Anti-NMDAR Encephalitis: a Single-Center, Longitudinal Study in China E633

Volume 7, Number 1, January 2020 Neurology.org/NN

A peer-reviewed clinical and translational neurology open access journal

ARTICLE Anti-NMDAR encephalitis: A single-center, longitudinal study in China e633

ARTICLE Antibodies to neurofascin, contactin-1, and Caspr1 in CIDP: Clinical relevance of IgG isotype e639

ARTICLE Neurologic complications of acute hepatitis E virus infection e643

VIEWS & REVIEWS An argument for broad use of highly effi cacy treatments in early multiple sclerosis e636 TABLE OF CONTENTS Volume 7, Number 1, January 2020 Neurology.org/NN

e637 Minimal breast milk transfer of rituximab, a monoclonal antibody used in neurological conditions K.M. Krysko, S.C. LaHue, A. Anderson, A. Rutatangwa, W.Rowles,R.D.Schubert,J.Marcus,C.S.Riley,C.Bevan, T.W. Hale, and R. Bove Open Access

e639 Antibodies to neurofascin, contactin-1, and contactin- associated protein 1 in CIDP: Clinical relevance of Cover Image IgG isotype Diagnostic work-up and treatment regimes in an alemtuzumab-treated A. Cortese, R. Lombardi, C. Briani, I. Callegari, L. Benedetti, patient with RRMS developing idiopathic multicentric Castleman’s F. Manganelli, M. Luigetti, S. Ferrari, A.M. Clerici, G.A. Marfia, disease (Panels I and J shown). A. Rigamonti, M. Carpo, R. Fazio, M. Corbo, A. Mazzeo, “(I) H&E staining of bone marrow puncture displaying F. Giannini, G. Cosentino, E. Zardini, R. Curro,` M. Gastaldi, E. Vegezzi, E. Alfonsi, A. Berardinelli, L. Kouton, C. Manso, megakaryocytosis. (J) Mediastinal lymph node biopsy consistent with C. Giannotta, P. Doneddu, P. Dacci, L. Piccolo, M. Ruiz, iMCD plasma-cell type. (J) H&E staining: regressive germinal centers A. Salvalaggio, C. De Michelis, E. Spina, A. Topa, G. Bisogni, (blue arrows), with small vessels reaching into germinal centers (“lollipop A. Romano, S. Mariotto, G. Mataluni, F. Cerri, C. Stancanelli, ” ” M. Sabatelli, A. Schenone, E. Marchioni, G. Lauria, vessels , black arrows). E. Nobile-Orazio, J. Devaux, and D. Franciotta See page e638 Open Access ’ Editor s Corner e640 Neuromyelitis optica spectrum disorders with unevenly clustered attack occurrence e644 N2 year in review T. Akaishi, I. Nakashima, T. Takahashi, M. Abe, T. Ishii, and M. Aoki J. Dalmau, M.C. Dalakas, D.L. Kolson, F. Paul, and S.S. Zamvil Open Access Open Access

Editorial e643 Neurologic complications of acute hepatitis E virus infection ﬁ e646 Treating muscle-speci c kinase myasthenia gravis P. Ripellino, E. Pasi, G. Melli, C. Staedler, M. Fraga, D. Moradpour, from the inside out R. Sahli, V. Aubert, G. Martinetti, F. Bihl, E. Bernasconi, B. Terziroli Beretta-Piccoli, A. Cerny, H.R. Dalton, C. Zehnder, M.G. Huijbers and J.J.G.M. Verschuuren B. Mathis, C. Zecca, G. Disanto, A. Kaelin-Lang, and C. Gobbi Open Access Companion article, p. e645 Open Access

Articles e645 SHP2 inhibitor protects AChRs from effects of e633 Anti-NMDAR encephalitis: A single-center, myasthenia gravis MuSK antibody longitudinal study in China S. Huda, M. Cao, A. De Rosa, M. Woodhall, P.M. Rodriguez Cruz, J. Cossins, M. Maestri, R. Ricciardi, A. Evoli, D. Beeson, and A. Vincent X. Xu, Q. Lu, Y. Huang, S. Fan, L. Zhou, J. Yuan, X. Yang, H. Ren, D. Sun, Y. Dai, H. Zhu, Y. Jiang, Y. Zhu, B. Peng, L. Cui, and H. Guan Open Access Editorial, p. e646 Open Access Clinical/Scientific Notes e634 Black African and Latino/a identity correlates with e638 A case of idiopathic multicentric Castleman disease in increased plasmablasts in MS an alemtuzumab-treated patient with MS K.M. Telesford, U.W. Kaunzner, J. Perumal, S.A. Gauthier, X. Wu, I. Diaz, M. Kruse-Hoyer, C. Engel, M. Marcille, and T. Vartanian L. Rolfes, S. Pfeuffer, T. Ruck, S. Windhagen, I. Oschlies, H.-J. Pavenstädt, L. Angenendt, H. Wiendl, J. Krämer, and Open Access S.G. Meuth e635 Lymphocyte pharmacodynamics are not associated Open Access with autoimmunity or efficacy after alemtuzumab e641 West Nile virus-associated vasculitis and intracranial H. Wiendl, M. Carraro, G. Comi, G. Izquierdo, H.J. Kim, B. Sharrack, C. Tornatore, N. Daizadeh, L. Chung, A.K. Jacobs, R.J. Hogan, hemorrhage L.V. Wychowski, and B. Van Wijmeersch, on behalf of the CARE-MS I, A. Harroud, A. Almutlaq, D. Pellerin, D. Paz, G.J. Linnell, and CARE-MS II, and CAMMS03409 Investigators D. Gendron Open Access Open Access

Continued TABLE OF CONTENTS Volume 7, Number 1, January 2020 Neurology.org/NN

e642 An inflammatory milieu: Optic perineuritis, Corrections retroperitoneal fibrosis, and giant cell arteritis e647 Identification of circulating MOG-specific B cells in D.M. Gold and S.L. Galetta patients with MOG antibodies Open Access e648 Cholecalciferol in relapsing-remitting MS: A Views & Reviews randomized clinical trial (CHOLINE) 358 ffi e650 IL6 receptor Ala variant and trans-signaling are e636 An argument for broad use of high e cacy treatments fi in early multiple sclerosis disease modi ers in amyotrophic lateral sclerosis J.M. Stankiewicz and H.L. Weiner e657 A case of idiopathic multicentric Castleman disease in Open Access an alemtuzumab-treated patient with MS 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 Executive Office, American Academy of Neurology Copyright and Permission Information: Please go to the journal website Catherine M. Rydell, 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. Kathleen M. Pieper, Senior Managing Editor, Neurology Advertising Sales Representatives: Wolters Kluwer, 333 Seventh Avenue, New York, NY 10001. Contacts: Eileen Henry, tel: 732-778-2261, fax: 973-215-2485, Sharon L. Quimby, Managing Editor, Neurology® Clinical Practice [email protected] and in Europe: Craig Silver, tel: +44 7855 Morgan S. Sorenson, Managing Editor, Neurology® Neuroimmunology 062 550 or e-mail: [email protected]. & Neuroinflammation Careers & Events: Monique McLaughlin, Wolters Kluwer, Two Commerce Square, 2001 Market Street, Philadelphia, PA 19103, tel: (215)-521- 8468, fax: Lee Ann Kleffman, Managing Editor, Neurology® Genetics 215-521-8801; [email protected]. Andrea Rahkola, Production Editor, Neurology Reprints: Meredith Edelman, Commercial Reprint Sales, Wolters Kluwer, Two Commerce Square, 2001 Market Street, Philadelphia, PA 19103, tel: Robert J. Witherow, Senior Editorial Associate 215-356-2721; [email protected]; reprintsolutions@ Karen Skaja, Senior Editorial Associate wolterskluwer.com. Kaitlyn Aman Ramm, Editorial Assistant Special projects: US & Canada: Alan Moore, Wolters Kluwer, Two Commerce Square, 2001 Market Street, Philadelphia, PA 19103, tel: 215-521-8638, Kristen Swendsrud, Editorial Assistant [email protected]. International: Andrew Wible, Senior Manager, Rights, Licensing, and Partnerships, Wolters Kluwer, translationrights@ Justin Daugherty, Editorial Assistant wolterskluwer.com. Madeleine Sendek, MPH, Editorial Assistant Rachel A. Anderson, Administrative Assistant

Publisher Wolters Kluwer Baltimore, MD

Publishing Staff Kim Jansen, Executive Publisher Jessica Heise, Production Team Leader, Neurology Journals Megen Miller, Production Editor Stacy Drossner, Production Associate

Neurology® Neuroimmunology & Neuroinflammation

Editor Neurology Neuroimmunology & JosepO.Dalmau,MD,PhD Vision ® 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

Neurology® Journals

Editor-in-Chief Classification of Evidence Review Team Robert A. Gross, MD, PhD, FAAN Melissa J. Armstrong, MD Richard L. Barbano, MD, PhD, FAAN Deputy Editor Richard M. Dubinsky, MD, MPH, FAAN Bradford B. Worrall, MD, MSc, FAAN 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, and Inclusion (EDI) RoyH.Hamilton,MD,MS,FAAN Holly E. Hinson, MD, MCR, FAAN

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

Ombudsman David S. Knopman, MD, FAAN

Scientiﬁc Integrity Advisor Robert B. Daroﬀ, MD, FAAN

Josep Dalmau, MD, PhD, FAAN, Marinos C. Dalakas, MD, FAAN, Dennis L. Kolson, MD, PhD, Friedemann Paul, MD and Scott S. Zamvil, MD, PhD N2 year in review

Neurol Neuroimmunol Neuroinﬂamm January 2020 vol. 7 no. 1 e644. doi:10.1212/NXI.0000000000000644e644

2019 has been a very good year for our journal. N2 was included in MEDLINE and in June received its first impact factor, a competitive 7.353. We have been fortunate to recruit Dr. Marinos Dalakas to our team of International Associate Editors. Dr. Dalakas is Professor of Neurology at the University of Athens (Greece) and Jefferson University (Philadelphia, PA). His remarkable knowledge on inflammatory and autoimmune neuromuscular diseases will be a great asset to the journal. These accomplishments and successes would not have been possible without the growing number of readers, the increasing number of manuscript submissions, and the generosity of our reviewers who freely give their time to the journal. To all, we send our appreciation.

In the 2018 N2 Year in Review,1 I wrote some comments related to how the immune checkpoint inhibitors (ICIs) had revolutionized the treatment of cancer. I noted the concern of several investigators about the immune-related adverse effects (irAEs) of these drugs, potentially leading to an increase in patients with autoimmune or paraneoplastic neurologic syndromes. This concern still exists, although a recent review indicated a relatively low number of irAEs that fulfilled criteria of paraneoplastic syndromes, including the presence of immune responses specifically directed against antigens expressed by the tumor and the nervous system (onconeuronal antigens).2 Indeed, the authors identified only 14 reported cases (2 with Ma2 antibody–associated syndromes) that fulfilled these criteria. Much more frequent, however, were neurologic irAEs unrelated to these mechanisms (e.g., without onconeuronal antibodies) and mediated by other inflammatory or autoimmune responses, including polyneuropathy, Guillain-Barré syndrome, myasthenia gravis, aseptic meningitis, myelitis, or myositis.2,3

During this past year, several studies on ICIs have shown “adverse effects” (e.g., facilitating the occurrence of paraneoplastic syndromes), whereas other studies have shown “beneficial effects” such as the use of ICIs as potential treatments for progressive multifocal leukoencephalopathy (PML). In the November issue of N2, Vogrig et al.4 retrospectively reviewed a cohort of 50 patients with Ma2 antibody–associated paraneoplastic syndromes and identified 6 who developed the syndrome after treatment with ICIs. None of these 6 patients had seminoma or testicular germ cell tumors, which were found in 25% of the rest of the cohort. The authors did not find differences between the neurologic features of the ICI-associated cases compared with the other anti-Ma2 cases. All 6 patients were treated with steroids and removal of the ICI, and some received plasmapheresis or rituximab. Four of the patients died (3 from the neurologic disease and associated complications), and the other 2 had moderate to severe disability. During the 12-month period in which the 6 patients with ICI-associated anti-Ma2 syndromes were MORE ONLINE identified, a total of 17 patients were diagnosed with anti-Ma2 syndrome. Before this, the annual number of patients with anti-Ma2 syndromes diagnosed in this reference center was relatively Editor Summary stable with a median of 4 cases per year. Although the reason for the overall increase in the NPub.org/N2/edsum number of patients with anti-Ma2 syndrome is unclear, it is remarkable that almost 1/3 had

From the ICREA-IDIBAPS Hospital Cl´ınic, University of Barcelona (J.D.), Spain; University of Pennsylvania (J.D., D.L.K.), Philadelphia; University of Athens Medical School (M.C.D.), Athens, Greece; Jefferson University (M.C.D.), Philadelphia, PA; Charite University Hospital (F.P.), Berlin, Germany; and Department of Neurology (S.S.Z.), University of California, San Francisco.

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 © 2019 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 received ICIs. Overall, the finding suggested that we should rigidity and myoclonus (PERM) in only 3. Among the remain vigilant for an increase of other paraneoplastic neuro- remaining 3 patients, 1 had global fatigable weakness with logic syndromes as ICIs are increasingly available for a wide sustained clonus, another had laryngeal dystonia, and the other variety of cancers. Given that low titers of paraneoplastic had hemiballismus with tics.14 Of interest, the frequency of the antibodies (e.g., Hu, Ri, or Yo, among others) occur in some symptoms in this study is actually different from that reported – patients with cancer without paraneoplastic syndromes,5 7 in a review by the same authors of 187 previously reported several investigators have suggested that testing for these patients, in which 48% had PERM, 22% epilepsy, and the antibodies before treating patients with ICIs may help to remaining 30% mixed phenotypes (cerebellar ataxia, move- identify those that are at risk of developing paraneoplastic ment disorders, demyelination, encephalitis, and cognitive syndromes. This can be particularly useful for patients with dysfunction).14 To understand the discrepancies of these 2 tumors that have propensity to associate with paraneoplastic studies,13,14 we need to examine the features they have in syndromes, such as small-cell lung cancer.2 common. For example, neither of them examined systematically patients’ CSF; in one study, the CSF was not examined,13 PML is a severe disease of the brain that usually affects and in the other, only 3 of 14 patients had the CSF examined (2 immunosuppressed patients and results from reactivation of of them without antibodies).14 Moreover, as soon as the GlyR the polyomavirus JC (JCV) and infection of oligodendrocytes antibodies were identified in the patients, both studies took at and, to a lesser extent, astrocytes. Because there are no specific face value that the symptoms were linked to the antibodies. anti-JCV drugs, the only treatment strategy is to restore the There was little consideration for the fact that GlyR antibodies function of the immunologic system. Although ICIs were ini- can be detected in serum of patients with many different dis- tially designed to enhance tumor cell–specific responses, orders (e.g., AQP4 or MOG autoimmunity, MS, patients with anti–program cell death-1 monoclonal antibodies (PD1-abs) cancer without neurologic diseases, opsoclonus-myoclonus, or – have recently been used to treat PML. In the first group of cerebellar ataxia).15 17 Thus, it is not surprising that if enough publications that included a total of 10 patients treated with patients with the same disorder are investigated, a small per- a PD1-ab, 7 showed mild to substantial improvement, 1 who centage of seropositive cases will be identified. For example, in was stable before initiating the treatment remained stable, and another series of 238 patients with epilepsy, 13 (5%) had GlyR – 2 deteriorated.8 10 Two additional patients were recently antibodies in serum.18 reported in N2, and both showed no improvement with PD1- ab. One was considered to be a good candidate for the drug These considerations are applicable to most autoimmune en- (young age, with a primary immunodeficiency syndrome, not cephalitides associated with neuronal surface antibodies. In- previously treated with immunosuppression, limited MRI deed, bearing in mind that (1) antibodies against neuronal lesions, and low CSF viral load), was treated with PD1-ab at cell-surface antigens associate with autoimmune encephalitis very early stage of the PML, and received more doses than or myelitis, (2) patients may harbor antibodies only in CSF, those reported in patients who improved, but had relentless and (3) CSF neuronal antibodies are more disease specific progression to death.11 The second case had several immu- than serum antibodies, it is remarkable that in 2019, there are nosuppressive disorders, a very high viral load, and died after 2 still studies on autoimmune encephalitis in which CSF testing infusions of the PD1-ab. An interesting observation was that is not considered. The lack of comprehensive studies (e.g., this patient had fewer progenitor-exhausted T cells and more study of serum and CSF and use of appropriate controls) terminally exhausted T cells compared with a previously interferes with the assessment of true associations between reported patient who had a favorable outcome.12 Thus, the antibodies and diseases, promotes uncertainty (any antibody authors postulated that patients with PML deprived of can cause any symptom), and may have important con- progenitor-exhausted T cells (which have the ability to revert sequences in the diagnosis and treatment of patients. These from exhausted to active) may be those who do not respond to may include not treating patients who should be treated (e.g., PD1-ab. In addition, preliminary data from both patients patients in whom antibodies are present in CSF but missed suggested that JCV-specific CD4+ T cells may be more im- with serum only testing19) or using immunotherapy in patients portant than CD8+ T cells in keeping JCV on check.12 who may not have an autoimmune disease (e.g., patients with antibodies only in serum and unrelated neurologic symptoms). Moving to a different topic, a study of a series of 17 patients This is reminiscent of the history of VGKC complex antibodies with serum glycine receptor (GlyR) antibodies was published and its ever increasing “expansion of the phenotype” due to this year in N2. Thirteen of these patients developed stiff- resistance to accept that their utility was more than question- person syndrome accompanied by parkinsonism or cerebellar able, leading to frequent misdiagnosis and exposure of patients signs. In addition, 10 patients had various visual symptoms to unwarranted therapies. (spider web-like images, palinopsia, photophobia, hallucina- tions, and synesthesia, among others), and another 3 presented In recent years, several autoimmune, multisystemic, or with primarily autoimmune epilepsy with psychiatric symp- fibroinflammatory disorders have been identified based on the toms.13 These findings are in contrast with a series of 14 presence of antigen-specific autoantibodies of the IgG4 sub- patients in which the most prevalent symptoms were seizures class.20 In Neurology, IgG4 antibodies are largely represented and epilepsy in 8 cases and progressive encephalomyelitis with by 2 immunologically distinct neuromuscular diseases, MuSK

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN antibody–positive myasthenia gravis (MuSK-MG) and useful observation that emerged from this study is that IgG4 antinodal/paranodal antibody-mediated CIDP. These dis- antinodal/paranodal antibodies were not only restricted to orders are clinically important because they exhibit poor patients with the originally described atypical CIDP response to IVIg or plasmapheresis, and they are immuno- phenotype,24,25 but they are relevant to all patients with CIDP logically unique because IgG4 antibodies cannot bind com- fulfilling the EFNS/PNS criteria. The same message was re- plement or many Fc receptors on immune cells, and they are iterated in the September issue by Carrera-Garcia et al.26 who able to exchange Fab-arms with other IgG4 molecules. As reported the first child with relapsing CIDP and IgG-4 anti- a result, IgG4 antibodies are viewed as functionally bispecific bodies against contactin-1, poorly responding to IVIg. Both and monovalent, unable to engage in antigen cross-linking and studies strengthen the view that testing for antibodies against internalization. Although their immunopathogenicity is not nodo-paranodal proteins is needed for all adults and pediatric completely understood, there is convincing evidence, mostly patients with CIDP refractory to conventional therapies. from autoimmune neuromuscular diseases studies, that IgG4 antibodies function by blocking enzymatic activity or protein- The complexity of nodal/paranodal antibodies was further protein interactions of their target antigens. The past year, the highlighted in 2 patients with autoantibodies against 2 different field has further advanced with new information regarding the neurofascin (NF) isoforms.27 In contrast to the aforemen- association of Ig4 antibodies with distinct neurologic pheno- tioned characteristic phenotype associated with antibodies types and their effect on antigenic targets. against the paranodal NF-155–specific Fn3Fn4 domain, a different phenotype was identified in patients who had IgG3 Huijbers et al.21 in the May issue provide new insights on the antibodies against all the NF isoforms. The anti–pan-NF- function of IgG4 antibodies in MuSK-MG. They generated associated CIDP was characterized by a more aggressive monovalent Fab fragments from patient-derived recombinant course, tetraplegia and cranial nerve involvement indicating IgG4-MuSK antibodies and investigated the functional effects that IgG3 antibodies directed against both NF epitopes, the of bispecificity and monovalency of Fab-arm–exchanged paranodal NF-155 and the nodal NF-140/186, define a differ- antibodies in a tissue culture model. They report that ent clinical phenotype. In contrast to IgG4, IgG3 autoantibodies recombinant monovalent MuSK IgG4 engages in bivalent activate complement and have a strong proinflammatory effect, monospecific antibody-antigen interactions. Although the hence their association with more severe disease. Although the monovalent anti-MuSK blocks MuSK signaling and AChR underlying mechanism of multiple epitopes remains unclear, clustering, the bivalent anti-MuSK stimulates MuSK phos- intramolecular epitope spreading could be an explanation, as phorylation and partially induces AChR clustering. It seems seen in other autoimmune diseases. Collectively, these studies that the IgG4 MuSK antibodies require Fab-arm exchange of strengthen the clinical importance of antinodal/paranodal IgG4 to be more functionally monovalent to become patho- proteins not only in defining phenotypes but also in highlight- genic; in other words, depending on the number of MuSK ing that these patients have a severe disease poorly responding binding sites, MuSK antibodies can either act as MuSK agonist to IVIg and plasmapheresis necessitating early intervention with or antagonist. The work has practical implications in antibody- more effective therapies aiming at downregulating the humoral mediated autoimmunity because inhibition of Fab-arm ex- immune response. Rituximab is currently the preferred agent for change might have therapeutic potential not only in MuSK all IgG4-related diseases,20,28 and it is likely effective by de- MG but also in other IgG4-mediated autoimmune disorders. pleting the Nfasc155, Nfasc140/186, CNTN1, Caspr1, and MuSK-reactive B cells. A breakthrough in CIDP autoimmunity has been the remarkable observation that some patients, especially with an atypical At least 12 other autoimmune multisystemic or lymphoproli- CIDP phenotype highlighted by severe subacute neuropathy, ferative diseases are hallmarked by the prototypic IgG4-related tremor, and sensory ataxia, do not respond to IVIg or plasma- syndrome (IgG4-RD).20 Neurologic manifestations can be pheresis and have IgG4 antibodies to nodal/paranodal antigens seen as part of the multiorgan fibroinflammatory involvement, directed against neurofascin-155 (Nfasc155), neurofascin-140/ mostly represented by meningeal and spinal cord disease and 186 (Nfasc140/186), contactin-1 (CNTN1), and contactin- often presented as hypertrophic pachymeningitis and hypo- associated protein-1 (Caspr1).22 These major observations physitis. In this context, Levraut et al.29 in the July issue report were now confirmed in a larger multicenter study providing on 2 cases of hypertrophic pachymeningitis and suggest that further data on the IgG4 phenotypes and their pathogenicity. PET imaging increased intrathecal IgG4 levels, and histo- Cortese et al.23 in the January issue found the incidence of pathologic studies are essential in arriving at proper diagnosis these antibodies in 5.5% among 342 Italian patients with and early therapy initiation. CIDP; of those, 9 had antibodies against the paranodal Nfasc155, 1 against both the nodal Nfasc140/186 and the In persons living with HIV (PLWH), the persistence of the paranodal Nfasc155, 3 against CNTN1, and 6 against Caspr1. virus within the CNS and in lymphoid tissues continues to be Anti-Caspr1 IgG4 was shown to penetrate paranodal regions a major challenge in the effort to “cure” HIV infection and to and disrupt the integrity of the Nfasc155/CNTN1/Caspr1 prevent complications such as HIV-associated neurocognitive complex, consistent with the effect of IgG4 in disrupting disorders.30 Although suppressive antiretroviral therapy – protein-protein interactions of the targeted antigens.21 23 A (ART) effectively prevents the development of CNS

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 3 opportunistic infections and fulminant HIV encephalitis, it neurocognitive disorders, Gisslen et al.37 (January 2019 N2 does not fully protect against the development of neuro- issue) have correlated CSF levels of soluble TREM2 (a specific cognitive dysfunction.31 Therefore, neuroprotective therapies macrophage/microglia activation marker) with CSF levels of to use in conjunction with ART are being sought, as are plasma NFL in PLWH. Archived CSF samples from 112 adult PLWH and CSF biomarkers that might identify patients at risk of HIV- and 11 HIV-negative controls (all collected between 1999 and driven neurocognitive dysfunction. 2014, irrespective of neurocognitive status and ART suppression) were analyzed for sTREM2, neopterin (a marker of ac- In the May 2019 issue of N2, a review by Ambrosius et al.32 tivation of macrophages, microglia, and astrocytes), and NFL. discusses the potential use for anti-neuroinflammatory drugs as CSF sTREM2 levels correlated strongly with neopterin and adjunctive agents for treatment and/or prevention of HIV- even more strongly with NFL. The correlation of CSF sTREM associated neurocognitive disorders in PLWH. Evidence sug- with severe neurocognitive dysfunction, seen typically in un- gests that even during effective ART suppression of HIV controlled CNS HIV infection and not in ART-suppressed infection (i.e., undetectable plasma and CSF HIV RNA), im- infection, suggests that it will not be a sensitive biomarker for mune activation, oxidative stress, and inflammation persist to neurocognitive dysfunction in ART-suppressed patients. Fur- some degree within the CNS and also within peripheral tissue thermore, although 36% of patients on suppressive ART had compartments.33 In this sense, the neuropathogenic mecha- elevated CSF neopterin compared with controls, none of those nisms that characterize virally suppressed HIV infection of the patients had elevated sTREM2. Because TREM2 is considered CNS resemble those of MS, and drugs that target these specific to cells of monocyte lineage, the investigators con- pathways and that can be used in conjunction with ART are cluded that a significant component of the residual CNS in- being investigated. Furthermore, such drugs might be expected flammation present in ART-suppressed PLWH may result to have beneficial effects not only within the CNS but also from activation of cells (astrocytes and lymphocytes) other within other tissue compartments (e.g., cardiovascular system) than macrophages and microglia. This study has specific value affected by immune activation, oxidative stress, and in- in providing evidence for multiple cellular contributors to flammation in PLWH. Among those MS drugs discussed by chronic CNS inflammation in PLWH, and it suggests that Ambrosius et al. as potential candidates are several developed assessing multiple CSF biomarkers (sTREM2, neopterin, for the treatment of MS, including dimethyl fumarate (anti- NFL, and others) may be necessary for accurately profiling inflammatory, antioxidant; use in PLWH also discussed in disease progression, clinical risk, and response to neuro- reference 34), fingolimod (inhibitor of T-lymphocyte traf- protective therapies. ficking), teriflunomide (inhibitor of T- and B-lymphocyte proliferation), and natalizumab (inhibitor of T-lymphocyte Finally, in the March 2019 issue of N2, Kamtchum-Tatuene trafficking). Each of these drugs indeed targets critical points in et al.38 presented an interesting study of ischemic stroke risk in the HIV neuropathogenesis cycle, including the trafficking of PLWH in sub-Saharan Africa, which linked elevated plasma infected T lymphocytes out of lymph nodes and across the levels of ICAM-1 (activation of endothelia) with HIV infection blood-brain barrier, proinflammatory signaling (NF-kB and independently of stroke and other risk factors (ART use, di- others), and production of reactive oxygen species, among abetes, and hypercholesterolemia). Because this study involved others. The ability of fingolimod to retain antiviral only 61 stroke cases (19 PLWH) and 168 nonstroke controls T lymphocytes in lymphoid tissues in the simian immunode- (32 PLWH), it might be underpowered to detect the expected ficiency virus (SIV) rhesus macaque model of HIV patho- association between HIV and stroke risk, but the detection of genesis has generated interest in its potential to limit SIV/HIV significantly elevated plasma ICAM-1 in individuals receiving persistence in situ.35 Additional studies are certain to follow. ART suggests persistent endothelial activation in ART-treated PLWH. However, this study adds to growing evidence sup- However, resistance to the use of such immunomodulating porting the potential pathogenic effects of chronic in- and/or immunosuppressive MS drugs associated with lym- flammation and activation of multiple CNS-relevant cell phocytopenia for treatment in PLWH appears to be common lineages (endothelia, macrophages, microglia, and astrocytes) among neurologists, and concerns about possible complica- in ART-suppressed PLWH, and it also supports the impor- tions deserve robust discussion. It should be noted, however, tance of studying of HIV-associated CNS disease risks as that PLWH who are virally suppressed by ART sustain CD4+ a worldwide issue. T-lymphocyte counts with long-standing immunosuppressive therapy after kidney and liver transplantation, without in- In 2019, several articles published in N2 have made relevant creased risk of opportunistic infections or premature death contributions to the areas of pathogenesis, biomarkers, and (reviewed in Ref. 36). As chronic inflammation in PLWH and treatment of MS. Individual MS risk is influenced both by associated end-organ effects become more aggressively tar- genetic susceptibility and environmental factors, such as EBV – geted, one may anticipate that immunomodulating MS drugs infection, low vitamin D, smoking, obesity, and others.39 42 will receive more and more attention for use in PLWH. Recently, alterations of the gut microbiome in MS through dietary habits have received increasing attention as possible Attending to the need for identification of associative and link between potentially modifiable environmental factors and – predictive biomarkers for the development of HIV-associated the immune system.43 46 However, previous human studies

