with SE with prominent motor symptoms. ILAE = In- ternational League Against Epilepsy; NORSE = new-onset refractory status epilepticus; SE = status epilepticus. considered as refractory when it continued longer than 60 only patients who had electroencephalographic correlates minutes despite adequate administration of benzodiazepines (such as spikes and waves or periodic discharges that explain and adequate loading of standard IV ASDs.2,6,15,16 The eti- prominent motor symptoms) were regarded to meet the first ology of NORSE was extensively investigated with CSF ex- clinical feature of the score. Accordingly, patients without amination, malignancy survey, and serologic testing including apparent electroencephalographic correlates despite convul- autoantibodies against neuronal surface and classical para- sive SE or epilepsia partialis continua were scored 0. Sym- neoplastic intracellular antigens. metric brain MRI abnormalities imply relatively symmetric increased diffusion-weighted image (DWI) or T2/fluid- C-NORSE score attenuated inversion recovery (FLAIR) signals in the hippo- C-NORSE score is a clinically based score (range 0–6) based campus, fimbria, amygdala, claustrum, insula, or perisylvian 5 on the following 6 clinical features usually obtained within 14 opercular cortex; these changes may not be seen at the onset days after admission in general hospital: (1) NORSE highly of SE-M but often subsequently develop associated with resistant to conventional ASD treatment, (2) previously persistent seizure activity.5 healthy individual before the onset of SE, (3) presence of prodromal high fever of unknown origin before the onset of Clinical assessments SE, (4) absence of prodromal psychobehavioral or memory We assessed the clinical features between patients with a alterations before the onset of SE, (5) absence of sustained high scale score (≥5) and those with a low scale score (≤4), orofacial-limb dyskinesias despite a profoundly decreased including sex, age at onset of SE-M, prodromal fever, pro- level of consciousness, and (6) symmetric brain MRI abnor- dromal psychobehavioral or memory alterations, in- malities (table 1). voluntary movements, mechanical ventilatory support, CSF and MRI findings, and presence of tumor. We reviewed the In the criteria, we previously defined that each feature rep- final diagnosis of these patients after extensive workup and resents 1 point, but the first 2 clinical features are mandatory.5 finally determined the sensitivity and specificity of the in- Accordingly, if either the first or second feature is absent, the dicated high scale score. In this study, to focus on the patient is scored 0. We applied 2015 ILAE criteria for SE6 to C-NORSE score, we did not assess the efficacy of treatment, include patients with SE-M, and all patients underwent EEG such as immunotherapy, or long-term outcome in these and MRI repeatedly during their hospitalization. However, patients.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 Table 1 Components of the C-NORSE score5

Clinical feature Value

1. New-onset refractory SE highly resistant to conventional ASD treatment 1

2. Previously healthy individual before the onset of SE 1

3. Presence of prodromal high fever of unknown origin before the onset of SE 1

4. Absence of prodromal psychobehavioral or memory alterations before the onset of SE 1

5. Absence of sustained orofacial-limb dyskinesias despite profoundly decreased level of consciousness 1

6. Symmetric DWI or T2/FLAIR hyperintensities 1

Total 6

Abbrevaitions: ASD = antiseizure drug; C-NORSE = cryptogenic new-onset refractory status epilepticus; DWI = diffusion-weighted image; FLAIR = fluid- attenuated inversion recovery; SE = status epilepticus. C-NORSE score is a clinically based score (range 0–6) based on the above 6 clinical features (slightly modified from the original one5). In the criteria, each feature represents 1 point, but the first 2 clinical features are mandatory. If either the first or second feature is absent, the patient is scored 0. In this scale score, refractory SE requires EEG correlates that explain prominent motor symptoms; a patient without EEG correlates is not regarded as C- NORSE. The sixth feature means relatively symmetric increased DWI or T2/FLAIR signals in the hippocampus, fimbria, amygdala, claustrum, insula or perisylvian opercular cortex; these changes may not be seen at the onset of SE but often subsequently develop associated with persistent seizure activity. The C-NORSE score should be used only to predict C-NORSE at the early state of SE-M of unclear etiology before antibody test results become available, but it should not be used to make a diagnosis (see Text).

Standard protocol approvals, registrations, patients with a high score had more frequent prodromal fever and patient consents (28/31 vs 24/52), mechanical ventilatory support (31/31 vs The study was approved by Institutional Review Boards of 36/52), and symmetric DWI or T2/FLAIR hyperintensities Kitasato University (B18-193). Written informed consent was (26/31 vs 12/52) than those with a low score. By contrast, they obtained from the patients or their family members. In- had less frequent involuntary movements (2/31 vs 30/52) and formation on symptoms, CSF, MRI, EEG, and treatments was absent prodromal psychobehavioral alterations (0/31 vs 27/ obtained from the authors or referring physicians. 52), CSF oligoclonal band (OCB) detection (0/27 vs 11/38), tumor association (0/31 vs 13/52), or NS-Abs (0/31 vs 29/52) Statistical analysis than those with a low score. There was no difference in pro- The Fisher exact test was performed for comparison of cat- dromal headache before the onset of SE, CSF pleocytosis, egorical variables, and the Mann-Whitney test was used for white blood cell (WBC) counts in CSF, CSF protein levels, or continuous variables. The statistical significance was set at p < elevated immunoglobulin G (IgG) index. 0.05. The sensitivity and specificity of the high C-NORSE score were determined with 2-way contingency table analysis. Final diagnosis 6 We used JMP, version 14 (SAS Institute Inc.), for statistical Of 83 patients with 2015 ILAE criteria for SE-M of unclear analyses. etiology on admission or early stage of SE, 29 (34.9%) patients were positive for NS-Abs, NMDAR in 26 patients (1 Data availability with concurrent AQP4 and 1 with MOG), LGI1 in 1, Any data not published within the article are available and will GABAbR in 1, and unknown antigens (not characterized yet) fi be shared anonymously by request from any qualified in 1. No AMPAR or GABAaR antibodies were identi ed. All investigator. antibody-positive patients had a low C-NORSE score: 24 patients had 0, and 5 patients had 3. The remaining 54 patients (65.1%) were negative for NS-Abs; 21 patients were Results diagnosed with miscellaneous disorders or syndrome including possible autoimmune encephalitis (AE)17 (n = 11), Clinical features in patients with a high score autoantibody-negative but probable AE17 (n = 5), antibody- and those with a low score negative autoimmune limbic encephalitis17 (n = 1), enceph- Of 83 patients, 31 (37.3%) had a high score (5–6); 17 patients alitis associated with systemic lupus erythematosus (n = 2), (54.8%) were female; median age at symptom onset was 27 and nonautoimmune neurologic disorders (n = 2). The years (range 5–73 years) (table 2). The remaining 52 patients remaining 33 patients were finally regarded as C-NORSE (62.7%) had a low score (0–4); 37 patients (71.2%) were based on the above criteria (figure 1). female; median age at symptom onset was 25 years (range 10–79 years). Other clinical information is shown in table 2. Clinical features of C-NORSE There was no difference between patients with a high score and Eighteen of 33 patients (54.5%) were female; median age at low score in female sex and median age at onset. However, onset was 27 years (range 5–73 years). Thirty-one patients

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Table 2 Clinical features in patients with a high score and those with a low score

High score (n = 31) Low score (n = 52) p Value

Female sex 17 (54.8%) 37 (71.2%) 0.1574

Median age at symptom onset (y) 27 (5–73) 25 (10–79) 0.4916

Prodromal fever of unknown origina 28 (90.3%) 24 (46.2%) <0.0001

Prodromal headachea 15/29 (51.7%) 23 (44.2%) 0.6431

Prodromal psychobehavioral or memory alterationsa 0 (0.0%) 27 (51.9%) <0.0001

Involuntary movements 2 (6.5%) 30 (57.7%) <0.0001

Mechanical ventilatory support 31 (100.0%) 36 (69.2%) 0.0003

Symmetric DWI or T2/FLAIR hyperintensities 26 (83.9%) 12 (23.1%) <0.0001

Tumor association 0 (0.0%) 13 (25.0%) 0.0014

CSF pleocytosis (>5 WBCs/μL) 22 (71.0%) 40 (76.9%) 0.6063

Median CSF WBC counts (WBCs/μL) 9(1–224) 17 (0–279) 0.2188

Median CSF protein (mg/dL) 41 (13–129) 35 (14–354) 0.2685

CSF OCB detection 0/27 (0.0%) 11/38 (28.9%) 0.0017

Elevated IgG index (≥0.74) 2/23 (8.7%) 10/38 (26.3%) 0.1113

NS-Ab detection 0 (0.0%) 29 (55.8%) <0.0001 a Prodromal symptoms mean symptoms/signs that developed before the onset of status epilepticus. Abbrevaitions: DWI = diffusion-weighted image; FLAIR = fluid-attenuated inversion recovery; IgG = immunoglobulin G; NS-Ab = neuronal surface antibody; OCB = oligoclonal band; WBC = white blood cell.

(93.9%) had a high score; 23 patients had 6, and 8 patients patients (45.5%), but follow-up MRIs showed abnormal had 5, but 2 patients had a low score (both 4). Of interest, 7 of findings in 30 patients (90.9%); in 27 patients (81.8%), brain the 33 patients (21.2%) had a past medical history (PMH) of MRIs showed symmetric DWI or T2/FLAIR hyperintensities febrile convulsion (FC), family history of FC or epilepsy, or in the medial temporal lobes, basal ganglia, fimbria, claustrum, both; 3 patients had a PMH of FC (one of them had a family or perisylvian opercular cortex (figure 2). None of these pa- history of FC); 4 patients had no PMH of FC but had a family tients had a tumor identified during the course of the disease. history of FC (n = 1) or epilepsy (n = 3). The sensitivity and specificity of the high Prodromal symptoms developed before the onset of SE in 31 C-NORSE score of 33 patients (93.9%), fever in 28 of 33 patients (84.8%), and The sensitivity and specificity of the high score (≥5) for headache in 15 of 31 patients (2 unknown). Only 1 patient predicting C-NORSE were 93.9% (95% CI 0.87–0.94) and (3.0%) developed psychobehavioral alterations before the 100% (95% CI 0.95–1.00), respectively. onset of SE, whereas 3 patients (9.1%) showed involuntary movements during the course of the disease, but only 1 patient developed sustained dyskinesias mimicking orofacial- Discussion limb dyskinesias. All patients required mechanical ventilatory support due to refractory SE. This study shows that (1) patients with the high score are more likely to have C-NORSE, (2) the clinically based score NS-Abs were not detected in either serum or CSF. Classical C-NORSE score has high sensitivity and specificity for pre- paraneoplastic antineuronal antibodies measured in serum in dicting the C-NORSE, and (3) patients with C-NORSE had 28 patients were negative but not measured in 5 (no serum distinctive clinical features. was available for examination). CSF examination revealed a median of 9 WBCs/μL (range 0–224 WBCs/μL) and a me- In clinical practice, it is important to estimate antibody status in dian protein level of 41 mg/dL (range 13–129 mg/dL). No patients with SE of unclear etiology and identify patients with CSF-restricted OCBs were detected in 29 examined patients, C-NORSE as early as possible because patients with C-NORSE whereas the IgG index was elevated in 2 of 25 examined are usually less responsive to first-line immunotherapy4,5 and patients (8.0%). Ten patients (30.3%) had no pleocytosis (>5 more likely to have poor long-term outcome with cognitive WBCs/μL). Initial brain MRI was unremarkable in 15 deficits and refractory partial seizures.5

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 5 Figure 2 Brain MRIs finding obtained from 3 patients with C-NORSE

Brain MRIs show symmetric increased DWI or FLAIR signals in the amygdala, hippocampus, fimbria, claustrum, insula, and frontotemporal cortex. Basal ganglia and perisylvian opercular cortex are also involved in patients with C-NORSE (not shown). (A) A 37-year- old man; (B) a 49-year-old woman; (C and D) a 39-year-old woman; (A–C) FLAIR, (D) DWI. C-NORSE = cryptogenic new-onset refractory status epilepticus; DWI = diffusion-weighted image; FLAIR = fluid- attenuated inversion recovery.

This scale score was originally developed based on our previous C-NORSE is a devastating epileptic syndrome of unknown – preliminary study5 that compared the clinical features of 11 adult causes, probably of diverse etiologies1 5 including autoim- patients with C-NORSE (aged ≥17 years) with those of 32 munity, neuroinflammation, or individual susceptibility to patients with anti-NMDAR encephalitis. We previously reported seizure. This study highlighted distinctive clinical features of that the C-NORSE score was higher in patients with C-NORSE C-NORSE phenotypically different from antibody-positive than those with anti-NMDAR encephalitis; however, the sensi- AE, such as anti-NMDAR, anti-LGI1, or anti-GABAaR en- tivity and specificity were not determined. After that, we had cephalitis. Patients with C-NORSE often present with high recruited additional patients since September 2016. In the fever of unknown cause, followed by sudden onset of mainly meantime, the international consensus definition of NORSE was convulsive seizures, leading to refractory SE (occasionally proposed in 20183; hence, the concept of C-NORSE was much super-refractory SE) requiring a mechanical ventilatory sup- more clearly defined than before. In this study, we adopted the port and continuous infusion of sedative drugs. Early brain concept of C-NORSE and included pediatric cases as well as MRI is often normal or may show symmetric DWI or T2/ newly identified adult cases. Accordingly, we increased the FLAIR hyperintensities in the medial temporal lobes,5 mim- number of patients with C-NORSE from 11 to 33. icking autoimmune limbic encephalitis. CSF examination often shows nonspecific mild pleocytosis; however, none of In this study, we assessed the sensitivity and specificity of the these patients had CSF-restricted OCBs, and the IgG index high score (≥5) in 83 patients with SE-M. In this cohort, the was not elevated in most of them. Of interest, prodromal sensitivity and specificity for predicting C-NORSE were psychobehavioral or memory alterations usually did not de- 93.9% and 100%, respectively, making it a useful diagnostic velop before the onset of SE or decreased level of con- tool at the early stage of SE-M of unclear etiology before sciousness. This is highly contrast to those with anti-NMDAR – antibody test results become available. encephalitis5,6,17 19 or autoimmune limbic encephalitis,17 in

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN whom these symptoms usually develop in the early course of variety of autoimmune inflammatory diseases.34 The effi- the disease, and often predominant. Thus, the lack of pro- cacy of tocilizumab, IL-6 receptor antagonist, has also been dromal psychobehavioral or memory alterations is an im- reported in patients with NORSE.35 Therefore, elevated portant feature in discrimination of C-NORSE from anti- CSF levels of proinflammatory cytokines may play an im- NMDAR encephalitis or limbic encephalitis. The follow-up portant role in neuroinflammation, leading to development brain MRIs often show symmetric brain lesions involving the of refractory partial seizures in NORSE or FIRES. In our hippocampus, amygdala, fimbria, claustrum, basal ganglia, cohort of patients with C-NORSE, none of them had au- insular cortex, and perisylvian opercular cortex presumably toantibodies binding to the neuronal surface membrane – associated with persistent seizure activity.5,20 22 Such neuro- with a rat brain immunohistochemistry in either CSF or imaging pattern is quite different from autoimmune limbic serum, indicating that autoantibodies may not play an im- encephalitis with highly restricted to bilateral medial temporal portant role in C-NORSE or FIRES, but rather innate im- lobes17 or anti-GABAaR encephalitis with multiple cortico- munity may be more important than adaptive immunity as subcortical lesions.13,23 Involuntary movements may develop previously described.5 in patients with C-NORSE due to secondary basal ganglia lesions, but not like NMDAR-associated orofacial-limb dys- Of interest, 21.2% of patients with C-NORSE had a PMH of kinesias18 or movement disorders,24 or LGI1-associated FC, family history of FC, or both. In a small group of patients, faciobrachial dystonic seizures.25 some genetic predisposition to epileptic seizure might contribute to development of NORSE following fever. Further – The etiology of C-NORSE remains unknown.1 5 It is also research is required to determine a role of genomic suscep- controversial whether it is of autoimmune origin.5 One might tibility to NORSE. argue that C-NORSE is an epileptic syndrome and should not be confused with AE; randomized controlled trial with im- This study has limitations of being retrospective studies and munotherapy has not been conducted yet; therefore, little based on the small number of patients included. Genomic information is available on the adequate dosage of other im- studies have not been performed yet in our cohort. Classical mune treatments to formulate any recommendation.3 How- paraneoplastic antineuronal antibodies were not examined in ever, it is not easy in clinical practice to exclude a possibility of all patients. Cytokine or chemokines were not examined in C-NORSE or antibody-positive AE particularly at the early either case. In an emergency situation, some of the compo- stage of SE before antibody test results become available; nents of the score may be difficult to assess historically due to therefore, many patients with NORSE may have been treated a variety of individual factors. A brain MRI is often difficult to with immunotherapy,5,26 although the first-line immuno- obtain in a ventilated patient with SE-M or cannot be per- therapy is presumed to be less effective. However, if the formed on a patient with contraindication (e.g., implanted C-NORSE score is high (≥5) on referral from other hospital, pacemakers, intracranial aneurysm clips, and iron-based metal it is suggested that the patient is more likely to be negative for implants). When early brain MRI is unremarkable, repeated neuronal antibodies, thus more likely to be less responsive to studies are required to see symmetric MRI abnormalities. first-line immunotherapy and have poor outcome. This However, a brain MRI within the first 24 hours is currently scoring strategy might help physicians to identify patients included in the diagnostic checklist for etiology of NORSE,36 with C-NORSE and their decision making in a patient with and follow-up MRI is also important in exclusion of alterna- the high score. tive diagnosis (multifocal corticosubcortical lesions may appear in the course of the disease in anti-GABAaR Although the underlying mechanism of C-NORSE is en- encephalitis). It is important to keep in mind that this score tirely unknown, inflammation-mediated epileptogenesis has was developed in patients with SE-M of unclear etiology. been proposed,27 in which a vicious cycle that involves in- Thus, the results should not be generalized for patients with flammationandseizureactivityisassumedtoleadtocell NCSE. death and network reorganization, ultimately causing refractory seizure. One previous study reported high levels of Despite these limitations, this study demonstrated that the cytokines (interleukin-6 [IL-6]) or chemokines (CXCL10 clinically based score is useful for early identification of pa- and IL-8) in serum and CSF in pediatric cases of febrile tients with C-NORSE. However, this score should not be infection-related epilepsy syndrome (FIRES),28 which is used to make the diagnosis of C-NORSE because NORSE is currently regarded as a subcategory of NORSE.3 Among not a specific diagnosis and exclusion of alternative diagnosis those, proinflammatory cytokines, such as IL-1β and IL-6, is mandatory. In patients with C-NORSE, irreversible brain have received attention as potential key molecules in damage is expected to occur quickly; thus, early recognition of C-NORSE. IL-1β has been implicated in seizure-induced C-NORSE is crucial. In addition to ASD treatment, we hope neuronal cell death,29 SE,30 and posttraumatic epilepsy.31 that this scoring strategy improves their functional outcome Anakinra, IL-1 receptor antagonist, has been reported to through facilitating early intervention with potential effective be effective in patients with FIRES.32,33 IL-6 secreted from drugs that break a vicious cycle of neuroinflammation- macrophages is also important mediator of fever and its induced neuronal damage that consequently increases seizure deregulated expression is responsible for development of a susceptibility.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 7 Acknowledgment The authors particularly thank Dr. Josep Dalmau (Service of Appendix (continued) ı Neurology, IDIBAPS Hospital Cl´nic, University of Barce- Name Location Contribution lona, Barcelona, Spain) for examining antibodies against neuronal surface antigens and critical comments on this study. Ryoko Iwase, Kitasato University Major role in the acquisition MD School of Medicine, of data and interpreted the They also thank Drs. Hiroki Asari (Shizuoka City Shimizu Sagamihara, Japan data Hospital), Junya Kaneko (Nippon Medical School Tama Hiroki Suga, Kitasato University Major role in the acquisition Nagayama Hospital), Kenji Yoshida, Kasumi Hattori, and MD School of Medicine, of data and interpreted the Yuya Itagaki (Fukushima Medical University Hospital), Sagamihara, Japan data

Arifumi Kosakai (Keiyu Hospital), Masashi Watanabe (Ehime Yutaka Kitasato University Major role in the acquisition Prefectural Central Hospital), Takeo Shishido (Hiroshima Nonoda, MD, School of Medicine, of data and interpreted the University Hospital), Hiroya Ohara (South Nara General PhD Sagamihara, Japan data Medical Center), Yasushi Hosoi (Hamamatsu University Yuya Kitasato University Major role in the acquisition Onozawa, Hospital, Sagamihara, of data and interpreted the Hospital), Mizuki Ayano (Kyorin University Hospital), Yuka PhD Japan data Terasawa (The Jikei University Hospital), Keisuke Imai (Japanese Red Cross Kyoto Daiichi Hospital), and Masamune Eiji Kitasato University Major role in the acquisition Kitamura, School of Medicine, of data and interpreted the Sakamoto (Yokohama City University Medical Center) for MD, PhD Sagamihara, Japan data providing clinical information. They are grateful to all Kazutoshi Kitasato University Major role in the acquisition participants and physicians for their contribution to this study. Nishiyama, School of Medicine, of data and interpreted the MD, PhD Sagamihara, Japan data

Study funding Takahiro Kitasato University Designed and This study was supported in part by a grant from the Japan Iizuka, MD School of Medicine, conceptualized the study; Sagamihara, Japan major role in the acquisition Epilepsy Research Foundation (JERFTENKAN 17002, TI). of data; analyzed and interpreted the data; and drafted and revised the Disclosure manuscript for intellectual A. Yanagida, N. Kanazawa, J. Kaneko, A. Kaneko, R. Iwase, H. content Suga, Y. Nonoda, Y. Onozawa, and E. Kitamura report no disclosure. K. Nishiyama received research support from Daiichi Sankyo Co., Ltd., Otsuka Pharmaceutical Co., Ltd., References Dainippon Sumitomo Pharma Co., Ltd., and Eisai Co., Ltd. T. 1. Wilder-Smith EP, Lim EC, Teoh HL, et al. The NORSE (new-onset refractory status epilepticus) syndrome: defining a disease entity. Ann Acad Med Singapore 2005;34: Iizuka received a grant from The Japan Epilepsy Research 417–420. Foundation and research support from Astellas Pharma Inc. 2. Gaspard N, Foreman BP, Alvarez V, et al. New-onset refractory status epilepticus: etiology, clinical features, and outcome. Neurology 2015;85:1604–1613. Go to Neurology.org/NN for full disclosures. 3. Hirsch LJ, Gaspard N, van Baalen A, et al. Proposed consensus definitions for new- onset refractory status epilepticus (NORSE), febrile infection-related epilepsy syndrome (FIRES), and related conditions. Epilepsia 2018;59:739–744. Publication history 4. Gaspard N, Hirsch LJ, Sculier C, et al. New-onset refractory status epilepticus Received by Neurology: Neuroimmunology & Neuroinflammation (NORSE) and febrile-infection-related epilepsy syndrome (FIRES): state of the art fi and perspectives. Epilepsia 2018;59:745–752. April 29, 2020. Accepted in nal form May 22, 2020. 5. Iizuka T, Kanazawa N, Kaneko J, et al. Cryptogenic NORSE: its distinctive clinical features and response to immunotherapy. Neurol Neuroimmunol Neuroinflamm 2017;4e:e396. doi: 10.1212/NXI.0000000000000396. 6. Trinka E, Cock H, Hesdorffer D, et al. A definition and classification of status Appendix Authors epilepticus-report of the ILAE task force on classification of status epilepticus. Epi- lepsia 2015;56:1515–1523. Name Location Contribution 7. Dalmau J, Tuzün̈ E, Wu HY, et al. Paraneoplastic anti-N-methyl-D-aspartate receptor encephalitis associated with ovarian teratoma. Ann Neurol 2007;61:25–36. Atsuko Kitasato University Designed and 8. Lai M, Hughes EG, Peng X, et al. AMPA receptor antibodies in limbic encephalitis Yanagida, School of Medicine, conceptualized the study; alter synaptic receptor location. Ann Neurol 2009;65:424–434. MD Sagamihara, Japan acquisition of data; analyzed 9. Lancaster E, Lai M, Peng X, et al. Antibodies to the GABA (B) receptor in limbic and interpreted the data; encephalitis with seizures: case series and characterisation of the antigen. Lancet and drafted and revised the Neurol 2010;9:67–76. manuscript for intellectual 10. Lai M, Huijbers MG, Lancaster E, et al. Investigation of LGI1 as the antigen in limbic content encephalitis previously attributed to potassium channels: a case series. Lancet Neurol 2010;9:776–785. Naomi Kitasato University Major role in the acquisition 11. Lancaster E, Huijbers MG, Bar V, et al. Investigations of caspr2, an autoantigen of Kanazawa, School of Medicine, of data; analyzed and encephalitis and neuromyotonia. Ann Neurol 2011;69:303–311. BS Sagamihara, Japan interpreted the data; and 12. Boronat A, Gelfand JM, Gresa-Arribas N, et al. Encephalitis and antibodies to revised the manuscript for dipeptidyl-peptidase-like protein-6, a subunit of Kv4.2 potassium channels. Ann intellectual content Neurol 2013;73:120–128. 13. Petit-Pedrol M, Armangue T, Peng X, et al. Encephalitis with refractory seizures, Juntaro Kitasato University Major role in the acquisition status epilepticus, and antibodies to the GABAA receptor: a case series, character- Kaneko, MD School of Medicine, of data and interpreted the isation of the antigen, and analysis of the effects of antibodies. Lancet Neurol 2014;13: Sagamihara, Japan data 276–286. 14. Titulaer MJ, Höftberger R, Iizuka T, et al. Overlapping demyelinating syndromes and Atsushi Kitasato University Major role in the acquisition anti–N-methyl-D-aspartate receptor encephalitis. Ann Neurol 2014;75:411–428. Kaneko, MD School of Medicine, of data and interpreted the 15. Mayer SA, Claassen J, Lokin J, Mendelsohn F, Dennis LJ, Fitzsimmons BF. Refractory Sagamihara, Japan data status epilepticus: frequency, risk factors, and impact on outcome. Arch Neurol 2002; 59:205–210.

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN 16. Brophy GM, Bell R, Claassen J, et al. Guidelines for the evaluation and management of 28. Sakuma H, Tanuma N, Kuki I, Takahashi Y, Shiomi M, Hayashi M. Intrathecal status epilepticus. Neurocrit Care 2012;17:3–23. overproduction of proinflammatory cytokines and chemokines in febrile infection- 17. Graus F, Titulaer JM, Balu R, et al. Clinical approach to diagnosis of autoimmune related refractory status epilepticus. J Neurol Neurosurg Psychiatry 2015;86:820–822. encephalitis. Lancet Neurol 2016;15:391–404. 29. Medel-Matus JS, Alvarez-Croda´ DM, Mart´ınez-Quiroz J, Beltrán-Parrazal L, Mor- 18. Iizuka T, Sakai F, Ide T, et al. Anti-NMDA receptor encephalitis in Japan: long-term gado-Valle C, López-Meraz ML. IL-1β increases necrotic neuronal cell death in the outcome without tumor removal. Neurology 2008;70:504–511. developing rat hippocampus after status epilepticus by activating type I IL-1 receptor 19. Dalmau J, Gleichman AJ, Hughes EG, et al. Anti-NMDA-receptor encephalitis: case (IL-1RI). Int J Dev Neurosci 2014;38:232–240. series and analysis of the effects of antibodies. Lancet Neurol 2008;7:1091–1098. 30. Tian DS, Peng J, Murugan M, et al. Chemokine CCL2-CCR2 signaling induces 20. Chevret L, Husson B, Nguefack S, Nehlig A, Bouilleret V. Prolonged refractory status neuronal cell death via STAT3 activation and IL-1β production after status epi- epilepticus with early and persistent restricted hippocampal signal MRI abnormality. lepticus. J Neurosci 2017;37:7878–7892. J Neurol 2008;255:112–116. 31. Sharma R, Leung WL, Zamani A, O’Brien TJ, Casillas Espinosa PM, Semple BD. 21. Chatzikonstantinou A, Gass A, Förster A, Hennerici MG, Szabo K. Features of acute Neuroinflammation in post-traumatic epilepsy: pathophysiology and tractable ther- DWI abnormalities related to status epilepticus. Epilepsy Res 2011;97:45–51. apeutic targets. Brain Sci 2019;9:318. 22. Meletti S, Slonkova J, Mareckova I, et al. Claustrum damage and refractory status 32. Kenney-Jung DL, Vezzani A, Kahoud RJ, et al. Febrile infection-related epilepsy epilepticus following febrile illness. Neurology 2015;85:1224–1232. syndrome treated with anakinra. Ann Neurol 2016;80:939–945. 23. Spatola M, Petit-Pedrol M, Simabukuro MM, et al. Investigations in GABAA receptor 33. Westbrook C, Subramaniam T, Seagren RM, et al. Febrile infection-related epilepsy antibody-associated encephalitis. Neurology 2017;88:1012–1020. syndrome treated successfully with anakinra in a 21-year-old woman. WMJ 2019;118: 24. Varley JA, Webb AJS, Balint B, et al. The movement disorder associated with NMDAR 135–139. antibody-encephalitis is complex and characteristic: an expert video-rating study. 34. Kishimoto T, Kang S, Tanaka T. IL-6: a new era for the treatment of autoimmune J Neurol Neurosurg Psychiatry 2019;90:724–726. inflammatory diseases. In: Nakao K, Minato N, Uemoto S, editors. Innovative 25. Irani SR, Michell AW, Lang B, et al. Faciobrachial dystonic seizures precede Lgi1 Medicine: Basic Research and Development [online]. Tokyo: Springer; 2015. antibody limbic encephalitis. Ann Neurol 2011;69:892–900. 35. Jun JS, Lee ST, Kim R, Chu K, Lee SK. Tocilizumab treatment for new onset re- 26. Gugger JJ, Husari K, Probasco JC, Cervenka MC. New-onset refractory status epi- fractory status epilepticus. Ann Neurol 2018;84:940–945. lepticus: a retrospective cohort study. Seizure 2020;74:41–48. 36. Gofton TE, Gaspard N, Hocker SE, Loddenkemper T, Hirsch LJ. New onset re- 27. Nabbout R, Vezzani A, Dulac O, Chiron C. Acute encephalopathy with inflammation- fractory status epilepticus research: what is on the horizon? Neurology 2019;92: mediated status epilepticus. Lancet Neurol 2011;10:99–108. 802–810.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 9 ARTICLE OPEN ACCESS High-throughput investigation of molecular and cellular biomarkers in NMOSD

Soumya S. Yandamuri, MSE, PhD, Ruoyi Jiang, MS, Aditi Sharma, MBBS, Elizabeth Cotzomi, BS, Correspondence ’ Chrysoula Zografou, PhD, Anthony K. Ma, MS, Jessica S. Alvey, MS, Lawrence J. Cook, PhD, Terry J. Smith, MD, Dr. O Connor [email protected] Michael R. Yeaman, PhD, and Kevin C. O’Connor, PhD, On behalf of the Guthy-Jackson Charitable Foundation CIRCLES Study Group

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e852. doi:10.1212/NXI.0000000000000852 Abstract Objective To identify candidate biomarkers associated with neuromyelitis optica spectrum disorder (NMOSD) using high-throughput technologies that broadly assay the concentrations of serum analytes and frequencies of immune cell subsets.

Methods Sera, peripheral blood mononuclear cells (PBMCs), and matched clinical data from participants with NMOSD and healthy controls (HCs) were obtained from the Collaborative International Research in Clinical and Longitudinal Experience Study NMOSD biorepository. Flow cytometry panels were used to measure the frequencies of 39 T-cell, B-cell, regulatory T-cell, monocyte, natural killer (NK) cell, and dendritic cell subsets in unstimulated PBMCs. In parallel, multiplex proteomics assays were used to measure 46 serum cytokines and chemokines in 2 independent NMOSD and HC cohorts. Multivariable regression models were used to assess molecular and cellular proﬁles in NMOSD compared with HC.

Results NMOSD samples had a lower frequency of CD16+CD56+ NK cells. Both serum cohorts and multivariable logistic regression revealed increased levels of B-cell activating factor associated with NMOSD. Interleukin 6, CCL22, and CCL3 were also elevated in 1 NMOSD cohort of the 2 analyzed. Multivariable linear regression of serum analyte levels revealed a correlation between CX3CL1 (fractalkine) levels and the number of days since most recent disease relapse.

Conclusions Integrative analyses of cytokines, chemokines, and immune cells in participants with NMOSD and HCs provide congruence with previously identiﬁed biomarkers of NMOSD and highlight CD16+CD56+ NK cells and CX3CL1 as potential novel biomarker candidates.

From the Department of Neurology (S.S.Y., A.S., E.C., C.Z., K.C.O.C.), Department of Immunobiology (R.J., K.C.O.C.), and Department of Pathology (A.K.M.), Yale School of Medicine, New Haven, CT; University of Utah School of Medicine (J.S.A., L.J.C.), Salt Lake City; Departments of Ophthalmology and Visual Sciences and Internal Medicine (T.J.S.), University of Michigan Medical School, Ann Arbor; Department of Medicine (M.R.Y.), David Geffen School of Medicine at the University of California, Los Angeles; Divisions of Molecular Medicine & Infectious Diseases (M.R.Y.), Harbor-UCLA Medical Center, Torrance; and Lundquist Institute for Biomedical Innovation at Harbor-UCLA Medical Center (M.R.Y.), Torrance.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

Guthy-Jackson Charitable Foundation CIRCLES Study Group coinvestigators are listed in the appendix 2 at the end of the article.

The Article Processing Charge was funded by the Guthy-Jackson Charitable Foundation. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary AQP4 = aquaporin-4; ARR = annual relapse rate; AUC = area under the curve; BAFF = B-cell activating factor; CIRCLES = Collaborative International Research in Clinical and Longitudinal Experience Study; HC = healthy control; IgG = immunoglobulin G; IL = interleukin; MOG = myelin oligodendrocyte glycoprotein; NK = natural killer; NMOSD = neuromyelitis optica spectrum disorder; PBMC = peripheral blood mononuclear cell; ROC = receiver operating characteristic.

Neuromyelitis optica spectrum disorder (NMOSD) is a rare au- Jackson Charitable Foundation CIRCLES study19 or through a toimmune demyelinating disease of the optic nerve and spinal biorepository established at the Yale University School of cord.1,2 Autoantibodies directed at the aquaporin-4 water channel Medicine, Department of Neurology. Written informed con- (AQP4), expressed on astrocytes in the CNS, have been identified sent was obtained from all study participants before sample as key contributors to NMOSD pathogenesis and are called collection. – AQP4 immunoglobulin G (IgG).3 7 Approximately 70% of patients with NMOSD are seropositive for AQP4-IgG, which appear Two sets of serum samples were separately obtained and to contribute to disease by causing internalization of AQP4 and analyzed independently to avoid batch effect. These 2 serum – complement fixation.8 10 Twenty to 50% of patients in whom sets were termed cohort 1 and cohort 2. AQP4-IgG cannot be detected are seropositive for myelin oligodendrocyte glycoprotein (MOG) autoantibodies (MOG- Flow cytometry immunophenotyping of – IgG).11 13 The remainder of NMOSD patients appear to be se- cell subsets ronegative for both AQP4-IgG and MOG-IgG. Neither autoan- Five flow cytometry panels, previously validated by the Human 20 tibody has been convincingly demonstrated to predict relapse, Immune Phenotyping Consortium, were used to examine the – response to therapy, or prognosis.14 18 Given the cumulative and frequencies of 39 immune cell subsets (table e-1, links.lww. potentially devastating nature of relapses, a prognostic biomarker com/NXI/A292) in unstimulated PBMCs. These panels were would be especially valuable to herald imminent relapses and developed to standardize routine immunophenotyping in hu- guide treatment interventions. Moreover, specificbiomarkersmay mans across various sites. All samples analyzed were verified to reveal disease mechanisms, inform disease status, provide insight have a PBMC viability greater than 80%, assessed by for development of therapies, and help monitor treatment efficacy. 7-aminoactinomycin D staining (Thermo Fisher, Waltham, MA). Briefly, samples were thawed and incubated in the dark Biomarker discovery in NMOSD is hampered by disease rarity. with viability dye for 20 minutes. Following washing, PBMCs 5 Furthermore, studies often use an a priori single-candidate ap- were split into 5 × 10 cells for each of the 5 panels. Cells were proach to biomarker discovery, limiting the probability of dis- incubated with an antibody cocktail respective of each panel covery of a potential novel biomarker. Although imaging (table e-1, links.lww.com/NXI/A292) at 4°C and then ana- techniques and interrogation of CSF may provide more direct lyzed on a BD Biosciences LSR Fortessa cytometer. information on CNS disease, the collection of blood samples is less invasive, time intensive, and costly. Therefore, we sought to Evaluation of serum cytokines perform an unbiased, discovery-based evaluation of candidate and chemokines biomarkers that can be accessed from peripheral blood samples. The concentrations of 46 soluble circulating cytokines/chemo- To this end, we acquired peripheral blood mononuclear cells kines were measured using customized multiplex proteomics (PBMCs), serum, and clinical data from the Collaborative In- assays (R&D Systems Human Magnetic Luminex Assay kit; ternational Research in Clinical and Longitudinal Experience vendor catalog no. CUST0I704/QT84038/2, LXSAHM) for Study (CIRCLES) NMOSD biorepository.19 We applied high- each cohort. The assays were conducted according to the man- ’ fl throughput technologies to assess these biospecimens using a ufacturer s instructions. Brie y, diluted serum samples or stan- simultaneous, hypothesis-generating strategy focused on a large dards were added to individual Luminex wells and incubated with set of immune cell populations, cytokines, and chemokines. the microparticle cocktail for 2 hours with agitation. Sample wells Statistical methods were then used to compare cell subset fre- were washed and incubated in the dark with diluted biotin- quencies and serum analyte concentrations in NMOSD vs antibody cocktail for 1 hour at room temperature with agitation. healthy controls (HC). In addition, clinical metadata were in- After washing, streptavidin-PE was added to each well and in- tegrated into statistical modeling to assess potential relationships cubated in the dark for 30 minutes with agitation. Last, magnetic between molecular and cellular profiles and NMOSD relapses. microparticles were resuspended and read using a Luminex analyzer. The concentration of each serum analyte was then quantified using its respective standard curve. Methods In cohort 1, B-cell activating factor (BAFF), which was not Study participants and biospecimens available in that Luminex kit, was analyzed using a commercial PBMC and serum specimens from participants with NMOSD ELISA kit (R&D Systems, Minneapolis, MN) following man- and HCs were collected and archived through the Guthy- ufacturer instructions.

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Statistical analysis in table 1. The mean age of each NMOSD and HC group for Flow cytometry frequencies were calculated using FlowJo v10.4. the 3 analyses ranged from 31.5 to 54.5 years. The mean disease The percent frequency of each cell subset was determined by duration ranged from 5.40 to 6.42 years. The majority of par- calculating the fraction of the subset population relative to its ticipants were female (between 67% and 84% per group). parent population. Univariate statistical analyses on immuno- NMO-IgG serostatus was distributed between seronegative, phenotyping and serum analyte data were performed using the seropositive, and unknown for all NMOSD groups. Welch t test, and 95% CIs of the differences of the means (with mean difference defined as meanNMOSD—meanHC) were Primary reported race or ethnicity was heterogeneous between calculated in R. To correct for multiple hypotheses testing, q- groups, as was treatment history. At least 50% of the partici- value or p-adjusted (p-adj) was calculated while maintaining a pants with NMOSD in each group had been treated with rit- 5% false discovery rate using the Storey algorithm in R. For uximab. Varying frequencies of participants had been treated univariate analyses of serum analytes, p-adj ≤ 0.05 was inter- with corticosteroid, IVIg, plasma exchange, or other immuno- preted as significant, whereas for immunophenotyping, p ≤ 0.05 therapies. Participants were also heterogeneous in terms of was interpreted as significant. Further evaluation of univariate primary symptoms and phenotypic presentation including analyses based on rituximab treatment was performed using the optic neuritis or longitudinally extensive transverse myelitis. Welch t test, and p ≤ 0.05 was interpreted as significant. CD16+CD56+ NK cell frequency is reduced in Multivariate statistics NMOSD vs HCs To permit comparisons across serum analyte cohorts, nor- We next sought to determine whether differences in immune malization was performed by z-score normalizing assay values cell frequencies between participants with NMOSD and HCs (subtracting the mean and dividing by the SD). could be identified. To that end, 5 validated flow cytometry panels that applied defined markers were used to delineate Multivariable logistic or linear regression was performed using subsets of T lymphocytes, B lymphocytes, monocytes, natural the glm function in base R v3.4.2 for the prediction of NMOSD killer (NK) cells, and dendritic cells in unstimulated PBMCs. vs HC or for detecting associations with NMOSD clinical data Panel 1 measured central, effector, and naive memory T cells; + such as days since relapse or annual relapse rate (ARR). All panel 2 measured CD4 Th1, Th2, and Th17 cells, as well as + available serum analyte or immunophenotyping frequencies their CD8 counterparts defined by the same markers, namely + were used as variables in these models, and no data were Th1-like, Th2-like, and Th17-like CD8 cells; panel 3 mea- missing during model training. R2 (1 subtracted by the ratio of sured regulatory T-cell subsets; panel 4 measured B-cell null deviance to residual deviance) was used to evaluate for subsets; and panel 5 measured monocyte, NK cell, and den- goodness of fit of multivariable linear regression models. Five- dritic cell subsets (table e-1, links.lww.com/NXI/A292). A fold cross validation was performed to assess models for total of 39 cell subset frequencies were analyzed in the 5 overfitting of multivariable logistic regression models. A re- panels (table 2). ceiver operating characteristic (ROC) curve was computed, and the area under the curve (AUC) was calculated to assess The frequency of 3 populations was significantly different + + performance using the pROC v1.15.3 R package. The R between participants with NMOSD and HCs. CD16 CD56 packagebootv1.3-22wasusedtocomputeCIsforAUC. NK cell frequency was decreased by 54% in NMOSD vs HC Principal component analysis was performed using base R. (95% CI −56.8 to −10.4, p = 0.0067). A reduction in + + Mean and 95% CIs of the mean (95% CImean) were calculated CD16 CD56 NK cell frequency was still seen in the partic- for days since relapse and ARR. All codes used in this study are ipants with NMOSD who had not been treated by rituximab available on request. A p value ≤0.05 was interpreted as sta- (95% CI −69.7 to −6.45, p = 0.024), though not in those who tistically significant. had been treated by rituximab (95% CI −62.9 to 4.72, p = 0.083). The frequency of Th1-like CD8+ T cells was also Data availability reduced in NMOSD vs HC by 53% (95% CI −3.98 to −0.192, Anonymized data will be shared on request from qualified p = 0.034). Stratification analyses based on rituximab treat- investigators. ment revealed that both rituximab-treated (95% CI −4.06 to −0.176, p = 0.035) and untreated (95% CI −4.15 to −0.451, p = 0.019) participants with NMOSD had reduced Th1-like CD8+ Results T cells compared with HCs. Last, the frequency of B cells overall was also lower in NMOSD than HC by 54% (95% CI Clinical and demographic characteristics of −6.33 to −0.702, p = 0.017). However, further analysis revealed participants with NMOSD and HCs that rituximab-untreated participants had a similar frequency PBMC samples were provided by 12 participants with of B cells as HCs (95% CI −4.72 to 3.28, p = 0.685), and it was NMOSD and HCs. Serum samples were provided by 27 par- the rituximab-treated group who had the reduced B-cell fre- − ticipants with NMOSD and 11 HCs for cohort 1 and 29 par- quency (95% CI −7.96 to −4.66, p = 3.5 × 10 6). All other cell ticipants with NMOSD and 11 HCs for cohort 2. Demographic subset frequencies were similar in NMOSD compared and clinical characteristics of study participants are summarized with HC.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 Table 1 Clinical and demographic characteristics of NMOSD and HC participants

Immunophenotyping Cohort 1: serum Cohort 2: serum

HC (n = 12) NMOSD (n = 12) HC (n = 11) NMOSD (n = 27) HC (n = 11) NMOSD (n = 29)

Age, y, mean ± SD 48.6 ± 16.7 44.3 ± 13.7/11 31.5 ± 13.1 39.5 ± 12.9/24 54.5 ± 17.5 31.6 ± 16.8

Disease duration, y, — 5.55 ± 4.36/11 — 6.42 ± 6.78/22 — 5.40 ± 5.74/26 mean ± SD

Sex, n (%)

Female 8 (67) 10 (83) 8 (73) 21/25 (84) 8 (73) 21 (72)

Male 4 (33) 2 (17) 3 (27) 4/25 (16) 3 (27) 8 (28)

Serostatus,a n(%)

AQP4-IgG+ — 5 (42) — 9 (33) — 17 (59)

2 AQP4-IgG — 3 (25) — 7 (26) — 7 (24)

AQP4-IgG unknown — 4 (33) — 11 (41) — 5 (17)

Race/ethnicity, n (%)

Caucasian 9 (75) 10 (83) 8 (73) 7/21 (33) 8 (72) 20 (69)

Black/African American 0 (0) 1 (8.3) 0 (0) 9/21 (43) 0 (0) 3 (10)

Latino/Hispanic 0 (0) 1 (8.3) 1 (9) 3/21 (14) 1 (9.0) 5 (17)

Asian 3 (25) 0 (0) 2 (18) 3/21 (14) 2 (18) 0 (0)

Other b 0 (0) 0 (0) 1 (9) 0/21 (0) 0 (0) 1 (3.4)

Treatment, n (%)

Rituximab — 6 (50) — 20 (74) — 16 (55)

Corticosteroids — 8 (67) — 13/23 (57) — 8 (28)

Other immunotherapies — 6 (50) — 1/11 (9.1) — 8 (28)

IVIg — 2 (17) — 0/11 (0) — 1 (3.4)

PLEX — 1 (8.3) — 1/11 (9.1) — 1 (3.4)

Abbreviations: AQP4-IgG = aquaporin-4 immunoglobulin G; HC = healthy control; IVIg = IV immunoglobulin; NMOSD = neuromyelitis optica spectrum disorder; PLEX = plasma exchange. If demographics or clinical data are unknown for some samples in a group, they are depicted as a fraction with the number of known samples indicated in the denominator. For example, 1/11 (9.1) for PLEX under treatment means that of the 11 samples for which PLEX treatment status is known, 1, or 9.1%, had undergone PLEX treatment. a Serostatus was determined using standard methods approved for NMOSD diagnosis and confirmed by the site investigator. b The “Other” category includes American Indian, First Canadian, Alaskan Native, Hawaiian Native, Pacific Islander, or another specified racial/ethnic identity.

− Analysis of serum analytes validates 95% CI 769–1,650, p =1.3×10 5; cohort 2: 95% CI 2,740–6,190, − NMOSD biomarkers p = 4.7 × 10 4) in comparison to HC. In cohort 2, the serum Serum analyte concentrations were measured using multiplex BAFF concentration was elevated in rituximab-treated NMOSD proteomics assays and ELISA in 2 independent cohorts. A total of compared with untreated (95% CI 1740–5,620, p = 0.00065). In 46 serum cytokines or chemokines were measured in each cohort; summary, the rituximab-treated and unstratified NMOSD groups 4 of these differed significantly between NMOSD and HC in one had significantly elevated serum BAFF in both cohorts. The use of or both cohorts (table 3). BAFF was elevated in NMOSD samples 2 independent cohorts and 2 different methods (ELISA and − in cohorts 1 (95% CI 671–1,430, p-adj = 1.6 × 10 4) and 2 (95% multiplex proteomics assay) validated BAFF as a significant result. CI 1,360–3,880, p-adj = 0.0028). As rituximab treatment has been shown to elevate serum BAFF,18 we evaluated whether BAFF The 3 other serum analytes that differed between NMOSD and serum concentration differed based on previous rituximab treat- HC were interleukin (IL)-6 (95% CI 0.500–1.97, p-adj = 0.013), ment. The groups, arranged in ascending order of BAFF con- CCL22 (95% CI 104–504, p-adj = 0.019), and CCL3 (95% CI centration, are HC, untreated NMOSD, all (unstratified) 8.49–50.7, p-adj = 0.027), which were all elevated in cohort 2 NMOSD, and treated NMOSD. In both cohorts, BAFF con- only. IL-17F, S100B, FGF-basic, IL-15, RANTES, and IL-1RA centration was elevated in rituximab-treated NMOSD (cohort 1: did not meet the threshold for significance (p-adj ≤ 0.05), but

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Table 2 Mean immune cell frequencies in NMOSD and HC PBMC samples

Cell subset HC NMOSD Difference (95% CI) p Value p-adj

CD16+CD56+ NK/NK cells 65.4 31.9 256.8 to 210.4 0.0067 0.31

2 B cells/viable CD3 6.46 2.94 26.33 to 20.702 0.017 0.31

Th1-like/CD8+ T cellsa 3.95 1.87 23.98 to 20.192 0.034 0.45

Central memory/CD8+ T cells 5.08 9.57 −1.03 to 9.99 0.10 0.70

2 CD16 CD56 bright 3.71 5.85 −0.479 to 4.75 0.10 0.70 NK/NK cells

Naive/B cells 61.7 45.8 −37.3 to 5.68 0.13 0.70

Memory CCR4+/Tregs 0.88 0.295 −1.39 to 0.225 0.14 0.70

2 CXCR5+PD1 /CD4+ Tcells 5.40 3.82 −3.77 to 0.603 0.15 0.70

Activated/CD8+ T cells 1.22 2.75 −0.771 to 3.82 0.17 0.70

T follicular 1.31 0.541 −1.96 to 0.414 0.18 0.70 helper/CD4+ T cells

Activated CCR4+/Tregs 0.129 0.383 −0.133 to 0.641 0.18 0.70

CD8+ T/viable CD3+ 26.2 22.4 −10.4 to 2.72 0.24 0.74

Treg/CD4+ T cells 1.96 2.66 −0.582 to 1.99 0.26 0.74

2 2 2 NK cells/viable CD3 19 20 31.5 25.6 −16.7 to 4.83 0.26 0.74

Effector/CD4+ T cells 5.97 10.8 −5.63 to 15.2 0.35 0.83

Effector memory/CD8+ T cells 42.4 36.2 −19.8 to 7.25 0.35 0.83

Th2-like/CD8+ T cellsa 0.563 0.340 −0.711 to 0.265 0.35 0.83

2 DCs/Lin NK-subset 12.9 17.3 −6.53 to 15.4 0.40 0.86

2 Plasmablasts/CD19+CD20 6.90 4.14 −9.78 to 4.24 0.42 0.86 2 viable CD3

2 IgD memory/B cells 9.51 12.8 −5.84 to 12.3 0.45 0.86

IgD+ memory/B cells 10.6 16.3 −10.5 to 21.9 0.46 0.86

Plasmacytoid DCs/DCs 25.9 22.6 −13.8 to 7.27 0.52 0.86

CD4+ T cells/viable CD3+ 56.7 53.3 −15.2 to 8.51 0.56 0.86

2 Monocytes/Lin NK-subset 43.3 39.7 −15.9 to 8.78 0.56 0.86

Th17-like/CD8+ T cellsa 1.45 1.76 −0.899 to 1.52 0.60 0.86

CD45RO+/CD4+ T cells 37.0 33.2 −19.9 to 12.3 0.62 0.86

Activated/CD4+ T cells 0.470 0.383 −0.445 to 0.278 0.63 0.86

Naive/CD4+ T cells 42.5 37.9 −23.8 to 14.8 0.63 0.86

Th1/CD4+ T cells 4.83 6.00 −4.45 to 6.79 0.66 0.86

Naive/CD8+ T cells 25.6 28.7 −12.2 to 18.5 0.67 0.86

Th17/CD4+ T cells 3.49 3.07 −3.26 to 2.43 0.76 0.91

Central memory/CD4+ 28.8 27.2 −14.6 to 11.3 0.79 0.91 T cells

Naive CCR4+/Tregs 1.74 2.01 −1.93 to 2.47 0.80 0.91

Myeloid DCs/DCs 41.9 43.4 −12.1 to 15.2 0.81 0.91

Effector/CD8+ T cells 27.0 25.6 −16.5 to 13.7 0.85 0.91

Effector memory/CD4+ T cells 22.7 24.2 −15.5 to 18.3 0.86 0.91

Continued

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 5 Table 2 Mean immune cell frequencies in NMOSD and HC PBMC samples (continued)

Cell subset HC NMOSD Difference (95% CI) p Value p-adj

Th2/CD4+ T cells 0.661 0.642 −0.499 to 0.460 0.93 0.94

CCR4+/Tregs 20.2 21.0 −17.4 to 19.0 0.93 0.94

Transitional/B cells 7.72 7.92 −5.60 to 6.00 0.94 0.94

Abbreviations: DC = dendritic cell; HC = healthy control; NK cells = natural killer cells; NMOSD = neuromyelitis optica spectrum disorder; p-adj = p-adjusted (q- value); PBMC = peripheral blood mononuclear cell; Tregs = regulatory T cells. Statistical analysis was performed using the Welch t test and Storey adjustment for multiple comparisons, and serum analytes with p ≤ 0.05 are bolded. Cell subsets are in ascending order of p value. a Th1-like, Th2-like, and Th17-like CD8+ T cells are the CD8+ correlates of CD4+ Th1, Th2, and Th17 subtypes, based on marker expression.

were elevated in NMOSD in 1 cohort (p ≤ 0.05). All other serum principal component analysis of immunophenotyping or serum analyte concentrations in NMOSD were similar to HC. analyte data sets could not reliably distinguish NMOSD vs HC.