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN were limited by small sample sizes, enrollment of patients with hypointense lesions on brain MRI, and EDSS progression) longer disease duration, and confounding effects of immuno- significantly favored the active intervention, whereas there was modulatory therapy, thus precluding conclusions regarding the no difference between both completer groups with regard to new causal influence of the gut microbiome on the MS immune T2 and gadolinium-enhancing lesions and brain gray and white system, or in other words, leaving the “chicken or egg di- matter volumes. The rate of adverse events was similar between lemma” unresolved.47 Katz Sand et al.48 investigated in a cross- groups. Although the study was negative in respect to the primary sectional study the effects of 2 widely used disease-modifying end point, presumably due to lack of statistical power, and is drugs, glatiramer acetate (GA) and dimethyl fumarate (DMF), therefore not able to provide an unambiguous answer to the on gut microbial composition. Stool samples from 168 par- question whether high-dose vitamin D supplementation bene- ticipants with MS from 2 MS centers (75 treatment naive, 33 ficially modifies MS disease course, it backs current clinical on DMF, and 60 on GA) were collected, and 16S rRNA management now adopted by many MS neurologists who sup- amplicon sequencing was performed in parallel with immu- plement low vitamin D levels in their patients. nophenotyping from patients’ whole blood (at 1 center only) to validate the expected effects of DMF and GA. Both drugs The year 2019 in N2 has also provided us with some new were associated with alterations of the fecal microbiota com- insights into biomarkers to monitor disease course and im- position, namely a decreased relative abundance of the Lach- munotherapy in MS. Brain atrophy that is believed to reflect at nospiraceae and Veillonellaceae families. Moreover, in patients least in part also the neurodegenerative component of MS and treated with DMF, there was a decreased relative abundance of is detectable on a group level from earliest disease stages56 has the phyla Firmicutes and Fusobacteria and the order Clos- long been proposed as marker to measure progressive tissue tridiales and an increase in the phylum Bacteroidetes. Both loss over the course of the disease. Azevedo et al.57 in a longi- drugs differentially affected metabolic pathways with some tudinal study have now investigated the effect of normal aging overlap. This study demonstrates that DMDs may have on brain atrophy in MS. Brain MRIs from 520 patients with a profound impact on the gut microbiome in MS, which has to relapse onset MS and from 130 healthy controls, most of them be taken into account for future studies. with more than 1 measurement time point, were investigated. The rate of whole-brain atrophy attributable to MS changed Two other studies have dealt with therapeutic modulation of significantly with age and decreased from −0.38% per year at environmental factors in MS. Modulation of diet was proposed age 30 years to −0.12% per year at age 60 years, whereas the to have beneficial impact on tissue damage and disease severity slope of normal aging atrophy increased from 0.01% per year at in animal models of MS.49,50 Brenton et al.51 have now con- age 30 years to −0.31% per year at age 60 years. Of interest, the ducted a pilot study to assess the safety and tolerability of a type rate of MS-specific thalamic atrophy decreased from −0.59% of ketogenic diet in patients with relapsing-remitting MS per year at age 30 years to −0.05% per year at age 60 years, (RRMS). Of 20 patients enrolled into this single-arm, open- whereas the rate of normal aging atrophy increased from label 6-month trial, 19 adhered to the dietary regimen for 3 −0.15% per year at age 30 years to −0.62% at age 60 years. By months and 15 for 6 months. Body mass index, total fat mass, contrast, in the putamen and the caudate nucleus, the con- fatigue, and depression scores were significantly improved at tributions of MS-specific atrophy and normal age did not the end of the study, and the proinflammatory adipokine leptin change substantially over the age span. This study suggests that was reduced after 3 months on diet. Although this study was the trajectories of tissue loss attributable to MS and normal not designed to prove a beneficial effect of a dietary in- aging, respectively, may differ across brain regions, which has tervention on MS disease course, it has shown that nutrition implications for the interpretation of brain volume changes in interventions are feasible in MS with good adherence, thus clinical trials and with immunotherapy. warranting subsequent larger, randomized trials. Another emerging imaging technique to visualize and quantify The second study by Camu et al.52 adds to the contentious neuroaxonal degeneration in autoimmune neuroinflammation issue as to whether vitamin D supplementation is able to is retinal optical coherence tomography that seems to be in – modify MS disease course.53 55 The CHOLINE trial was closer proximity to clinical use in individual patients than brain – a randomized, double-blind, placebo-controlled, parallel-group atrophy measurements.58 60 Cordano et al.61 have now in- study in 181 patients with RRMS on stable immunomodula- vestigated the value of OCT-derived measures of retinal neu- tory treatment with interferon beta-1a 44 μg SC 3 times weekly rodegeneration to predict disability worsening in patients with and at least 1 documented relapse during the previous 2 MS. This retrospective study in 305 patients with various years.52 Patients with low serum vitamin D (<75 nmol/L 25- subforms of MS and a median follow-up time of 7.9 years hydroxy vitamin D) were eligible and were randomized 1:1 to between OCT scan and most recent EDSS grading evaluated either 100,000 IU of high-dose oral cholecalciferol or placebo the association of the baseline peripapillary retinal nerve fiber every other week add-on to interferon beta over 96 weeks. The layer thickness (pRNFL) and the subsequent EDSS score. The primary end point (change in the annualized relapse rate authors report an increase in the EDSS score of 0.024 points, [ARR] at 96 weeks) was not met. However, in 90 patients (45 with each 1-μm decrease in the baseline pRNFL. Similar results per group) who completed the 2-year follow-up, efficacy were obtained when adjusting for the presence of previous parameters (ARR, new T1 hypointense lesions, volume of T1 optic neuritis episodes. In line with a previous report,62 this

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 5 study shows that a pRNFL measurement may be useful to influenza, as well as certain bacteria. When comparing serum prognosticate disability as long as 6–9 years later. Despite some and CSF antibody levels, they detected evidence of intrathecal obvious limitations such as the use of the older time-domain synthesis of anti-JCV antibodies in 20% of patients with MS OCT technology, the retrospective nature of the study, and the before natalizumab treatment. During natalizumab treatment, lack of a baseline EDSS score in all study participants, this work intrathecal production of JCV antibodies was lost more fre- supports the use of OCT in clinical patient management as quently in comparison to antigen-specific antibodies to other long as acquisition procedures comply with established quality neurotropic viruses and bacteria tested. Thus, their data sug- criteria.63 gest that there is intrathecal production of JCV-specificanti- bodies in a minority of patients and that there may be selective Initial investigations leading to the development of the anti- reduction of intrathecal JCV-specific humoral immunity dur- VLA4 monoclonal antibody natalizumab (Tysabri) in MS ing natalizumab treatment. Previous clinical investigations, and therapy focused on its role in inhibiting CNS recruitment of recent experimental evidence from Hussain et al.67 and T cells through the blood-brain barrier (BBB).64 Because Lehmann-Horn et al.,66,68 demonstrating that CNS penetra- VLA4 is also expressed on B cells and monocytes, natalizumab tion of B cells, like T cells, is VLA-4 dependent, raise the treatment may affect function of those cells and possibly possibility that reduction of intrathecal production of JCV- contribute to its therapeutic benefit. It is known that natali- specific antibodies could reflect decreased CNS recruitment of zumab treatment of patients with MS reduces accumulation of JCV-specific B cells and/or T cells (i.e., T follicular helper), B cells in CSF.65 Selective genetic deficiency of B-cell VLA-4 which are required for B-cell differentiation into antibody- expression reduces CNS accumulation of B cells, proin- producing plasmablasts and plasma cells. Further studies inflammatory Th17 cells, and monocytes in experimental auto- cluding replication of these findings by Largey et al.71 in a larger immune encephalomyelitis (EAE) induced by MOG protein, cohort will be important to determine whether reduction of a model that requires B cells and leads to antigen-specific B-cell intrathecal humoral JCV-specific immunity translates to higher activation.66 In the July issue of N2, Hussain et al.67 reported risk of PML. on selective B-cell VLA-4 deficiency in a model of EAE that is dependent on T cells, but not B cells, and observed that al- For several years, MOG antibodies have been well recog- though the absence of VLA-4 on B cells reduced CNS B-cell nized in acute disseminated encephalomyelitis and bi- accumulation, it did not alter EAE susceptibility. Their exciting lateral optic neuritis. In 2014 and 2015, several groups findings are consistent with previous work indicating that VLA- described patients with opticospinal disease mimicking 4 expression on B cells is important in regulatory B-cell (Breg) NMOSD that had MOG-specific antibodies, but not – control of EAE68,69 and collectively highlight how B-cell VLA-4 AQP4-specific antibodies.72 74 At that time, there was con- expression may promote pathogenic and regulatory roles of cern in classifying MOG antibody–associated disease as a form different B-cell subsets in CNS autoimmune disease, in- of NMOSD, as the pathophysiology of MOG-targeted and cludingMS.InthesameissueofN2, Sucksdorff et al.70 AQP4-targeted diseases is distinct, the clinical course of these evaluated how natalizumab treatment of patients with MS conditions may not be identical, and patients with these 2 influenced activation of microglia, resident CNS innate im- disorders may not respond to therapeutics in the same mune cells. Microglial activation was measured in 10 patients manner.75,76 Since that time, the number of MOG antibody– with MS using the 18-kDa translocator protein (TSPO)- associated diseases has increased. In the March issue of N2,2 binding radioligand [11C]PK11195 and PET imaging before groups described cases of patients with additional clinical and after 1-year treatment with natalizumab. Natalizumab conditions associated with MOG antibodies.77,78 In 1 study, treatment was associated with reduced microglial activation in Patterson et al.77 reported on 2 patients exhibiting signs of normal-appearing white matter and at the rim of chronic MS small vessel CNS vasculitis. Both patients presented with fever, lesions. Thus, their study demonstrated how natalizumab headaches, and mental status changes, and 1 had cranial nerve treatment of MS may reduce activation of resident CNS innate palsies. They had abnormal brain MRIs and brain biopsies immune cells and established how TSPO-PET imaging can be demonstrating lymphocytic infiltration of small vessels. With used as a tool to assess longitudinal changes in microglial ac- presumed diagnosis of small vessel primary vasculitis, these tivation in NAWM and in perilesional areas in the MS brain in patients initially received cyclophosphamide and steroids. vivo. While on that regimen, 1 patient developed optic neuritis. MOG antibodies were detected in stored serum samples from Although natalizumab is recognized as a very effective MS both patients, and rereview of the biopsies revealed absence of therapy, its use can be associated with PML, a CNS infection fibrinoid necrosis, a pathologic requirement for the diagnosis caused by the opportunistic JCV. Serum antibodies to JCV of small vessel primary vasculitis. Treatment was changed upon serve as a surrogate marker for exposure to JCV, and the risk of recognition these patients had a MOG antibody–associated PML is increased with elevated titers. In the November issue of disease. One patient had no more relapses on treatment with N2, Largey et al.71 conducted a longitudinal observational rituximab, azathioprine, and low-dose prednisone, whereas the study analyzing serum and CSF samples before and during other patient stabilized on a slow prednisone taper. In the natalizumab treatment of 15 patients with MS for antibodies to same issue of N2, Cobo-Calvo et al.78 identified 3 MOG JCV and other viruses, including measles, mumps, rubella, and antibody–positive patients with radiologic and/or clinical

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN involvement of their cranial nerves. One patient had in- Last, we want to thank our reviewers. We are able to accept volvement at the root exit of the oculomotor nerve with only a minority of submitted manuscripts and must make gadolinium enhancement that extended beyond the transition difficult decisions regarding which articles will most benefit our between the CNS and the peripheral myelin. Labeled purified readers and improve patient care. Your thoughtful comments serum IgG from each patient was tested on tissue specimens regarding experimental research investigations, the uniqueness from nonhuman primates (NHPs) (Cynomolgus macaques), of study populations, novel methods and techniques, studies a species that expresses MOG protein that is highly homolo- that are particularly educational, or new strategies for di- gous with human MOG. The serum IgG samples reacted to agnosing and treating neurologic disease are enormously myelin in the NHP brain tissue with the same pattern as helpful and highly appreciated. Our gratitude for your dedi- a MOG-specific monoclonal antibody used as a positive con- cation to reviewing for Neurology® Neuroimmunology & Neu- trol and did not react with cranial nerves. As MOG is produced roinflammation cannot be adequately conveyed. by oligodendrocytes within the CNS, absence of staining of cranial nerves was not surprising. However, at this time, the Please send an email to [email protected] if you pathophysiologic mechanism(s) responsible for the cranial would like to do more reviews or if you have never reviewed for nerve involvement in these patients with anti-MOG antibodies the journal but are interested in doing so. Please include a de- is puzzling. Further studies are needed to determine whether scription of your credentials and expertise in the areas in which the MOG antibodies are pathogenic or if such antibodies are you are qualified to review. The reviewers listed below with 1 a surrogate for a humoral or cellular immune response tar- asterisk have reviewed 5 or more manuscripts. Two asterisks geting an antigen expressed in cranial nerves. Regardless, it is indicate that the reviewer has reviewed 10 or more manu- important to be aware that the spectrum of conditions asso- scripts. This list includes those reviewers who returned a re- ciated with anti-MOG antibodies has grown, some of those view or reviews of initial submissions (rereviews of the same conditions may mimic other neurologic disorders, and treat- manuscript are not included) between January 1, 2019, and ments may differ. October 22, 2019.

Ronaldo Abraham Renaud A. Du Pasquier Douglas Craig Hooper Elisabeth Maillart Albert Saiz Stanley H. Appel Richard M. Dubinsky Maartje G. Huijbers Romain Marignier Joan Santamaria Thais Armangue Massimo Filippi Hanneke E. Hulst Silva Markovic-Plese Stefano Sartori Melissa J. Armstrong Eoin P. Flanagan Ellen Viveka Iacobaeus Roland Martin Ulf Schulze- Christina Azevedo Jeff Fletcher Takahiro Iizuka Andrew McKeon* Topphoff Francesca Bagnato Thomas Forsthuber Carolina Ionete Steven R. Messe Johann Sellner Rohit Bakshi Manuel A. Friese Sarosh R. Irani Augusto Miravalle Tim Sinnecker Konstantin E. Balashov Elliot M. Frohman Maciej Jurynczyk Nicolas Molnarfi Raymond A. Sobel Sergio Baranzini Carles Gaig Peter B. Kang Ichiro Nakashima Collin Spencer Richard L. Barbano Steven Galetta Ho Jin Kim Ram Narendra Israel Steiner David Bearden Hector H. Garcia Samantha M. Kimball Narayan Olaf Stuve Tamir Ben-Hur Christian Geis Ilya Kister Pushpa Björn Tackenberg Joseph R. Berger Jeffrey Marc Gelfand* Ingo Kleiter Narayanaswami Kevin Tan Antonio Bertolotto Benjamin Gelman Luisa Klotz Theodore E. Nash Silvia N. Tenembaum Christian G. Bien Jeroen J. G. Geurts Igor J. Koralnik Scott Douglas Maarten J. Titulaer Dennis Bourdette Gavin Giovannoni Stephen C. Krieger Newsome Keith P. Van Haren Simon Broadley Norbert Goebels Krister Kristensson Kevin O’Connor Jan J. Verschuuren Wolfgang Brück Ralf Gold Jens Kuhle Jorge Oksenberg Pablo Villoslada Helmut Butzkueven Paloma Gonzalez- Mark J. Kupersmith Andrew R. Pachner Patrick J. Waters Peter A. Calabresi Perez Eric Lancaster* Francesco Patti Martin Weber Jonathan L. Carter Rebecca F. Gottesman Klaus Lehmann-Horn Fredrik Piehl Brian G. Qingqing Chai Joan M. Goverman Michael Levy Sean J. Pittock Weinshenker Felicia C. Chow Dennis Grab Frank Leypoldt Anne-Katrin Proebstel Patrick Weydt David B. Clifford Francesc Graus** Teerin Liewluck Harald Prüss Heinz Wiendl Jeffrey A. Cohen Gary Gronseth** Volker Limmroth Helena Radbruch Jörg Wischhusen Bruce Anthony Markus Gschwind Jenny Linnoila Alex D. Rae-Grant Gregory Wu Campbell Cree Yael Hacohen Robert P. Lisak Markus Reindl E. Ann Yeh Anne H. Cross David A. Hafler Erin Longbrake Victor M. Rivera Xinhua Yu Russell C. Dale* John J. Halperin Querol Luis Leonardo Roever Hanna Suhayl S. Dhib-Jalbut Romana Höftberger Jan D. Lunemann Bhaskar Roy Zimmermann Jan Dörr Reinhard Hohlfeld Anja Mähler Klemens Ruprecht Robert Zivadinov

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 7 Study funding 27. Stengel A, Vural A, Brunder AM, et al. Anti-pan-neurofascin IgG3 as a marker of fulminant autoimmune neuropathy. Neurol Neuroimmunol Neuroinflamm 2019;6: No targeted funding reported. e603. doi:10.1212/NXI.0000000000000603. 28. Querol L, Rojas-Garc´ıa R, Diaz-Manera J, et al. Rituximab in treatment-resistant CIDP Disclosure with antibodies against paranodal proteins. Neurol Neuroimmunol Neuroinflamm 2015;2:e149. doi: 10.1212/NXI.0000000000000149. J. Dalmau is editor of N2. Drs, Dalakas, Kolson, and Paul are 29. Levraut M, Cohen M, Bresch S, et al. Immunoglobulin G4-related hypertrophic Associate Editors of N2. Dr. Zamvil is Deputy Editor of N2.Go pachymeningitis: a case-oriented review. Neurol Neuroimmunol Neuroinflamm 2019;6:e568. doi: 10.1212/NXI.0000000000000568. to Neurology.org/NN for full disclosures. 30. Veenhuis RT, Clements JE, Gama L. HIV eradication strategies: implications for the central nervous system. Curr HIV/AIDS Rep 2019;16:96–104. 31. Heaton RK, Franklin DR Jr, Deutsch R, et al. Neurocognitive change in the era of References HIV combination antiretroviral therapy: the longitudinal CHARTER study. Clin 1. Dalmau J, Kolson DL, Paul F, Zamvil SS. N2 year in review and message from the Infect Dis 2015;60:473–480. editor to our reviewers. Neurol Neuroimmunol Neuroinflamm 2019;6:e525. doi:10. 32. Ambrosius B, Gold R, Chan A, Faissner S. Antineuroinflammatory drugs in HIV- 1212/NXI.0000000000000525. associated neurocognitive disorders as potential therapy. Neurol Neuroimmunol 2. Graus F, Dalmau J. Paraneoplastic neurological syndromes in the era of immune- Neuroinflamm 2019;6:e551. doi: 10.1212/NXI.0000000000000551. checkpoint inhibitors. Nat Rev Clin Oncol 2019;16:535–548. 33. Zicari S, Sessa L, Cotugno N, et al. Immune activation, inflammation, and non-AIDS 3. Yshii LM, Hohlfeld R, Liblau RS. Inflammatory CNS disease caused by immune Co-morbidities in HIV-infected patients under long-term ART. Viruses 2019;11: checkpoint inhibitors: status and perspectives. Nat Rev Neurol 2017;13:755–763. E200. 4. Vogrig A, Fouret M, Joubert B, Picard G, Rogemond V, Pinto A. Increased frequency 34. Gill AJ, Kolson DL. Dimethyl fumarate modulation of immune and antioxi- of anti-Ma2 encephalitis associated with immune checkpoint inhibitors. Neurol dant responses: application to HIV therapy. Crit Rev Immunol 2013;33: Neuroimmunol Neuroimmfl 2019;6:e604. doi: 10.1212/NXI.0000000000000604. 307–359. 5. Graus F, Dalmau J, Rene R, et al. Anti-Hu antibodies in patients with small-cell lung 35. Pino M, Paganini S, Deleage C, et al. Fingolimod retains cytolytic T cells and limits T cancer: association with complete response to therapy and improved survival. J Clin follicular helper cell infection in lymphoid sites of SIV persistence. PLoS Pathog 2019; Oncol 1997;15:2866–2872. 15:e1008081. 6. Monstad SE, Storstein A, Dorum A, et al. Yo antibodies in ovarian and breast cancer 36. Blumberg EA, Rogers CC; American Society of Transplantation Infectious Diseases patients detected by a sensitive immunoprecipitation technique. Clin Exp Immunol Community of P. Solid organ transplantation in the HIV-infected patient: guidelines 2006;144:53–58. from the American Society of Transplantation Infectious Diseases Community of 7. Knudsen A, Monstad SE, Dorum A, et al. Ri antibodies in patients with breast, ovarian Practice. Clin Transpl 2019;33:e13499. or small cell lung cancer determined by a sensitive immunoprecipitation technique. 37. Gisslen M, Heslegrave A, Veleva E, et al. CSF concentrations of soluble TREM2 as Cancer Immunol Immunother 2006;55:1280–1284. a marker of microglial activation in HIV-1 infection. Neurol Neuroimmunol Neu- 8. Cortese I, Muranski P, Enose-Akahata Y, et al. Pembrolizumab treatment for pro- roinflamm 2019;6:e512. doi: 10.1212/NXI.0000000000000512. gressive multifocal leukoencephalopathy. N Engl J Med 2019;380:1597–1605. 38. Kamtchum-Tatuene J, Mwandumba H, Al-Bayati Z, et al. HIV is associated with 9. Rauer S, Marks R, Urbach H, et al. Treatment of progressive multifocal leukoence- endothelial activation despite ART, in a sub-Saharan African setting. Neurol Neu- phalopathy with pembrolizumab. N Engl J Med 2019;380:1676–1677. roimmunol Neuroinflamm 2019;6:e531. doi: 10.1212/NXI.0000000000000531. 10. Walter O, Treiner E, Bonneville F, et al. Treatment of progressive multifocal leu- 39. Reich DS, Lucchinetti CF, Calabresi PA. Multiple sclerosis. N Engl J Med 2018;378: koencephalopathy with nivolumab. N Engl J Med 2019;380:1674–1676. 169–180. 11. Kupper C, Heinrich J, Kamm K, Bucklein V, Rothenfusser S, Straube A. Pem- 40. Koduah P, Paul F, Dörr JM. Vitamin D in the prevention, prediction and treatment of brolizumab for progressive multifocal leukoencephalopathy due to primary immu- neurodegenerative and neuroinflammatory diseases. EPMA J 2017;8:313–325. nodeficiency. Neurol Neuroimmunol Neuroinflamm 2019;6:e628. doi: 10.1212/NXI. 41. Behrens JR, Rasche L, Gieß RM, et al. Low 25-hydroxyvitamin D, but not the bio- 0000000000000628. available fraction of 25-hydroxyvitamin D, is a risk factor for multiple sclerosis. Eur J 12. Pawlitzki M, Schneider-Hohendorf T, Rolfes L, et al. Ineffective treatment of PML with Neurol 2016;23:62–67. pembrolizumab: exhausted memory T-cell subsets as a clue? Neurol Neuroimmunol 42. Moreno MA, Or-Geva N, Aftab BT, et al. Molecular signature of Epstein-Barr virus Neuroinflamm 2019;6:e627. doi: 10.1212/NXI.0000000000000627. infection in MS brain lesions. Neurol Neuroimmunol Neuroinflamm 2018;5:e466. 13. PiquetAL,KhanM,WarnerJEA,etal.Novelclinicalfeaturesofglycinereceptorantibody doi: 10.1212/NXI.0000000000000466. syndrome: a series of 17 cases. Neurol Neuroimmunol Neuroinflamm 2019;6:e592. doi: 43. Jangi S, Gandhi R, Cox LM, et al. Alterations of the human gut microbiome in multiple 10.1212/NXI.0000000000000592. sclerosis. Nat Commun 2016;7:12015. 14. Swayne A, Tjoa L, Broadley S, et al. Antiglycine receptor antibody related disease: 44. Chen J, Chia N, Kalari KR, et al. Multiple sclerosis patients have a distinct gut a case series and literature review. Eur J Neurol 2018;25:1290–1298. microbiota compared to healthy controls. Sci Rep 2016;6:28484. 15. Martinez-Hernandez E, Sepulveda M, Rostasy K, et al. Antibodies to aquaporin 4, 45. Miyake S, Kim S, Suda W, et al. Dysbiosis in the gut microbiota of patients with myelin-oligodendrocyte glycoprotein, and the glycine receptor alpha1 subunit in multiple sclerosis, with a Striking depletion of species belonging to clostridia XIVa and patients with isolated optic neuritis. JAMA Neurol 2015;72:187–193. IV clusters. PLoS One 2015;10:e0137429. 16. Armangue T, Sabater L, Torres-Vega E, et al. Clinical and immunological features of 46. Tremlett H, Fadrosh DW, Faruqi AA, et al; US Network of Pediatric MS Centers. Gut opsoclonus-myoclonus syndrome in the era of neuronal cell surface antibodies. JAMA microbiota in early pediatric multiple sclerosis: a case-control study. Eur J Neurol Neurol 2016;73:417–424. 2016;23:1308–1321. 17. Arino H, Gresa-Arribas N, Blanco Y, et al. Cerebellar ataxia and glutamic acid 47. Ochoa-Repáraz J, Magori K, Kasper LH. The chicken or the egg dilemma: intestinal decarboxylase antibodies: immunologic profile and long-term effect of immuno- dysbiosis in multiple sclerosis. Ann Transl Med 2017;5:145. therapy. JAMA Neurol 2014;71:1009–1016. 48. Katz Sand I, Zhu Y, Ntranos A, et al. Disease-modifying therapies alter gut microbial 18. Ekizoglu E, Baykan B, Sezgin M, et al. Follow-up of patients with epilepsy harboring composition in MS. Neurol Neuroimmunol Neuroinflamm 2018;6:e517. doi: 10. antiglycine receptor antibodies. Epilepsy Behav 2019;92:103–107. 1212/NXI.0000000000000517. 19. Gresa-Arribas N, Titulaer MJ, Torrents A, et al. Antibody titres at diagnosis and during 49. Choi IY, Piccio L, Childress P, et al. A diet mimicking fasting promotes regeneration follow-up of anti-NMDA receptor encephalitis: a retrospective study. Lancet Neurol and reduces autoimmunity and multiple sclerosis symptoms. Cell Rep 2016;15: 2014;13:167–177. 2136–2146. 20. Koneczny I. A new classification system for igG4 autoantibodies. Front Immunol 50. Kim DY, Hao J, Liu R, Turner G, Shi FD, Rho JM. Inflammation-mediated memory 2018;9:97. dysfunction and effects of a ketogenic diet in a murine model of multiple sclerosis. 21. Huijbers MG, Vergoossen DL, Fillié-Grijpma YE, et al. MuSK myasthenia gravis PLoS One 2012;7:e35476. monoclonal antibodies: valency dictates pathogenicity. Neurol Neuroimmunol 51. Brenton JN, Banwell B, Bergqvist AGC, et al. Pilot study of a ketogenic diet in Neuroinflamm 2019;6:e547. doi: 10.1212/NXI.0000000000000547. relapsing-remitting MS. Neurol Neuroimmunol Neuroinflamm 2019;6:e565. doi: 10. 22. Querol L, Devaux J, Rojas-Garcia R, Illa I. Autoantibodies in chronic inflammatory neu- 1212/NXI.0000000000000565. ropathies: diagnostic and therapeutic implications. Nat Rev Neurol 2017;13:533–547. 52. Camu W, Lehert P, Pierrot-Deseilligny C, et al. Cholecalciferol in relapsing-remitting 23. Cortese A, Lombardi R, Briani C, et al. Antibodies to neurofascin, contactin-1, and MS: a randomized clinical trial (CHOLINE). Neurol Neuroimmunol Neuroinflamm contactin-associated protein 1 in CIDP: Clinical relevance of IgG isotype. Neurol 2019;6:e597. doi: 10.1212/NXI.0000000000000597. Neuroimmunol Neuroinflamm 2019;7:e639. doi: 10.1212/NXI.0000000000000639. 53. Dörr J, Döring A, Paul F. Can we prevent or treat multiple sclerosis by individualised 24. Querol L, Nogales-Gadea G, Rojas-Garcia R, et al. Antibodies to contactin-1 in vitamin D supply? EPMA J 2013;29:4. chronic inflammatory demyelinating polyneuropathy. Ann Neurol 2013;73:370–380. 54. Feng X, Wang Z, Howlett-Prieto Q, Einhorn N, Causevic S, Reder AT. Vitamin D 25. Miura Y, Devaux JJ, Fukami Y, et al. Contactin 1 IgG4 associates to chronic inflammatory enhances responses to interferon-β in MS. Neurol Neuroimmunol Neuroinflamm demyelinating polyneuropathy with sensory ataxia. Brain 2015;138:1484–1491. 2019;6:e622. doi: 10.1212/NXI.0000000000000622. 26. Carrera-Garcia L, Natera-de Benito D, Lleixà C, et al. Chronic inflammatory de- 55. Rotstein DL, Healy BC, Malik MT, et al. Effect of vitamin D on MS activity by disease- myelinating polyneuropathy associated with contactin-1 antibodies in a child. Neurol modifying therapy class. Neurol Neuroimmunol Neuroinflamm 2015;2:e167. doi: 10. Neuroimmunol Neuroinflamm 2019;6:e602. doi:10.1212/NXI.0000000000000602. 1212/NXI.0000000000000167.

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN 56. Azevedo CJ, Overton E, Khadka S, et al. Early CNS neurodegeneration in radiolog- 68. Lehmann-Horn K, Sagan SA, Winger RC, et al. CNS accumulation of regulatory ically isolated syndrome. Neurol Neuroimmunol Neuroinflamm 2015;2:e102. doi: 10. B cells is VLA-4-dependent. Neurol Neuroimmunol Neuroinflamm 2016;3:e212. doi: 1212/NXI.0000000000000102. 10.1212/NXI.0000000000000212. 57. Azevedo CJ, Cen SY, Jaberzadeh A, Zheng L, Hauser SL, Pelletier D. Contribution of 69. Glatigny S, Wagner CA, Bettelli E. Cutting edge: integrin alpha4 is required for normal aging to brain atrophy in MS. Neurol Neuroimmunol Neuroinflamm 2019;6: regulatory B cell control of experimental autoimmune encephalomyelitis. J Immunol e616. doi:10.1212/NXI.0000000000000616. 2016;196:3542–3546. 58. Oertel FC, Zimmermann HG, Brandt AU, Paul F. Novel uses of retinal imaging with optical 70. Sucksdorff M, Tuisku J, Matilainen M, et al. Natalizumab treatment reduces microglial coherence tomography in multiple sclerosis. Expert Rev Neurother 2019;19:31–43. activation in the white matter of the MS brain. Neurol Neuroimmunol Neuroinflamm 59. Oberwahrenbrock T, Traber GL, Lukas S, et al. Multicenter reliability of semi- 2019;6:e574. doi: 10.1212/NXI.0000000000000574. automatic retinal layer segmentation using OCT. Neurol Neuroimmunol Neuro- 71. Largey F, Jelcic I, Sospedra M, Heesen C, Martin R, Jelcic I. Effects of natalizumab inflamm 2018;5:e449. doi: 10.1212/NXI.0000000000000449. therapy on intrathecal antiviral antibody responses in MS. Neurol Neuroimmunol 60. Cellerino M, Cordano C, Boffa G, et al. Relationship between retinal inner nuclear Neuroinflamm 2019;6:e621. doi: 10.1212/NXI.0000000000000621. layer, age, and disease activity in progressive MS. Neurol Neuroimmunol Neuro- 72. Sato DK, Callegaro D, Lana-Peixoto MA, et al. Distinction between MOG antibody- inflamm 2019;6:e596. doi: 10.1212/NXI.0000000000000596. positive and AQP4 antibody-positive NMO spectrum disorders. Neurology 2014;82: 61. Cordano C, Nourbakhsh B, Devereux M, et al. pRNFL as a marker of disability 474–481. worsening in the medium/long term in patients with MS. Neurol Neuroimmunol 73. Kitley J, Waters P, Woodhall M, et al. Neuromyelitis optica spectrum disorders with Neuroinflamm 2019;6:e533. doi:10.1212/NXI.0000000000000533. aquaporin-4 and myelin-oligodendrocyte glycoprotein antibodies: a comparative 62. Martinez-Lapiscina EH, Arnow S, Wilson JA, et al; IMSVISUAL consortium. Retinal study. JAMA Neurol 2014;71:276–283. thickness measured with optical coherence tomography and risk of disability wors- 74. Hoftberger R, Sepulveda M, Armangue T, et al. Antibodies to MOG and AQP4 in ening in multiple sclerosis: a cohort study. Lancet Neurol 2016;15:574–584. adults with neuromyelitis optica and suspected limited forms of the disease. Mult Scler 63. Schippling S, Balk LJ, Costello F, et al. Quality control for retinal OCT in multiple 2015;21:866–874. sclerosis: validation of the OSCAR-IB criteria. Mult Scler 2015;21:163–170. 75. Zamvil SS, Slavin AJ. Does MOG Ig-positive AQP4-seronegative opticospinal 64. Yednock TA, Cannon C, Fritz LC, Sanchez-Madrid F, Steinman L, Karin N. Pre- inflammatory disease justify a diagnosis of NMO spectrum disorder? vention of experimental autoimmune encephalomyelitis by antibodies against alpha 4 Neurol Neuroimmunol Neuroinflamm 2015;2:e62. doi: 10.1212/NXI. beta 1 integrin. Nature 1992;356:63–66. 0000000000000062. 65. Stuve O, Marra CM, Jerome KR, et al. Immune surveillance in multiple sclerosis 76. Reindl M, Rostasy K. MOG antibody-associated diseases. Neurol Neuroimmunol patients treated with natalizumab. Ann Neurol 2006;59:743–747. Neuroinflamm 2015;2:e60. doi: 10.1212/NXI.0000000000000060. 66. Lehmann-Horn K, Sagan SA, Bernard CC, Sobel RA, Zamvil SS. B-cell very late 77. Patterson K, Iglesias E, Nasrallah M, et al. Anti-MOG encephalitis mimicking small antigen-4 deficiency reduces leukocyte recruitment and susceptibility to central ner- vessel CNS vasculitis. Neurol Neuroimmunol Neuroinflamm 2019;6:e538. doi: 10. vous system autoimmunity. Ann Neurol 2015;77:902–908. 1212/NXI.0000000000000538. 67. Hussain RZ, Cravens PD, Miller-Little WA, et al. α4-integrin deficiency in B cells does 78. Cobo-Calvo A, Ayrignac X, Kerschen P, et al. Cranial nerve involvement in patients not affect disease in a T-cell-mediated EAE disease model. Neurol Neuroimmunol with MOG antibody-associated disease. Neurol Neuroimmunol Neuroinflamm 2019; Neuroinflamm 2019;6:e563. doi: 10.1212/NXI.0000000000000563. 6:e543. doi: 10.1212/NXI.0000000000000543.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 9 EDITORIAL OPEN ACCESS Treating muscle-speciﬁc kinase myasthenia gravis fromtheinsideout

Maartje G. Huijbers, PhD, and Jan J.G.M. Verschuuren, MD, PhD Correspondence Dr. Verschuuren Neurol Neuroimmunol Neuroinﬂamm 2020;7:e646. doi:10.1212/NXI.0000000000000646 [email protected]

fi Myasthenia gravis (MG) with antibodies to muscle-speci c kinase (MuSK) is an autoimmune RELATED ARTICLE disease of the neuromuscular junction with some remarkable characteristics. Although MuSK is expressed in all skeletal muscles, patients with this disease present with particularly cranial, bulbar, SHp2 inhibitor protects AChRs from effects of and respiratory muscle weakness, leading to a relatively high frequency of respiratory crises.1 myasthenia gravis MuSK Despite several available immunosuppressive or immunomodulatory treatments, patients with antibody MuSK MG often retain residual muscle weakness and experience more often severe muscle atrophy than patients with the more frequent MG with acetylcholine receptor (AChR) anti- Page e645 bodies.2 Symptomatic treatment with acetylcholinesterase inhibitors is usually ineffective or even detrimental in MuSK MG,1 although 3,4-diaminopyridine might be effective in some patients.3 Therefore, treatment of MuSK MG today still mostly relies on immunosuppressive treatment.1

At the basis of the development of fatigable muscle weakness in patients with MuSK MG is the trophic signaling cascade that enables the motor nerve terminal to instruct the muscle membrane to differentiate into a mature postsynapse. MuSK is at the heart of this signaling cascade. Several groups have shown that perturbation of MuSK signaling by binding of patients’ IgG4 MuSK antibodies leads to loss of this differentiation signal (MuSK phosphorylation), subsequent synaptic disintegration, failure in neuromuscular transmission, and ultimately fatigue.4,5 The predominance of IgG4 subclass autoantibodies is particularly important, as the anti-inflammatory characteristics of this type of antibody imply that the disease mechanism, response to therapy, and possibly also the etiology are different from other forms of MG with IgG1 or IgG3 autoantibodies.

In this issue of Neurology®: Neuroimmunology & Neuroinflammation, Huda et al.6 use the current knowledge on the pathophysiology of MuSK MG to test a new therapeutic strategy by targeting the muscle directly. Normal MuSK signaling/phosphorylation is normally halted by intracellular Rous sarcoma gene (SRC) homology 2 domain-containing phosphotyrosine phosphatase 2 (SHP2). Inhibition of this phosphatase results in spontaneous and agrin-independent MuSK phosphorylation and AChR clustering in vitro. Huda et al. tested whether an SHP2 inhibitor, NSC- 87877, could overcome the inhibitory effect of MuSK autoantibodies in mouse myotube cultures. The study clearly shows improved MuSK activation levels, indicated by increased MuSK phosphorylation and restoration of AChR cluster formation in MuSK autoantibody– treated myofibers. AChR clusters in myotubes are generally considered immature, as they do not express the adult epsilon subunit nor contain the complex morphology required for natural neuromuscular junction transmission. In vivo, agrin (in combination with several other extracellular proteins) is considered critical for induction of maturation of the postsynaptic structure and AChR clusters. A next exciting step will be to extend these observations in vivo and confirm whether SHP2 inhibition can induce formation and maintenance of mature postsynaptic structures also after prolonged exposure to patients autoantibodies. Downstream of kinase 7 overexpression, another form of intracellular activation of MuSK phosphorylation, was shown successful in maintaining synapses in several neuromuscular disorders.7 Huda et al. show that their SHP2 inhibitor also increased the number of AChR clusters in

From the Department of Neurology (M.G.H., J.J.G.M.V.) and Department of Human Genetics (M.G.H.), Leiden University Medical Center, The Netherlands.

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 © 2019 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 downstream of kinase 7-overexpressing myotubes. A negative received by the LUMC. The LUMC receives royalties for effect of chronic activation of MuSK through SHP2 inhibition a MuSK assay. Go to Neurology.org/NN for full disclosures. in vivo is not expected, as chronic activation of MuSK phosphorylation with a monoclonal antibody in a mouse model for Publication history amyotrophic lateral sclerosis showed positive effects on in- Received by Neurology: Neuroimmunology & Neuroinflammation nervation and did not show a deleterious effect.8 Importantly, October 28, 2019. Accepted in final form October 30, 2019. SHP2 is expressed in many cell types of the human body. The authors discuss the potential off-target effects, especially the References 1. Guptill JT, Sanders DB, Evoli A. Anti-MuSK antibody myasthenia gravis: clinical findings risk of tumorigenesis, and suggest that based on current and response to treatment in two large cohorts. Muscle Nerve 2011;44:36–40. knowledge, the risk of tumorigenesis is not increased. 2. Farrugia ME, Robson MD, Clover L, et al. MRI and clinical studies of facial and bulbar muscle involvement in MuSK antibody-associated myasthenia gravis. Brain 2006; 129(pt 6):1481–1492. When these technical challenges are overcome, SHP2 in- 3. Skjei KL, Lennon VA, Kuntz NL. Muscle specific kinase autoimmune myasthenia hibition has the potential to become a new symptomatic gravis in children: a case series. Neuromuscul Disord 2013;23:874–882. 4. Koneczny I, Cossins J, Waters P, Beeson D, Vincent A. MuSK myasthenia gravis IgG4 treatment for patients with MuSK MG. It is tempting to disrupts the interaction of LRP4 with MuSK but both IgG4 and IgG1-3 can disperse speculate which other disorders with neuromuscular junction preformed agrin-independent AChR clusters. PLoS One 2013;8:e80695. impairment may benefit from strengthening synapses from 5. Huijbers MG, Zhang W, Klooster R, et al. MuSK IgG4 autoantibodies cause myas- “ ” 9 thenia gravis by inhibiting binding between MuSK and Lrp4. Proc Natl Acad Sci U S A the inside out using SHP2 inhibitors. 2013;110:20783–20788. 6. Huda S, Cao M, De Rosa A, et al. SHp2 inhibitor protects AChRs from effects of Study funding myasthenia gravis MuSK antibody. Neurol Neuroimmunol Neuroinflamm 2019;7: e645. doi: 10.1212/NXI.0000000000000645. No targeted funding reported. 7. Arimura S, Okada T, Tezuka T, et al. Neuromuscular disease. DOK7 gene therapy benefits mouse models of diseases characterized by defects in the neuromuscular junction. Science 2014;345:1505–1508. Disclosure 8. Cantor S, Zhang W, Delestrée N, Remédio L, Mentis GZ, Burden SJ. Preserving LUMC, M.G. Huijbers, and J.J.G.M. Verschuuren have ap- neuromuscular synapses in ALS by stimulating MuSK with a therapeutic agonist plied for patents in the field of MuSK myasthenia gravis. antibody. Elife 2018;7:e34375. 9. Ohno K, Ohkawara B, Ito M. Agrin-LRP4-MuSK signaling as a therapeutic target for J.J.G.M. Verschuuren has been involved in consultancies for myasthenia gravis and other neuromuscular disorders. Expert Opin Ther Targets Argenx, Alexion, and Rapharma. All reimbursements were 2017;21:949–958.

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN ARTICLE OPEN ACCESS Anti-NMDAR encephalitis A single-center, longitudinal study in China

Xiaolu Xu, MD,* Qiang Lu, MD,* Yan Huang, MD, Siyuan Fan, MD, Lixin Zhou, MD, Jing Yuan, MD, Correspondence Xunzhe Yang, MD, Haitao Ren, BS, Dawei Sun, MD, Yi Dai, MD, Huadong Zhu, MD, Yinan Jiang, MD, Dr. Guan [email protected] Yicheng Zhu, MD, Bin Peng, MD, Liying Cui, MD, and Hongzhi Guan, MD

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e633. doi:10.1212/NXI.0000000000000633 Abstract Objective To describe the detailed clinical characteristics, immunotherapy, and long-term outcomes of patients with anti-NMDA receptor (NMDAR) encephalitis in China.

Methods A single-center, prospective study. Patients who met the diagnostic criteria were enrolled from 2011 to 2017 and followed up. The clinical features, treatment, and long-term outcomes were collected prospectively. Factors aﬀecting the long-term prognosis were analyzed.

Results The study included 220 patients. The most common clinical presentations were psychosis (82.7%) and seizures (80.9%). Of the patients, 19.5% had an underlying neoplasm; of which ovarian teratoma was 100% of tumors in females and only one male had lung cancer. Most patients (99.5%) received first-line therapy (glucocorticoids, IV immunoglobulin, or plasmapheresis alone or combined), and only 7.3% received second-line immunotherapy (rituximab, cyclophosphamide alone, or combined). Long-term immunotherapy (mycophenolate mofetil or azathioprine >1 year) was administered to 53.2% of patients. During the first 12 months, 207 (94.1%) patients experienced improvement, and 5 (2.3%) died, whereas 38 (17.3%) experienced relapses. At 12-month follow-up, 92.7% had favorable clinical outcomes (modified Rankin Scale score ≤2).

Conclusions Patients in China present with psychosis and seizure frequently but have a low percentage of underlying neoplasms. Re-enforced first-line immunotherapy is effective in managing anti- NMDAR encephalitis in the acute phase. Although relapse is relatively common, with combined first-line and long-term immunotherapy, most patients reached favorable outcomes.

*These authors contributed equally to the manuscript.

From the Department of Neurology (X.X., Q.L., Y.H., S.F., L.Z., J.Y., X.Y., H.R., Y.Z., B.P., L.C., H.G.), Department of Gynecology and Obstetrics (D.S., Y.D.), Department of Emergency (H.Z.), and Department of Psychology (Y.J.), Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China; and Center of Neuroscience (L.C.), Chinese Academy of Medical Sciences.

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 National Key Research and Development Program of China (Grant no. 2016YFC0901500); Center for Rare Diseases Research, Chinese Academy of Medical Sciences, Beijing, China (Grant no. 2016ZX310174-4); and Beijing Municipal Science and Technology Foundation (Grant no. Z161100000516094). 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 © 2019 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary AE = autoimmune encephalitis; AQP4 = aquaporin-4; AZA = azathioprine; CTX = cyclophosphamide; HSV = herpes simplex virus; ICU = intensive care unit; IQR = interquartile range; IVIG = IV immunoglobulin; MMF = mycophenolate mofetil; MOG = myelin oligodendrocyte glycoprotein; mRS = modiﬁed Rankin Scale; MTX = methotrexate; NMDAR = NMDA receptor; PE = plasmapheresis; PUMCH = Peking Union Medical College Hospital; RTX = rituximab.

Anti-NMDA receptor (NMDAR) encephalitis is the most Demographic data and ancillary tests results were recorded, common type of autoimmune encephalitis (AE), which is including age at onset, sex, disease course, CSF tests results, associated with autoantibodies against neurosurface or syn- MRI, and EEG results. All patients were screened at least once – aptic antigens.1 3 Since its first report in 2007,1 with the for systemic tumors at onset. Patients with tumors underwent proposed clinical approach,3 increasing numbers of anti- tumor removal. Immunotherapy included first-line (cortico- NMDAR encephalitis cases were identified. steroids, IV immunoglobulin [IVIG], or plasmapheresis [PE] alone or combined) and second-line (rituximab Differences in clinical characteristics and treatment strategies [RTX] and cyclophosphamide [CTX] alone or combined) of anti-NMDAR encephalitis were reported among races and immunotherapies.6,7 Long-term immunotherapy (mycophe- – countries.4 10 Main barriers to AE management in China nolate mofetil [MMF] or azathioprine [AZA] >1 year) and consist of the availability of referral centers, the timeliness of other immunotherapy (intrathecal methotrexate [MTX]) correct diagnosis, and financial concerns.11 In 2017, China were also administered.15,16 proposed a domestic consensus on the management of AE, aiming to increase awareness of the disease and determine the Patients were followed regularly in local hospitals or PUMCH optimal treatment for Chinese patients.12 However, limited neurology clinics. Treatment effects and long-term outcomes data of clinical characteristics and long-term prognosis of were assessed using the modified Rankin Scale (mRS). A poor Chinese anti-NMDAR patients are available owing to few response was defined as no improvement in the mRS score or reports with small sample size.9,11,13,14 as an mRS score ≥4 for 4 weeks; clinical improvement was defined as a decrease in the mRS score ≥1 point from that at Taking the advantage of Peking Union Medical College the previous visit; relapse was defined as an exacerbation of Hospital (PUMCH) as the national referral center for com- previous symptoms or the occurrence of new symptoms after plicated disease, a prospective anti-NMDAR encephalitis being stable for 2 months. Long-term favorable outcome was disease cohort was established to describe the clinical char- defined as an mRS score ≤2, and poor outcome was defined as acteristics, treatment regimen, and long-term outcomes of an mRS score >2 at the end of follow-up. patients with anti-NMDAR encephalitis in China. Statistical analysis Statistical analyses were performed using SPSS IBM 20.0. Methods GraphPad Prism 6.0 was used to generate figures. Quantita- tive data with normal distributions are presented as mean ± Study design and population SD, otherwise as medians with the interquartile range (IQR). In this study, patients with anti-NMDAR encephalitis were The mRS scores before and after treatment were compared enrolled consecutively at PUMCH between May 2011 and using the Wilcoxon test. Symptoms and demographic data December 2017. The inclusion criteria were as follows: were analyzed using the χ2 test or Fisher exact test for cate- (1) acute onset of 1 or more of the 8 major groups of mani- gorical variables and Mann-Whitney U test for continuous fi festations: psychosis, memory de cit, speech disturbance, variables. Factors affecting outcome were assessed using bi- seizures, movement disorder, loss of consciousness, auto- nary logistic regression analysis. Kaplan-Meier curves with nomic dysfunction, and central hypoventilation; (2) CSF log-rank were used to analyze relapse frequency. p < 0.05 was tested positive for NMDAR antibodies (cell-based assay considered significant. [EUROIMMUN, Lübeck, Germany]); and (3) reasonable exclusion of other disorders. Standard protocol approvals, registrations, and patient consents To better understand the clinical characteristics, we also This study was approved by the Ethics Committee of recorded the co-occurrence of fever, headache, arrhythmia, in- PUMCH (JS-891), and informed consent was obtained from tensive care unit (ICU) admission, and other atypical symp- each patient. All the data analyzed in the study were strictly toms. Because of limited resources of the hospital and financial anonymous. concerns the patients, the absolute indications for ICU admission included severe anti-NMDAR encephalitis with ovarian Data availability teratoma requiring surgical operation, status epilepticus, me- Anonymized data not published within this article will be chanical ventilation requirement, and hemodynamic instability. made available by request from any qualified investigator.

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN Results Table 1 Clinical characteristics of patients with anti- Clinical characteristics NMDA receptor encephalitis A total of 220 patients were enrolled, and all were tested for No. of patients (percentage) anti-NMDAR antibody with paired CSF and serum. All Clinical All Age <18 y Age >18 y p patients (100%) were positive for anti-NMDAR antibodies in characteristics (n = 220) (n = 69) (n = 151) Value CSF, and 157 (71.4%) were positive in serum. Review of the Median age, 21 (5–72) 15 (5–17) 26 (18–72) enrolled patients indicated that all patients met the diagnostic range (y) criteria proposed by Graus et al.3 Female 143 (65.0) 41 (59.4) 102 (67.5) 0.24

The median age at onset was 21 (range 5–72) years, with 69 Psychosis 182 (82.7) 57 (82.6) 125 (82.8) 0.98 (31.4%) of the patients younger than 18 years. Overall, 143 Seizures 178 (80.9) 56 (81.2) 122 (80.8) 0.95 (65.0%) patients were females, and 77 (35.0%) were males. Tumors were found in 43 (19.5%) patients: 42 females with Fever 126 (57.3) 41 (59.4) 85 (56.3) 0.67 ovarian teratomas and 1 male with lung cancer. All the female Decreased level of 117 (53.2) 35 (50.7) 82 (54.3) 0.62 patients with ovarian teratomas underwent tumor removal, consciousness whereas the patient with lung cancer was treated with pallia- Memory deficit 106 (48.2) 28 (40.6) 78 (51.7) 0.13 tive therapy by an internist and later died of cancer. In addition, 8 patients were found to have prominent moles on the Speech disturbance 100 (45.5) 30 (43.5) 70 (46.4) 0.69 body surface, which were later resected. Pathologically, the Movement disorder 94 (42.7) 38 (55.1) 56 (37.1) 0.01 moles were compound or intradermal nevi.17,18 Sleep disorder 93 (42.3) 27 (12.3) 66 (30.0) 0.52 The most common clinical manifestations of the anti- Headache 86 (39.1) 21 (30.4) 65 (43.0) 0.08 NMDAR encephalitis were psychosis (182, 82.7%) and Admission to 68 (30.9) 18 (26.1) 50 (33.1) 0.30 seizures (178, 80.9%). Movement disorder presented more the ICU often in patients younger than 18 years than adult patients Mechanical 44 (20.0) 12 (17.4) 32 (21.2) 0.51 (38/69, 55.1% vs 56/151, 37.1%, p = 0.01). Table 1 describes ventilation ’ the patients clinical characteristics in detail. Central 39 (17.7) 9 (13.0) 30 (19.9) 0.22 hypoventilation

Figure 1 shows the distribution of female patients according Arrhythmia 34 (15.5) 12 (17.4) 22 (14.6) 0.59 to age and presence vs absence of tumor; the peak co- occurrence of ovarian teratoma was between 19 to 24 years. Limb weakness 24 (10.9) 9 (13.0) 15 (9.9) 0.49 Ataxia 11 (5.0) 4 (5.8) 7 (4.6) 0.74 Ancillary test results Ovarian teratoma of 42 (29.4) 9 (22.0) 33 (32.4) 0.22 All patients underwent brain MRI at onset, and 79 (35.9%) female patients had abnormal fluid-attenuated inversion recovery sequence signals, including 31 (14.1%) in the medial temporal lobe. Abbreviation: ICU = intensive care unit. Other involved areas included the frontal, parietal, and occipital cortices, diencephalon, cerebellum, and brainstem. Eleven (5.0%) patients had demyelinating lesions, of whom 4 Improvements in diagnosis accuracy and were positive for aquaporin-4 (AQP4) antibody and 5 for general hospitalization status myelin oligodendrocyte glycoprotein (MOG) antibody. Ab- Underrecognition and misdiagnosis were inevitable in the normal EEG findings were seen in 113 (51.4%) patients: 102 past, occurring at the initial visit to a local physician or ad- (46.4%) had slow activity and 14 (6.4%) epileptic discharges. mission to our hospital. In 2011, 4 patients were identified, However, as most patients received short-duration EEG in- taking a median duration of 9 (IQR 1–36) months before stead of video EEG of 24 hours or longer, the percentage of diagnosis was made. The misdiagnosis rate then was 75.0%, delta-brush abnormality was unable to assess. with most cases misdiagnosed as viral encephalitis.

Repeated lumbar punctures were performed for diagnosis and With increased awareness of AE, screening for anti-NMDAR evaluation, and the CSF results at onset before immunotherapy antibodies at admission is now required for all cases with were collected and analyzed. The analysis showed a median suspected encephalitis. By 2017, the misdiagnosis rate had – opening pressure of 170 (IQR 150 280) mmH2O. Of note, decreased to 15.4%. Overall, 30 patients were misdiagnosed 81.3% of the patients had pleocytosis, the median white blood with viral encephalitis, 10 with schizophrenia, 2 with epi- cell count was 14.0 (IQR 7.0–22.5) × 106/L, and 90.9% were of lepsy, and 1 each with cerebral angiitis, cerebral vascular mononuclear cells (i.e., lymphocyte and monocytes). The events, and tuberculosis. The percentages of correct di- protein level was elevated in 29.7% of the patients. Table 2 agnosis at the initial hospital visit over misdiagnosis by cal- summarizes the main ancillary tests of the patients. endar year were shown in ﬁgure 2, indicating a growing

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 3 Figure 1 Distribution of female patients by age and presence or absence of ovarian teratoma

number of referred patients and increased correct diagnosis shown in figure 3. The median mRS score at the last follow-up identified over the years. was 0 (IQR 0–1), which was significantly lower than the score of 4 (IQR 3–5) at onset (Z = −12.67, p < 0.0001). The median duration from onset to diagnosis was 2 (IQR 1–4) weeks. Immunotherapy was initiated the same day di- Table 3 summarizes the comparisons between patients with agnosis was made and sometimes even before diagnosis at favorable and poor clinical outcomes. Of interest, more local hospitals as empirical treatment. The median length of patients with speech disturbance were identified in the fa- hospitalization was 26 (IQR 14–42) days. However, severe vorable outcome group (p = 0.03). Further analysis indicated patients who required intensive care stayed in the hospital for that patients with speech disturbance presented to the neu- up to 117 days. rologist earlier than those without (20 days [IQR 12–34 days] vs 31 days [IQR 19–58 days], p = 0.008, Z = −2.665). Treatment outcomes However, age at onset, the rates of decreased consciousness Overall, 219 (99.5%) patients received first-line immuno- level, central hypoventilation, ICU admission, mechanical therapy, in most cases a combined regimen of repeated ste- ventilation, ovarian teratoma, and relapse were higher in the roids and IVIG. A total of 208 (94.5%) patients received poor outcome subgroup, although none of these differences steroids, of whom 103 (46.8%) received pulsed IV methyl- reached significance. prednisolone. IVIG was administered to 199 (90.5%) patients, and 7 (3.2%) patients underwent PE. Second-line Relapse immunotherapy was administered to only a small proportion During the first 12 months, 38 (17.3%) patients experienced of the patients, usually because of the off-label use of RTX in a first relapse. Their median age was 21 (IQR 16–37) years, AE, cost, IV and hospitalization requirements, and concerns and the median duration from onset to the first relapse was about side effects. Twelve (5.5%) patients received RTX, and 7 (IQR 5–10) months. Of the relapsed patients, 25 (65.8%) 4 (1.8%) received CTX. Long-term immunotherapy was ad- were female, and 5 had ovarian teratomas at disease onset. ministered mainly in patients who were enrolled later, as an add-on therapy for severe or refractory patients in the acute Thorough clinical and laboratory examinations were conducted phase, or as maintenance therapy to prevent and manage to rule out other etiologies and to validate the diagnosis. Two relapses. In general, MMF was administered in 109 (49.5%) patients were treated at local hospitals before antibodies were patients, 55 of whom at onset and 54 after relapse, and AZA tested, whereas 10/36 (27.8%) patients showed elevated anti- was administered in 8 (3.6%) patients. In addition, intrathecal body titer at relapse during serial serum antibody monitoring. MTX was given to 8 (3.6%) severe patients. Of the 17 patients who underwent lumbar puncture, CSF antibodies were detected in 15 (88.2%) patients. MRI was During the first 12 months, 207 (94.1%) patients experienced abnormal in 8 (21.1%) patients, and an ovarian teratoma was improvement, 8 (3.6%) were stable, and 5 (2.3%) patients died. detected in only 1 (2.6%) patient at relapse. A delay in treat- All survival patients were followed for at least a year (range ment was associated with relapse (p < 0.01). 12–72 months). At 12-month follow-up, 204 (92.7%) patients had attained satisfactory neurologic function (mRS score of 0, However, neither tumor status (p = 0.58) nor treatment 1, and 2 in 144, 47, and 13 patients, respectively) compared regimen (p = 0.34) was statistically associated with relapse with 23 (10.5%) patients with an mRS score ≤2 at onset, as frequency (figure e-1, links.lww.com/NXI/A157).

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN duration was 25 (IQR 14–43) days, and the subsequent Table 2 Summary of the main ancillary tests results hospitalizations were all shorter than 14 days.

Number (percentage) Between 2011 and 2017, 80 (36.4%) patients experienced Brain MRI relapse, and 21 (26.3%) experienced multiple relapses (range Total with abnormal findings 79 (35.9) 2–4 episodes). Most patients had a ﬁrst relapse during the ﬁrst 24 months (64/80, 80.0%). However, relapses up to 6 years Medial temporal lobe 31 (14.1) after onset were also reported in our cohort. Frontal lobe 20 (9.1) Parietal lobe 13 (5.9) Discussion Occipital lobe 13 (5.9) To our knowledge, this is the largest Chinese anti-NMDAR Diencephalon 2 (0.9) encephalitis cohort to date. In our study, anti-NMDAR en- Cerebellum 3 (1.4) cephalitis is predominantly found in females (65.0%) with the Brainstem 6 (2.7) median age at onset of 21 years. Most patients presented with psychiatric symptoms and seizures, and younger patients Demyelinating lesion 11 (5.0) presented more often with movement disorders, which are 6,7 Anti-AQP4 Ab(+) 4 (1.8) consistent with previous studies.

Anti-MOG Ab(+) 5 (2.3) However, our study reported a low ICU admission rate. The EEG large cohort study of Titulaer et al.7 reported that 75% Total with abnormal findings 113 (51.4) (435/563) of patients with mRS score ≥4 were admitted to the ICU. In our study, because of limited availability of medical Epileptic discharges 14 (6.4) recourses and concern of expenses, only 68/133 (51.1%) of the Slow activity 102 (46.4) severe patients (mRS score ≥4) were admitted to the ICU. CSF analysis (information of Number 64 patients) Median (IQR) (percentage) The prevalence of an underlying neoplasm varied among studies. Titulaer et al.7 reported that 38% of the patients had Opening pressure (mmH2O) 170 (150–280) a tumor, and Asian patients were more likely (45%) to have 6 – WBC (×10 /L) 14.0 (7.0 22.5) a teratoma. However, only 19.5% of the patients in our cohort Pleocytosis 52 (81.3) had a tumor, with 29.4% of the females had an ovarian teratoma. Other studies of Chinese or Asian patients have also Percentage of mononuclear 90.9 (83.1–99)% cells (%) reported low prevalence of tumors (Lim et al., 22.7%; Wang et al., 8%; Liu et al., 6.7%; and Zhang et al., 8.1%).8,11,13,14 The Protein (g/L) 0.30 (0.21–0.42) heterogeneity among reports may be due to sample sizes and ≥0.45 19 (29.7) selection bias or other factors including genetic backgrounds

Positive SOB 16 (25.0) and epidemiologic reasons, and future studies are required.

– Glucose (mmol/L) 3.5 (3.1 4.2) In tumor-negative patients, herpes simplex virus (HSV) in- Chloride (mmol/L) 123 (120–125) fection has been reported to be the possible trigger. However, as the gold standard for the infection is PCR testing, which is Abbreviations: AQP4 = aquaporin-4; IQR = interquartile range; MOG Ab = myelin oligodendrocyte glycoprotein antibody; SOB = specific oligoclonal expensive and time consuming, most patients received em- band; WBC = white blood cell count. pirical treatment before or even without a deﬁnite diagnosis of Mononuclear cells: lymphocytes and monocytes. CSF-SOB is defined as bands that are present in CSF but are absent in serum. HSV encephalitis. Thus post-HSV anti-NMDAR encephalitis Positive SOB is defined as at least 1 specific band in the CSF. was unable to assess in our current study.

Brain MRI findings provide further evidence that anti- All relapsed patients underwent reinitiation of the first-line im- NMDAR encephalitis is a “diffuse encephalopathy.”6 Abnor- munotherapy, and 18 patients were also given long-term MMF. mal signals were reported in 35.9% of the patients, Subsequently, 12 (31.6%) patients experienced further relapses predominantly in the medial temporal lobe. Signals in other (range 2–4 episodes). There was no significant difference in the areas of the cortex, diencephalon, brainstem, and cerebellum occurrence of subsequent relapse between MMF-treated patients were also reported. and other patients (4/18, 22.2% vs 8/20, 40.0%, p = 0.31). Notably, “overlapping syndrome” was identified in 11 (5.0%) Subsequent relapses were similar to, or worse than, the initial patients with both MRI demyelinating lesions and anti-AQP4 episodes in only 3 (7.9%) patients. The initial hospitalization or anti-MOG antibodies.19 Studies have suggested more

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 5 Figure 2 Number of patients of correct diagnosis over misdiagnosis at the initial hospital visit by calendar year

intense immunotherapy requirements and more residual was associated with relapse frequency statistically, possibly due deficits in these patients.19,20 In our series, all these patients to the low prevalence of tumor and variety in the treatment received long-term immunotherapy with MMF, and regimens used in our cohort. Long-term MMF did not prevent 10 (90.9%) had favorable clinical outcomes. further relapses, possibly because of the relatively small sample size, and the role of long-term immunotherapy warrants further In the management of anti-NMDAR encephalitis, repeated investigation. Although most patients experienced a first relapse first-line immunotherapy was frequently used in our cohort, within 24 months, relapse 6 years after onset was also reported. whereas second-line immunotherapy was administered in Other reports also suggested that AE relapse could occur years a small portion of patients owing to the off-label use of RTX for after the initial episode.21,23,24 Therefore, extended follow-up is AE in China, cost, hospitalization requirements, and concerns essential. Our study has several limitations. As the national about side effect.9,11,14,15 However, long-term immunotherapy referral center for complicated disease, our cohort may be bi- was administered to 117 (53.2%) patients, including MMF to ased by more refractory cases. The analysis for each individual 109 (49.5%). With combined therapy of re-enforced first-line treatment, particularly for long-term immunotherapy, and the therapy and long-term immunotherapy, 204 (92.7%) patients assessment of relapses warrants further study. Parameters other reached favorable clinical outcomes, and the median mRS score decreased significantly from 4 to 0 at a follow-up of 12 months. Compared with other reports (Dalmau et al., 77%; Figure 3 Distribution of mRS scores at onset and last fol- Tituaer et al., 79%; Wang et al., 80.4%; Liu et al., 64%; and low-up Zhang et al., 89.2%),6,7,11,13,14 we observed more satisfactory clinical outcomes. Speech disturbance was found to be more frequent in the group with favorable outcome. Further analysis indicated that patients with speech distance were diagnosed earlier. Thus, this could be a confounder reflecting better recognition and therefore quicker treatment.