Multivariable logistic regression identifies Serum CX3CL1 concentration correlates with BAFF association with NMOSD days since relapse We performed multivariable logistic regression analyses to We next sought to investigate whether measured serum analyte model the association of serum analyte concentrations or concentrations correlated with temporal measures of clinical immunophenotyping frequencies with NMOSD. The ability outcome, particularly the number of days since relapse or ARR. to predict NMOSD was assessed using 5-fold cross validation The mean number of days since relapse was 590 (95% CImean evaluated through quantification of the AUC of the associated 350–820), and the mean ARR was 0.8 (95% CImean 0.5–1.2). ROC curve (figure 1, A–D; table 4). Models constructed on Multivariable linear regression models were used to identify as- immunophenotyping and serum analyte concentrations from sociations between cohort 1 and 2 serum analyte levels and the 2 both cohorts had a greater predictive ability (0.7 < AUC < outcome variables. These models fit weakly, though better for days 0.8) than models constructed on serum analyte concentra- since relapse (R2 = 0.602) than ARR (R2 = 0.361). Statistical tions from cohort 1 or 2 alone (0.5 < AUC < 0.7). The model analysis of serum analyte parameters (table e-2, links.lww.com/ based on serum analytes from both cohorts had the highest NXI/A292) revealed that the chemokine CX3CL1 (fractalkine; p AUC (0.734) and standard accuracy value (0.700) and was = 0.031) was associated with days since relapse. This association further analyzed to identify significant parameters. indicated that CX3CL1 levels in serum decline as a function of time since previous relapse. No other serum analyte was significant In the multivariate serum analyte model based on both co- in our multivariable linear regression analyses for either outcome. horts, BAFF concentration was found to be associated with NMOSD (p = 0.022, figure 1E), consistent with observations from univariate analyses. BAFF concentration remained as- Discussion sociated with NMOSD in this model when rituximab treatment was included as a covariate (p = 0.038). A multivariable The goal of this exploratory study was to assess circulating logistic regression model constructed on cohort 1 and tested immune cell and cytokine/chemokine profiles using an un- on cohort 2 performed with an AUC of 0.66 and accuracy of biased and hypothesis-generating strategy. This approach ach- 0.58, suggesting weak generalizability between cohorts (figure ieved 3 overarching goals, which were to (1) verify the utility of 1, F and G). The cohort 1–trained model was a relatively poor the CIRCLES biospecimen repository in supporting molecular predictor (AUC <0.7) of NMOSD in cohort 2. and cellular NMOSD research, (2) reference known biomarkers as a means of validation using external standards, and (3) A separation of NMOSD and HC was observed when visual- identify potentially novel candidate NMOSD biomarkers. With izing components 1 and 2 derived from principal component the high-throughput evaluation of a large set of circulating im- analysis (figure e-1, A–D, links.lww.com/NXI/A291). These mune cell frequencies and serum cytokine and chemokine components explained only 45% and 25%, respectively, of the concentrations, we have validated previously identified bio- total variance from immunophenotyping and serum analytes markers. Beyond this concordance, the current studies have also (both cohorts). We did not observe distinct clustering based on identified CD16+CD56+ NK cells and CX3CL1 as intriguing these analyses. Principal components 1 and 2 explained 73% and novel biomarker candidates to be further evaluated for and 78% of the total variance of cohorts 1 and 2, respectively. possible NMOSD predictive or prognostic applications. Although clustering was still not observed in these decompo- sitions, some NMOSD samples did separate. No differentiation Asignificant decrease in peripheral NK cell frequency has been of samples was seen based on institutional source or cohort reported in patients with NMOSD in comparison to patients (figure e-1, E–J, links.lww.com/NXI/A291). In general, with MS and HCs.21 However, stratificationbyclinicalstatus

Table 3 Mean NMOSD and HC serum analyte concentrations in 2 cohorts

Cohort 1 Cohort 2

Analyte HC, pg/mL NMOSD, pg/mL Difference (95% CI) p Value (p-adj) HC, pg/mL NMOSD, pg/mL Difference (95% CI) p Value (p-adj)

2 2 2 BAFF 527 1,580 671 to 1,430 4.4 × 10 6 (1.6 × 10 4 ) 1,350 3,970 1,360 to 3,880 1.8 × 10 4 (0.0028)

IL-6 —— — BL 1.81 3.04 0.500 to 1.97 0.0018 (0.013)

CCL22 332 314 −85.9 to 48.9 0.58 (0.78) 469 773 104 to 504 0.0038 (0.019)

CCL3 116 87.4 −145 to 87.5 0.61 (0.78) 98.6 128 8.49 to 50.7 0.0071 (0.027)

S100B —— — BL 46.9 90.1 6.12 to 80.4 0.024 (0.072)

FGF-basic —— — BL 1.11 1.57 0.0262 to 0.884 0.038 (0.089)

IL-15 —— — BL 4.00 4.57 0.00840 to 1.14 0.047 (0.089)

IL-1RA 534 571 −186 to 259 0.74 (0.88) 702 1,130 6.65 to 856 0.047 (0.089)

IL-2 —— BL 29.6 47.9 −0.678 to 37.4 0.058 (0.099)

TGF-a —— BL 5.00 5.89 −0.102 to 1.88 0.077 (0.12)

CCL2 1,378 163 −16.6 to 67.9 0.23 (0.51) 530 410 −303 to 64.1 0.19 (0.21)

CX3CL1 553 566 −143 to 168 0.87 (0.89) 932.9 1,147 −136 to 564 0.22 (0.21)

IP-10 14.1 15.5 −4.55 to 7.43 0.63 (0.78) 19.3 25.2 −4.01 to 15.9 0.23 (0.21)

APRIL —— NA 1,830 2,200 −260 to 1,010 0.24 (0.21)

FLT-3L 35.1 36.2 −6.42 to 8.55 0.77 (0.88) 73.2 88.5 −10.8 to 41.5 0.24 (0.21)

IL-1a 1.71 1.75 −0.0297 to 0.125 0.22 (0.51) 4.20 4.28 −0.0543 to 0.210 0.24 (0.21)

IL-13 —— — BL 419 686 −261 to 794 0.31 (0.21)

IL-21 —— — NA 9.90 12.2 −2.26 to 6.76 0.32 (0.21)

CCL4 130 115 −187 to 155 0.85 (0.89) 366 208 −509 to 195 0.35 (0.21)

G-CSF —— BL 14.4 18.3 −4.89 to 12.8 0.37 (0.21)

IL-8 13.1 62.1 −15.8 to 114 0.13 (0.46) 18.4 27.6 −12.8 to 31.2 0.40 (0.23)

IL-19 —— — BL 253 249 −16.1 to 7.24 0.45 (0.24)

CCL7 —— — BL 91.0 89.2 −7.07 to 3.46 0.48 (0.24)

Eotaxin 55.9 71.2 −0.199 to 30.8 0.053 (0.32) 120 105 −59.9 to 29.6 0.49 (0.24)

Continued 7 8 erlg:Nuomuooy&Nuonlmain|Vlm ,Nme etme 00Neurology.org/NN 2020 September | 5 Number 7, Volume | Neuroinflammation & Neuroimmunology Neurology:

Table 3 Mean NMOSD and HC serum analyte concentrations in 2 cohorts (continued)

Cohort 1 Cohort 2

Analyte HC, pg/mL NMOSD, pg/mL Difference (95% CI) p Value (p-adj) HC, pg/mL NMOSD, pg/mL Difference (95% CI) p Value (p-adj)

IL-31 —— — BL 138 165 −51.2 to 106 0.49 (0.24)

IL-16 —— — NA 159 175 −43.7 to 76.0 0.59 (0.26)

IL-7 2.83 3.03 −0.280 to 0.672 0.41 (0.67) 8.61 9.64 −2.82 to 4.88 0.59 (0.26)

IL-27 —— — BL 164 158 −39.8 to 27.7 0.72 (0.30)

Eotaxin-2 535 463 −276 to 131 0.47 (0.73) 1,760 1,570 −1,480 to 1,100 0.76 (0.31)

IL-4 2.63 3.2 −0.525 to 1.72 0.29 (0.58) 23.60 23.73 −0.835 to 1.08 0.80 (0.32)

CD40L 2,370 2,470 −638 to 834 0.79 (0.88) 5,120 5,200 −2,040 to 2,220 0.93 (0.36)

IL-33 —— — BL 16.5 16.6 −3.84 to 4.08 0.95 (0.36)

IL-17F 335 507 38.4 to 306 0.013 (0.24) —— — NA

RANTES 679 1,120 8.68 to 875 0.046 (0.32) —— — NA

IL-3 5.65 6.72 −0.189 to 2.32 0.093 (0.42) —— — BL

PDGF-BB 11.3 46.2 −42.1 to 112 0.36 (0.62) —— — NA

PDGF-AA 94.6 109 −40.6 to 69.5 0.60 (0.78) —— — NA

Abbreviations: BL = concentration below limit for analysis; HC = healthy control; NA = not available; NMOSD = neuromyelitis optica spectrum disorder; p-adj = p-adjusted (q-value). Statistical analysis was performed using the Welch t test and Storey adjustment for multiple comparisons, and serum analytes with p ≤ 0.05 are bolded. Serum analytes are in ascending order of cohort 2 p value and then cohort 1 p value. Ten serum analytes were below limit for analysis for both cohorts (IFNγ, IL-10, TNFα, IL-17A, IL-17E/IL-25, IL-5, IL-36β, GM-CSF, IL-22, and IL-12 p70). Three serum analytes measured below detection limit in cohort 1 and were not analyzed in cohort 2 (LT-α, IL-12 p40, and IL-11), and 3 serum analytes measured below detection limit in cohort 2 and were not analyzed in cohort 1 (IL-9, TNFβ, and IL-23). Figure 1 Multivariable logistic regressions based on immunophenotyping frequencies and serum analyte concentrations

ROC curves for 5-fold cross validation of multivariable logistic regression models were constructed for NMOSD predictability based on (A) immunophenotyping, (B) cohort 1, (C) cohort 2, and (D) cohorts 1 and 2 serum analyte concentrations. Combined serum cohort and immunophenotyping models have better predictive value than models constructed on cohorts 1 or 2 alone. (E) Coefficients for the combined cohort model are depicted with SD. BAFF concentration was significantly associated with NMOSD (p = 0.022, in orange). Cohort 2 was tested on a multivariable regression model for cohort 1, for which the (F) ROC curve and (G) confusion matrix are shown. They revealed an AUC of 0.66 and accuracy of 0.58. AUC = area under the curve; NMOSD = neuromyelitis optica spectrum disorder; ROC = receiver operating characteristic; p ≤ 0.05 was considered statistically significant. revealed that only patients with relapsing NMOSD had sig- specificity. In this study, we did not find a significant change in nificantly less NK cells; those in remission did not.21 Therefore, total NK cell frequency, although we found a decrease in NK cell frequency may be a candidate biomarker with temporal CD16+CD56+ NK cell subset frequency in participants with

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 9 syndrome, and rheumatoid arthritis, and it correlates with dis- Table 4 AUC and ACC for multivariable logistic regression ease status.38,39 In the present study, BAFF was significantly curves elevated in both study cohorts in unstratified NMOSD groups. fi Data AUC, mean ± SD ACC, mean ± SD Furthermore, BAFF was signi cantly associated with NMOSD in a multivariable logistic regression model, even with rituximab Immunophenotyping 0.700 ± 0.217 0.530 ± 0.157 treatment used as a covariate. Therefore, our analyses are con- Serum analytes: cohort 1 0.634 ± 0.239 0.464 ± 0.163 sistent with previous studies identified BAFF as an NMOSD

Serum analytes: cohort 2 0.503 ± 0.206 0.444 ± 0.193 biomarker.

Serum analytes: both 0.734 ± 0.143 0.700 ± 0.142 The current ﬁndings were supported by the observation that

Abbreviations: ACC = standard accuracy value; AUC = area under the curve; treatment history is associated with immunophenotyping ROC = receiver operating characteristic. frequencies and serum analyte concentrations. For example, we observed a significant reduction in B-cell frequency in NMOSD samples in relation to rituximab treatment. This NMOSD. Of note, CD16+ NK cells express perforin and are outcome was anticipated, given that rituximab depletes cytotoxic; they are involved in antibody-dependent cellular CD20+ B cells from peripheral circulation. However, ritux- cytotoxicity, which is a mechanism associated with NMOSD imab also depletes non–B-cell CD20+ cells, such as a small pathogenesis.22,23 Although histopathologic examination of population of CD3+ T lymphocytes.40 Of interest, CD20+ NMOSD lesions has not identified NK cells as a predominant T cells are more likely to be CD8+,40 and we observed that the infiltrating cell type, studies conducted in vitro or with mouse frequency of Th1-like CD8+ T cells was significantly reduced models have demonstrated NMO-IgG–dependent astrocyte in participants with NMOSD with a history of rituximab killing by NK cells, highlighting a possible mechanism for therapy. Thus, this study may provide support for the hy- – NMOSD pathology.24 26 Other studies elucidate a protective pothesis that rituximab treatment affects cell subsets beyond role of NK cells in an autoimmune demyelinating disease CD20+ B cells and therefore may contribute to its mecha- model and a reduction in NK cell frequency in patients with nism(s) of action. This concept deserves additional in- MS.27,28 Our findings provide impetus for further inquiry into vestigation. Steroids and other preventive treatments used for CD16+CD56+ NK cells and their functions in association with their immunosuppressive properties are also likely to affect NMOSD relapses in a larger patient sample. immune cell frequencies and serum analyte concentrations. Here, sample sizes for other treatments were not sufficient for CX3CL1 is also a viable novel biomarker candidate for further analysis by stratification, given that stratification by treatment study in NMOSD. We found that CX3CL1 was significantly requires sizable cohorts. Compounded by the rarity of associated with the number of days since relapse in a multi- NMOSD, treatment heterogeneity and limited treatment- variable linear regression model. CX3CL1, also known as naive patients are challenges to biomarker discovery in fractalkine, was originally called neurotactin due to its high NMOSD. However, treated patients should be evaluated for expression in the CNS, including by astrocytes.29,30 CX3CL1 is prognostic biomarkers, given such biomarkers would have important for chemotaxis of leukocytes to the CNS and is greater value if they predicted a potential relapse regardless of commonly studied in the context of neurologic disease. For treatment. example, CX3CL1 is elevated in the serum of patients with MS.31 Its receptor, CX3CR1, is expressed on microglia, Taken together, the present findings provide impetus for fur- monocytes, dendritic cells, and especially NK cells.32 Notably, ther study of CX3CL1 and CD16+CD56+ NK cells as potential CD16+ NK cells highly express CX3CR1,33 and variations in novel biomarker candidates specifically associated with CX3CR1 are correlated with age-related macular de- NMOSD and its relapse patterns. Considering that the eleva- generation.34 These collective findings reinforce our current tion of serum CCL22, IL-6, and CCL3 was not validated by a data suggesting that CX3CL1 is an intriguing candidate for second cohort, it remains to be determined whether signifi- further study in the context of NMOSD relapse prediction. cantly different concentrations would be found in larger NMOSD cohorts. Therefore, these serum analytes should be BAFF elevation as a result of rituximab treatment is a recognized further validated for consideration as biomarkers, as should phenomenon and has been implicated in relapses occurring Th1-like CD8+ T cells. Future studies involving larger cohorts immediately subsequent to rituximab dosing.18 Ostensibly, the may reveal that these or other immunologic constituents, either depletion of B cells causes a compensatory elevation in BAFF, alone or as components of composite signatures, represent promoting B-cell activation and disease recrudescence. Elevated important indicators of disease status. In any case, the current BAFF levels have been observed in the CSF and serum of studies affirm the use of high-throughput techniques for bio- patients with NMOSD.35,36 Autoreactive B cells are especially marker identification in larger NMOSD cohorts. It would be dependent on BAFF for survival; BAFF selectively rescues informative to reevaluate the predictive value of multivariable autoreactive B cells from elimination.37 Elevated BAFF con- models and principal component analyses using a larger cohort centration has been observed in other autoimmune diseases as and further stratification based on treatment, relapse recency, well, including systemic lupus erythematosus, Sjogren and other clinical factors. Larger sample sizes, assessment of

10 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN biospecimens from clinical trials, and investigation of samples Inc.; is a member of the Genentech-Roche Inc. Scientific drawn immediately before and after relapse events might dis- Advisory Committee; and is Chair medical advisor to the play improved delineation of NMOSD from HC in multivari- Guthy-Jackson Charitable Foundation. K. C. O’Connor has able models. We were not able to assess the correlation received research support from Ra Pharma; is a consultant between immunophenotyping and clinical outcomes due to and equity shareholder of Cabaletta Bio; and is the recipient limited relapse data in these study participants (only 4 partic- of a sponsored research subaward from the University of ipants had paired relapse data). Pennsylvania, the primary financial sponsor of which is Ca- baletta Bio. Go to Neurology.org/NN for full disclosures. Our findings provide validation for the use of high-throughput technologies, particularly multipanel flow cytometry and multi- Publication history plex proteomics assay, for identification of candidate NMOSD Received by Neurology: Neuroimmunology & Neuroinflammation biomarkers. Furthermore, we have identified CX3CL1 and April 12, 2020. Accepted in final form June 19, 2020. CD16+CD56+ NK cells as intriguing and novel prognostic biomarker candidates for further study. Based on these findings, for future evaluation of large NMOSD cohorts with high-throughput Appendix 1 Authors techniques, we propose (1) appropriate and standardized collection and storage of samples, as in the CIRCLES program or Name Location Contribution fi clinical trials; (2) veri cation of sample quality, such as PBMC Soumya S. Yale University, Analyzed data, interpreted data, viability quantification; (3) collection of paired information on Yandamuri, New Haven acquired data, and drafted MSE, PhD manuscript for intellectual content treatment and clinical characteristics, such as relapse data; and (4) controlling for or stratification analyses based on treatment his- Ruoyi Jiang, MS Yale University, Performed multivariate statistical New Haven analyses and modeling and tory, current treatment regimen, and/or clinical characteristics. contributed to manuscript writing

Aditi Sharma, Yale University, Acquired data, analyzed data, and Acknowledgment MBBS New Haven contributed to manuscript writing The authors thank Dr. Lesley Devine, Scientiﬁc Director of the Immune Monitoring Core Facility at Yale School of Elizabeth Yale University, Acquired data and analyzed data Cotzomi, BS New Haven Medicine, for assisting with the FACS and proteomics assay. The authors also appreciate the eﬀorts of the following Chrysoula Yale University, Analyzed data and revised the Zografou, PhD New Haven manuscript individuals: Yanet Babcock, J. Michael Dean, Colleen Farrell, Haojun Feng, Susan Filomena, Toni Ganaway, Myka Barnes- Anthony K. Ma, Yale University, Performed multivariate statistical MS New Haven analyses and modeling Garcia, Samuel Glaisher, Rivka Green, Elizabeth Gonzales, Elaine Hsu, Catherine J. Johnson, Ilana Katz Sand, Ilya Kister, Jessica S. University of Coordinated biorepository Alvey, MS Utah, Salt Lake participant information Marlene Keymolen Ramirez, Allyson Reid, Pavle Repovic, City Gloria Rodriguez, Jennifer L. Sedlak, Nancy L. Sicotte, Angela Lawrence J. University of Managed biorepository participant Stangorone, Ben W. Thrower, Anthony L. Traboulsee, and Cook, PhD Utah, Salt Lake information Gabriella Tosto. The CIRCLES project team including site City

PIs and coordinators, the DCC team, and the GJCF and its Terry J. Smith, University of Designed and conceptualized the advisors are grateful to study participants, caregivers, and MD Michigan, Ann study, interpreted data, and families who participated in this study. The CIRCLES project Arbor revised the manuscript for intellectual content is supported by the Guthy-Jackson Charitable Foundation. Michael R. University of Designed and conceptualized the Yeaman, PhD California, Los study, interpreted data, and Study funding Angeles revised the manuscript for Supported by the Guthy-Jackson Charitable Foundation and intellectual content its CIRCLES research and biorepository program. Kevin C. Yale University, Designed and conceptualized the O’Connor, PhD New Haven study, supervised data acquisition, interpreted data, and revised the Disclosure manuscript for intellectual content S. S. Yandamuri, R. Jiang, A. Sharma, E. Cotzomi, C. Zografou, A. K. Ma, J. S. Alvey, and L. J. Cook report no disclosures. T. J. Smith is a paid consultant for Horizon Therapeutics and Immunovant; has been issued US patents for the diagnosis Appendix 2 Coinvestigators from the Guthy-Jackson and treatment of autoimmune diseases with IGF-I receptor Charitable Foundation CIRCLES Study Group monoclonal antibodies; and is a medical advisor to the Name Location Contribution Guthy-Jackson Charitable Foundation. Michael R. Yeaman is Lilyana Department of Neurology, Keck Reviewed and revised founder of NovaDigm Therapeutics Inc. developing novel Amezcua, School of Medicine, University the manuscript for vaccines and immunotherapeutics and Metacin, Inc. de- MD, MS of Southern California, Los intellectual content veloping novel neuroscience pattern recognition tools; has Angeles received honoraria from Alexion Inc. and Genentech-Roche Continued

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 11 Appendix 2 (continued) Appendix 2 (continued)

Name Location Contribution Name Location Contribution

Erika Department of Neurology, Keck Reviewed and revised Mason Weill Cornell Medicine, New Reviewed and revised Amundson, School of Medicine, University the manuscript for Kruse- York, NY the manuscript for BA of Southern California, Los intellectual content Hoyer, MD, intellectual content Angeles MA

Jeffrey L. Departments of Neurology and Reviewed and revised Michael Department of Neurology, Reviewed and revised Bennett, Ophthalmology, University of the manuscript for Levy, MD, Massachusetts General the manuscript for MD, PhD Colorado School of Medicine, intellectual content PhD Hospital, Harvard Medical intellectual content Aurora School, Boston

Jacinta M. The Guthy-Jackson Charitable Reviewed and revised Libby Levine, Department of Neurology, Reviewed and revised Behne, MA Foundation, Beverly Hills, CA the manuscript for RN, ANP-BC Columbia University Medical the manuscript for intellectual content Center, New York, NY intellectual content

Megan The Guthy-Jackson Charitable Reviewed and revised Melanie Weill Cornell Medicine, New Reviewed and revised Kenneally Foundation, Beverly Hills, CA the manuscript for Marcille, BA York, NY the manuscript for Behne, AS intellectual content intellectual content

Terrence M. Stanford University School of Reviewed and revised Katrina Department of Medicine & Reviewed and revised Blaschke, Medicine. Palo Alto, CA the manuscript for McMullen, Neurology, University of British the manuscript for MD, PhD intellectual content PhD Columbia, Vancouver, Canada intellectual content

Robert L. Department of Medicine & Reviewed and revised Maureen A. Viela Bio, Gaithersburg, MD Reviewed and revised Carruthers, Neurology, University of British the manuscript for Mealy, PhD (formerly Johns Hopkins the manuscript for MD Columbia, Vancouver, Canada intellectual content University School of Medicine, intellectual content Baltimore, MD) Tanuja Department of Neurology, Reviewed and revised Chitnis, MD Brigham and Women’s the manuscript for Stephanie Mellen Center for MS Reviewed and revised Hospital, Harvard Medical intellectual content Moore, MPH Treatment and Research, the manuscript for School, Boston, MA Neurological Institute, intellectual content Cleveland Clinic, OH Jeffrey A. Mellen Center for MS Reviewed and revised Cohen, MD Treatment and Research, the manuscript for Devin S. Department of Neurology, Reviewed and revised Neurological Institute, intellectual content Mullin, BS Brigham and Women’s the manuscript for Cleveland Clinic, OH Hospital, Harvard Medical intellectual content School, Boston, MA Jessica Department of Neurology, Reviewed and revised Coleman, BA Johns Hopkins University the manuscript for Kaho B. Department of Neurology, Reviewed and revised School of Medicine, Baltimore, intellectual content Onomichi, Columbia University Medical the manuscript for MD MS Center, New York, NY intellectual content

Casey Engel, Weill Cornell Medicine, New Reviewed and Nancy M. Weill Cornell Medicine, New Reviewed and revised BA York, NY revised the Nealon, MD York, NY the manuscript for manuscript for intellectual content intellectual content Katherine E. Department of Neurology, Reviewed and revised Nelson, BA Columbia University Medical the manuscript for Lisa Department of Medicine & Reviewed and revised Center, New York, NY intellectual content Eunyoung Neurology, University of the manuscript for Lee, BS, MSc British Columbia, intellectual content Sarah M. Mellen Center for MS Reviewed and revised Vancouver, Canada Planchon, Treatment and Research, the manuscript for PhD Neurological Institute, intellectual content May H. Han, Department of Neurology and Reviewed and revised Cleveland Clinic, OH MD Neurological Sciences, Division the manuscript for of Neuroimmunology and intellectual content Ana Pruitt, Department of Neurology, Keck Reviewed and revised Multiple Sclerosis Center, BS School of Medicine, University the manuscript for Stanford University, CA of Southern California, Los intellectual content Angeles Ruth Departments of Neurology Reviewed and revised Johnson, BS and Ophthalmology, the manuscript for Claire S. Department of Neurology, Reviewed and revised University of Colorado intellectual content Riley, MD Columbia University Medical the manuscript for School of Medicine, Aurora Center, New York, NY intellectual content

Diane Mellen Center for MS Reviewed and revised Zoe Rimler, NYU Langone Health Reviewed and revised Ivancic, Treatment and Research, the manuscript for BS the manuscript for CCRP Neurological Institute, intellectual content intellectual content Cleveland Clinic, OH Andrew W. Department of Neurology, Reviewed and revised Eric C. Department of Neurology, Reviewed and revised Russo, BS Massachusetts General the manuscript for Klawiter, MD Massachusetts General the manuscript for Hospital, Harvard Medical intellectual content Hospital, Harvard Medical intellectual content School, Boston School, Boston Julia J. Department of Medicine & Reviewed and revised Alexandra Weill Cornell Medicine, New Reviewed and revised Schubert, BS, Neurology, University of the manuscript for Kocsik, BS York, NY the manuscript for BCS British Columbia, intellectual content intellectual content Vancouver, Canada

12 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN 18. Pellkofer HL, Krumbholz M, Berthele A, et al. Long-term follow-up of patients with Appendix 2 (continued) neuromyelitis optica after repeated therapy with rituximab. Neurology 2011;76: 1310–1315. 19. Cook LJ, Rose JW, Alvey JS, et al. Collaborative International Research in Clinical and Name Location Contribution Longitudinal Experience Study in NMOSD. Neurol Neuroimmunol Neuroinflamm 2019;6:e583. Judy Sheard, The Guthy-Jackson Charitable Reviewed and revised 20. Finak G, Langweiler M, Jaimes M, et al. Standardizing flow cytometry immunophenotyping MPH, MA Foundation, Beverly Hills, CA the manuscript for analysis from the Human ImmunoPhenotyping Consortium. Sci Rep 2016;6:20686. CCRA intellectual content 21. Ding J, Zhu DS, Hong RH, et al. The differential expression of natural killer cells in NMOSD and MS. J Clin Neurosci 2020;71:9–14. Anna J. Department of Neurology and Reviewed and revised 22. Ratelade J, Asavapanumas N, Ritchie AM, Wemlinger S, Bennett JL, Verkman AS. In- Tomczak, MS Neurological Sciences, Division the manuscript for volvement of antibody-dependent cell-mediated cytotoxicity in inflammatory demyelination of Neuroimmunology and intellectual content in a mouse model of neuromyelitis optica. Acta Neuropathol 2013;126:699–709. Multiple Sclerosis Center, 23. Lanier LL, Le AM, Civin CI, Loken MR, Phillips JH. The relationship of CD16 (Leu- Stanford University, CA 11) and Leu-19 (NKH-1) antigen expression on human peripheral blood NK cells and cytotoxic T lymphocytes. J Immunol 1986;136:4480–4486. 24. Saadoun S, Bridges LR, Verkman AS, Papadopoulos MC. Paucity of natural killer and cytotoxic T cells in human neuromyelitis optica lesions. Neuroreport 2012;23:1044–1047. References 25. Ratelade J, Zhang H, Saadoun S, Bennett JL, Papadopoulos MC, Verkman AS. 1. Wingerchuk DM, Lennon VA, Lucchinetti CF, Pittock SJ, Weinshenker BG. The Neuromyelitis optica IgG and natural killer cells produce NMO lesions in mice spectrum of neuromyelitis optica. Lancet Neurol 2007;6:805–815. without myelin loss. Acta Neuropathol 2012;123:861–872. 2. Hinson SR, Lennon VA, Pittock SJ. Autoimmune AQP4 channelopathies and neu- 26. Vincent T, Saikali P, Cayrol R, et al. Functional consequences of neuromyelitis optica- romyelitis optica spectrum disorders. Handb Clin Neurol 2016;133:377–403. IgG astrocyte interactions on blood-brain barrier permeability and granulocyte re- 3. Lennon VA, Wingerchuk DM, Kryzer TJ, et al. A serum autoantibody marker of neu- cruitment. J Immunol 2008;181:5730–5737. romyelitis optica: distinction from multiple sclerosis. Lancet 2004;364:2106–2112. 27. Rodriguez-Martin E, Picon C, Costa-Frossard L, et al. Natural killer cell subsets in 4. Lennon VA, Kryzer TJ, Pittock SJ, Verkman AS, Hinson SR. IgG marker of optic-spinal cerebrospinal fluid of patients with multiple sclerosis. Clin Exp Immunol 2015;180: multiple sclerosis binds to the aquaporin-4 water channel. J Exp Med 2005;202:473–477. 243–249. 5. Bennett JL, Lam C, Kalluri SR, et al. Intrathecal pathogenic anti-aquaporin-4 anti- 28. Zhang B, Yamamura T, Kondo T, Fujiwara M, Tabira T. Regulation of experimental bodies in early neuromyelitis optica. Ann Neurol 2009;66:617–629. autoimmune encephalomyelitis by natural killer (NK) cells. J Exp Med 1997;186: 6. Asavapanumas N, Ratelade J, Papadopoulos MC, Bennett JL, Levin MH, Verkman 1677–1687. AS. Experimental mouse model of optic neuritis with inflammatory demyelination 29. Hatori K, Nagai A, Heisel R, Ryu JK, Kim SU. Fractalkine and fractalkine receptors in produced by passive transfer of neuromyelitis optica-immunoglobulin G. human neurons and glial cells. J Neurosci Res 2002;69:418–426. J Neuroinflammation 2014;11:16. 30. Pan Y, Lloyd C, Zhou H, et al. Neurotactin, a membrane-anchored chemokine 7. Kinoshita M, Nakatsuji Y, Kimura T, et al. Neuromyelitis optica: passive transfer to upregulated in brain inflammation. Nature 1997;387:611–617. rats by human immunoglobulin. Biochem Biophys Res Commun 2009;386:623–627. 31. Kastenbauer S, Koedel U, Wick M, Kieseier BC, Hartung HP, Pfister HW. CSF and 8. Weinshenker BG, Wingerchuk DM. Neuromyelitis spectrum disorders. Mayo Clin serum levels of soluble fractalkine (CX3CL1) in inflammatory diseases of the nervous Proc 2017;92:663–679. system. J Neuroimmunol 2003;137:210–217. 9. Saadoun S, Waters P, Bell BA, Vincent A, Verkman AS, Papadopoulos MC. Intra- 32. Imai T, Hieshima K, Haskell C, et al. Identification and molecular characterization of cerebral injection of neuromyelitis optica immunoglobulin G and human complement fractalkine receptor CX3CR1, which mediates both leukocyte migration and adhe- produces neuromyelitis optica lesions in mice. Brain 2010;133:349–361. sion. Cell 1997;91:521–530. 10. Hinson SR, Romero MF, Popescu BF, et al. Molecular outcomes of neuromyelitis 33. Campbell JJ, Qin S, Unutmaz D, et al. Unique subpopulations of CD56+ NK and NK- optica (NMO)-IgG binding to aquaporin-4 in astrocytes. Proc Natl Acad Sci USA T peripheral blood lymphocytes identified by chemokine receptor expression rep- 2012;109:1245–1250. ertoire. J Immunol 2001;166:6477–6482. 11. Kitley J, Woodhall M, Waters P, et al. Myelin-oligodendrocyte glycoprotein antibodies in 34. Tuo J, Smith BC, Bojanowski CM, et al. The involvement of sequence variation and adults with a neuromyelitis optica phenotype. Neurology 2012;79:1273–1277. expression of CX3CR1 in the pathogenesis of age-related macular degeneration. 12. Hamid SHM, Whittam D, Mutch K, et al. What proportion of AQP4-IgG-negative FASEB J 2004;18:1297–1299. NMO spectrum disorder patients are MOG-IgG positive? A cross sectional study of 35. Okada K, Matsushita T, Kira J, Tsuji S. B-cell activating factor of the TNF family is 132 patients. J Neurol 2017;264:2088–2094. upregulated in neuromyelitis optica. Neurology 2010;74:177–178. 13. Sato DK, Callegaro D, Lana-Peixoto MA, et al. Distinction between MOG antibody-positive 36. Wang H, Wang K, Zhong X, et al. Cerebrospinal fluid BAFF and APRIL levels in and AQP4 antibody-positive NMO spectrum disorders. Neurology 2014;82:474–481. neuromyelitis optica and multiple sclerosis patients during relapse. J Clin Immunol 14. Wingerchuk DM, Banwell B, Bennett JL, et al. International consensus diagnostic 2012;32:1007–1011. criteria for neuromyelitis optica spectrum disorders. Neurology 2015;85:177–189. 37. Lesley R, Xu Y, Kalled SL, et al. Reduced competitiveness of autoantigen-engaged 15. Takahashi T, Fujihara K, Nakashima I, et al. Anti-aquaporin-4 antibody is involved in B cells due to increased dependence on BAFF. Immunity 2004;20:441–453. the pathogenesis of NMO: a study on antibody titre. Brain 2007;130:1235–1243. 38. Groom J, Kalled SL, Cutler AH, et al. Association of BAFF/BLyS overexpression and 16. Matiello M, Lennon VA, Jacob A, et al. NMO-IgG predicts the outcome of recurrent altered B cell differentiation with Sjogren’s syndrome. J Clin Invest 2002;109:59–68. optic neuritis. Neurology 2008;70:2197–2200. 39. Pers JO, Daridon C, Devauchelle V, et al. BAFF overexpression is associated with auto- 17. Kang H, Chen T, Li H, Xu Q, Cao S, Wei S. Prognostic factors and disease course in antibody production in autoimmune diseases. Ann NY Acad Sci 2005;1050:34–39. aquaporin-4 antibody-positive Chinese patients with acute optic neuritis. J Neurol 40. Schuh E, Berer K, Mulazzani M, et al. Features of human CD3+CD20+ T cells. 2017;264:2130–2140. J Immunol 2016;197:1111–1117.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 13 ARTICLE OPEN ACCESS Regional microglial activation in the substantia nigra is linked with fatigue in MS

Tarun Singhal, MD, Steven Cicero, BS, Hong Pan, PhD, Kelsey Carter, BS, Shipra Dubey, PhD, Renxin Chu, PhD, Correspondence Bonnie Glanz, PhD, Shelley Hurwitz, PhD, Shahamat Tauhid, MD, Mi-Ae Park, PhD, Marie Kijewski, DSc, Dr. Singhal [email protected] Emily Stern, MD, Rohit Bakshi, MD, MA, David Silbersweig, MD, and Howard L. Weiner, MD

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e854. doi:10.1212/NXI.0000000000000854 Abstract Objective The goal of our study is to assess the role of microglial activation in MS-associated fatigue (MSAF) using [F-18]PBR06-PET.

Methods Fatigue severity was measured using the Modiﬁed Fatigue Impact Scale (MFIS) in 12 subjects with MS (7 relapsing-remitting and 5 secondary progressive) and 10 healthy control participants who underwent [F-18]PBR06-PET. The MFIS provides a total fatigue score as well as physical, cognitive, and psychosocial fatigue subscale scores. Standardized Uptake Value (SUV) 60–90 minute frame PET maps were coregistered to 3T MRI. Voxel-by-voxel analysis using Statistical Parametric Mapping and atlas-based regional analyses were performed. SUV ratios (SUVRs) were global brain normalized.

Results Peak voxel-based level of significance for correlation between total fatigue score and PET uptake was localized to the right substantia nigra (T-score 4.67, p = 0.001). Similarly, SUVRs derived from atlas-based segmentation of the substantia nigra showed significant correlation with MFIS (r = 0.76, p = 0.004). On multiple regression, the right substantia nigra was an independent predictor of total MFIS (p = 0.02) and cognitive MFIS subscale values (p = 0.007), after adjustment for age, disability, and depression. Several additional areas of significant correlations with fatigue scores were identified, including the right parahippocampal gyrus, right precuneus, and juxtacortical white matter (all p < 0.05). There was no correlation between fatigue scores and brain atrophy and lesion load in patients with MS.

Conclusion Substantia nigra microglial activation is linked to fatigue in MS. Microglial activation across key brain regions may represent a unifying mechanism for MSAF, and further evaluation of neuroimmunologic basis of MSAF is warranted.

From the Partners MS Center (T.S., S.C., K.C., B.G., R.B., H.L.W.), Ann Romney Center for Neurological Diseases, Department of Neurology, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA; PET Imaging Program in Neurologic Diseases (T.S., S.C., K.C.), Ann Romney Center for Neurological Diseases, Department of Neurology, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA; Functional Neuroimaging Laboratory (H.P., R.B., D.S.), Department of Psychiatry, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA; Division of Nuclear Medicine and Molecular Imaging (S.D., M.-A.P., M.K.), Department of Radiology, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA; Laboratory for Neuroimaging Research (R.C., S.T.), Ann Romney Center for Neurological Diseases, Department of Neurology, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA; Department of Medicine (S.H.), Brigham and Women’s Hospital, Harvard Medical School, Boston, MA; Ceretype Neuromedicine (E.S.)Department of Radiology (R.B.), Brigham and Women’s Hospital, Harvard Medical School, Boston, MA.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary AAL = automated anatomic labeling; BPV = brain parenchymal volume; CFS = chronic fatigue syndrome; CIS = clinically isolated syndrome; DMN = default mode network; EDSS = Expanded Disability Status Scale; ES =effect size; HAB = high- affinity binding; HC = healthy control; HDRS = Hamilton Depression Rating Scale; MAB = medium-affinity binding; RRMS = relapsing-remitting MS; SN = substantia nigra; SPM = Statistical Parametric Mapping; SPMS = secondary progressive MS; TSPO = translocator protein.

Defined as an overwhelming sense of tiredness, lack of energy, participants with MS and 2 HC participants were low-affinity or feeling of exhaustion,1,2 fatigue is reported as the most binders and were excluded from the study (see the Geno- disabling symptom in up to 60% of patients with MS3 and is typing section below). Four additional patients with MS estimated to have a lifetime prevalence of 80%.4 Moreover, subsequently changed their minds about participating in the the prevalence of fatigue increases with disease progression study or were lost to follow-up following consent. Twelve (46% in patients with clinically isolated syndrome [CIS] and patients with MS (5 SP and 7 RR; 8 women and 4 men) and 80% in patients with secondary progressive MS [SPMS]).5 10 HC participants (4 women and 6 men) completed the Importantly, fatigue has been described as the leading cause of study. Table 1 contains a summary of characteristics for all absence from work in MS.6 Fatigue may also have prognostic study participants, and table 2 contains detailed individualized implications in patients with MS. For example, in patients participant characteristics. Our findings regarding the re- with CIS, up to 46% of patients may present with fatigue, lationship of the [F-18]PBR06 PET scans from these patients which has recently been shown to predict conversion to with MS with disability and brain atrophy were previously clinically definite MS.7 Among patients with relapsing- reported.16 remitting MS (RRMS), fatigue scores were higher in patients who converted to a confirmed Expanded Disability In terms of the inclusion criteria, we included patients with Status Scale (EDSS) score ≥3 after at least 3 years of follow-up SPMS if they experienced a worsening of their EDSS score by compared with nonconverters. This association remained at least 0.5 (if their baseline EDSS score was ≥5.5) or at least significant after adjusting for depression scales and baseline 1.0 (if their baseline EDSS score was <5) over a period of 1 EDSS.8 Despite the high prevalence and significance of fatigue year before PET scanning. In addition, we included patients in MS, its anatomic and physiologic substrate and its mech- with RRMS who had a relapse or if there was evidence of a anism are not clear. new or enlarging T2 bright lesion on MRI or a gadolinium- enhancing lesion on T1-weighted MRI in the last year. Pa- Inflammation is proposed as a potential mechanism for fa- tients who were treated with corticosteroids during a period of tigue in MS but has the lack of sufficient evidence.3 Abnor- 1 month before the PET scan were excluded. PET scanning malities in functional connectivity of the resting default mode was performed within a median time interval of 5.1 (range network have been linked with fatigue in MS.9 However, the 1.1–25.6) weeks in relation to the subjects’ clinical assess- biological processes underlying these abnormalities are not ment. Patients did not experience a relapse between the known. clinical and PET imaging visits.

Microglial activation may play a role in the pathogenesis of Standard protocol approvals, registrations, MS,10,11 but it has not been systematically studied in relation and patient consents to MS-associated fatigue (MSAF). [F-18]PBR06 is a second- The clinicaltrials.gov ID for our study is NCT02649985. The generation, longer half-life PET radioligand, targeting the study was approved and overseen by the Institutional Review 18-kDa translocator protein for noninvasive assessment of ce- Board, Radiation Safety Committee, and Radioactive Drug rebral microglial activation that we have recently reported to be Research Committee at our hospital. Written informed consent increased in subcortical gray matter and normal-appearing was obtained from all study participants before participation. – white matter in MS.12 17 Our aim is to assess the role of microglial activation in MSAF using [F-18]PBR06 PET. Genotyping Blood samples were collected during the initial visits, and genotyping was performed for screening purposes. Geno- Methods typing used a TaqMan assay to identify the DNA polymorphism of the translocator protein (TSPO) gene on Participants chromosome 22q13.2. Study participants with high-affinity Participants were recruited from the Partners Multiple Scle- (HAB) and medium-affinity binding (MAB) were included in rosis Center at the Brigham and Women’s Hospital and the study, whereas ones with low-affinity binding were ex- through an online recruitment portal for healthy control cluded. Among the 12 patients with MS who completed the participants (HCs). Initially, 18 participants with MS and 12 study, 7 were HABs and 5 were MABs. Among the HC par- HC participants were consecutively enrolled. Two ticipants, 6 were HABs and 4 were MABs (table 1). The

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Table 1 Summary of participant characteristics

MS (N = 12) HC (N = 10) p Value

Age (y) 42 ± 11.7 46 ± 15.5 0.54

Sex distribution 7 F, 5 M 4 F, 6 M 0.39

TSPO binding affinity 7 high, 5 medium 6 high, 4 medium 0.93

Median EDSS score 3.5

MFIS score 36 ± 21

MFIS physical score 17 ± 10

MFIS cognitive score 17 ± 10

MFIS psychosocial score 3±2

HDRS score 5±6.5

Abbreviations: EDSS = Expanded Disability Status Scale; HC = healthy control; HDRS = Hamilton Depression Rating Scale; MFIS = Modified Fatigue Impact Scale; TSPO = translocator protein. Data are mean ± SD, unless otherwise indicated. proportion of HABs and MABs in the MS and HC groups was acquired in a list mode acquisition mode using a high- not significantly different (table 1). resolution PET/CT scanner (GE Discovery ECAT, Waukesha, WI). Statistical Parametric Mapping (SPM) was used as the Radiopharmaceutical production primary analysis technique for this study. SPM12 software (the We have previously described the radiopharmaceutical pro- Wellcome Institute of Cognitive Neurology, London, United duction methodology at our center.16 Kingdom; fil.ion.ucl.ac.uk/spm/software/) was used to process the [F-18]PBR06 PET SUV images.24 PET images were ster- MRI acquisition and corresponding analysis eotactically normalized to the Montreal Neurologic Institute Each study participant completed an MRI scan on the same version of Talairach space. Whole-brain multiple linear re- scanner (Siemens 3T Skyra, Erlangen, Germany) according to a gression modeling of the within-group effects of SUV images – previously described acquisition protocol.18 20 The protocol was used on a voxel-by-voxel basis to examine their association included a 2D T1-weighted spin-echo axial series (repetition with Modified Fatigue Impact Scale (MFIS) as the main re- time = 611–943ms,echotime=7.9ms,andvoxelsize=0.43× gressor, and age, sex, and global SUV as covariates of no in- 0.43 × 3 mm3)anda3Dfluid-attenuated inversion recovery and terest, in an analysis of covariance (ANCOVA) setting. These magnetization-prepared rapid gradient-echo series (voxel sizes group-level correlation effect estimates generated statistical forboth1×1×1mm3). Using a previously reported technique, parametric maps of the t-statistic that demonstrated the age- normalized whole-brain parenchymal volume (BPV) was de- adjusted and sex-adjusted correlations of MFIS total and sub- rived by applying the latter images to a fully automated algo- scale scores with PET uptake in the MS group. rithm (SIENAX, v. 5.0; Analysis Group, fsl.fmrib.ox.ac.uk.)21 Furthermore, we coregistered the summed [F-18]PBR06- Hyperintense lesions on 3D T2 FLAIR and hypointense le- PET images acquired 60–90 minutes following tracer in- sions on T1-weighted spin-echo images were marked by a jection to the individual MRIs and segmented the images into trained observer (K.C.) and verified by a senior observer gray and white matter regions of interest, and subregions as (S.T. or T.S.). We determined that to qualify as a T1 hypo- defined by the automated anatomical labeling (AAL) atlas intense lesion, the lesion had to be both hypointense on template,25 using PNEURO 3.8 software (PMOD Technol- T1-weighted images and hyperintense on Fluid Attenuation ogies, Zurich, Switzerland; pmod.com/web/). This is an au- Inversion Recovery (FLAIR) images.22,23 A semiautomated tomated pipeline and reduces risk of operator-dependent bias edge-finding tool was then used for volumetric lesion con- in region-of-interest delineation. To account for inter- touring using Jim (version 7; Xinapse Systems, West Bergholt, participant differences, partial volume–corrected SUV ratios United Kingdom; xinapse.com). Our previous work has shown (SUVRs) were calculated for participants based on normali- ’ high reliability for this semiautomated method of measuring zation of the individual region s SUV60-90 to the global brain 22,23 16,26–28 MS cerebral lesion burden. SUV60-90, similar to our previous reports. SUVRs for individual supratentorial AAL template regions of interest PET acquisition and corresponding analysis (ROIs) were also obtained and assessed for correlations with [F-18]PBR06 was injected as a bolus injection for PET scan- fatigue scores. Because the substantia nigra (SN) showed the ning using an IV catheter into an arm or hand vein; images were strongest significant correlations with fatigue scores on SPM

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 Table 2 Detailed participant characteristics

TSPO Disease Participant Age binding duration Current EDSS MFIS MFIS MFIS number Sex (y) affinity Group (y) DMT score MFIS Physical Cognitive Psychosocial HDRS

1 F 37 HAB RRMS 5.2 Fingolimod 4 64 25 35 4 13

2 M 37 HAB RRMS 19.6 Rituximab 3 33 12 19 2 1

3 F 34 MAB RRMS 4.8 Fingolimod 1 16 7 9 0 0

4 M 32 MAB RRMS 11.9 Fingolimod 1 15 13 2 0 1

5 F 23 HAB RRMS 8.2 Natalizumab 1.5 6 2 3 1 0

6 F 41 HAB RRMS 9.6 Fingolimod 1.5 21 9 9 3 17

7 F 27 MAB RRMS 3.5 Fingolimod 2 37 15 20 2 0

8 M 52 MAB SPMS 14.9 Rituximab 6.5 15 8 6 1 0

9 M 53 HAB SPMS 19 Rituximab 6 37 21 15 1 2

10 F 59 MAB SPMS 19 Glatiramer 65423256 7 acetate

11 F 59 HAB SPMS 19 None 6.5 71 34 29 8 17

12 F 50 HAB SPMS 17 Rituximab 4.5 62 31 26 5 2

13 M 25 MAB HC

14 F 45 HAB HC

15 M 60 MAB HC

16 F 25 HAB HC

17 F 34 MAB HC

18 M 33 HAB HC

19 M 70 MAB HC

20 M 65 HAB HC

21 M 48 HAB HC

22 F 54 HAB HC

Abbreviations: DMT = disease-modifying treatment; EDSS = Expanded Disability Status Scale; HAB = high-affinity binding; HDRS = Hamilton Depression Rating Scale; MAB = medium-affinity binding; MFIS = Modified Fatigue Impact Scale, HC = healthy control; RRMS = relapsing-remitting MS; SP = secondary progressive MS; TSPO = translocator protein. Disease duration refers to time from first MS symptom onset.

analysis, the SN region was specifically segmented using the Patient-reported outcome measures Hammers atlas applied to PET images in the standard space Participants with MS completed the following patient-reported using PVIEW tool of PMOD 3.8 platform. SN SUVRs were outcome measures: MFIS29 andHamiltonDepressionRating calculated based on Hammers atlas brain parenchymal ROIs Scale (HDRS).30 The MFIS is a 21-item fatigue scale with 5 delineated in PVIEW without partial volume correction owing scaled responses ranging from “never” to “almost always”.Total to the small size of the SN ROI. PET uptake in right, left and fatigue scores as well as physical, cognitive, and psychosocial average of right and left SN ROIs were further investigated for subscale scores are derived. The HDRS is a 17-item rating scale group differences between participants with MS with fatigue that evaluates the severity of depression symptoms using a and participants with MS without fatigue (defined as MFIS >37 semistructured interview. and MFIS ≤37, respectively) and HCs and for correlations with MFIS and its subscales in participants with MS. In addition, for Statistical analysis illustration purposes, individualized maps of increased sub- The exact Wilcoxon-Mann-Whitney test with accompanying stantia nigra PET uptake in a patient with MS with fatigue and a exact Hodges-Lehmann 95% CIs was used to evaluate group patient with MS without fatigue, represented as voxel-based differences in SUVRs. Associations were evaluated using z-score values >2 compared with a healthy data set of 9 healthy Pearson correlations and partial correlations, with 95% CIs volunteers were generated, using the PNEURO platform. calculated using the Fisher transformation. The study’s nature

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN was exploratory as a relatively large number of regions were 4.18 for the right and left side, respectively), right pre- evaluated; therefore, results should be considered hypothesis cuneus (T-score 4.21), left premotor and supplemental generating. motor area (BA6) (T-score 4.08), and juxtacortical white matter (T-score 4.06, figure 1B). A complete list of regions Data availability with voxels demonstrating peak positive correlations be- After due anonymization, we will make any unpublished study tween PET uptake and MFIS with p <0.01isprovidedin data available for sharing with other qualified investigators, if table 3. requested. Strongest voxel-wise peaks for negative correlations with MFIS were seen in the right orbital inferior frontal gyrus (BA Results 47) (T-score 4.03) and the right temporal pole (BA 38) Relationship of MFIS with MRI and (T-score 3.14). clinical parameters BPV was lower in patients with MS (n = 12) compared with Regional analysis HCs (n = 9) (1,402.8 ± 57.1 mL vs 1,475.1 ± 66.9 mL), but Relationship of substantia nigra PET uptake with MFIS there was no correlation between BPV and total MFIS or its in MS physical, cognitive, or psychosocial subscale scores (r = −0.05, fi − − The nding of strongest correlation between fatigue scores 0.09, 0.02, and 0.004, n = 12). There was no correlation and PET in the SN was endorsed by a regional analysis as well. between MFIS and T1 lesion volume (r = 0.002, n = 12) or T2 fi − On regional analysis, there was a signi cant correlation be- lesion volume (r = 0.11, n = 12). In terms of clinical cor- tween the SUVR in the SN and total MFIS score (r = 0.76, relations, MFIS increased with increasing age and EDSS (both 95% CI 0.33 to 0.93, figure 2A). The correlation coefficients r = 0.58) and a positive correlation with increasing HDRS of right and left SN SUVRs with total MFIS (figure 2B) were (r = 0.52). 0.795 (95% CI 0.41 to 0.94) and 0.427 (95% CI −0.20 to SPM analysis 0.80), respectively.