Relapses were relatively common in our cohort. The definition of relapse in our study, along with other proposed definitions,7,21,22 is based more on observations and descrip- tions of clinical symptoms. Nevertheless, thorough examinations are needed to rule out other disorders and validate the diagnosis. When monitoring and evaluating the relapses, MRI was frequently unremarkable. The serum antibody titer did not correlate with the clinical severity perfectly, and some relapsed antibodies were detected only in the CSF, as previously reported.6,22 However, serial CSF monitoring may be imprac- tical during follow-up, and better indicators should be identified mRS = modified Rankin Scale. in future studies. Neither tumor status nor treatment regimen

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN Table 3 Comparisons of clinical data of patients with anti-NMDA receptor encephalitis

Favorable outcome (n = 204) Poor outcome (n = 16) Z/χ2 value p Value

Age (y) 21 (16–28) 27 (20–32) −1.79 0.07

Sex (n, %)

Male 71 (34.8) 6 (37.5) 0.05 0.83

Female 133 (65.2) 10 (62.5)

Symptoms (n, %)

Fever 118 (57.8) 7 (43.8) 1.20 0.27

Headache 80 (39.2) 5 (31.3) 0.40 0.53

Psychosis 170 (83.3) 11 (68.8) 2.16 0.17

Memory deficit 100 (49.0) 5 (31.3) 1.88 0.17

Seizures 165 (80.9) 12 (75.0) 0.33 0.52

Decreased level of consciousness 106 (52.0) 10 (62.5) 0.66 0.42

Movement disorder 87 (42.7) 6 (37.5) 0.16 0.69

Speech disturbance 96 (47.1) 3 (18.8) 4.80 0.03

Central hypoventilation 35 (17.2) 4 (25.0) 0.63 0.49

ICU admission 61 (29.9) 7 (43.8) 1.33 0.27

Mechanical ventilation 40 (19.6) 4 (25.0) 0.27 0.53

Arrhythmia 32 (15.7) 2 (12.5) 0.12 0.73

Ataxia 9 (4.4) 2 (12.5) 2.04 0.19

Limb weakness 22 (10.8) 2 (12.5) 0.05 0.69

Ovarian teratomaa 37 (27.8) 5 (50.0) 2.21 0.16

CSF analysis of acute phase (n, %)

Pleocytosisb 88 (58.7) 4 (44.4) 0.70 0.40

Increased level of proteinc 33 (24.6) 3 (33.3) 0.34 0.56

Positive SOBd 48 (50.5) 2 (28.6) 1.26 0.26

Abnormal brain MRI signal (n, %) 72 (35.3) 7 (43.8) 0.46 0.50

Abnormal EEG findings (n, %)e 110 (80.9) 6 (75.0) 0.17 0.68

Diagnosis duration (mo) 1.0 1.0 / /

Immunotherapy (n, %)

Steroids 194 (95.1) 14 (87.5) 1.66 1.20

IVIG 184 (90.2) 15 (93.8) 0.22 0.64

MMF 103 (50.5) 6 (37.5) 1.51 0.17

Baseline mRS score >2 (n, %) 183 (89.7) 14 (87.5) 0.08 0.78

Relapsed patient (n, %) 73 (35.8) 7 (43.8) 0.41 0.52

Abbreviations: ICU = intensive care unit; IVIG = IV immunoglobulin; MMF = mycophenolate mofetil; mRS = modified Rankin Scale; SOB = specific oligoclonal band. a Analysis of 143 female patients. b Analysis of 159 patients with results available. c Analysis of 143 patients with results available. d Analysis of 102 patients with results available. e Analysis of 144 patients with results available.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 7 than the mRS score are required to describe fully the outcomes of anti-NMDAR encephalitis, especially in evaluating the cog- Appendix (continued)

nitive or behavioral function, which usually remains last in these Name Location Role Contribution patients. Despite these limitations, our study adds to the present knowledge of anti-NMDAR encephalitis, and acts as a precursor Yan Peking Union Author Major role in the Huang, Medical College acquisition of data and for future multicenter studies with more comprehensive eval- MD Hospital, Beijing, analyzed the data uations, and studies to further evaluate the eﬃcacy of each China

individual treatment. Siyuan Peking Union Author Major role in the Fan, MD Medical College acquisition of data and Hospital, Beijing, analyzed the data We describe the clinical characteristics, immunotherapy reg- China imens, and long-term outcomes of patients with anti- fi Lixin Peking Union Author Analyzed the data and NMDAR encephalitis in China. Repeated rst-line therapy Zhou, MD Medical College revised the manuscript for is effective in managing acute phase encephalitis, and the ef- Hospital, Beijing, intellectual content ficacy of long-term immunotherapy warrants further study. China Although relapses are relatively common, most patients Jing Peking Union Author Analyzed the data and reached favorable outcomes. Further multicenter studies with Yuan, MD Medical College revised the manuscript for Hospital, Beijing, intellectual content more advanced study design, more detailed evaluation, and China extended follow-up are required. Xunzhe Peking Union Author Interpreted the data and Yang, MD Medical College revised the manuscript for Acknowledgment Hospital, Beijing, intellectual content The authors thank Prof. Jianming Wang of Peking Union China Medical College Hospital and Prof. Jiawei Wang of Beijing Haitao Peking Union Author Major role in the Tongren Hospital Affiliated to Capital Medical University for Ren Medical College acquisition of data; Hospital, Beijing, interpreted the data; and providing advice for the article. They also thank Dr. Yan Zhang China revised the manuscript for of Xuanwu Hospital Affiliated to Capital Medical University, intellectual content Dr. Yongqiang Hu of Beijing Fengtai Youanmen Hospital, and Dawei Peking Union Author Major role in the Beijing Encephalitis Group for contributing participants. Sun, MD Medical College acquisition of data; Hospital, Beijing, interpreted the data; and China revised the manuscript for Study funding intellectual content

The study was supported by (1) National Key Research and Yi Dai, Peking Union Author Major role in the Development Program of China (Grant no. 2016YFC0901500); MD Medical College acquisition of data and Hospital, Beijing, interpreted the data (2) Center for Rare Diseases Research, Chinese Academy of China Medical Sciences, Beijing, China (Grant no. 2016ZX310174-4); and (3) Beijing Municipal Science and Technology Foundation Huadong Peking Union Author Major role in the Zhu, MD Medical College acquisition of data and (Grant no. Z161100000516094). Hospital, Beijing, interpreted the data China

Disclosure Yinan Peking Union Author Major role in the The authors report no disclosures. Go to Neurology.org/NN Jiang, MD Medical College acquisition of data Hospital, Beijing, for full disclosures. China

Yicheng Peking Union Author Designed and Publication history Zhu, MD Medical College conceptualized study and Received by Neurology: Neuroimmunology & Neuroinﬂammation Hospital, Beijing, revised the manuscript for April 21, 2019. Accepted in ﬁnal form September 12, 2019. China intellectual content Bin Peng, Peking Union Author Designed and MD Medical College conceptualized study and Hospital, Beijing, revised the manuscript for Appendix Authors China intellectual content

Name Location Role Contribution Liying Peking Union Author Designed and Cui, MD Medical College conceptualized the study Xiaolu Peking Union Author Analyzed the data and Hospital, Beijing, and revised the Xu, MD Medical College drafted the manuscript for China manuscript for intellectual Hospital, Beijing, intellectual content content China Hongzhi Peking Union Author Designed and Qiang Lu, Peking Union Author Analyzed the data and Guan, Medical College conceptualized the study; MD Medical College drafted the manuscript for MD Hospital, Beijing, interpreted the data; and Hospital, Beijing, intellectual content China revised the manuscript for China intellectual content

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN References 14. Zhang Y, Liu G, Jiang M, Chen W, He Y, Su Y. Clinical characteristics and prognosis of 1. Dalmau J, Tuzun E, Wu HY, et al. Paraneoplastic anti-N-methyl-D-aspartate receptor severe anti-N-methyl-D-aspartate receptor encephalitis patients. Neurocrit Care – encephalitis associated with ovarian teratoma. Ann Neurol 2007;61:25–36. 2018;29:264 272. 2. Dalmau J, Gleichman AJ, Hughes EG, et al. Anti-NMDA-receptor encephalitis: case 15. Guan HZ, Xu XL, Zhu YC, et al. Clinical and immunological analysis of mycophe- series and analysis of the eﬀects of antibodies. Lancet Neurol 2008;7:1091–1098. nolate mofetil treatment in anti-leucine-rich glioma-inactivated 1 encephalitis. Chin J – 3. Graus F, Titulaer MJ, Balu R, et al. A clinical approach to diagnosis of autoimmune Neurol 2018;51:281 287. encephalitis. Lancet Neurol 2016;15:391–404. 16. Yang XZ, Cui LY, Ren HT, Qu T, Guan HZ. Anti-NMDAR encephalitis after re- 4. Iizuka T, Sakai F, Ide T, et al. Anti-NMDA receptor encephalitis in Japan: long-term section of melanocytic nevi: report of two cases. BMC Neurol 2015;15:165. outcome without tumor removal. Neurology 2008;70:504–511. 17. Yang XZ, Zhu HD, Ren HT, et al. Utility and safety of intrathecal methotrexate 5. Irani SR, Bera K, Waters P, et al. N-methyl-D-aspartate antibody encephalitis: tem- treatment in severe anti-N-methyl-D-aspartate receptor encephalitis: a pilot study. – poral progression of clinical and paraclinical observations in a predominantly non- Chin Med J (Engl) 2018;131:156 160. paraneoplastic disorder of both sexes. Brain 2010;133:1655–1667. 18. Yin H, Zhu C, Ren H, et al. Resection of melanocytic nevi as a potential treatment of 6. Dalmau J, Lancaster E, Martinez-Hernandez E, Rosenfeld MR, Balice-Gordon R. anti-NMDAR encephalitis patients without tumor: report of three cases. Neurol Sci Clinical experience and laboratory investigations in patients with anti-NMDAR en- 2018;39:165–167. cephalitis. Lancet Neurol 2011;10:63–74. 19. Titulaer MJ, Hoftberger R, Iizuka T, et al. Overlapping demyelinating syndromes 7. Titulaer MJ, McCracken L, Gabilondo I, et al. Treatment and prognostic factors for and anti-N-methyl-D-aspartate receptor encephalitis. Ann Neurol 2014;75: long-term outcome in patients with anti-NMDA receptor encephalitis: an observa- 411–428. tional cohort study. Lancet Neurol 2013;12:157–165. 20. Fan S, Xu Y, Ren H, et al. Comparison of myelin oligodendrocyte glycoprotein 8. Lim JA, Lee ST, Jung KH, et al. Anti-N-methyl-d-aspartate receptor encephalitis in (MOG)-antibody disease and AQP4-IgG-positive neuromyelitis optica spectrum Korea: clinical features, treatment, and outcome. J Clin Neurol 2014;10:157–161. disorder (NMOSD) when they co-exist with anti-NMDA (N-methyl-D-aspartate) 9. Yuan J, Peng B, Guan H, et al. Immunotherapy strategy for 35 cases of severe anti-N- receptor encephalitis. Mult Scler Relat Disord 2018;20:144–152. methyl-D-aspartate receptor encephalitis. Natl Med J China 2016;96:1035–1039. 21. Gabilondo I, Saiz A, Galan L, et al. Analysis of relapses in anti-NMDAR encephalitis. 10. Bartolini L, Muscal E. Diﬀerences in treatment of anti-NMDA receptor encephalitis: Neurology 2011;77:996–999. results of a worldwide survey. J Neurol 2017;264:647–653. 22. Gresa-Arribas N, Titulaer MJ, Torrents A, et al. Antibody titres at diagnosis and during 11. Wang W, Li JM, Hu FY, et al. Anti-NMDA receptor encephalitis: clinical character- follow-up of anti-NMDA receptor encephalitis: a retrospective study. Lancet Neurol istics, predictors of outcome and the knowledge gap in southwest China. Eur J Neurol 2014;13:167–177. 2016;23:621–629. 23. ArinoH,ArmangueT,Petit-PedrolM,et al. Anti-LGI1-associated cognitive 12. Association NBoCM. Chinese expert consensus on the diagnosis and management of impairment: presentation and long-term outcome. Neurology 2016;87: autoimmune encephalitis. Chin J Neurol 2017;50:91–98. 759–765. 13. Liu L, Song ZH, Guo J, et al. Clinical analysis of 45 Chinese patients with anti-N- 24. van Sonderen A, Thijs RD, Coenders EC, et al. Anti-LGI1 encephalitis: clinical methyl-D-aspartate receptor encephalitis. Chin J Neurol 2014;47:474–481. syndrome and long-term follow-up. Neurology 2016;87:1449–1456.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 9 ARTICLE OPEN ACCESS Black African and Latino/a identity correlates with increased plasmablasts in MS

Kiel M. Telesford, PhD, Ulrike W. Kaunzner, MD, PhD, Jai Perumal, MD, Susan A. Gauthier, DO, MPH, Correspondence Xian Wu, MPH, Ivan Diaz, PhD, Mason Kruse-Hoyer, MD, Casey Engel, BA, Melanie Marcille, BA, and Dr. Telesford [email protected] Timothy Vartanian, MD, PhD or Dr. Vartanian [email protected] Neurol Neuroimmunol Neuroinﬂamm 2020;7:e634. doi:10.1212/NXI.0000000000000634 Abstract Objective To determine the inﬂuence of self-reported Black African and Latin American identity on peripheral blood antibody-secreting cell (ASC) frequency in the context of relapsing-remitting MS.

Methods In this cross-sectional study, we recruited 74 subjects with relapsing-remitting MS and 24 age-, and self-reported ethno-ancestral identity-matched healthy donors (HDs) to provide peripheral blood study samples. Subjects with MS were either oﬀ therapy at the time of study draw or on monthly natalizumab therapy infusions. Using ﬂow cytometry, we assessed peripheral blood mononuclear cells for antibody-secreting B-cell subsets.

Results When stratified by self-reported ethno-ancestry, we identified significantly elevated frequencies of circulating plasmablasts among individuals with MS identifying as Black African or Latin American relative to those of Caucasian ancestry. Ethno-ancestry–specificdifferences in ASC frequency were observed only among individuals with MS. By contrast, this differential was not observed among HDs. ASCs linked with poorer MS prognosis and active disease, including IgM+- and class-switched CD138+ subsets, were among those significantly increased.

Conclusion The enhanced peripheral blood plasmablast signature revealed among Black African or Latin American subjects with MS points to distinct underlying mechanisms associated with MS immunopathogenesis. This dysregulation may contribute to the disease disparity experienced by patient populations of Black African or Latin American ethno-ancestry.

From the Brain and Mind Research Institute (K.M.T., T.V.), Weill Cornell Medicine; Department of Neurology (K.M.T., U.W.K., J.P., S.A.G., M.K.-H., C.E., M.M., T.V.), Weill Cornell Medicine; and Department of Healthcare Policy and Research (X.W., I.D.), Weill Cornell Medicine, New York.

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

Copyright © 2019 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary ASC = antibody-secreting cell; BALAwMS = Subjects with MS of Black African or Latin American self-identity; CAwMS = Subjects with MS of Caucasian self-identity; DMT = disease-modifying therapy; HD = healthy donors without MS; Ig = immunoglobulin; MSSS = MS Severity Scale Score; NAT = natalizumab therapy; PBMC = peripheral blood mononuclear cell; SLE = systemic lupus erythematosus; T25-FW = timed 25-foot walk.

Individuals with MS of Black African or Latin American self- ASC-derived intrathecal IgG correlates with CNS atrophy,9 identity (BALAwMS) are more likely to experience a severe and IgM-producing ASCs are associated with aggressive disease disease course compared with individuals with MS of Cauca- course.15,e11 The present cross-sectional study therefore sian self-identity (CAwMS).1,2 Paraclinical measures of CNS investigates whether the peripheral blood of subjects with MS inflammation (T2 lesion accumulation and lesion volume) may identifying with ethno-ancestral categories more likely to ex- be pronounced,3,4 whereas atrophy metrics, including brain and hibit poorer prognosis reflects an identity-based differential retinal degeneration, appear accelerated among BALAwMS ASC signature. – compared with CAwMS.5 7 Ethno-ancestry is clearly an important consideration in MS. Unfortunately, the paradox of ethno-ancestry being simultaneously relevant in MS yet un- Methods derrepresented in both clinical and translational investigation Standard protocol approvals, registrations, is apparent both in clinical trials and observational research. and patient consents We calculated an average of 2.7% of African Americans among All study subjects were recruited according to Weill Cornell total subjects enrolled in 7 clinical trials conducted between – e1-e10 Medicine Institutional Review Board approved protocol 2006 and 2017 for which demographic data were avail- #1508016490R003. Subjects provided informed written able (table 1). Furthermore, there are essentially no reports consent at the Weill Cornell Medicine Multiple Sclerosis providing direct biological evidence of potential mechanisms Center before study inclusion. underlying ethno-ancestry–based clinical disparity. Retrospec- tive chart review implicates the contribution of antibody- Subject recruitment and study cohorts secreting cells (ASCs), highlighting a relationship between el- Study subjects represent a convenience sample comprising evated intrathecal IgG among African American patients with individuals with clinically definite MS according to the 2010 MS relative to Caucasian patients8 and linking this differential McDonald criteria and healthy donors without MS (HD). We to gray matter atrophy.9 Indeed, plasmablasts and plasma cells, recruited subjects on natalizumab therapy (NAT) as a pri- as ASCs derived from antigen-experienced B cells, appear to be mary study subject population. This enabled study re- – important drivers of both inflammatory10 12 and neurode- cruitment from among a relatively accessible patient generative aspects of MS pathogenesis.9,13,14 ASCs are enriched population. Furthermore, this facilitated our investigation of within the CSF during active gadolinium-enhancing disease,12 intact ASC biology (despite disease-modifying therapy)e12

Table 1 Underrepresentation of individuals of Black African self-identity in Phase III clinical trials for relapsing MS in the last decade

Clinical trial (year reported) Total no. of subjects No. (%) of CA No. (%) of BAA

AFFIRMe1 (2006) 627 603 (96) 10 (1.6)

SENTINALe2 (2006) 1,171 1092 (93) 39 (3.3)

FREEDOMSe3 (2010) 1,272 Not reported 77 (6.1)

TRANSFORMSe4 (2010) 1,292 1216 (94.1) Not reported

TEMSOe5 (2011) 1,088 1058 (97.3) Not reported

DEFINEe6 (2012) 1,234 969 (78.5) 26 (2.1)

CONFIRMe7 (2012) 1,414 1191 (84.1) 27 (1.9)

CARE MS Ie8 (2012) 563 (IIT population) 532 (94.5) Not reported

CARE MS IIe9 (2012) 798 (IIT population) 714 (89.5) Not reported

OPERA I/OPERA IIe10 (2017) 1,656 (IIT population) Not reported 72 (4.3%)

Abbreviations: BAA = Black or African American self-identity; CA = Caucasian self-identity.

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN inclusive of prospective pathogenic lymphocytes within the the associations between ethno-ancestry and outcome meas- peripheral blood. MS NAT subjects required at least 3 prior ures after adjusting for a priori hypothesized confounders age, doses for study inclusion. No subject had received lymphocyte- sex, and MS disease duration. To assess the potential in- depleting therapies at any point before study draw. Study teraction (i.e., effect-measure modification) between ethno- subjects were asked to self-identify according to ethno-ancestral ancestry and disease status in relation to outcome measures, categories for subsequent stratification. Individuals identifying we performed subgroup analyses for HD subjects and subjects with “Black African” or “Latin American/Hispanic” ancestry with MS separately. All statistical tests were 2 sided, with were combined into a single cohort as groups (BALA) based on a significance level of p < 0.05. All analyses were conducted risk for greater disease severity2,16 relative to those identifying with SAS version 9.4 (SAS Institute, Inc, Cary, NC). with “Caucasian/European ancestry” (CA). In addition to self- reported identity, we collected other demographic and clinical Data availability data for each cohort, including age (at study involvement), sex, Anonymized data not shown will be shared by request from fi disease duration, timed 25-foot walk (T25-FW), date since last a quali ed investigator. clinical flare, and MS Severity Scale scores (MSSS).

Sample collection, ASC frequency, and Results count determination Demographic and clinical characteristics We isolated peripheral blood mononuclear cells (PBMCs) We enrolled a total of 54 subjects on natalizumab (“NAT”; through density-gradient Ficoll centrifugation. Ficoll-spun 27 BALAwMS, 27 CAwMS), 20 subjects who were off drug ff bu y coats were harvested within hours of peripheral blood atthetimeofstudydraw(“no disease-modifying drug ff draws and resuspended in volumes of standard staining bu er [DMT]”;12BALAwMS,8CAwMS),and24HDsubjects equal to the original whole blood. We stained cells according who lacked an MS diagnosis (“HD”;11BALAHD,13CA to a standard protocol using the following BioLegend anti- HD). Subjects were comparable in age, both within and be- bodies: CD19 PE/Cy7 (clone HIB19), CD20 PerCP (clone tween cohorts, whereas sex ratio varied. BALAwMS subjects 2H7), CD27 brilliant violet 421 (clone M-T271), CD38 exhibited a longer on average disease duration compared with APC(clone HIT2), CD138 PE (clone MI15), IgD APC/ CAwMS subjects, with this difference more pronounced in the Cy7(clone IA6-2), and IgM FITC (clone MHM-88). We no-DMT cohort (table 2) Clinically, there was some (albeit, fl performed ow cytometric analysis using a Becton Dickinson nonsignificant) indication of greater acute disease activity and FACSVerse cytometer. We determined cell frequencies as the disability within BALAwMS compared with CAwMS subjects percent of a gated cell population among a parent cell pop- at the time of the draw. Thus, MSSS scores were greater, and fl ulation. Cell count was determined as the number of ow T25-FW measures longer, among BALAwMS compared with fi cytometry events within each de ned analysis gate. CAwMS subjects in both NAT and no-DMT groups. Of these measures, only T25-FW in the NAT cohort reached statistical Statistical analysis significance. Among our no-DMT cohort, BALAwMS subjects The primary outcomes of this study are percent frequency of + + exhibited on average a shorter period since their last acute total CD19 ASCs, class-switched ASC subsets, and IgM ASC relapse than CAwMS. This was paralleled by a commensurate subsets. In planning, the estimated sample size needed for measure of months since previous high-dose steroid adminis- a 2-sample means test was calculated based on preliminary data tration. The majority of subjects with MS in our study did not from natalizumab-treated subjects. Trends in the analysis for our possess comorbidities. However, there were several reported additional study cohorts should be treated as exploratory in autoimmune, rheumatoid, or oncologic indications: CAwMS nature. Twenty-six subjects in each group were needed to detect NAT—eczema, mononucleosis, asthma, fibromyalgia, celiac ff ameandi erence of 2 (SD: 2.5) in class-switched frequency disease, Graves disease, and malignant thyroid neoplasm; between ethno-ancestral groups; 17 subjects in each group were BALAwMS NAT—thyroid (diffuse goiter), asthma, irritable ff + needed to detect a mean di erence of 1 (SD: 1) in IgM ASC bowel syndrome, pituitary microadenoma, and positive anti- frequency, assuming alpha of 0.05 and power of 0.8. We cal- nuclear antibodies; CAwMS no DMT—ulcerative colitis, irri- culated summary statistics using frequencies and proportions for table bowel syndrome, hypothyroidism, and psoriasis; and categorical variables and mean, SDs, medians, and interquartile BALAwMS no DMT—positive antinuclear antibodies. ranges for continuous variables. Demographic and clinical characteristics were compared between ethno-ancestry groups Subjects with MS of Black African or Latin using the 2-sample t test, the Wilcoxon rank-sum test, or the American ethno-ancestral identity exhibit Fisher exact test as appropriate. Bivariate analyses were used to enhanced ASC frequencies over those of determine the associations between age, sex, ethno-ancestry, MS Caucasian identity disease status, MS disease duration, and outcome measures To test our hypothesis that stratifying our study sample (i.e., percent frequency and absolute counts of total CD19+ according to ethno-ancestral cohorts at risk of severe disease ASCs, class-switched ASC subsets, and IgM+ ASC subsets) (BALAwMS relative to CAwMS) would reveal a differential using simple linear regression or 2-sample t test. We further ASC frequency within the peripheral blood compartment, constructed multivariable linear regression models to assess we first interrogated the PBMCs of subjects with MS on

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 3 Table 2 Demographic and clinical characteristics for patients with MS and HD study participants

MS NAT MS no DMT HD

CA BALA CA BALA CA BALA p p p N = 27 N = 27Value N = 8 N = 12Value N = 13 N = 11 Value

Age, y 35.37 (11.86) 32.63 (9.12) 0.35 34.88 (6.10) 37.17 (6.66) 0.45 29.33 32.82 0.30 (6.97) (8.85)

Sex, no. (%) 0.1 0.17 0.36

Female 21 (77.8) 26 (96.3) 3 (37.5) 9 (75.0) 2 (15.4) 4 (36.4)

Male 6 (22.2) 1 (3.7) 5 (62.5) 3 (25.0) 11 (84.6) 7 (63.6)

Disease duration, median (IQR), 62 (26–106) 99 (45–148) 0.23 3 (0–70) 56 (13–132) 0.12 ——— mo

MSSS, median (IQR) 0.49 1.92 0.28 1.16 1.86 0.85 ——— (0.23–2.60) (0.23–4.79) (0.89–2.64) (0.59–5.90)

Undetermined values 59 ———

T25-FW, s 3.70 (0.49) 4.29 (0.55) 0.009 3.85 (1.06) 5.30 (1.92) 0.35 ———

Undetermined values 11 17 6 5 ———

Days since last clinical flare, ———146 56 0.90 ——— median (IQR) (52.5–364.5) (23–659)

Months since high-dose steroids, ———8.80 3.30 0.35 ——— median (IQR) (4.20–73.85) (1.60–11.00)

None ———43 ———

Abbreviations: BALA = Black African or Latin American self-identity; CA = Caucasian self-identity; HD = healthy donors without MS; IQR = interquartile range; MSSS = MS Severity Scale Score; MS NAT = subjects with MS on natalizumab; MS no DMT = subjects with MS not on any disease-modifying therapy at the time of study participation; T25-FW = timed 25-foot walk. Both demographic and clinical data were derived from chart review. Mean duration of MS is calculated from years since diagnosis to study draw. Data are given as mean (SD) except where indicated. The p values displayed were derived from 2-sample t test, Wilcoxon rank-sum test, or Fisher exact test. These were employed when comparing demographic and clinical characteristics between ethno-ancestry groups.

natalizumab. One of our most effective therapies, natalizu- and perhaps bound for long-term residence in the bone mab, functions in MS by binding to the alpha-4 integrin marrow. subunit, essentially inhibiting the binding and trafficking of lymphocytes across the blood-brain barrier. In theory, this We uncovered a significant differential, with BALAwMS restricts otherwise pathogenic CNS-destined cells to the exhibiting on average a greater frequency of total ASCs periphery. compared with CAwMS (figure 1A). The majority of these − ASCs were class-switched IgD (figure 1, B and D left ASCs represent a continuum of antigen-activated B cells panel), with a smaller proportion of IgM+ ASCs (figure 1, C that, through differentiation into plasmablasts and mature and E left panel). We further assessed CD38+CD138+ sub- plasma cells, dedicate metabolic and protein processing populations based on recent work, which implicates these mechanisms to antibody production. For this study, we use ASCs as significant intrathecal contributors to inflammatory the term ASCs as inclusive of plasmablasts and maturing gadolinium-enhancing MS.12 Both class-switched and IgM+- plasma cells found in the blood. Despite heterogeneity CD38+CD138+ ASCs were significantly enhanced among across the ASC spectrum, these cells may be collectively BALAwMScomparedwithCAwMS(figure 1, D and E, re- identified as brightly positive for CD27 and CD38.17 We spectively, right panels). Our findings were similar after have thus used this general definition to clarify the distinct adjusting for age, sex, and disease duration (table e-1, links. total ASC population among blood-borne CD19+ cells. We lww.com/NXI/A159) found the majority of circulating ASCs to express high levels of HLA-DR (and CD86), (figure e-1, links.lww.com/NXI/ Ethno-ancestry–based differential ASC A158), consistent with recently generated plasmablast frequency is present among those with MS and identity as shown by others.18 Also in agreement with not HDs others, the majority of circulating CD27hiCD38+ cells lack Although natalizumab possesses a targeted mechanism of CD20 expression, whereas a substantial portion express action,wewishedtoexamineourfindings without its in- CD138, possibly distinguishing these as apoptosis resistant, fluence. To investigate the generalizability of our findings

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN Figure 1 Subjects with MS self-identifying with Black African or Latin American ethno-ancestry exhibit heightened peripheral blood ASC frequencies

Peripheral blood samples from subjects with MS on natalizumab were stratified according to self-identified ethno-ancestry at risk of more severe MS, CAwMS (n = 27), and BALAwMS (n = 27). Percent frequency of different ASC populations was compared between these cohorts. (A) Representative gating strategy for − ASCs; CD27hi CD38+ total ASCs were selected from CD19+ cells for downstream phenotypic subset analysis. (B) Representative gates for class-switched IgD CD27+ total ASCs and CD38+ CD138+ subpopulation. (C) Representative gates for IgM+ ASCs. (D and E), Average frequencies of previously described class- switched (D) and IgM+ (E) ASC populations; error bars represent SD, p values determined the by 2-sided t test. ASC = antibody-secreting cell; BALAwMS = subjects with MS of Black African or Latin American self-identity; CAwMS = subjects with MS of Caucasian self-identity; NAT = natalizumab therapy. beyond the inﬂuence of natalizumab, and to examine include those not on any established DMT. Among these whether our observations were present in individuals subjects, approximately 60% were treatment-naive, having without MS, we extended MS subject enrollment to never been exposed to any DMT. Those with any previous

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 5 exposure to therapy had been off of DMT for at least 2 to our findings in subjects with MS, there was virtually no months before study participation. difference in ASC frequency found among HDs (figure 2, A, D, and E). Our analysis of these cohorts demonstrated an ethno- ancestry–dependent discrepancy in ASC frequency, with Class-switched ASCs are differentially enriched this difference present across essentially all ASC subsets among subjects with MS of Black African or found among natalizumab and off-treatment MS subjects Latin American ethno-ancestral identity (figure 2, A–C). To determine whether this ethnicity- Frequencies of both class-switched and IgM+ ASCs were dependent difference was associated with MS, we performed significantly enhanced among subjects with MS of Black the same phenotypic analysis among HDs. In stark contrast African or Latin American compared with Caucasian ethno-

Figure 2 Ethno-ancestry–based differences in ASC frequency are present among subjects with MS but not HDs

− Average frequencies of previously described total CD19+(A), class-switched CD19+IgD (B and D) and unswitched CD19+IgM+ (C and E) ASC populations. CAwMS (n = 8) and BALAwMS (n = 12); CAHD (n = 13) and BALAHD (n = 11). Error bars represent SD, p values determined by the 2-sided t test. ASC = antibody- secreting cell; BALA HD = healthy donors without MS of Black African or Latin American self-identity; BALAwMS = Subjects with MS of Black African or Latin American self-identity; CAwMS = Subjects with MS of Caucasian self-identity; CA HD = healthy donors without MS of Caucasian self-identity.