Voxel-level correlations of PET uptake with MFIS An overall model consisting of right SN SUVR, age, EDSS, Strongest voxel-wise peaks for age- and sex-adjusted cor- and HDRS explained the majority of variability of MFIS in relation with MFIS (ﬁgure 1A) were seen in the right SN participants with MS (r2 = 0.7751, adjusted r2 = 0.6466, n = (T-score 4.67), left SN (T-score 4.25), cerebellar vermis, 12) with right SN SUVR remaining an independent predictor right inferior cerebellar cortex (T-scores 4.23 and 4.15, of MFIS (r = 0.73, 95% CI 0.14 to 0.94, p = 0.02) after respectively), bilateral angular gyri (T-scores 4.22 and adjustment for the other covariates.

Figure 1 SPM analysis

(A) Statistical Parametric Mapping (SPM)-based, voxel-by-voxel maps of correlation between Modified Fatigue Impact Scale (MFIS) and PET uptake. The strongest positive correlation was seen in the right substantia nigra (cross-bars). Additional regions of significant correlations were seen in the left substantia nigra, periaqueductal gray, pons, medulla, precuneus, midcingulate, cerebellar vermis, and insular cortex regions. (B) SPM-based, voxel-by-voxel maps demonstrate widespread positive correlation in juxtacortical white matter between MFIS scores and PET uptake.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 5 Table 3 SPM voxel level correlations with MFIS scores

MNI coordinates Brodmann area description (x, y, and z) Peak-level t value Peak-level z value

Positive correlations

Brainstem

Right substantia nigra 10, −16, −14 4.67 3.05

Left substantia nigra −10, −10, −10 4.25 2.89

Pons −8, −24, −28 3.98 2.79

Medulla −2, −34, −48 3.90 2.75

Midbrain: periaqueductal gray (PAG) 4, −30, −18 3.44 2.55

Frontal lobe

Left premotor cortex and supplementary motor area (BA6) −12, −24, 48 4.08 2.83

Left pars opercularis of the inferior frontal gyrus, Broca area (BA44) −50, 20, 26 3.84 2.73

Right Broca area (BA45) 40, 30, 2 3.13 2.40

Right insula (BA13) 32, 28, 0 3.20 2.43

Right dorsolateral and medial prefrontal cortex (BA9) 8, 46, 36 3.04 2.35

Parietal lobe

Right parietal lobe including the angular gyrus near the TPO junction (BA39) 40, −62, 40 4.22 2.88

Right precuneus (BA7) 42, −42, 50 4.21 2.88

Left parietal lobe including the angular gyrus near the TPO junction (BA39) −50, −54, 42 4.18 2.87

Left dorsal posterior cingulate area (BA31) −6, −56, 34 3.93 2.77

Left precuneus (BA7) −20, −68, 38 3.23 2.45

Right inferior parietal lobe/supramarginal gyrus, parietal operculum (BA40) 58, −32, 26 3.03 2.35

Temporal lobe

Right fusiform (BA37) 46, −54, −2 3.63 2.64

Right PrimAuditory (BA41) 54, −12, 0 3.41 2.54

Left superior temporal gyrus, included in the Wernicke area (BA22) −60, −6, −10 3.21 2.44

Right parahip (BA36) 28, −16, −26 3.18 2.42

Left middle temporal gyrus/auditory cortex (BA21) −64, −18, −12 3.12 2.39

Occipital lobe

Right PrimVisual (BA17) 12, −66, 10 3.42 2.54

Right extra striate cortex; receives input from pulvinar (BA19) 30, −82, −14 3.38 2.52

Right VisualAssoc (BA18) 16, −92, 10 3.17 2.41

Cerebellum

Cerebellar vermis −2, −70, −20 4.23 2.89

Right inferior cerebellar cortex 14, −74, −46 4.15 2.85

Left superior cerebellar cortex −38, −52, −24 3.86 2.74

Right superior cerebellar cortex 18, −62, −18 3.25 2.46

Left cerebellar cortex −46, −60, −32 3.18 2.42

Left inferior cerebellar cortex −24, −42, −54 3.09 2.38

Continued

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Table 3 SPM voxel level correlations with MFIS scores (continued)

MNI coordinates Brodmann area description (x, y, and z) Peak-level t value Peak-level z value

White matter

Subcortical/juxtacortical WM 26, −62, 34 4.06 2.82

Juxtacortical/deep WM in the left occipital lobe −20, −76, 16 3.44 2.55

JWM: right frontal (inferior occipitofrontal fascicle) 24, 24, −10 3.39 2.52

Right inferior occipitofrontal fasciculus 32, −6, −10 3.27 2.47

Right frontal JWM 32, −20, 42 3.22 2.44

Negative correlations

Orbital part of the right inferior frontal gyrus (BA47) 30, 22, −26 4.03 2.81

Right temporal pole (BA38) 20, 6, −46 3.14 2.40

Abbreviations: JWM = juxtacortical white matter; MFIS = Modified Fatigue Impact Scale; SPM = Statistical Parametric Mapping; TPO = temporoparietooccipital. p < 0.01.

There was no significant correlation between MFIS and global cognitive subscale of the MFIS in the right precuneus (r = 0.71, brain microglial activation (r = −0.01, 95% CI −0.58 to 0.57) 95% CI 0.22 to 0.91) and the right parahippocampus (r = 0.67, or between MFIS and total brainstem microglial activation (r 95% CI 0.16 to 0.90). Among these regions, after adjustment = 0.24, 95% CI −0.39 to 0.71). When analyzed separately, for age and EDSS, the right precuneus remained significantly both patients with RRMS and SPMS showed at least a trend correlated with cognitive MFIS (r = 0.78, 95% CI 0.29 to 0.95) for an increase in the MFIS increasing right SN SUVR (r = in patients with MS. 0.70, 95% CI −0.11 to 0.95 for RRMS [n = 7], and r = 0.94, 95% CI 0.33 to 0.99 for SPMS [n = 5], respectively). Physical subscale of the MFIS The SN SUVR also correlated significantly with the physical Additional regional correlations of PET with subscale of the MFIS (r = 0.774, 95% CI 0.36 to 0.93, figure 2D) total MFIS that remained significant after adjustment for age and EDSS (r = Inaddition,onsegmentationofthebrainusingtheAALtem- 0.64, 95% CI 0.03 to 0.91). There was a significant correlation fi plate, signi cant increases in MFIS were seen with increasing between the right SN SUVR and the physical subscale of the SUVRs in the right parahippocampus (r = 0.75, 95% CI 0.31 to MFIS (0.68, 95% CI 0.17 to 0.90) that also remained significant 0.93), right precuneus (r = 0.65, 95% CI 0.13 to 0.89), and left after adjustment for age (r = 0.68, 95% CI 0.36 to 0.93) but not putamen (r = 0.62, 95% CI 0.07 to 0.88) in patients with MS. after adjustment for both age and EDSS (r = 0.59, 95% CI −0.06 Among these regions, after adjustment for age and EDSS, right to 0.89). The correlation between the left SN SUVR and the fi precuneus remained signi cantlycorrelatedwithMFIS(r= MFIS physical subscale showed a trend but did not attain sta- 0.65, 95% CI 0.03 to 0.91) in patients with MS. tistical significance (r = 0.56, 95% CI −0.02 to 0.86). Correlations between MFIS subscales and PET On segmentation of the brain, using the AAL template, sig- in MS on regional analysis nificant increases were seen in PET uptake with increasing Cognitive subscale of the MFIS values of the physical subscale of MFIS in the left calcarine The SN SUVR correlated significantly with the cognitive gyrus (r = 0.59, 95% CI 0.02 to 0.87), right parahippocampus subscale of the MFIS (r = 0.71, 95% CI 0.23 to 0.91, figure 2C) (r = 0.77, 95% CI 0.34 to 0.93), left putamen (r = 0.66, 95% CI that remained significant after adjustment for age and EDSS (r 0.13 to 0.89), right thalamus (r = 0.64, 95% CI 0.10 to 0.89), = 0.65, 95% CI 0.04 to 0.91). Again, the right SN SUVR and vermis 10 (r = 0.66, 95% CI 0.14 to 0.90). A decrease in showed a significant correlation with the cognitive subscale PET uptake with increasing values of the physical subscale of (0.84, 95% CI 0.51 to 0.95) that also remained significant after MFIS was seen in the left superomedial frontal gyrus (r = adjustment for age and EDSS (r = 0.84, 95% CI 0.45 to 0.96). −0.62, 95% CI −0.88 to −0.07). The left SN SUVR did not show a significant correlation with cognitive MFIS (r = 0.30, 95% CI −0.33 to 0.75). Psychosocial subscale of the MFIS The correlation between the average SN SUVR and the On segmentation of the brain using the AAL template, sig- psychosocial MFIS subscale did not attain statistical signifi- nificant increases were seen in PET uptake increases in the cance (r = 0.556, 95% CI −0.03 to 0.86), but the right SN

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 7 Figure 2 Correlational analysis between regional PET SUVR and fatigue scores

Correlations between (A) average substantia nigra (SN) SUVR and total MFIS, (B) right SN SUVR and total MFIS, (C) average SN SUVR and cognitive MFIS, (D) average SN SUVR and physical MFIS, and (E) right SN SUVR and psychosocial MFIS. MFIS = Modified Fatigue Impact Scale.

SUVR showed a significant correlation with the psychosocial Group comparisons of SN PET uptake between subscale of the MFIS (r = 0.67, 95% CI 0.16 to 0.90, figure patients with MS with fatigue and patients 2E). The left SN SUVR did not show a significant correlation with MS without fatigue and HCs with the psychosocial subscale of the MFIS (r = 0.22, 95% CI MFIS scores were 6–37 in patients with MS without fatigue −0.40 to 0.71). and 54 to 71 in patients with MS with fatigue. The SN SUVR was higher in participants with MS with fatigue compared On segmentation of the brain using the AAL template, with participants with MS without fatigue and HC partici- significant increases were seen in PET uptake with in- pants (1.053 ± 0.031 vs 0.976 ± 0.045 vs 0.932 ± 0.069; 95% creasing values of the psychosocial subscale of MFIS in the CI for effect size [ES] 0.03 to 0.14 and 0.03 to 0.21, re- left cuneus (r = 0.58, 95% CI 0.01 to 0.87), right para- spectively, figures 3, A and B). Similarly, the right SN SUVR hippocampus (r = 0.66, 95% CI 0.14 to 0.90), and vermis was also higher in participants with MS with fatigue compared 10 (r = 0.58, 95% CI 0.02 to 0.87). A decrease in PET with participants with MS without fatigue and HC partici- uptake with increasing values of the psychosocial subscale pants (1.066 ± 0.034 vs 0.971 ± 0.059 vs 0.933 ± 0.102; 95% of MFIS was seen in vermis 12 (r = −0.59, 95% CI −0.87 CI for ES 0.02 to 0.16 and 0.03 to 0.26, respectively, figure to −0.04). 3C). The left SN SUVR was higher in participants with MS

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Figure 3 Comparison of substantia nigra PET uptake between subjects with MS with fatigue and subjects with MS without fatigue and healthy participants

(A) Individualized z-score maps showing increased [F-18]PBR06 PET uptake in the bilateral substantia nigra in a patient with MS with fatigue with a high total MFIS score (total MFIS score = 64) compared with a patient with MS without fatigue with a comparable EDSS score (3.5 vs 4) and a low total MFIS score (MFIS score = 33). For the latter patient, the ROIs for the substantia nigra are delineated but do not demonstrate an increased z-score of >2 compared with a healthy control group. (B) Increased average substantia nigra SUVR in patients with MS with fatigue compared with patients with MS without fatigue and healthy participants. (C) Increased right substantia nigra SUVR in patients with MS with fatigue compared with patients with MS without fatigue and healthy participants. (D) Increased left substantia nigra SUVR in patients with MS with fatigue compared with healthy controls. MFIS = Modified Fatigue Impact Scale. *p < 0.05; **p < 0.01. with fatigue compared with HC participants (1.041 ± 0.041 Discussion vs 0.930 ± 0.085, 95% CI for ES 0.03 to 0.24) but not as compared to participants with MS without fatigue (1.041 ± The major findings of our study are that substantia nigra 0.041 vs 0.980 ± 0.073, 95% CI for ES −0.03 to 0.13) microglial activation is linked to fatigue scores in patients with (figure 3D). MS and that patients with MS with fatigue have a higher substantia nigra microglial activation than HCs. Specifically, Group comparisons of PET uptake in AAL right substantia nigra microglial activation correlated with template regions between patients with MS fatigue scores in patients with MS, independent of age, dis- with fatigue and patients with MS without ability, and depression severity, highlighting the potential fatigue and HCs specificity of this observation. On subsequent exploratory Among the AAL template regions that showed correlations analysis, we found correlation of fatigue scores with microglial with MFIS or its subscales in patients with MS, only the right activation in widespread cortical and subcortical gray matter parahippocampus and right thalamus showed increased regions, including the right precuneus, parahippocampal gy- SUVRs in patients with MS with fatigue compared with HCs rus, putamen, thalamus, and juxtacortical white matter. (1.0 ± 0.056 vs 0.92 ± 0.13 and 1.24 ± 0.08 vs 1.12 ± 0.11, respectively), but they were not statistically significant (95% It has been previously proposed that fatigue in MS is linked to CI for ES −0.04 to 0.21 and −0.05 to 0.27). regional changes rather than global brain damage.31 Rocca

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 9 et al.31 found that total lesional load measured on T2- and T1- The association of [F-18]PBR06 PET uptake in the para- weighted MRIs and global brain atrophy did not distinguish hippocampal gyrus, posterior cingulate, and precuneus ce- patients with MS with fatigue from patients with MS without rebral cortical areas, with fatigue scores, supports the role of fatigue. Instead, injury to strategic gray and white matter re- microglial activation in these regions in the manifestation of gions manifesting as atrophy and microstructural changes fatigue in MS. Precuneus is a hub of the default mode net- measured on diffusion tensor imaging significantly contrib- work (DMN).44 Microglial activation may play a role in the uted to fatigue in MS. The lack of correlations between BPV reported abnormalities in functional connectivity of the and total T2- and T1 lesional load with fatigue scores in our DMN, which has been linked with fatigue in MS.9 Para- study is consistent with these observations. Other studies have hippocampal gyrus has been reported to connect the default found, in terms of anatomic substrates, abnormalities in cor- mode network with the memory system in the medial tem- ticocortical connections, corticostriatal networks, deep gray poral lobe.45 Inflammation of the parahippocampus, pre- matter structures, and a cortico-striato-thalamo-cortical loop cuneus, and posterior cingulate may interfere with in relation to fatigue in MS.32 In terms of functional and physiologic activation and deactivation of the DMN and molecular changes, abnormalities in the dopaminergic system retrieval of memory, contributing to fatigue in patients with (dopamine hypothesis),33 the neuroendocrine system in- MS.45 The association of right parahippocampal microglial volving the hypothalamo-pituitary axis (neuroendocrine hy- activity with psychosocial fatigue is consistent with its role in pothesis),34 and altered functional connectivity of the resting affective and cognitive empathy46 and detecting sarcasm in default mode network (functional disconnection hypothesis) interpersonal interactions.47 More studies are needed to 35,36 have been linked to fatigue in MS in various studies, but assess the relationship between gray matter inflammation the underlying biological bases of these abnormalities are not and fatigue in MS and their modification by various treat- known.32,37 Microglial activation may represent a unifying ment approaches. mechanism underlying these myriad abnormalities detected in patients with MS with fatigue. It is also possible, however, Our study has several limitations. The sample size is small, and that some of these changes are linked to fatigue in general our results need to be confirmed in larger studies. The cross- rather than being specific for MS-related fatigue. Notably, sectional and observational design of our study allows us to inflammation in widespread cortical areas has been associated explore an association between microglial activation and fa- with severity of symptoms in patients with chronic fatigue tigue, but does not allow us to establish causality. We did not syndrome (CFS)/myalgic encephalomyelitis.38 formally assess cognition and sleep disturbances, although none of our participants with MS were diagnosed with a sleep Abnormal functional activation of the SN in association with disorder. Our conclusions rest on the validity of the MFIS, fatigue39 has been previously reported in patients with MS. This which is a subjective scale but has been otherwise validated in is consistent with our results, and the microglial activation in the the assessment of fatigue in MS. Other brainstem nuclei such SN may underly the reported abnormal functional activation in as locus coeruleus may be involved in fatigue pathogenesis in this population. The SN is a major seat of dopaminergic neurons MS, but our approach may lack the resolution and statistical in the brain, and microglial activation in the SN may be linked to power to detect changes in this small-sized structure.48 TSPO dopaminergic imbalance, which has also been proposed as a is not completely specific for micrgolia, and approximately potential mechanism for MS-related fatigue.33 Increased iron 20% of the PET signal may originate from astrocytes accumulation in the SN40 has previously been reported in MS, expressing glial fibrillary acid protein.49 but its association with fatigue in MS has not been studied. Future research studies are needed to compare the fatigue- Moreover, a brainstem fatigue generator model has been related microglial changes in MS with other diseases such as proposed in postviral fatigue syndromes and fatigue that CFS and Parkinson disease and identify common and disease- follows poliomyelitis41 that may be relevant for un- specific mechanisms of fatigue in neurologic disorders. derstanding fatigue in MS. In this model, a central role of Studies with a longitudinal design can help evaluate a causal substantia nigra lesions in inhibiting the functional activity relationship between microglial activation and fatigue in MS. of thalamus, cerebral cortex, and reticular formation via Link of microglial activation with neurochemical and regional decreased dopaminergic stimulation of the putamen and neurodegenerative changes in the context of MS-related fa- resultant excitation of the globus pallidus externa and in- tigue also needs to be further investigated. Furthermore, the hibition of the globus pallidus interna has been prognostic value of fatigue and related microglial activation in proposed.41,42 Recently, a fatigue nucleus that is triggered by the context of progressive MS is also a potentially exciting area cytokines produced by neuroinflammation has been postu- for further research. lated to be responsible for the behavioral changes seen in CFS.43 Further studies are albeit needed to directly assess Widespread microglial activation, particularly in the SN, is the associations of substantia nigra microglial activation with linked to fatigue in MS. Such findings, if replicated and ex- regional and widespread metabolic, electrophysiologic, and panded, could provide a foundation for novel therapeutic neurochemical alterations and to assess whether the SN strategies and targets for fatigue in MS. Further evaluation of could be the fatigue nucleus in MS. neuroimmunologic basis of fatigue in MS is warranted.

10 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Acknowledgment The authors gratefully acknowledge the participation of our Appendix (continued) ﬀ patients and contributions of the sta of Biomedical Imaging Name Location Contribution Research Core and PET technologists in the Division of Nuclear Medicine and Molecular Imaging, Department of Renxin Chu, Brigham and Women’s Analyzed and interpreted PhD Hospital, Harvard Medical the data Radiology at Brigham and Women’s Hospital. School, Boston, MA

Bonnie Brigham and Women’s Analyzed and interpreted Study funding Glanz, PhD Hospital, Harvard Medical the data and revised the The authors gratefully acknowledge research grants from Nancy School, Boston, MA manuscript for intellectual content Davis Foundation’s “Race to Erase MS” program, Ann Romney Center for Neurologic Diseases, Harvard Neuro-Discovery Shelley Brigham and Women’s Analyzed and interpreted Hurwitz, Hospital, Harvard Medical the data and revised the Center, and Water Cove Charitable Foundation for their sup- PhD School, Boston, MA manuscript for intellectual port of this work. Funding agencies did not have any role in the content performance of the research or preparation of the manuscript. Shahamat Brigham and Women’s Analyzed and interpreted Tauhid, MD Hospital, Harvard Medical the data Disclosure School, Boston, MA T. Singhal, S. Cicero, H. Pan, K. Carter, S. Dubey, R. Chu, Mi-Ae Park, Brigham and Women’s Analyzed and interpreted B. Glanz, S. Hurwitz, S. Tauhid, M. Park, M. Kijewski, and PhD Hospital, Harvard Medical the data and revised the School, Boston, MA manuscript for intellectual E. Stern report no disclosures relevant to the manuscript. content Dr. Bakshi has received consulting fees from Bayer, Biogen, fi Marie Brigham and Women’s Designed and Celgene, EMD Serono, Genentech, Guerbet, Sano - Kijewski, Hospital, Harvard Medical conceptualized the study; Genzyme, and Shire and research support from EMD DSc School, Boston, MA analyzed and interpreted fi the data; and revised the Serono and Sano -Genzyme. D. Silbersweig reports no dis- manuscript for intellectual closures relevant to the manuscript. H.L. Weiner has received content consulting fees from Biogen, Tiziana, Novartis, Merck Emily Stern, Brigham and Women’s Designed and Serono, and Teva Neurosciences and has received grant MD Hospital, Harvard Medical conceptualized the study; fi School, Boston, MA analyzed and interpreted support from Merck Serono and Sano -Genzyme and Verily the data; and revised the Life Sciences. Go to Neurology.org/NN for full disclosures. manuscript for intellectual content

Publication history Rohit Brigham and Women’s Designed and Received by Neurology: Neuroimmunology & Neuroinﬂammation Bakshi, MD, Hospital, Harvard Medical conceptualized the study; ﬁ MA School, Boston, MA revised the manuscript for March 24, 2020. Accepted in nal form June 18, 2020. intellectual content; and provided funding

David Brigham and Women’s Analyzed and interpreted Silbersweig, Hospital, Harvard Medical the data and revised the Appendix Authors MD School, Boston, MA manuscript for intellectual content Name Location Contribution Howard L. Brigham and Women’s Designed and Weiner, MD Hospital, Harvard Medical conceptualized the study; Tarun Brigham and Women’s Designed and School, Boston, MA revised the manuscript for Singhal, MD Hospital, Harvard Medical conceptualized the study; intellectual content; and School, Boston, MA major role in the acquisition provided funding of data; analyzed and interpreted the data; drafted and revised the manuscript for intellectual content; and provided References funding 1. Calabrese M, Rinaldi F, Grossi P, et al. Basal ganglia and frontal/parietal cortical atrophy is associated with fatigue in relapsing-remitting multiple sclerosis. Mult Scler Steven Brigham and Women’s Major role in the acquisition 2010;16:1220–1228. Cicero, BS Hospital, Harvard Medical of data and analyzed and 2. Comi G, Leocani L, Rossi P, Colombo B. Physiopathology and treatment of fatigue in School, Boston, MA interpreted the data multiple sclerosis. J Neurol 2001;248:174–179. 3. Newland P, Starkweather A, Sorenson M. Central fatigue in multiple sclerosis: a Hong Pan, Brigham and Women’s Analyzed and interpreted review of the literature. J Spinal Cord Med 2016;39:386–399. PhD Hospital, Harvard Medical the data and revised the 4. Krupp LB, Seraﬁn DJ, Christodoulou C. Multiple sclerosis-associated fatigue. Expert School, Boston, MA manuscript for intellectual Rev Neurother 2010;10:1437–1447. content 5. Feinstein A, Freeman J, Lo AC. Treatment of progressive multiple sclerosis: what works, what does not, and what is needed. Lancet Neurol 2015;14:194–207. Kelsey Brigham and Women’s Major role in the acquisition 6. Fischer A, Heesen C, Gold SM. Biological outcome measurements for behavioral Carter, BS Hospital, Harvard Medical of data and analyzed and interventions in multiple sclerosis. Ther Adv Neurol Disord 2011;4:217–229. School, Boston, MA interpreted the data 7. van der Vuurst de Vries RM, van den Dorpel JJ, Mescheriakova JY, et al. Fatigue after a ﬁrst attack of suspected multiple sclerosis. Mult Scler 2018;24:974–981. Shipra Brigham and Women’s Designed and 8. Cavallari M, Palotai M, Glanz BI, et al. Fatigue predicts disease worsening in relapsing- Dubey, PhD Hospital, Harvard Medical conceptualized the study remitting multiple sclerosis patients. Mult Scler 2016;22:1841–1849. School, Boston, MA and major role in the 9. Bisecco A, Nardo FD, Docimo R, et al. Fatigue in multiple sclerosis: the contribution acquisition of data of resting-state functional connectivity reorganization. Mult Scler 2018;24: 1696–1705.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 11 10. Gandhi R, Laroni A, Weiner HL. Role of the innate immune system in the patho- 29. Kos D, Kerckhofs E, Carrea I, Verza R, Ramos M, Jansa J. Evaluation of the modified genesis of multiple sclerosis. J Neuroimmunol 2010;221:7–14. fatigue impact scale in four different European countries. Mult Scler 2005;11:76–80. 11. Weiner HL. The challenge of multiple sclerosis: how do we cure a chronic hetero- 30. Williams JB. A structured interview guide for the Hamilton Depression Rating Scale. geneous disease? Ann Neurol 2009;65:239–248. Arch Gen Psychiatry 1988;45:742–747. 12. Fujimura Y, Kimura Y, Simeon FG, et al. Biodistribution and radiation dosimetry in 31. Rocca MA, Parisi L, Pagani E, et al. Regional but not global brain damage contributes humans of a new PET ligand, (18)F-PBR06, to image translocator protein (18 kDa). to fatigue in multiple sclerosis. Radiology 2014;273:511–520. J Nucl Med 2010;51:145–149. 32. Chalah MA, Riachi N, Ahdab R, Creange A, Lefaucheur JP, Ayache SS. Fatigue in 13. Fujimura Y, Zoghbi SS, Simeon FG, et al. Quantification of translocator protein (18 multiple sclerosis: neural correlates and the role of non-invasive brain stimulation. kDa) in the human brain with PET and a novel radioligand, (18)F-PBR06. J Nucl Med Front Cell Neurosci 2015;9:460. 2009;50:1047–1053. 33. Dobryakova E, Genova HM, DeLuca J, Wylie GR. The dopamine imbalance hy- 14. James ML, Belichenko NP, Nguyen TV, et al. PET imaging of translocator protein (18 pothesis of fatigue in multiple sclerosis and other neurological disorders. Front Neurol kDa) in a mouse model of Alzheimer’s disease using N-(2,5-dimethoxybenzyl)-2-18F- 2015;6:52. fluoro-N-(2-phenoxyphenyl)acetamide. J Nucl Med 2015;56:311–316. 34. Burfeind KG, Yadav V, Marks DL. Hypothalamic dysfunction and multiple sclerosis: 15. Lartey FM, Ahn GO, Shen B, et al. PET imaging of stroke-induced neuroinflammation implications for fatigue and weight dysregulation. Curr Neurol Neurosci Rep 2016;16: in mice using [18F]PBR06. Molecular imaging and biology. Mol Imaging Biol 2014; 98. 16:109–117. 35. Finke C, Schlichting J, Papazoglou S, et al. Altered basal ganglia functional connec- 16. Singhal T, O’Connor K, Dubey S, et al. Gray matter microglial activation in relapsing tivity in multiple sclerosis patients with fatigue. Mult Scler 2015;21:925–934. vs progressive MS: a [F-18]PBR06-PET study. Neurol Neuroimmunol Neuro- 36. Hidalgo de la Cruz M, d’Ambrosio A, Valsasina P, et al. Abnormal functional con- inflamm 2019;6:e587. doi:10.1212/NXI.0000000000000587. nectivity of thalamic sub-regions contributes to fatigue in multiple sclerosis. Mult Scler 17. Singhal T, O’Connor K, Dubey S, et al. 18F-PBR06 versus 11C-PBR28 PET for 2018;24:1183–1195. assessing white matter translocator protein binding in multiple sclerosis. Clin Nucl 37. Chalah MA, Ayache SS. Is there a link between inflammation and fatigue in multiple Med 2018;43:e289–e295. sclerosis? J Inflamm Res 2018;11:253–264. 18. Dupuy SL, Tauhid S, Hurwitz S, Chu R, Yousuf F, Bakshi R. The effect of dimethyl fumarate 38. Nakatomi Y, Mizuno K, Ishii A, et al. Neuroinflammation in patients with chronic on cerebral gray matter atrophy in multiple sclerosis. Neurol Ther 2016;5:215–229. fatigue syndrome/myalgic encephalomyelitis: an (1)(1)C-(R)-PK11195 PET study. 19. Kim G, Chu R, Yousuf F, et al. Sample size requirements for one-year treatment effects J Nucl Med 2014;55:945–950. using deep gray matter volume from 3T MRI in progressive forms of multiple scle- 39. Engstrom M, Flensner G, Landtblom AM, Ek AC, Karlsson T. Thalamo-striato- rosis. Int J Neurosci 2017;127:971–980. cortical determinants to fatigue in multiple sclerosis. Brain Behav 2013;3:715–728. 20. Meier DS, Guttmann CRG, Tummala S, et al. Dual-sensitivity multiple sclerosis lesion 40. Blazejewska AI, Al-Radaideh AM, Wharton S, et al. Increase in the iron content of the and CSF segmentation for multichannel 3T brain MRI. J Neuroimaging 2018;28:36–47. substantia nigra and red nucleus in multiple sclerosis and clinically isolated syndrome: 21. Chu R, Tauhid S, Glanz BI, et al. Whole brain volume measured from 1.5T versus 3T a 7 Tesla MRI study. J Magn Reson Imaging 2015;41:1065–1070. MRI in healthy subjects and patients with multiple sclerosis. J Neuroimaging 2016;26: 41. Chaudhuri A, Behan PO. Fatigue in neurological disorders. Lancet 2004;363: 62–67. 978–988. 22. Bermel RA, Sharma J, Tjoa CW, Puli SR, Bakshi R. A semiautomated measure of 42. Bruno RL, Creange SJ, Frick NM. Parallels between post-polio fatigue and chronic whole-brain atrophy in multiple sclerosis. J Neurol Sci 2003;208:57–65. fatigue syndrome: a common pathophysiology? Am J Med 1998;105:66S–73S. 23. Ceccarelli A, Jackson JS, Tauhid S, et al. The impact of lesion in-painting and regis- 43. Komaroff AL. Advances in understanding the pathophysiology of chronic fatigue tration methods on voxel-based morphometry in detecting regional cerebral gray syndrome. JAMA . Epub 2019 Jul 5. matter atrophy in multiple sclerosis. AJNR Am J Neuroradiol 2012;33:1579–1585. 44. Raichle ME. The brain’s default mode network. Annu Rev Neurosci 2015;38: 24. Friston KJ, Penny WD, Nichols TE, Kiebel SJ, Ashburner JT. Statistical Parametric 433–447. Mapping. The Analysis of Functional Brain Images. London, UK: Elsevier Ltd.; 2007. 45. Ward AM, Schultz AP, Huijbers W, Van Dijk KR, Hedden T, Sperling RA. The 25. Tzourio-Mazoyer N, Landeau B, Papathanassiou D, et al. Automated anatomical parahippocampal gyrus links the default-mode cortical network with the medial labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI temporal lobe memory system. Hum Brain Mapp 2014;35:1061–1073. MRI single-subject brain. NeuroImage 2002;15:273–289. 46. Toller G, Adhimoolam B, Rankin KP, Huppertz HJ, Kurthen M, Jokeit H. Right 26. Singhal T, O’Connor K, Pan H, Dubey S, et al. The relationship of microglial acti- fronto-limbic atrophy is associated with reduced empathy in refractory unilateral vation and multiple sclerosis-associated fatigue: a [F-18]PBR06 PET study. Berlin, mesial temporal lobe epilepsy. Neuropsychologia 2015;78:80–87. Germany: ECTRIMS; 2018. 47. Rankin KP, Salazar A, Gorno-Tempini ML, et al. Detecting sarcasm from para- 27. Singhal T, O’Connor K, Chu R, Tauhid S, et al. [F-18]PBR06 vs. [C-11]PBR28 PET linguistic cues: anatomic and cognitive correlates in neurodegenerative disease. for Assessing White Matter TSPO Binding in Multiple Sclerosis. Paris, France: NeuroImage 2009;47:2005–2015. ECTRIMS/ACTRIMS; 2017. 48. Niepel G, Bibani RH, Vilisaar J, et al. Association of a deficit of arousal with fatigue in 28. Singhal T, O’Connor K, Chu R, Tauhid S, et al. [F-18]PBR06 PET to Assess TSPO multiple sclerosis: effect of modafinil. Neuropharmacology 2013;64:380–388. Binding and its Association with Brain Atrophy and Disability in Multiple Sclerosis. 49. Kaunzner UW, Kang Y, Zhang S, et al. Quantitative susceptibility mapping identifies Paris, France: ECTRIMS/ACTRIMS; 2017. inflammation in a subset of chronic multiple sclerosis lesions. Brain 2019;142:133–145.

12 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN ARTICLE OPEN ACCESS CLASS OF EVIDENCE Long-term prognostic value of longitudinal measurements of blood neuroﬁlament levels

Dieter A. H¨aring, PhD, Harald Kropshofer, PhD, Ludwig Kappos, MD, Jeffrey A. Cohen, MD, Anuja Shah, PhD, Correspondence Rolf Meinert, PhD, David Leppert, MD, Davorka Tomic, DVM, PhD, and Jens Kuhle, MD, PhD Dr. Kuhle [email protected] Neurol Neuroimmunol Neuroinﬂamm 2020;7:e856. doi:10.1212/NXI.0000000000000856

Abstract MORE ONLINE Objective Class of Evidence To assess the long-term prognostic value of an integral of longitudinal measurements of plasma neu- Criteria for rating ﬁ ﬁ therapeutic and diagnostic ro lament light chain levels (NfLlong) over 12 and 24 months vs single neuro lament light chain (NfL) measurements in patients with relapsing-remitting MS (RRMS) and its additional value when combined studies with clinical and MRI measures. NPub.org/coe

Methods This analysis included continuously fingolimod-treated patients with RRMS from the 24-month FTY720 Research Evaluating Effects of Daily Oral therapy in Multiple Sclerosis (FREEDOMS)/12- month Trial Assessing Injectable Interferon vs FTY720 Oral in Relapsing–Remitting Multiple Sclerosis (TRANSFORMS) phase 3 trials and their long-term extension, LONGTERMS. Patients were classified into high (≥30 pg/mL, n = 110) and low (<30 pg/mL, n = 164) NfL categories based on the baseline (BL) NfL value or the geometric mean NfLlong calculated over 12 and 24 months to predict disability-related outcomes and brain volume loss (BVL). The additional prognostic value of NfL was quantified using the area under the receiver operating characteristic (ROC) curve.

Results A single high (vs low) NfL measure at BL was prognostic of a higher risk of reaching Expanded Disability Status Scale (EDSS) score ≥4 earlier (hazard ratio [HR] = 2.19; 95% CI = 1.21–3.97) and ﬀ − − − higher BVL over 120 months (di erence: 1.12%; 95% CI = 2.07 to 0.17). When NfLlong was measured over 24 months, high NfL was associated with a higher risk of reaching EDSS score ≥4(HR= 7.91; 95% CI = 2.99–20.92), accelerated 6-month conﬁrmed disability worsening (HR = 3.14; 95% CI =1.38–7.11), and 20% worsening in the Timed 25-Foot Walk Test (HR = 3.05; 95% CI = 1.38–6.70). Area under the ROC curve was consistently highest in models combining NfL with clinical and MRI measures.

Conclusions NfLlong had a higher prognostic value than single NfL assessments on long-term outcomes in RRMS. Combining it with clinical and MRI measures increased sensitivity and speciﬁcity to predict long-term disease outcomes.

Classification of evidence This study provides Class I evidence that NfLlong was more strongly associated with long-term outcomes than single NfL assessments in patients with RRMS.

From the Novartis Pharma AG (D.A.H., H.K., D.T.), Basel; Research Center for Clinical Neuroimmunology and Neuroscience Basel (L.K., D.L., J.K.), Departments of Medicine, Clinical Research, Biomedicine and Biomedical Engineering, University Hospital and University of Basel, Switzerland; Department of Neurology (J.A.C.), Mellen MS Center, Neurological Institute, Cleveland Clinic, OH; Novartis Healthcare Pvt. Ltd. (A.S.), Hyderabad, India; and DATAMAP GmbH (R.M.), Freiburg, Germany.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

The Article Processing Charge was funded by Novartis Pharma AG, Basel, Switzerland. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary ARBA = annualized rate of brain atrophy; AUC = area under the curve; BVL = brain volume loss; BL = baseline; CM = clinical model; EDSS = Expanded Disability Status Scale; Gd+ = gadolinium enhancing; HR = hazard ratio; NfL = neuroﬁlament light chain; PBVC = percentage brain volume change; ROC = receiver operating characteristic; RRMS = relapsing-remitting MS; PASAT = Paced Auditory Serial Addition Test; T25FWT = Timed 25-Foot Walk Test; 9HPT = 9-Hole Peg Test; 6m-CDW = 6-month conﬁrmed disability worsening.

MS is a chronic autoimmune disorder, characterized by CNS the same treatment and dose in the respective extension studies inflammation and neurodegeneration, leading to accumulation (NCT00662649 and NCT00340834),16,17 and thereafter of disability.1 The clinical disease course of MS is heteroge- transitioned into the open-label long-term extension LONG- neous and remains a challenge for prognosis and therapeutic TERMS study for up to 10 years (NCT01201356).18 Details of decision making in individual patients based on clinical and the individual study design and patient population are reported – MRI measures.2,3 elsewhere.14 18 NfL analysis was performed in all samples where informed consent was granted, irrespective of clinical outcomes. Neurofilament light chain (NfL) is a cytoskeletal protein exclusively expressed by neurons4,5; its release into the CSF and NfL assessments blood4,6 is a highly specific sign of neuronal injury. The strong NfL was analyzed using a single molecule array (SIMOA) im- correlation of CSF and serum/plasma levels of NfL has allowed munoassay (Quanterix Corporation, Billerica, MA) in all patients its establishment as a blood biomarker for monitoring disease who gave consent for an exploratory analysis of their stored – 8,10 activity and treatment response.6 13 Furthermore, single mea- ethylenediaminetetraacetic acid–treated plasma samples. surements of elevated NfL concentrations at baseline (BL) are Plasma samples of NfL were collected during the core study associated with on-study relapses, MRI lesions, brain volume period at BL and at months 6, 12, and 18, and 24 (FREEDOMS – loss (BVL), spinal cord atrophy, and disability worsening.7 9 only) and analyzed later at the University Hospital, Basel, Swit- However, the prognostic value of NfL related to long-term zerland. Laboratory personnel were blinded to treatment alloca- disability outcomes, and particularly the added value when tion with no access to clinical data. The biostatistical analyses combined with clinical and MRI markers, has so far not been were performed at DATAMAP GmbH, Freiburg, Germany. explored in the long-term follow-up of phase 3 studies. Outcome measures We hypothesized that an integral of longitudinal measure- The prognostic value of NfL was tested separately for per- ments of NfL (NfL ) over 12 or 24 months would have centage brain volume change (PBVC), time to Expanded long ≥ fi superior prognostic value for long-term outcomes over single Disability Status Scale (EDSS) score 4.0, time to rst fi (i.e., BL) NfL measures in relapsing-remitting MS (RRMS). 6-month con rmed disability worsening (6m-CDW), time to fi The present analysis of data from 2 phase 3 clinical studies and 6-month con rmed 20% worsening in the Timed 25-Foot fi their extensions aimed to quantify the long-term prognostic Walk Test (T25FWT), time to 6-month con rmed 20% value of an integral of NfL over 12 or 24 months in patients worsening in the 9-Hole Peg Test (9HPT), and time to long fi with RRMS under fingolimod (Gilenya; Novartis Pharma AG, 6-month con rmed 20% worsening in the Paced Auditory Basel, Switzerland) therapy for disability worsening over a 10- Serial Addition Test (PASAT). year follow-up. Furthermore, we assessed whether NfL provides additional value when combined with conventional The prognostic value of NfL, clinical measures with/without clinical and MRI measures to improve long-term prognosis of MRI, or NfL in combination with clinical and MRI measures disability outcomes and BVL in patients with MS. was measured for all the long-term outcomes up to month 84 based on the following combinations of different predictor sets: clinical model (CM), CM plus MRI predictor set (CM + Methods MRI), CM plus NfL predictor set (CM + NfL), and CM plus MRI predictor set and NfL predictor set (CM + MRI + NfL). Study design and patient population The present post hoc biomarker analysis included pooled data Standardized MRI scans were obtained at the screening visit from patients with RRMS who were randomly assigned to re- and at months 6, 12, and 24 (FREEDOMS only) during the ceive fingolimod 0.5 mg once daily during the core period of the core phase and yearly in the extension phase. Brain volume 24-month FTY720 Research Evaluating Effects of Daily Oral change was measured using structural image evaluation using therapy in Multiple Sclerosis (FREEDOMS) (NCT00289978) normalization of the atrophy (SIENA; v3.3 [TRANSFORMS], 14 or 12-month Trial Assessing Injectable Interferon vs FTY720 and v4.2 [FREEDOMS]) software (FMRIB [Oxford Centre Oral in Relapsing -Remitting Multiple Sclerosis (TRANS- for Functional Magnetic Resonance Imaging of the Brain], FORMS) (NCT00340834)15 phase 3 trials (both trials had Oxford UK) using the provider’s default settings (in all cases, essentially the same inclusion/exclusion criteria), continued on the MS MRI lesions were not masked in the process). The

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN annualized rate of brain atrophy (ARBA) was calculated from The prognostic value of NfL for PBVC was analyzed using a the PBVC, as = ([PBVC/100 + 1]^[365.25/#days)] −1) × 100 linearmixedmodelforrepeatedmeasureswithadjustmentsfor where “days” stands for the scan date relative to day 1 for the sex, age, duration of MS, number of relapses in the 2 years primary analysis and relative to the date of the month 6 scan for before the study, BL NBV, gadolinium-enhancing (Gd+) T1 the sensitivity analysis. lesion number at the beginning of the analysis period, T2 lesion volume at the beginning of the analysis period, and geometric EDSS scores were determined every 3 months. T25FWT, mean NfL by category (high vs low) according to the analysis 9HPT, and PASAT scores were measured every 6 months in period (BL, month 12, and month 24). For analysis periods the core phase and yearly in the extension phase.14,15,17,19 starting at month 12 or 24, the model also included change from BLtomonth12or24inT2lesionvolume.Furthermore,the Level of evidence repeated measures model included interaction terms between This post hoc analysis provides level I evidence for long-term visit and NfL category and between visit and NBV at BL to prognostic value of an integral of NfLlong over 24 months to account for the possibility that the relevance of BL assessments determine disability worsening in patients with RRMS over 10 might decrease for PBVC observations measured post-BL. years, using data from phase 3 fingolimod studies and their extensions. To investigate whether NfL has additional prognostic value over clinical and MRI measures, all outcomes measured up to month Statistical analysis 84 were dichotomized (long-term disability event: yes/no; BVL The present analysis classified patients into high (≥30 pg/mL) 10 >0.4%/y: yes/no) and analyzed using logistic regression mod- or low (<30 pg/mL) NfL level categories, based on 3 classi- els. The CM contained the following covariates: sex, age, disease fi cations of NfL as follows: (1) a single measurement at BL duration, number of relapses in the 2 years before the study, and (before study treatment initiation; NfL [BL]), (2) the geo- a reference value (at month 12 or 24) of the respective outcome – metric mean over 1 year (2 3 values per patient at BL, month 6, (EDSS, T25FWT, 9HPT, or PASAT score) taken at the start of and month 12; NfL [BL-month 12]), and (3) the geometric the analysis period (at BL, month 12, or month 24). The MRI – mean over 2 years (3 5 values per patient at BL, months 6, 12, predictor set for analysis of the period from BL onward con- — 18, and 24 at least 1 value from month 18 or 24; NfL [BL- sisted of BL assessments of normalized brain volume (NBV), month 24]). Patients without a BL NfL assessment still could number of Gd+ T1 lesions, and T2 lesion volume. The MRI contribute to the integral measurements over 12 and/or 24 predictor set for analyses of the period from month 12 or 24 months. The analysis was performed in all patients who received onward consisted of NBV at BL, T2 lesion volume at BL, T2 fi fi ngolimod during the respective studies and remained on n- lesion volume change from BL to month 12 or 24, number of golimod in the extension study (patients who discontinued Gd+ T1 lesions at month 12 or 24, and PBVC from BL to fi from ngolimod and switched to other disease-modifying month 12 or 24. The prognostic value of the various models was therapies had to discontinue from the study and were censored compared by the area under the receiver operating characteristic at this time point). All patients who had at least 1 NfL assess- (ROC) curve. In the area under the ROC curve, the true pos- ment (at BL) and the respective demographic and disease itive rate (sensitivity) is plotted against the false positive rate (1 characteristic data could contribute to the analysis. Only events − specificity) across all possible cutoff values; the higher the area that occurred post-BL, or after the interval used for the cate- under the ROC curve, the better the model.20 In the best case, gorization of patients by NfL levels, were counted in the sta- the area under the ROC curve is one, corresponding to 100% tistical analysis. When patients were categorized by BL NfL, all sensitivity and 100% specificity; in contrast, a random classifi- post-BL outcome events were considered; when patients were cation would lead to an area under the curve (AUC) of 0.5. categorized by the geometric mean NfL level in the first (or second) year, only outcome events with an onset after the first Standard protocol approvals, registrations, (or second) year were included in the statistical analysis. and patient consents The protocols and amendment of studies included in the pre- The prognostic value of NfL for patients reaching EDSS score sent analysis were originally reviewed and approved by the In- ≥4.0, 6m-CDW, and 20% worsening on the T25FWT, 9HPT, dependent Ethics Committees and Institutional Review Boards or PASAT was analyzed using the log-rank test and the Cox for each center per local regulations. All patients or legally ac- proportional hazards model with adjustments for sex, age, cepted representatives of patients provided written informed disease duration, number of relapses in the 2 years before the consent before study entry for the present analysis. The study study, a reference value of the respective analysis outcome was conducted in compliance with the ethical principles of the (EDSS, T25FWT, 9HPT, or PASAT) according to the analysis Declaration of Helsinki and the International Conference on period (BL, month 12, and month 24), and geometric mean Harmonisation Good Clinical Practice Guidelines. NfL by category (high vs low) according to the analysis period. Furthermore, Kaplan-Meier plot results of time-to-event anal- Data availability yses are reported with p values from the log-rank test across Anonymized data will be made available to qualified external NfL categories, with hazard ratios (HRs) and 95% CIs from the researchers, with requests reviewed and approved by an in- Cox proportional hazards model. dependent review panel on the basis of scientific merit.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 Results Prognosis of long-term outcomes by NfL Patient disposition and BL characteristics Disability-related outcomes Of the full FREEDOMS/TRANSFORMS analysis set of pa- A single high (compared with low) NfL measurement at BL tients who received fingolimod 0.5 mg once daily during the was associated with a 2-fold increase in the hazard of reaching core period, 301 patients had at least 1 NfL value, and 274 had EDSS ≥4.0 (HR = 2.19; 95% CI = 1.21–3.97; figure 1–1.1A), an available BL assessment; patients without a BL NfL value but was not predictive of the risk of reaching 6m-CDW could still be included in analyses on NfL measured over 12 (n (figure 1–1.2A), or 20% worsening in the T25FWT = 274) or 24 (n = 132) months. The numbers of NfL values of (figure 1–1.3A), 9HPT (figure 1–1.4A), or PASAT (figure patients contributing to the average geometric mean over 12 1–1.5A). and 24 months are presented in table 1. Demographic and BL characteristics of patients who had an evaluable NfL assessment When using the geometric mean of NfLlong collected over 12 at BL aligned with the overall trial populations of FREEDOMS months (up to 3 measurements), a higher predictive value for and TRANSFORMS (table 2). At BL, the geometric mean of reaching EDSS ≥4 was observed (HR = 2.78; 95% CI = 1.51 NfL was 29.7 pg/mL (table 2), and 110 patients (37%) had –5.10; figure 1–1.1B). Moreover, the geometric mean of ≥ high NfL levels ( 30 pg/mL). NfLlong collected over 12 months predicted 20% worsening in the PASAT (HR = 2.59; 95% CI = 1.04–6.47; figure 1–1.5B. The mean age and sex distribution of patients were similar However, it was not predictive of the risk of reaching 6m-CDW between the high and low NfL categories (table 3). At BL, (HR = 1.53; 95% CI = 0.89–2.62; figure 1–1.2B) or 20% however, patients with high NfL had experienced a higher worsening in the T25FWT (figure 1–1.3B) or 9HPT (figure number of relapses before study entry, had higher EDSS scores, 1–1.4B). more Gd+ lesions, and higher T2 lesion volume compared with patients with low NfL. Patients with high BL NfL had higher A high (compared with low) geometric mean of NfLlong col- EDSS scores at months 12 and 24 and lower PASAT scores at lected over 24 months (up to 5 measurements) was associated month 24; the loss of brain volume over the follow-up was with an 8-fold increase in the hazard of reaching EDSS score ≥4 more pronounced in high NfL patients. The percentage of (HR = 7.91; 95% CI = 2.99–20.92; figure 1–1.1C) and a 3-fold patients completing months 24, 48, 84, and 96 was similar increase in the hazard of reaching 6m-CDW (HR = 3.14; 95% between the high and low NfL categories. CI = 1.38–7.11; figure 1–1.2C) and 20% worsening in the

Table 1 Visit patterns of patients contributing to the average geometric mean NfL values over 12 and 24 months

BL M6 M12 M18 M24 Frequency N = 274 N = 260 N = 269 N = 122 N = 130 (patients, n)

No. of NfL values of patients contributing to the average geometric mean over 12 months

✓✓ 19

✓ ✓ 16

✓ ✓ 11

✓ ✓✓ 228

Total: 274

No. of NfL values of patients contributing to the average geometric mean over 24 months

✓✓✓1

✓✓ ✓1

✓✓✓✓4

✓ ✓✓✓5

✓ ✓✓ ✓9

✓ ✓✓✓ 2

✓ ✓✓✓✓110

Total: 132

Abbreviations: BL = baseline; M = month; NfL = neurofilament light chain. ✓Indicates that the NfL assessment was available at that particular time point.

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Table 2 Patient demographics and disease characteristics (total population)

Fingolimod 0.5 mg FREEDOMS (full TRANSFORMS (full (NfL set) analysis set) analysis set)

Characteristics N = 301 N = 1,272 N = 1,280

Age (y) 37.0 (30, 44) 37.0 (30, 43) 36.0 (30, 43)

Female, n (%) 198 (65.8) 889 (69.9) 861 (67.3)

Duration of MS since first symptoms (y) 6.6 (3.2, 12.4) 6.7 (3.0, 11.9) 5.9 (2.4, 10.7)

No. of relapses in the 2 y before screening 2.0 (1.0, 3.0) 2.0 (1.0, 3.0) 2.0 (1.0, 3.0)

Prior MS treatment, n (%) 166 (55.1) 520 (40.9) 745 (58.2)

EDSS scores at BL 2.0 (1.5, 3.5) 2.0 (1.5, 3.5) 2.0 (1.5, 3.0)

EDSS score at M12 2.0 (1.5, 3.5) 2.0 (1.5, 3.5) 2.0 (1.5, 3.0)

EDSS score at M24 2.0 (1.5, 3.5) 2.0 (1.5, 3.5) 2.0 (1.5, 3.0)

PASAT score at BL 52.0 (44.0, 57.0) 52.0 (43.0, 57.0) 52.0 (42.0, 57.0)

PASAT score at M12 53.0 (47.0, 58.0) 53.0 (44.0, 58.0) 53.0 (45.0, 57.0)

PASAT score at M24 55.0 (49.0, 58.0) 54.0 (46.0, 58.0) 54.0 (45.0, 58.0)

NBV (cm3) 1,521.9 (1,466.3, 1,575) 1,520.4 (1,461.3, 1,574.0) 1,529.5 (1,473.5, 1,577.5)

Change in brain volume from BL to M12 (%) −0.35 (−0.81, 0.08) −0.40 (−1.0, 0.07) −0.30 (−0.7, 0.1)

Change in brain volume from BL to M24 (%) −0.60 (−1.3, −0.2) −0.78 (−1.7, −0.2) −0.50 (−1.2, −1.0)

Presence of Gd+ T1 lesions at BL, n (%) 121 (40.5) 480 (38.1) 437 (34.7)

Number of Gd+ T1 lesions at BL 0 (0, 1) 0 (0, 1) 0 (0, 1)

T2LV at BL (cm3) 3.2 (1.5, 7.6) 3.4 (1.3, 8.3) 2.8 (1.1, 6.7)

Change in T2LV from BL to M12 (cm3) 0.014 (−0.3, 0.3) −0.002 (−0.2, 0.3) 0.059 (−0.13, 0.49)

Change in T2LV from BL to M24 (cm3) 0.009 (−0.32, 0.37) −0.003 (−0.23, 0.48) 0.13 (−0.12, 0.64)

Follow-up duration (y) 8.8 (3.7, 9.2) 8.5 (2.2, 9.4) 6.1 (1.7, 8.9)

Patients who completed M24, n (%) 268 (89.0) 1,127 (88.6) 983 (76.8)

Patients who completed M48, n (%) 229 (76.1) 833 (65.5) 793 (62.0)

Patients who completed M84, n (%) 199 (66.1) 697 (54.8) 633 (49.5)

Patients who completed M96, n (%) 188 (62.5) 658 (51.7) 588 (45.9)

Patients with ≥1 NfL assessment, N = 301 N = 277 N = 473 geometric mean (pg/mL)

BL 29.70 30.09 26.00

M12 17.72 21.67 17.15

M24 17.96 21.27 ··

BL-M12 21.42 23.98 20.12

BL-M24 20.50 22.88 ··

Abbreviations: BL = baseline; EDSS = Expanded Disability Status Scale; Gd+ = gadolinium enhancing; M = month; NBV = normalized brain volume; NfL = neurofilament light chain; PASAT = Paced Auditory Serial Addition Test; Q = quartile; T2LV = T2 lesion volume. Summary statistics are presented as median (Q1, Q3), unless stated otherwise; 301 patients in the fingolimod 0.5 mg group had at least 1 NfL assessment, but only 274 had a BL NfL assessment.

T25FWT (HR = 3.05; 95% CI = 1.38–6.70; figure 1–1.3C). Change in brain volume However, it was not predictive of reaching 20% worsening on Patients with high (compared with low) NfL levels at BL lost the 9HPT (figure 1–1.4C) or PASAT (figure 1–1.5C) in this more brain volume over 120 months (least square mean dif- data set. ference between the high and low category: −1.12%; 95% CI =

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 5 Table 3 Patient demographics and disease characteristics at BL, M12, and M24 (by NfL category at BL)

NfL category

<30 pg/mL ≥30 pg/mL

n = 164 n = 110 p Value

Age (y) 37.0 (31.0, 44.5) 35.5 (29.0, 43.0) NS

Female, n (%) 111 (67.7) 73 (66.4) NS

MS duration since first symptoms (y) 7.2 (3.2, 13.2) 5.8 (3.2, 10.1) NS

Number of relapses in the 2 ys 2.0 (1.0, 2.0) 2.0 (2.0, 3.0) ≤0.0001 before screeninga

Prior MS treatment, n (%) 85 (51.8) 65 (59.1) NS

EDSS score at BL 2.0 (1.5, 3.5) 2.5 (1.5, 3.5) ≤0.05

EDSS score at M12 2.0 (1.5, 3.0) 2.3 (1.5, 3.5) ≤0.05

EDSS score at M24 2.0 (1.5, 3.0) 2.0 (1.5, 3.5) ≤0.05

PASAT score at BL 52.0 (45.0, 57.0) 52.0 (44.0, 57.0) NS

PASAT score at M12 54.0 (48.0, 58.0) 52.0 (46.5, 58.0) NS

PASAT score at M24 56.0 (49.0, 59.0) 54.0 (44.0, 57.0) ≤0.05

T25FWT score at BL 4.7 (4.1, 5.9) 5.1 (4.3, 6.8) ≤0.05

T25FWT score at M12 4.8 (4.2, 6.0) 4.9 (4.3, 6.4) NS

T25FWT score at M24 4.8 (4.2, 5.6) 5.0 (4.2, 6.3) NS

9HPT score at BL 19.7 (18.1, 22.7) 21.3 (19.0, 24.8) ≤0.05

9HPT score at M12 19.6 (17.8, 22.3) 20.7 (18.0, 24.3) NS

9HPT score at M24 19.3 (17.7, 22.1) 20.4 (17.6, 24.0) NS

NBV (cm3) 1,524.4 (1,475.2, 1,572.3) 1,520.0 (1,453.9, 1,585.3) NS

Change in brain volume from BL to M12 (%) −0.20 (−0.60, 0.10) −0.55 (−1.1, −0.17) ≤0.001

Change in brain volume from BL to M24 (%) −0.44 (−1.0, −0.1) −1.10 (−1.8, −0.5) ≤0.0001

Presence of Gd+ T1 lesions at BL, n (%) 39 (23.9) 69 (63.3) ≤0.0001

Number of Gd+ T1 lesions at BL 0 (0, 0) 1 (0, 3.0) ≤0.0001

T2LV at BL (cm3) 1.97 (0.82, 4.86) 6.12 (2.72, 12.48) ≤0.0001

Change in T2LV from BL to M12 (cm3) 0.01 (−0.12, 0.15) 0.01 (−0.51, 0.49) NS

Change in T2LV from BL to M24 (cm3) 0.02 (−0.13, 0.28) −0.06 (−0.76, 0.42) NS

Follow-up duration (y) 8.8 (3.6, 9.2) 8.7 (3.7, 9.3) NS

Patients who completed M24, n (%) 149 (90.9) 95 (86.4) NS

Patients who completed M48, n (%) 125 (76.2) 84 (76.4) NS

Patients who completed M84, n (%) 109 (66.5) 70 (63.6) NS

Patients who completed M96, n (%) 104 (63.4) 66 (60.0) NS

NfL, geometric mean (pg/mL)

BL 19.07 57.47 ≤0.0001

M12 15.67 21.79 ≤0.0001

M24 15.31 21.76 ≤0.05

BL-M12 16.82 32.22 ≤0.0001

BL-M24 16.42 27.36 ≤0.0001

Abbreviations: 9HPT = 9-Hole Peg Test; BL = baseline; EDSS = Expanded Disability Status Scale; Gd+ = gadolinium enhancing; M = month; NBV = normalized brain volume; NfL = neurofilament light chain; NS = not significant; PASAT = Paced Auditory Serial Addition Test; T25FWT = Timed 25-Foot Walk Test; T2LV=T2 lesion volume; Q, quartile. Data are presented as median (Q1, Q3), unless stated otherwise. a Mean ± SD number of relapses in the 2 years before screening: 1.9 ± 0.90 in the <30 pg/mL group and 2.6 ± 1.38 in the ≥30 pg/mL group

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Figure 1 Kaplan-Meier plots of time to event by NfL assessment for disability outcomes (A) at BL, (B) over 12 months, and (C) over 24 months

The reference visit is defined as BL for A, M12 for B, and M24 for C. (1.1) EDSS score ≥4 (only patients with an initial BL EDSS <4 were analyzed); (1.2) 6m-CDW (change of ≥1.5 in EDSS score if initial EDSS = 0, ≥1 if initial EDSS between 1 and 5, or ≥0.5 if initial EDSS >5); (1.3) 20% worsening in the T25FWT; (1.4) 20% worsening in the 9HPT; and (1.5) 20% worsening in the PASAT (only patients with an initial PASAT score >0 were analyzed). 6m-CDW = 6-month confirmed disability worsening; 9HPT = 9-Hole Peg Test; BL = baseline; EDSS = Expanded Disability Status Scale; M = month; NfL = neurofilament light chain; PASAT = Paced Auditory Serial Addition Test; T25FWT = Timed 25-Foot Walk Test.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 7 −2.07 to −0.17), with the difference being statistically signif- measuring NfL only once (table 4). Models that used only a icant from year 3 onward (figure 2, A). Qualitatively similar, single assessment of NfL at BL had a lower AUC compared though not always significant, trends were observed when with models that used an integral measure of NfL over time patients were stratified by the geometric mean of NfL taken (BL-month 12 and BL-month 24). over either 12 or 24 months (figure 2, B–C). It is of note that the number and proportion of patients categorized as having The best prognostic results for the long-term outcomes were high NfL was higher at BL before initiation of study treatment achieved when an integral measure of serial NfL was taken (figure 2, A) compared with NfL assessments taken during the over 24 months in combination with clinical and MRI pa- fingolimod treatment phase (figure 2, B–C). rameters. The area under the ROC curve for the CM + MRI + NfL model was 0.834 for ARBA, 0.954 for reaching EDSS ≥4, Prognostic value of NfL in different predictor 0.849 for 6m-CDW, 0.868 for 20% worsening in the sets for long-term outcomes T25FWT, 0.777 for 20% worsening in the 9HPT, and 0.875 for 20% worsening in the PASAT. Combination of NfL with clinical and/or MRI measures The additional value of NfL to predict changes in brain volume and long-term clinical outcomes over conventional clinical Discussion and/or MRI measures is illustrated in table 4 and figure 3. Regardless of NfL measured at a single time point or integral NfL has been established as the first blood-based biomarker for measures over 12 or 24 months, the area under the ROC curve MS to reflect current disease activity (relapses and lesion for- was generally lowest for models that used only clinical mea- mation) and therapy response; moreover, NfL is able to sures (CM; AUC range, 0.599–0.873), intermediate for models predict—on the group level—the degree of long-term disability that combined clinical measures and NfL (CM + NfL; AUC and features of neuronal degeneration based on BL measure- – range, 0.623–0.927) or clinical and MRI measures (CM + ments before starting disease-modifying therapies.6,8 12,21,22 MRI; AUC range, 0.653–0.939), and highest for models that However, this approach does not factor in post-BL treatment used clinical and MRI measures in combination with NfL (CM effects for the prediction of long-term outcomes, and the ac- +MRI+NfL;AUCrange,0.658–0.954). curacy of single-time NfL assessments could be limited by their short-term fluctuations due to intercurrent acute disease The best prognostic results for the long-term outcomes were activity. achieved when an integral of NfLlong over 24 months was combined with clinical and MRI parameters, indicating that This analysis from the pooled fingolimod phase 3 clinical both MRI and NfL have additional value when each is com- program is the first to address these issues and demonstrates bined with clinical measures, but that NfL has additional, that NfLlong over 12 or 24 months is superior to single BL NfL qualitatively different prognostic value over conventional measures. The combination of NfLlong with clinical and MRI clinical and MRI measures. measures further improves the ability to predict the 10-year disability outcomes for patients with RRMS. Single NfL at BL vs integral measures over 12 and 24 months NfL reflects different disease features compared with MRI; An integral measure of serial NfL assessments was superior for 36.7% of patients whose brain scans were free of Gd+ lesions the prognosis of long-term outcomes in MS compared with at BL had NfL concentrations categorized as high. Plausible

Figure 2 Estimated mean PBVC from BL by NfL assessment (A) at BL, (B) over 12 months, and (C) over 24 months

The reference visit is defined as BL for (A), M12 for (B), and M24 for (C). In (A), where the categorization was performed by BL NfL (before study drug initiation), more patients were categorized as having high NfL (n = 110) compared with (B) (n = 61) and (C) (n = 22) where patients were categorized by a geometric mean under fingolimod treatment. *p ≤ 0.05, **p ≤ 0.001, and ***p ≤ 0.0001 for high vs low NfL. BL = baseline; M = month; NfL = neurofilament light chain; PBVC = percentage brain volume change; SE = standard error.

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN causes for this constellation include lesion formation in the Table 4 Area under the ROC curve for different predictor spinal cord, disease activity that escapes detection on routine sets of clinical, NfL, and MRI parameters for the MRI,21 brain diﬀuse pathology in the gray and/or white prognosis of long-term clinical outcomes and matter, or early phases of lesion formation not yet visible in brain volume change at M84 routine MRI.23 More research with high-frequency MRI and Geometric Geometric NfL measurements is needed to better understand the ki- mean mean Outcomes (predictor sets) BL over 12 mo over 24 mo netics of change in NfL concentration in blood in relation to MS lesion formation in the CNS. EDSS score ≥4

CM 0.840 0.873 0.849 The BL features of all patients who contributed to this NfL CM + NfL 0.874 0.877 0.927 analysis were not notably diﬀerent from the overall population

CM + MRI 0.842 0.939 0.867 of FREEDOMS and TRANSFORMS and representative of a typical RRMS population. However, patients with high NfL at CM + MRI + NfL 0.882 0.945 0.954 BL represented a more active and advanced MS population. 6m-CDW

CM 0.699 0.631 0.681 Patients with high NfL at BL had a higher on-study PBVC for up to 120 months. This is clinically relevant because NBV has been CM + NfL 0.715 0.642 0.781 – showntopredictlong-termoutcomesinMS.24 27 The prog- CM + MRI 0.718 0.678 0.756 nostic value of BL NfL for on-study BVL observed in the current CM + MRI + NfL 0.739 0.683 0.849 study was broadly in line with recently published work, partly

20% worsening in using the same data from FREEEDOMS and TRANS- 10 the T25FWT FORMS. Furthermore, the present results are largely consis-

CM 0.652 0.599 0.617 tent with the Comprehensive Longitudinal Investigations in MS at Brigham and Women’s Hospital (CLIMB)22 and Expression, CM + NfL 0.659 0.623 0.778 Proteomics, Imaging, Clinical (EPIC)28 studies. The CLIMB CM + MRI 0.653 0.697 0.686 study reported a correlation of early annual and averaged yearly

CM + MRI + NfL 0.658 0.720 0.868 serum NfL levels with 10-year MRI outcomes and worsening of fatigue measures.22 In the EPIC study, BL serum NfL levels were 20% worsening in the 9HPT predictive of brain atrophy in the following 2–10 years.28 Of CM 0.652 0.603 0.605 interest, we identiﬁed a lag in time between NfL and BVL in our CM + NfL 0.691 0.682 0.702 study, suggesting that these measures diﬀer in their kinetic

CM + MRI 0.694 0.674 0.618 change. Although the curves of BVL separated almost immediately when categorizing patients by their BL NfL values, a longer CM + MRI + NfL 0.746 0.745 0.777 ﬁ lag time was identi ed when categorizing patients by NfLlong 20% worsening in the PASAT assessment over 12 or 24 months. Acute MS disease activity (e.g.,

CM 0.644 0.641 0.702 Gd+ lesions) could be one explanation for high NfL values in a single NfL assessment at BL, and Gd+ lesions have been iden- CM + NfL 0.667 0.635 0.740 tiﬁed as a strong predictor for on-study BVL in 3 phase 3 trials.29 CM + MRI 0.704 0.733 0.789

CM + MRI + NfL 0.715 0.758 0.875 Consistently, the geometric mean of NfLlong was found to be superior for the prognosis of unfavorable disability outcomes ARBAa ≤ 20.4 compared with single NfL measures at BL. The prediction of CM 0.737 0.709 0.711 the long-term outcomes based on elevated NfLlong is less CM + NfL 0.760 0.733 0.761 inﬂuenced by an intermittent increase of disease activity and ﬂ CM + MRI 0.743 0.734 0.790 hence may better re ect the chronic process of neuronal injury and eventual tissue loss. CM + MRI + NfL 0.762 0.781 0.834

Abbreviations: 6m-CDW = 6-month confirmed disability worsening; 9HPT = The area under the ROC curve analysis demonstrated that 9-Hole Peg Test; ARBA = annualized rate of brain atrophy; BL = baseline; CM long-term outcomes were better predicted when MRI and = clinical model; EDSS = Expanded Disability Status Scale; NfL = neurofilament light chain; PASAT = Paced Auditory Serial Addition Test; PBVC, per- clinical features were combined with NfLlong compared with centage brain volume change; ROC = receiver operating characteristic; when the former 2 were used alone, indicating that NfL SIENA = structural image evaluation using normalization of atrophy; T25FWT long = Timed 25-Foot Walk Test. identifies an additional pathogenesis that escapes the current a SIENA PBVC was converted to ARBA by ([SIENA/100 + 1] [365.25/days] −1) × 100, where “SIENA” represents the PBVC obtained between 2 scans and standard. The conceptual advantage of NfLlong over single “days” means the days for the scan date relative to day 1,32 categorized as NfL measures at BL is the inclusion of the disease-modifying ≤−0.4% vs >0.4% to correct for differences in the distance between MRI ff fi between patients. e ect of therapies as an additional factor de ning the long- term outcomes. Based on these findings, NfL has been

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 9 Figure 3 ROC curves for analyses of ARBA, EDSS ≥4, 6m-CDW, and 20% worsening in the T25FWT, 9HPT, and PASAT at M84

Models with/without predictor NfL (A) at BL, (B) over 12 months, and (C) over 24 months, for (3.1) ARBA up to −0.4%, (3.2) EDSS score ≥4, (3.3) 6m-CDW, (3.4) 20% worsening in the T25FWT, (3.5) 20% worsening in the 9HPT, and (3.6) 20% worsening in the PASAT. 6m-CDW = 6-month confirmed disability worsening; 9HPT = 9-Hole Peg Test; ARBA = annualized rate of brain atrophy; BL = baseline; EDSS = Expanded Disability Status Scale; M = month; NfL = neurofilament light chain; PASAT = Paced Auditory Serial Addition Test; ROC = receiver operating characteristic; T25FWT = Timed 25-Foot Walk Test.

10 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN suggested as an end point for phase 2 studies in progressive Convelo and Population Council; speaking for Mylan; and MS,30 where we currently lack established trial paradigms.31 serving as an Editor of Multiple Sclerosis Journal. A. Shah is an employee of Novartis Healthcare Pvt. Ltd. R. Meinert is an The sample size of this post hoc analysis was limited by the employee of DATAMAP GmbH, which provides services to availability of blood samples, and the current study was not Novartis Pharma AG. D. Leppert was an employee of Novartis powered to show the effects on long-term outcomes in some Pharma AG at the time of outline development (until January of the disability measures. Despite these limitations, a con- 2019); he has received personal compensation for consulting sistent trend toward unfavorable long-term outcomes in pa- and speaking and travel reimbursement from Quanterix, tients with high NfL with HRs up to a factor of 8 between Orion, and Sanofi. D. Tomic was an employee of Novartis patients with high compared with low NfL was found, sug- Pharma AG at the time of submission of the manuscript. J. gesting that NfL is a promising biomarker to stratify patients Kuhle’s institution (University Hospital Basel) received and into risk groups. Given the limited sample size, the focus on used exclusively for research support: consulting fees from only 1 disease-modifying therapy (fingolimod), the post hoc Biogen, Novartis, Roche, and Teva; speaker fees from the nature of this study, and disease heterogeneity, confirmatory Swiss MS Society, Biogen, Novartis, Roche, and Sanofi; travel evidence for the value of NfL for the long-term prognosis of expenses from Merck, Novartis, and Roche; and grants from patients with MS is needed from future prospective clinical ECTRIMS Research Fellowship Programme, University of studies with long-term data collection. Basel, Swiss MS Society, Swiss National Research Foundation (320030_189140/1), Bayer, Biogen, Celgene, Genzyme, An integral measure of longitudinal NfL assessments collected Merck, Novartis, Roche, and Sanofi.Go to Neurology.org/NN over 12 or 24 months might improve the accuracy of the for full disclosures. prognosis of long-term disability outcomes in patients with MS. In the current study, the highest prognostic value was Publication history achieved when an integral measure of NfL in combination Received by Neurology: Neuroimmunology & Neuroinflammation with clinical and MRI features was used. The prognostic value March 30, 2020. Accepted in final form June 19, 2020. of low NfL concentrations for beneficial long-term outcomes on the group level also supports the need to keep NfL levels low in individual patients. Thus, NfL in blood fulfills a critical requirement as a prognostic biomarker for disability wors- Appendix Authors ening and could be useful in monitoring treatment success in Name Location Contribution patients with MS. Dieter A. Novartis Pharma AG, Basel, Study concept, design, Häring, PhD Switzerland execution, data acquisition, Study funding analysis and interpretation, outline The study was funded by Novartis Pharma AG. The study review, critical revision of sponsor participated in the design and conduct of the study, the manuscript, and statistical analysis data collection, data management, data analysis and interpretation, and preparation, review, and approval of the Harald Novartis Pharma AG, Basel, Study concept, design, Kropshofer, Switzerland execution, data acquisition, manuscript. The biostatistical analyses were performed at PhD analysis and interpretation, DATAMAP GmbH, Freiburg, Germany. critical revision of the manuscript, obtaining study funding, and supervising Disclosure the research

D.A. Häring and H. Kropshofer are employees of Novartis Ludwig Research Center for Study concept, design, data Pharma AG. L. Kappos’ institution (University Hospital Basel) Kappos, MD Clinical Neuroimmunology analysis and and Neuroscience Basel interpretation, and critical has received the following in the last 3 years and used exclu- and Departments of revision of the manuscript sively for research support: steering committee, advisory Medicine, Clinical Research, Biomedicine and board, and consultancy fees from Actelion, Addex, Bayer Biomedical Engineering, HealthCare, Biogen Idec, Biotica, Genzyme, Lilly, Merck, University Hospital and Mitsubishi, Novartis, Ono Pharma, Pfizer, Receptos, Sanofi, University of Basel, Switzerland Santhera, Siemens, Teva, UCB, and XenoPort; speaker fees from Bayer HealthCare, Biogen Idec, Merck, Novartis, Sanofi, Jeffrey A. Department of Neurology, Study concept, design, Cohen, MD Mellen MS Center, execution, data acquisition, and Teva; support for educational activities from Bayer Neurological Institute, analysis and HealthCare, Biogen, CSL Behring Genzyme, Merck, Novartis, Cleveland Clinic, OH interpretation, outline fi ; review, and critical revision Sano , and Teva; royalties from Neurostatus products license of the manuscript fees for Neurostatus products; and grants from Bayer Anuja Shah, Novartis Healthcare Pvt. Conducted literature HealthCare, Biogen Idec, European Union, INNO-Swiss, PhD Ltd. Hyderabad, India search, manuscript Merck, Novartis, Roche Research Foundation, Swiss MS So- drafting, revising, and ciety, and the Swiss National Research Foundation. J.A. Cohen editing has received personal compensation for consulting for Continued

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 11 9. Barro C, Benkert P, Disanto G, et al. Serum neurofilament as a predictor of disease Appendix (continued) worsening and brain and spinal cord atrophy in multiple sclerosis. Brain 2018;141: 2382–2391. 10. Kuhle J, Kropshofer H, Haering DA, et al. Blood neurofilament light chain as a Name Location Contribution biomarker of MS disease activity and treatment response. Neurology 2019;92: e1007–e1015. Rolf DATAMAP GmbH, Freiburg, Data analysis and 11. Novakova L, Zetterberg H, Sundstrom P, et al. Monitoring disease activity in multiple Meinert, Germany interpretation, outline sclerosis using serum neurofilament light protein. Neurology 2017;89:2230–2237. PhD review, critical revision of 12. Piehl F, Kockum I, Khademi M, et al. Plasma neurofilament light chain levels in the manuscript, and patients with MS switching from injectable therapies to fingolimod. Mult Scler 2018; statistical analysis 24:1046–1054. 13. Siller N, Kuhle J, Muthuraman M, et al. Serum neurofilament light chain is a bio- David Research Center for Study concept, execution, marker of acute and chronic neuronal damage in early multiple sclerosis. Mult Scler Leppert, MD Clinical Neuroimmunology data acquisition and 2019;25:678–686. and Neuroscience Basel interpretation, critical 14. Kappos L, Radue EW, O’Connor P, et al. A placebo-controlled trial of oral fingolimod and Departments of revision of the manuscript, in relapsing multiple sclerosis. N Engl J Med 2010;362:387–401. Medicine, Clinical obtaining study funding, 15. Cohen JA, Barkhof F, Comi G, et al. Oral fingolimod or intramuscular interferon for Research, Biomedicine and technical support, and relapsing multiple sclerosis. New Engl J Med 2010;362:402–415. Biomedical Engineering, supervising the research 16. Kappos L, O’Connor P, Radue EW, et al. Long-term effects of fingolimod in multiple University Hospital and sclerosis: the randomized FREEDOMS extension trial. Neurology 2015;84:1582–1591. University of Basel, 17. Cohen JA, Khatri B, Barkhof F, et al. Long-term (up to 4.5 years) treatment with Switzerland fingolimod in multiple sclerosis: results from the extension of the randomised TRANSFORMS study. J Neurol Neurosurg Psychiatry 2016;87:468–475. Davorka Novartis Pharma AG, Basel, Study concept, design, 18. Cohen JA, Tenenbaum N, Bhatt A, Zhang Y, Kappos L. Extended treatment with Tomic, DVM, Switzerland execution, data analysis fingolimod for relapsing multiple sclerosis: the 14-year LONGTERMS study results. PhD and interpretation, outline Ther Adv Neurol Disord 2019;12:1756286419878324. review, critical revision of 19. Kappos L, Mehling M, Arroyo R, et al. Randomized trial of vaccination in fingolimod- the manuscript, obtaining treated patients with multiple sclerosis. Neurology 2015;84:872–879. study funding, technical 20. Park SH, Goo JM, Jo CH. Receiver operating characteristic (ROC) curve: practical support, and supervising review for radiologists. Korean J Radiol 2004;5:11–18. the research 21. Ruggieri SLA, Tinelli E, Giglio De L, Prosperini L, Gasperini C, Pozzilli C. Measuring disease activity in multiple sclerosis: the essential role of spinal cord MRI monitoring. Jens Kuhle, Neurologic Clinic and Study concept, data Mult Scler J 2018;24:121–327. MD, PhD Policlinic, Departments of acquisition and 22. Chitnis T, Gonzalez C, Healy BC, et al. Neurofilament light chain serum levels Medicine, Biomedicine and interpretation, outline correlate with 10-year MRI outcomes in multiple sclerosis. Ann Clin Transl Neurol Clinical Research, review, critical revision of 2018;5:1478–1491. University Hospital and the manuscript, obtaining 23. Werring DJ, Brassat D, Droogan AG, et al. The pathogenesis of lesions and normal- University of Basel, study funding, technical appearing white matter changes in multiple sclerosis: a serial diffusion MRI study. Switzerland support, and supervising Brain 2000;123(pt 8):1667–1676. the research 24. Sormani MP, Kappos L, Radue EW, et al. Defining brain volume cutoffs to identify clinically relevant atrophy in RRMS. Mult Scler 2017;23:656–664. 25. Traboulsee AL, Cornelisse feminine P, Sandberg-Wollheim M, et al. Prognostic factors for long-term outcomes in relapsing-remitting multiple sclerosis. Mult Scler J References Exp Transl Clin 2016;2:2055217316666406. 1. Antel J, Antel S, Caramanos Z, Arnold DL, Kuhlmann T. Primary progressive multiple 26. Miller DH, Lublin FD, Sormani MP, et al. Brain atrophy and disability worsening in sclerosis: part of the MS disease spectrum or separate disease entity? Acta Neuro- primary progressive multiple sclerosis: insights from the INFORMS study. Ann Clin pathol 2012;123:627–638. Transl Neurol 2018;5:346–356. 2. Lassmann H, van Horssen J, Mahad D. Progressive multiple sclerosis: pathology and 27. Gaetano L, Haring DA, Radue EW, et al. Fingolimod effect on gray matter, thalamus, and pathogenesis. Nat Rev Neurol 2012;8:647–656. white matter in patients with multiple sclerosis. Neurology 2018;90:e1324–e1332. 3. Lublin FD, Reingold SC, Cohen JA, et al. Defining the clinical course of multiple 28. Canto E, Barro C, Zhao C, et al. Association between serum neurofilament light chain sclerosis: the 2013 revisions. Neurology 2014;83:278–286. levels and long-term disease course among patients with multiple sclerosis followed 4. Yan Y, Jensen K, Brown A. The polypeptide composition of moving and stationary neu- up for 12 years. JAMA Neurol 2019;76:1359–1366. rofilaments in cultured sympathetic neurons. Cell Motil Cytoskeleton 2007;64:299–309. 29. Radue EW, Barkhof F, Kappos L, et al. Correlation between brain volume loss and 5. Teunissen CE, Khalil M. Neurofilaments as biomarkers in multiple sclerosis. Mult clinical and MRI outcomes in multiple sclerosis. Neurology 2015;84:784–793. Scler 2012;18:552–556. 30. Sormani MP, Haering DA, Kropshofer H, et al. Blood neurofilament light as a potential 6. Khalil M, Teunissen CE, Otto M, et al. Neurofilaments as biomarkers in neurological endpoint in phase 2 studies in MS. Ann Clin Translational Neurol 2019;92:e1007–e1015. disorders. Nat Rev Neurol 2018;14:577–589. 31. Leppert D, Kuhle J. Blood neurofilament light chain at the doorstep of clinical ap- 7. Kuhle J, Barro C, Disanto G, et al. Serum neurofilament light chain in early relapsing plication. Neurol Neuroimmunol Neuroinflamm 2019;6:e599. doi: 10.1212/NXI. remitting MS is increased and correlates with CSF levels and with MRI measures of 0000000000000599. disease severity. Mult Scler 2016;22:1550–1559. 32. Jeffery DR, Di Cantogno EV, Ritter S, Meier DP, Radue EW, Camu W. The re- 8. Disanto G, Barro C, Benkert P, et al. Serum Neurofilament light: a biomarker of lationship between the rate of brain volume loss during first 24 months and disability neuronal damage in multiple sclerosis. Ann Neurol 2017;81:857–870. progression over 24 and 48 months in relapsing MS. J Neurol 2016;263:299–305.

12 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN ARTICLE OPEN ACCESS CLASS OF EVIDENCE Validation of the NEOS score in Chinese patients with anti-NMDAR encephalitis

Yujing Peng, MD, Feifei Dai, MD, Lei Liu, MD, Weiqi Chen, MD, Hongyi Yan, MD, Aihua Liu, MD, Correspondence Xinghu Zhang, MD, PhD, Xiaohui Wang, MD, Junying He, MD, Yatong Li, MD, Chenxi Li, MD, Liuxi Chen, MD, Dr. Wang [email protected] Yan Zhao, MD, Lin Li, MD, Qiuying Ma, MD, and Jiawei Wang, MD, PhD

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e860. doi:10.1212/NXI.0000000000000860

Abstract MORE ONLINE Objective Class of Evidence The performance of anti-NMDAR Encephalitis One-Year Functional Status (NEOS) in pre- Criteria for rating dicting the 1-year functional status in Chinese patients with anti-NMDAR encephalitis is therapeutic and diagnostic unknown. studies NPub.org/coe Methods We recruited patients with anti-NMDAR encephalitis from the Multicenter and Prospective Clinical Registry Study of Anti-NMDAR Encephalitis in Beijing Area. Patients were followed up for 1 year. We defined the poor functional status as a modified Rankin Scale score of more than 2 and good functional status as a modified Rankin Scale score of no more than 2. We performed a receiver-operator characteristic analysis to assess the discriminatory power of the NEOS score in predicting the 1-year functional status by using the area under the curve (AUC). Calibration was assessed by Pearson correlation coefficient and Hosmer-Lemeshow tests.

Results Among the 111 patients with anti-NMDAR encephalitis recruited from 364 potentially eligible participants, 87 (78.4%) had good functional status at 1 year, whereas the remaining 24 (21.6%) had poor functional status. The AUC of the NEOS score for 1-year poor functional status was 0.86 (95% CI 0.78–0.93, p < 0.001). The increased NEOS was associated with higher risk of 1-year poor functional status in patients with anti-NMDAR encephalitis.

Conclusions The NEOS score is considered a reliable predictor of the risk of 1-year poor functional status in Chinese patients with anti-NMDAR encephalitis. This score could help to estimate the velocity of clinical improvement in advance.

Clinicaltrial.gov identifier NCT02443350.

Classification of evidence This study provides Class III evidence that in patients with anti-NMDAR encephalitis, the NEOS score predicts 1-year functional status.

From the Department of Neurology, Beijing Tong Ren Hospital, Capital Medical University (Y.P., F.D., L. Liu, Y.L., C.L., L.C., Y.Z., L. Li, Q.M., J.W.); Medical Research Center, Beijing Tong Ren Hospital, Capital Medical University (Y.P., F.D., L. Liu, Y.L., C.L., L.C., Y.Z., L. Li, Q.M., J.W.); Department of Neurology, Beijing Tiantan Hospital, Capital Medical University (W.C., H.Y., X.Z.); Department of Neurology, Xuanwu Hospital, Capital Medical University (A.L.), Beijing; Beijing Children Hospital, Capital Medical University (X.W.); Department of Neurology, the Second Hospital of Hebei Medical University (J.H.), Shijiazhuang; China National Clinical Research Center for Neurological Diseases (W.C., H.Y.), Beijing; Center of Stroke, Beijing Institute for Brain Disorders (W.C., H.Y.); and Beijing Key Laboratory of Translational Medicine for Cerebrovascular Disease (W.C., H.Y.), China.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

The Article Processing Charge was funded by the authors. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work, provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary AUC = area under the receiver operating characteristic curve; ICU = intensive care unit; mRS = modiﬁed Rankin Scale; NEOS = anti-NMDAR Encephalitis One-Year Functional Status; WBC = white blood cell.

Anti-NMDAR encephalitis is a rare, debilitating, and potentially treatable condition that is characterized by acute Figure 1 Trial profile to subacute psychiatric and/or neurologic complaints.1 Early identification of patients with poor prognosis remains to be a major concern in clinical practice.2,3 Some – predictive factors, such as delayed treatment,4 6 intensive – care unit (ICU) admission,7 9 and abnormal CSF inflammation,3,10 might be considered useful in the early identification of patients with poor prognosis. The anti- NMDAR Encephalitis One-Year Functional Status (NEOS) score, including not only the aforementioned factors, has been developed and assisted in predicting the risk of 1-year poor functional status, which in turn is useful in deciding whether early second-line immunotherapy or other novel salvage therapies should be offered to those patients with anti-NMDAR encephalitis.11 However, it has not been validated in Chinese population to date. This study aimed to validate the performance of the NEOS score in Chinese patients with anti-NMDAR encephalitis for predicting poor functional status at 1 year.

Data collection and follow-up Methods We collected detailed information of baseline demographics, time of symptom onset, clinical features, therapeutic regimen, Data sources and the 5 variables involved in the NEOS score (see below). The Multicenter and Prospective Clinical Registry Study of The follow-up duration was at least 1 year, and the in- Anti-NMDAR Encephalitis in Beijing Area (Clinicaltrials. formation of functional status (quantified using the modified gov number: NCT02443350) was a multicenter clinically Rankin Scale [mRS]) was collected through face to face or registered study with consecutive suspected patients with telephone by neurologists who were not aware of this study. encephalitis conducted at 5 clinical centers in China. The Poor functional status was defined as a mRS score of more inclusion criteria were as follows: patients (1) older than 6 than 2 (mRS score of 6 represents death), whereas good months; (2) with at least one or more clinical features of functional status was defined as a mRS score of no more than the following: fever, epilepsy, focal neurologic deficiency 2. We defined relapse of encephalitis as the new onset or symptoms, changes in CSF, changes in EEG, and radio- worsening of symptoms occurring after at least 2 months of graphic abnormalities; (3) with confirmed anti-NMDAR improvement or stabilization.4 Mortality occurs because of encephalitis whose CSF or serum showing a characteristic encephalitis or its associated complications. The outcomes pattern of reactivity in rat brain tissues and specific discussed above were determined by at least 2 neurologists immunolabeling of HEK293 cells expressing GluN1 sub- based on the clinical features. If there was any disagreement, units of NMDAR12,13; and (4) screened at least once for we would resort to a third senior neurologist to reach a systemic tumors. consensus decision. EEG was classified as abnormal when the presence of any of the following was recorded: electrographic Study population seizures, rhythmic slowing, epileptic form discharges, extreme We enrolled patients with anti-NMDAR encephalitis with delta brush,14 focal or diffuse slowing, or abnormal state the available information between July 15, 2014, and changes. Brain MRI scans were classified as abnormal if the February 20, 2019. All participants signed written in- images determined by both radiologists and neurologists were formed consent before study initiation. This study was consistent or suggestive of encephalitis.7 approved by the ethics committee of each study center. A total of 364 patients were included, and 245 (67%) among NEOS score these were excluded because they were diagnosed with The NEOS score11 was derived from the study conducted by other diseases and 8 (7%) patients were lost to follow up Balu and his colleagues, in which 382 patients with confirmed (figure 1). anti-NMDAR encephalitis were recruited. A multivariable

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN logistic regression model was constructed by entering vari- AUC of the NEOS score was presented in figure 2. The ables of ICU admission, the absence of treatment for more NEOS score was shown to be a significant predictor of poor than 4 weeks, improvement delay of more than 4 weeks after functional status (AUC 0.86, 95% CI 0.78–0.93, p < 0.001). starting treatment, abnormal MRI, and CSF white blood cell (WBC) count of more than 20 cells/μL. Each variable was Calibration ability of the NEOS score given 1 point, and the score ranges from 0 to 5 points. The Calibration analysis of the NEOS score showed a moderate NEOS score was strongly associated with the probability of correlation between the predicted and observed probabilities poor functional status at 1 year. of 1-year poor functional status, and the coefficient of r was 0.53 (p < 0.001). The significance level of the Hosmer- Statistical analysis Lemeshow test for prediction of poor functional status was The categorical variables were presented as frequencies 0.35 (figure 3). (percentages). The baseline variables were analyzed by χ2 statistics or Fisher exact tests. OR with 95% CI was used to measure the effect of the NEOS score. We tested the performance of the NEOS score by estimating its discrimination Discussion and calibration. The discriminatory power of the NEOS score Our study showed that the NEOS score well predicted the was assessed by the area under the receiver operating char- probability of 1-year poor functional status after initial acteristic curves (AUCs) and 95% CI. An AUC of 1.0 in- symptom presentation in Chinese patients with anti-NMDAR dicated perfect prediction, and 0.5 indicated no better than encephalitis. random prediction. Calibration was assessed by using Pearson correlation coefficient and Hosmer-Lemeshow tests. The α For patients with poor status, the slow and variable trajectory level of significance was p < 0.05 on both sides. All analyses of recovery and step-by-step treatment meant delay in more were performed using SPSS (version 25; IBM Corp., Armonk, effective treatment, imparting significant stress onto the NY) and SAS software version 9.3 (SAS Institute Inc., family members of patients.15 Understanding the possible Cary, NC). prognosis is essential in providing information to clinicians, patients, and families, as well as potentially influencing the Classification of evidence future treatment strategies. The NEOS score can be easily The primary aim was to explore whether the NEOS score was calculated at bedside within 4 weeks of treatment initiation appropriate for Chinese patients with anti-NMDAR encepha- and capable of discriminating the differences in the probability litis to predict the poor functional status at 1 year. The classi- of poor functional status at 1 year over a wide range of score fi cation of evidence assigned to these questions is Class III. value. To some extent, this might help clinicians to counsel patients and their families. Furthermore, to our knowledge, Data availability our study is the first prospective study to externally validate All data are available to researchers on request for the purpose the NEOS score, except the original study. of reproducing the results or replicating the procedure by directly contacting the corresponding author. This study showed several features that are not consistent with the original NEOS study. Patients with anti-NMDAR encephalitis in our study demonstrated a lower incidence in Results women, with tumors, central hypoventilation, and a better Patient characteristics prognosis, and this is consistent with the results of previous 16–18 Of the 111 patients, 87 (78.4%) had good functional status at studies in China. There was no sex difference, but a 1 year, whereas the remaining 24 (21.6%) had poor functional tendency to women was found. The prevalence of tumors was status. All the 24 patients with poor functional status at 1 year also lower in the Korean population than in the Western study 19 were followed up for 2 years after initial presentation. In this populations. Experts have suggested that a race-specific subset after follow-up for 2 years, 7 (29.2%) were recovered to factor, the human leucocyte antigen, or other genetic factors 20 good functional status. might play a significant role. It is unclear as to whether the discrepancy of central hypoventilation incidence occurs be- The baseline characteristics of patients are shown in table 1. cause of earlier diagnosis or if they showed location differ- The patients with poor functional status were more likely to ences in the brainstem control of breathing or expression of be younger, had central hypoventilation, had abnormal MRI, NMDAR between different populations. The specific mech- had CSF WBC counts of more than 20 cells μL, had no anism requires further exploration on brain imaging and an- treatment within 4 weeks of symptom onset, had treatment imal model. In addition, more than half of the patients (57%) delay of >4 weeks, and had first-line immunotherapy (table 1). were juveniles in our sample. An anti-NMDAR encephalitis study on children and adolescents has reported that the in- NEOS score and risk of poor functional status cidence of central hypoventilation in juveniles is lower, and As shown in table 2, patients with higher NEOS scores have the symptoms seemed to be less severe than that in adults.21 significantly higher rate of 1-year poor functional status. The As a matter of fact, the main reasons for ICU admission in our

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 Table 1 Characteristics of patients with anti-NMDA receptor encephalitis

Good status (N = 87), n (%) Poor status (N = 24), n (%) p Value

Age, y 0.016

<45 82 (94.3) 18 (75)

≥45 5 (5.7) 6 (25)

Female 48 (55.2) 15 (62.5) 0.52

Tumor 4 (4.6) 0 (0) 0.29

Altered behavior 72 (82.8) 22 (91.7) 0.45

Memory impairment 37 (42.5) 15 (62.5) 0.08

Speech disorder 53 (60.9) 15 (62.5) 0.89

Seizure 72 (82.8) 18 (75) 0.57

Reduced consciousness 42 (48.3) 8 (36.4) 0.32

Movement disorder 60 (69) 19 (79.2) 0.33

Sleep disturbance 51 (58.6) 15 (62.5) 0.73

Autonomic dysfunction 27 (31) 11 (45.8) 0.18

Central hypoventilation 0 (0) 7 (29.2) <0.001

Requires ICU admission 26 (29.9) 9 (37.5) 0.48

Abnormal EEG 60 (70.6) 19 (79.2) 0.41

Abnormal MRI 36 (41.4) 20 (83.3) <0.001

CSF WBC count 0.045

≤20 cells/μL 56 (64.4) 10 (41.7)

>20 cells/μL 31 (35.6) 14 (58.3)

Time to start of treatment after symptom onset <0.001

≤4wk 71 (81.6) 8 (33.3)

>4 wk 16 (18.4) 16 (66.7)

Time to initial improvement <0.001

≤4wk 69 (79.3) 4 (16.7)

>4 wk 18 (20.7) 20 (83.3)

Line of immunotherapy 0.038

First 66 (75.9) 13 (54.2)

Second 21 (24.1) 11 (45.8)

Abbreviations: ICU = intensive care unit; WBC = white blood cell.

cohort were coma, seizures, agitation/confusion, and acute with larger sample size are warranted to validate external validity respiratory failure in turn. Multiple studies reported better of the NEOS score and to better identify the subgroup of pa- prognosis in the Chinese population, suggesting that the tients with poor functional status. Second, the study included prognosis of anti-NMDAR encephalitis might be pre- variations in treatment approaches. For pediatric patients, dominantly related to race and natural history of the disorder. second-line immunotherapy might be limited because of security reasons when ﬁrst-line therapies fail. Third, Western populations However, our study has some limitations. First, our study in- were not included, and so inter-racial diﬀerences cannot be excluded only 5 major hospitals with more medical resources and cluded. Finally, our study lacked biomarkers that are connected experts than other hospitals in rural areas and included small with response to treatment, which might decrease the scores’ sample size. Thus, selection bias can arise in the study. Studies ability to predict the ultimate clinical outcome.

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Table 2 NEOS score and risk of poor functional status Figure 3 Calibration plot of the NEOS score for poor functional status NEOS p Outcome score n/N OR (95% CI) Value

Poor 0–1 1/53 (1.89) Reference — status

2 3/21 (14.29) 8.67 (0.85–88.70) 0.069

3 11/23 47.67 <0.001 (47.83) (5.60–405.62)

4–5 8/14 (57.14) 69.33 <0.001 (7.35–653.91)

Abbreviation: NEOS = anti-NMDAR Encephalitis One-Year Functional Status.

In conclusion, our study showed that the NEOS score was a reliable and accurate tool for physicians to predict the risk of poor functional status in Chinese patients with anti- NMDAR encephalitis at 1 year. This score could be helpful to estimate the velocity of clinical improvement and might allow clinicians to stratify patients who could beneﬁtfrom The vertical lines indicate 95% CIs of predicted rates of clinical status. NEOS = novel therapies in the future clinical trials. anti-NMDAR Encephalitis One-Year Functional Status.

Acknowledgment The authors thank Yuesong Pan from Beijing Tiantan Study funding Hospital, Capital Medical University, Beijing, China, for Supported by grants from the following: National Natural Sci- providing statistics assistance, and Jiejie Li from Beijing ence Foundation of China (81771313, 81870950); Youth Pro- Tiantan Hospital, Capital Medical University, Beijing, gram of National Natural Science Foundation of China China, for revising the manuscript. In addition, they (81301029); Beijing Municipal Natural Science Foundation appreciated all the investigators and patients who partici- (19G11041, 7182077); Beijing Hospitals Authority Youth Pro- pated in the registry. gram (QML20150206); Key Research and Development Plan of the Ministry of Science and Technology of the People’s Republic of China (2016YFC0904502); Beijing Science and Technology Project “Capital Characteristics” (Z171100001017039); Key Figure 2 Receiver operator characteristic curve for pre- Projects of Medical Development in Capital (2014-1-1101); diction of 1-year prognosis of the NEOS score Beijing Tongren Hospital, Capital Medical University, Key Medical Development Plan (TRYY-KYJJ-2017-054).

Disclosure The authors have no conﬂicts interest to declare as regards this study. Go to Neurology.org/NN for full disclosures.

Publication history Received by Neurology: Neuroimmunology & Neuroinﬂammation March 30, 2020. Accepted in ﬁnal form June 29, 2020.

Appendix Authors

Name Location Contribution

Yujing Capital Medical Study concept and design, Peng, MD University, China acquisition of data, data analysis and interpretation, and manuscript drafting

Feifei Dai, Beijing Tong Ren Acquisition of data and study NEOS = anti-NMDAR Encephalitis One-Year Functional Status. MD Hospital, China supervision or coordination

Continued

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 5 References Appendix (continued) 1. Dalmau J, Graus F. Antibody-mediated encephalitis. N Eng J Med 2018;378:840–851. 2. Dalmau J, Gleichman AJ, Hughes EG, et al. Anti-NMDA-receptor encephalitis: case ff – Name Location Contribution series and analysis of the e ects of antibodies. Lancet Neurol 2008;7:1091 1098. 3. Broadley J, Seneviratne U, Beech P, et al. Prognosticating autoimmune encephalitis: a systematic review. J Autoimmun 2019;96:24–34. Lei Liu, Beijing Tong Ren Acquisition of data and 4. Titulaer MJ, McCracken L, Gabilondo I, et al. Treatment and prognostic factors for MD Hospital, China manuscript revising long-term outcome in patients with anti-NMDA receptor encephalitis: an observational cohort study. Lancet Neurol 2013;12:157–165. Weiqi Beijing Tiantan Hospital, Manuscript revising 5. Titulaer MJ, McCracken L, Gabilondo I, et al. Late-onset anti-NMDA receptor en- Chen, MD China cephalitis. Neurology 2013;81:1058–1063. 6. Irani SR, Stagg CJ, Schott JM, et al. Faciobrachial dystonic seizures: the influence of Hongyi Beijing Tiantan Hospital, Data analysis and immunotherapy on seizure control and prevention of cognitive impairment in a Yan, MD China interpretation broadening phenotype. Brain 2013;136:3151–3162. 7. Graus F, Titulaer MJ, Balu R, et al. A clinical approach to diagnosis of autoimmune Aihua Liu, Xuanwu Hospital, China Acquisition of data and study encephalitis. Lancet Neurol 2016;15:391–404. MD supervision or coordination 8. Jaquet P, de Montmollin E, Dupuis C, et al. Functional outcomes in adult patients with herpes simplex encephalitis admitted to the ICU: a multicenter cohort study. Xinghu Beijing Tiantan Hospital, Acquisition of data and study Intensive Care Med 2019;45:1103–1111. Zhang, China supervision or coordination 9. de Montmollin E, Demeret S, Brulé N, et al. Anti-N-methyl-d-aspartate receptor MD, PhD encephalitis in adult patients requiring intensive care. Am J Respir Crit Care Med 2017;195:491–499. Xiaohui Beijing Children Hospital, Acquisition of data and study 10. Pillai SC, Hacohen Y, Tantsis E, et al. Infectious and autoantibody-associated en- Wang, MD China supervision or coordination cephalitis: clinical features and long-term outcome. Pediatrics 2015;135:e974–e984. 11. Balu R, McCracken L, Lancaster E, Graus F, Dalmau J, Titulaer MJ. A score that Junying The Second Hospital of Acquisition of data and study predicts 1-year functional status in patients with anti-NMDA receptor encephalitis. He, MD Hebei Medical supervision or coordination Neurology 2019;92:e244–e252. University, China 12. Dalmau J, Lancaster E, Martinez-Hernandez E, Rosenfeld MR, Balice-Gordon R. Clinical experience and laboratory investigations in patients with anti-NMDAR en- Yatong Li, Beijing Tong Ren Acquisition of data cephalitis. Lancet Neurol 2011;10:63–74. MD Hospital, China 13. Li L, Sun L, Du R, et al. Application of the 2016 diagnostic approach for autoimmune encephalitis from Lancet Neurology to Chinese patients. BMC Neurol 2017;17:195. Chenxi Li, Beijing Tong Ren Acquisition of data 14. Schmitt SE, Pargeon K, Frechette ES, Hirsch LJ, Dalmau J, Friedman D. Extreme delta MD Hospital, China brush: a unique EEG pattern in adults with anti-NMDA receptor encephalitis. Neu- rology 2012;79:1094–1100. Liuxi Beijing Tong Ren Acquisition of data 15. Vora NM, Holman RC, Mehal JM, Steiner CA, Blanton J, Sejvar J. Burden of encephalitis- Chen, MD Hospital, China associated hospitalizations in the United States, 1998–2010. Neurology 2014;82:443–451. 16. Wang W, Li JM, Hu FY, et al. Anti-NMDA receptor encephalitis: clinical character- Yan Zhao, Beijing Tong Ren Acquisition of data istics, predictors of outcome and the knowledge gap in southwest China. Eur J Neurol MD Hospital, China 2016;23:621–629. 17. Wang Y, Zhang W, Yin J, et al. Anti-N-methyl-d-aspartate receptor encephalitis in Lin Li, MD Beijing Tong Ren Acquisition of data children of Central South China: clinical features, treatment, influencing factors, and Hospital, China outcomes. J Neuroimmunol 2017;312:59–65. 18. Zhang Y, Liu G, Jiang M, Chen W, He Y, Su Y. Clinical characteristics and prognosis of severe Qiuying Beijing Tong Ren Acquisition of data anti-N-methyl-D-aspartate receptor encephalitis patients. Neurocrit Care 2018;29:264–272. Ma, MD Hospital, China 19. Lim JA, Lee ST, Jung KH, et al. Anti-N-methyl-d-aspartate receptor encephalitis in Korea: clinical features, treatment, and outcome. J Clin Neurol 2014;10:157–161. Jiawei Beijing Tong Ren Study concept and design, data 20. Irani SR, Bera K, Waters P, et al. N-methyl-D-aspartate antibody encephalitis: tem- Wang, MD, Hospital, China analysis and interpretation, poral progression of clinical and paraclinical observations in a predominantly non- PhD obtaining funding, and study paraneoplastic disorder of both sexes. Brain 2010;133:1655–1667. supervision or coordination 21. Florance NR, Davis RL, Lam C, et al. Anti-N-methyl-D-aspartate receptor (NMDAR) encephalitis in children and adolescents. Ann Neurol 2009;66:11–18.

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN ARTICLE OPEN ACCESS Progressive multifocal leukoencephalopathy and sarcoidosis under interleukin 7 The price of healing

Aurélien Guffroy, MD, PhD, Morgane Solis, PharmD, PhD, Vincent Gies, PharmD, PhD, Correspondence Yannick Dieudonne, MD, Cornelia Kuhnert, MD, Cédric Lenormand, MD, PhD, Laurent Kremer, MD, PhD, Dr. Guffroy [email protected] Anne Molitor, PhD, Raphaël Carapito, PhD, Yves Hansmann, MD, PhD, Vincent Poindron, MD, PhD, Thierry Martin, MD, PhD, Sandrine Hirschi, MD, and Anne-Sophie Korganow, MD, PhD

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e862. doi:10.1212/NXI.0000000000000862 Abstract Objective To report the association of JC virus infection of the brain (progressive multifocal encephalopathy [PML]) during the course of sarcoidosis and the challenging balance between immune reconstitution under targeted cytokine interleukin 7 (IL7) therapy for PML and immunosuppression for sarcoidosis.

Methods Original case report including deep sequencing (whole-exome sequencing) to exclude a primary immunodeﬁciency (PID) and review of the literature of cases of PML and sarcoidosis.

Results We report and discuss here a challenging case of immune reconstitution with IL7 therapy for PML in sarcoidosis in a patient without evidence for underling PID or previous immunosuppressive therapy.

Conclusions New targeted therapies in immunology and infectiology open the doors of more specific and more specialized therapies for patients with immunodeficiencies, autoimmune diseases, or cancers. However, before instauration of these treatments, the risk of immune reconstitution inflammatory syndrome and potential exacerbation of an underlying disease must be considered. It is particularly true in case of autoimmune disease such as sarcoidosis or lupus.