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN ancestry. To ascertain whether these frequencies (repre- Discussion sentative of ASCs derived from precursor B cells) affected overall circulating numbers of these cells, we compared flow The major finding of our investigation was the significant cytometry event counts of class-switched and IgM ASCs ethno-ancestry–based differential of peripheral blood ASC among our cohorts. In contrast to heightened frequency subsets examined among subjects with MS but not HD sub- differences across both switched and unswitched classes in jects. Both interventional and observational clinical studies the context of MS, only class-switched (figure 3A) but not reveal a clear pathogenic role for blood-borne, CNS-infiltrating 19–21 IgM+ ASCs were differentially enriched in subjects of Black B cells in MS. There appears to be crosstalk between the African or Latin American identity compared with those of peripheral blood and CNS as demonstrated by clonal se- Caucasian ethno-ancestry (figure 3B). quencing approaches.22,23 For instance, recent observations by Eggers et al.12 suggest that it is predominantly peripheral blood Taken together, our observations show an identity-dependent memory B cells and derived CD138+ plasma cells that are differential of ASC frequency in the context of MS, and associated with CNS pathology during active inflammatory support the existence of alternative underlying immunopa- MS. Continued crosstalk between the blood and CNS may lead thogenesis across different ethno-ancestries. to an accumulation CNS-resident ASCs. Several lines of

Figure 3 Numbers of class-switched but not IgM+ ASCs are elevated according to ethno-ancestry among subjects with MS

Average number of circulating ASCs for each subject cohort as obtained through event counts derived from flow cytometry gates described prior. Class- − switched CD19+IgD ASCs (A); unswitched CD19+IgM+ ASCs (B). Error bars represent SD, p values determined by the 2-sided t test. ASC = antibody-secreting cell; BALA HD = healthy donors without MS of Black African or Latin American self-identity; BALAwMS = Subjects with MS of Black African or Latin American self-identity; CAwMS = Subjects with MS of Caucasian self-identity; CA HD = healthy donors without MS of Caucasian self-identity; NAT = natalizumab therapy.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 7 evidence point to a pernicious ectopic humoral response within ultimately poor reflections of the subtle differences between CNS tissue, likely maintained by CD138+ apoptotic-resistant individuals,30 whereas their use may be also confounded by plasma cells.12,24,e13 These include the existence of pathogenic sociohistoric narratives.e15 Future studies, particularly those autoantibodies,13,14,25,e14 persistence of intrathecal antibodies seeking to unravel the interplay between ethno-ancestry and despite peripheral CD20+ memory B-cell depletion,26 and disease, should emphasize molecular genetic and epigenetic observations of follicle-like aggregates in progressive MS.27 Our delineations. In addition to demonstrably transcending broad study reveals remarkable trends consistent with, and additive ancestral categories, underlying molecular genetic factors may to, previous observations that implicate differential intrathecal be affected by socioeconomic31 and behavioral32,33 influences humoral responses in patients consistently shown to be at risk important to shaping disease incidence and prognosis. of aggressive manifestation of MS.8,9 We note that although our BALAwMS cohort exhibited an increased disease manifesta- Ethno-ancestry–based immunologic differences are long- tion, this current study is not powered to assess the clinical established observations.34 However, these are largely differences. Nonetheless, having been conducted using pe- described in immunity and vaccination studies among pre- ripheral blood, our work highlights the potential value of this sumably healthy individuals.35,36 For instance, total serum accessible compartment for subsequent research into questions IgG concentrations are more likely elevated among “black” of ethno-ancestry and MS. relative to “white” individuals34; African American recipients of HIV gp120 vaccination generated significantly greater Taking into account data supporting a B cell-centric view neutralizing antibody responses compared with whites of MS immunopathogenesis, and previous work showing assessed in a separate study35; similarly, according to National enhanced intrathecal humoral responses among African Health and Nutritional Examination Survey data, seropreva- Americans with MS, we suggest that among BALAwMS, lence of IgG antibody specific to mumps virus was highest CNS-infiltrating B cells may be more likely driven toward among individuals identifying as “non-Hispanic black” and ASC differentiation. Our findings provide plausibility for this lowest among “non-Hispanic whites.”36 Our work is the first to scenario, as ASC frequency used as the primary outcome in directly demonstrate ethno-ancestry–based immunologic dif- our study, alludes to the tendency of B cells to adopt a plas- ference in the context of MS. Nonetheless; there are several mablast or plasma cell fate. Given the importance of migration limitations to our study. These include a cross-sectional design, between peripheral blood and CNS in MS, it is possible that convenience sample, and the limited resolution provided by the tendencies observed in the periphery may be recapitulated broad ethno-ancestral categories. Very few subjects had un- in the CNS. In their retrospective study, Rinker et al.8 do not dergone lumbar puncture procedures as part of standard of care; distinguish leukocyte categories among the overall increase in as such, we were unable to directly relate peripheral differentials intrathecal white blood cells among African American patients with intrathecal antibody burden as reported prior. Despite with MS over Caucasian counterparts. Thus, in addition to these limitations, we provide evidence of an ethno-ancestry– examining the question of intrathecal plasma cell burden, dependent peripheral blood ASC differential concordant with future work with BALAwMS should examine transcriptional observations of intrathecal humoral responses. mechanisms associated with plasma cell fate. In addition, although intrathecal IgM is associated with a poor MS course,15 Over the past 16 years,2,7,37 observational studies revealed there is no study to our knowledge that examines whether greater clinical severity and poorer prognosis among indi- IgM oligoclonal bands are enriched among BALAwMS rela- viduals of the Black African diaspora, including those iden- tive to other subpopulations. Our finding that frequencies of tifying within the spectrum of Latin American identity.2,16 IgM+ ASCs are significantly enhanced in this population Both overt and indirect evidence support ethno-ancestry as warrants such an investigation. However, only class-switched a critical factor in MS. A primary role for B cells in MS and not IgM+ ASC numbers were differentially elevated pathogenesis unequivocally shown in 200811,38 is not only among subjects with MS, pointing to the likely involvement of accentuated in individuals of Black African heritage8 but class-switch recombination mechanisms alongside ASC dif- parallels similar ethno-ancestry–based B cell–driven dysre- ferentiation biology. Our work thus facilitates numerous gulation in other conditions.39,40 For instance, memory avenues for future study in an effort to better understand and B cells from African American patients with systemic lupus treat aggressive MS. erythematosus (SLE) exhibit an activated phenotype relative to Caucasian patients.39 Furthermore, plasmablasts are as- In this report, we used the term “ethno-ancestry” to both sociated with disease activity in SLE. African Americans with highlight and encapsulate the complexity around societal (eth- SLE exhibit a pronounced plasmablast transcriptomic sig- nicity) and biologic (ancestry) heritage. As with other studies nature in conjunction with heightened clinical severity rel- examining the influence of ethno-ancestral categories and MS, ative to Caucasian subjects.40 Thus, our results have we find some value in the use of self-identification according to interdisciplinary implications, as they are consistent with broad labels to establish study cohorts. However, we recognize a burgeoning trend of possibly differentially regulated B-cell that all such labels, whether ethno-ancestry, race, or other var- biology as a nexus through which ethno-ancestry–based iants, ultimately represent combinations of socially constructed disease severity disparities may manifest.39,40 Understanding institutional and cultural designations.28,29 Such categories are what mediates these differences may be key to unlocking

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN major advancements in diagnostics and therapy, both for MS and other maladies. Appendix (continued)

Name Location Role Contribution Acknowledgment The authors thank all individuals consenting to participate in Ivan Diaz, Weill Author Conducted biostatistical data PhD Cornell analysis and manuscript revision this research. They appreciate the members of the Vartanian Medicine for intellectual content Lab for critical scientific discourse and logistic support: Drs. Yinghua Ma and Jennifer Linden, Paige Winokur, Samantha Mason Weill Author Contributed to data acquisition, Kruse- Cornell data analysis, patient/subject Shetty, and Baohua Zhao. They are also appreciative of the Hoyer, MD Medicine recruitment, and manuscript physicians, nurses, and staff of the Weill Cornell Multiple revision Sclerosis Center that helped make this study possible; in Casey Engel, Weill Author Contributed to data acquisition, particular, Dr. Nancy Nealon for patient recruitment and BA Cornell data analysis, patient/subject Medicine recruitment, and manuscript Stacyann Foster for conducting subject blood draws. Finally, revision this project was supported by the Mortimer Sackler Neuro- Melanie Weill Author Contributed to data acquisition, regeneration Fund. The authors are also grateful to the Weill Marcille, BA Cornell data analysis, and patient/ Cornell Clinical and Translational Science Center for Medicine subject recruitment fi additional nancial provision through NIH/NCATS Grants Timothy Weill Author Contributed to study design and TL1-TR-002386 and UL1-TR002384-01. Vartanian, Cornell conceptualization, data MD, PhD Medicine interpretation, and manuscript revision for intellectual content Study funding This study was funded by the Mortimer Sackler Neuro- regeneration Fund and Weill Cornell Clinical and Translational References Science Center NIH/NCATS Grants TL1-TR-002386 and 1. Cree BAC, Khan O, Bourdette D, et al. Clinical characteristics of African Americans vs UL1-TR002384-01. Caucasian Americans with multiple sclerosis. Neurology 2004;63:2039–2045. 2. Ventura RE, Antezana AO, Bacon T, Kister I. Hispanic Americans and African Americans with multiple sclerosis have more severe disease course than Caucasian Disclosure Americans. Mult Scler J 2017;23:1554–1557. K.M. Telesford, U.W. Kaunzner, and J. Perumal report no 3. Weinstock-Guttman B, Ramanathan M, Hashmi K, et al. Increased tissue damage and lesion volumes in African Americans with multiple sclerosis. Neurology 2010;74: disclosures. S.A. Gauthier has received research grant support 538–544. from Genzyme and Mallinckrodt in the fiscal year preceding 4. Howard J, Battaglini M, Babb JS, et al. MRI correlates of disability in African- Americans with multiple sclerosis. PLoS One 2012;7:e43061. the date of the original submission. X. Wu, I. Diaz, M. Kruse- 5. Kimbrough DJ, Sotirchos ES, Wilson JA, et al. Retinal damage and vision loss in Hoyer, C. Engel, M. Marcille, and T. Vartanian report no African American multiple sclerosis patients. Ann Neurol 2015;77:228–236. 6. Al-Kawaz M, Monohan E, Morris E, et al. Differential impact of multiple sclerosis on disclosures. Go to Neurology.org/NN for full disclosures. cortical and deep gray matter structures in African Americans and Caucasian Amer- icans. J Neuroimaging 2017;27:333–338. 7. Caldito NG, Saidha S, Sotirchos ES, et al. Brain and Retinal Atrophy in African- Publication history Americans versus Caucasian-Americans with Multiple Sclerosis: A Longitudinal Received by Neurology: Neuroimmunology & Neuroinflammation Study. Brain 2018;141:3115–3129. fi 8. Rinker JR, Trinkaus K, Naismith RT, Cross AH. Higher IgG index found in African December 13, 2018. Accepted in nal form September 5, 2019. Americans versus Caucasians with multiple sclerosis. Neurology 2007;69:68–72. 9. Seraji-Bozorgzad N, Khan O, Cree BAC, et al. Cerebral gray matter atrophy is associated with the CSF IgG index in African American with multiple sclerosis. J Neuroimaging 2017;65:239. 10. Cepok S, Rosche B, Grummel V, et al. Short-lived plasma blasts are the main B cell Appendix Authors effector subset during the course of multiple sclerosis. Brain 2005;128:1667–1676. 11. Hauser SL, Waubant E, Arnold DL, et al. B-cell depletion with rituximab in relapsing- Name Location Role Contribution remitting multiple sclerosis. N Engl J Med 2008;358:676–688. 12. Eggers EL, Michel BA, Wu H, et al. Clonal relationships of CSF B cells in treatment- Kiel M. Weill Author Designed and conceptualized the naive multiple sclerosis patients. JCI Insight 2017;2:92724. Telesford, Cornell study; analyzed the data; and 13. Elliott C, Lindner M, Arthur A, et al. Functional identification of pathogenic auto- PhD Medicine drafted the manuscript for antibody responses in patients with multiple sclerosis. Brain 2012;135:1819–1833. intellectual content 14. Blauth K, Soltys J, Matschulat A, et al. Antibodies produced by clonally expanded plasma cells in multiple sclerosis cerebrospinal fluid cause demyelination of spinal Ulrike W. Weill Author Contributed to project design, cord explants. Acta Neuropathol 2015;130:765–781. Kaunzner, Cornell rationale, and patient/subject 15. Thangarajh M, Gomez-Rial J, Hedstrom A, et al. Lipid-specific immunoglobulin M in MD, PhD Medicine recruitment CSF predicts adverse long-term outcome in multiple sclerosis. Mult Scler J 2008;14: 1208–1213. Jai Perumal, Weill Author Contributed to data acquisition, 16. Khan O, Williams MJ, Amezcua L, Javed A, Larsen KE, Smrtka JM. Multiple sclerosis MD Cornell data analysis, manuscript writing, in US minority populations: clinical practice insights. Neurol Clin Pract 2015;5: Medicine and patient/subject recruitment 132–142. 17. Kaminski DA, Wei C, Qian Y, Rosenberg AF, Sanz I. Advances in human B cell Susan A. Weill Author Contributed to data acquisition, phenotypic profiling. Front Immunol 2012;3:302. Gauthier, Cornell study design, and patient/subject 18. Halliley JL, Tipton CM, Liesveld J, et al. Long-Lived plasma cells are contained within DO, MPH Medicine recruitment the CD19- CD38hi CD138+ subset in human bone marrow. Immunity 2015;43: 132–145. Xian Wu, Weill Author Conducted biostatistical data 19. Krumbholz M, Meinl I, Kümpfel T, Hohlfeld R, Meinl E. Natalizumab dispropor- MPH Cornell analysis, manuscript writing, and tionately increases circulating pre-B and B cells in multiple sclerosis. Neurology 2015; Medicine manuscript revision for 43:132–145. intellectual content 20. Hauser SL, Bar-Or A, Comi G, et al. Ocrelizumab versus interferon beta-1a in relapsing multiple sclerosis. N Engl J Med 2017;376:221–234.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 9 21. Baker D, Marta M, Pryce G, Giovannoni G, Schmierer K. Memory B cells are major 31. Marrie RA, Cutter G, Tyry T, Vollmer T, Campagnolo D. Does multiple sclerosis- targets for effective immunotherapy in relapsing multiple sclerosis. EBioMedicine associated disability differ between races? Neurology 2006;66:1235–1240. 2017;16:41–50. 32. McGowan PO, Roth TL. Epigenetic pathways through which experiences become 22. Bankoti J, Apeltsin L, Hauser SL, et al. In multiple sclerosis, oligoclonal bands connect linked with biology. Dev Psychopathol 2015;27:637–648. to peripheral B-cell responses. Ann Neurol 2014;75:266–276. 33. Saadi A, Himmelstein DU, Woolhandler S, Mejia NI. Racial disparities in neuro- 23. Palanichamy A, Apeltsin L, Kuo TC, et al. Immunoglobulin class-switched B cells logic health care access and utilization in the United States. Neurology 2017;88: form an active immune axis between CNS and periphery in multiple sclerosis. Sci 2268–2275. Translational Med 2014;6:248ra106. 34. Shackelford PG, Granoff DM, Nahm MH, et al. Relation of age, race, and allotype to 24. Winges KM, Gilden DH, Bennett JL, Yu X, Ritchie AM, Owens GP. Analysis of multiple immunoglobulin subclass concentrations. Pediatr Res 1985;19:846–849. sclerosis cerebrospinal fluid reveals a continuum of clonally related antibody-secreting 35. Montefiori DC, Metch B, McElrath MJ, et al. Demographic factors that influence the cells that are predominantly plasma blasts. J Neuroimmunol 2007;192:226–234. neutralizing antibody response in recipients of recombinant HIV-1 gp120 vaccines. 25. Ligocki AJ, Rivas JR, Rounds WH, et al. A distinct class of antibodies may be an J Infect Dis 2004;190:1962–1969. indicator of gray matter autoimmunity in early and established relapsing remitting 36. Kutty PK, Kruszon Moran DM, Dayan GH, et al. Seroprevalence of antibody to multiple sclerosis patients. ASN Neuro 2015;7:175909141560961. mumps virus in the US population, 1999–2004. J Infect Dis 2010;202:667–674. 26. Cross AH, Stark JL, Lauber J, Ramsbottom MJ, Lyons JA. Rituximab reduces B cells and 37. Kaufman MD, Johnson SK, Moyer D, Bivens J, Norton HJ. Multiple sclerosis severity T cells in cerebrospinal fluid of multiple sclerosis patients. J Neuroimmunol 2006;180:63–70. and progression rate in African Americans compared with whites. Am J Phys Med 27. Lovato L, Willis SN, Rodig SJ, et al. Related B cell clones populate the meninges and Rehabil 2003;82:582–590. parenchyma of patients with multiple sclerosis. Brain 2011;134:534–541. 38. Bar-Or A, Calabresi PAJ, Arnold D, et al. Rituximab in relapsing-remitting multiple 28. Foster MW, Sharp RR. Race, ethnicity, and genomics: social classifications as proxies sclerosis: a 72-week, open-label, phase I trial. Ann Neurol 2008;63:395–400. of biological heterogeneity. Genome Res 2002;12:844–850. 39. Menard LC, Habte S, Gonsiorek W, et al. B cells from African American lupus patients 29. Collins FS. What we do and don’t know about “race,”“ethnicity,” genetics and health exhibit an activated phenotype. JCI Insight 2016;1;e87310. at the dawn of the genome era. Nat Genet 2004;36:S13–S15. 40. Banchereau R, Hong S, Cantarel B, et al. Personalized immunomonitoring uncovers 30. Bryc K, Durand EY, Macpherson JM, Reich D, Mountain JL. The genetic ancestry of molecular networks that stratify lupus patients. Cell 2016;165:551–565. African Americans, Latinos, and European Americans across the United States. Am J Hum Genet 2015;96:37–53. References e1-e15 are available at: links.lww.com/NXI/A162.

10 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN ARTICLE OPEN ACCESS Lymphocyte pharmacodynamics are not associated with autoimmunity or eﬃcacy after alemtuzumab

Heinz Wiendl, MD, PhD, Matthew Carraro, MD, Giancarlo Comi, MD, Guillermo Izquierdo, MD, PhD, Correspondence Ho Jin Kim, MD, PhD, Basil Sharrack, MD, PhD, FRCP, FAAN, Carlo Tornatore, MD, Nadia Daizadeh, PhD, Dr. Wiendl [email protected] Luke Chung, MD, MPH, Alan K. Jacobs, MD, FAAN, Richard J. Hogan, PhD, Linda V. Wychowski, PhD, and Bart Van Wijmeersch, MD, PhD, On behalf of the CARE-MS I, CARE-MS II, and CAMMS03409 Investigators

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e635. doi:10.1212/NXI.0000000000000635 Abstract Objective To examine the association between peripheral blood lymphocyte pharmacodynamics and autoimmune adverse events (AEs) or return of disease activity in alemtuzumab-treated patients with relapsing-remitting MS.

Methods Patients received 2 alemtuzumab courses (12 mg/d IV; 5 days at baseline, 3 days 12 months later) in the 2-year Comparison of Alemtuzumab and Rebif Efficacy in Multiple Sclerosis studies (NCT00530348 and NCT00548405) and could then receive as-needed alemtuzumab or other disease-modifying therapy in a 4-year extension (NCT00930553). Lymphocytes were phenotyped quarterly over 2 years using fluorescence-activated cell sorting. Pharmacodynamic assessments included counts of total lymphocytes, CD3+ T cells, CD4+/CD8+ T cells (total/ + naive/memory/regulatory [Treg]), and CD19 B cells (total/immature/mature/memory) and + + + ratios of CD19 (total/immature/mature/memory) to Treg (CD4 /CD8 ) counts. Assessed autoimmune AEs included immune thrombocytopenia, nephropathies, and thyroid events. Efficacy assessments included relapses, 6-month confirmed disability worsening (CDW), and MRI disease activity.

Results Lymphocyte repopulation patterns, including ratios between distinct lymphocyte subsets (e.g., + ﬁ ﬀ CD19 to Treg cell count ratios), showed no signi cant di erences over 2 years in patients developing/not developing autoimmune AEs, relapses, CDW, or MRI activity through 6 years following alemtuzumab. Lymphocyte kinetics were also unrelated to multiple autoimmune AEs or extreme clinical phenotypes.

Conclusions Repopulation kinetics of the evaluated peripheral lymphocyte subsets did not predict autoimmune AE occurrence or disease activity, including return of disease activity after 2 alemtuzumab courses. Further study is needed to investigate potential antigen-level markers of treatment response.

From the University of Munster¨ (H.W.), Munster,¨ Germany; Novant Health (M.C.), Charlotte, NC; University Vita-Salute San Raffaele (G.C.), Milan, Italy; Virgen Macarena University Hospital (G.I.), Seville, Spain; Research Institute and Hospital of National Cancer Center (H.J.K.), Goyang, South Korea; NIHR Sheffield Biomedical Research Centre, Sheffield Teaching Hospitals, University of Sheffield (B.S.), Sheffield, United Kingdom; Georgetown University Medical Center (C.T.), Washington, DC; Sanofi (N.D., L.C., A.K.J.), Cambridge, MA; Eloquent Scientific Solutions (R.J.H.), Sydney, NSW, Australia; Eloquent Scientific Solutions (L.V.W.), Philadelphia, PA; and Rehabilitation & MS-Centre Overpelt (B.V.W.), BIOMED, Hasselt University, Hasselt, Belgium.

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 Sanofi.

CARE-MS I, CARE-MS II, and CAMMS03409 coinvestigators are listed at links.lww.com/NXI/A163. 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 © 2019 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary ffi AE = adverse event; Breg = regulatory B; CARE-MS = Comparison of Alemtuzumab and Rebif E cacy in Multiple Sclerosis; CDW = confirmed disability worsening; EDSS = Expanded Disability Status Scale; MMRM = mixed-effects model for repeated measures; NEDA = no evidence of disease activity; RRMS = relapsing-remitting MS; Treg = regulatory T.

Alemtuzumab is a humanized monoclonal antibody that se- response to previous therapy (CARE-MS II; NCT00548405; lectively depletes circulating CD52-expressing B and T aged 18–55 years).6,7 In the 2-year CARE-MS studies (conducted lymphocytes.1,2 Following depletion, a distinctive pattern at 178 academic medical centers or clinical practices in 23 of lymphocyte repopulation potentially leads to a rebalanced countries; starting in September 2007), patients in the alemtu- – immune system.3 5 zumab arm received 2 courses of alemtuzumab 12 mg/d IV on 5 consecutive days at baseline and on 3 consecutive days 12 In phase III trials, patients with relapsing-remitting MS months later.6,7 Patients who completed the phase III studies (RRMS) receiving alemtuzumab experienced significant clini- could enter the 4-year CARE-MS extension (CAMMS03409; cal and MRI efficacy improvements vs subcutaneous interferon NCT00930553), in which they could receive additional beta-1a over 2 years.6,7 Efficacy was maintained over 5 addi- courses of alemtuzumab (12 mg/d on 3 consecutive days ≥12 – tional years in 2 extension studies.8 14 The most frequent ad- months after the most recent dose) as needed for relapse or verse events (AEs) with alemtuzumab were infusion-associated MRI activity or receive other licensed disease-modifying reactions; autoimmune AEs also occurred, including thyroid therapy at the investigator’s discretion.10,11 Patients com- – events, immune thrombocytopenia, and nephropathies.6 13,15 pleting the CARE-MS extension study could enroll in an additional extension, the 5-year long-Term follow-up study for Pharmacodynamic changes after lymphocyte depletion, in- multiple sclerOsis Patients who have completed the Alemtu- cluding different repopulation patterns among cell subsets, Zumab extension (TOPAZ) study (NCT02255656), in which may account for the overall efficacy of alemtuzumab in RRMS further evaluation is ongoing.8,9,12,13 – and its associated AE profile.1,16 18 Furthermore, differences among patients’ lymphocyte repopulation patterns have been Post hoc analysis hypothesized to explain individual differences in drug re- Two-year lymphocyte pharmacodynamics were assessed in fi sponse and create the environment for autoimmune AEs in alemtuzumab-treated patients (N = 802), strati ed by whether some patients.19 This may include establishment of perma- they experienced autoimmune AEs, relapse, 6-month con- fi nent vs more transient influences on tolerance-associated rmed disability worsening (CDW), or MRI disease activity at immune regulatory network dynamics.20 any time point within 6 years of follow-up. fi However, biomarkers that would predict response to alemtu- Autoimmune AEs were de ned as any of the following, zumab or selection of patients at risk for development of au- documented at any point within 6 years of follow-up: thyroid toimmune events have not been identified.21 Although AEs (excluding asymptomatic abnormal laboratory inves- fi increased serum interleukin-21 levels before alemtuzumab have tigations), immune thrombocytopenia (de ned according to been associated with autoimmune disorders posttreatment, the diagnostic criteria outlined by an international working 23 widespread applicability of such an assay has not been estab- group ), or autoimmune nephropathy (in particular, anti- lished.22 Furthermore, no biomarkers exist for predicting re- glomerular basement membrane disease or membranous glo- currence of disease activity after 2 alemtuzumab courses. The merulonephropathy). Monitoring for autoimmune AEs (as current post hoc analysis methodically assesses whether phar- part of the safety monitoring program) included quarterly macodynamic patterns of major peripheral blood lymphocyte thyroid function tests, monthly hematology tests for immune populations are associated with autoimmune AEs or MS dis- thrombocytopenia, and monthly serum creatinine tests and ease activity over 6 years after initiating alemtuzumab. urinalysis with microscopy for nephropathy, as well as spontaneous reporting of AEs and serious AEs. Education on signs and symptoms of autoimmune AEs was provided to health care Methods providers and patients. All monitoring began at baseline and continued until 4 years after the last alemtuzumab adminis- Design of CARE-MS and extension studies tration, or until study end, whichever occurred later. The 4-year of alemtuzumab monitoring period restarted if patients received additional The efficacy and safety of alemtuzumab were established in 2 courses of alemtuzumab. Treatment for autoimmune AEs was phase III studies against subcutaneous interferon beta-1a in at the discretion of the treating neurologist in consultation with patients with active RRMS who were either treatment naive local endocrinologists, hematologists, or nephrologists. (Comparison of Alemtuzumab and Rebif Efficacy in Multi- ple Sclerosis [CARE-MS] I; ClinicalTrials.gov trial identifier: Relapses were defined as new neurologic symptoms attrib- NCT00530348; aged 18–50 years) or had an inadequate utable to MS lasting ≥48 hours with an objective change in

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN neurologic examination. Six-month CDW was defined as an defined in the table. To control for any effect of changes in ≥ fi + increase of 1.0 Expanded Disability Status Scale (EDSS) Treg cell levels over time, the kinetic pro les of CD19 B cells point (or ≥1.5 points if the baseline EDSS score = 0), con- (total, immature, mature, and memory) were also assessed fi + + rmed over 6 months. The EDSS score was assessed quarterly relative to CD4 and CD8 Treg cell counts. and at the time of suspected relapse by raters who were blinded throughout the follow-up period to core study Statistical analyses treatment assignment and treatment history. MRI disease All data were analyzed by alemtuzumab treatment course. activity was defined as new gadolinium-enhancing T1 lesions Patients were categorized as “active” or “nonactive” based on on current MRI or new/enlarging T2 hyperintense lesions defined events: presence or absence of autoimmune AEs, since last MRI. MRI was assessed annually by imaging spe- relapses, CDW, and MRI disease activity. To compare the cialists blinded to core study treatment assignment and depletion at month 1 and month 13, and the differential retreatment history. If patients experienced autoimmune events, constitution between groups, linear mixed-effects models for relapse, CDW, or MRI disease activity at any time during the repeated measures (MMRMs) were undertaken with lym- follow-up period, they were classified as having had that type phocyte parameters as the outcome variable. Explanatory of event from month 0 onward. variables included age and baseline value and either autoimmune AEs, relapse, CDW, or MRI disease activity. Separate In support of the primary analyses, additional analyses MMRMs were fit for separate events as explanatory variables. assessed lymphocyte pharmacodynamics in patients with Multiple hypothesis testing was adjusted using the Bonferroni 24 multiple autoimmune AEs and patients with various “extreme correction and the Benjamini-Hochberg procedure. efficacy” phenotypes, including patients with or without sustained “no evidence of disease activity” (NEDA; absence of Standard protocol approvals, registrations, relapse, CDW, and MRI disease activity, sustained over years and patient consents 0–6); with or without both relapse and CDW (clinical NEDA); The CARE-MS trials and CAMMS03409 extension are regis- with or without both severe relapse and CDW; and with CDW, tered with ClinicalTrials.gov (NCT00530348, NCT00548405, 6-month confirmed disability improvement (≥1.0-point EDSS and NCT00930553). All procedures were approved by the score decrease from baseline [assessed in patients with baseline institutional ethics review boards of participating sites. Patients EDSS score ≥2.0]), or stable EDSS score. provided written informed consent. Data availability Lymphocyte dynamics Qualified researchers may request access to patient-level data Blood cell counts in the CARE-MS studies were performed and related study documents including the clinical study re- monthly. Lymphocytes were phenotyped using fluorescence- port, study protocol with any amendments, blank case report activated cell sorting (Quest Diagnostics; Exton, PA) at base- form, statistical analysis plan, and data set specifications. line and quarterly thereafter, as well as at months 1 and 13 Patient-level data will be anonymized and study documents (i.e., 1 month after receiving alemtuzumab courses 1 and 2, will be redacted to protect the privacy of trial participants. respectively). Peripheral blood mononuclear cells were stained Further details on Sanofi’s data-sharing criteria, eligible using a T-cell panel of monoclonal antibodies against studies, and process for requesting access can be found at CD45RA-FITC (clone L48), CD27-PE (clone L128), CD3- clinicalstudydatarequest.com. PerCP-Cy5.5 (clone SK7), CD25-PE-Cy7 (clone 2A3), CD127-APC (clone 40131), and CD4-APC-Cy7 (clone SK3), and a B-cell panel of monoclonal antibodies against CD27- Table Lymphocyte subset phenotypes FITC (clone L128), IgD-PE (clone IA6-2), CD19-PerCP- Lymphocyte Lymphocyte Cy5.5 (clone SJ25C1), CD10-PE-Cy7 (clone HI10a), CD38- designation subset name Phenotype APC (clone HB7), and CD69-APC-Cy7 (clone FN50) (all + + + + + + antibodies from BD Biosciences; San Jose/San Diego, CA Naive CD4 Tcells Naive CD3 4 T cells CD4 45RA 27 − [except CD127-APC: from R&D Systems; Minneapolis, Memory CD4+ T cells Memory CD3+4+ T cells CD4+45RA

MN]). Lymphocyte data from the CARE-MS I and II studies + + + + int/br+ lo/− CD4 Treg cells CD3 4 regulatory T cells CD4 25 127 were pooled for analysis. Absolute cell counts were analyzed for total levels of lymphocytes; all CD3+ T lymphocytes, CD4+ Naive CD8+ Tcells Naive CD3+8+ T cells CD8+45RA+27+ + + − T lymphocytes, and CD8 T lymphocytes; and CD19 Memory CD8+ T cells Memory CD3+8+ T cells CD8+45RA

B lymphocytes. + + + + int/br lo/− CD8 Treg cells CD3 8 regulatory T cells CD8 25 127

A lymphocyte substudy was conducted at select study centers, Immature B cells Immature B cells IgD+CD19+27-38+10+

− in which additional lymphocyte phenotype analyses were Mature B cells Mature naive B cells IgD+CD19+27-38+10 performed for the naive, memory, and regulatory (Treg) + + subsets of CD4+ and CD8+ T lymphocytes and the immature, Memory B cells Memory B cells CD19 27 mature, and memory subsets of CD19+ B lymphocytes Abbreviation: Treg = regulatory T. (n = 146). The phenotypes of all cell subsets assessed are

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 3 Results CD4+, CD8+, and CD19+ lymphocyte depletion and repopulation patterns over 2 years are not Patients and overall lymphocyte profiles associated with clinical and MRI efficacy A total of 811 patients were treated with alemtuzumab 12 mg parameters over 6 years in CARE-MS I and II; baseline characteristics were reported Total lymphocyte, CD4+ and CD8+ T-lymphocyte, and 6,7 previously (mean age 34.0 [SD 8.24] years). Total lym- CD19+ B-lymphocyte depletion and repopulation patterns + + + phocyte, CD4 and CD8 T-lymphocyte, and CD19 were also similar in patients who did or did not experience B-lymphocyte counts following alemtuzumab treatment were relapses, CDW, or MRI disease activity through 6 years fol- fi assessed in 802 patients ( gure 1). lowing alemtuzumab treatment (figure 3, A–C). Among the various efficacy subgroups, the timing and magnitude of + + + CD4 , CD8 , and CD19 lymphocyte depletion the changes in total CD4+ and CD8+ T- and CD19+ and repopulation patterns over 2 years are not B-lymphocyte levels after treatment, as well as the repo- associated with risk of an autoimmune event pulation pattern, were comparable (figure 3, A–C). Results over 6 years for CD3+ T-lymphocyte counts were similar (figure e-1, Of the 802 patients, 447 (56%) developed autoimmune links.lww.com/NXI/A160). AEs by definition. There was no significant overall difference in depletion or repopulation patterns of Analysis of repopulation of total lymphocytes, CD4+ and CD8+ T lymphocytes, or specific lymphocyte subsets did not show CD19+ B lymphocytes (assessed over 2 years) in patients association with the risk of an autoimmune who did or did not experience autoimmune AEs over 6 event or disease activity over 6 years years (figure 2). The timing and magnitude of the changes In the expanded phenotype analysis of CD4+ and CD8+ in total T- and B-lymphocyte levels after treatment, as well T-lymphocyte subsets (naive, memory, and Treg cells) and as the repopulation patterns, were comparable between CD19+ B-lymphocyte subsets (immature, mature, and memory patients with or without autoimmune AEs (figure 2). cells) in a subpopulation of patients (n = 146), repopulation Results for CD3+ T-lymphocytecountsweresimilar kinetics over 2 years did not differ overall in patients with or between the subgroups (figure e-1, links.lww.com/NXI/ without autoimmune AEs, relapses, CDW, or MRI disease A160). activity through 6 years (figure 4, A–C).