From the Department of Clinical Immunology and Internal Medicine (A.G., V.G, Y.D., V.P., T.M., A.-S.K.), National Reference Center for Systemic Autoimmune Diseases (CNR RESO), Tertiary Center for Primary Immunodeficiency, Strasbourg University Hospital; Université de Strasbourg (A.G., M.S., V.G., Y.D., T.M., A.-S.K.), INSERM UMR - S1109; Universitéde Strasbourg (A.G., M.S., Y.D., C.L., Y.H., T.M., A.-S.K.), Faculty of Medicine; Virology Laboratory (M.S.), Strasbourg University Hospital; Université de Strasbourg (V.G.), Faculty of Pharmacy, Illkirch, France; Internal Medicine and Intensive Care (C.K.), Strasbourg University Hospital; Department of Dermatology (C.L.), Strasbourg University Hospital; Department of Neurology (L.K.), INSERM U1119, Biopathologie de La Myéline, Neuroprotection et Stratégies Thérapeutiques, Université de Strasbourg, Fédération de Médecine Translationnelle de Strasbourg (FMTS); Université de Strasbourg (A.M., R.C.), INSERM UMR-S1109, GENOMAX Platform, Fédération Hospitalo-Universitaire OMICARE, FacultédeMédecine, Fédération de Médecine Translationnelle de Strasbourg (FMTS), LabEx TRANSPLANTEX; Department of Infectious Diseases (Y.H.), Strasbourg University Hospital; and Departement of Pneumology (S.H.), Strasbourg University Hospital, Strasbourg, France.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

The Article Processing Charge was funded by Hopitauxˆ Universitaires de Strasbourg. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary BREF = “Batterie rapide d’efficience frontale” test (also called Dubois’ test); IRIS = immune reconstitution inflammatory syndrome; JCV = JC virus; MMSE = Mini Mental State Examination; PID = primary immunodeficiency; PML = progressive multifocal encephalopathy; STAT1 GOF = gain-of-function in STAT1 (signal transducer and activator of transcription 1) gene.

Progressive multifocal encephalopathy (PML) is a devasting de- lymphadenopathy on the chest CT (figure). A bronchial bi- myelinating disease of the brain white matter described for the first opsy showed a noncaseating granuloma. He developed severe time by Astr¨¨ om and colleagues in 1958 in a context of hematologic lymphopenia up to 350/mm3 (normal > 1,500/mm3). Con- malignancy. PML is the consequence of the glial cell opportunistic comitantly, cognitive impairment (BREF 10/18; MMSE 25/ infection by the human JC virus (JCV). Although asymptomatic 30), frontal syndrome with perseverance, aphasia, and speech JCV infection usually occurs in childhood and remains clinically disturbance were noticed. First referred in psychiatry for un- silent in adult throughout life, active JCV replication in the brain usual behavior, he was then hospitalized in the neurology could occurs in primary or secondary immunodeficiencies (mainly department where the diagnosis of PML was made based on AIDS and hematologic malignancies) leading to PML. Immuno- suggestive MRI demyelinating lesions of the white matter and suppressive or immunomodulatory drugs such as natalizumab in positive JCV load in serum and CSF (figure and table 1). At MS or rituximab have already been linked to PML, and some cases this time, he did not receive any therapy for sarcoidosis. His have been reported in sarcoidosis with severe CD4+ lymphope- cognitive impairment became worse (BREF 7/18; MMSE nia.1 Currently, achievement of immune reconstitution is the only 17/30) during the first days of hospitalization, and a com- curative option. In this view, interleukin 7 (IL7) or anti-PD1 passionate regiment of IL7 was started. Thanks to a 4-week therapies have been suggested to help the control of JCV in regiment of IL7 at 10 mg/kg/wk, the neurologic symptoms – standing the immune response to the virus.2 4 improved, and the viral load felt to undetectable value (figure and table 1). Recent sharply demarcated erythema with fine Case report scaling of seborrheic areas (presternal region and hairy zones of the face, scalp, and groin) along with erythemato-squamous A 45-year-old man with PML was referred for acute re- plaques of the elbows was noticed, strongly suggestive of ei- spiratory failure and hypercalcemia after IL7 treatment in the ther sebopsoriasis or profuse seborrheic dermatitis. Skin bi- context of underlying thoracic sarcoidosis. His medical history opsy with histopathologic examination demonstrated only started 3 months with progressive dyspnea and cough re- nonspecific dermatitis without any granuloma, thus ruling out vealing a mild micronodular infiltrate and mediastinal psoriasiform sarcoidosis. At the same time, the patient had

Figure PML and sarcoidosis evolution

MRI (axial, fluid-attenuated inversion recovery [FLAIR] sequences) of the brain at diagnosis of PML (A) and after IL7 therapy (B) showing a regressionof inflammatory lesions of the white matter of the frontoparietal lobes predominant in the left hemisphere of the brain. TDM of the chest before IL7 showing light micronodular interstitial infiltrate related to sarcoidosis (C), dramatic worsening after 4 doses of IL7 therapy related to immune reconstituting syndrome (D), and improvement 3 months later after IL7 discontinuation, steroids, and hydroxychloroquine (E).

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Table 1 Clinical and biological features

Diagnosis After 4 wk of IL7 M+1 M+3 M+6

Clinical symptoms

Weight 62 63 65 65 65

Fever (>38.2°C) No Yes Yes No No

Dyspnea (NYHA) I/IV II/IV I/IV No No

Confusion Yes No No No No

Aphasia Yes No No No No

BREF 7/18 ND ND ND 18/18

MMSE 17/30 ND ND ND 30/30

Splenomegaly No No Yesa No No

Autoimmune manifestations No No Yes (psoriasis flare-up) Yes (psoriasis flare-up) No

Therapies None IL7 Stero¨ıds + HCQ Stero¨ıds + HCQ HCQ alone

Biological parameters

Lymphocyte count (/mm3) 340 400 400 540 560

CD4+ T cells 120 NA 141 150 150

Hypercalcemia (mmol/L) No NA Yes (3.1) No No

1-25-OH-VitD3 NA NA 143 ng/L (N < 90) NA Normal

Conversion enzyme activity 65.8 NA Normal Normal Normal

Gamma-globulinemia (g/L) NA NA 4.1g/L NA 7.4g/L

JCV load (copies/mL)

Serum 15,000 1,400 712 112 Undetectable

CNS 150,000 NA NA NA NA

Radiology

MRI Yes No Yes (improve) No Yes (improve)

18F-FDG-TEP-TDM Yes No Yesa No No

Chest CT Yes No Yes (worsening) No Yes (improve)

Abbreviations: HCQ = hydroxychloroquine; IFN = interferon; JCV = JC virus; NA = not available; ND = no data; NYHA = New York Heart Association. a 18F-FDG-TEP-TDM show an uptake of the spleen, the lung, and hypermetabolism of the chest and coealio-mesenteric adenopathies. hypercalcemia and required oxygen with a serious worsening common variable immunodeficiency for example). In this of his micronodular lung infiltrate (figure). Bronchoalveolar way, we decide to perform thorough immunophenotyping lavage showed a predominant lymphocytosis (86%) with 5,8 and genetic explorations. Analysis of his immune status as well CD4:CD8 ratio and negative microbiology. One 25-OH-vi- as whole-exome sequencing analysis of the patient with his 2 tamin D3 activity was elevated (143 ng/L, normal <90). Flare- parents did not reveal any evidence for PID. up of sarcoidosis was suspected. He was then transferred to the intensive care unit for acute respiratory failure with a huge IL7 was stopped. An immunosuppressive treatment of sar- worsening of his interstitial lung disease treated with nasal coidosis was introduced and associated with a tight control of oxygen, and hyperhydration was given with bisphosphonates blood JCV load (table 1). Solumedrol (40 mg, twice a day) (pamidronate 90 mg IV) to control calcemia (figure). was started, carefully relayed after 5 days by moderate dose of oral steroids (prednisone 30 mg/d) with a decreasing pro- Several recent observations and cases series support the hy- tocol for 6 months and prophylactic treatments (cotrimox- pothesis of underlined inherited immunodeficiency in PML. azole and valaciclovir). In addition, hydroxychloroquine (400 Moreover, granuloma is a frequent symptom revealing a pri- mg/d) was started as a steroid-sparing agent. Under these mary immunodeficiency (PID) (in almost 20% of cases in therapies, sarcoidosis and psoriasis improved. PML did not

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 relapse during the 6 months of follow-up. However, the PML is rarer than other opportunistic infections in sar- lymphopenia persisted around 600/mm3. coidosis with a very poor prognosis. Our review of literature (n = 41 cases) identifies a median age of 47 years (range 24–74 years) with a sex ratio (M/F) of 1.9 (table 2) and Methods highlights a mortality rate of 61.5% at last follow-up. Unlike other opportunistic infections, PML may occur in untreated Whole-exome sequencing patients (table 2).1 In this context, it is of importance to Genomic DNA was isolated from patient and parents’ pe- make the differential diagnosis with neurosarcoidosis, which ripheral blood or saliva using standard protocols. Exome se- required the use of intensive immunosuppressive therapy. quencing libraries were prepared with the Twist Library Patients with sarcoidosis and opportunistic infections, es- Preparation Kit and captured with Human Core Exome pecially with no history of immunosuppressive treatment, probes extended by Twist Human RefSeq Panel (Twist Bio- may belong to a more sensitive subgroup of patients with science, San Francisco, CA) following the manufacturer’s inherited susceptibility factors. In this view, such factors recommendations. Paired-end (2 × 75 bp) sequencing was described in inherited pediatric and familial cases of sar- performed on a NextSeq500 sequencer (Illumina, San Diego, coidosis involving autophagy pathway or T-cell activation CA). Before any processing, quality control was performed 7 pathway could be crucial for response against pathogens. using FastQC. The raw reads data were next mapped using Moreover, it is also important to consider the hypothesis of the Burrows-Wheeler Alignment tool. For each sample, av- PID with granulomatous manifestations and opportunistic erage target read coverage was at least 60-fold. After read infections (PML) in some combined immunodeficiencies mapping, further quality indicators were calculated from the (i.e., Immunodeficiency, Centromeric region instability, resulting BAM file using SAMtools, Qualimap. Variant calling Facial anomalies syndrome, DOCK8 deficiency, STAT1 was performed using the GATK HalotypeCaller of the GATK 8 GOF or other combined immunodeficiencies). In our case, software suite. The annotation was performed by VEP, the we excluded the eventuality of an inherited error of immu- Ensembl Variant Effect Predictor. We focused only on nity by several arguments: (1) the age at onset of the op- protein-altering variants (missense, nonsense, splice site var- portunistic infection and the past medical history of the iants, and coding indels) with alternative allele frequencies patient and his family; (2) the late onset of lymphopenia <0.005 in the 1000 Genomes Project, the Genome Aggrega- with prior normal blood examinations; and (3) the whole- tion Database, the Exome Aggregation Consortium, and an exome analysis that excluded a known mutation in PID- internal exome database including ;700 exomes. To identify related genes. potential causal variants, we further filtered the variants based on a de novo and recessive mode of inheritance. PML, caused by invasive JCV infection of the brain, is a ff Statistics poor prognosis a ection with a fatal outcome in weeks or Data are presented as median (range) or frequency (%) with months in the absence of immune reconstitution. Recently, 95% CIs. Statistical analyses were performed using JMP 7.0. IL7 cytokine therapy and anti-PD1 monoclonal therapy have been proposed to restore an immunity against JCV in a context of secondary immunodeficiency, with or without Data availability 2,4 Anonymized data will be shared by request from any qualified vaccination strategies. Before using such therapies, it is investigator. very important to consider, however, (1) the risk of immune reconstitution inflammatory syndrome (IRIS) and (2) the potential exacerbation of an underlying disease or Discussion the emergence of secondary immune/inflammatory manifestations. It is particularly true in case of autoimmune To date, only few cases of sarcoidosis and PML have been disease such as sarcoidosis or lupus in which imbalance in reported (table 2). The main cause of the immunodeficiency immune-activating cytokine such as IL7 (one of the most leading to such opportunistic infections in this situation re- important cytokines to T-cell expansion and activation) mains unclear. Indeed, lymphopenia, and mainly the CD4+- could favor a flare. Mechanisms leading to the worsening of lymphocytes decreased, is usually considered as a redistribution apreviousinflammatory/autoimmune situation have al- of cells (demargination) in the organs involved by the disease ready been described not only in PML associated with and is not considered at risk for infections. Nonetheless, some AIDS but also in apparently non-immunocompromised case reports or series report severe or opportunistic infections patients.9 occurring in less than 5% of cases of sarcoidosis.5,6 The most frequent infections are Cryptococcus, mycobacterial infection, In our case, the challenging point was to restore the immunity nocardiosis, histoplasmosis, pneumocystis, and Aspergillus in- against JCV and to control the sarcoidosis activity. IRIS has fections. Their occurrence is closely linked not only to the been well described in HIV-infected patients as a paradoxical severity (neurologic form of sarcoidosis) and activity of the reaction after introduction of the treatment. Even if rare, IRIS disease but also to the immunosuppressive therapy (steroids in can potentially occur in all granulomatous diseases and not first and cyclophosphamide).5 just infectious ones.

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Table 2 Literature review of PML cases and sarcoidosis

Time from Cases (sex/ Therapies diagnosis of Lymphocyte CD4+ age at PML before sarcoidosis to count count Outcome References diagnosis, y) PML PML (mo) (/mm3) (/mm3) Treatment Tolerance of PML

Christensen E et al. M/59 No 0 NA NA No NA Death Acta Psychiatr Neurol Scand. 1955

Headington JT. M/41 NA NA NA NA NA NA Neurology. 1962

Mariott PJ et al. J F/52 CS 72 NA NA Cytarabine (2 mg/ Good Improve Neurol Neurosurg. kg/d) 1975

Smith CR et al. 1982 M/32 CS 108 NA NA Cytarabine NA Death

Rosenbloom MA F/59 CS 0 820 NA CS (increased) NA Death et al. Chest. 1983

Iannarella G et al. F/68 No 26 NA 500 CS and acyclovir NA Death Ann Med Int. 1992

Steiger MJ et al. M/37 NA 0 NA NA Cytarabine, NA Improve Ann Neurol. 1993 acyclovir, and IFN-a

Heide W. Ann M/47 CS 132 NA NA Cytarabine and NA Death Neurol. 1995 IFN-a

Hammarin AL et al. M/51 NA NA NA NA NA NA NA J Clin Microbiol. 1996

Jochum W et al. M/54 NA NA NA NA NA NA NA Acta Neuropathol. 1997

Olindo S. Rev Neuro F/47 CS 168 NA NA NA NA Death (Paris). 2000

Mackowiak- F/70 CS/MTX NA 1,000 <300 Cidofovir and IL2 NA Death Cordoliani MA et al. Rev Neurol (Paris). 2001

Volker¨ HU et al. Clin M/49 CS 0 580 258 IVIg and cidofovir NA Death Neurol Neurosurg. 2007

Owczarczyk et al. F/48 CS/MTX 258 580 NA Cidofovir, Good Improve Rheumatology. mirtazapine, and 2007 HCQ

De Raedt S et al. M/43 No 0 NA 88 Cidofovir and Stable Stabilization Clin Neurol steroids Neurosurg. 2008

Le Guilloux J et al. M/42 CS/HCQ 144 NA NA Cidofovir and IL2 Bad Death Rev Neurol (Paris). (4,5MIU 2/d) (worsening) 2009

Granot R et al. J Clin M/49 CS 11 NA NA Cidofovir Bad Death Neurosci. 2009 (worsening)

Neeb L et al. J F/56 228 310 NA Cidofovir Bad Death Neurol. 2009 (worsening)

Yagi T et al. Clin M/34 No 156 NA NA Cidofovir Good Improve Neurol Neurosurg. 2010

Goldbecker A et al. F/50 No 240 Normal Normal CS and acyclovir Good Int J Infect Dis. 2010

Continued

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 5 Table 2 Literature review of PML cases and sarcoidosis (continued)

Time from Cases (sex/ Therapies diagnosis of Lymphocyte CD4+ age at PML before sarcoidosis to count count Outcome References diagnosis, y) PML PML (mo) (/mm3) (/mm3) Treatment Tolerance of PML

Gofton TE et al. J F/54 CS 72 310 NA Cidofovir Bad Death Neurol Neurosurg (worsening) Psychiatr. 2011

Park JH et al. Case M/45 CS 4 NA NA Cidofovir and Good Improve Rep Neurol. 2011 mirtazapine

Hohfeld SK et al. M/39 No 0 600 171 Mirtazapine Bad Death BMJ Case Rep. 2012 (worsening)

Keith J et al. M/74 CS/MTX/ NA 300 260 No Bad Death Neuropathology. HCQ 2012

Pallin M et al. QJM. M/47 CS/HCQ 180 488 131 Cytarabine, Good Improve 2012 cidofovir, mirtazapine, and mefloquine

Davis MJ et al. The F/68 No 0 NA 182 No NA Death Neurologist. 2013

Jamilloux Y et al. M/40 CS 146 300 167 Cidofovir and Good Improve Neurology. 2014 mirtazapine

M/24 CS, MTX, 96 1,299 NA No NA Improve CYC, and aTNFa

M/41 No 0 1,530 354 Cytarabine and Good Improve mirtazapine

F/32 No 171 1,620 426 No NA Death

M/42 CS, MTX, 67 NA 131 Cidofovir and IL2 Bad Death and HCQ (worsening of PML)

M/35 CS 38 900 NA Cidofovir and Good Stabilization mirtazapine

M/27 CS and 138 720 NA Mefloquine NA Death MTX

M/37 CS 0 500 115 IL2 Medium Death (cutaneous rash)

F/36 CS 332 830 101 Cidofovir Good Death

M/35 CS, MTX, 147 900 NA Cidofovir and IL2 NA Death and HCQ

Lefaucheur R et al. M/30 CS 60 NA NA Mirtazapine, NA Improve Neurology. 2014 mefloquine, and IVIg

Scholten P et al. M/57 No 0 NA 240 Mirtazapine Bad Death BMJ Case Rep. 2017 (worsening)

Dur´eault et al. M/69 No 3 <200 34 Cidofovir NA Death Medicine. 2017

Gamperl I et al. Clin F/73 No 24 680 NA Steroids (pulse) Good Alive Case Rep. 2018 and mirtazapine

F/54 Steroids 72 NA NA Mirtazapine IRIS and Alive and MMF then improve

F/63 No NA NA NA No Bad Death (worsening)

Continued

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Table 2 Literature review of PML cases and sarcoidosis (continued)

Time from Cases (sex/ Therapies diagnosis of Lymphocyte CD4+ age at PML before sarcoidosis to count count Outcome References diagnosis, y) PML PML (mo) (/mm3) (/mm3) Treatment Tolerance of PML

Total 27M/14F 85,9 (0;332) 750 210 24/39 47 (24;74) (300;1620) (34;500) (61.5%)

Abbreviations: aTNFa = anti-TNFa; CS = corticosteroids; CYC = cyclophosphamide; HCQ = hydroxychloroquine; IRIS = immune reconstitution inflammatory syndrome; IVIg = IV immunoglobulin; MIU = million international unit; MMF = mycophenolate mofetil; MTX = methotrexate; NA = not available; PML = progressive multifocal encephalopathy; TNF = tumor necrosis factor.

In sarcoidosis, CD4 T-cell depletion is mainly linked to the margination process. Thus, the use of IL7 as a therapy for Appendix (continued)

PML has probably favored the expansion and activation of Name Location Contribution CD4 T cells in tissue, explaining the IRIS in involved tissue (lung and lymph nodes) and hypercalcemia. We chose to Vincent Gies, Strasbourg Performed the study and wrote, PharmD, PhD University read, and accepted the paper in introduce low dose of steroids and hydroxychloroquine to Hospital, France the final version balance the related risk of PML resurgence and immune Yannick Strasbourg Performed the study and wrote, restoration syndrome associated with granuloma activity Dieudonne, University read, and accepted the paper in (hypercalcemia and acute lung injury) and sebopsoriasis MD Hospital, France the final version fl 10 are. Cornelia Strasbourg Involved in patient care and Kuhnert, MD University collecting the data and read and Hospital, France accepted the paper in the final New targeted therapies in immunology and infectiology version open the doors of more specific and more specialized therapies for patients with immunodeficiencies, autoim- Cédric Strasbourg Involved in patient care and Lenormand, University collecting the data and read and mune diseases, or cancers. Nonetheless, some imbalance MD, PhD Hospital, France accepted the paper in the final has to be finely found to avoid some severe complications. version This is the beginning of a new era for physicians involved in Laurent Strasbourg Involved in patient care and these fields. Kremer, MD, University collecting the data and read and PhD Hospital, France accepted the paper in the final version Study funding Anne Molitor, Strasbourg Interpreted the data of whole- Supported by the European regional development fund PhD University exome sequencing and read and (European Union) INTERREG V program (project PER- Hospital, France accepted the paper in the final SONALIS) and the MSD Avenir grant (Autogen project). version Raphaël Strasbourg Interpreted the data of whole- Carapito, PhD University exome sequencing and read and Disclosure Hospital, France accepted the paper in the final The authors report no disclosures relevant to the manuscript. version Go to Neurology.org/NN for full disclosures. Yves Strasbourg Involved in patient care and Hansmann, University collecting the data and read and Publication history MD, PhD Hospital, France accepted the paper in the final version Received by Neurology: Neuroimmunology & Neuroinflammation April 27, 2020. Accepted in final form June 15, 2020. Vincent Strasbourg Involved in patient care and Poindron, MD, University collecting the data and read and PhD Hospital, France accepted the paper in the final version

Appendix Authors Thierry Strasbourg Involved in patient care and Martin, MD, University collecting the data and read and Name Location Contribution PhD Hospital, France accepted the paper in the final version Aur´elien Strasbourg Designed and performed the Guffroy, MD, University study, wrote the paper, Sandrine Strasbourg Performed the study, involved in PhD Hospital, France coordinated the care of the Hirschi, MD University patient care, and wrote, read, and patient, and read and accepted Hospital, France accepted the paper in the final the paper in the final version version

Morgane Solis, Strasbourg Performed the study, involved in Anne-Sophie Strasbourg Performed the study, involved in PharmD, PhD University patient care, and read and Korganow, University patient care, and wrote, read, and Hospital, France accepted the paper in the final MD, PhD Hospital, France accepted the paper in the final version version

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 7 References 6. Jamilloux Y, Valeyre D, Lortholary O, et al. The spectrum of opportunistic diseases – 1. Jamilloux Y, Néel A, Lecouffe-Desprets M, et al. Progressive multifocal leukoence- complicating sarcoidosis. Autoimmun Rev 2015;14:64 74. ı phalopathy in patients with sarcoidosis. Neurology 2014;82:1307–1313. 7. In the frame of GSF (Groupe Sarco¨dose France), Calender A, Rollat Farnier PA, 2. Cortese I, Muranski P, Enose-Akahata Y, et al. Pembrolizumab treatment for Buisson A, et al. Whole exome sequencing in three families segregating a pediatric case progressive multifocal leukoencephalopathy. N Engl J Med 2019;380: of sarcoidosis. BMC Med Genomics 2018;11:23. 1597–1605. 8. Zerbe CS, Marciano BE, Katial RK, et al. Progressive multifocal leukoencephalopathy fi 3. Walter O, Treiner E, Bonneville F, et al. Treatment of progressive multifocal leu- in primary immune de ciencies: stat1 gain of function and review of the literature. – koencephalopathy with nivolumab. N Engl J Med 2019;380:1674–1676. Clin Infect Dis 2016;62:986 994. 4. Sospedra M, Schippling S, Yousef S, et al. Treating progressive multifocal leu- 9. Krey L, Raab P, Sherzay R, et al. Severe progressive multifocal leukoencephalopathy fl koencephalopathy with interleukin 7 and vaccination with JC virus capsid protein (PML) and spontaneous immune reconstitution in ammatory syndrome (IRIS) in an VP1. Clin Infect Dis 2014;59:1588–1592. immunocompetent patient. Front Immunol 2019;10:1188. 5. Duréault A, Chapelon C, Biard L, et al. Severe infections in sarcoidosis: incidence, 10. Tripathi SV, Leslie KS, Maurer TA, Amerson EH. Psoriasis as a manifestation of HIV- fl predictors and long-term outcome in a cohort of 585 patients. Medicine 2017;96: related immune reconstitution in ammatory syndrome. J Am Acad Dermatol 2015; – e8846. 72:e35 e36.

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN CLINICAL/SCIENTIFIC NOTES OPEN ACCESS COVID-19 in MS Initial observations from the Paciﬁc Northwest

James D. Bowen, MD, Justine Brink, DO, MPH, Ted R. Brown, MD, MPH, Elisabeth B. Lucassen, MD, Correspondence Kyle Smoot, MD, Annette Wundes, MD, and Pavle Repovic, MD, PhD Dr. Repovic [email protected] Neurol Neuroimmunol Neuroinﬂamm 2020;7:e783. doi:10.1212/NXI.0000000000000783

Coronavirus disease 2019 (COVID-19), caused by the severe acute respiratory syndrome MORE ONLINE coronavirus-2 (SARS-CoV-2), is marked by a wide range and severity of symptoms. Most patients have mild to moderate symptoms, whereas 10%–15% require hospitalization. Mortality is higher COVID-19 Resources with cardiovascular disorders, diabetes, and older age.1 For the latest articles, invited commentaries, and Little is known about the clinical features of COVID-19 in the context of multiple sclerosis (MS), blogs from physicians where some disease-modifying therapies (DMTs) may lead to immunocompromised state. around the world Regarding the use of MS DMTs in the context of the COVID-19 pandemic, clinicians have had to NPub.org/COVID19 make inferences, based on the DMT’s mechanism of action and clinical trial data on infections, whether to continue, stop, or change the therapy in their patients with MS.2

Here, we report our initial experience with COVID-19 among 8 patients with MS (table). The earliest case in our series occurred abroad in the early days of the COVID-19 outbreak. Other infections occurred locally in Washington and Oregon. The source of infection was a close contact in 2 cases, residential nursing facility in 2 cases, travel in 1, but not identified in 3 cases. Our case series consists mostly of female patients (6 of 8), aged 35–74, and most of whom had relapsing- remitting MS (5 of 8). Most of the patients had lower disability (Expanded Disability Status Scale [EDSS] score 1–3.5), with 2 exceptions (EDSS 7.5 and 8.5). Two patients were on injectable agents (1 interferon and 1 glatiramer), 2 on dimethyl fumarate (DMF), 1 on teriflunomide (TFL), and 2 on fingolimod (FNG), whereas 1 patient was not on DMT. None had a relapse or systemic corticosteroids recently. Neither of the patients on DMF had lymphopenia at baseline. Patients with FNG had lymphopenia (0.8 and 0.6 K/ul) at presentation and 6 months before COVID-19 infection, respectively.

The most common presenting symptoms of COVID-19 in this cohort were fever (7 of 8), cough (6 of 8), and headache (4 of 8). Less common symptoms were sneezing (2), anosmia (2), fatigue (2), nausea (1), chills (1), and disequilibrium (1). One patient never developed fever. Two patient had dyspnea and altered mental status. We suspect that their altered mental status was probably due to hypoxia, although direct eﬀect of SARS-CoV-2 on CNS could not be excluded.

COVID-19 diagnosis was confirmed by RT-PCR on a sample obtained by nasopharyngeal swab between 1 and 14 days after symptom onset, reflecting the challenges with laboratory testing at the time. One patient’s diagnosis was not confirmed by RT-PCR, but her spouse, who had the same symptoms, tested positive 2 days earlier, so we believe that COVID-19 diagnosis is most likely accurate.

COVID-19 symptoms lasted 6–28 days. Three patients were hospitalized, one of them primarily for observation. Although symptomatic, 2 patients on FNG stopped taking their medication for 2 and 4 days (while febrile). The patient on interferon missed 1 dose. Patients on glatiramer acetate and DMF continued their treatment without interruption.

From the Swedish Multiple Sclerosis Center (J.D.B., P.R), Seattle, WA; Providence Multiple Sclerosis Center (J.B., E.B.L., K.S.), Portland, Oregon; EvergreenHealth Multiple Sclerosis Center (T.R.B.), Kirkland; and University of Washington Multiple Sclerosis Center (A.W.), Seattle.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

The Aricle Processing Charge was funded by the authors. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Days from symptom onset DMT MS COVID-19 to + modification/ Age Gender diagnosis EDSS DMT symptoms test (PCR) Comorbidities interruption Outcome

35 F RRMS 2.0 DMF Fever and dry cough 2 DM and asthma None Full recovery

43 M RRMS 2.5 FNG Fever, cough, 12 None Held 4 d while Full headache, febrile recovery and sneezing

46 F RRMS 2.0 IFN Fever, cough, and Not testeda None Held one dose Recovering fatigue (cough)

50 F RRMS 1.0 DMF Fever, dry cough, 8 None None Full headache, fatigue, recovery anosmia, nausea, and disequilibrium

53 F RRMS 2.0 GA Headache, 14 Obesity None Full sneezing, recovery and anosmia

55 F SPMS 7.5 TFL Fever, cough, 7 Myotonic dystrophy Stopped Fatal dyspnea, and altered mental status

56 F SPMS 3.5 FNG Fever, headache, 4 None Held 2 d while Full and chills febrile recovery

74 M SPMS 8.5 None Fever, cough, 1 CAD, HTN, DM, COPD, Not applicable Fatal dyspnea, and cardiomyopathy and altered mental status

Abbreviations: CAD = coronary artery disease; COPD = chronic obstructive pulmonary disease; DM = diabetes mellitus; DMF = dimethyl fumarate; DMT = disease-modifying therapy; FNG = fingolimod; GA = glatiramer acetate; HTN = hypertension; IFN = interferon; RRMS = relapsing-remitting MS; SPMS = secondary progressive MS; TFL = teriflunomide. a Partner had the same symptoms as the patient confirmed COVID-19 2 days before patient’s symptom onset.

Two patients died. Both of these patients were severely af- general (non-MS) population. At the same time, the fatal fected by both MS (EDSS 7.5 17 and 8.5) and COVID-19 outcome in our most disabled patients portends significant (hypoxia,fever,andalteredmentalstatus),inadditionto risks for patients with advanced MS. Most of our patients having significant comorbidities. On admission, both remained on their DMTs with no interruption during the patients had low absolute lymphocyte counts (0.6 and 0.58 COVID-19 infection. However, the generalizability of this K/uL) and one had increased liver function tests (AST 93 finding is limited because none of these were cell-depleting and ALT 66), probably because of the COVID-19 infection1 therapies, and most infections were mild. We hope that because laboratory test results were in the normal range 10 larger studies will provide more definitive information on months before for the TFL-treated patient. Both patients additional risks associated with MS DMTs in COVID-19 were placed on supplemental oxygen, but continued to de- and hospitalization outcomes to better inform our care for teriorate. As per advanced directive of both patients, they this population. were not intubated, and they died 3 and 4 days after the admission, respectively. Autopsy was not performed. Study funding No targeted funding reported. The full scope of COVID-19 manifestations in the MS population remains to be defined. To that end, we en- Disclosure courage all clinicians to follow our example and report their J.D. Bowen, J. Brink, T. Brown, E.B. Lucassen, K. Smoot, cases of COVID-19 in MS and related disorders in North A. Wundes, and P. Repovic report no financial disclosures rel- America (covims.org) and elsewhere (msdataalliance. evant to the topic of this manuscript. Go to Neurology.org/NN com). In publishing this initial report, we wanted to share for full disclosures. our experiences and observations among patients from a region with early community spread of SARS-CoV-2 in Publication history 3 the United States. We were relieved that most of these Received by Neurology: Neuroimmunology & Neuroinflammation infections were mild and in line with observations in April 14, 2020. Accepted in final form May 7, 2020.

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Appendix Authors Appendix (continued)

Name Location Contribution Name Location Contribution

Pavle Swedish MS Center, Designed and conceptualized Kyle Smoot, Providence MS Major role in the acquisition Repovic, Seattle, WA study, analyzed the data, and MD Center, Portland, of data and revised the MD, PhD drafted the manuscript for Oregon manuscript for intellectual intellectual content content

James D. Swedish MS Center, Major role in the acquisition of Annette University of Major role in the acquisition Bowen, MD Seattle, WA data and revised the manuscript Wundes, MD Washington, of data and revised the for intellectual content Seattle manuscript for intellectual content Justine Providence MS Major role in the acquisition of Brink, DO, Center, Portland, data and revised the manuscript MPH Oregon for intellectual content

Ted Brown, Evergreen Major role in the acquisition of References MD, MPH Healthcare, data and revised the manuscript 1. Guan WJ, Ni ZY, Hu Y, et al. Clinical characteristics of coronavirus disease 2019 in Kirkland, WA for intellectual content China. N Engl J Med 2020;382:1708–1720. 2. Giovannoni G, Hawkes C, Lechner-Scott J, Levy M, Waubant E, Gold J. The COVID- Elisabeth B. Providence MS Major role in the acquisition of 19 pandemic and the use of MS disease-modifying therapies. Mult Scler Relat Disord Lucassen, Center, Portland, data and revised the manuscript 2020;39:102073. MD Oregon for intellectual content 3. Bhatraju PK, Ghassemieh BJ, Nichols M, et al. Covid-19 in critically ill patients in the Seattle region—case series. N Engl J Med 2020;382:2012–2022.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 CLINICAL/SCIENTIFIC NOTES OPEN ACCESS Guillain-Barr´e syndrome related to SARS-CoV-2 infection

K´evin Bigaut, MD,* Martial Mallaret, MD, PhD,* Seyyid Baloglu, MD, Benjamin Nemoz, MD, Correspondence Patrice Morand, MD, PhD, Florent Baicry, MD, Alexandre Godon, MD, Paul Voulleminot, MD, Dr. Bigaut [email protected] Laurent Kremer, MD, PhD, Jean-Baptiste Chanson, MD, PhD, and J´eromeˆ de Seze, MD, PhD

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e785. doi:10.1212/NXI.0000000000000785

The outbreak of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) responsible for MORE ONLINE coronavirus disease 2019 (COVID-19) led to the death of thousands of people around the world.1 Neurologic manifestations are not much specific apart from acute anosmia, and post- COVID-19 Resources infectious manifestation data are missing.2 We described the cases of 2 patients exhibiting de- For the latest articles, myelinating form of Guillain-Barré syndrome (GBS) and summarized neurologic manifestations invited commentaries, and and investigations results in table 1. blogs from physicians around the world A 43-year-old man presented with cough, asthenia, and myalgia in legs, followed by acute anosmia NPub.org/COVID19 and ageusia with diarrhea the next day. Symptoms resolved spontaneously after 2 weeks. Twenty-one days after the beginning of respiratory symptoms, he presented with in a rapidly progressive manner paraesthesia, hypoesthesia, and distal weakness in the lower limbs. In the following 2 days, these symptoms extended to the midthigh and tip of the fingers associated with ataxia, and he was hospitalized at day 4 because a right peripheral facial palsy had occurred. His body temperature was 36.9°C and oxygen saturation was 99%. Neurologic examination disclosed decreased light touch from midthigh to feet and the tip of the fingers; decreased vibration sense in the lower limbs, symmetric weakness for dorsiflexion and extension of the toes (Medical Research Council [MRC] score = 3/5) and flexionofthethigh(MRC=4/5);andareflexia in the forelimbs apart from the left biceps reflex.

Laboratory results at day 4 were unremarkable (normal blood cell count, negative C-reactive protein, negatives HIV, Lyme, and syphilis serologies). Antigangliosides antibodies were negatives. Nasopharyngeal swab test was positive for SARS-CoV-2 on reverse transcription- polymerase chain reaction (RT-PCR) assay. CT of the chest showed ground-glass opacities in 10–25% on both lungs (figure e-1, links.lww.com/NXI/A267). CSF results showed normal cell count (1 × 106/L), increased protein level (0.94 g/L), and negative SARS-CoV-2 on RT-PCR assay. MRI at day 7 showed multiple cranial neuritis (in nerves III, V, VI, VII, and VIII), radiculitis, and plexitis on both the brachial and lumbar plexus (figure e-2, links.lww.com/NXI/A267). Nerve conduction studies at day 9 showed 2 conduction blocks (>50%) in both peroneal nerves, decreased motor conduction velocities in both peroneal and tibial nerves approximately 30–37 m/s, a sural sparing pattern, abolition of the H-reflex, and slightly increased of F-wave latencies supporting demyelinating pattern (table e-1, links.lww.com/NXI/A267). The patient was diagnosed with GBS, and IV immunoglobulin infusions (IVIg) were started at day 5 (2 g/kg). He was rapidly discharged home with progressive improvement.

An obese 70-year-old woman presented with anosmia and ageusia, followed by diarrhea for 2 days. She complained of mild asthenia and myalgia without fever. All symptoms resolved excepted anosmia and ageusia. Nasopharyngeal swab test was positive for SARS-CoV-2 on RT-PCR assay. Seven days later, she presented with acute proximal tetraparesis and distal forelimb and perioral

*These authors contributed equally to the manuscript.

From the Service de Neurologie (K.B., P.V., L.K., J.-B.C., J.S.), Hopitauxˆ Universitaires de Strasbourg; Service de Neurologie (M.M.), Centre Hospitalo-Universitaire de Grenoble Alpes, La Tronche; Service de Neuroradiologie (S.B.), Hopitauxˆ Universitaires de Strasbourg; Institut de Biologie Structurale (IBS) (B.N., P.M.), Universit´e de Grenoble Alpes, CEA, CNRS; Laboratoire de virologie (B.N., P.M.), Centre Hospitalo-Universitaire de Grenoble Alpes, La Tronche; Service d’Accueil des Urgences (F.B.), Hopitauxˆ Universitaires de Strasbourg; and Service de R´eanimation Polyvalente Chirurgicale (A.G.), Centre Hospitalo-Universitaire de Grenoble Alpes, La Tronche, France.

The Article Processing Charge was funded by the authors.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Table 1 Clinical characteristics and investigations result for 2 patients with GBS related to SARS-CoV-2 infection

Days between the onset of Nerve Treatments COVID-19 GBS symptoms conduction and Patient and GBS and signs CSF findings studies MRI results evolution

1. Man 21 d Flaccid paraparesis, Day 4; protein Day 9; demyelinating Day 7; radiculitis and Day 5; IVIg 48 yo generalized areflexia, level: 0.95 g/L; cell pattern with motor plexitis on both brachial (2 g/kg); lower-limb and distal count: 1 × 106/L; decreased conduction and lumbar plexus; progressive upper-limb paresthesia, negative SARS- velocities and multiple cranial neuritis (in improvement ataxia; right facial palsy CoV-2 RT-PCR conduction blocks nerves III, VI, VII, and VIII) (day 4)

2. 10 d Flaccid tetraparesis, Day 3; protein Day 7; demyelinating Not performed Day 4; IVIg Woman generalized areflexia, level: 1.6 g/L; cell pattern with motor and (2 g/kg); slow 70 yo forelimb paresthesia; count: 6 × 106/L; sensitive decreased progressive respiratory failure (day 3) negative SARS- conduction velocities improvement CoV-2 on RT-PCR and conduction blocks

Abbreviations: COVID-19 = coronavirus disease 2019; GBS = Guillain-Barr´e syndrome; IVIg = IV immunoglobulin infusions; SARS-CoV-2 = severe acute respiratory syndrome coronavirus 2.

paraesthesia. She was hospitalized for dyspnea and loss of am- syndrome coronavirus infection in 4 patients with Bick- bulation 3 days later and was rapidly transferred to an intensive erstaff’s encephalitis overlapping with GBS.5 care unit for noninvasive ventilation for acute respiratory failure with hypercapnia. She was discharged from the intensive care unit Thus, our cases add to several other reported cases and 9 days later, without requiring invasive mechanical ventilation. strengthen the view that GBS occurs with COVID-19. Neurologic examination disclosed proximal lower-limb weakness (MRC 2/5), distal weakness (MRC 4/5), and diffuse areflexia. Study funding No targeted funding reported. At admission, C-reactive protein was slightly increased at 22 mg/L. Antigangliosides antibodies were negative. CSF results Disclosure showed subnormal cell count (6 × 106/L), increased protein K. Bigaut, M. Mallaret, S. Baloglu, B. Nemoz, P. Morand, level (1.06 g/L), and negative SARS-CoV-2 on RT-PCR assay. F. Baicry, A. Godon, P. Voulleminot, L. Kremer, J.-B. Chan- CT of the chest showed moderate ground-glass opacities in son, and J. de Seze report no disclosures relevant to the both lungs (figure e-1, links.lww.com/NXI/A267). Nerve manuscript. Go to Neurology.org/NN for full disclosures. conductions studies at day 7 showed a typical demyelinating Publication history pattern with a conduction block in the left median nerve, Received by Neurology: Neuroimmunology & Neuroinflammation temporal dispersion, upper limb increased motor distal laten- April 29, 2020. Accepted in final form May 6, 2020. cies, diffuse decreased motor and sensory conduction velocities lower than 38 m/s in 9 nerves of 10 tested (table e-1, links.lww. com/NXI/A267), and neurogenic pattern on EMG. IVIg (2 g/ kg) were started at day 4 after the onset of the first neurologic Appendix Authors symptoms. Left peripheral facial palsy occurred in a delayed Name Location Contribution manner at day 9. Her clinical condition improved slowly with Kévin Bigaut, Service de Neurologie, Conception, organization physiotherapy, needing a transfer in a rehabilitation center. MD Hopitauxˆ Universitaires de and execution of the Strasbourg, France research project, and writing of the first draft We reported here 2 cases of GBS related to SARS-CoV-2 and the review and infection with neurologic improvement on IVIg, adding to critique of the manuscript

few cases of GBS, one case of Miller Fisher syndrome, and one Martial Service de Neurologie, Conception, organization case of polyneuritis cranialis already published. Mallaret, MD, Centre Hospitalo- and execution of the PhD Universitaire de Grenoble research project, and Alpes, France writing of the first draft The first case report described a patient with GBS whose and the review and symptoms began 7 days before COVID-19, which asks the critique of the manuscript 3 question of parainfectious profile or coincidence. However, Seyyid Service de Conception, organization Baloglu, MD Neuroradiologie, and execution of the previous reports and our cases suggest that GBS associated Hopitauxˆ Universitaires de research project, and with SARS-CoV-2 infection could start between 5 and 21 days Strasbourg, France writing of the first draft after the SARS-CoV-2 clinical symptoms.4 It could follow and the review and critique of the manuscript a postinfectious profile as reported on Middle East respiratory

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Appendix (continued) Appendix (continued)

Name Location Contribution Name Location Contribution

Benjamin Institut de Biologie Conception and Laurent Service de Neurologie, Conception and Nemoz, MD Structurale (IBS), organization of Kremer, MD, Hopitauxˆ Universitaires de organization of the Universit´e de Grenoble the research PhD Strasbourg, France research project and Alpes, CEA, CNRS, France project and review review and critique of the and critique of the manuscript manuscript Jean-Baptiste Service de Neurologie, Conception, organization Patrice Institut de Biologie Conception and Chanson, Hopitauxˆ Universitaires de and execution of the Morand, MD, Structurale (IBS), organization of the MD, PhD Strasbourg, France research project, and PhD Universit´e de Grenoble research project and writing of the first draft Alpes, CEA, CNRS, France review and critique of the and the review and manuscript critique of the manuscript

Florent Service d’Accueil des Conception and J´eromeˆ de Service de Neurologie, Conception, organization Baicry, MD Urgences, Hopitauxˆ organization of Seze, MD, Hopitauxˆ Universitaires de and execution of the Universitaires de the research project PhD Strasbourg, France research project, and Strasbourg, France and review and writing of the first draft critique of the and the review and manuscript critique of the manuscript

Alexandre Service de Réanimation Conception and Godon, MD Chirurgicale, Centre organization of Hospitalo-Universitaire de the research project References Grenoble Alpes, France and review and 1. Guan W, Ni Z, Hu Y, et al. Clinical characteristics of coronavirus disease 2019 in critique of the China. N Engl J Med 2020;382:1708–1720. manuscript 2. Mao L, Jin H, Wang M, et al. Neurologic manifestations of hospitalized patients with coronavirus disease 2019 in Wuhan, China. JAMA Neurol Epub 2020 April 10. doi: Paul Service de Neurologie, Conception and 10.1001/jamaneurol.2020.1127. Voulleminot, Hopitauxˆ Universitaires de organization of 3. Zhao H, Shen D, Zhou H, et al. Guillain-Barré syndrome associated with SARS-CoV-2 MD Strasbourg, France the research project infection: causality or coincidence? Lancet Neurol 2020;19:383–384. and review and 4. Toscano G, Palmerini F, Ravaglia S, et al. Guillain–Barré syndrome associated with critique of the SARS-CoV-2. N Engl J Med Epub 2020 Apr 17. doi: 10.1056/NEJMc2009191. manuscript 5. Kim JE, Heo JH, Kim HO, et al. Neurological complications during treatment of Middle East respiratory syndrome. J Clin Neurol 2017;13:227–233.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 CLINICAL/SCIENTIFIC NOTES OPEN ACCESS A case of cerebral vasculitis due to neurobartonellosis

Meryim Poursheykhi, MD,* Farhan Mithani, BS,* Tanu Garg, MD,* Christian Cajavilca, MD, Siraya Jaijakul, MD, Correspondence Steve Fung, MD, Richard Klucznik, MD, and Rajan Gadhia, MD Dr. Poursheykhi mpoursheykhi@ Neurol Neuroimmunol Neuroinﬂamm 2020;7:e791. doi:10.1212/NXI.0000000000000791 houstonmethodist.org

We report a case of a 60-year-old right-handed woman with hypertension, hyperlipidemia, and hypothyroidism who presented with a three-week history of recurrent thunderclap headaches accompanied by photophobia, phonophobia, nausea, and vomiting. She reported one brief episode of slurred speech, expressive aphasia, right facial droop, and right hemiparesis suggestive of a TIA. Family history was remarkable for primary angiitis of the CNS (PACNS) in the mother. Neurologic examination was unremarkable. CT of the head was negative; CT angiography (CTA) of the head and neck suggested fibromuscular dysplasia in bilateral cervical internal carotid arteries and distal right vertebral artery. MRI of the brain showed no correlating abnormalities. A digital subtraction angiography (DSA) revealed multivessel intracranial medium and large vessel narrowing and fusiform dilatations, suggestive of reversible cerebral vasoconstriction syndrome (RCVS) vs vasculitis. Subsequent MR intracranial vessel wall imaging (IVWI) showed multifocal concentric vessel wall thickening and enhancement consistent with vasculitis (figure). Transcranial Doppler showed no evidence of elevated intracranial velocities. CSF studies were unremarkable with an opening pressure of 10 cm H2O, 2 white blood cells (normal 0–5/mm3), 2 red blood cells (normal 0–1/mm3), 58 mg/dL glucose (normal 40–70, serum glucose 87), 41 mg/dL protein (normal 15–45), normal Q-albumin ratio, normal IgG synthetic rate, and IgG index. Serum inflammatory and infectious studies had been negative thus far. Empiric high-dose IV steroids lead to complete symptom resolution. Final infectious workup revealed strongly positive serum Bartonella IgM titer of 1:256 and negative IgG, consistent with her reported cat exposure. She was started on an outpatient two- week course of doxycycline, rifampin, and oral steroids. Four weeks later, repeat vessel wall MRI and Bartonella serologies (IgM titer 1:80) showed improvement.

Discussion We present an individual with symptoms initially concerning for RCVS vs vasculitis who was subsequently found to have secondary CNS vasculitis due to cat-scratch disease (CSD). To our knowledge, this is the ﬁrst adult case of Bartonella henselae-associated CNS vasculitis, particularly without encephalopathy as the presenting symptom.

CSD typically presents with self-limited regional lymphadenopathy and fever.1 Neurologic complications are rare, occurring in 2% of cases with encephalopathy as the most common manifestation.2 Neuroretinitis, seizures, coma, myelopathies, and cranial and peripheral nerve involvement have also been reported. CNS vasculitis associated with CSD, however, has only been reported in 2 pediatric cases which presented with strokes.3,4

Diagnostically, identifying primary and secondary CNS vasculitis can be challenging both clinically and radiographically. No speciﬁc studies in serum or CSF are available for the diagnosis of CNS

*These authors contributed equally to the manuscript.

From the Department of Neurology (M.P., T.G., R.G.), Department of Infectious Diseases (S.J.), Department of Neuroradiology (S.F.), and Department of Interventional Neurora- diology (R.K.), Houston Methodist Hospital, TX; Texas A&M Health Science Center College of Medicine (F.M.), Bryan; and Department of Interventional Neurology (C.C.), Texas Stroke Institute, Plano.

The Article Processing Charge was funded by the authors.

(A) Angiogram of the left internal carotid artery showing multifocal narrowing and fusiform dilatations (arrows) pretreatment. (B) Intracranial vessel wall MRI showing multifocal concentric vessel wall thickening and enhancement (arrows) pretreatment. (C) Intracranial vessel wall MRI showing reduction in vessel wall enhancement (arrows) posttreatment. (D) Magnetic resonance angiography (MRA) head showing multifocal stenoses (arrows) pretreatment. (E) MRA head showing improvement of stenoses (arrows) posttreatment.

vasculitis. As in neurobartonellosis, CSF may be unremarkable including CNS vasculitis.1 We recommend concomitant or reveal nonspecific mild lymphocytic pleocytosis. Cerebral treatment of the infection with antibiotics and secondary vasculopathies can present with similar luminal patterns, and vasculitis with high-dose steroids. Our patient received therefore, imaging modalities such as DSA, magnetic resonance a 2-week combination of doxycycline 100 mg and rifampin angiography (MRA), and CTA provide nonspecificresults 300 mg twice daily per current expert opinion.6 In addition, leading to difficulties identifying and differentiating between we initiated 5 days of high-dose IV steroids, followed by a 1- common etiologies of intracranial disease including vasospasm, week oral steroid taper. To avoid recurrent invasive testing, atherosclerosis, and inflammation. Although DSA remains the we repeated IVWI 4 weeks later for treatment monitoring and gold standard for vessel imaging, it is an invasive study that found significant reduction in vessel wall enhancement provides information limited to the vessel lumen. Conversely, (figure). IVWI allows direct visualization of the vessel wall by subtracting the signal of blood in the vessel lumen and has shown to im- Our case reiterates the importance of ruling out rare causes of prove diagnostic specificity.5 In CNS vasculitis, IVWI shows CNS vasculitis including assessing animal exposure before multifocal concentric vessel wall enhancement and thickening diagnosing PACNS. Detection of the etiology of vasculitis is as seen in our patient. In RCVS, vessel wall thickening may be essential to guide treatment and for prognostication. Non- present but with minimal or no enhancement.5 invasive imaging such as an IVWI provides valuable diagnostic information and can be useful in assessing the treatment re- At this time, there is no clear evidence-based treatment regi- sponse over time by minimizing the need for repeat in- men or duration for neurologic manifestations of CSD vasive DSA.

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Acknowledgment The authors thank Dr. Gadhia for his mentorship. Appendix (continued)

Name Location Contribution Study funding No targeted funding reported. Christian Texas Stroke Interpreted data Cajavilca, MD Institute, Plano and revised the manuscript for Disclosure intellectual content M. Poursheykhi, F. Mithani, T. Garg, C. Cajavilca, S. Jaijakul, S. Fung, R. Klucznik, and R. Gadhia report no disclosures. Go Siraya Houston Methodist Interpreted data Jaijakul, MD Hospital, TX and revised the to Neurology.org/NN for full disclosures. manuscript for intellectual Publication history content Received by Neurology: Neuroimmunology & Neuroinﬂammation Steve Fung, Houston Methodist Interpreted data and April 19, 2020. Accepted in ﬁnal form May 7, 2020. MD Hospital, TX revised the manuscript for intellectual content

Richard Houston Methodist Interpreted data and Klucznik, MD Hospital, TX revised the manuscript Appendix Authors for intellectual content

Name Location Contribution Rajan Gadhia, Houston Methodist Interpreted data and MD Hospital, TX revised the manuscript Meryim Houston Methodist Conceptualized the study, for intellectual content Poursheykhi, Hospital, TX conducted literature MD review, interpreted the data, and drafted and revised the manuscript for References intellectual content 1. Breitschwerdt E, Sontakke S, Hopkins S. Neurological manifestations of bartonellosis in immunocompetent patients: a composite of reports from 2005–2012. J Neuro- Farhan Texas A&M Health Conceptualized the study, parasitol 2012;3:1–15. Mithani, BS Science Center conducted literature 2. Carithers HA, Margileth AM. Cat-scratch disease: acute encephalopathy and other College of Medicine, review, interpreted the neurologic manifestations. Am J Dis Child 1991;145:98–101. Bryan data, and drafted and 3. Selby G, Walker GL. Cerebral arteritis in cat-scratch disease. Neurology 1979;29: revised the manuscript for 1413–1418. intellectual content 4. Balakrishnan N, Ericson M, Maggi R, Breitschwerdt EB. Vasculitis, cerebral infarction and persistent Bartonella henselae infection in a child. Parasit Vectors Tanu Garg, Houston Methodist Conceptualized the study, 2016;9:254. MD Hospital, TX conducted literature 5. Mossa-Basha M, Alexander M, Gaddikeri S, Yuan C, Gandhi D. Vessel wall review, interpreted the imaging for intracranial vascular disease evaluation. J Neurointerv Surg 2016;8: data; drafted and revised 1154–1159. the manuscript for 6. Rolain JM, Brouqui P, Koehler JE, Maguina C, Dolan MJ, Raoult D. Recom- intellectual content mendations for treatment of human infections caused by Bartonella species. Anti- microb Agents Chemother 2004;48:1921–1933.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 CLINICAL/SCIENTIFIC NOTES OPEN ACCESS Therarecaseofa20-year-oldmalewithrapidly progressive primary angiitis of the CNS with a good outcome

Edith Graham, MD, Thomas Shoemaker, MD, Dusan Stefoski, MD, Marinos Kontzialis, MD, Correspondence Anam Naumaan, MD, and Rajeev K. Garg, MD, MS Dr. Graham [email protected] Neurol Neuroimmunol Neuroinﬂamm 2020;7:e793. doi:10.1212/NXI.0000000000000793

Rapidly progressive primary angiitis of the CNS (PACNS) is recognized as a clinical subtype of PACNS that is almost uniformly fatal. PACNS is usually characterized by subacute progression over several months. The most common symptoms are headaches and cognitive dysfunction, followed by focal neurologic deﬁcits.1 Compared with this more indolent presentation, patients with rapidly progressive PACNS are more likely to present in coma with quadriparesis and MR imaging notable for bilateral infarcts.2 Various treatment combinations including high-dose corticosteroids, cyclophosphamide, azathioprine, methotrexate, rituximab, and plasma exchange have been described1,3; however, no good outcomes have been reported to date. The largest case series of 11 patients had only 1 survivor who at 14 months had a modiﬁed Rankin Scale (mRS) of 5 (bedridden, requiring constant nursing care).2 One other survivor of rapidly progressive PACNS had an mRS of 5 at the 3-month follow-up.4 We describe the treatment and outcome of a patient with rapidly progressive PACNS who returned to his premorbid baseline with an mRS of 0.

Case description A 20-year-old previously healthy man presented to a local hospital in acute coma. He had become progressively somnolent over the previous 8 hours. He had no prodromal illness, drug use, travel, or toxic exposures. His mother died suddenly in her forties of unknown causes. On arrival, he was unresponsive with pinpoint pupils and decorticate posturing to painful stimuli. Initial noncontrast head CT showed diﬀuse cerebral edema with hypodensities of the brainstem and thalami (ﬁgure 1, A–C). He was intubated for airway protection and transferred to our institution.

Noncontrast brain MRI 3 hours after admission revealed multifocal T2 FLAIR hyperintensities and effacement of the third and fourth ventricles (figure 1, D–F). Broad-spectrum antibiotics and dexamethasone were started for possible meningitis. Ten hours later, contrasted MRI showed worsening edema, new T2 FLAIR hyperintensities, and diffuse perivascular enhancement, suggestive of cerebral vasculitis (figure 1, G–J).

An external ventricular drain was placed for intracranial pressure monitoring and CSF sampling. This demonstrated 42 white blood cells (88% polymorphonuclear leukocytes), 8,000 red blood cells, glucose of 108, and protein of 285. Infectious studies were negative. The patient remained afebrile. On hospital day 1, he was started on methylprednisolone 1 g daily and plasmapheresis every other day for 3 treatments. Urgent brain biopsy on hospital day 2 revealed small vessel necrotizing cerebral vasculitis (ﬁgure 1, K and J). Rheumatologic evaluation was negative for systemic in- ﬂammatory disease. He was given cyclophosphamide 1,000 mg/m2 on days 6–8. Hospital course was complicated by ventilator-associated pneumonia, thrombocytopenia, and cerebral salt wasting. The patient also developed malignant intracranial hypertension requiring treatment with

From the Department of Neurological Sciences (E.G., T.S., D.S., R.K.G.), Department of Radiology (M.K.) and Department of Pathology (A.N.), Rush University Medical Center, Chicago, IL.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

The article processing charge was funded by the authors. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Initial head CT noncontrast showed diffuse cerebral edema, effacement of the ambient cistern, and hypodensities of the pons, midbrain, and thalami (A–C). Noncontrast brain MRI 5 hours later with multifocal, supratentorial, and infratentorial T2 FLAIR hyperintensities and diffuse cerebral edema (D–F). Ten hours later, contrasted brain MRI showed worsening cerebral edema, new T2 FLAIR hyperintensities (G–I), and diffuse perivascular enhancement suggestive of cerebral vasculitis (J). Brain tissue contains small vessel perivascular dense lymphohistiocytic infiltrate (K). Higher power view of the vessel wall shows fibrinoid necrosis and a mixed inflammatory infiltrate consisting of neutrophils, lymphocytes, and histiocytes (L).

hyperosmolar therapy on days 1 through 22, pentobarbital and quadriparesis in less than 24 hours and imaging evidence of methohexital coma on days 3–17, and hypothermia protocol on multiple bilateral cerebral infarctions. Future research into days 4–10 (figure e-1, links.lww.com/NXI/A269). optimal diagnostic and therapeutic strategies may lead to more patients with a favorable outcome in this disease. Repeat contrasted MRI brain on day 22 showed decreasing T2 FLAIR hyperintensities and resolution of perivascular en- Study funding hancement. He required tracheostomy and percutaneous gas- No targeted funding. trostomy. He received 1 dose of rituximab 1,000 mg on hospital day 30 and was discharged to a long-term acute care facility on Disclosure day 31. On discharge, he was awake and texting on his cell E. Graham reports no financial disclosures. T. Shoemaker phone. He returned to work at 5 months and remains with an receives royalties from Genentech. D. Stefoski reports no fi- mRS of 0 at 14 months. nancial disclosures. M. Kontzialis receives royalties from Elsevier. A. Naumaan and R. K. Garg report no financial dis- Discussion closures. Go to Neurology.org/NN for full disclosures. fi To our knowledge, this is the rst case described in the lit- Publication history erature of a patient with rapidly progressive PACNS who Received by Neurology: Neuroimmunology & Neuroinflammation survived with good outcome. Our report demonstrates that April 22, 2020. Accepted in final form May 12, 2020. prompt diagnosis, aggressive neurocritical intervention, and early immunosuppressant therapy can lead to a good outcome in rapidly progressive PACNS. Appendix Authors

The terminology to describe rapidly progressive PACNS is not Author Institution Contribution consistent in the literature. Terms such as “fulminant” and “cat- Edith Rush University Medical Primary author astrophic” are used interchangeably to describe this subset of Graham, MD Center, Chicago, IL PACNS. “Necrotizing” has also been used, but has been reserved 5 ’ Thomas Rush University Medical Contributed to case mainly to describe the histopathology. Our patient s biopsy Shoemaker, Center, Chicago, IL management and reviewed showed a necrotizing pattern. Some studies have suggested that MD the article

necrotizing and granulomatous patterns are associated with worse Dusan Rush University Medical Contributed to case 2 outcomes ; however, this is controversial because these patterns Stefoski, MD Center, Chicago, IL management and reviewed may also be seen in less aggressive PACNS.6 A lymphocytic the article 3,5 pattern has been associated with a better prognosis in PACNS ; Marinos Rush University Medical Reviewed the imaging and however, there has also been a case of rapidly progressive PACNS Kontzialis, MS Center, Chicago, IL article 7 with lymphocytic pattern. Therefore, it is diﬃcult to draw con- Anam Rush University Medical Obtained pathology images clusions on prognosis based on histopathology alone. Naumaan, Center, Chicago, IL and reviewed the article MD

We recommend a high index of suspicion for rapidly pro- Rajeev K Rush University Medical Project advisor and edited Garg, MD, MS Center, Chicago, IL the manuscript gressive PACNS in patients with progression to coma and

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN References 4. Safouris A, Stricker J, Michotte A, Voumvourakis K, Gazagnes MD, Tsivgoulis G. Biopsy- 1. Beuker C, Schmidt A, Strunk D, et al. Primary angiitis of the central nervous system: proven fulminant primary angiitis of the central nervous system with normal arteriography: – diagnosis and treatment. Ther Adv Neurol Disord 2018;11:1–16. a challenging diagnosis of recurrent ischemic strokes. Neurol Sci 2014;35:135 137. 2. Salvarani C, Brown RD Jr, Calamia KT, et al. Rapidly progressive primary central 5. Salvarani C, Brown RD Jr, Christianson T, et al. An update of the Mayo Clinic cohort nervous system vasculitis. Rheumatology (Oxford) 2011;50:349–358. of patients with adult primary central nervous system vasculitis: description of 163 3. de Boysson H, Zuber M, Naggara O, et al; French Vasculitis Study Group and the patients. Medicine (Baltimore) 2015;94:e738. French NeuroVascular Society. Primary angiitis of the central nervous system: de- 6. Sundaram S, Menon D, Khatri P, et al. Primary angiitis of the central nervous system: scription of the first fifty-two adults enrolled in the French cohort of patients with clinical profiles and outcomes of 45 patients. Neurol India 2019;67:105–112. primary vasculitis of the central nervous system. Arthritis Rheumatol 2014;66: 7. Spence S, Ng D, Casault C. Atypical presentation of fulminant primary central ner- 1315–1326. vous system angiitis. J Neuroimmunol 2019;330:1–4.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 CLINICAL/SCIENTIFIC NOTES OPEN ACCESS Acute disseminated encephalomyelitis after SARS-CoV-2 infection

Giovanni Novi, MD, Tommaso Rossi, MD, Enrico Pedemonte, MD, Laura Saitta, MD, Claudia Rolla, MD, Correspondence Luca Roccatagliata, MD, PhD, Matilde Inglese, MD, PhD, and Daniele Farinini, MD Dr. Novi [email protected] Neurol Neuroimmunol Neuroinﬂamm 2020;7:e797. doi:10.1212/NXI.0000000000000797

Acute disseminated encephalomyelitis (ADEM) is a rare autoimmune disease of the CNS that MORE ONLINE often after viral infections and mainly affecting children. ADEM is characterized by the onset of multifocal neurologic symptoms, encephalopathy, with brain MRI showing demyelinating COVID-19 Resources abnormalities in the acute phase.1 Coronavirus disease 2019 (COVID-19) is a novel entity For the latest articles, caused by the pandemic severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2), invited commentaries, and which is characterized by influenza-like symptoms, pneumonia, and in severe cases respiratory blogs from physicians insufficiency.2 Many neurologic complications occurring in patients with COVID-19 have been around the world described,3 and it has been hypothesized that, in some cases, SARS-CoV-2 might exhibit NPub.org/COVID19 a neurotropic behavior.4

Results We report on a 64-year-old woman with a history of vitiligo, hypertension, and monoclonal gammopathy of undetermined significance who developed an influenza-like syndrome in mid- March 2020 that lasted for 2 weeks. The patient also reported the development of smell and taste deficit that rapidly reached anosmia and ageusia. Starting from April 10, the patient developed bilateral vision impairment associated with sensory deficit on her right leg. She reached the emergency department of our hospital and was referred for ophthalmologic evaluation. Visual acuity was hand motion bilaterally, and relative afferent pupillary defect was detected. Ocular motility and fundus examination were unremarkable, whereas visual field test showed profound defects bilaterally. Neurologic examination, performed on hospitalization, showed mild behavioral abnormalities (irritability), headache, bilateral relative afferent pupillary defect, ageusia and anosmia, severe visual loss, right abdominal sensory level, and left-sided lower limb hyper-reflexia with the Babinski sign. The patient underwent a brain and spine MRI scan (figure), with evidence of multiple T1 post-Gd enhancing lesions of the brain, associated with a single spinal cord lesion at the T8 level and with bilateral optic nerve enhancement. A lumbar puncture was performed and showed lymphocytic pleocytosis with 22 cells/mm3 (reference range: 0–5 cells/mm3), mainly represented by CD3+CD4+ T-cells, with mild hyperproteinorrachia (452 mg/L, reference range: 150–450 mg/L), and identical immunoglobulin G oligoclonal bands were present in the CSF and serum (mirror pattern). PCR for SARS-CoV-2 tested negative on nasal swab and positive on CSF sample. The patient’s serum tested positive for anti-SARS-CoV-2 immunoglobulin G and negative for antiaquaporin-4 (AQ4) antibody (ab) and antimyelin oligodendrocyte glycoprotein (MOG) ab; the levels of interleukins were not assessed in serum and CSF.

ADEM disease was suspected, and high-dose steroids (IV methylprednisolone 1 g/d for 5 days tapered with oral prednisone 75 mg/d) associated with IV immunoglobulins (2 g/kg in 5 days) were administered. The patient reported signiﬁcant improvement in visual symptoms and progressive recovery of visual acuity. After 14 days of treatment, vision was 20/30 in the

From the Department of Neuroscience (G.N., E.P., D.F.), Ospedale Policlinico San Martino, IRCCS; Ospedale Policlinico San Martino (T.R., L.R., M.I.), IRCCS; Department of Neuro- radiology (L.S., C.R.), Ospedale Policlinico San Martino, IRCCS, Genova, Italy; Department of Health Sciences (DISSAL) (L.R.), University of Genova; and Department of Neuroscience (M.I.), Rehabilitation, Ophthalmology, Genetics, Maternal and Child Health (DINOGMI) and Center of Excellence for Biomedical Research (CEBR), University of Genova, Italy.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

(A–C) Postgadolinium (Gd) T1-weighted (T1w) sequence of the brain in the axial plane showing 6 enhancing lesions (green arrows), most of which with ring enhancement and some of which with nodular enhancement. Incomplete ring enhancement is shown about the right temporal lesion (dashed green arrow). (D–F) T2-weighted (T2w) fat saturated sequence in the coronal plane with evidence of (D) hyperintense signal of the optic nerves bilaterally (white arrows), which is best seen when compared with the normal signal within (E) the optic chiasm and (F) the optic tracts (dashed white arrows). (G) T2w sequence of the thoracic spine in the sagittal plane showing a hyperintense spindle-like T8 lesion, involving less than 2 metameric levels (dashed yellow arrow). (H) Post-Gd T1w sequence of the spine in the axial plane at the level of T8 showing eccentric areas of enhancement respectively located poste- riorly to the right and anteriorly to the left (dashed yellow arrow).

right eye and 20/25 in the left eye. Visual-evoked potential the development of neurologic symptoms do not support this showed increased latency in both eyes (p100 wave latency: hypothesis, favoring ADEM diagnosis. 114 ms right eye, 120 ms left eye, reference: <100 ms). A follow-up brain MRI scan also showed a partial improvement Since COVID-19 is currently a pandemic disease, neurolo- with a reduction in the number of Gd-enhancing lesions. On gists should be aware that autoimmune neurologic compli- April 27, the patient was discharged with oral prednisone cations involving the CNS might occur and should be tapering. promptly recognized and treated to reduce permanent neurologic disability.

Discussion Study funding We report a rare case of an immune-mediated CNS disease No targeted funding reported. that occurs after SARS-CoV-2 infection. Phenotypically, the disease resembled an atypical form of neuromyelitis optica Disclosure spectrum disorder; however, (1) the hyperacute dynamic of G. Novi: received speaker honoraria from Merck, Novartis, and the disease, (2) the presence of multiple, synchronous, en- Roche. T. Rossi: reports no disclosures relevant to the manu- hancing brain lesions, (3) the lack of anti-AQ4 or anti-MOG script. Enrico Pedemonte: received speaker honoraria or con- abs, (4) the absence of longitudinally extended transverse sultation fees from Biogen and Merck-Serono. L. Saitta: reports myelitis, and (5) the presence of a viral infection preceding no disclosures relevant to the manuscript. C. Rolla: reports no

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN disclosures relevant to the manuscript. L. Roccatagliata: reports no disclosures relevant to the manuscript. M. Inglese: received Appendix (continued) honoraria or consultation fees from Roche, Biogen, Merck- Name Location Contribution Serono, Novartis, and Genzyme and research grants from NIH, NMSS, FISM, and Teva Neuroscience. D. Farinini: reports no Laura Saitta, MD San Martino Major role in the acquisition of Hospital, Genova data disclosures relevant to the manuscript. Go to Neurology.org/ NN for full disclosures. Claudia Rolla, MD San Martino Major role in the acquisition of Hospital, Genova data

Publication history Luca University of Analyzed the data and revised ﬂ Roccatagliata, Genova, Genova, the manuscript for intellectual Received by Neurology: Neuroimmunology & Neuroin ammation MD, PhD Italy content May 4, 2020. Accepted in ﬁnal form May 13, 2020. Matilde Inglese, University of Interpreted the data, design MD, PhD Genova, Genova, and conceptualized study, Italy analyzed the data, and revised the manuscript for intellectual Appendix Author content

Name Location Contribution Daniele Farinini, San Martino Revised the manuscript for MD Hospital, Genova intellectual content Giovanni Novi, San Martino Major role in the acquisition of MD Hospital, Genova data, interpreted the data, design and conceptualized study, analyzed the data, and References drafted the manuscript 1. Pohl D, Alper G, Van Haren K, et al. Acute disseminated encephalomyelitis: updates on an inﬂammatory CNS syndrome. Neurology 2016;87:S38–S45. Tommaso Rossi, San Martino Major role in the acquisition of 2. Guan W, Ni Z, Hu Y, et al. Clinical characteristics of coronavirus disease 2019 in MD Hospital, Genova data China. N Engl J Med 2020;382:1708–1720. 3. Helms J, Kremer S, Merdji H, et al. Neurologic features in severe SARS-CoV-2 Enrico San Martino Major role in the acquisition of infection. New Engl J Med Epub 2020 April 15. Pedemonte, MD Hospital, Genova data 4. Li YC, Bai WZ, Hashikawa T. The neuroinvasive potential of SARS-CoV2 may play a role in the respiratory failure of COVID-19 patients. J Med Virol Epub 2020 Feb 27.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 CLINICAL/SCIENTIFIC NOTES OPEN ACCESS Severe paradoxical disease activation following alemtuzumab treatment for multiple sclerosis

Jamie Brannigan, Joanne L. Jones, MB ChB, PhD, and Sybil R. L. Stacpoole, MB BChir, PhD Correspondence Dr. Stacpoole Neurol Neuroimmunol Neuroinﬂamm 2020;7:e799. doi:10.1212/NXI.0000000000000799 [email protected]

A 39-year-old right-handed agricultural service engineer developed rapidly evolving severe relapsing-remitting multiple sclerosis (MS). MRI showed multiple T2 hyperintensities throughout his neuroaxis (ﬁgure, A). Several lesions showed restricted diﬀusion, and 2 enhanced. He received steroids for each relapse, making a full recovery (Expanded Disability Status Score [EDSS] 0).

There was clinical equipoise between induction therapy with alemtuzumab or natalizumab. He was JC virus positive. He chose treatment with alemtuzumab, partly because of the even chance that after 2 cycles, he might not require further treatment, plus the ease of administration, vs the longer-term risks with natalizumab.

Nine months after the first cycle of alemtuzumab, he represented with an encephalopathic picture and progressive focal neurology. He had a headache for 12 days, diplopia for 8 days, and became confused and unsteady on his feet over 5 days. He was apyrexial, looked unwell, was drowsy, and disorientated. Diplopia was present in all directions, and he had mild left upper motor neurone facial weakness and proximal weakness of his lower limbs. Reflexes were difficult to elicit, with upgoing plantars. He deteriorated markedly over the subsequent week, developing ophthalmoplegia, tetraparesis, inability to communicate, and respiratory compromise requiring intubation (EDSS 9.5).

Serum and CSF markers of infection were all negative, including listeria, but his CSF was active with signiﬁcantly elevated lymphocytes (200 cells/mm3; 90% lymphocytes) and protein (0.85 g/L); glucose was normal. CSF showed unmatched oligoclonal bands. Aquaporin-4 and MOG antibodies were negative.

CSF cytology conﬁrmed reactive lymphocytes (85%). Flow cytometry reported 59% T cells (CD4:CD8 2:1), 35% B cells (kappa:lambda 1.4:1), and 3% NK cells. An additional 13% had a larger morphology and phenotype proﬁle consistent with plasma cells, strongly expressing CD45 and CD38, moderately expressing CD81+, and weakly expressing CD19 and light chains.

Repeat brain MRI revealed numerous new white matter lesions. Most showed restricted diffusion, suggestive of acute demyelination (figure, B), but there was no enhancement (postgadolinium images not shown). EEG showed nonspecific generalized slowing, consistent with generalized cerebral dysfunction/encephalopathy.

His presentation was consistent with encephalitis, with an infective or inﬂammatory cause. Acyclovir, ceftriaxone, and amoxicillin were administered early, followed by 3,000 mg IV methylprednisolone over 3 days and 10 cycles of alternate day plasma exchange.

There was no clinical response, and repeat MRI brain showed progression (ﬁgure, C), still without contrast enhancement (not shown). Severe B cell–mediated demyelination was

From Jesus College (J.B., S.R.L.S.), Cambridge University, UK; Department of Neurology (J.L.J., S.R.L.S.), Addenbrooke’s Hospital, Cambridge University Hospitals NHS Foundation Trust, Cambridge, UK; and Department of Neurology (S.R.L.S.), Peterborough City Hospital, North West Anglia NHS Foundation Trust, Peterborough, UK.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

The Article Processing Charge was funded by the Wellcome Trust. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (CC BY), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Figure Evolving imaging features at diagnosis, representation, 3 and 5 weeks later, and after 17 months

(A) MRI of the brain and spinal cord at the point of diagnosis with rapidly evolving relapsing-remitting MS. His initial demyelinating event involved the brainstem and medial longitudinal fasciculus with several areas of involvement in the cerebral cortex. The second event localized to the lower spinal cord, with no evidence of involvement of the cervical spine. (B) MRI of the brain the day after representation, 9 months after the first cycle of alemtuzumab. T2-weighted imaging showed multiple new areas throughout the brain. Diffusion-weighted imaging confirmed appearances suggestive of acute demyelination with multiple areas of restricted diffusion. (C) MRI of the brain and cervical spine 3 weeks after representation. T2-weighted images show extensive demyelination in the brainstem, cerebral cortex, and cervical spine. (D) MRI of the brain at nadir, 5 weeks after representation. T2-weighted imaging illustrates extensive demyelination with areas of T1 signal change. (E) MRI of the brain 17 months later showing remarkable resolution over time.

suspected. Rituximab was commenced and continued 6 alemtuzumab is associated with poor thymic T-cell recovery monthly, alongside 3 cycles of cyclophosphamide. Sub- and exaggerated CD4+ T-cell homeostatic proliferation.2 CNS- sequent gradual improvement resulted in a remarkable func- directed alemtuzumab-induced autoimmunity is also unlikely tional recovery (EDSS 3.5 two years later with return to full to be due to numerical differences in B- and T-cell counts. time work) along with significant remyelination and resolu- Instead, this acute disseminated encephalomyelitis-like illness tion of T1 black holes (figures, D and E). after alemtuzumab is more likely to result from the complex interplay of multiple reconstituting immune cell subsets. Discussion Eleven of the other 17 reported cases of disease activation The first 2 cases of paradoxical disease activation in patients after alemtuzumab occurred in those switching therapies di- with MS treated with alemtuzumab were reported in 2017.1 rectly from fingolimod,1,3,4 suggesting confounding of path- The authors proposed that a secondary B cell–driven auto- ologic mechanisms. Following treatment with fingolimod, immune disease targeting the CNS and appearing similar to many lymphocytes remain hidden from the intravascular MS could occur after alemtuzumab therapy due to the ob- therapeutic effects of alemtuzumab due to selective lymphoid servation that B-cell numbers recover far more quickly after sequestration. Later egression could initiate rebound activity, alemtuzumab than CD4+ and CD8+ T cells, sometimes which also occurs if fingolimod treatment is stopped without overshooting pretreatment levels.1 However, the relative ki- substituting a different therapy.3 This disease activity would netics of B- and T-cell recovery have no effect on the risk of be due to a failure of alemtuzumab to bind to CD52+- non–CNS-directed autoimmunity. Rather, the risk of de- lymphocytes, not the repopulation kinetics of B and T cells veloping Graves disease and immune thrombocytopenia after after treatment.

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN One question on encountering paradoxical disease activation after alemtuzumab is whether to proceed to the second Appendix Authors treatment cycle or switch to rituximab (or an alternative anti- Name Location Contribution CD20 B-cell therapy such as ocrelizumab). Five patients in the literature received rituximab; 6 including ours. All have Jamie Jesus College, Cambridge Drafted the manuscript, Brannigan University, UK analyzed data, and responded well. Willis’ case series of disease activity 4–5 literature review months post-alemtuzumab in patients switched from ﬁn- Joanne Department of Neurology, Interpreted the data and golimod responded well to the second cycle of alemtuzu- Jones, MB Addenbrooke’s Hospital, revised the manuscript mab, as might be expected from the mechanisms discussed ChB, PhD Cambridge University for intellectual content Hospitals NHS Foundation above. Trust, Cambridge, UK

Sybil Jesus College, Cambridge Lead physician caring for We hope that the remarkable recovery of our patient, despite Stacpoole, University, UK the patient, data extensive brainstem involvement, will provide clinicians with MB BChir, Department of Neurology, acquisition, and revised fi PhD Addenbrooke’s Hospital, the manuscript for the con dence to treat in such severe situations. Cambridge University intellectual content Hospitals NHS Foundation Trust, Cambridge, UK Study funding Department of Neurology, No targeted funding was provided. J.L. Jones is supported by Peterborough City Hospital, North West Anglia NHS the Wellcome Trust (RG49413). Foundation Trust, Peterborough, UK Disclosure J. Brannigan and S.R.L. Stacpoole report no disclosures relevant to the manuscript. J.L. Jones has received honoraria References 1. Haghikia A, Dendrou CA, Schneider R, et al. Severe B-cell-mediated CNS disease from Sanofi-Genzyme.GotoNeurology.org/NNforfull secondary to alemtuzumab therapy. Lancet Neurol 2017;16:104–106. disclosures. 2. Jones JL, Thompson SAJ, Loh P, et al. Human autoimmunity after lymphocyte depletion is caused by homeostatic T-cell proliferation. Proc Natl Acad Sci U S A 2013;110:20200–20205. 3. Willis M, Pearson O, Illes Z, et al. An observational study of alemtuzumab following Publication history fingolimod for multiple sclerosis. Neurol Neuroimmunol Neuroinflamm 2017;4: fl e320. doi:10.1212/NXI.0000000000000320. Received by Neurology: Neuroimmunology & Neuroin ammation 4. Wehrum T, Beume LA, Stich O, et al. Activation of disease during therapy with January 6, 2020. Accepted in final form April 22, 2020. alemtuzumab in 3 patients with multiple sclerosis. Neurology 2018;90:e1–e5.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 CLINICAL/SCIENTIFIC NOTES OPEN ACCESS Acute necrotizing encephalopathy and myocarditis in a young patient with COVID-19

Ahmed Elkady, MSc, and Alejandro A. Rabinstein, MD Correspondence Dr. Elkady Neurol Neuroimmunol Neuroinﬂamm 2020;7:e801. doi:10.1212/NXI.0000000000000801 [email protected]

A 33-year-old woman, previously healthy was admitted with generalized status epilepticus. MORE ONLINE Four days before, she had developed generalized fatigue, fever, headache, and nasal con- gestion. On hospital arrival, she was comatose (Glasgow coma scale 5 [E1/V1/M3]), febrile COVID-19 Resources (38.6°C), tachycardic (145/min) with a blood pressure of 100/60 mm Hg, and tachypneic For the latest articles, (35/min) with hypoxemia (pulse oximetry 80%). She was emergently intubated and started invited commentaries, and on mechanical ventilation and received IV midazolam and valproic acid for seizure control. blogs from physicians Chest x-ray showed mild edema. The initial blood test results showed elevated myocardial around the world enzymes; high-sensitivity troponin-I, 2,210 pg/mL; probrain natriuretic peptide, 992 pg/ NPub.org/COVID19 mL; creatine phosphokinase, 1858 mcg/L; and creatine kinase-MB, 22.5 ng/mL. Blood count revealed a white blood cell count of 14550/μL with absolute lymphopenia of 0.92/μL. C-reactive protein and erythrocyte sedimentation rate levels were high, whereas procalcitonin and lactic acid were normal. ECG showed sinus tachycardia and diffuse ST segment elevation. Bedside echocardiography showed diffuse myocardial dyskinesia with low left ventricular ejection fraction (29%) and small pericardial effusion. Head CT showed diffuse brain edema.

On day 2, brain MRI showed bilateral hemorrhagic thalamic (ﬁgure, A–C) and cerebellar lesions (ﬁgure, E–G). Noninvasive and invasive cerebral angiographies excluded venous or arterial occlusions and showed no signs of vasculitis. Lumbar puncture had an opening μ pressure of 22 cmH2O, and CSF showed a white cell count of 26 cells/ L (90% lymphocytes), protein of 541 mg/dL, no oligoclonal bands, and normal glucose; PCR tests results for herpes simplex types 1 and 2, varicella-zoster, cytomegalovirus, Epstein-Barr virus, West Nile virus, Coxsackie viruses, echovirus, and dengue were negative. Testing for the presence of SARS-CoV2 in the CSF was unable to be performed. Follow-up chest x-ray showed worsening bilateral alveolar and interstitial lung edema. Serum interleukin-6 (IL-6) was 378 pg/mL, whereas IL-6 in the CSF was not performed. High-dose IV methylprednisolone (1,000 mg/d) was started for presumptive diagnosis of acute necrotizing encephalitis (ANE).

On day 3, a nasopharyngeal swab specimen was tested for severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2). Follow-up brain MRI with gadolinium showed rim enhancement of the bilateral thalamic and cerebellar lesions (figure, D and H); it also disclosed signs of acute pansinusitis (figure, H). Cardiac MRI showed diffuse myocardial signal hyperintensity in both ventricles and associated pericarditis. Blood tests for autoimmune diseases were negative; moreover, blood, urine, and sputum cultures revealed no organism growth.

On day 4, the patient remained comatose and hypoxemic. Blood tests revealed new thrombocytopenia and elevated liver enzymes. The patient then had sudden neurologic deterioration with anisocoria and extensor posturing. Emergent head CT demonstrated diﬀuse brain swelling with right thalamic and right cerebellar hemorrhages.

From the Department of Neurology (A.E.), Mataria Teaching Hospital, Cairo, Egypt; and Department of Neurology (A.A.R.), Mayo Clinic, Rochester, MN.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

The article processing charge was funded by the authors. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work, provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Figure Brain MRI showing bilateral inflammatory lesions in thalami (arrows) (A–D) and cerebellum (headarrows) (E–H) with hemorrhagic components and rim contrast enhancement within brain; (D) left occipital minimal leptomeningeal enhancement (headarrows); (H) right maxillary sinus showing circumferential mucosal enhancement (arrows). Sequences shown are diffusion-weighted imaging (A and E), gradient recall echo (B and F), T2 FLAIR (C and G) and postcontrast T1-weighted (D and H)

On day 5, she suffered a cardiopulmonary arrest, and re- Disclosure suscitation attempts were unsuccessful. One day later, the A. Elkady and A. A. Rabinstein reports no disclosures relevant to results of her nasopharyngeal swab confirmed the detection of the manuscript. Go to Neurology.org/NN for full disclosures. SARS-CoV2 by PCR. Publication history Received by Neurology: Neuroimmunology & Neuroinflammation Discussion May 11, 2020. Accepted in final form May 19, 2020. Neurologic complications from SARS-Cov2 infection are in- 1 sufficiently understood. ANE has been described in children Appendix Authors and adults after various acute viral respiratory infections, most Name Location Contribution notably influenza A (H1N1).2 Concomitant extrapulmonary manifestations includes myocarditis and pericarditis.3 Our Ahmed Mataria Teaching Major role in the acquisition of fi Elkady, MSc Hospital, Cairo, data; coordinated imaging patient had con rmed SARS-Cov2 infection and presented Egypt with features consistent with fulminant ANE and myocarditis Alejandro A. Mayo Clinic, Conceptualized case; drafted the in addition to hypoxemic respiratory failure from ARDS. We Rabinstein, Rochester, MN manuscript for intellectual know of only 1 previously reported case of ANE associated MD content with SARS-Cov2 infection,4 and myocarditis has been documented.5 The frequency of these complications is unclear. References Necropsy was not performed in our patient, but her neuro- 1. Mao L, Jin H, Wang M, et al. Neurologic manifestations of hospitalized patients with radiologic findings were quite characteristic for the diagnosis coronavirus disease 2019 in Wuhan, China. JAMA Neurol Epub 2020 Apr 10. doi: 10.1001/jamaneurol.2020.1127. of ANE. This diagnosis must be considered in patients with 2. Mizuguchi M, Yamanouchi H, Ichiyama T, Shiomi M. Acute encephalopathy associated SARS-CoV-2infection presenting with or developing altered with influenza and other viral infections. Acta Neurol Scand 2007;115(4 suppl):45–56. consciousness. 3. Sellers SA, Hagan RS, Hayden FG, Fischer WA. The hidden burden of influenza: a review of the extra-pulmonary complications of influenza infection. Influenza Other Respir Viruses 2017;11:372–393. 4. Poyiadji N, Shahin G, Noujaim D, Stone M, Patel S, Griffith B. COVID-19-associated acute hemorrhagic necrotizing encephalopathy: CT and MRI features. Radiology Study funding Epub 2020 Mar 31. doi: 10.1148/radiol.2020201187. fi 5. Inciardi RM, Lupi L, Zaccone G, et al. Cardiac involvement in a patient with The authors received no nancial support for the research, coronavirus disease 2019 (COVID-19). JAMA Cardiol Epub 2020 Mar 27. doi: authorship, and/or publication of this article. 10.1001/jamacardio.2020.1096.

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN CLINICAL/SCIENTIFIC NOTES OPEN ACCESS COVID-19-associated acutenecrotizingmyelitis

Javier Sotoca, MD, and Yensa Rodr´ıguez-Alvarez,´ MD Correspondence Dr. Sotoca Neurol Neuroimmunol Neuroinﬂamm 2020;7:e803. doi:10.1212/NXI.0000000000000803 [email protected]

A 69-year-old otherwise healthy woman presented to our clinic with irradiated cervical pain, MORE ONLINE imbalance, and motor weakness and numbness in the left hand, which had been ongoing for 7 days. Eight days before the onset of these symptoms, she had fever and dry cough. COVID-19 Resources For the latest articles, On admission, her neurologic examination showed right facial and left hand hypoesthesia, invited commentaries, and subtle left hand interosseous weakness, and general hyperreflexia. blogs from physicians around the world MRI of the brain was normal, whereas spinal cord images (figure 1, A and D) showed T2- NPub.org/COVID19 hyperintensity extending from the medulla oblongata to C7, involving most of the cord with diffuse patchy enhancing lesions, suggesting acute transverse myelitis.

Extensive diagnostic workup was performed, showing negative results in blood test for infectious, autoimmune diseases (including myelin oligodendrocyte glycoprotein and aquaporin- 4 antibodies), and other potential causes such as vitamin deﬁcits or antiphospholipid syndrome.

The patient’s CSF analysis showed a traumatic puncture (75 erythrocytes/μL), mild lymphocytic pleocytosis (75 cells/μL, 98% lymphocytes), hyperproteinorraquia of 2.83 g/L, normal adenosine deaminase, and glucose results; IgG index was normal, no oligoclonal bands were present, and bacterial culture and viral multi-PCR test were also negative. The presence of neuronal surface antibodies was also ruled out. Thoracoabdominal CT scan was negative for tumor and lymphadenopathy. SARS-CoV-2 PCR was positive in nasopharyngeal swab and negative in CSF. No cytokine levels were examined neither in serum nor CSF.

Treatment with methylprednisolone 1 g IV for 5 days resulted in initial improvement. However, a few days later, her clinical condition worsened markedly: she developed sensory motor deﬁcits in both hands and paraparesis with sphincter incontinence.

A new spinal MRI was performed (ﬁgure 1, B and E), showing transversally and caudally progression until T6 level with similar enhancement and a new area of central necrosis at the T1 level with peripheral enhancement.

Treatment with plasma exchange and other course of methylprednisolone pulses for 5 days with posterior slow oral prednisone tapering resulted in improvement of strength and motor function, until being able to walk with assistance, use electronic devices such as typing on her mobile phone or write with some diﬃculties, remaining left leg moderate weakness and no sphincter control. After 4 weeks from the clinical onset, she continues improving slowly and performing physical and occupational therapy.

Spinal MRI postplasmapheresis (ﬁgure 1, C and F) showed substantial decrease in myelitis extension and enhancement, but central necrosis at the C7-T1 level remained unchanged.

Acute necrotizing myelitis (ANM) is a rare inflammatory disorder of the spinal cord. Only a few cases have been associated to inflammatory diseases such as neuromyelitis optica or vasculitis, – paraneoplastic mechanisms, or adverse effect of new oncologic treatments.1 3

From the Neurology Department (J.S.) and Radiology Department (Y.R.-A.),´ Hospital Universitari Mutua´ Terrassa, Spain.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

Sagittal T2-weighted (A–C) and sagittal postcontrast T1-weighted (D–F). Baseline performed (A and D), follow-up 1 week (B and E), and plasmapheresis posttreatment (C and F). A LETM is seen in the cervical spinal cord (A) with patchy enhancement (D). Pro- gressive swelling of the spinal cord and a new necrotizing area (B: arrow) that shows peripheral enhancement (E: arrow). Significant decreases of both myelitis extension (C) and enhancement (F) after plasmapheresis treatment, with necrosis area in evolution. LETM = longitudinal extensive transverse myelitis.

Spine MRI in ANM usually shows hypointense T1 signal and Acknowledgment corresponding increase in T2 signal and, characteristically, The authors thank the patient and all the physicians and hemorrhage, cavitation, and postcontrast enhancement could health personnel involved in the care of patients in the also be seen. COVID-19 pandemic.

The exact pathogenesis of ANM remains obscure, and anal- Study funding ogously to acute necrotizing encephalitis (ANE), an in- Not related to this manuscript.. flammatory response (“cytokine storm”) secondary to a viral infection has been postulated as a possible cause.4 Human Disclosure coronaviruses are a group of respiratory viruses that can J. Sotoca reports no disclosures relevant to the manuscript. naturally reach the CNS in humans through hematogenous or Y. Rodr´ıguez-Alvarez´ reports no disclosures relevant to the neuronal retrograde route and could potentially be associated manuscript. Go to Neurology.org/NN for full disclosures. with neurologic symptoms.5 A systemic cytokine production due to the SARS-CoV-2 infection has been suggested to Publication history contribute to the pathophysiology of severe coronavirus dis- Received by Neurology: Neuroimmunology & Neuroinflammation ease 2019 (COVID-19).6 Recently, a case of ANE caused by May 11, 2020. Accepted in final form May 18, 2020. the new coronavirus infection has been reported.7

In our patient, the presence of a longitudinal extensive transverse myelitis with subsequent worsening along with development of a necrotic area, associated with focal swelling, Appendix Authors

peripheral enhancement, and hypointense foci on T2 images, Name Location Contribution led us to the diagnosis of ANM in a patient COVID-19 positive. Javier Neurology Department. Drafting and revising the Immunomodulatory treatment such as steroids or plasma- Sotoca, MD Hospital Universitari manuscript, data pheresis can result in neurologic improvement as the patient MutuaTerrassa,´ Terrassa acquisition, analysis and (Barcelona), Spain interpretation of data, and reported here. edited and approved the final draft Additional studies are needed to better deﬁne the potential Yensa Radiology Department. Drafting and revising the role of human coronaviruses in the pathogenesis and the ef- Rodr´ıgyez- Hospital Universitari manuscript, edited fectiveness of any therapeutic measure in the management Alvarez,´ MutuaTerrassa,´ Terrassa images, and edited and MD (Barcelona), Spain approved final draft of ANM.

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN References 4. Kansagra SM, Gallentine WB. Cytokine storm of acute necrotizing encephalopathy. – 1. Katz JD, Ropper AH. Progressive necrotic myelopathy. Arch Neurol 2000;57: Pediatr Neurol 2011;45:400 402. 355. 5. Desforges M, Le Coupanec A, Dubeau P, et al. Human coronaviruses and other respiratory viruses: underestimated opportunistic pathogens of the central nervous 2. Okai AF, Muppidi S, Bagla R, Leist TP. Progressive necrotizing myelopathy: system? Viruses 2019;12:1–28. Part of the spectrum of neuromyelitis optica? Neurol Res 2006;28: 6. Moore JB, June CH. Cytokine release syndrome in severe COVID-19. Science 2020; 354–359. 368:473–474. 3. Abdallah AO, Herlopian A, Ravilla R, et al. Ipilimumab-induced necrotic myelopathy 7. Poyadji N, Shahin G, Noujaim D, et al. COVID-19 associated acute hemorrhagic in a patient with metastatic melanoma: a case report and review of literature. J Oncol necrotizing encephalopathy: CT and MRI features. Radiology 2020:201187. doi: Pharm Pract 2016;22:537–542. 10.1148/radiol.2020201187.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 CLINICAL/SCIENTIFIC NOTES OPEN ACCESS Should interferons take front stage as an essential MS disease-modifying therapy in the era of coronavirus disease 2019?

Cole Maguire, Teresa Frohman, PA-C, MSCS, FANA, Scott S. Zamvil, MD, PhD, Correspondence Elliot Frohman, MD, PhD, FANA, FAAN, and Esther Melamed, MD, PhD Dr. Melamed [email protected] Neurol Neuroimmunol Neuroinﬂamm 2020;7:e811. doi:10.1212/NXI.0000000000000811

In the unprecedented pandemic of the coronavirus disease 2019 (COVID-19) along with MORE ONLINE a limited clinical understanding on eﬀective vaccines and therapies, there are currently many unknowns for patients with autoimmune conditions, such as MS, who require ongoing treat- COVID-19 Resources ment with immunotherapies. As information is currently lacking on the immune eﬀects of For the latest articles, COVID-19 in the context of MS disease-modifying therapies (DMTs), a challenging clinical invited commentaries, and question being faced by patients and neurologists is whether to continue current DMTs for blogs from physicians patients with MS and risk potentially greater morbidity and mortality due to COVID-19 around the world infection vs discontinue DMT therapy and risk MS disease relapse. At this time, we desperately NPub.org/COVID19 need data to guide which DMTs may best treat both COVID-19 and MS.

Impressively, there has been a rapid international medical response in repurposing several antiviral therapies toward COVID-19 treatment, including remdesivir, lopinavir, ritonavir, ribavirin, interferon-alpha (IFN-α), and interferon-beta (IFN-β)1,2 (NCT04315948). Type I interferons such as IFN-β are of particular interest for patients with MS because these DMTs have been Food and Drug Administration-approved for use in MS since the 1990s.3 Interferons were originally discovered to “interfere” with viral replication and are classified as type I (IFN-α, β, and ω), type II (IFN-γ), or type III. In the human IFN-α family, there are 13 genes encoding several isoforms, many of which have been popularized for treatment. By contrast, human IFN- β is encoded by only 2 genes with a more limited number of commercially available recombinant isoforms. Between the type I interferons, IFN-β has been primarily used in MS, whereas IFN-α has been commonly used in the treatment of viral infections, such as herpes zoster, hepatitis B and C, and HIV. However, there is also evidence to suggest that IFN-α may have beneficial effects in MS. For example, treatment with IFN-α2a can lead to a reduction in MRI disease activity in patients with MS with neutralizing antibodies to IFN-β.4

Different IFN-α and IFN-β isoforms are currently being evaluated and compared for the treatment of COVID-19. Interferon signaling pathways seem to be significantly upregulated, especially during the most critical stages of pulmonary disease, with IFN-α serum levels correlating with disease severity during the peak of disease.5,6 Similarly, circulating IFN-β has also been reported to increase during peak disease stages although IFN-β levels remain elevated even after symptomatic improvement.5,6 These data suggest that IFN-α may have a key role in the reduction of the viral load during the peak of the COVID-19 disease, whereas IFN-β may have a potential role in the reduction of viral replication over the entire course of the disease. Because interferons are known to induce different cytokines and genes in different cell types and organs, one important caveat to consider is the timing of interferon administration in COVID-19 and the possibility of dichotomous interferon effects during early vs later stages of the disease.

From the Department of Neurology (C.M., T.F., E.F., E.M.), Dell Medical School, Austin, TX; and Department of Neurology and Program in Immunology (S.S.Z.), University of California, San Francisco, CA.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

The Article Processing Charge was funded by the authors. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work, provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 IFN-α has already been used as an active treatment for Acknowledgment COVID-19 in China.1 Both nebulized and subcutaneous The authors thank Dr. William Schwartz for critical reading of forms of recombinant IFN-α, IFN-β1a, and IFN-β1b are also the manuscript. currently tested in COVID-19 clinical trials in combination with other therapeutics (NCT04315948, NCT04276688, Study funding NCT04293887, NCT04251871, and NCT04275388). This work was supported by the University of Texas Dell Medical School Start Up funding to E. Melamed. The cor- One of the largest current trials, the World Health Organi- responding author, E. Melamed, had the final responsibility zation multinational SOLIDARITY trial, is currently evalu- for the decision to submit this manuscript for publication. ating 4 therapeutic regiments including (1) IFN-β with ritonavir/lopinavir, (2) ritonavir/lopinavir without IFN-β, Disclosure (3) chloroquine or hydroxychloroquine, and (4) remdesivir.2 None to report. Go to Neurology.org/NN for full disclosures. Of note, nebulized preparation of IFN-α has been reported to improve delivery to the respiratory tract while also reducing Publication history common interferon side effects, such as fatigue and myalgias, Received by Neurology: Neuroimmunology & Neuroinflammation symptoms that can also be a part of COVID-19 infection. April 19, 2020. Accepted in final form May 12, 2020.

Although interferon therapies revolutionized MS management in the 1990s, more recently, interferons have been rel- Appendix Authors egated to the backstage of the MS DMT armamentarium Name Location Contribution because of the emergence of more effective and better toler- fi Cole Maguire Dell Medical School at Designed and ated DMTs. Nevertheless, given the potential bene ts of the University of conceptualized study, interferons in the treatment of COVID-19, an important Texas at Austin performed literature searches, and wrote the question now is whether interferons should be again brought manuscript to the front stage of MS management. If data from ongoing fi Teresa Dell Medical School at Reviewing and editing clinical trials continue to support the bene t of interferons in Frohman, PA- the University of the treatment of COVID-19, it is perhaps prudent to consider C, MSCS, FANA Texas at Austin interferons as a first-line therapy in newly diagnosed patients Scott S. Zamvil, University of Reviewing and editing with MS. In patients with MS on other DMTs, we could MD, PhD California, San consider changing therapy to interferons or adding interferon Francisco therapy to the patient’s current DMT. For patients with active Elliot Frohman, Dell Medical School at Reviewing and editing MS, other potential treatment strategies may be to combine MD, PhD, the University of FANA, FAAN Texas at Austin interferon therapy with pulsed intermittent corticosteroids or IV immunoglobulin given monthly to quarterly or with Esther Dell Medical School at Designed and fl 7 Melamed, MD, the University of conceptualized study, a DMT that has antiviral properties (e.g., teri unomide ). In PhD Texas at Austin performed literature the latter case, a carefully formulated treatment plan tailored searches, and wrote the to the patient could provide a strategic advantage over cell manuscript depleting MS DMTs by virtue of the inherent multiplicity of actions of type I IFNs when used as part of combination therapy. References 1. Dong L, Hu S, Gao J. Discovering drugs to treat coronavirus disease 2019 (COVID-19). There are of course many unknowns regarding whether IFN-β Drug Discoveries Ther 2020;14:58–60. 2. Kupferschmidt K, Cohen J. WHO launches global megatrial of the four most promising or IFN-α therapies would be effective in the management of coronavirus treatments [online]. Science; 2020. Available at: https://www.sciencemag. COVID-19 and MS and which specific IFN-β and IFN-α iso- org/news/2020/03/who-launches-global-megatrial-four-most-promising-coronavirus- treatments. Accessed March 30, 2020. forms may best create this balance. It will also be important 3. Axtell RC, Steinman L. Type 1 interferons cool the inflamed brain. Immunity 2008;28: to continue to assess how treatments such as IVIG and steroids 600–602. 4. Myhr K, Riise T, Lilleås FG, et al. Interferon-α2a reduces MRI disease activity in may impact MS and COVID-19. Furthermore, treatment relapsing-remitting multiple sclerosis. Neurology 1999;52:1049. decisions will vary depending on patient age, disease activity, 5. Huang L, Shi Y, Gong B, et al. Blood single cell immune profiling reveals the interferon-MAPK pathway mediated adaptive immune response for COVID-19. comorbidities, and current DMT. Nevertheless, based on the medRxiv 2020.03.15.20033472; doi: 10.1101/2020.03.15.20033472. long-standing clinical experience and long-term safety pro- 6. Wei L, Ming S, Zou B, et al. Viral invasion and type I interferon response char- fi β acterize the immunophenotypes during COVID-19 infection. CELL-D-20-00887. le of IFN- use in patients with MS, interferons could be Available at SSRN: https://ssrn.com/abstract=3564998 or http://dx.doi.org/ a critical therapy to treat both COVID-19 and MS during the 10.2139/ssrn.3564998. 7. Gilli F, Li L, Royce DB, DiSano KD, Pachner AR. Treatment of Theiler’s virus- unprecedented international pandemic, and we should re- induced demyelinating disease with teriflunomide. J Neurovirol 2017;23: consider interferons’ place in the MS DMT armamentarium. 825–838.

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN CLINICAL/SCIENTIFIC NOTES OPEN ACCESS CNS inﬂammatory vasculopathy with antimyelin oligodendrocyte glycoprotein antibodies in COVID-19

Ashwin A. Pinto, DPhil, FRCP, Liam S. Carroll, PhD, MRCP, Vijay Nar, MRCP, Aravinthan Varatharaj, MRCP, and Correspondence Ian Galea, PhD, FRCP Dr. Pinto [email protected] Neurol Neuroimmunol Neuroinﬂamm 2020;7:e813. doi:10.1212/NXI.0000000000000813

A 44-year-old right-handed woman reported a gradual onset of right hand incoordination seven MORE ONLINE days after the onset of minor respiratory symptoms and pruritus due to COVID-19 infection. Over 48 hours, the patient developed word-finding difficulties and progression in right arm COVID-19 Resources weakness leading to presentation to the emergency department as a suspected stroke. For the latest articles, invited commentaries, and Neurologic examination confirmed a mild expressive and receptive dysphasia, visual and sensory blogs from physicians inattention, and Medical Research Council grade 4/5 weakness in the right arm and right leg. around the world There was a rash on the chest wall bilaterally but no abnormal respiratory findings. NPub.org/COVID19

Blood workup conﬁrmed normal results for full blood count (lymphocytes 1.5 × 109/L), C-reactive protein, lactate dehydrogenase, and ferritin. In addition, the blood tests for antinuclear antibody, antineutrophil cytoplasmic antibody, anticardiolipin immunoglobulin G and immunoglobulin M, lupus anticoagulant and cold agglutinins were negative. The HIV and syphilis serologies were also negative. Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) PCR from nasopharyngeal swab was positive.

MRI of the head with gadolinium and magnetic resonance angiography at presentation showed T2-hyperintensity within the centrum semiovale bilaterally in a periventricular location, extending along the left temporal and occipital horns and into the subcortical deep white matter bilaterally, more extensive in the left hemisphere. There was perivascular enhancement within the lesions, although no diffusion restriction, hemorrhage, or mass effect was found, and magnetic resonance angiography was normal (figure, A and D). MRI scan of the spinal cord was unremarkable with no radiologic signs of myelitis. CT of the chest, abdomen, and pelvis was normal with no evidence of pulmonary COVID-19 involvement.

CSF analysis showed 13/mm3 white cells (all mononuclear), red cells <1, protein 507 mg/L, glucose 2.9 mmol/L (serum glucose 6.3 mmol/L) with negative PCR for SARS-CoV-2, herpes simplex types 1 and 2, and JC virus. Oligoclonal bands were absent in the CSF. Although of potential relevance, serum and CSF cytokine analysis was unavailable for practical reasons during the university restrictions on activity in the laboratory.

There was clinical deterioration over the next 6 days with the development of severe aphasia and no antigravity movements of the right upper limb or at the right hip and knee. Repeat MRI brain scan, 6 days after presentation, showed progression of the bilateral centrum semiovale and white matter changes with extension into both hemispheres and more pronounced perivascular enhancement (ﬁgure, B and E). There were multiple, new cystic spaces without CSF signal consistent with enlarged perivascular spaces. Repeat CSF analysis on day 6 showed 8 mononuclear cells only and negative repeat SARS-CoV-2 PCR.

From the Wessex Neurological Centre (A.A.P., L.S.C., V.N., A.V., I.G.), Southampton General Hospital, Southampton, UK; and Clinical Neurosciences (A.A.P., A.V., I.G.), Clinical and Experimental Sciences, Faculty of Medicine, University of Southampton, UK.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

The Article Processing Charge was funded by the MRC (UKRI). This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (CC BY), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Figure MRI appearances of CNS inflammatory vasculopathy with antimyelin oligodendrocyte glycoprotein antibodies in COVID-19

T2-weighted axial images at day 1 (A), day 6 (B), and post-treatment day 17 (C). Postcontrast axial images at day 1 (D), day 6 (E), and post-treatment day 17 (F).