Figure 1 Median lymphocyte counts in patients treated with alemtuzumab 12 mg

Median total lymphocyte, CD4+ and CD8+ T-lymphocyte, and CD19+ B-lymphocyte counts in the overall pooled Comparison of Alemtuzumab and Rebif Efficacy in Multiple Sclerosis population treated with alemtuzumab 12 mg. N value range over 24 months: 752–802 patients. Vertical lines indicate administration of alemtuzumab. Dashed horizontal lines represent LLN for each of the lymphocyte subsets presented. IQR = interquartile range; LLN = lower limit of normal.

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN Figure 2 Median lymphocyte counts in alemtuzumab-treated patients with or without autoimmune AEs

(A) Median total lymphocyte, (B) CD4+ and (C) CD8+ T-lymphocyte, and (D) CD19+ B-lymphocyte counts in patients from the pooled Compari- son of Alemtuzumab and Rebif Efficacy in Multiple Sclerosis trials treated with alemtuzumab 12 mg, stratified by the presence or absence of autoimmune events. N value range over 24 months: 415–447 patients without autoimmune AEs; 337–355 patients with autoimmune AEs. Vertical lines indicate administration of alemtuzumab. AE = adverse event; IQR = interquartile range; LLN = lower limit of normal.

For both the autoimmune AE and efficacy assessments, the key posttreatment blood count (month 1) and reconstituted pharmacodynamic characteristics of the lymphocyte subsets steadily after each treatment course (figure 4, A and B). CD8+ following alemtuzumab treatment were comparable regardless Treg cells showed less pronounced depletion compared with ffi + fi of autoimmune event or e cacy status. Absolute counts of CD4 Treg cells ( gure 4, A and B). Absolute counts of mature CD4+ and CD8+ T-lymphocyte subsets (naive, memory, and and memory CD19+ B lymphocytes decreased posttreatment fi Treg cells) were typically lowest at the time of rst and were lowest at month 1, whereas absolute counts of

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 5 Figure 3 Median lymphocyte counts in alemtuzumab-treated patients with or without disease activity

Median total lymphocyte, CD4+ and CD8+ T-lymphocyte, and CD19+ B-lymphocyte counts in Comparison of Alemtuzumab and Rebif Efficacy in Multiple Sclerosis patients treated with alemtuzumab 12 mg, stratified by the presence or absence of (A) relapse, (B) 6-month CDW, and (C) MRI disease activity. N value range over 24 months: 356–388 patients without relapse; 395–414 patients with relapse; 567–606 patients without CDW; 184–196 patients with CDW; 263–280 patients without MRI activity; 485–516 patients with MRI activity. Vertical lines indicate administration of alemtuzumab. CDW = confirmed disability worsening; IQR = interquartile range; LLN = lower limit of normal.

immature B lymphocytes increased posttreatment and were Additional analyses in patients with or without multiple au- highest at month 3 (figure 4C). Mature B cells repopulated to toimmune AEs or “extreme efficacy” phenotypes (e.g., with or approach baseline levels by month 6, followed by a rise above without sustained NEDA) also failed to find any significant baseline levels; memory B cells repopulated more slowly overall differences in depletion or repopulation profiles (data fi (figure 4C). not shown). Of note, our analysis did not reveal a speci c lymphocyte pattern in the first 2 years that would allow In light of a recently put forward hypothesis that variations in a prediction of the return of disease activity after the second the efficacy and safety of alemtuzumab may reflect B-cell course of alemtuzumab. hyper-repopulation in the absence of effective T-cell regulation in some patients,19 we additionally visualized our data in an alternative manner, by normalizing the kinetic profiles of Discussion + + ff B cells against CD4 and CD8 Treg cell levels. No di erences B and T lymphocytes play an integral role in the disease in depletion or repopulation kinetics were seen in patients process in MS.25 By selectively depleting B and with or without autoimmune events, relapse, CDW, or MRI T lymphocytes, alemtuzumab induces a lymphopenia that is activity through 6 years (figure e-2, A and B, links.lww.com/ measurable in the peripheral blood. The distinct repopulation NXI/A161). patterns of B and T subsets following alemtuzumab

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN Figure 4 Lymphocyte subset counts in alemtuzumab-treated patients with or without autoimmune AEs or disease activity

(A) CD4+ T-, (B) CD8+ T-, and (C) CD19+ B-lymphocyte subset counts in Comparison of Alemtuzumab and Rebif Efficacy in Multiple Sclerosis patients treated with alemtuzumab 12 mg, stratified by the presence or absence of autoimmune AEs, relapse, 6-month CDW, and MRI disease activity. N value range over 24 months: 29–72 patients without autoimmune AEs; 30–76 patients with autoimmune AEs; 23–72 patients without relapse; 27–75 patients with relapse; 44–108 patients without CDW; 14–38 patients with CDW; 23–63 patients without MRI activity; 35–84 patients with MRI activity. AE = adverse event; CDW = confirmed disability worsening; IQR = interquartile range; Treg = regulatory T.

26,27 28,29 treatment, which have been described previously, lead to MS, a relative increase in Treg and memory T-cell counts, quantitative and qualitative changes in immune regulatory and a potential shift from a pro- to anti-inﬂammatory envi- networks. These changes include suppression of memory ronment (driven by diﬀerential reconstitution of T-cell B cells, which may play several roles in the pathogenesis of subsets).26,27,30,31 Despite this knowledge, few studies have

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 7 attempted to correlate aspects of these pharmacodynamic T lymphocytes (CD3+, CD4+, and CD8+), and CD19+ changes with either safety or efficacy or the need for additional B lymphocytes were similar at months 1 and 13 in patient courses of alemtuzumab. subgroups stratified by whether they experienced a relapse or CDW event subsequent to the time point being assessed. The The present study assessed lymphocyte dynamics over the rate of lymphocyte repopulation was also similar between first 2 years following alemtuzumab initiation in relation to the patients who did or did not experience a relapse or CDW after development of autoimmune events or the return of disease receiving alemtuzumab.34 activity over 6 years. This enabled a methodical evaluation of the prognostic value of lymphocyte depletion and repopula- The findings of this study imply that the pharmacodynamic tion patterns, specifically focusing on the putatively most patterns of major lymphocyte populations in the peripheral important populations (CD4+/CD8+ T cells and CD19+ blood within the 2 years following alemtuzumab initiation do B cells), including key subsets such as Treg cells. The questions not have prognostic value. Peripheral immune phenotyping addressed were whether differences in the depletion or may be too superficial or not sufficiently sensitive to detect repopulation patterns of these cell types are associated with differences among individuals in immune regulatory and tol- (1) development of autoimmune AEs or (2) efficacy (in- erance networks that likely occur after alemtuzumab treat- cluding the relevant question of the need for retreatment). In ment. Of particular interest, a recent study of patients with MS answer to the first question, our analysis demonstrated the demonstrated that some self-antigens that are expressed on repopulation kinetics of the tested lymphocyte populations peripheral B memory cells are also expressed in the brain, and did not differ in alemtuzumab-treated patients with or without T cells that become autoreactive to the antigen can also mi- autoimmune AEs over 6 years. Our findings are consistent grate to brain MS lesions.35 Further research on the effects of with a previous study demonstrating that T-cell regulation alemtuzumab at the antigen-specific level and within the CNS does not play a role in the development of autoimmunity in will likely provide additional insights into its overall mecha- alemtuzumab-treated patients; instead, the autoimmunity nism of action and underlying mechanisms of autoimmune may be largely driven by reduced thymopoiesis, a restricted AEs, clinical response, and potential prognosis. T-cell repertoire, and homeostatic expansion of T cells that have escaped depletion.32 This is also in line with a study by Limitations of our study include the caveats that are inherent Jones et al.33, which suggested that increased levels of in any post hoc analysis. In addition, the lymphocyte data in interleukin-21 may drive cycles of T-cell expansion and apo- this study only extend for the first 2 years after initiation of ptosis after alemtuzumab, thus increasing the opportunities alemtuzumab treatment; longer-term follow-up may have for T cells to encounter self-antigen that leads to autoim- provided additional insights. As the depth of immune phe- mune AEs. notyping in our study was limited, the pharmacodynamics of other leukocyte subsets that are linked to response to alem- Importantly, our study could not confirm what one of the tuzumab (i.e., natural killer cells, dendritic cells, and regula- prevailing hypotheses in this area of research would suggest: tory B [Breg] cells) need to be further investigated. For namely, that an overshoot of B cells is a likely contributor to example, relative increases in CD56bright natural killer 36,37 the risk of disease exacerbation and/or autoimmune de- cells and Breg cells have been reported following initiation regulation. That hypothesis is based on the observation that of alemtuzumab, along with decreased proportions of den- after initial depletion with alemtuzumab treatment, B cells in dritic cell subsets that are capable of eliciting detrimental many patients begin to repopulate to levels exceeding those immune responses in MS.37 Differences in the pharmacody- observed before treatment, with implications for safety based namics of these subsets could also underlie autoimmune on B-cell hyper-repopulation in the absence of effective T-cell events or differences in efficacy responses in patients treated regulation.19 The results from our analysis demonstrated that with alemtuzumab. Of particular interest in the case of pe- the repopulation kinetics of B cells, including memory B cells, ripheral Breg cells, a recent study showed levels of the did not differ in alemtuzumab-treated patients with or without CD19+CD24hiCD38hi cell subset to deplete with the onset of autoimmune AEs over 6 years. Furthermore, there were no severe relapses and repopulate during recovery, although this ff di erences either in the B to Treg cell ratios or the repopu- observation was based on data from a single alemtuzumab- lation patterns of T-cell subsets between patients with or treated patient.38 Further analysis of the pharmacodynamics without autoimmune AEs. of B-cell clones may provide additional insight into autoimmune events in alemtuzumab-treated patients. Finally, in Our study also provides evidence that more robust or addition to studying peripheral blood, further research on accelerated lymphocyte repopulation does not predict the lymphoid organs and the CSF is needed to improve un- return of clinical or MRI disease activity in alemtuzumab- derstanding of the interrelation of different compartments treated patients over 6 years. These findings extend those contributing to immune regulation after alemtuzumab from a previous analysis that examined the relationship treatment. between lymphocyte pharmacodynamics and clinical efficacy over the course of the CARE-MS core (2-year) studies. In that The current analyses suggest that differences in depletion analysis, median cell counts for total lymphocytes, or repopulation kinetics of the major subpopulations of

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN peripheral blood lymphocytes have no prognostic value in Group. B. Van Wijmeersch reports receiving research and travel alemtuzumab-treated patients with RRMS. However, grants, honoraria for MS-expert advice, and speaking fees from repopulation patterns of specific B-cell clones with the Bayer Schering, Biogen,MerckSerono,Novartis,Roche,Sanofi, potential to express autoimmune antibodies, and the rele- and Teva. Go to Neurology.org/NN for full disclosures. Funding vance of those patterns to autoimmune AE occurrence, information is provided at the end of the article. remain to be determined. Lymphocyte repopulation kinetics, at the level in the present analysis, likely cannot be Publication history used to predict the need for administration of additional Received by Neurology: Neuroimmunology & Neuroinflammation courses of alemtuzumab. May 15, 2019. Accepted in final form September 5, 2019.

Acknowledgment The authors and Sanoﬁ thank the patients for their participation in the CARE-MS I, CARE-MS II, and Appendix Authors

CAMMS03409 studies and the steering committees and the Name Location Role Contribution investigators. Critical review of the manuscript for medical Heinz University of Author Acquisition and accuracy was provided by Darren P. Baker, PhD, Ericka M. Wiendl, MD, Munster,¨ Munster,¨ interpretation of Bueno, PhD, and Colin P. Mitchell, PhD, of Sanoﬁ. PhD Germany the data and critical revision of the manuscript Study funding This study was supported by Sanoﬁ and Bayer HealthCare Matthew Novant Health, Author Acquisition and Carraro, MD Charlotte, NC interpretation of Pharmaceuticals. Prof. H. Wiendl was supported by the the data and critical Deutsche Forschungsgemeinschaft (DFG) Grant CRC128 revision of the manuscript Project A09, and the Kompetenznetz Multiple Sklerose (Competence Network for Multiple Sclerosis) funded by the Giancarlo University Vita- Author Acquisition and Comi, MD Salute San Raffaele, interpretation of Federal Ministry of Education and Research (FKZ Milan, Italy the data and critical 01GI1308B 01GI0907). revision of the manuscript

Disclosures Guillermo Virgen Macarena Author Acquisition and Izquierdo, University Hospital, interpretation of H. Wiendl reports receiving consulting and/or speaking fees and MD, PhD Seville, Spain the data and critical grant/research support from Bayer, Bayer Schering Pharma, revision of the manuscript Biogen, Elan Corporation, Lilly, Lundbeck, Merck Serono, Novartis, Novo Nordisk, Sanofi, and Teva Neuroscience. M. Ho Jin Kim, Research Institute Author Acquisition and MD, PhD and Hospital of interpretation of the Carraro reports receiving speaking and consulting fees and ad- National Cancer data and critical visory honoraria from Biogen, Genentech, Genzyme, and Mal- Center, Goyang, revision of the linckrodt. G. Comi reports receiving consulting fees from South Korea manuscript Actelion,Bayer,MerckSerono,Novartis,Sanofi, and Teva and Basil NIHR Sheffield Author Acquisition and Sharrack, Biomedical Research interpretation of lecture fees from Bayer, Biogen Dompé, Merck Serono, MD, PhD, Centre, Sheffield the data and critical Novartis, Sanofi, Serono, Symposia International Foundation, FRCP, FAAN Teaching Hospitals, revision of the University of manuscript and Teva. G. Izquierdo reports receiving speaking and advisory Sheffield, Sheffield, fees from Almirall, Bayer, Biogen, Merck Serono, Novartis, United Kingdom Roche, Sanofi, and Teva. H.J. Kim reports receiving consulting Carlo Georgetown Author Acquisition and and/or speaking fees from Bayer, Biogen,Celltrion,Eisai,Gen- Tornatore, University Medical interpretation of zyme,HanAllBioPharma,MedImmune,MerckSerono, MD Center, the data and critical Washington, DC revision of the Novartis, Teva-Handok, and UCB; research support from manuscript Genzyme, Merck Serono, Ministry of Science & ICT, Teva- Nadia Sanofi, Cambridge, Author Design and Handok, and UCB; serving as a steering committee member for Daizadeh, MA and conceptualization MedImmune; and serving as a coeditor for Multiple Sclerosis PhD statistician of the analysis; Journal—Experimental, Translational, and Clinical analysis of the data; and as an as- interpretation of sociate editor for the Journal of Clinical Neurology.B.Sharrack the data; and critical revision of reports receiving research and travel grants, honoraria for expert the manuscript advice on MS, and speaking fees from Biogen, Merck, Novartis, fi Luke Chung, Sanofi, Cambridge, Author Design and Roche, Sano , and Teva. C. Tornatore reports receiving hono- MD, MPH MA conceptualization raria for attending advisory boards and research funding from of the analysis; Bayer, Biogen, Novartis, and Sanofi.N.Daizadeh,L.Chung,and interpretation of the data; and A.K. Jacobs report receiving personal compensation as employ- critical revision of ees of Sanofi. R.J. Hogan and L.V. Wychowski report receiving the manuscript personal compensation as employees of Envision Pharma Continued

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 9 14. Ziemssen T, Thomas K. Alemtuzumab in the long-term treatment of relapsing- Appendix (continued) remitting multiple sclerosis: an update on the clinical trial evidence and data from the real world. Ther Adv Neurol Disord 2017;10:343–359. 15. Wiendl H, Bourdette D, Ciccarelli O. Can immune reprogramming with alemtuzu- Name Location Role Contribution mab induce permanent remission in multiple sclerosis? Neurology 2017;89: 1098–1100. Alan K. Sanofi, Cambridge, Author Design and 16. Genzyme Therapeutics, Ltd. LEMTRADA (alemtuzumab): summary of product Jacobs, MD, MA conceptualization characteristics. 2017. Available at: ema.europa.eu/docs/en_GB/document_library/ FAAN of the analysis; EPAR_-_Product_Information/human/003718/WC500150521.pdf. Accessed Feb- interpretation of ruary 23, 2018. the data; and 17. Fox EJ. Alemtuzumab in the treatment of relapsing-remitting multiple sclerosis. Ex- critical revision of pert Rev Neurother 2010;10:1789–1797. the manuscript 18. Turner MJ, Lamorte MJ, Chretien N, et al. Immune status following alemtuzumab treatment in human CD52 transgenic mice. J Neuroimmunol 2013;261:29–36. Richard J. Eloquent Scientific Author Interpretation of 19. Baker D, Herrod SS, Alvarez-Gonzalez C, Giovannoni G, Schmierer K. Interpreting Hogan, PhD Solutions, Sydney, the data and lymphocyte reconstitution data from the pivotal phase 3 trials of alemtuzumab. JAMA NSW, Australia drafted the Neurol 2017;74:961–969. manuscript for 20. Wiendl H, Calabresi PA, Meuth SG. Defining response profiles after alemtuzumab: intellectual content rare paradoxical disease exacerbation. Neurology 2018;90:309–311. 21. Dörr J, Baum K. Alemtuzumab in the treatment of multiple sclerosis: patient selection Linda V. Eloquent Scientific Author Interpretation of and special considerations. Drug Des Devel Ther 2016;10:3379–3386. Wychowski, Solutions, the data and 22. Azzopardi L, Thompson SA, Harding KE, et al. Predicting autoimmunity after PhD Philadelphia, PA drafted the alemtuzumab treatment of multiple sclerosis. J Neurol Neurosurg Psychiatry 2014;85: manuscript for 795–798. intellectual content 23. Rodeghiero F, Stasi R, Gernsheimer T, et al. Standardization of terminology, definitions and outcome criteria in immune thrombocytopenic purpura of adults Bart Van Rehabilitation & Author Acquisition and and children: report from an international working group. Blood 2009;113: Wijmeersch, MS-Centre interpretation of 2386–2393. MD, PhD Overpelt, BIOMED, the data and critical 24. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and Hasselt University, revision of the powerful approach to multiple testing. J R Stat Soc Ser B Stat Methodol 1995;57: Hasselt, Belgium manuscript 289–300. 25. Frohman EM, Eagar T, Monson N, Stuve O, Karandikar N. Immunologic mechanisms of multiple sclerosis. Neuroimaging Clin N Am 2008;18:577–588. 26. Hartung HP, Arnold DL, Cohen JA, et al. Lymphocyte subset dynamics following References alemtuzumab treatment in the CARE-MS I study. Presented at the 28th Congress of 1. Freedman MS, Kaplan JM, Markovic-Plese S. Insights into the mechanisms of the the European Committee for Treatment and Research in Multiple Sclerosis; October therapeutic efficacy of alemtuzumab in multiple sclerosis. J Clin Cell Immunol 2013;4: 10–13, 2012; Lyon, France. 1000152. 27. Kasper LH, Arnold DL, Coles AJ, et al. Lymphocyte subset dynamics following 2. Hu Y, Turner MJ, Shields J, et al. Investigation of the mechanism of action of alemtu- alemtuzumab treatment in the CARE-MS II study. Presented at the 29th Congress of zumab in a human CD52 transgenic mouse model. Immunology 2009;128:260–270. the European Committee for Treatment and Research in Multiple Sclerosis; October 3. Cox AL, Thompson SA, Jones JL, et al. Lymphocyte homeostasis following 2–5, 2013; Copenhagen, Denmark. therapeutic lymphocyte depletion in multiple sclerosis. Eur J Immunol 2005;35: 28. Blauth K, Owens GP, Bennett JL. The ins and outs of B cells in multiple sclerosis. 3332–3342. Front Immunol 2015;6:565. 4. Havari E, Turner MJ, Campos-Rivera J, et al. Impact of alemtuzumab treatment on the 29. von Büdingen HC, Palanichamy A, Lehmann-Horn K, Michel BA, Zamvil SS. Update survival and function of human regulatory T cells in vitro. Immunology 2014;141:123–131. on the autoimmune pathology of multiple sclerosis: B-cells as disease-drivers and 5. Wiendl H, Kieseier B. Multiple sclerosis: reprogramming the immune repertoire with therapeutic targets. Eur Neurol 2015;73:238–246. alemtuzumab in MS. Nat Rev Neurol 2013;9:125–126. 30. De Mercanti S, Rolla S, Cucci A, et al. Alemtuzumab long-term immunologic effect: 6. Cohen JA, Coles AJ, Arnold DL, et al. Alemtuzumab versus interferon beta 1a as first- Treg suppressor function increases up to 24 months. Neurol Neuroimmunol Neu- line treatment for patients with relapsing-remitting multiple sclerosis: a randomised roinflamm 2016;3:e194. doi:10.1212/NXI.0000000000000194. controlled phase 3 trial. Lancet 2012;380:1819–1828. 31. Durelli L, De Mercanti S, Rolla S, et al. Alemtuzumab long term immunological study: 7. Coles AJ, Twyman CL, Arnold DL, et al. Alemtuzumab for patients with relapsing the immunosuppressive effect does not last more than 48 months. Neurology 2016; multiple sclerosis after disease-modifying therapy: a randomised controlled phase 3 86(suppl 16):S2.008. trial. Lancet 2012;380:1829–1839. 32. Jones JL, Thompson SA, Loh P, et al. Human autoimmunity after lymphocyte de- 8. Arnold DL, Barnett M, Comi G, et al. Durable reduction in MRI disease activity and pletion is caused by homeostatic T-cell proliferation. Proc Natl Acad Sci U S A 2013; slowing of brain volume loss with alemtuzumab in patients with active RRMS: 7-year 110:20200–20205. follow-up of CARE-MS I patients (TOPAZ study). Mult Scler J 2017;23(suppl 3):P1189. 33. Jones JL, Phuah CL, Cox AL, et al. IL-21 drives secondary autoimmunity in patients 9. Coles AJ, Boyko AN, De Seze J, et al. Alemtuzumab durably improves clinical out- with multiple sclerosis, following therapeutic lymphocyte depletion with alemtuzu- comes in patients with active RRMS in the absence of continuous treatment: 7-year mab (Campath-1H). J Clin Invest 2009;119:2052–2061. follow-up of CARE-MS I patients (TOPAZ study). Mult Scler J 2017;23(suppl 3): 34. Coles AJ, Palmer J, Margolin DH. Lymphocyte counts do not predict risk of sub- P1188. sequent relapse or disability accumulation in alemtuzumab-treated relapsing- 10. Coles AJ, Cohen JA, Fox EJ, et al. Alemtuzumab CARE-MS II 5-year follow-up: remitting multiple sclerosis patients: an analysis of the CARE-MS studies. Neurol- efficacy and safety findings. Neurology 2017;89:1117–1126. ogy 2014;82(suppl 10):P3.181. 11. Havrdova E, Arnold DL, Cohen JA, et al. Alemtuzumab CARE-MS I 5-year follow-up: 35. Jelcic I, Al Nimer F, Wang J, et al. Memory B cells activate brain-homing, autoreactive durable efficacy in the absence of continuous MS therapy. Neurology 2017;89: CD4+ T cells in multiple sclerosis. Cell 2018;175:85–100. 1107–1116. 36. Gilmore W, Lund BT, Traboulsee A, et al. Leukocyte repopulation following alem- 12. Pelletier D, Traboulsee A, Barnett M, et al. Patients with active RRMS experience tuzumab treatment in relapsing-remitting MS contains multiple regulatory immune durable reductions in MRI disease activity and slowing of brain volume loss with cell types. Mult Scler 2017;23(suppl 3):P979. alemtuzumab: 7-year follow-up of CARE-MS II patients (TOPAZ study). Mult Scler J 37. Gross CC, Ahmetspahic D, Ruck T, et al. Alemtuzumab treatment alters circulating 2017;23(suppl 3):P741. innate immune cells in multiple sclerosis. Neurol Neuroimmunol Neuroinflamm 13. Singer BA, Alroughani R, Brassat D, et al. Durable improvements in clinical outcomes 2016;3:e289. doi:10.1212/NXI.0000000000000289. with alemtuzumab in patients with active RRMS in the absence of continuous 38. Kim Y, Kim G, Shin HJ, et al. Restoration of regulatory B cell deficiency following treatment: 7-year follow-up of CARE-MS II patients (TOPAZ study). Mult Scler J alemtuzumab therapy in patients with relapsing multiple sclerosis. J Neuroinflammation 2017;23(suppl 3):P736. 2018;15:300.

10 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN ARTICLE OPEN ACCESS Minimal breast milk transfer of rituximab, a monoclonal antibody used in neurological conditions

Kristen M. Krysko, MD, MAS,* Sara C. LaHue, MD,* Annika Anderson, BA, Alice Rutatangwa, DO, MSc, Correspondence William Rowles, BA, Ryan D. Schubert, MD, Jacqueline Marcus, MD, Claire S. Riley, MD, Carolyn Bevan, MD, MS, Dr. Bove [email protected] Thomas W. Hale, RPh, PhD, and Riley Bove, MD, MSc

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e637. doi:10.1212/NXI.0000000000000637 Abstract Objective To determine the transfer of rituximab, an anti-CD20 monoclonal antibody widely used for neurologic conditions, into mature breast milk.

Methods Breast milk samples were collected from 9 women with MS who received rituximab 500 or 1,000 mg intravenous once or twice while breastfeeding from November 2017 to April 2019. Serial breast milk samples were collected before infusion and at 8 hours, 24 hours, 7 days, and 18–21 days after rituximab infusion in 4 patients. Five additional patients provided 1–2 samples at various times after rituximab infusion.

Results The median average rituximab concentration in mature breast milk was low at 0.063 μg/mL (range 0.046–0.097) in the 4 patients with serial breast milk collection, with an estimated median absolute infant dose of 0.0094 mg/kg/d and a relative infant dose (RID) of 0.08% (range 0.06%–0.10%). Most patients had a maximum concentration at 1–7 days after infusion. The maximum concentration occurred in a woman with a single breast milk sample and was 0.29 μg/mL at 11 days postinfusion, which corresponds with an estimated RID of 0.33%. Rituximab concentration in milk was virtually undetectable by 90 days postinfusion.

Conclusions We determined minimal transfer of rituximab into mature breast milk. The RID for rituximab was less than 0.4% and well below theoretically acceptable levels of less than 10%. Low oral bioavailability would probably also limit the absorption of rituximab by the newborn. In women with serious autoimmune neurologic conditions, monoclonal antibody therapy may aﬀord an acceptable beneﬁt to risk ratio, supporting both maternal treatment and breastfeeding.

*These authors contributed equally to the manuscript.

From the Department of Neurology (K.M.K., S.C.L., A.A., A.R., W.R., R.D.S., R.B.), University of California, San Francisco; Weill Institute for Neurosciences (K.M.K., S.C.L., A.A., A.R., W.R., R.D.S., R.B.), Department of Neurology, University of California, San Francisco; Department of Neurology (J.M.), Kaiser Permanente, San Francisco; Department of Neurology (C.S.R.), Columbia University, New York, New York; Department of Neurology (C.B.), Northwestern University, Chicago, Illinois; and Department of Pediatrics (T.W.H.), Texas Tech University School of Medicine, Amarillo.

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

Copyright © 2019 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary ASQ3 = Ages and Stages Questionnaire; AUC = area under the curve; CDC = Centers for Disease Control and Prevention; EDSS = Expanded Disability Status Scale; GI = gastrointestinal; ICC = intraclass correlation coeﬃcient; IV = intravenous; NMOSD = neuromyelitis optica spectrum disorder; RID = relative infant dose; UCSF = University of California, San Francisco.

Neurologic diseases including multiple sclerosis (MS), neuro- Patient demographics, medical history, and peripartum details myelitis optica spectrum disorder (NMOSD), autoimmune were obtained through questionnaires, interview, and review of encephalitis, myasthenia gravis, and migraine commonly medical records. Medical records were reviewed to collect affect women of childbearing age, and monoclonal antibody demographic data, treatment history, rituximab infusion therapies are becoming increasingly available for these date(s)/time and dose(s), and Expanded Disability Status – conditions.1 5 However, limited data on the safety of medi- Scale (EDSS) score before pregnancy. Prospective ques- cations during breastfeeding force many women to choose tionnaires collected pregnancy information including treat- between treating their neurologic disease or breastfeeding ment before and during pregnancy, pregnancy complications, their infant, despite the many benefits of breastfeeding.6 delivery information, and breastfeeding status. Interview was used to supplement missing information including breast- Monoclonal antibodies are appealing options during breastfeeding information and infusion dates as needed. Data from feeding because immunoglobulin Gs (IgGs) are large mole- all sites were collected in the same manner by the UCSF cules with low expected breast milk transfer after the study team after referral to the pregnancy registry by their colostrum phase, with IgA the main immunoglobulin in hu- neurologist. man breast milk.7,8 There have been recent efforts to bridge the gap in safety data for monoclonal antibodies during Standard protocol approvals, registrations, breastfeeding,9,10 given their widespread use in neurologic, and patient consents rheumatologic, gastrointestinal (GI), and oncologic diseases. The UCSF Institutional Review Board approved this study – In fact, the American College of Rheumatology presented (17 22422). Patients provided informed consent to enter the guidelines in 2018 discussing use of biologics while breast- study and provide breast milk samples. feeding,11 and in 2019, the American Gastrointestinal Asso- Breast milk collection ciation recommended biologics can be continued while 12 breastfeeding. In neurology, there are limited data and no Feasibility samples guidelines for monoclonal antibody use during breastfeeding. To evaluate feasibility and validate rituximab concentration assay techniques, we initially collected 1–2 breast milk sam- Rituximab is an anti-CD20 monoclonal IgG1 antibody used ples from 4 mothers (5 samples total) within 2 weeks after 1 widely for neurologic autoimmune conditions and is effective or 2 rituximab infusions of 500 or 1,000 mg intravenous (IV) – and used off-label for MS and myasthenia gravis.13 15 Only (feasibility samples). a single case report measured rituximab concentration in breast milk, for a woman with vasculitis, and levels were re- Serial samples assuringly low (≤0.6 μg/mL). However, sample collection We then systematically collected serial breast milk samples occurred on days 7–10 after infusion, and thus, peak con- preinfusion (drug-naive) and at 8, 24, and 48 hours as well as centration and decline in concentration were not captured.16 at 7, 18–21, and 30 days after rituximab 500 or 1,000 mg Herein, we aimed to evaluate the transfer of rituximab into infusion once or twice when possible in 4 mothers (1 of whom mature breast milk in postpartum patients with MS. had 2 samples as part of feasibility samples).