Treatment was initiated at day 6 with IV methylprednisolone unique clinicoradiologic CNS presentation. Vascular com- (IVMP) 1 g daily for 5 consecutive days, followed by oral plications are increasingly recognized in COVID-19. The prednisolone 60 mg daily. The patient did not receive experi- angiotensin-converting enzyme 2 receptors targeted by mental antiviral treatment of COVID-19. On day 8, plasma SARS-CoV-2 are expressed by endothelial cells in multiple exchange (PLEX) at 3.5L/d (1.5 plasma volumes) was com- organs including the brain.5 Recent histopathology from menced. There was a rapid clinical improvement in the neu- patients with COVID-19 has demonstrated a lymphocytic rologic deficit after the patient started immunomodulatory endotheliitis in the lungs, heart, kidney, small intestine, and treatment. The patient had normal speech, almost full power in liver with evidence of infarction.6 The blood-brain barrier the right arm and leg, and no visual or sensory inattention at breakdown secondary to endotheliitis, as suggested by the day 18 after 5 sessions of plasma exchange and IVMP. The MRI linear and punctate enhancement, may have facilitated the of the brain scan, after PLEX treatment on day 17 (figure, C entry of anti-MOG antibodies to initiate the disease process and F), showed residual white matter vasogenic edema but no and resulted in the unusual clinical and radiologic picture. The evidence of residual perivascular contrast-enhanced changes. enlarged perivascular spaces returned signal higher than the Two weeks after discharge from hospital, an antimyelin oli- CSF on fluid-attenuated inversion recovery sequences, which godendrocyte glycoprotein (MOG) antibody test requested on may represent distension by leucocytes migrating across the admission was reported as positive. cerebral endothelium before traversing the glia limitans.7 The twice negative CSF SARS-CoV-2 PCR supports the idea that Several classic autoimmune neurologic sequelae following the CNS pathology was not because of parenchymal infection. COVID-19 have been described to date.1 However, this case was The response to IVMP and PLEX was striking and is in unusual for classic anti-MOG disease for a number of reasons. keeping with the hypothesis of an immune-mediated process. When solitary brain involvement occurs in the absence of opticospinal disease, the clinical and radiologic presentation is Study funding usually similar to that of acute disseminated encephalomyelitis,2 Medical Research Council (UK)—A. Varatharaj and I. Galea. unlike here. In addition, perivascular enhancement is exceedingly rare in anti-MOG syndromes,3 with only one case reported.4 Disclosure We hypothesize that a parainfectious anti-MOG antibody No relevant disclosures. Go to Neurology.org/NN for full response combined with endothelial dysfunction to cause this disclosures.

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Publication history Received by Neurology: Neuroimmunology & Neuroinﬂammation Appendix (continued) ﬁ May 20, 2020. Accepted in nal form May 26, 2020. Name Location Contribution

Ian Galea, University of Interpreted the data and revised PhD, FRCP Southampton, UK the manuscript for intellectual content Appendix Authors

Name Location Contribution

Ashwin Wessex Neurological Design and conceptualized the References Pinto, DPhil, Centre, study, analyzed the data, and 1. Vonck K, Garrez I, De Herdt V, et al. Neurological manifestations and neuro-invasive FRCP Southampton, UK drafted the manuscript for mechanisms of the severe acute respiratory syndrome coronavirus type 2. Eur J Neurol intellectual content Epub 2020 May 16. 2. Jurynczyk M, Messina S, Woodhall MR, et al. Clinical presentation and prognosis in Liam Carroll, Wessex Neurological Major role in the acquisition of MOG-antibody disease: a UK study. Brain 2017;140:3128–3138. PhD, MRCP Centre, data 3. Den`eve M, Biotti D, Patsoura S, et al. MRI features of demyelinating disease asso- Southampton, UK ciated with anti-MOG antibodies in adults. J Neuroradiol 2019;46:312–318. 4. Komatsu T, Matsushima S, Kaneko K, Fukuda T. Perivascular enhancement in anti- Vijay Nar, Wessex Neurological Major role in the acquisition of MOG antibody demyelinating disease of the CNS. J Neurol Neurosurg Psychiatry MRCP Centre, data 2019;90:111–112. Southampton, UK 5. Xia H, Lazartigues E. Angiotensin-converting enzyme 2 in the brain: properties and future directions. J Neurochem 2008;107:1482–1494. Aravinthan University of Interpreted the data and revised 6. Varga Z, Flammer AJ, Steiger P, et al. Endothelial cell infection and endotheliitis in Varatharaj, Southampton, UK the manuscript for intellectual COVID-19. Lancet 2020;395:1417–1418. MRCP content 7. Bechmann I, Galea I, Perry VH. What is the blood-brain barrier (not)? Trends Immunol 2007;28:5–11.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 CLINICAL/SCIENTIFIC NOTES OPEN ACCESS Increased CSF levels of IL-1β,IL-6,andACE in SARS-CoV-2–associated encephalitis

Marta Bodro, MD, PhD,* Yaroslau Compta, MD, PhD,* Laura Llanso,´ MD, Diana Esteller, MD, Correspondence Antonio Doncel-Moriano, MD, Alex Mesa, MD, Alejandro Rodr´ıguez, MD, Jordi Sarto, MD, Dr. Bodro [email protected] ´ı Eugenia Mart nez-Hernandez, MD, PhD, Alexandru Vlagea, MD, PhD, Natalia Egri, MD, Xavier Filella, MD, PhD, or Dr. Compta Manuel Morales-Ruiz, MD, PhD, Jordi Yague,¨ MD, PhD, Alex´ Soriano, MD, PhD, Francesc Graus, MD, PhD, and [email protected] Felipe Garc´ıa, MD, PhD, on behalf of the “Hospital Cl´ınic Infecto-COVID-19” and “Hospital Cl´ınic Neuro-COVID-19” groups

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e821. doi:10.1212/NXI.0000000000000821

A national outbreak of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) MORE ONLINE emerged in December 2019 in Wuhan, China, rapidly evolving to the coronavirus disease 2019 (COVID-19) pandemic. Neurologic complications of COVID-19 include headache, confusion, COVID-19 Resources hyposmia, and dysgeusia,1 with encephalitis being rarely reported. Coronaviruses can poten- For the latest articles, tially invade the CNS through trans-synaptic propagation via nasal entry, likely causing invited commentaries, and hyposmia. Alternatively, CNS dysfunction may result from the systemic hyperinﬂammatory blogs from physicians response to the virus. We report 2 patients supporting this hypothesis. around the world NPub.org/COVID19 Patient 1 A 25-year-old healthy man presented with 1-day history of headache, left-sided paresthesias, and ipsilateral paresis progressing within 12 hours to confusion and agitation. His axillary temperature was 38.2°C. Brain CT and MRI scans were normal. CSF showed lymphocytic pleocytosis and increased proteins. He was started on IV acyclovir, ampicillin, and ceftriaxone, which were discontinued when CSF cultures and PCR ruled bacterial or viral etiologies. PCR of SARS-CoV-2 was negative in CSF but positive in the nasopharyngeal swab. On day 2, he fully recovered except for amnesia of the previous 2 days.

Patient 2 A healthy 49-year-old man presented with fever, myalgias, and dry cough lasting 1 week. A few hours after admission, he developed diﬃculty naming objects, temporospatial disorientation, confusion, and agitation. A thoracic CT scan showed bilateral peripheral opacities suggestive of COVID-19 pneumonia. Brain CT and MRI scans obtained 2 days later were unremarkable. CSF showed lymphocytic pleocytosis and increased proteins. The patient was empirically started on acyclovir, ampicillin, and ceftriaxone, which were discontinued once CSF cultures and PCR returned negative. PCR for SARS-CoV-2 was positive in the nasopharyngeal swab, but negative in CSF. Three days later, he was back to normal except for amnesia of the previous days.

None of the patients developed severe respiratory problems or required intensive care support.

*Both these authors contributed equally to this work.

From the Infectious Diseases Service (M.B., A.S.,´ F. Garc´ıa), Hospital Cl´ınic de Barcelona, Catalonia; Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS) (M.B., Y.C., X.F., M.M.-R., J.Y., A.S.,F.Graus,F.Garc´ ´ıa), Barcelona, Catalonia; Department of Medicine (M.B., Y.C., A.S.,´ F. Garc´ıa), Universitat de Barcelona, Catalonia; Neurology Service (Y.C., L.L., D.E., A.D.-M., A.R., J.S.), Hospital Cl´ınic de Barcelona, Catalonia; Institut de Neurociències (Y.C.), Maria de Maeztu excellence center, Universitat de Barcelona, Catalonia; Endocrinology and Nutricion Service (A.M.), Hospital Cl´ınic de Barcelona, Catalonia; Immunology Service (A.V., N.E., J.Y.), Centre de Diagnostic` Biomèdic (CDB), Hospital Cl´ınic de Barcelona, Functional Unit of Clinical Immunology, Hospital Sant Joan de Déu & Hospital Cl´ınic, Barcelona, Catalonia; CDB (X.F., M.M.-R.), Hospital Cl´ınic de Barcelona, Barcelona, Catalonia; and Department of Biomedicine (M.M.-R.), Universitat de Barcelona; CIBERehd; Catalonia, Spain.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

The Article Processing Charge was funded by the authors.

“Hospital Cl´ınic Infecto-COVID-19” group and “Hospital Cl´ınic Neuro-COVID-19” group coinvestigators are listed in appendix 2 at the end of the article. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Case 1 Case 2

Age 25 49

Sex Male Male

Comorbiditiesa No No

Respiratory symptoms No Yes

Rash Yes No

Anosmia and ageusia No No

Headache Yes No

Neck stiffness No No

C-reactive protein (CRP) [<1 mg/dL] <0.40 1.70

Ferritin (20–400 ng/mL) 151 428

Procalcitonin (<0.50 ng/mL) — <0,03

LDH (<234 U/L) 190 254

Dimer D (<500 ng/mL) 600 600

Leukocyte count (4–11 × 10^9/dL) 9 5.74

Lymphocyte count (0.9–4.5 × 10^9/dL) (17–55%) 1.9 (21%) 0.8 (13.1%)

Platelets (130–400 × 10^9/dL) 147 113

CSF proteins (150–450 mg/L) 1,055 1,155

CSF glucose (40–80 mg/dL) (2.8–4.2 mmol/L) 80 (3.6) 54 (2.99)

CSF nucleated cells/mm3 95 neutrophils 0% 90 neutrophils 0% Lymphocytes 98% Lymphocytes 99% Macrophages 2% Monocytes 1%

CSF erythrocytes/mm3 0 260

CSF IL-1β (pg/mL) 14.8 <2.56

CSF IL-6 (pg/mL) 190 25

CSF IFNα (pg/mL) <0.58 <0.58

CSF IFNβ (pg/mL) <8.78 <8.78

CSF ACE (U/L) 15.5 10.9

Abbreviations: ACE = angiotensin-converting enzyme; IL = interleukin. a Hypertension, respiratory chronic disease, cardiovascular disease, diabetes, cancer, chronic hepatopathy, or immunosuppression. CSF IL-1β and IL-6 were considered increased when greater than 2.56 pg/mL and 7 pg/mL, respectively. As for ACE, the normal range was 0–2.5 U/L.

Clinical features, serum, and CSF characteristics including cyto- but confusion, disorientation, and aphasia rapidly dominated kines and angiotensin-converting enzyme (ACE) profile from the clinical picture. both cases are shown in the table. Three previous case reports of CNS involvement in COVID- 19 suggest different pathogenic mechanisms: direct CNS Discussion infection demonstrated by detection of SARS-CoV-2 RNA These patients suggest that encephalitis may be the first or in CSF,2 recrudescence of symptoms related to previous dominant manifestation of COVID-19. For patient 1, the focal lesions (e.g., brain infarction) in the context of systemic neurologic deficits were the first symptom manifestation; his infection,3 and inflammatory-mediated mechanisms result- young age, absence of risk factors, and comprehensive studies ing in acute hemorrhagic necrotizing encephalopathy. This ruling out other etiologies suggest a link between the neuro- is a rare complication of viral infections, usually influenza, logic symptoms and systemic SARS-CoV-2 infection. By con- considered to result from severe systemic inflammation as- trast, patient 2 presented with typical COVID-19 symptoms, sociated with elevated cytokine levels, such as interleukin

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN (IL)-6 and tumor necrosis factor-α.4 No information on serum or CSF cytokine levels was provided for any of these 3 Appendix 1 Authors patients. Name Location Contribution

In our 2 patients, we cannot completely rule out a direct in- Marta Bodro Hospital Cl´ınic de Designed and conceptualized the Barcelona, study; analyzed the data; fectious mechanism, despite the negative CSF testing of SARS- Barcelona, Spain interpreted the data; and drafted CoV-2, but the rapid recovery in less than 3 days makes it un- the manuscript for intellectual content likely. Alternatively, and in keeping with the abovementioned cytokine-mediated systemic inflammation, there is evidence that Yaroslau Hospital Cl´ınic de Designed and conceptualized Compta Barcelona, the study; analyzed the data; patients with coronaviruses can develop a cytokine storm syn- Barcelona, Spain interpreted the data; and drafted drome with increased IL-1 and IL-6 among other inflammatory the manuscript for intellectual content mediators. Hence, in a study of children with acute encephalitis–like syndrome, serum anti–human coronavirus- Laura Llanso´ Hospital Cl´ınic de Acquisition of data and revised Barcelona, the manuscript for intellectual OC43 immunoglobulin M antibodies were present in 12% of Barcelona, Spain content patients and levels of IL-6, IL-8, monocyte chemotactic protein-1, Diana Hospital Cl´ınic de Acquisition of data and revised and Granulocyte Macrophage Colony-Stimulating Factor were Esteller Barcelona, the manuscript for intellectual 5 increased in their CSF. The elevated CSF levels of IL-6 in our 2 Barcelona, Spain content patients, and IL-1β in 1 of them, are in line with those studies. Antonio Hospital Cl´ınic de Acquisition of data and revised Biological anti-IL treatments targeting IL-1 (anakinra) or IL-6 Doncel- Barcelona, the manuscript for intellectual (tocilizumab and siltuximab) are useful to treat symptoms of Moriano Barcelona, Spain content CNS involvement related to the cytokine storm triggered by Alex Mesa Hospital Cl´ınic de Acquisition of data and revised chimeric antigen receptor T-cell therapy.6 Our patients improved Barcelona, the manuscript for intellectual Barcelona, Spain content spontaneously, but these treatments could be considered in more severe cases of COVID-19–associated encephalitis with increased Alejandro Hospital Cl´ınic de Acquisition of data and revised Rodr´ıguez Barcelona, the manuscript for intellectual CSF levels of ILs. Finally, our 2 patients also had increased CSF Barcelona, Spain content levels of ACE. It has been postulated that SARS-CoV-2 enters the 7 ffi Jordi Sarto Hospital Cl´ınic de Acquisition of data and revised cell using the ACE2 receptor. Although it is di cult to interpret Barcelona, the manuscript for intellectual the meaning of increased CSF levels of ACE in these 2 patients, Barcelona, Spain content they could be linked to the postulated alteration in the ACE Eugenia Hospital Cl´ınic de Acquisition of data and revised pathway in COVID-19. Mart´ınez- Barcelona, the manuscript for intellectual Hernández Barcelona, Spain content

The main implication of these 2 patients is that physicians should Alexandru Hospital Cl´ınic de Acquisition of data and revised be aware of COVID-19 infections presenting or predominantly Vlagea Barcelona, the manuscript for intellectual Barcelona, Spain content manifesting as encephalitis, likely resulting from activation of inflammatory pathways with increased ILs and ACE in CSF. Natalia Egri Hospital Cl´ınic de Acquisition of data and revised Barcelona, the manuscript for intellectual Barcelona, Spain content Acknowledgment Xavier Filella Hospital Cl´ınic de Acquisition of data and revised The authors acknowledge the patients for their generosity and Barcelona, the manuscript for intellectual to all other people suffering the COVID-19 pandemics, as well Barcelona, Spain content ff as to all the health professionals and other sta crucially Manuel Hospital Cl´ınic de Acquisition of data and revised involved in their care and the handling of this unprecedented Morales- Barcelona, the manuscript for intellectual global crisis. Ruiz Barcelona, Spain content Jordi Yague¨ Hospital Cl´ınic de Acquisition of data and revised Barcelona, the manuscript for intellectual Study funding Barcelona, Spain content No targeted funding reported. Alex Soriano Hospital Cl´ınic de Interpreted the data and revised Barcelona, the manuscript for intellectual Disclosure Barcelona, Spain content The authors report no disclosures relevant to the manuscript. Francesc Hospital Cl´ınic de Interpreted the data and revised Go to Neurology.org/NN for full disclosures. Graus Barcelona, the manuscript for intellectual Barcelona, Spain content

Publication history Felipe Garc´ıa Hospital Cl´ınic de Interpreted the data and revised ﬂ Barcelona, the manuscript for intellectual Received by Neurology: Neuroimmunology & Neuroin ammation Barcelona, Spain content May 1, 2020. Accepted in ﬁnal form May 19, 2020.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 Appendix 2 Coinvestigators Appendix 2 (continued)

Name Location Role Contribution Name Location Role Contribution

Laura Hospital Clinic de Site Collecting Nicole Hospital Clinic de Site Collecting Morata Barcelona, Barcelona, investigator data Garc´ıa Barcelona, Barcelona, investigator data Spain Spain

Juan Hospital Clinic de Site Collecting Pedro Hospital Clinic de Site Collecting Ambrosioni Barcelona, Barcelona, investigator data Puerta Barcelona, Barcelona, investigator data Spain Spain

Estela Hospital Clinic de Site Collecting David Hospital Clinic de Site Collecting Moreno- Barcelona, Barcelona, investigator data Nicol´as Barcelona, Barcelona, investigator data Garc´ıa Spain Spain

Veronica Hospital Clinic de Site Collecting Montserrat Hospital Clinic de Site Collecting Rico Barcelona, Barcelona, investigator data Sol´a Barcelona, Barcelona, investigator data Spain Spain

Berta Hospital Clinic de Site Collecting Antonio Hospital Clinic de Site Collecting Torres Barcelona, Barcelona, investigator data Moreno Barcelona, Barcelona, investigator data Spain Spain

Alexy Hospital Clinic de Site Collecting Manuel Hospital Clinic de Site Collecting Inciarte Barcelona, Barcelona, investigator data Torres Barcelona, Barcelona, investigator data Spain Spain

Laia Albiac Hospital Clinic de Site Collecting Jose Hospital Clinic de Site Collecting Barcelona, Barcelona, investigator data Antonio Barcelona, Barcelona, investigator data Spain Mart´ınez Spain

Daiana Hospital Clinic de Site Collecting Jose Mensa Hospital Clinic de Site Collecting Aguero¨ Barcelona, Barcelona, investigator data Barcelona, Barcelona, investigator data Spain Spain

Lorna Leal Hospital Clinic de Site Collecting Carolina Hospital Clinic de Site Collecting Barcelona, Barcelona, investigator data Garc´ıa-Vidal Barcelona, Barcelona, investigator data Spain Spain

Mariana Hospital Clinic de Site Collecting Aida Hospital Clinic de Site Collecting Chumbita Barcelona, Barcelona, investigator data Alejandre Barcelona, Barcelona, investigator data Spain Spain

Celia Hospital Clinic de Site Collecting Sergio Hospital Clinic de Site Collecting Cardozo Barcelona, Barcelona, investigator data Amaro Barcelona, Barcelona, investigator data Spain Spain

Montserrat Hospital Clinic de Site Collecting Sergi Hospital Clinic de Site Collecting Laguno Barcelona, Barcelona, investigator data Borrego- Barcelona, Barcelona, investigator data Spain Ecija Spain

Laura Hospital Clinic de Site Collecting Jordi Hospital Clinic de Site Collecting Linares Barcelona, Barcelona, investigator data Casanova Barcelona, Barcelona, investigator data Spain Spain

Irene Hospital Clinic de Site Collecting Maria Hospital Clinic de Site Collecting Macaya Barcelona, Barcelona, investigator data Centeno Barcelona, Barcelona, investigator data Spain Spain

Fernanda Hospital Clinic de Site Collecting Estefania Hospital Clinic de Site Collecting Meira Barcelona, Barcelona, investigator data Conde Barcelona, Barcelona, investigator data Spain Spain

Ana Hospital Clinic de Site Collecting Jose Miguel Hospital Clinic de Site Collecting Gonzalez- Barcelona, Barcelona, investigator data Contador Barcelona, Barcelona, investigator data Cordon´ Spain Spain

Marta Hospital Clinic de Site Collecting Andrea Hospital Clinic de Site Collecting Hernandez Barcelona, Barcelona, investigator data Garc´ıa- Barcelona, Barcelona, investigator data Spain Ortega Spain

Jhon Rojas Hospital Clinic de Site Collecting Alicia Hospital Clinic de Site Collecting Barcelona, Barcelona, investigator data Garrido Barcelona, Barcelona, investigator data Spain Spain

Lorena de la Hospital Clinic de Site Collecting Mar Guasp Hospital Clinic de Site Collecting Mora Barcelona, Barcelona, investigator data Barcelona, Barcelona, investigator data Spain Spain

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Appendix 2 (continued) Appendix 2 (continued)

Name Location Role Contribution Name Location Role Contribution

Sara Llufriu Hospital Clinic de Site Collecting Imma Hospital Clinic de Site Collecting Barcelona, Barcelona, investigator data Zardoya Barcelona, Barcelona, investigator data Spain Spain

Laura Hospital Clinic de Site Collecting Carlos Hospital Clinic de Site Collecting Molina- Barcelona, Barcelona, investigator data Castillo Barcelona, Barcelona, investigator data Porcel Spain Spain

Carmen Hospital Clinic de Site Collecting Carmen Hospital Clinic de Site Collecting Montejo Barcelona, Barcelona, investigator data Perez Barcelona, Barcelona, investigator data Spain Spain

Amaia Hospital Clinic de Site Collecting Soledad Hospital Clinic de Site Collecting Muñoz Barcelona, Barcelona, investigator data Barreiro Barcelona, Barcelona, investigator data Spain Spain

Judith Hospital Clinic de Site Collecting Olga Hospital Clinic de Site Collecting Navarro- Barcelona, Barcelona, investigator data Bertran Barcelona, Barcelona, investigator data Otano Spain Spain

Celia Hospital Clinic de Site Collecting Ana Camara Hospital Clinic de Site Collecting Painous Barcelona, Barcelona, investigator data Barcelona, Barcelona, investigator data Spain Spain

David Reyes Hospital Clinic de Site Collecting Guadalupe Hospital Clinic de Site Collecting Barcelona, Barcelona, investigator data Fernandez Barcelona, Barcelona, investigator data Spain Spain

Almudena Hospital Clinic de Site Collecting Idoia Zaro Hospital Clinic de Site Collecting S´anchez Barcelona, Barcelona, investigator data Barcelona, Barcelona, investigator data Spain Spain

Raquel Hospital Clinic de Site Collecting Mar´ıa Hospital Clinic de Site Collecting S´anchez Barcelona, Barcelona, investigator data Teresa Barcelona, Barcelona, investigator data Spain Blazquez Spain

Mar´ıa Hospital Clinic de Site Collecting Sepulveda´ Barcelona, Barcelona, investigator data Spain References 1. Chen N, Zhou M, Dong X, et al. Epidemiological and clinical characteristics of 99 Ana Isabel Hospital Clinic de Site Collecting cases of 2019 novel coronavirus pneumonia in Wuhan, China: a descriptive study. Tercero Barcelona, Barcelona, investigator data Lancet 2020;395:507–513. Spain 2. Moriguchi T, Harii N, Goto J, et al. A first Case of Meningitis/Encephalitis associated with SARS-Coronavirus-2. Int J Infect Dis 2020;94:55–58. Dolors Hospital Clinic de Site Collecting 3. Poyiadji N, shahin G, Noujaim D, Stone M, Patel S, Griffith B. COVID-19-associated Rambla Barcelona, Barcelona, investigator data acute hemorragic necrotizing encephalopathy: CT and MRI features. Radiology Epub Spain 2020 Mar 31. 4. Filatov A, Sharma P, Hindi Fawzi, Espinosa PS. Neurological complications of Xenia Hospital Clinic de Site Collecting coronavirus disease (COVID-19): encephalopathy. Cureus 2020;12:e7352. Barcons Barcelona, Barcelona, investigator data 5. Li Y, Li H, Fan R, et al. Coronavirus infections in the central nervous system and respiratory Spain tract show distinct features in hospitalized children. Intervirology 2017;59:163–169. 6. Hays P, Costello C, Asudani D. Clinical care of chimeric antigen receptor T-cell Maria Hospital Clinic de Site Collecting patients and managing immune-related adverse effects in the ambulatory and hos- Dolores Sag Barcelona, Barcelona, investigator data pitalized setting: a review. Future Oncol 2019;15:4235–4246. Spain 7. Patel AB, Verma A. COVID-19 and angiotensin-converting enzyme inhibitors and angiotensin receptor blockers. JAMA Epub 2020 Mar 24.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 5 CLINICAL/SCIENTIFIC NOTES OPEN ACCESS COVID-19-associated ophthalmoparesis and hypothalamic involvement

Elba Pascual-Goñi, MD, Juan Fortea, MD, PhD, Alejandro Mart´ınez-Domeño, MD, Nuria Rabella, MD, PhD, Correspondence Mario Tecame, MD, Cristina Gomez-Oliva,´ MD, PhD, Luis Querol, MD, PhD, and Dr. Querol [email protected] Beatriz Gomez-Ans´ on,´ MD, PhD

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e823. doi:10.1212/NXI.0000000000000823

SARS-CoV-2 is an emergent virus responsible for the coronavirus disease 2019 (COVID-19) MORE ONLINE outbreak. Reported neurologic manifestations associated with SARS-CoV-2 include hyposmia, headache, and consciousness disturbances.1 We describe 2 patients with COVID-19 presenting COVID-19 Resources with ophthalmoparesis and characteristic MRI findings. For the latest articles, invited commentaries, and blogs from physicians Case description around the world NPub.org/COVID19 Case 1 A 60-year-old woman presented with diplopia and right hemicranial headache after 10 days of fever, hyposmia, nausea, and cough. Neurologic examination revealed right abducens nerve palsy. The patient was hypoxemic (PaO2 67 mm Hg) in the absence of dyspnea. Blood investigations showed lymphopenia (430/μL) and increased C-reactive protein (C-RP; 85 mg/ L), with thiamine, pyridoxine, and D-dimer levels within the normal range. Chest x-ray revealed bilateral pneumonia. CSF showed 1 cell/mm3, normal protein levels (0.32 g/L), and elevated lactic acid dehydrogenase (LDH; 54 U/L). SARS-CoV-2 RNA was detected by RT-PCR in a nasopharyngeal swab specimen but not in CSF. Antiganglioside antibodies were negative, and oligoclonal bands were not detected in the CSF. Brain MRI (figure 1, A–D and figure e-1 links. lww.com/NXI/A276) showed diffuse fluid-attenuated inversion recovery (FLAIR)/T2- hyperintensity (HI) in the pontine tegmentum and focal HI in the right VI cranial nerve nuclei. The mammillary bodies and hypothalamus were HI, the pituitary gland was enlarged, and the upper pituitary stalk seemed globular. Treatment with hydroxychloroquine and azi- thromycin was started. Diplopia persisted 1 month after admission.

Case 2 A 35-year-old woman with a history of bulimia and a 3-week history of vomiting was admitted. During her hospital stay, she developed diplopia and paresthesia. Vomiting disappeared few days after admission, and she was discharged. Three days later, the patient presented with progressive encephalopathy and was again admitted. She had neither fever nor respiratory symptoms. Neurologic examination showed decreased arousal, disorientation, episodic memory deficits, bilateral abducens nerve palsy, and mild paraparesis with normal reflexes. Hyp- μ oxemia (PaO2 61 mm Hg), decreased lymphocytes (740/ L), increased C-RP (42 mg/L), and creatinine kinase (346 U/L) were detected; LDH, D-dimer, thiamine, pyridoxine, and ion levels were within the normal range. CSF revealed 0.53 g/L proteins, 2 cell/mm3, and normal LDH. Chest x-ray was normal. SARS-CoV-2 RNA was detected by RT-PCR in a nasopharyngeal swab specimen but not in CSF. Anti-AQP4 and antiganglioside antibodies were negative, and oligoclonal bands were not detected in the CSF. MRI (figure 1, E–H and figure e-1) demonstrated T2/FLAIR HI in the brainstem, including the VI cranial nerve nuclei, thalami, medial temporal lobes, mammillary bodies, and hypothalamus. The upper pituitary stalk seemed swollen and HI. The patient was diagnosed with Wernicke encephalopathy (WE) and was supplemented with

From the Department of Neurology (E.P.-G., J.F., A.M.-D., L.Q.), Neuroradiology Unit (M.T., B.G.-A.), Radiology Department, Microbiology Department (N.R.), and Gastroenterology Department (C.G.-O.), Hospital de la Santa Creu i Sant Pau, Universitat Autonoma` de Barcelona, Spain.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

The Article Processing Charge was funded by the Institut de Investigacions Biom`ediques Sant Pau. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Case 1: (A) Sagittal fluid-attenuated inversion recovery (FLAIR) demonstrates an enlarged, hyperintense pituitary gland, thickened pituitary stalk, and hyperintense hypothalamus. T2/FLAIR-hyperintensity (HI) is also seen in the dorsal midbrain and pons. (B) Sagittal postcontrast 3D magnetization-prepared rapid gradient echo imaging T1 shows a heterogeneously enhancing pituitary and upper stalk. (C) Axial FLAIR shows HI in the tegmentum, involving the right abducens nucleus. (D) Coronal FLAIR shows a hyperintense pituitary gland, prominent HI in the periventricular region of the III ventricle, and subtleHIinthe medial temporal lobes. Case 2: (E) Sagittal FLAIR demonstrates extensive HI in the hypothalamus and dorsal brainstem. (F) Axial FLAIR image shows HI and swelling of the hypothalamus, mammillary bodies, and dorsal midbrain. (G) Axial FLAIR demonstrates HI in the tectum of the midbrain and subtle HI in the medial temporal lobes. (H) Coronal FLAIR shows a prominent HI in the periventricular region of the III ventricle and in the mammillary bodies, and subtle HI in the medial temporal lobes.

thiamine and pyridoxine despite normal vitamin levels. The epileptic seizures in which SARS-CoV-2 was detected in the patient’s mental status partially improved during the following CSF, and MRI demonstrated T2-HI in the right medial days. A control MRI (1 week after) also demonstrated partial temporal lobe.6 Our current cases show prominent hypo- improvement of the MRI findings. After 1 month, oph- thalamus and pituitary stalk involvement on MRI. Indeed, thalmoparesis and paraparesis as well as her mental status SARS-CoV (a related coronavirus causing the 2002–2004 improved, but she had persistent episodic memory loss and SARS outbreak) has also been detected in hypothalamic depression. neurons of SARS autopsies,7 and hypothalamic-pituitary- adrenal axis dysfunction has been observed in almost 40% of SARS survivors, suggesting that SARS-CoV could cause Discussion hypophysitis and/or hypothalamic damage.8 Furthermore, To our knowledge, these are the 2 first patients with COVID- animal models have shown that after intranasal inoculations, 19 presenting with ophthalmoparesis and involvement of the SARS-CoV enters the brain through the olfactory bulb and via hypothalamus and mesencephalic tegmentum with some ra- the olfactory nerve with subsequent trans-synaptic spread, diologic features resembling those of WE.2 However, thia- causing selective neuronal infection and death in the absence fl mine levels in both our patients were normal, and there was of in ammation in the amygdala, raphe nuclei, and paramedial no condition leading to thiamine deficiency in patient 1 and hypothalamus, a topographic distribution of lesions similar to 9 some radiologic features in case 2 (limbic involvement) were that from our cases. The olfactory bulb involvement is sup- not typical of WE. Concurrency of both patients, both di- ported by the high rates of hyposmia in patients with COVID- agnosed with COVID-19, and the important hypothalamic 19 and would enable the trans-synaptic spread hypothesis. involvement suggested a relationship with the underlying However, other mechanisms, arising from the inflammatory infection. Other cases of ophthalmoparesis associated with response or the metabolic demands to the susceptible regions, COVID-19 have recently been reported,3,4 one of them with could also be playing a role in the development of our enlargement and contrast enhancement of the oculomotor patients’ lesions. nerve that we did not observe in our patients. Acute hemorrhagic necrotizing encephalopathy has also been reported in In conclusion, we report 2 cases of COVID-19-associated neu- association with confirmed COVID-19 on MRI.5 Similarly, rologic manifestations with abducens palsy, encephalopathy, and encephalomyelitis associated with SARS-CoV-2 has been re- characteristic MRI findings that suggest selective vulnerability of cently reported in a patient with unconsciousness and the involved regions. Whether this vulnerability is related to

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN direct viral tropism, the inﬂammatory cascade, or to an increase of local metabolic/vitamin demands remains to be elucidated. Appendix (continued)

Study funding Name Location Contribution No targeted funding reported. Mario Hospital de la Santa Interpreted the data and Tecame, MD Creu i Sant Pau, revised the manuscript for Barcelona, Spain intellectual content Disclosure E.P.-G. received speaking fees from Roche and Biogen. J.F., A. M.- Cristina Hospital de la Santa Interpreted the data and fl Gomez-Oliva,´ Creu i Sant Pau, revised the manuscript for D., N.R., M.T., and C.G.-O. declare no con icts. L.Q. has received MD, PhD Barcelona, Spain intellectual content speaker honoraria from Merck, Sanofi-Genzyme, Roche, Biogen, Luis Querol, Hospital de la Santa Conceptualized the study, Grifols, and CSL Behring; provided expert testimony for Grifols, MD, PhD Creu i Sant Pau, interpreted the data, and Johnson and Johnson, Annexin Pharmaceuticals, Alexion, Sanofi- Barcelona, Spain drafted and revised the manuscript for intellectual Genzyme, Novartis, and CSL Behring; and received research content funds from Roche and Grifols. B.G.-A. declares no conflicts. Go to Neurology.org/NN for full disclosures. Beatriz Hospital de la Santa Conceptualized the study, Gomez-´ Creu i Sant Pau, interpreted the data, and Anson,´ MD, Barcelona, Spain drafted and revised the Publication history PhD manuscript for intellectual content Received by Neurology: Neuroimmunology & Neuroinflammation May 4, 2020. Accepted in final form May 22, 2020. References 1. Mao L, Wang M, Chen S, et al. Neurological manifestations of hospitalized patients with COVID-19 in Wuhan, China: a retrospective case series study. SSRN Electron J Appendix Authors Epub 2020 Feb 25. 2. Zuccoli G, Cruz DS, Bertolini M, et al. MR imaging findings in 56 patients with Name Location Contribution wernicke encephalopathy: nonalcoholics may differ from alcoholics. Am J Neuro- radiol 2009;30:171–176. Elba Pascual- Hospital de la Santa Conceptualized the study, 3. Gutiérrez-Ortiz C, Méndez A, Rodrigo-Rey S, et al. Miller Fisher Syndrome and Goñi, MD Creu i Sant Pau, acquired and analyzed the polyneuritis cranialis in COVID-19. Neurology Epub 2020 Apr 27. doi: 10.1212/ Barcelona, Spain data, and drafted the WNL.0000000000009619. manuscript for intellectual 4. Dinkin M, Gao V, Mbbs JK, et al. COVID-19 presenting with ophthalmoparesis from content cranial nerve palsy. Neurology Epub 2020 May 1. doi: 10.1212/WNL. 0000000000009700. Juan Fortea, Hospital de la Santa Acquired and interpreted the 5. Poyiadji N, Shahin G, Noujaim D, Stone M, Pate P, Griffith B. COVID-19–associated MD, PhD Creu i Sant Pau, data and revised the acute hemorrhagic necrotizing encephalopathy: CT and MRI features. Radiology Barcelona, Spain manuscript for intellectual 2019;2:1–19. content 6. Moriguchi T, Harii N, Goto J, et al. A first case of meningitis/encephalitis associated with SARS-Coronavirus-2. Int J Infect Dis 2020;94:55–58. doi: 10.1016/j.ijid.2020. Alejandro Hospital de la Santa Interpreted the data and 03.062. Mart´ınez- Creu i Sant Pau, revised the manuscript for 7. Gu J, Gong E, Zhang B, et al. Multiple organ infection and the pathogenesis of SARS. Domeño, MD Barcelona, Spain intellectual content J Exp Med 2005;202:415–424. 8. Leow MKS, Kwek DSK, Ng AWK, Ong KC, Kaw GJL, Lee LSU. Hypocortisolism in Nuria Hospital de la Santa Acquired and interpreted the survivors of severe acute respiratory syndrome (SARS). Clin Endocrinol (Oxf) 2005; Rabella, MD, Creu i Sant Pau, data and revised the 63:197–202. PhD Barcelona, Spain manuscript for intellectual 9. Netland J, Meyerholz DK, Moore S, Cassell M, Perlman S. Severe acute respiratory content syndrome coronavirus infection causes neuronal death in the absence of encephalitis in mice transgenic for human ACE2. J Virol 2008;82:7264–7275.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 CLINICAL/SCIENTIFIC NOTES OPEN ACCESS High proportion of Guillain-Barr´e syndrome associated with chikungunya in Northeast Brazil

Aline de Moura Brasil Matos, MD, Fernanda Martins Maia Carvalho, MD, PhD, Danielle Lima Malta, PhD, Correspondence Cleonisio Leite Rodrigues, MD, PhD, Alvina Clara Félix, MSc, Claudio Sergio Pannuti, PhD, Dr. Romano [email protected] Amanda Dias da Rocha Lima, MSc, Danilo Lucas Alves Esposito,´ PhD, Leonilda Maria Barbosa dos Santos, PhD, Felipe von Glehn, MD, PhD, FAAN, Jeová Keny Baima Colares, MD, PhD, Benedito Antonioˆ Lopes da Fonseca, MD, PhD, Augusto César Penalva de Oliveira, MD, PhD, and Camila Malta Romano, PhD

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e833. doi:10.1212/NXI.0000000000000833

From 2013 to 2015, sanitary authorities reported an increased incidence of Guillain-Barr´e syndrome (GBS) associated with Zika virus (ZIKV) in French Polynesia, Caribbean, and – Brazil.1 3 After the end of ZIKV epidemics, GBS cases where still above the usual limits in countries where the arrival of chikungunya virus (CHIKV) was also a concern.3

Here, we report the ﬁndings from Hospital Geral de Fortaleza (HGF), a neuroinvasive arboviral disease vigilance center in Cear´a State, Northeast Brazil.

Methods We performed a prospective observational study that enrolled patients aged 15 year or older with the diagnosis of GBS4 (Brighton criteria I or II). All consecutive patients fulﬁlling the inclusion criteria from May 2015 to December 2017 were invited to participate.

Patients were evaluated for demographics, clinics (at admission and 6 months later), serum, and CSF complementary tests, and EMG. Owing to the local epidemics, besides investigating for classic GBS triggers, virologic tests for dengue virus, ZIKV, and CHIKV (i.e., ELISA IgM, IgG, and speciﬁc real-time PCR) were performed by a researcher blinded to clinical results. Neu- roﬁlament light chain5 (NfL) was measured in CSF and related to death, need for mechanical ventilation (MV), and incomplete recovery. CSF from 10 healthy subjects served as the control for NfL.

For comparison, oﬃcial reports of arboviral systemic infections and total GBS cases/year from 2013 to 2017 were requested to the local state government according the Law #12.527 No- vember 18, 2011.

Continuous data were summarized as median and interquartile range (IQR). Categorical data were presented as counts and percentages. Kendall τ was used for correlations and Mann- Whitney test for comparisons. p values <0.05 were deemed statistically signiﬁcant. Data were analyzed using SPSS version 25.0. Graphs were constructed using Sigma Plot version 11.0. The study was approved by the HGF ethics committee (CAAE: 56572316.9.0000.5040) and the USP ethics committee (CAAE: 00274418.7.0000.0068) and conducted according to appropriate Brazilian regulations. All subjects provided written informed consent.

From the Universidade de São Paulo (A.d.M.B.M., A.C.F., C.S.P., C.M.R.), Instituto de Medicina Tropical, Faculdade de Medicina; Hospital Geral de Fortaleza (F.M.M.C., C.L.R.), Serviço de Neurologia, CE; Universidade de Fortaleza (F.M.M.C., D.L.M., J.K.B.C.), Programa de Pos-Graduação´ em Ciˆencias M´edicas, CE; Universidade de Campinas (A.D.d.R.L., L.M.B.d.S., F.V.G.), Instituto de Biologia, SP; Universidade de São Paulo (D.L.A.E., B.A.L.d.F.), Faculdade de Medicina de Ribeirão Preto; Instituto de Infectologia Emilio Ribas (A.C.P.d.O.), São Paulo; and Laboratorio´ de Virologia (LIMHC 52) (C.M.R.), Hospital das Clinicas HCFMUSP, Faculdade de Medicina, Universidade de São Paulo, Brazil.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

The Article Processing Charge was funded by PROFAP-HCLIM. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Figure 1 Notified cases of dengue, Zika, and chikungunya in Cear´a state and GBS cases in the study site

CHIKV = Chikungunya virus; DENV = Dengue virus; GBS = Guillain-Barr´e syndrome; ZIKV = Zika virus.

Results Discussion A total of 42 patients with GBS suspicion were admitted. We found that 26% of the cases enrolled were associated with Eight patients were excluded (figure e-1, links.lww.com/ CHIKV as an infectious trigger. The increase was coincident NXI/A278). From the 34 remaining, GBS trigger was at- with the first local epidemics of CHIKV and followed a ZIKV tributed to CHIKV for 9 patients.6 For those, median age epidemics. The association might be the result of a molecular was 47 (IQR 31–55), 56% were men and most without mimicry autoimmune mechanism because CHIKV E1 gly- comorbidities (56%). The mean number of GBS cases coprotein shares homology with contactin-2,7 a protein in the previous 2 years of the study was 52 cases/year; present in the juxtaparanode.8 from 2015 to 2017, it was 88 cases/year. Most CHIKV- GBS cases concentrated in the CHIKV epidemic peak In one of the larger GBS cohort available,9 although no labora- (figure 1). tory tests are mentioned, there are no reports of rash/arthralgia as prodromal symptoms. Despite that, reports of GBS-ZIKV At admission, patients were confine to bed/wheelchair (56%), associations are well known,1,3 unlike GBS-CHIKV which are walking with support (33%), or unable to run (11%). During rarely reported. Regarding clinical outcomes, differences are also in hospital stay, 22% required MV and no patient died. Me- apparent from our GBS-CHIKV to this same cohort. For re- dian days in hospital were 16 (IQR 12–25.5). Recovery after 6 covery, all of our patients achieved Hugues score of “0” or “1” in months was complete for 33%, whereas 67% remained with 6 months (vs 61%). No patient with GBS-CHIKV died (vs 7%). minor signs or symptoms (table e-1, links.lww.com/NXI/ As for MV, we have similar numbers (22% vs 19%). A278). A limitation of our study is the small sample size. This was Major laboratory results can be found in supplementary ma- unavoidable because of the rare nature of GBS disease and the terial (links.lww.com/NXI/A278). EMG was primary de- seasonality of arbovirus infections. In addition, we were not able myelinating (75%) or primary axonal (25%). Although NfL to access antiganglioside tests. However, we performed a clinical presented higher titles than control (figure e-2), there was no follow-up of at least 6 months and adopted strict criteria for GBS, correlation with death, MV or, recovery. allowing proper exclusion of diagnostic mimics.

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Our findings suggest CHIKV as an important trigger for GBS during epidemics, overcoming classic triggers as Campylobacter Appendix (continued) fl jejuni, Epstein-Barr virus, and in uenza virus. Good outcomes Name Location Contribution were a commonplace in our study; however, sanitary authorities of areas affected by CHIKV should be aware of a possible increase Danielle Lima Universidade de Study concept and design, Malta, PhD Fortaleza data acquisition, and in GBS incidence and as a consequence an increased necessity for critical revision of the intensive care unit beds and rehabilitation treatments. manuscript from important intellectual content. Acknowledgment The authors would like to thank Professor Steven S. Witkin Cleonisio Leite Hospital Geral de Study concept and design, Rodrigues, MD, Fortaleza data acquisition, and from Weill Cornell Medicine for constructive comments and PhD critical revision of the kindness in editing the manuscript, Vaniaˆ Maria Alves de manuscript from important intellectual Araujo from the Hospital Geral de Fortaleza for the laboratory content. assistance, and Rosa Maria Nascimento Marcusso from Alvina Clara Universidade de São Data acquisition and Instituto de Infectologia Emilio Ribas for the organizational Félix, MSc Paulo critical revision of the support and revision of statistical analysis. manuscript from important intellectual content. Study funding Claudio Sergio Universidade de São Critical revision of the CNPq/Brazil under the grant number 2016/407429 and Funda- Pannuti, PhD Paulo manuscript from ção de Amparo àPesquisadoEstadodeSãoPaulo-FAPESP/Brazil important intellectual under the grants numbers 2014/26431-0 and 2019/03859-9. content. Amanda Dias da Universidade de Data acquisition and Disclosure Rocha Lima, Campinas critical revision of the MSc manuscript from A.d.M.B. Matos reports no disclosures. F.M. Maia Carvalho important intellectual received research financial support from Conselho Nacional content. ıfi de Desenvolvimento Cient´ co e Tecnológico CNPq/Brazil Danilo Lucas Universidade de São Data acquisition and under the grant number 2016/407429. D.L. Malta, C.L. Alves Esposito,´ Paulo, Ribeirão Preto critical revision of the PhD manuscript from Rodrigues, A.C. Félix, C.S. Pannuti, A.d.R. Lima, and D.L.A. important intellectual Espósito report no disclosures. L.M.B. dos Santos received content. fi research nancial support from Fundação de Amparo à Pes- Leonilda Maria Universidade de Data acquisition and quisa FAPESP/Brazil under the grant number 2014/26431-0. Barbosa dos Campinas critical revision of the Santos, PhD manuscript from F. von Glehn, J.K.B. Colares, B.A.L. da Fonseca, and A.C.P. de important intellectual Oliveira report no disclosures. C.M. Romano received re- content. fi search nancial support from Programa de fomento as ativi- Felipe von Universidade de Data acquisition and dades de lideranças cient´ıficas dos LIMs do Hospital das Glehn, MD, PhD, Campinas critical revision of the Cl´ınicas (PROFAP-LIM/HCFMUSP) #10/2020. Go to FAAN manuscript from important intellectual Neurology.org/NN for full disclosures. content.

Jeová Keny Universidade de Critical revision of the Publication history Baima Colares, Fortaleza manuscript from Received by Neurology: Neuroimmunology & Neuroinflammation MD, PhD important intellectual content. January 21, 2020. Accepted in final form May 26, 2020. Benedito Universidade de São Data acquisition and Antonioˆ Lopes Paulo, Ribeirão Preto critical revision of the da Fonseca, MD, manuscript from Appendix Authors PhD important intellectual content. Name Location Contribution Augusto César Instituto de Study concept and design, Aline de Moura Universidade de São Study concept and design, Penalva de Infectologia Dr. Emilio analysis and Brasil Matos, Paulo data acquisition, analysis Oliveira, MD, Ribas, São Paulo interpretation of the data, MD and interpretation of the PhD and critical revision of the data, and critical revision of manuscript from the manuscript from important intellectual important intellectual content. content. Camila Malta Universidade de São Study concept and design, Fernanda Hospital Geral de Study concept and design, Romano, PhD Paulo and hospital das data acquisition, analysis Martins Maia Fortaleza and data acquisition, analysis and Clinicas, São Paulo and interpretation of the Carvalho, MD, Universidade de interpretation of the data, data, and critical revision PhD Fortaleza and critical revision of the of the manuscript from manuscript from important important intellectual intellectual content. content.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 References 5. Rajabally YA, Uncini A. Outcome and its predictors in Guillain-Barre syndrome. – 1. Cao-Lormeau VM, Blake A, Mons S, et al. Guillain-Barre syndrome outbreak asso- J Neurol Neurosurg Psychiatry 2012;83:711 718. ciated with Zika virus infection in French polynesia: a case-control study. Lancet 2016; 6. Center for Disease Control and Prevention. Arboviral diseases, neuroinvasive and fi 387:1531–1539. non-neuroinvasise 2015 case de nition. In: NNDSS, ed. Available at: wwwn.cdc.gov/ 2. Arias A, Torres-Tobar L, Hernández G, et al. Guillain-Barré syndrome in patients with nndss/conditions/arboviral-diseases-neuroinvasive-and-non-neuroinvasive/case-def- a recent history of Zika in Cucuta, Colombia: a descriptive case series of 19 patients inition/2015/2015. Accessed January 2020. from December 2015 to March 2016. J Crit Care 2017;37:19–23. 7. Reddy V, Desai A, Krishna SS, Vasanthapuram R. Molecular mimicry between chi- 3. Mehta R, Soares CN, Medialdea-Carrera R, et al. The spectrum of neurological kungunya virus and host components: a possible mechanism for the arthritic mani- disease associated with Zika and chikungunya viruses in adults in Rio de Janeiro, festations. PLoS Negl Trop Dis 2017;11:e0005238. Brazil: a case series. PLoS Negl Trop Dis 2018;12:e0006212. 8. Masuda T. Contactin-2/TAG-1, active on the front line for three decades. Cell Adhes 4. Fokke C, van den Berg B, Drenthen J, Walgaard C, van Doorn PA, Jacobs BC. Migration 2017;11:524–531. Diagnosis of Guillain-Barre syndrome and validation of Brighton criteria. Brain 2014; 9. Doets AY, Verboon C, van den Berg B, et al. Regional variation of Guillain-Barre 137:33–43. syndrome. Brain 2018;141:2866–2877.

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN VIEWS & REVIEWS OPEN ACCESS Interleukin-6 in neuromyelitis optica spectrum disorder pathophysiology

Kazuo Fujihara, MD, Jeffrey L. Bennett, MD, PhD, Jerome de Seze, MD, PhD, Masayuki Haramura, PhD, Correspondence Ingo Kleiter, MD, Brian G. Weinshenker, MD, Delene Kang, Tabasum Mughal, PhD, and Dr. Fujihara [email protected] Takashi Yamamura, MD, PhD

Neurol Neuroimmunol Neuroinflamm 2020;7:e841. doi:10.1212/NXI.0000000000000841 Abstract Neuromyelitis optica spectrum disorder (NMOSD) is a rare autoimmune disorder that preferentially affects the spinal cord and optic nerve. Most patients with NMOSD experience severe relapses that lead to permanent neurologic disability; therefore, limiting frequency and severity of these attacks is the primary goal of disease management. Currently, patients are treated with immunosuppressants. Interleukin-6 (IL-6) is a pleiotropic cytokine that is significantly elevated in the serum and the CSF of patients with NMOSD. IL-6 may have multiple roles in NMOSD pathophysiology by promoting plasmablast survival, stimulating the production of antibodies against aquaporin-4, disrupting blood-brain barrier integrity and functionality, and enhancing proinflammatory T-lymphocyte differentiation and activation. Case series have shown decreased relapse rates following IL-6 receptor (IL-6R) blockade in patients with NMOSD, and 2 recent phase 3 randomized controlled trials confirmed that IL-6R inhibition reduces the risk of relapses in NMOSD. As such, inhibition of IL-6 activity represents a promising emerging therapy for the management of NMOSD manifestations. In this review, we summarize the role of IL-6 in the context of NMOSD.

From the Department of Multiple Sclerosis Therapeutics (K.F.), Fukushima Medical University School of Medicine; and Multiple Sclerosis and Neuromyelitis Optica Center, Southern TOHOKU Research Institute for Neuroscience, Koriyama, Japan; Departments of Neurology and Ophthalmology (J.L.B.), Programs in Neuroscience and Immunology, School of Medicine, University of Colorado, Aurora; Department of Neurology (J.S.), Hopitalˆ de Hautepierre, Strasbourg Cedex, France; Chugai Pharmaceutical Co. (M.H.), Ltd, Tokyo, Japan; Department of Neurology (I.K.), St. Josef Hospital, Ruhr University Bochum; Marianne-Strauß-Klinik (I.K.), Behandlungszentrum Kempfenhausen fur¨ Multiple Sklerose Kranke gGmbH, Berg, Germany; Department of Neurology (B.G.W.), Mayo Clinic, Rochester, MN; ApotheCom (D.K., T.M.), London, UK; and Department of Immunology (T.Y.), National Institute of Neuroscience, National Center of Neurology and Psychiatry, Tokyo, Japan.

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

The Article Processing Charge was funded by Chugai Pharmaceutical. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary ADEM = acute disseminated encephalomyelitis; AQP4 = aquaporin-4; ARR = annualized relapse rate; BBB = blood-brain barrier; CD59 = complement regulatory protein; CDC = complement-dependent cytotoxicity; CDCC = complement- dependent cellular cytotoxicity; EDSS = Expanded Disability Status Scale; GFAP = glial ﬁbrillary acidic protein; GRP78 = 78- kDa glucose-regulated protein; HR = hazard ratio; ICAM-1 = intracellular adhesion molecule 1; IgG = immunoglobulin G; IL-6 = interleukin-6; IL-6R = interleukin-6 receptor; IPND = International Panel for NMO Diagnosis; LETM = longitudinal extensive transverse myelitis; MAC = membrane attack complex; mIL-6R = membrane-bound IL-6R; MOG = myelin oligodendrocyte glycoprotein; MS = multiple sclerosis; NF-κB = nuclear factor kappa-light-chain-enhancer of activated B cells; NMO = neuromyelitis optica; NMOSD = neuromyelitis optica spectrum disorder; ON = optic neuritis; SE = standard error; SEM = standard error of the mean; sgp130 = soluble glycoprotein 130; sIL-6R = soluble IL-6R; SLE = systemic lupus erythematosus; TGF-β1 = transforming growth factor beta 1; Th = T helper cell; Treg = regulatory T cell.

Interleukin-6 (IL-6) is a soluble, pleiotropic cytokine that mesangial cells.6 IL-6 is involved in many physiologic processes, plays a key part in many biologic processes.1 It is a primary including inflammation, antigen-specificimmuneresponses, regulator of acute and chronic inflammation and hemato- host defense mechanisms, hematopoiesis, and production of poiesis, contributing to the onset and maintenance of various acute phase proteins.6 Outside the immune system, IL-6 can autoimmune and inflammatory disorders.1 promote angiogenesis, osteoclast differentiation, and kerati- nocyte and mesangial cell proliferation.7 Within the immune Neuromyelitis optica spectrum disorder (NMOSD) is an system, IL-6 plays a key part in the adaptive immune response uncommon, often debilitating, inflammatory condition of the by stimulating antibody production and effector T-cell de- CNS.2 Neuromyelitis optica (NMO) was previously consid- velopment.7 Furthermore, IL-6 has an important role in regu- ered a rare, severe variant of multiple sclerosis (MS); however, lating the balance between proinflammatory T helper (Th) 17 it is now recognized as a distinct autoimmune disorder.2 An cells and regulatory T cells (Treg).7 International Panel for NMO Diagnosis codified clinical, radiologic, and laboratory features collectively distinguishing IL-6 binds to the IL-6 receptor (IL-6R), which is expressed as NMOSD from MS and other CNS inflammatory disorders.2 membrane-bound (mIL-6R) and soluble (sIL-6R) forms.8 The Left untreated, patients with NMOSD experience new attacks sIL-6R binds to IL-6 with a similar affinity as the mIL-6R, and both or relapses, often leading to permanent disability.3 receptors interact with glycoprotein 130 (gp130, also known as IL- 6R subunit β) to initiate cellular signaling through the Src ho- The pathophysiologic processes and inflammatory cascade in mology region 2-containing protein tyrosine phosphatase-2/ NMOSD are complex and not fully understood. Circulating mitogen-activated protein kinase and Janus kinase/signal trans- immunoglobulin G (IgG) 1 antibodies targeting the astrocyte ducer and activator of transcription 3 protein pathways (figure B).8 water channel aquaporin-4 (AQP4) have been found almost Importantly, cells that do not express IL-6R and are therefore not exclusively in patients with NMOSD.3 AQP4-IgG binding to responsive to IL-6 can be stimulated by the complex of sIL-6R–IL- astrocytic AQP4 leads to classical complement cascade acti- 6 (IL-6 trans-signaling). IL-6 trans-signaling can be selectively vation and granulocyte and lymphocyte infiltration that blocked by the soluble form of gp130 (sgp130Fc)—which is combine to damage neural tissues.4,5 dimerized by a human immunoglobulin IgG1-Fc—without affecting IL-6 signaling via the membrane-bound IL-6R.8 IL-6 may drive disease activity in NMOSD by promoting plasmablast survival, stimulating AQP4-IgG secretion, re- Pathogenic role ducing blood-brain barrier (BBB) integrity and functionality, Dysregulation of IL-6 expression or signaling contributes to the and enhancing proinflammatory T-lymphocyte differentia- pathogenesis of various human diseases and is linked to in- fl tion and activation (figure A). CSF and serum IL-6 levels are ammatory and/or lymphoproliferative disorders, such as rheu- significantly elevated in patients with NMOSD, and IL-6 in- matoid arthritis, Castleman disease, multiple myeloma, giant cell 8 hibition has been shown to improve disease control (table 1). arteritis, and systemic lupus erythematosus (SLE). The path- Therefore, the IL-6 receptor (IL-6R) represents a promising ways driving IL-6 secretion from CNS-resident cells are complex. therapeutic target for NMOSD relapse prevention. This re- Neurons, astrocytes, microglia, and endothelial cells produce IL-6 view summarizes the role of IL-6 in NMOSD. following injury, and CSF IL-6 levels are elevated in multiple neuroinflammatory diseases.9 Interleukin-6 Dysregulation of IL-6 signaling may aggravate the inflammatory response in some CNS diseases; however, intrathecal IL-6 pro- Biological activities duction may have variable effects. Because the IL-6R is expressed IL-6 is produced by diverse cell types, such as T cells, B cells, on both oligodendrocyte progenitor cells and microglia, CNS IL- monocytes, fibroblasts, keratinocytes, endothelial cells, and 6 signaling may have both direct and indirect effects on

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Figure Potential roles of IL-6 signaling and inhibition in NMOSD pathophysiology and treatment

(A) Potential roles for IL-6 signaling in NMOSD pathophysiology. IL-6 induces differentiation of inflammatory Th17 cells from naive T cells, which in turn provide support to AQP4-dependent activated B cells. IL-6 also promotes differentiation of B cells into plasmablasts, inducing production of pathogenic AQP4-IgG. These events are followed by increased BBB permeability to antibodies and proinflammatory cell infiltration into the CNS, leading to binding of AQP4-IgG to AQP4 channels on the astrocytes. In response to stimulation by proinflammatory cytokines, astrocytes produce IL-6, which promotes demyelination and contributes to oligodendrocyte and axon damage. (B) Schematic of IL-6 signaling with potential modes of therapeutic inhibition. IL-6 can bind either to the membrane-bound (classic signaling) or soluble form (trans- signaling) of the IL-6R α receptor. IL-6 trans-signaling allows for the activation of cells that do not express the IL-6R α receptor. Classic signaling may be blocked by antibodies against IL-6 and IL-6R. Trans-signaling may be blocked by antibodies against IL-6R or the soluble form of glycoprotein 130 (sgp130). IL-6 signaling is mediated at theplasmamembranethroughthehomodimerizationofgp130,whichactivates the intracellular JAK-STAT and SHP2-MAPK signaling pathways. AQP4 = aquaporin-4; AQP4-IgG = aquaporin-4 immunoglobulin G; BBB = blood-brain barrier; CDC = complement-dependent cytotoxicity; CDCC = complement-dependent cellular cytotoxicity; D1-D3 = subdomain of IL-6Rα;IL-1β = interleukin-1β; IL-6 = interleukin-6; JAK/STAT = Janus kinase/signal transducers and activators of transcription; MAC = membrane attack complex; mAbs = monoclonal antibodies; MAPK = mitogen-activated protein kinase; NMOSD = neuromyelitis optica spectrum disorder; sgp130 = soluble glycoprotein 130; SHP2/MAPK = Src homology region 2 domain-containing phosphatase-2/mitogen-activated protein kinase; STAT3 = signal transducers and activators of transcription 3; Th = T helper cell; TNF-α = tumor necrosis factor α; Treg = regulatory T cell.

microglial and macroglial survival. In rat cerebral ischemic injury, predominantly in the spinal cord and optic nerves.4 Pa- it was shown that IL-6 modulates a decrease in neuronal and tients with NMOSD may present with a variety of symp- BBB integrity in the CNS.10 In addition, intrathecal IL-6 pro- toms, but most commonly with optic neuritis and myelitis. duction during CNS inflammation has been demonstrated to Optic neuritis or the inflammation of the contiguous optic ff 11 a ect BBB permeability. How these roles interact to promote chiasm causes acute visual impairment and eye pain. My- CNS injury in autoimmune disease requires further research. elitis causes varying degrees of motor paralysis, sensory loss, pain, or bladder and bowel dysfunction associated with Neuromyelitis optica MRI evidence of longitudinal extensive transverse myelitis (LETM) lesions. NMOSD may also lead to intractable spectrum disorder hiccups, nausea, or vomiting due to area postrema lesion Overview inflammation; brainstem dysfunction; or encephalopa- NMOSD is a rare, debilitating, autoimmune condition thy.12 Most patients with NMOSD experience a more se- fl of the CNS, characterized by in ammatory lesions vere disease course than do patients with MS due to

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 3 Table 1 Clinical case reports on IL-6R blockade in the treatment of patients with NMOSDa

ARR (IL-6 block) Case No. Age (y)/sex DD (y) before/during Other effects of anti–IL-6 AEs associated with anti–IL-6

154 31/F 12.1 1.6/0.5 No new or active SL (47 months) Transient diarrhea, deep venous thrombosis

254 18/F 8.8 2.1/0.3 No new or active SL (33 months) R-UTI during self-catheterization

354 30/F 5.5 2.5/0 No new or active SL (41 months) None

454 37/F 2.8 1.8/0.6 No new or active SL (34 months) Headache, fatigue

554 22/F 8.9 1.2/0 No new or SL (28 months) None

654 24/F 24 0.7/0 No new but still active SL (14 months) Transient mild fatigue

754 24/F 0.9 5.5/0.8 No new or active SL (12 months) Mild post-infusion nausea, transient gastritis, R-UTI

854 49/F 0.5 6/2.4 No new or active SL (3 months) R-UTI

955 32/F 8.8 1.3b/0 EDSS score: 9.0 to 2.5. Anti–AQP4-Ab None titer dropped from 1:800 to 1:20

1056 37/F 14 3/1.5b Oral PSL and AZA were tapered URIs, AEC, acute pyelonephritis, LKP and/or LPP

1156 38/F 11 2/0 NA NA

1256 26/F 5 2/0 Oral PSL was tapered Anemia

1356 31/M 19 2/0 Oral PSL and AZA were tapered AEC, LKP and/or LPP

1456 55/F 17 3/0.77b NA NA

1556 62/F 2 3/0 NA NA

1656 23/F 2 5/0 Oral PSL was tapered URIs, LKP and/or LPP, anemia

1757 36/F 1.7 4.3b/0 EDSS score: 8.0 to 2.5. MRI has No safety signals have occurred remained free of Gd activity

18c,d,58 40/F 9.4 2.6/0.6 EDSS score: 6.5 to 6.5. No new No serious AEs were observed lesions, no contrast enhancement

19c,d,58 26/F 8.2 2.7/0 EDSS score: 5.0 to 4.0. No new No serious AEs were observed lesions, no contrast enhancement

20c,d,58 39/F 2.5 1.7/1.3 No new lesions, no contrast UTI. Mild oral mucosis. No enhancement serious AEs

21c,59 36/F 14 5.3b/2b EDSS score: improved from 3.5 to 2.0. Decline in systolic blood pressure. LPP. No significant changes of lesions on MRI Enteritis caused by a norovirus. A URI

Abbreviations: AE = adverse event; AEC = acute enterocolitis; AZA = azathioprine; CS = corticosteroids; DD = disease duration; Gd = gadolinium; LKP = leukocytopenia; LPP = lymphocytopenia; NA = not applicable; PSL = prednisolone; RTX = rituximab; R-UTI = recurrent urinary tract infection; SL = spinal lesions; URI = upper respiratory infection. All patients are serum AQP4-IgG positive. a Please see the supplemental table (links.lww.com/NXI/A288) for full version of the table. b Calculated from the number of relapses, please see the supplemental table. c No oligoclonal IgG bands found. d No concomitant autoimmune diseases.

frequent and severe relapses that lead to early and in- 32%, depending on age, relapse rate, and recovery from cremental disability. A benign disease course of NMOSD is attacks.16 rare.13 Limiting frequency and severity of relapses represents a primary goal in NMOSD management.14 In 2004, the identification of AQP4-IgG greatly facilitated differentiation of NMO from MS,17 and in 2006, If patients are not treated, 49% and 70% of patients relapse AQP4-IgG serology was incorporatedintotherevised within 1 and 2 years, respectively.15 The mean annualized NMO diagnostic criteria.18 An international expert panel relapse rate (ARR) in patient cohorts ranges from 0.82 to in 2015 proposed a unifying term for the disease as 1.34, with a median time to first relapse of 14 months.15 NMOSD, with further stratification by AQP4-IgG sero- The mortality rate without treatment ranges from 9% to logic status.2

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN Pathophysiology A recent study demonstrated that some serum samples taken NMOSD is recognized as a humoral immune disease driven in from patients with NMOSD and SLE harbor autoantibodies most patients by the presence of AQP4-IgG.19 As a result, against the 78-kDa glucose-regulated protein (GRP78-IgG). initial attention has focused on B-cell autoimmunity in These autoantibodies bind to brain microvascular endothelial NMOSD. However, polymorphonuclear infiltration into the cells, resulting in nuclear factor kappa-light-chain-enhancer of CNS is a prominent feature of active disease and is charac- activated B-cell nuclear translocation, intercellular adhesion teristic of Th17-mediated pathology.20 Furthermore, clinical molecule 1 induction, reduced tight junction expression, and worsening in response to interferon-β therapy in patients with barrier permeability.26 Peripheral administration of recombi- NMOSD is also characteristic of Th17-mediated in- nant GRP78-IgG to mice resulted in increased BBB perme- flammatory processes. Thus, T-cell autoimmunity may also ability, suggesting that GRP78 autoantibodies may induce play a part in disease pathogenesis.20 BBB permeability and may contribute to disease activity in some patients.26 Further research is needed to confirm Role of AQP4 whether GRP78-IgG has a role in NMOSD pathogenesis. AQP4 is the most abundant water channel in the CNS and has a key role in transcellular water transport.21 AQP4 function Role of complement may also affect neuroinflammation, astrocyte migration, and Complement activation plays an essential part in NMOSD neuroexcitation. Within the CNS, AQP4 is expressed in the lesion formation.4 Activation of the classical complement endfeet of astrocytes that surround the blood vessels and cascade is thought to result from engagement of complement subarachnoid space and in retinal Mueller cells. It is particu- component C1q with AQP4-IgG bound to perivascular as- larly enriched in brain parenchyma interfacing with the CSF.21 trocyte endfeet.4 This leads to classical complement cascade activation and membrane attack complex formation.4 AQP4-IgG has a critical role in mediating CNS injury in NMOSD. AQP4-IgG is detected in ;70% of patients with Nytrova et al. showed that levels of the complement com- NMOSD, but not in patients with MS or other neurologic ponent C3a were higher in patients with NMO than in diseases.22 In AQP4-IgG–seropositive patients with NMOSD, healthy control subjects and that C3a levels in these patients CNS injury initiates with the binding of AQP4-IgG to AQP4 on also correlated with disease activity, neurologic disability, and perivascular astrocyte endfeet.4 Autoantibody binding results in AQP4-IgG.27 activation of the classical complement cascade, granulocyte and macrophage infiltration, secondary oligodendrocyte damage, Recent in vitro studies have identified alternative complement and neuronal death.4 CSF analysis of patients with NMOSD pathways, such as the bystander mechanism, where, following suggests that the majority of AQP4-IgG may transit passively to AQP4-IgG binding to astrocyte AQP4, activated, soluble the CNS through an open BBB.22 However, molecular tech- complement proteins were implicated in early oligodendrocyte niques show that AQP4-IgG is also produced intrathecally injury in NMOSD.28 In addition, complement regulatory during acute NMOSD exacerbations.23 protein (CD59) may confer a protective role in AQP4- IgG–seropositive NMOSD tissues outside of the CNS and thus Whether AQP4-IgG–seronegative patients with NMOSD have explain why peripheral AQP4-expressing cells in NMOSD re- the same disease as AQP4-IgG–seropositive patients remains main mostly unaffected.29 controversial. The demographics, clinical presentation, and prognosis differ between AQP4-IgG–seropositive and AQP4- Relationship between gut microbiota and IgG–seronegative patients.24 Some AQP4-IgG–seronegative humoral and cellular immunity patients are seropositive for myelin oligodendrocyte glyco- The AQP4 epitope p63-76 displays sequence homology with protein (MOG)–IgG; however, MOG-IgG–seropositive pa- p204-217 of Clostridium perfringens, a ubiquitous Gram- 30 tients show differences in natural history, neuroimaging, positive bacterium found in the human gut. This observa- and lesion histopathology from AQP4-IgG–seropositive pa- tion provided new perspectives for investigating NMOSD tients, supporting a distinct pathophysiology between these pathogenesis. Gut microbiome analysis of patients with NMO fi 30 disorders.25 identi ed an overabundance of C. perfringens. Because specific gut clostridia can regulate Treg and Th17 cell balance,31 Role of the BBB an excess of C. perfringens could theoretically evoke proin- Disruption of the BBB is important in the pathophysiology of flammatory AQP4-specific T- and B-cell responses driving NMOSD and correlates modestly with disease activity. In- NMOSD development. C. perfringens may also enhance Th17 trathecal production of AQP4-IgG23 may initiate disease ac- differentiation by promoting the secretion of IL-6 by antigen- tivity by causing focal BBB breakdown and precipitating a presenting cells in the gut.30 large influx of serum AQP4-IgG and serum complement into the CNS compartment. Alternatively, systemic inflammation Pathophysiologic role of IL-6 may disrupt the BBB, allowing entry of serum AQP4-IgG and IL-6 levels are associated with key NMOSD disease autoreactive B cells into the CNS and subsequent lesion markers.32,33 Studies have shown that CSF and serum levels of formation.19 IL-6 correlate with CSF cell counts and the Expanded Disability

Table 2 Spearman rank correlation coefficient (ρ) of the association among IL-6 levels, CSF cell count, and disability

CSF cell count, Mean (SD) CSF IL-6 ρ (p value) between Mean (SD) serum Baseline EDSS ρ (p value) correlation between Reference N Female, n cells/mm3 levels, pg/mL IL-6 levels and CSF cell count IL-6 levels, pg/mL score, mean (SD) IL-6 levels and EDSS score

Wang et al.33 22 16 NR 24.95 (25.57) NR NR 3.5 (1.0–8.5)a 0.372 (0.088)b

Uzawa et al.34 31 31 26.8 757.3 (1179.6) 0.638 (<0.001) NR 7.5 (2.0–9.0)a 0.258 (0.161)b

Uchida et al.35 29 27 8 (12.0)a,c 89.3 (737.7)a 0.4497 (0.021) 2.32 (0.00)d 6.0 (3)a NR

Matsushita et al.36 20 17 11.5 (15.3)e NR 0.75 (0.012) NR 5.6 (2.5) 0.72 (0.012)b

Icoz et al.37 23 17

AQP4–IgG+ 12 10 12f (3.63)g 781f (207.98)g NR 48f (9.44)g 6f (0.73)g NR

AQP4–IgG– 11 7 18f (3.41)g 39f (15.64)g NR 11f (4.11)g 4f (0.59)g NR

Barros et al.38 20 16 NR NR NR 514.1 (213.1)h 4.93 (1.91)i 0.5880 (0.0064)j 898.4 (411)k

Uzawa et al.39 17 17 9.5 281.0 (212.4)g 0.5 (0.002) NR NR NR

Abbreviations: AQP4-IgG = aquaporin-4-immunoglobulin G; EDSS = Expanded Disability Status Scale; IL-6 = interleukin-6; NMOSD = neuromyelitis optica spectrum disorder; NR = not reported. a Median (interquartile range). b In CSF. c /μL. d n = 29. e n = 19. f Mean/median not specified. g Mean (standard error). h Patients who did not relapse. i The EDSS score was determined either at the time the blood was analyzed or after 2 years of follow-up. j In plasma. k Patients who relapsed. – Status Scale (EDSS) score (table 2).33 39 CSF levels of IL-6 are pregerminal center–naive and postgerminal center memory shown to correlate with AQP4-IgG levels and glial fibrillary B cells, but not the circulating plasmablast compartment. IL-6R acidic protein levels, an indicator of astrocyte damage.34 CSF blockade may act indirectly to modulate AQP4-IgG production and serum IL-6 concentrations are significantly elevated in and relapse risk through its effects on T-cell development.44 patients with NMOSD and are higher than in healthy indi- Further immunologic studies of serum AQP4-IgG titers and viduals and patients with MS or other noninflammatory neu- circulating B-cell profiles via T-cell development in patients with – rologic disorders.34 37 In 1 study, an IL-6 CSF concentration of NMOSD receiving IL-6R inhibitors are needed. 7.8 pg/mL was proposed as an optimal cutoff value for diagnosing NMOSD.32 Furthermore, CSF IL-6 levels are also Effect of IL-6 on T cells significantly elevated in the acute phase of NMOSD compared Naive T cells are differentiated into Th17 cells through the with MS, myelitis, ON, and other inflammatory and non- combination of transforming growth factor beta 1 (TGF-β1), inflammatory neurologic disorders.40 IL-6 levels may be in- IL-6, and IL-23.45 Th17 cells mediate disruption of the BBB and dicative of NMOSD relapse because IL-6 serum and CSF subsequent CNS inflammation.45 Levels of Th17-related cyto- concentrations in patients with NMOSD are significantly kines, IL-17, IL-21, and IL-23 are increased in NMOSD, as is the higher during relapse than in periods of remission.36,38 level of IL-6, particularly in AQP4-IgG–seropositive patients with NMOSD.46 AQP4-specific T cells in AQP4- Barros et al. also demonstrated a relationship between serum IgG–seropositive patients with NMO exhibit Th17 polariza- IL-6 levels with occurrence of clinical relapses in patients with tion.47 In addition, levels of Th17-related cytokines, including IL- NMO.38 Furthermore, during a 2-year disease follow-up in 6, are significantly elevated in the CSF and peripheral blood of these patients, baseline IL-6 levels correlated with the risk of patients with NMOSD during relapse.34 The release of IL-6 and clinical relapses and severity. An 8-fold increase in relapse risk other Th17-related cytokines by activated T lymphocytes from was observed in patients with IL-6 serum concentrations above patients with NMOSD was also found to directly correlate with the baseline median value of 58.5 pg/mL during remissions.38 neurologic disability and risk of relapse.34,38

Higher IL-6 levels are also associated with greater disability in Effect of IL-6 in the BBB NMOSD following relapse. One study evaluated recovery from Proinflammatory cytokines, including IL-6, increase the per- relapse-related impairment in patients with NMOSD and meability of the BBB allowing antibodies and proin- demonstrated that patients with NMOSD and higher CSF IL-6 flammatory cells to infiltrate the CNS, leading to binding of levels experienced more modest improvements than did pa- AQP4-IgG to AQP4 channels on the astrocytes. In response tients with lower IL-6 levels.41 Furthermore, elevated IL-6 to stimulation by proinflammatory cytokines, astrocytes levels have been linked with short relapse-free durations in produce IL-6, which promotes demyelination and contributes NMOSD after relapse. Uzawa et al.41 evaluated the relapse-free to oligodendrocyte and axon damage (figure A). In vitro BBB duration after relapse in patients with NMOSD and found that models, including cocultures of human brain microvascular patients with high CSF IL-6 levels tended to have shorter endothelial cells and human astrocyte cell lines with or relapse-free durations than did those with low IL-6 levels fol- without AQP4 expression, showed that AQP4-IgG induced lowing relapse (p =0.079). IL-6 production by AQP4-positive astrocytes, and IL-6 signaling to endothelial cells impaired BBB function, increased Effect of IL-6 on B cells production of other chemokines, and enhanced flow-based B cells appear to have a central proinflammatory role in leukocyte transmigration.48 These effects were reversed with NMOSD immunopathology.19 This is mediated through an IL-6R-neutralizing antibody.48 several proposed mechanisms, including production of AQP4-IgG, antigen presentation, increased B-cell and plas- MOG-IgG-seropositive patients and IL-6 mablast activity, and reduced regulatory function of B cells.19 MOG is a glycoprotein localized on the surface of the myelin sheath and represents a potential target for the treatment of IL-6 was originally identified as B cell stimulatory factor-2, which demyelinating diseases. Mice immunized with MOG peptides induced activated B cells into antibody-producing cells.7 Indeed, may develop experimental autoimmune encephalomyelitis with − circulating short-lived, CD19intCD27high CD38highCD180 an NMO phenotype. MOG-IgG is found in a subset of patients plasmablasts may be increased in the peripheral blood of some with optic neuritis, acute disseminated encephalomyelitis/ patients with NMOSD and are capable of producing AQP4-IgG multiphasic disseminated encephalomyelitis, myelitis, AQP4- when stimulated with IL-6.42 However, these activated antibody- IgG–seronegative NMOSD, or encephalitis.49 Although MOG- secreting cells may not play a major part in the maintenance of IgG–seropositive and AQP4-IgG–seropositive NMOSD share a circulating serum AQP4-IgG. Detailed molecular studies have similar serum and CSF cytokine profile characterized by the demonstrated that AQP4-specific, peripheral blood B cells in the coordinated upregulation of multiple cytokines, especially Th17- − CD27+ classical and CD27 IgD memory compartments show related, these diseases have different pathologies.50 AQP4- the closest lineal relationship to AQP4-IgG–secreting cells in the IgG–seropositive NMOSD is primarily an astrocytopathic dis- CSF.43 In addition, human cell culture experiments showed that ease, whereas MOG-IgG–seropositive NMOSD is a de- AQP4-IgG–secreting cells were readily generated from myelinating disease.50

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 7 Therapeutic approaches AQP4-positive female patients with NMO and a history of Current clinical practice involves treatment of exacerba- treatment with rituximab and other immunosuppressive and tions with high-dose corticosteroids, apheresis therapies, immunomodulating agents experienced a reduction in me- and oﬀ-label use of immunotherapies for relapse pre- dian ARR (from 3.0 at baseline to 0.6 at the end of the vention, particularly with azathioprine, methotrexate, study), but not in impairment scores after switching to an and prednisone (immunosuppressants); mycophenolate anti–IL-6R antibody.58 There was no serious infection, mofetil (inhibits lymphocyte proliferation); anti-CD20 malignancy, hypersensitivity reaction, or elevation of trans- monoclonal antibody; and anti–IL-6R monoclonal anti- aminase levels. One patient experienced urinary tract in- body. Until recently, NMO/NMOSD treatment studies fection in the fourth month and mild oral mycosis in the were mostly retrospective. seventh month of therapy.58

There is a need for alternative therapeutic options for pa- Harmel et al.60 reported favorable outcomes following treat- tients with NMOSD who do not respond to ﬁrst-line therapy ment with an anti–IL-6R antibody after inadequate response and for treatments that manage other manifestations of the to interferon-β therapy in a patient misdiagnosed with MS. disease, including pain and fatigue. Diﬀerent therapeutic Under interferon-β treatment, the patient showed persisting targets have been investigated in international randomized relapse activity. Treatment with rituximab was initiated, but controlled trials. These include monoclonal antibodies tar- severe LETM attacks occurred shortly after. Following switch geting the B-cell antigen CD19 (inebilizumab),51 CD20 to an anti-IL-6R antibody, the patient experienced no further (rituximab), the complement component 5 (eculizumab),52 relapses for the following 12 months. Other single patient and IL-6R (satralizumab).53 reports included no adverse events during treatment.

IL-6 blockade in the treatment of NMOSD Satralizumab is a novel, humanized IgG2 recycling mono- Targeting the IL-6 signaling pathway inhibits both the hu- clonal antibody targeting the IL-6R. It has been investigated moral immune response as well as T-cell pathway and dys- in 2 phase 3 studies for the treatment of NMOSD.53,61 Both function of the BBB, providing a comprehensive treatment for results from the monotherapy trial (no other chronic base- the different NMOSD manifestations. Several case reports line immunosuppression allowed) and data from the add-on showed decreased relapse rate and reduced neurologic dis- to other immunosuppressive therapy trial were recently ability following off-label treatment with an anti-IL-6R anti- published.53,61 Satralizumab as an add-on to baseline treat- – body in patients with NMOSD (table 1).54 59 ment significantly reduced the risk of protocol-defined relapse by 62% vs placebo; hazard ratio (HR) 0.38; 95% CI Araki et al.56 reported that the mean (standard error of the 0.16–0.88; p = 0.02. Prespecified additional analyses showed mean [SEM]) ARR decreased from 2.9 (1.1) to 0.4 (0.8; a 79% relapse risk reduction with satralizumab vs placebo in p < 0.005) and mean EDSS scores (SEM, range) from 5.1 (1.7, AQP4-IgG–seropositive patients (n = 55; HR 0.21; 95% CI 3.0–6.5) to 4.1 (1.6, 2.0–6.0), following 12-month treatment 0.06–0.75) and 34% in AQP4-IgG–seronegative patients with an anti–IL-6R antibody as add-on therapy to either (HR 0.66; 95% CI 0.20–2.24). In the overall study pop- prednisolone or immunosuppressants in 7 patients with ulation, the ARR was 0.11 (95% CI 0.05–0.21) when treated NMOSD. In the same report, mean (SEM) pain decreased with satralizumab vs 0.32 (95% CI 0.19–0.51) with placebo. from 3.0 (1.5) at baseline to 0.9 (1.2) at 12 months, and mean There were no deaths or anaphylactic reactions in either (SEM) general fatigue decreased from 6.1 (2.0) at baseline to group during the double-blind period. The rate of serious 3.0 (1.4) at 12 months, respectively. Adverse events in this adverse events and infections did not differ between the report included upper respiratory infections (2 patients), acute satralizumab and placebo groups (table 3).53 The results of enterocolitis (2 patients), acute pyelonephritis (1 patient), the monotherapy trial were similar to the results of the add- leukocytopenia and/or lymphocytopenia (3 patients), anemia on therapy trial (table 3).61 These results further support an (2 patients), and a slight decline in systolic blood pressure (1 important role of IL-6 in NMOSD pathophysiology. patient). However, none of the events were severe. Ringelstein et al.54 reported that patients with NMOSD who Summary and future directions were not responsive to rituximab and other immunosuppressants experienced significant improvements in the me- NMOSDisrecognizedasaB-andT-cell–mediated immune dian ARR (from 4.0 at baseline to 0.4 at the end of the study, disease. IL-6 appears to have a key role in the immune re- p = 0.008), EDSS score (from 7.3 at baseline to 5.5 at the end sponse by stimulating antibody production in B cells and the of the study, p = 0.03), and pain scores (from 6.5 at baseline development of effector T cells, by disruption of BBB to 2.5 at the end of the study, p = 0.02) following treatment function, and by regulating the balance between Th17 and with an anti–IL-6R antibody. No major adverse events or Treg cells. Inhibition of IL-6 activity was shown to be ef- laboratory abnormalities were reported except for elevation fective in 2 phase 3 clinical trials, particularly in AQP4- of cholesterol levels in 6 patients and neutropenia in IgG–seropositive individuals, and represents a promising 1 patient.54 Similarly, Ayzenberg et al. reported that 3 emerging therapy for the management of NMOSD.

Table 3 Summary of phase 3, add-on therapy and monotherapy clinical trial of satralizumab in NMOSD

Product name Satralizumab

Mode of action IL-6 inhibition (anti–IL-6 receptor monoclonal antibody)

Study name SAkuraSky (NCT02028884) SAkuraStar (NCT02073279)

Study design Add-on therapy to baseline Monotherapy compared with placebo treatment; azathioprine, mycophenolate mofetil, and/or OCs compared with placebo

Every 4 wk, subcutaneously Every 4 wk, subcutaneously

Satralizumab Placebo Satralizumab Placebo Administration (n = 41) (n = 42) (n = 63) (n = 32)

Characteristics of the patient at baseline

Mean age, y (range) 40.8 ± 16.1 (13–73) 43.4 ± 12.0 (14–65) 45.3 ± 12.0 (21–70) 40.5 ± 10.5 (20–56)

Female sex, n (%) 37 (90) 40 (95) 46 (73) 31 (97)

AQP4-IgG–seropositive 27 (66) 28 (67) 41 (65) 23 (72) status, n (%)

Annualized relapse rate in 1.5 ± 0.5 1.4 ± 0.5 1.4 ± 0.6 1.5 ± 0.7 previous 2 y

Annualized relapse rate 0.11 (0.05–0.21) 0.32 (0.19–0.51) 0.17 (0.10–0.26) 0.41 (0.24–0.67) during the double-blind period

Relapse prevention efficacy (vs placebo)

AQP4-IgG (+ and 2) HR = 0.38 (0.16–0.88) HR = 0.45 (0.23–0.89)

AQP4-IgG (+) HR = 0.21 (0.06–0.75) HR = 0.26 (0.11–0.63),

AQP4-IgG (2) HR = 0.66 (0.20–2.24) HR = 1.19 (0.30–4.78)

AEs in the double-blind period Patients, Events (95% CI), Patients, Events (95% CI), Patients, n (%) Events (95% CI), Patients, n (%) Events (95% CI), (safety population) n(%) n/100 pt-y n(%) n/100 pt-y n/100 pt-y n/100 pt-y

Infection 28 (68) 132.5 26 (62) 149.6 (120.1–184.1) 34 (54) 99.8 (82.4–119.8) 14 (44) 162.6 (125.8–206.9) (108.2–160.5)

Serious infection 2 (5) 2.6 (0.3–9.2) 3 (7) 5.0 (1.0–14.7) 6 (10) 5.2 (1.9–11.3) 3 (9) 9.9 (2.7–25.2)

Injection-related reaction 5 (12) 21.7 (12.6–34.7) 2 (5) 3.4 (0.4–12.1) 8 (13) 13.9 (7.9–22.6) 5 (16) 17.3 (6.9–35.5)

Anaphylactic reaction 0 0 0 0 0 0 (NE–3.2) 0 0 (NE–9.1)

Abbreviations: AE = adverse event; AQP4-IgG = aquaporin-4-immunoglobulin G; HR = hazard ratio; IgG = immunoglobulin G; IL-6 = interleukin-6; NMOSD = neuromyelitis optica spectrum disorders; OC = oral corticosteroid; pt-y = patient-year. 9 Study funding This study was funded by Chugai Pharmaceutical Co, Ltd. Medical Appendix Authors

and medical writing support was provided by ApotheCom, UK, Name Location Contribution and was funded by Chugai Pharmaceutical Co, Ltd. Kazuo Department of Multiple Design and Fujihara, MD Sclerosis Therapeutics, conceptualization of the Disclosure Fukushima Medical review, manuscript University School of drafting, and critical K. Fujihara has received grants from the Ministry of Edu- Medicine and Multiple revision of the manuscript cation of Japan, Ministry of Health, Welfare and Labor of Sclerosis and for important intellectual Japan, Alexion Pharmaceuticals, Chemo-Sero-Therapeutic Neuromyelitis Optica content Center, Southern Research Institute, Teva, Teijin, and Genzyme Japan; re- TOHOKU Research ceived grants and personal fees from Chugai Pharmaceutical Institute for Neuroscience, Koriyama, Co,Ltd,Bayer,Biogen,MitsubishiTanabe,Ono,Nihon Japan Pharmaceutical, and Asahi Kasei; and received personal fees Jeffrey L. Departments of Design of the review and from Novartis, Merck Serono, MedImmune, Eisai, Astellas, Bennett, MD, Neurology and critical revision of the Takeda, Daiichi Sankyo, and Cosmic Corporation. J.L. PhD Ophthalmology, manuscript for important Programs in intellectual content Bennett holds patents for compositions and methods for the Neuroscience and treatment of neuromyeltis optica and novel blocking Immunology, School of Medicine, University of monoclonal therapy for neuromyelitis optica; has consulted Colorado, Aurora for Frequency Therapeutics, Roche, Viela Bio, AbbVie, Jerome de Department of Design of the review and Clene Nanomedicine, Alexion Pharmaceuticals, EMD- Seze, MD, PhD Neurology, Hopitalˆ de critical revision of the Serono, Chugai Pharmaceutical Co, Ltd, Reistone, and Hautepierre, Strasbourg manuscript for important Genentech; received research support from Mallinckrodt Cedex, France intellectual content Pharmaceuticals, NIH, Novartis, and Guthy-Jackson Foun- Masayuki Chugai Pharmaceutical Design of the review and dation; and receives license fee and royalty payments for Haramura, Co., Ltd, Tokyo, Japan critical revision of the PhD manuscript for important aquaporumab. J. de Seze has received grants and personal intellectual content fees from Roche; received personal fees from Chugai Phar- Ingo Kleiter, Department of Design of the review and maceutical Co, Ltd; and served on advisory boards and ex- MD Neurology, St. Josef critical revision of the pert committees for the clinical trial conducted by Chugai Hospital, Ruhr-University manuscript for important Bochum, Germany intellectual content Pharmaceutical Co, Ltd. M. Haramura is an employee of Chugai Pharmaceutical Co., Ltd. I. Kleiter has received Brian G. Department of Design of the review and Weinshenker, Neurology, Mayo Clinic, critical revision of the personal fees from Alexion Pharmaceuticals, Bayer, Biogen, MD Rochester, MN manuscript for important Celgene, IQVIA, Novartis, Merck, Mylan, Roche, Sanoﬁ intellectual content

Genzyme, grants and personal fees from Chugai Pharma- Delene Kang ApotheCom, London, UK Critical revision of the ceutical Co, Ltd, and grants from Diamed. B.G. Weinshenker manuscript for important holds a patent for and receives royalties from an NMO-IgG intellectual content for diagnosis of neuromyelitis optica from RSR Ltd, Oxford Tabasum ApotheCom, London, UK Critical revision of the University, Hospices Civil de Lyon, and MVZ Labor PD Dr. Mughal, PhD manuscript for important intellectual content Volkmann und Kollegen GbR.; is an adjudication committee member of MedImmune Pharmaceuticals and Alexion Takashi Department of Design of the review and Yamamura, Immunology, National critical revision of the Pharmaceuticals; consults for Chugai Pharmaceutical Co., MD, PhD Institute of Neuroscience, manuscript for important Ltd and Mitsubishi Tanabe; and served on the editorial National Center of intellectual content Neurology and Psychiatry, board of the Canadian Journal of Neurological Sciences and Tokyo, Japan Turkish Journal of Neurology.T.Yamamurahasservedon scientific advisory boards for Biogen, Takeda, Ono, Sumi- tomo Dainippon, Novartis, and Chugai Pharmaceutical Co, Ltd; received research grants from Chugai Pharmaceutical References Co, Ltd, Biogen, Novartis, Takeda, Teva, and Nihon Phar- 1. Tanaka T, Narazaki M, Kishimoto T. IL-6 in inflammation, immunity, and disease. Cold Spring Harb Perspect Biol 2014;6:a016295. maceutical; and received speaker honoraria from Chugai 2. Wingerchuk DM, Banwell B, Bennett JL, et al. International consensus diagnostic Pharmaceutical Co, Ltd, Ono, Takeda, Biogen, Sumitomo criteria for neuromyelitis optica spectrum disorders. Neurology 2015;85:177–189. 3. Zekeridou A, Lennon VA. Aquaporin-4 autoimmunity. Neurol Neuroimmunol Dainippon, Mitsubishi Tanabe, Human Metabolome Neuroinflamm 2015;2:e110. doi: 10.1212/NXI.0000000000000110. Technologies, Bayer, Japan Blood Products Organization, 4. Papadopoulos MC, Bennett JL, Verkman AS. Treatment of neuromyelitis optica: state-of-the-art and emerging therapies. Nat Rev Neurol 2014;10:493–506. Otsuka, Kissei, Novartis, and Daiichi Sankyo. Go to Neu- 5. Duan T, Smith AJ, Verkman AS. Complement-independent bystander injury in rology.org/NN for full disclosures. AQP4-IgG seropositive neuromyelitis optica produced by antibody-dependent cellular cytotoxicity. Acta Neuropathol Commun 2019;7:112. 6. Ataie-Kachoie P, Pourgholami MH, Morris DL. Inhibition of the IL-6 signaling Publication history pathway: a strategy to combat chronic inflammatory diseases and cancer. Cytokine – fl Growth Factor Rev 2013;24:163 173. Received by Neurology: Neuroimmunology & Neuroin ammation 7. Tanaka T, Kishimoto T. Targeting interleukin-6: all the way to treat autoimmune and April 1, 2020. Accepted in final form June 5, 2020. inflammatory diseases. Int J Biol Sci 2012;8:1227–1236.

10 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 Neurology.org/NN 8. Garbers C, Heink S, Korn T, Rose-John S. Interleukin-6: designing specific thera- 36. Matsushita T, Tateishi T, Isobe N, et al. Characteristic cerebrospinal fluid cytokine/ peutics for a complex cytokine. Nat Rev Drug Discov 2018;17:395–412. chemokine profiles in neuromyelitis optica, relapsing remitting or primary progressive 9. Erta M, Quintana A, Hidalgo J. Interleukin-6, a major cytokine in the central nervous multiple sclerosis. PLoS One 2013;8:e61835. system. Int J Biol Sci 2012;8:1254–1266. 37. Icoz S, Tuzun E, Kurtuncu M, et al. Enhanced IL-6 production in aquaporin-4 anti- 10. Feng Q, Wang YI, Yang Y. Neuroprotective effect of interleukin-6 in a rat model of body positive neuromyelitis optica patients. Int J Neurosci 2010;120:71–75. cerebral ischemia. Exp Ther Med 2015;9:1695–1701. 38. Barros PO, Cassano T, Hygino J, et al. Prediction of disease severity in neuromyelitis 11. Zhang J, Sadowska GB, Chen X, et al. Anti-IL-6 neutralizing antibody modulates optica by the levels of interleukin (IL)-6 produced during remission phase. Clin Exp blood-brain barrier function in the ovine fetus. FASEB J 2015;29:1739–1753. Immunol 2016;183:480–489. 12. Weinshenker BG, Wingerchuk DM. Neuromyelitis spectrum disorders. Mayo Clin 39. Uzawa A, Mori M, Ito M, et al. Markedly increased CSF interleukin-6 levels in Proc 2017;92:663–679. neuromyelitis optica, but not in multiple sclerosis. J Neurol 2009;256:2082–2084. 13. Uzawa A, Mori M, Muto M, et al. Benign neuromyelitis optica is rare in Japanese 40. Uzawa A, Mori M, Sawai S, et al. Cerebrospinal fluid interleukin-6 and glial fibrillary patients. Mult Scler 2015;21:1204–1208. acidic protein levels are increased during initial neuromyelitis optica attacks. Clin 14. Wingerchuk DM, Lennon VA, Lucchinetti CF, Pittock SJ, Weinshenker BG. The Chim Acta 2013;421:181–183. spectrum of neuromyelitis optica. Lancet Neurol 2007;6:805–815. 41. Uzawa A, Mori M, Sato Y, Masuda S, Kuwabara S. CSF interleukin-6 level predicts recovery 15. Kitley J, Leite MI, Nakashima I, et al. Prognostic factors and disease course in from neuromyelitis optica relapse. J Neurol Neurosurg Psychiatry 2012;83:339–340. aquaporin-4 antibody-positive patients with neuromyelitis optica spectrum disorder 42. Chihara N, Aranami T, Sato W, et al. Interleukin 6 signaling promotes anti-aquaporin from the United Kingdom and Japan. Brain 2012;135:1834–1849. 4 autoantibody production from plasmablasts in neuromyelitis optica. Proc Natl Acad 16. Mealy MA, Kessler RA, Rimler Z, et al. Mortality in neuromyelitis optica is strongly Sci USA 2011;108:3701–3706. associated with African ancestry. Neurol Neuroimmunol Neuroinflamm 2018;5:e468. 43. Kowarik MC, Astling D, Gasperi C, et al. CNS Aquaporin-4-specific B cells connect doi: 10.1212/NXI.0000000000000468. with multiple B-cell compartments in neuromyelitis optica spectrum disorder. Ann 17. Lennon VA, Wingerchuk DM, Kryzer TJ, et al. A serum autoantibody marker of Clin Transl Neurol 2017;4:369–380. neuromyelitis optica: distinction from multiple sclerosis. Lancet 2004;364: 44. Dienz O, Eaton SM, Bond JP, et al. The induction of antibody production by IL-6 is 2106–2112. indirectly mediated by IL-21 produced by CD4+ T cells. J Exp Med 2009;206:69–78. 18. Wingerchuk DM, Lennon VA, Pittock SJ, Lucchinetti CF, Weinshenker BG. Revised 45. Dos Passos GR, Sato DK, Becker J, Fujihara K. Th17 cells pathways in multiple diagnostic criteria for neuromyelitis optica. Neurology 2006;66:1485–1489. sclerosis and neuromyelitis optica spectrum disorders: pathophysiological and ther- 19. Kowarik MC, Dzieciatkowska M, Wemlinger S, et al. The cerebrospinal fluid im- apeutic implications. Mediators Inflamm 2016;2016:5314541. munoglobulin transcriptome and proteome in neuromyelitis optica reveals central 46. Wang HH, Dai YQ, Qiu W, et al. Interleukin-17-secreting T cells in neuromyelitis nervous system-specific B cell populations. J Neuroinflammation 2015;12:19. optica and multiple sclerosis during relapse. J Clin Neurosci 2011;18:1313–1317. 20. Mitsdoerffer M, Kuchroo V, Korn T. Immunology of neuromyelitis optica: a T cell-B 47. Varrin-Doyer M, Spencer CM, Schulze-Topphoff U, et al. Aquaporin 4-specific T cells cell collaboration. Ann N Y Acad Sci 2013;1283:57–66. in neuromyelitis optica exhibit a Th17 bias and recognize Clostridium ABC trans- 21. Day RE, Kitchen P, Owen DS, et al. Human aquaporins: regulators of transcellular porter. Ann Neurol 2012;72:53–64. water flow. Biochim Biophys Acta 2014;1840:1492–1506. 48. Takeshita Y, Obermeier B, Cotleur AC, et al. Effects of neuromyelitis optica-IgG at the 22. Jarius S, Franciotta D, Paul F, et al. Cerebrospinal fluid antibodies to aquaporin-4 in blood-brain barrier in vitro. Neurol Neuroimmunol Neuroinflamm 2017;4:e311. doi: neuromyelitis optica and related disorders: frequency, origin, and diagnostic rele- 10.1212/NXI.0000000000000311. vance. J Neuroinflammation 2010;7:52. 49. Fujihara K, Sato DK, Nakashima I, et al. Myelin oligodendrocyte glycoprotein im- 23. Bennett JL, Lam C, Kalluri SR, et al. Intrathecal pathogenic anti-aquaporin-4 anti- munoglobulin G-associated disease: an overview. Clin Exp Neuroimmunol 2018;9: bodies in early neuromyelitis optica. Ann Neurol 2009;66:617–629. 48–55. 24. Jarius S, Ruprecht K, Wildemann B, et al. Contrasting disease patterns in seropositive 50. Kaneko K, Sato DK, Nakashima I, et al. Myelin injury without astrocytopathy in and seronegative neuromyelitis optica: a multicentre study of 175 patients. neuroinflammatory disorders with MOG antibodies. J Neurol Neurosurg Psychiatry J Neuroinflammation 2012;9:14. 2016;87:1257–1259. 25. Jarius S, Ruprecht K, Kleiter I, et al. MOG-IgG in NMO and related disorders: a 51. Cree BAC, Bennett JL, Kim HJ, et al. Inebilizumab for the treatment of neuromyelitis multicenter study of 50 patients. Part 2: epidemiology, clinical presentation, radio- optica spectrum disorder (N-MOmentum): a double-blind, randomised placebo- logical and laboratory features, treatment responses, and long-term outcome. controlled phase 2/3 trial. Lancet 2019;394:1352–1363. J Neuroinflammation 2016;13:280. 52. Pittock SJ, Berthele A, Fujihara K, et al. Eculizumab in aquaporin-4-positive neuro- 26. Shimizu F, Schaller KL, Owens GP, et al. Glucose-regulated protein 78 autoantibody myelitis optica spectrum disorder. N Engl J Med 2019;381:614–625. associates with blood-brain barrier disruption in neuromyelitis optica. Sci Transl Med 53. Yamamura T, Kleiter I, Fujihara K, et al. Trial of satralizumab in neuromyelitis optica 2017;9:eaai9111. spectrum disorder. N Engl J Med 2019;381:2114–2124. 27. Nytrova P, Potlukova E, Kemlink D, et al. Complement activation in patients with 54. Ringelstein M, Ayzenberg I, Harmel J, et al. Long-term therapy with interleukin 6 neuromyelitis optica. J Neuroimmunol 2014;274:185–191. receptor blockade in highly active neuromyelitis optica spectrum disorder. JAMA 28. Tradtrantip L, Yao X, Su T, Smith AJ, Verkman AS. Bystander mechanism for Neurol 2015;72:756–763. complement-initiated early oligodendrocyte injury in neuromyelitis optica. Acta 55. Lauenstein AS, Stettner M, Kieseier BC, Lensch E. Treating neuromyelitis optica with Neuropathol 2017;134:35–44. the interleukin-6 receptor antagonist tocilizumab. BMJ Case Rep 2014;2014: 29. Yao X, Verkman AS. Complement regulator CD59 prevents peripheral organ injury in bcr2013202939. rats made seropositive for neuromyelitis optica immunoglobulin G. Acta Neuropathol 56. Araki M, Matsuoka T, Miyamoto K, et al. Efficacy of the anti-IL-6 receptor antibody Commun 2017;5:57. tocilizumab in neuromyelitis optica: a pilot study. Neurology 2014;82:1302–1306. 30. Zamvil SS, Spencer CM, Baranzini SE, Cree BAC. The gut microbiome in neuro- 57. Kieseier BC, Stuve O, Dehmel T, et al. Disease amelioration with tocilizumab in a myelitis optica. Neurotherapeutics 2018;15:92–101. treatment-resistant patient with neuromyelitis optica: implication for cellular immune 31. Omenetti S, Pizarro TT. The Treg/Th17 Axis: a dynamic balance regulated by the gut responses. JAMA Neurol 2013;70:390–393. microbiome. Front Immunol 2015;6:639. 58. Ayzenberg I, Kleiter I, Schroder A, et al. Interleukin 6 receptor blockade in patients 32. Uzawa A, Mori M, Masuda H, et al. Interleukin-6 analysis of 572 consecutive CSF with neuromyelitis optica nonresponsive to anti-CD20 therapy. JAMA Neurol 2013; samples from neurological disorders: a special focus on neuromyelitis optica. Clin 70:394–397. Chim Acta 2017;469:144–149. 59. Araki M, Aranami T, Matsuoka T, Nakamura M, Miyake S, Yamamura T. Clinical 33. Wang H, Wang K, Zhong X, et al. Notable increased cerebrospinal fluid levels of improvement in a patient with neuromyelitis optica following therapy with the anti- soluble interleukin-6 receptors in neuromyelitis optica. Neuroimmunomodulation IL-6 receptor monoclonal antibody tocilizumab. Mod Rheumatol 2013;23:827–831. 2012;19:304–308. 60. Harmel J, Ringelstein M, Ingwersen J, et al. Interferon-beta-related tumefactive brain 34. Uzawa A, Mori M, Arai K, et al. Cytokine and chemokine profiles in neuromyelitis lesion in a Caucasian patient with neuromyelitis optica and clinical stabilization with optica: significance of interleukin-6. Mult Scler 2010;16:1443–1452. tocilizumab. BMC Neurol 2014;14:247. 35. Uchida T, Mori M, Uzawa A, et al. Increased cerebrospinal fluid metalloproteinase-2 61. Traboulsee A, Greenberg B , Bennett JL, et al. Safety and efficacy of satralizumab and interleukin-6 are associated with albumin quotient in neuromyelitis optica: their monotherapy in neuromyelitis optica spectrum disorder: a randomised, double-blind, possible role on blood-brain barrier disruption. Mult Scler 2017;23:1072–1084. multicentre, placebo-controlled phase 3 trial. Lancet Neurol 2020;19:402–412.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 5 | September 2020 11 CORRECTION Guillain-Barr´e syndrome related to SARS-CoV-2 infection Neurol Neuroimmunol Neuroinﬂamm 2020;7:e850. doi:10.1212/NXI.0000000000000850

In the Article “Guillain-Barr´e syndrome related to SARS-CoV-2 infection” by K. Bigaut et al.,1 there are several errors in table 1. The age noted for the patient described in the ﬁrst case should be 43; additionally, the protein level in CSF should be listed as 0.94 g/L for this patient. Lastly, for the patient described in the second case, the protein level in CSF should be listed as 1.06 g/L. The authors regret the errors.

Reference 1. Bigaut K, Mallaret M, Baloglu S, et al. Guillain-Barr´e syndrome related to SARS-CoV-2 infection. Neurol Neuroimmunol Neuroinﬂamm 2020;7:e785. doi:10.1212/NXI.0000000000000785.