Delayed samples Methods In 2 cases, eligible patients were identiﬁed at a timeframe more remote from their rituximab infusion, and to maximize Study design and population available data, delayed samples from these patients were col- In this prospective cohort, we enrolled women with MS re- lected up to 90 days from infusion. ceiving rituximab clinically while breastfeeding or shortly after weaning and who were willing to provide breast milk samples. The feasibility, serial, and delayed samples were all assayed in All participants were enrolled in the University of California, the same batch and are all reported to maximize data. San Francisco (UCSF) pregnancy registry for women with demyelinating diseases, which includes patients followed up at All samples included mature breast milk, which occurs be- the UCSF MS Center and women followed up at other ginning 14 days postpartum, after the colostrum and transi- centers if they seek enrollment in the registry, often facilitated tional phases. Mothers collected breast milk samples (2–5 cc) by their neurologist. The samples included in this analysis by pumping milk from the breast into sterile freezer safe breast were collected from November 2017 to April 2019. milk bags (NUK brand). The samples were stored by mothers

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN at 4°C immediately after collection and then transported lo- the time of the last breast milk sample for each woman. cally on dry ice (<1 hour) or shipped overnight in insulated We also determined the maximum concentration (Cmax) PolarTech shipping containers lined with ThermoSafe and the time of the maximum concentration for each PolarPacks to the UCSF, where they were stored at −80°C. woman. Samples were transported on dry ice (<1 hour) to Marin Biologic Laboratories, Inc., for analysis. The absolute average and maximum rituximab dose to the infant in a 24-hour period and the relative infant dose (RID) Quantification of rituximab in breast milk were calculated using methods described by Bennett.17 The To determine the breast milk concentration of rituximab, sam- actual maternal weight at the time or closely after the in- ples were analyzed by Marin Biologic Laboratories (Novato, CA, fusion was used for each woman. Calculation of the absolute US) using the Eagle Biosciences rituximab ELISA kit. This average and maximum dose and the RID assumes the infant ELISA is based on a rituximab-specific mouse monoclonal an- will ingest approximately 150 cc/kg/d of breast milk over tibody. Standards and samples were incubated in the microtiter a typical 24-hour period. The absolute average infant dose plate coated with a capture antibody. After incubation, the wells over 24 hours is calculated by multiplying the Cavg by 150 were washed, and a horseradish peroxidase-conjugated antihu- cc/kg/d, whereas the maximum absolute infant dose is man IgG monoclonal antibody was added that bound to ritux- calculated by multiplying Cmax by 150 cc/kg/d. The RID imab. After incubation, the wells were washed, and the bound estimates the infant’s exposure to rituximab as a percentage enzymatic activity was detected by addition of chromogen- of the maternal dose over a 24-hour period and is calculated substrate. The color was proportional to the amount of ritux- by taking the Cavg in mg/L, multiplying by 0.15 L/kg/d of imab in the sample or standard; the optical density was breast milk and by the maternal weight, and then dividing by measured using a spectrophotometer at 450 nm. Rituximab the maternal dose. The RID was also calculated using Cmax concentrations of samples were determined by using the to determine the maximum RID based on data collected. It standard curve and measured in ng/mL and were then con- has previously been reported by Bennett17 that a RID less verted to μg/mL for analyses and reporting for ease of than 10% is generally safe for an infant to breastfeed, interpretation. although toxicity of the specificdrugshouldalsobe considered. In an initial qualification experiment to test the ELISA kit, 3 standard curves were prepared with drug-naive milk Median and range were then calculated for these measures specimens using 3 different conditions (kit standard pre- across the women who provided serial samples. Although the pared in dilution buffer, in neat naive milk, and in naive milk rituximab concentrations in the women who provided 1–2 diluted 1:10 in a dilution buffer). Recovery of the rituximab samples (feasibility and delayed samples) were not included standard was similar in all conditions. The method of in these calculations, they also contributed to our overall testing clinical samples neat or at various dilutions and understanding of rituximab transfer into breast milk and were determining ng/mL values from the standard curve in kit qualitatively compared with those of the women who pro- buffer was also qualified by testing 5 feasibility clinical vided serial samples. samples neat and at 1:1, 1:10, and 1:30 dilution in kit buffer, with analysis against standard curves prepared in buffer in Infant outcomes neat milk or in 1:10 milk. Because there was no statistical Medical records were reviewed to identify birth outcomes, difference between the methods of analysis, the final ex- infection, vaccination, growth, and developmental concerns in periment testing concentrations reported in this study the infants who breastfed while their mothers were receiving tested the samples neat. rituximab from the time of birth to last well-child follow-up available. Weight, length, and head circumference recorded in Two breast milk concentrations were measured on each medical records were compared with Centers for Disease sample in duplicate, and the intraclass correlation coefficient Control and Prevention (CDC) clinical growth charts and (ICC) was calculated to evaluate reproducibility. The 2 rep- reported as percentiles for age.18 The Ages and Stages licates were averaged for the final breast milk concentration Questionnaire (ASQ3)19 was undertaken prospectively by for each sample. parents at 4, 8, and 12 months of age. The ASQ3 is a questionnaire that evaluates 5 domains of child development, Statistical analysis including gross motor, fine motor, communication, problem- For the 4 women who provided serial (at least 5) samples, solving, and personal-social development. Each domain is the breast milk concentrations of rituximab were analyzed scored and categorized as “below cutoff,”“monitoring zone,” with pharmacokinetic methods using STATA 15 (College or “above cutoff.”“Above cutoff” was considered normal be- Station, TX). The trapezoidal rule was used to calculate the cause this indicates a child’s development is on schedule. area under the breast milk concentration-time curves “Monitoring zone” means the child is close to the cutoff, and (AUC) for each of these women. The average concentra- providing learning materials and monitoring is suggested. “ ff” tion of rituximab in breast milk (Cavg) was calculated by Below cuto suggests professional assessment should be dividing the AUC by the number of days from infusion at completed.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 3 Data availability median time between delivery and rituximab treatment in Anonymized data are available to qualified investigators on the samples provided was 6 months, with a range from 1.5 request for the purposes of replicating procedures or results to 11 months. Most patients received a single rituximab by contacting the corresponding author. infusion of either 500 or 1,000 mg IV. One individual received 500 mg IV twice (19 days apart) with samples provided up to 19 days from the first infusion. Another had Results previously received rituximab 1,000 mg twice 0.5 months postpartum and provided samples after the second dose Demographics and clinical characteristics of her second cycle of 1,000 mg twice at 7 months Nine women with MS who entered the UCSF pregnancy postpartum. registry for women with demyelinating diseases were treated clinically with rituximab while breastfeeding or Rituximab concentration in breast milk shortly after weaning and agreed to provide breast milk The 9 women provided a total of 30 breast milk samples, with samples. All had relapsing remitting MS with relatively 4 patients providing at least 5 serial breast milk samples. The mild disability (maximum EDSS of 2.0), with a mean age of ICC calculated with 2 replicates was excellent (0.997). Con- 34 years and median disease duration of 4 years (table 1). centrations of rituximab in breast milk measured in all samples Four women had also received rituximab in the 6 months are displayed in figure 1, and pharmacokinetic curves for the 4 before pregnancy, 1 received dimethyl fumarate before patients with serial samples are shown in figure 2. The median fi pregnancy, and 1 received ngolimod before pregnancy, average rituximab concentration was low at 0.063 μg/mL followed by glatiramer acetate which was continued during (range 0.046–0.097) in the 4 patients who provided serial pregnancy. The other 3 women had not received disease- samples. Within these patients, the maximum rituximab modifying therapy in the 6 months before pregnancy. The concentration occurred between 1 and 7 days from the infusion, with a median maximum concentration of 0.074 μg/ mL (range 0.061–0.12). Based on the average concentration, Table 1 Patient demographics and clinical characteristics the median absolute 24-hour daily dose of rituximab was (n = 9) 0.0094 mg/kg/d, and the median RID was 0.08% (range – Postpartum women 0.06% 0.10%) (table 2). Characteristic (n = 9) – Age, mean y (SD) 33.7 (3.2) Among the patients who provided 1 2 breast milk samples (feasibility and delayed samples), measured rituximab Weight, mean kg (SD) 73.9 (15.8) concentrations were generally similar to those of patients Disease duration, median y (range) 4(0–11) with serial samples apart from a single outlier with a rit- μ MS subtype RRMS, n 9 uximab concentration of 0.29 g/mL at 11 days postinfusion (figure 1). This single outlier was a single sample EDSS, median (range) 0.0 (0.0–2.0) provided by a mother who received rituximab 1,000 mg at Rituximab in 6 mo before pregnancy, n 4 9 months postpartum. To ensure that the RID was not underestimated, this higher value was also used to calculate Rituximab infusion to conception, median 2.5 (1.6–3.7) mo (range) the RID, and it remained low at 0.33%. One patient provided samples more remote from the rituximab infusion Postpartum rituximab received, median mo 6(1.5–11) (range) at 60 and 90 days postinfusion. These values were reassuringly very low with a nearly undetectable level by 90 Rituximab dose, n days. Rituximab levels in breast milk were not higher in 500 mg once 2 mothers who also received rituximab before pregnancy or

500 mg twice (2.7 wk apart) 1 in the mother who received 1,000 mg twice instead of a single infusion. Patient 3 (table 2 and ﬁgure 2C) had 1,000 mg once 5 increasing rituximab levels at the last sample on day 19 1,000 mg twice (2 wk apart) 1 with her second dose of rituximab on that same day, without capture of peak and decline after the second No. of breast milk samples provided, n infusion. 1 3

2 2 Infant outcomes Of the 9 women, 5 continued to breastfeed after receiving 5 1 rituximab, whereas 4 discontinued breastfeeding before rit- 6 3 uximab infusion. Of the 5 mothers who breastfed after receiving rituximab postpartum, 3 also received rituximab Abbreviations: EDSS = Expanded Disability Status Scale; RRMS = relapsing remitting MS. within the 6 months before pregnancy. Detailed information is available for 4 infants who breastfed after their mothers

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN Figure 1 Rituximab concentrations in breast milk (n = 9)

Rituximab concentrations (μg/mL) in 9 patients up to 90 days from infusion of rituximab 1,000 or 500 mg intravenous once or twice. Lines connect samples provided from a given individual. There was 1 outlier with a concentration of 0.29 μg/mL at 11 days postinfusion, although this level is still low. received rituximab, whereas data were not available for the 1 none were serious and all were common infections seen in other exposed infant (table 3). There were a few minor but infancy. There were no preliminary concerns identiﬁed with no serious infections. The infant exposed to rituximab growth or development up to 8–12 months postpartum in through breast milk earliest and with a higher total dose to these infants, and routine vaccinations per CDC guidelines20 the mother had more infections (table 3, infant D), although were given.

Figure 2 Pharmacokinetic curves for rituximab concentration in breast milk for patients who provided serial samples (n = 4)

Rituximab concentrations (μg/mL) in 4 patients up to 30 days after infusion of rituximab 1,000 mg IV once (A: patient 1; B: patient 2), 500 mg IV twice day 0 and 19 (C: patient 3), or 500 mg IV once (D: patient 4). IV = intravenous.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 5 Table 2 AUC, average concentration, maximum concentration, and RID of rituximab in breast milk among 4 patients with serial samples

Parameters Patient 1 Patient 2 Patient 3 Patient 4 Median (range)

AUC0-Tmax, μg.d/mL 2.23 1.08 1.84 1.37 1.61 (1.08–2.23)

Tmax, d 30 21 19 30 25.5 (19–30)

Cavg, μg/mL 0.074 0.052 0.097 0.046 0.063 (0.046–0.097)

Cmax, μg/mL 0.083 0.065 0.12 0.061 0.074 (0.061–0.12)

Time of Cmax, d 72174.5(1–7)

Absolute infant dose by Cavg, mg/kg/d 0.011 0.0077 0.015 0.0068 0.0094 (0.0068–0.015)

Absolute infant dose by Cmax, mg/kg/d 0.012 0.0098 0.018 0.0092 0.011 (0.0092–0.018)

RID from Cavg, % 0.065 0.057 0.092 0.099 0.079 (0.057–0.099)

RID from Cmax, % 0.072 0.072 0.11 0.13 0.094 (0.072–0.13)

Rituximab dose, mg 1,000 × 1 1,000 × 1 500 × 2 500 × 1 1,000 (500–1,000)

Time of rituximab postpartum, mo 646116(4–11)

Maternal weight, kg 58.4 73.5 63.5 72.6 68.1 (58.4–73.5)

Breast milk samples, n 66566(5–6)

Timing of breast milk samples after 0,8h,24h,7d, 0,8h,24h,48h, 0,8h,24h,7d, 0,8h,24h,7d, rituximab infusion 18 d, 30 d 7d,18d 19 d 18 d, 30 d

Abbreviations: AUC = area under the drug-concentration-time curve in milk; Cavg = average drug concentration over the interval; Cmax = maximum drug concentration measured; RID = relative infant dose; Tmax = maximum time of milk collection. RID values are in bold because these are values of greatest importance.

Discussion 3) compared with others who received 1,000 mg rituximab, the concentrations were overall quite similar and reassuringly In this prospective cohort study of postpartum women with low. It is possible there may be minor interindividual vari- MS receiving rituximab, we observed very low transfer ability in the pharmacokinetics and breast milk transfer of of rituximab into mature breast milk. The average and rituximab. maximum concentrations of rituximab in breast milk were reassuringly low and reached nearly undetectable levels by Recent studies in inflammatory bowel disease evaluated 60–90 days from the infusion. The median RID was 0.08% concentrations of other monoclonal antibodies in breast and maximum RID was 0.10%, both less than the acceptable milk, including tumor necrosis factor-alpha inhibitors, and 17 value of less than 10%. Furthermore, the small amount of found reassuringly low peak concentrations.23,24 Natalizu- rituximab reaching the breast milk is likely degraded in the mab, an IgG4 monoclonal antibody therapy for MS and infant’sGItract,8 with less than 25% of ingested IgG inflammatory bowel disease, represents 1 notable exception reaching the stool and the remainder presumably diges- because of more repeated dosing and the higher relative 21 ted. The neonatal Fc receptor may allow passage of these proportion of IgG4 in mature breast milk than in undigested IgG molecules from the GI tract to circulation,22 serum.25,26 Although low levels of natalizumab were andsoasmallamountofinfant exposure cannot be reported in breast milk in 2 patients,23 serial collection of excluded. breast milk over 50 days from infusion in another single patient suggested accumulation of natalizumab in breast The low concentrations of rituximab measured in mature milk.27 By contrast, rituximab can be dosed once every – breast milk in our study were consistent with the low 6 12 months and is an IgG1 subclass with a large molecular 28 concentration reported in the single case of vasculitis weight (145 kD). IgG1 transfers in low levels into ma- treated with rituximab during breastfeeding.16 With serial ture milk,26 which is consistent with our finding of lower sampling of breast milk from 4 mothers, we were able to breast milk transfer of rituximab than that reported with calculate area under the breast milk concentration-time natalizumab. curve, allowing calculation of the average concentration and RID which were reassuring. Although the levels varied be- Regarding pregnancy and neonatal outcomes with maternal tween patients, with a slightly higher maximum concen- treatment with rituximab, a case series of 11 pregnancies in trationin1womantreatedwith500mg(figure 2C, patient women who received rituximab within 6 months of

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN Table 3 Outcomes for infants born to mothers treated with rituximab during breastfeeding

Characteristics Infant A Infant B Infant C Infant D

Time from last rituximab to 1.6 2.3 3.7 None conception, mo

Time of postpartum 2.6 4 6 0.5 and 7 rituximab infusion, moa

DMT in pregnancy None None None Glatiramer acetate

Sex FFMF

Gestational age 38 wk 6 d 39 wk 4 d 41 wk 37 wk 3 d

Delivery method SVD SVD SVD assisted SVD assisted

Perinatal complications None None 2 BRUEb None

Breastfeeding mo after Discarded 2 mo then BF 5 mo 2 mo 11.5 mo postpartum rituximab 4mo

Growth, timing 9moc 9moc 6mo1wkc 5moc

Weight, CDC percentile 50–75th 50–75th 75–90th 50th

Height, CDC percentile 50th 75–90th 75–90th 75–90th

Head circumference, CDC 75–90th 25–50th 90–95th 25–50th percentile

Infections 2wkd: Conjunctivitis 7moc: 5mod: URTI and 3wkc: GERD and esophagitis; 2 moc: URTI; 3 moc: URTI and and dacryostenosis Fever otitis media diarrhea; 11 moc: Fever; 12 moc: URTI

Routine vaccinationse Complete Complete Complete Complete except RV

Development at well-child Normal Normal Normal Normal visits

ASQ3f 4mod/8 moc 4moc 4mod/8moc 8moc/12 moc

Gross motor Above/above Above Above/above Above/above

Fine motor Above/above Above Above/above Above/above

Communication Monitor/above Above Above/above Above/above

Problem-solving Above/above Above Above/above Above/above

Personal-social Above/above Above Above/above Above/above

Follow-up, mo 99812

Abbreviations: ASQ3 = Ages and Stages Questionnaire; BF = breastfed; BRUE = brief resolved unexplained events; CDC = Centers for Disease Control and Prevention; DMT = disease-modifying therapy; GERD = gastroesophageal reflux disease; RV = rotavirus; SVD = spontaneous vaginal delivery; URTI = upper respiratory tract infection. a 1,000 mg × 1 in infants A, B, C; 1,000 mg × 2 at 0.5 mo and 7 mo in infant D. b 2 brief resolved unexplained events of color changes while breastfeeding during the first d of life, but was monitored with no additional complications. c Time point after postpartum rituximab received by the mother and breastfeeding. d Time point before postpartum rituximab received by the mother and breastfeeding. e Vaccinations as per CDC guidelines20 up to 9 mo of age include diphtheria, tetanus, and acellular pertussis (DTaP), Haemophilus influenzae type b (Hib), pneumococcal conjugate (PCV13), inactivated poliovirus (IPV), hepatitis B (HepB), Rotavirus (RV1 or RV5), and influenza (IIV). Only infant D had reached the age at which CDC recommends live vaccination against measles, mumps, and rubella (MMR), and this was given routinely. None had yet received varicella vaccination because of their age. f The ASQ3 evaluates 5 domains of child development: gross motor, fine motor, communication, problem-solving, and personal-social development. Each domain is scored and categorized as “below cutoff,”“monitoring zone” (Monitor), or “above cutoff” (Above). “Above cutoff” means child development is on schedule in the given domain. Eight-month ASQ3 is missing for infant B and 4-month ASQ3 is missing for infant D. conception for MS and NMOSD, as well as a systematic report of a low breast milk rituximab concentration in review of women treated for other conditions, did not a patient treated for vasculitis.16 Because patients were identify any major safety concerns.29 The patients included being treated clinically, this provided the opportunity to in our study were clinically treated with rituximab while measure the breast milk concentration of rituximab to help breastfeeding, despite its oﬀ-label use in MS, likely based on guide future treatment decision-making in the postpartum pregnancy safety data29 and the reassuring single case period.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 7 Four mothers in our cohort breastfed after receiving rituximab molecule of the same molecular size as rituximab (145 and provided growth, development, and medical records for kDa),32 and therefore, we would expect similarly low ocreli- their infants, which demonstrated no serious infections and zumab transfer into breast milk. To investigate this, our on- normal growth and development in these infants followed up going registry efforts will enable direct measurement of to 8–12 months of age. Even the infant whose mother re- ocrelizumab concentration in breast milk. ceived the highest total dose of rituximab early in the postpartum period and who experienced the most infections was Altogether, our findings suggest that at least in the mature well within the normal range of expected infections.30 Un- milk phase in term infants, a single rituximab infusion results fortunately, the small sample size and lack of a comparison in a RID in breast milk of less than 0.4% in all cases, which is group precluded rigorous determination as to whether the far less than the commonly accepted threshold of 10%. These number of infections differed from expected. The effect of findings are appealing for women with a higher risk of disease postpartum rituximab timing and dose on infant outcomes activity in the postpartum state (such as MS or NMOSD requires further study. relapse, which may be associated with disability progression,33 and refractory autoimmune myasthenia gravis or autoimmune Limitations of this study include the relatively small number encephalitis). In individual cases, the significant benefitof of breast milk samples, although collection of serial milk maintaining neurologic stability in the mother with the help of samples in 4 women did allow calculation of fairly consistent monoclonal antibodies could offset theoretical risks to the AUC and RID. Unfortunately, serial breast milk collection at infant, allowing both mother and infant to benefit from the preferred time points was not available in all treated breastfeeding.6 Areas requiring further study include evalu- women because data were collected during real-world use of ating drug concentrations and lymphocyte counts in infants rituximab and not all participants were identified at the time of exposed to drug-containing breast milk and studying longer- rituximab treatment. In addition, we were unable to capture term developmental data in these infants. Larger prospective the declining concentration of rituximab over time after a cy- registry studies are ongoing. cle containing 2 infusions (patient 3) because of a lack of sample collection after the second infusion. There were also Acknowledgment limited safety data in infants who continued to breastfeed after The authors would like to thank the women who donated rituximab treatment, with a relatively short follow-up in 4 their breast milk and participated in this study. infants. Unfortunately, blood samples from infants who breastfed after their mothers received rituximab were not Study funding available, so we could not assess rituximab concentration or National Multiple Sclerosis Society Career Transition Award. the biological effect of any possible rituximab exposure in the K.M. Krysko is funded by a Sylvia Lawry Physician Fellowship infants, such as blood counts, CD19+ B cell counts, or re- from the National Multiple Sclerosis Society (FP-1605-08753 sponse to vaccination. (Krysko)).

We expect that our findings would generalize to the use of Disclosures rituximab more broadly during breastfeeding; however, K.M. Krysko receives fellowship support from the National some special cases must be considered. Maternal complica- MS Society and from Biogen. S.C. LaHue reports no dis- tions such as mastitis might allow greater transfer of large closures. A. Anderson reports no disclosures. A. Rutatangwa molecules such as rituximab into breast milk.7 Our findings receives fellowship support from Biogen. W. Rowles reports may also not apply to breast milk in the first 14 days of life, no disclosures. R.D. Schubert receives fellowship support given the differences in the contents of colostrum and from the National Multiple Sclerosis Society and American transitional milk compared with mature breast milk. Finally, Brain Foundation Clinician Scientist Development Award in preterm infants, gut immaturity could affect absorption of (FAN-1608-25607) and has received advisory board fees rituximab. from Genzyme-Sanofi. J. Marcus reports no disclosures. C.S. Riley has received consulting and advisory board fees from Further studies are required to evaluate the long-term infant Biogen, Roche-Genentech, TG Therapeutics, and EMD safety of rituximab treatment during breastfeeding, including Serono. C. Bevan reports no disclosures. T.W. Hale is a con- evaluation of the biological effect on lymphocyte subsets and sultant for Biohaven Pharmaceuticals. R. Bove has received response to vaccination. Further study is also required to consulting and advisory board fees from Alexion, Biogen, understand whether maternal treatment with rituximab alters EMD Serono, Genzyme-Sanofi, Novartis, and Roche- the composition of breast milk because IgA is a major im- Genentech. The authors have no conflicts of interest rele- munologic component of breast milk,8 and immunoglobulins vant to this article to disclose. Go to Neurology.org/NN for may be depleted with long-term rituximab treatment.31 It is full disclosures. also not known whether the neonatal gut microbiome could be affected by rituximab exposure in breast milk. Ocrelizumab, Publication history another anti-CD20 monoclonal antibody that is Food and Received by Neurology: Neuroimmunology & Neuroinflammation August fi Drug Administration-approved for MS, is also an IgG1 26, 2019. Accepted in nal form October 1, 2019.

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN 6. Breastfeeding SO. Breastfeeding and the use of human milk. Pediatrics 2012;129: – Appendix Authors e827 e841. 7. Anderson PO, Sauberan JB. Modeling drug passage into human milk. Clin Pharmacol Ther 2016;100:42–52. Name Location Role Contribution 8. Hurley WL, Theil PK. Perspectives on immunoglobulins in colostrum and milk. Nutrients 2011;3:442–474. Kristen M. University of Author Designed and 9. Wang J, Johnson T, Sahin L, et al. Evaluation of the safety of drugs and biological Krysko, MD, California, San conceptualized the study, products used during lactation: workshop summary. Clin Pharmacol Ther 2017;101: MAS Francisco analyzed and interpreted 736–744. the data, and drafted the 10. Pistilli B, Bellettini G, Giovannetti E, et al. Chemotherapy, targeted agents, antie- manuscript for intellectual content metics and growth-factors in human milk: how should we counsel cancer patients about breastfeeding? Cancer Treat Rev 2013;39:207–211. Sara C. University of Author Designed and 11. Reproductive Health in Rheumatic Diseases. Available at: rheumatology.org/Practice- LaHue, MD California, San conceptualized the study, Quality/Clinical-Support/Clinical-Practice-Guidelines/Reproductive-Health-in- Francisco interpreted the data, and Rheumatic-Diseases. Accessed May 31, 2019. fl drafted the manuscript for 12. Mahadevan U, Robinson C, Bernasko N, et al. In ammatory bowel disease in preg- intellectual content nancy clinical care pathway: a report from the American gastroenterological association IBD parenthood project working group. Gastroenterology 2019;156: – Annika University of Author Major role in the 1508 1524. ff Anderson, BA California, San acquisition of data and 13. Granqvist M, Boremalm M, Poorghobad A, et al. Comparative e ectiveness of rit- Francisco revised the manuscript for uximab and other initial treatment choices for multiple sclerosis. JAMA Neurol 2018; intellectual content 75:320–327. 14. Hauser SL, Waubant E, Arnold DL, et al. B-cell depletion with rituximab in relapsing- Alice University of Author Designed and remitting multiple sclerosis. N Engl J Med 2008;358:676–688. Rutatangwa, California, San conceptualized the study 15. Iorio R, Damato V, Alboini PE, Evoli A. Efficacy and safety of rituximab for myasthenia DO, MSc Francisco and revised the manuscript gravis: a systematic review and meta-analysis. J Neurol 2015;262:1115–1119. for intellectual content 16. Bragnes Y, Boshuizen R, de Vries A, Lexberg Å, Østensen M. Low level of Rituximab in human breast milk in a patient treated during lactation. Rheumatology (Oxford) William University of Author Major role in the 2017;56:1047–1048. Rowles, BA California, San acquisition of data 17. Bennett P. Drugs and Human Lactation: A Comprehensive Guide to the Content and Francisco Consequences of Drugs, Micronutrients, Radiopharmaceuticals and Environmental and Occupational Chemicals in Human Milk, 2nd ed. Amsterdam, The Netherlands: Ryan University of Author Designed and Elsevier; 1996. Schubert, California, San conceptualized the study 18. Growth Charts—Clinical Growth Charts; 2000. Available at: cdc.gov/growthcharts/ MD Francisco and revised the manuscript clinical_charts.htm. Accessed June 4, 2019. for intellectual content 19. Squires J, Bricker D. Ages & Stages Questionnaires[R], 3rd ed, (ASQ-3[TM]): A Parent-Completed Child-Monitoring System. Baltimore, MD: Brookes Publishing Jacqueline Kaiser Author Major role in the Company; 2009. Marcus, MD Permanente acquisition of data and 20. Centers for Disease Control and Prevention (CDC). Immunization Schedules. revised the manuscript for Centers for Disease Control and Prevention (CDC). Available at: cdc.gov/vaccines/ intellectual content schedules/hcp/imz/child-adolescent.html. Published February 5, 2019. Accessed September 3, 2019. Claire S. Columbia Author Major role in the 21. Jasion VS, Burnett BP. Survival and digestibility of orally-administered immuno- Riley, MD University acquisition of data and globulin preparations containing IgG through the gastrointestinal tract in humans. revised the manuscript for Nutr J 2015;14:22. intellectual content 22. Giragossian C, Clark T, Piché-Nicholas N, Bowman CJ. Neonatal Fc receptor and its role in the absorption, distribution, metabolism and excretion of immunoglobulin Carolyn Northwestern Author Major role in the G-based biotherapeutics. Curr Drug Metab 2013;14:764–790. Bevan, MD, University acquisition of data and 23. Matro R, Martin CF, Wolf D, Shah SA, Mahadevan U. Exposure concentrations of MS revised the manuscript for infants breastfed by women receiving biologic therapies for inflammatory bowel intellectual content diseases and effects of breastfeeding on infections and development. Gastroenterology 2018;155:696–704. Thomas W. Texas Tech Author Reviewed and interpreted 24. Clowse ME, Förger F, Hwang C, et al. Minimal to no transfer of certolizumab pegol Hale, RPh, University the data and revised the into breast milk: results from CRADLE, a prospective, postmarketing, multicentre, PhD School of manuscript for intellectual pharmacokinetic study. Ann Rheum Dis 2017;76:1890–1896. Medicine content 25. Keller MA, Heiner DC, Kidd RM, Myers AS. Local production of IgG4 in human colostrum. J Immunol 1983;130:1654–1657. Riley Bove, University of Author Designed and 26. Rodr´ıguez-Camejo C, Puyol A, Fazio L, et al. Antibody profile of colostrum and the MD, MSc California, San conceptualized the study, effect of processing in human milk banks: implications in immunoregulatory prop- Francisco major role in the erties. J Hum Lact 2018;34:137–147. acquisition of data, 27. Baker TE, Cooper SD, Kessler L, Hale TW. Transfer of natalizumab into breast milk interpreted the data, and in a mother with multiple sclerosis. J Hum Lact 2015;31:233–236. revised the manuscript for 28. Genentech Rituxan (rituximab) [package insert]. U.S. Food and Drug Administra- intellectual content tion. Available at: accessdata.fda.gov/drugsatfda_docs/label/2010/103705s5311lbl. pdf. Published Revised 2010. Accessed May 21, 2019. 29. Das G, Damotte V, Gelfand JM, et al. Rituximab before and during pregnancy: References a systematic review, and a case series in MS and NMOSD. Neurol Neuroimmunol fl 1. Gelfand JM, Cree BAC, Hauser SL. Ocrelizumab and other CD20+ B-cell-depleting Neuroin ammation 2018;5:e453. doi:10.1212/NXI.0000000000000453. therapies in multiple sclerosis. Neurotherapeutics 2017;14:835–841. 30. McInerny TK, Adam HM, Campbell DE, DeWitt TG, Foy JM, Kamat DM, eds. 2. Nguyen A-L, Gresle M, Marshall T, Butzkueven H, Field J. Monoclonal antibodies in American Academy of Pediatrics Textbook of Pediatric Care. 2nd ed. Elk Grove the treatment of multiple sclerosis: emergence of B-cell-targeted therapies. Br J Village, IL: American Academy of Pediatrics; 2017. Pharmacol 2017;174:1895–1907. 31. Kado R, Sanders G, McCune WJ. Suppression of normal immune responses after 3. Shin YW, Lee ST, Park KI, et al. Treatment strategies for autoimmune encephalitis. treatment with rituximab. Curr Opin Rheumatol 2016;28:251–258. Ther Adv Neurol Disord 2017;11:1756285617722347. 32. Genentech. Ocrevus(ocrelizumab) [package insert].U.S. Food and Drug Adminis- 4. Mantegazza R, Antozzi C. When myasthenia gravis is deemed refractory: clinical tration website. Available at: accessdata.fda.gov/drugsatfda_docs/label/2017/ signposts and treatment strategies. Ther Adv Neurol Disord 2018;11: 761053lbl.pdf. Published Revised 2017. Accessed August 30, 2019. 1756285617749134. 33. Portaccio E, Ghezzi A, Hakiki B, et al. Postpartum relapses increase the risk of 5. Loder EW, Robbins MS. Monoclonal antibodies for migraine prevention: progress, disability progression in multiple sclerosis: the role of disease modifying drugs. but not a panacea. JAMA 2018;319:1985–1987. J Neurol Neurosurg Psychiatry 2014;85:845–850.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 9 ARTICLE OPEN ACCESS CLASS OF EVIDENCE Antibodies to neurofascin, contactin-1, and contactin-associated protein 1 in CIDP Clinical relevance of IgG isotype

Andrea Cortese, MD, PhD, Raffaella Lombardi, MA, Chiara Briani, MD, Ilaria Callegari, MD, Correspondence Luana Benedetti, MD, PhD, Fiore Manganelli, MD, Marco Luigetti, MD, PhD, Sergio Ferrari, MD, Dr. Cortese [email protected] Angelo M. Clerici, MD, Girolama Alessandra Marfia, MD, Andrea Rigamonti, MD, Marinella Carpo, MD, PhD, Raffaella Fazio, MD, Massimo Corbo, MD, Anna Mazzeo, MD, Fabio Giannini, MD, Giuseppe Cosentino, MD, PhD, Elisabetta Zardini, MA, Riccardo Curro,` MD, Matteo Gastaldi, MD, PhD, Elisa Vegezzi, MD, Enrico Alfonsi, MD, Angela Berardinelli, MD, Ludivine Kouton, MD, Constance Manso, MSc, Claudia Giannotta, MA, Pietro Doneddu, MD, Patrizia Dacci, MD, PhD, Laura Piccolo, MD, Marta Ruiz, MD, Alessandro Salvalaggio, MD, Chiara De Michelis, MD, Emanuele Spina, MD, Antonietta Topa, MD, Giulia Bisogni, MD, Angela Romano, MD, Sara Mariotto, MD, Giorgia Mataluni, MD, PhD, Federica Cerri, MD, PhD, Claudia Stancanelli, MD, Mario Sabatelli, MD, Angelo Schenone, MD, Enrico Marchioni, MD, Giuseppe Lauria, MD, Eduardo Nobile-Orazio, MD, J´eromeˆ Devaux, PhD,* and Diego Franciotta, MD*

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e639. doi:10.1212/NXI.0000000000000639

Abstract MORE ONLINE

Objective Class of Evidence To assess the prevalence and isotypes of anti-nodal/paranodal antibodies to nodal/para- Criteria for rating nodal proteins in a large chronic inﬂammatory demyelinating polyradiculoneuropathy therapeutic and diagnostic (CIDP) cohort, compare clinical features in seronegative vs seropositive patients, and studies gather evidence of their isotype-speciﬁcpathogenicrole. NPub.org/coe

Methods Antibodies to neurofascin-155 (Nfasc155), neurofascin-140/186 (Nfasc140/186), contactin-1 (CNTN1), and contactin-associated protein 1 (Caspr1) were detected with ELISA and/or cell- based assay. Antibody pathogenicity was tested by immunohistochemistry on skin biopsy, intraneural injection, and cell aggregation assay.

Results Of 342 patients with CIDP, 19 (5.5%) had antibodies against Nfasc155 (n = 9), Nfasc140/ 186 and Nfasc155 (n = 1), CNTN1 (n = 3), and Caspr1 (n = 6). Antibodies were absent from healthy and disease controls, including neuropathies of diﬀerent causes, and were mostly detected in patients with European Federation of Neurological Societies/ Peripheral Nerve Society (EFNS/PNS) deﬁnite CIDP (n = 18). Predominant antibody isotypes were immunoglobulin G (IgG)4 (n = 13), IgG3 (n = 2), IgG1 (n = 2), or

*These authors contributed equally to this work.

From the Department of Brain and Behavioral Sciences (A.C., I.C., G.C., R.C., E.V.), University of Pavia, Pavia, Italy; Department of Neuromuscular Disease (A.C.), UCL Queen Square Institute of Neurology, London, United Kingdom; Neuroalgology Unit (R.L., P.D., L.P., G.L.), IRCCS Fondazione Istituto Neurologico “Carlo Besta,” Milan, Italy; Department of Neurosciences (C.B., M.R., A.S.), University of Padova, Padova, Italy; IRCCS Mondino Foundation (I.C., G.C., E.Z., R.C., M.G., E.V., E.A., A.B., D.F.), Pavia, Italy; Department of Neuroscience (L.B., C.D.M., A.S.), Rehabilitation, Ophthalmology, Genetics, Maternal and Child Health (DiNOGMI), University of Genova, Genova, Italy; IRCCS Ospedale Policlinico San Martino (L.B., C.D.M.,A.S.),Genova,Italy;Departmentof Neurosciences (F.M., E.S., A.T.), Odontostomatological and Reproductive Sciences, University of Naples “Federico II,” Naples, Italy; Fondazione Policlinico Universitario Agostino Gemelli-IRCCS. UOC Neurologia (M.L., A.R., M.S.), Rome, Italy; Universit`a Cattolica del Sacro Cuore (M.L., A.R., M.S.), Rome, Italy; Section of Neurology (S.F., S.M.), Department of Neuroscience, Biomedicine and Movement Sciences, University of Verona, Verona, Italy; Department of Neurology and Stroke Unit (A.M.C.), Ospedale di Circolo/Fondazione Macchi, Varese, Italy; Department of Systems Medicine (G.A.M., G.M.), University of Rome Tor Vergata, Rome, Italy; Neurological Department (A.R.), ASST Lecco; Ospedale Treviglio ASST Bergamo Ovest (M.C.), Italy; Department of Neurology (R.F., F.C.), San Raffaele Scientific Institute, Milan, Italy; Department of Neurorehabilitation Sciences (M.C.), Casa Cura Policlinico (CCP), Milan, Italy; Department of Clinical and Experimental Medicine (A.M.), University of Messina, Messina, Italy; Department of Medicine, Surgery and Neurosciences (F.G.), University of Siena, Italy; Referral Center for Neuromuscular Diseases and ALS (L.K., E.M.), AP-HM, Timone University Hospital, Marseille, France; Universit´e de Bordeaux (C.M.), Interdisciplinary Institute for Neuroscience, Bordeaux, France; Humanitas Clinical and Research Center (C.G., P.D., E.N.-O.), Milan University, Milan, Italy; IRCCS Centro Neurolesi “Bonino Pulejo” (C.S.), Messina, Italy; Department of Biomedical and Clinical Sciences “Luigi Sacco” (G.B., G.L.), University of Milan, Milan, Italy; and Institute for Neurosciences of Montpellier (J.D.), INSERM U1051, Montpellier University, Hopital Saint Eloi, Montpellier, 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 the MRC and 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 © 2019 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary Caspr1 = contactin-associated protein 1; CBA = cell-based assay; CIDP = chronic inflammatory demyelinating polyradiculoneuropathy; CNTN1 = contactin-1; EFNS/PNS = European Federation of Neurological Societies/Peripheral Nerve Society; GBS = Guillain-Barré syndrome; GFP = green fluorescent protein; HC = healthy control; IVIG = intravenous immunoglobulin; MMN = multifocal motor neuropathy; MRC = medical research council; Nfasc155 = neurofascin-155; ONLS = Overall Neuropathy Limitation Scale; PN = peripheral neuropathy.

undetectable (n = 2). IgG4 antibody-associated phenotypes included onset before 30 years, severe neuropathy, subacute onset, tremor, sensory ataxia, and poor response to intravenous immunoglobulin (IVIG). Immunosuppressive treatments, including rituximab, cyclophosphamide, and methotrexate, proved eﬀective if started early in IVIG-resistant IgG4- seropositive cases. Five patients with an IgG1, IgG3, or undetectable isotype showed clinical features indistinguishable from seronegative patients, including good response to IVIG. IgG4 autoantibodies were associated with morphological changes at paranodes in patients’ skin biopsies. We also provided preliminary evidence from a single patient about the pathogenicity of anti-Caspr1 IgG4, showing their ability to penetrate paranodal regions and disrupt the integrity of the Nfasc155/CNTN1/Caspr1 complex.

Conclusions Our findings confirm previous data on the tight clinico-serological correlation between antibodies to nodal/paranodal proteins and CIDP. Despite the low prevalence, testing for their presence and isotype could ultimately be part of the diagnostic workup in suspected inflammatory demyelinating neuropathy to improve diagnostic accuracy and guide treatment.

Classification of evidence This study provides Class III evidence that antibodies to nodal/paranodal proteins identify patients with CIDP (sensitivity 6%, speciﬁcity 100%).

Chronic inflammatory demyelinating polyradiculoneuropathy patients with CIDP,3,4,15,16 and there is little information on the (CIDP), the most commonly acquired inflammatory neuropa- role and associated clinical features of antibodies of the IgG1-3 thy worldwide, is clinically heterogeneous. Proven treatments isotype. Moreover, the prevalence of anti-Caspr1 antibodies is still for CIDP include corticosteroids, plasma exchange, and in- largely unknown, as well as their possible pathogenic role. travenous immunoglobulin (IVIG). The response rates to treatments were reported to be heterogeneous in subgroups of Here, we assessed the prevalence and isotypes of antibodies patients, and the availability of specific biomarkers could provide against Nfasc155, Nfasc140/186, CNTN1, and Caspr1 in guidance for patient-tailored immunotherapeutic options.1 a large cohort of Italian patients affected by CIDP to better characterize their clinical associations. We also investigated Antibodies to cell adhesion molecules of the paranodal complex, skin biopsies from antibody-reactive patients for morpho- neurofascin-155 (Nfasc155), contactin-1 (CNTN1), and logical abnormalities and the pathogenicity of anti-Caspr1 contactin-associated protein 1 (Caspr1), and to nodal antibodies using in vitro and in vivo models. neurofascin-140/186 (Nfasc140/186) have been identified in various percentages of patients with CIDP, with IgG4 being the – predominant isotype of these antibodies.2 8 Moreover, IgG4- Methods seropositive patients show specific clinical features and a poor response to IVIG. Patients and sera The primary research question of this study is to test the frequency The prevalence of anti-CNTN1 and Nfasc155 IgG4 has been of antibodies to nodal/paranodal proteins in CIDP (Class III ev- well documented in cohorts of Japanese patients.2,5,6,9,10 It has idence). Sera from 342 patients fulfilling the diagnostic criteria for been shown that anti-CNTN1 IgG4 antibodies are pathogenic in CIDP17 (306 typical and 36 atypical) were collected in 11 Italian animal models and have a function-blocking activity.11,12 More- centers with specific expertise in neuromuscular disorders and were over, anti-CNTN1 and Nfasc155 IgG4 are associated with spe- all tested by ELISA and cell-based assay (CBA) for antibodies to cific alterations of the paranodal axo-glial contacts in nerve Nfasc155 and CNTN1 and by CBA only for antibodies to biopsies,13,14 suggesting that these antibodies induce conduction Nfasc140/186 and Caspr1 by 2 independent laboratories (Istituto defects in patients by altering paranode integrity, hence the term di Ricovero e Cura a Carattere Scientifico [IRCCS] Mondino paranodopathy. In Europe, the study of these antibodies has been Foundation, National Neurological Institute of Pavia, Italy, and so far limited to small cohorts of Spanish, French, and German Institute for Neurosciences of Montpellier, France). As controls,

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN samples from healthy controls (HCs, n = 60); patients with Results Guillain-Barré syndrome (GBS, n = 31), multifocal motor neuropathy (MMN, n = 13), inherited neuropathy (n = 18), other Serologic findings noninflammatory neuropathy (n = 52) including para- Of 342 sera from patients with CIDP tested, 10 (3%) were proteinemic neuropathy (n = 18) and toxic-metabolic neuropa- positive for antibodies to anti-Nfasc155 using ELISA and thy (n = 34), and patients with MS (n = 60) were tested. CBA (figure 1A). IgG isotypes were IgG4 in 7, IgG3 in 1, and Methods for ELISA18 and CBA6,10 were previously reported and undetectable in 2 patients (figure 1B). One of these latter are available as supplementary methods, links.lww.com/NXI/ patients showed IgG4 reactivity against Nfasc140/186 and A165. Clinical information was collected from seropositive was thus considered a Nfasc140/186-reactive patient. Three patients with CIDP and from 64 randomly selected seronegative patients (1%) tested positive on ELISA and CBA for anti- patients with CIDP. Weakness was graded as mild if the medical CNTN1 antibodies (figure 1A). IgG isotypes were IgG4 in 2 research council (MRC) score of the weakest muscle considered patients and mixed IgG3/IgG4 in 1 patient (figure 1B). All of either proximal or distal muscle groups was <5 and >3, sera reactive to Nfasc155 or CNTN1 by ELISA were also moderate if MRC = 3, or severe if MRC <3. Disability was graded confirmed by CBA, and no nonspecific reactivities by ELISA using the Overall Neuropathy Limitation Scale (ONLS).19 Re- only were identified. Six patients (2%) showed reactivity sponse to treatment was investigated post hoc and was defined as against Caspr1 on CBA (figure 1C). Three had an IgG4- a persistent improvement of disability with an ONLS change after predominant isotype, 2 had an IgG1 isotype, and 1 did not treatment ≥1. Reactivity of sera to teased fibers from murine have a detectable isotype. Overall, 19 patients (6%) had sciatic nerves was also evaluated as previously described.2 antibodies against one of these 4 targets. None of the HCs and patients with GBS, MMN, hereditary neuropathy, other Immunohistochemistry on skin biopsy noninflammatory neuropathies, or MS tested positive, in- Three patients with anti-Nfasc155 IgG4 antibodies, 1 patient dicating that the presence of antibodies to nodal/paranodal with anti-CNTN1 IgG3/IgG4 antibodies, 1 patient with anti- proteins accurately identify patients with CIDP with sensi- Caspr1 IgG4 antibodies, 1 patient with anti-Nfasc155 antibodies tivity 6% and specificity 100%. Of note, none of the sera from of an undetectable isotype, and 6 seronegative patients with 36 atypical CIDP cases tested positive. CIDP underwent 3-mm punch skin biopsies at 1 distal site of the lower limb as per EFNS/PNS guidelines.20 Nodal/internodal Sera from patients with IgG4 anti-Nfasc155, anti-CNTN1, or morphological analysis was performed by double immunofluo- anti-Caspr1 antibodies, but not anti-Nfasc155 of the IgG3 rescence staining for myelin basic protein and Nfasc or myelin isotype or undetectable isotype, consistently showed re- basic protein and Caspr1. Detailed methods are available in activity against paranodes when tested on teased fibers from supplementary methods, links.lww.com/NXI/A165. murine sciatic nerves (figure 1D). Sera with anti-Nfasc140/ 186 IgG4 that co-reacted against Nfasc155 stained the nodes Purification of patients’ antibodies of Ranvier on teased fibers. For in vitro and in vivo studies, IgG1 and IgG4 were purified from the serum of 1 Caspr1-reactive patient and the plasma of Clinical features a HC and their pathogenicity was tested by intraneural in- Patients with anti-Nfasc155 IgG4 antibodies jection on murine sciatic nerve and cell aggregation assay. Of patients with anti-Nfasc155 IgG4 antibodies, 4 were men Methods for in vitro and in vivo studies are detailed in sup- and 3 were women, with an average age at onset of 31 ± 18 plementary methods, links.lww.com/NXI/A164. years (table e-1, links.lww.com/NXI/A164). The onset of Statistics neuropathy was subacute in 5 patients. One case was initially Clinical data were presented as number (%) or mean (min-max). diagnosed with GBS. Categorical variables were compared in seronegative and seropositive patients using the χ2 test or Fisher exact test; unpaired All patients had symmetric 4-limb muscle weakness, which was 2-tailed Student t test was used for comparing continuous vari- mild to moderate in the upper limbs and proximal muscles of ables. Analysis of clinical variables was performed using STATA. the lower limbs but severe in 5 patients in the lower limb distal p values inferior to 0.05 were considered significant. muscles. Pinprick and position sense at the hallux were reduced or abolished in all patients. Tremor was also invariably present, Standard protocol approvals, registrations, and 6 had sensory ataxia. One patient reported neuropathic pain and patient consents at onset, and 1 had dysphagia. One of them presented optic The study was approved by local institutional ethical com- neuritis during steroid tapering. Brain MRI was thus performed mittees. Written informed consent was obtained from all but did not show evidence of disseminated inflammatory CNS patients (or guardians of patients) participating in the study. disease. The neuropathy was moderately to severely disabling with an ONLS at onset of 5.8 ± 2.2. Two patients required Data availability a wheelchair 1 month after the disease onset. The patients Anonymized data from this study will be shared by request received IVIG (n = 6), steroids (n = 7), azathioprine (n = 1), or from any qualified investigator. rituximab (n = 1). Only 1 patient responded to IVIG and

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 3 Figure 1 Reactivity to Nfasc155, CNTN1, and Caspr1 in CIDP by ELISA and CBA

(A) Serum samples from patients with CIDP (n = 342), MMN (n = 13), GBS (n = 31), genetic PN (n = 18), other noninflammatory PN (n = 52), MS (n = 60) and from HCs (n = 60) were tested for autoantibodies to human Nfasc155 (left) and CNTN1 (right) by ELISA. OD are shown after subtraction of the baseline OD readingto bovine serum albumin. The red line represents the mean OD in HCs plus 3 standard deviations. (B) IgG isotype in Nfasc155- and CNTN1-positive patients. (C) The sera (here case 14) were tested on living HEK cells transfected with CNTN1 and Caspr1 (red) and then revealed with mouse antihuman IgG1, IgG2, IgG3, or IgG4 (green) as indicated. Nuclei were stained with DAPI (blue). (D) These are teased fibers from mouse sciatic nerves immunostained for CNTN1 (red) and the serum from case 14, then revealed with mouse antihuman IgG1, IgG2, IgG3, or IgG4 (green) as indicated. IgG1 and IgG4 from this patient reacted against Caspr1 and bound to the paranodal regions. Scale bars: 10 μm. Caspr1 = contactin-associated protein 1; CBA = cell-based assay; CIDP = chronic inflammatory demyelinating polyradiculoneuropathy; CNTN1 = contactin-1; GBS = Guillain-Barr´e syndrome; HC = healthy control; HEK = human embryonic kidney; MMN = multifocal motor neuropathy; Nfasc155 = neurofascin-155; OD = Optical density; PN = peripheral neuropathy.

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN steroids (ONLS 4 → 2), 1 patient showed response to steroids subacute in 1 patient only. If present, weakness was graded as (ONLS 6 → 2),and1patientunresponsivetoIVIGandste- mild to moderate in upper limbs and proximal muscles of lower roids had sustained response after rituximab (ONLS 9 → 2). limbs. All 5 patients had weakness in lower limb distal muscles, which was severe in 2 of them. Pinprick sensation was reduced Patients with anti-CNTN1 IgG4 antibodies in 3 and position sense was reduced at the hallux in 2 patients, Of patients with anti-CNTN1 IgG4 antibodies, 3 were men but none had absent proprioception. Sensory ataxia was with a mean age at onset of 56 ± 26 years (table e-1, links.lww. reportedin4patients,2withanti-Nfasc155 antibodies and 2 com/NXI/A164). The onset of neuropathy was subacute in 2 with anti-Caspr1 antibodies, whereas tremor was observed in 2 patients. Strength was impaired in both proximal and distal patients with anti-Caspr1 antibodies. Disability was usually low muscles in the upper and lower limbs, often to a severe degree, with a mean ONLS at onset of 3.2 ± 1.3. Only 1 case with anti- and 2 of them required a wheelchair. Pinprick sensation and Caspr1 antibodies needed walking aids. The patients were proprioception were either severely reduced or abolished, and treatedwithIVIG(n=4)andsteroids (n = 2). One patient with all of them had sensory ataxia. One patient showed cranial minimal disability did not require treatment. Three of 4 patients nerve involvement and respiratory failure. One patient with treated with IVIG had a good response. anti-CNTN1 IgG3/IgG4 antibodies had a concurrent onset of CIDP and nephrotic syndrome because of membranous glo- Comparison of seropositive vs seronegative patients merulonephritis. A kidney biopsy showed subepithelial We compared the clinical features and responses to treat- deposits of immune complexes and complement deposition. ments of patients with IgG4 antibodies against Nfasc155, CNTN1 patients had the highest disability at onset with an CNTN1, or Caspr1 with those of 64 randomly selected ONLS of 7.6 ± 2.5. One patient died of cardiac disease 6 patients with CIDP who were negative for antibodies against months after the disease onset. The patients were treated with Nfasc155, Nfasc140/186, CNTN1, or Caspr1. We also IVIG (n = 3), steroids (n = 3), plasma exchange (n = 2), and compared seropositive patients with an IgG1, IgG3, or un- immunosuppressive drugs (n = 2). None of them had a sus- detectable IgG isotype with seronegative patients (table 1). tained response to either IVIG or steroids. Cyclophosphamide was effective in 2 patients, leading to significant improvement Anti-Nfasc155 and anti-Caspr1 IgG4 seropositive patients had of the neuropathy in 1 case and complete remission of the more frequently an early disease onset, before 30 years of age, neuropathy together with regression of the nephrotic syn- as opposed to that of anti-CNTN1 IgG4 seropositive patients drome in the other case, who did not require further treatment. and seronegative patients. A subacute onset was reported in up to two-thirds of anti-Nfasc155, anti-CNTN1, and anti-Caspr1 Patients with anti-Caspr1 IgG4 antibodies IgG4 seropositive patients compared with 28% of seronegative Of seropositive patients with anti-Caspr1 IgG4 antibodies, 2 patients. According to the EFNS/PNS criteria, CIDP pheno- were men and 1 was a woman, with a mean age at onset of 43 ± type was typical in all IgG4-seropositive patients. Tremor was 17 years (table e-1, links.lww.com/NXI/A164). The onset was invariably present in patients with anti-Nfasc155 IgG4 anti- subacute in 2 patients. Weakness was mild to moderate in bodies but was also observed in anti-CNTN1 and anti-Caspr1 proximal muscle groups and moderate to severe in distal IgG4 seropositive patients. Severe proprioceptive loss and muscle groups in both upper and lower limbs. Pinprick sen- sensory ataxia were significantly more frequent in IgG4- sation was reduced, and position sense was abolished at the seropositive patients and were observed in all patients with hallux in all of them. None of them reported neuropathic pain. anti-CNTN1 antibodies, 2 patients with anti-Caspr1 IgG4 One patient complained of dysphagia, and 1 patient had antibodies, and 3 patients with anti-Nfasc155 IgG4 antibodies tremor. The disease was severe, with a mean ONLS of 8.3 ± 0.6. (table 1). In anti-CNTN1 and anti-Caspr1 IgG4 seropositive The patients were treated with IVIG (n = 3), steroids (n = 3), patients, weakness was usually graded as moderate to severe plasma exchange (n = 1), methotrexate (n = 1), and interferon- across proximal and distal muscle groups, whereas in patients alfa (n = 1). A sustained response was observed only in 1 with anti-Nfasc155 IgG4 antibodies, motor impairment patient receiving steroids and methotrexate (ONLS 9 → 6) and showed a characteristic distal predominance in the lower limbs. 1 patient treated with interferon-alfa (ONLS 8 → 7). A higher level of CSF total protein and the presence of tem- Seropositive patients with an IgG1-3 isotype or poral dispersion on nerve conduction studies were also char- undetectable IgG isotype acteristic features of patients with autoantibodies of the IgG4 Six patients had IgG1 or IgG3 antibodies or an undetectable IgG isotype. IgG4-seropositive patients had a higher disability at isotype against Nfasc155 (n = 3) or Caspr1 (n = 3) (table e-1, onset: 5 patients needed a walking aid, and 5 patients required links.lww.com/NXI/A164). One of these patients presented a wheelchair. In particular, patients with anti-CNTN1 or anti- antibodies reacting to both Nfasc155 and Nfasc140/186. The Caspr1 IgG4 antibodies showed the highest disability at onset. clinical features of this patient, mainly consisting in a very early onset IVIG-responsive neuropathy, have been previously de- Patients with anti-Nfasc155, CNTN1, or Caspr1 IgG4 anti- tailed.16 We will thus focus here on the 5 other patients, 2 men bodies had a lower response rate to IVIG compared with and 3 women. In 2 of them, both reactive against Caspr1, the seronegative patients (8% vs 67%, p < 0.001), but there was disease started in the firstdecadeoflife.Themodeofonsetwas no remarkable difference in the response to steroids or

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 5 Table 1 Clinical features of patients with CIDP and antibodies to Nfasc155, CNTN1, and Caspr1

Anti-IgG1-3 or undetectable anti-Nfasc155 or anti-Caspr1 Autoantibody Anti-paranodal Anti-Nfasc155 Anti-CNTN1 Anti-Caspr1 IgG subclass seronegative (N = 64) proteins IgG4 (N = 13) IgG4 (N = 7) IgG4 (N = 3) IgG4 (N = 3) (N = 5)

Age at onset 39 (18–64) 36 (13–82) 22 (13–63) 58 (30–82) 46 (24–57) 56 (7–74)

Early onset (below 7 (11%) 5 (38%)a 4c 0 1 2 (40%) 30 years of age)

Male sex 42 (66%) 9 (69%) 4 3 2 2 (40%)

Other autoimmune 8 (12.5%) 3 (23%) 2 1 0 1 (20%) disease

M protein 15 (23%) 0 (0%) 0 0 0 0 (0%)

Triggering infection/ 11 (17%) 3 (23%) 2 0 1 0 (0%) vaccination

Clinical phenotype

Typical 52 (81%) 13 (100%) 7 3 3 5 (100%)

Atypical 12 (19%) 0 (0%) 0 0 0 0 (0%)

Subacute onset 18 (28%) 9 (69%)b 5a 2 2 1 (20%)

Weakness moderate/severe

UL proximal 12 (19%) 4 (31%) 0 2 2 0 (0%)

UL distal 32 (50%) 8 (61%) 3 2 3 1 (20%)

LL proximal 16 (25%) 6 (46%) 1 2 3 1 (20%)

LL distal 37 (58%) 12 (92%)a 7a 2 3 3 (60%)

Pinprick sensation at the hallux

Reduced 43 (67%) 10 (83%) 5 3 3 3 (60%)

Abolished 2 (3%) 1 (8%) 1 0 0 0 (0%)

Position sensation at the hallux

Reduced 27 (43%) 4 (30%)b 3 0 1 2 (40%)

Abolished 8 (13%) 8 (61%)b 33a 2 0 (0%)

Sensory ataxia 22 (35%) 11 (85%)b 6a 3a 2 4 (80%)

Tremor 8 (12%) 9 (69%)c 7c 1 1 2 (40%)

Pain 23 (36%) 3 (23%) 1 2 0 0 (0%)

CSF protein 96 (24–55) 350 (128–679)c 278 (142–679) 148 (128–350) 426 (343–510) 68 (45–586)

ONLS 5(1–11) 8 (3–10)a 5(3–9) 8 (5–10) 8 (8–9) 3 (2–5)

NCS features

Prolonged DML 30 (51%) 9 (70%) 5 1 3 5 (100%)

Reduced CV 52 (88%) 11 (85%) 5 3 3 5 (100%)

Prolonged F wave 35 (62%) 4 (36%) 3 0 1 4 (100%)

Conduction 30 (51%) 5 (42%) 3 1 1 3 (60%) blocks

Temporal 15 (25%) 7 (64%)a 4 2 1 4 (80%)a dispersion

Continued 6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN Table 1 Clinical features of patients with CIDP and antibodies to Nfasc155, CNTN1, and Caspr1 (continued)

Response to IVIG

No 4 (7%) 4 (31%)c 1 2 1 0 (0%)

Partial/transitory 14 (26%) 8 (61%)c 5 1 2 1 (25%)

Good 37 (67%) 1 (8%)c 1a 0b 0 3 (75%)

Response to steroids

No 14 (28.5%) 3 (25%) 0 2 1 1 (50%)

Partial/transitory 21 (43%) 6 (50%) 4 1 1 1 (50%)

Good 14 (28.5%) 3 (25%) 2 0 1 0 (0%)

Response to PEX

No 1 (8%) 2 (67%) 0 1 1 0 (0%)

Partial/transitory 6 (46%) 0 (0%) 0 0 0 0 (0%)

Good 6 (46%) 1 (33%) 0 1 0 0 (0%)

Response to immune suppressors

No 5 (19%) 0 (0%) 0 0 0 0 (0%)

Partial/transitory 13 (50%) 1 (20%) 1 0 0 0 (0%)

Good 8 (31%) 4 (80%) 1 2 1 0 (0%)

Abbreviations: Caspr1 = contactin-associated protein 1; CIDP = chronic inflammatory demyelinating polyradiculoneuropathy; CNTN1 = contactin-1;CV= conduction velocity; DML = distal motor latency; IVIG = intravenous immunoglobulin; LL = lower limb; NCS = nerve conduction study; Nfasc155 = neurofascin- 155; ONLS = Overall Neuropathy Limitation Scale; PEX = plasma exchange; UL = upper limb. a p < 0.05. b p < 0.01. c p < 0.005.

immunosuppressive treatment. In particular, 1 patient with anti- antibodies, 1 patient with anti-Caspr1 IgG4 antibodies, 1 pa- Nfasc155 IgG4 antibodies was treated with rituximab, 1 patient tient with anti-Nfasc155 antibodies of an undetectable isotype, with anti-Nfasc155 IgG4 antibodies was treated with azathio- and 6 seronegative patients with CIDP. Analysis of myelinated prine, 2 patients with anti-CNTN1 IgG4 antibodies were treated fibers from patients with anti-Nfasc155 and anti-Caspr1 IgG4 with cyclophosphamide, and 1 patient with anti-Caspr1 IgG4 showed elongation of the nodes of Ranvier and loss of para- antibodies was treated with methotrexate. Four of them who were nodal Nfasc155 and Caspr1 staining. Moderate elongation of started on immunosuppressors within 1 year from disease onset the nodes of Ranvier and loss of Nfasc155 paranodal staining showed long-lasting good response. were also observed in myelinated fibers of the patient with anti- CNTN1 IgG3/IgG4 antibodies. Contrarily, we did not observe Patients with antibodies to Nfasc155, CNTN1, or Caspr1 of the similar changes in the patient with anti-Nfasc155 antibodies IgG1, IgG3, or undetectable isotype did not show distinct clinical with an undetectable isotype, seronegative patients CIDP, or features or response to treatments compared with seronegative HCs (figure 2). patients, except for more frequent temporal dispersion on nerve conduction studies. Pathogenic effects of anti-Caspr1 IgG4 and IgG1 antibodies Morphological changes of the nodes of Ranvier in skin Anti-CNTN1 IgG4 antibodies were previously found to have biopsy of seropositive patients function-blocking activity and to disrupt the interaction be- We evaluated skin biopsies from 3 patients with anti-Nfasc155 tween CNTN1 and its glial partner Nfasc155.12 We have IgG4 antibodies, 1 patient with anti-CNTN1 IgG3/IgG4 previously shown that anti-CNTN1 IgG4 and anti-Nfasc155

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 7 Figure 2 Morphological alterations of the nodes of Ranvier in patients with CIDP with IgG4 autoantibodies

We evaluated skin biopsies from 3 patients with IgG4 anti-Nfasc155 antibodies, 1 patient with IgG3/IgG4 anti-CNTN1 antibodies, 1 patient with IgG4 anti- Caspr1 antibodies, 1 patient with undetectable isotype IgG anti-Nfasc155, and 6 seronegative patients with CIDP. Analysis of myelinated fibers showed elongation of the nodes of Ranvier and loss of paranodal Nfasc155 staining in skin biopsies from patients with anti-Nfasc155 (C) and Caspr1 (E) IgG4. Moderate elongation of the nodes of Ranvier and loss of Nfasc155 paranodal staining were also observed in myelinated fibers of a CNTN1 IgG3/IgG4-positive patient (D). Contrarily, we did not observe similar changes in the patient with undetectable isotype IgG anti-Nfasc155 antibodies (F), in seronegative patients with CIDP (B), or in HCs (A). A complete loss of Caspr1 staining was observed in biopsies from patients with IgG4 antibodies to paranodal proteins (I, L, M), but not in an Nfasc155 seropositive patient with an undetectable isotype (N) and in seronegative CIDP (H) or healthy patients (G). Caspr1 = contactin-associated protein 1; CIDP = chronic inflammatory demyelinating polyradiculoneuropathy; CNTN1 = contactin-1; HC = healthy control; Nfasc155 = neurofascin-155.

IgG4 antibodies are pathogenic and disrupt the paranodal express only green or red cells. The expression of CNTN1/ regions.11,21 We thus examined whether anti-Caspr1 antibodies Caspr1 significantly increases the number of clusters with con- also affect the integrity of the CNTN1/Caspr1/Nfasc155 tacts, and most cell clusters show an average of 50% of CNTN1/ complex. In particular, we purified IgG1 and IgG4 and tested Caspr1- and 50% of Nfasc155-expressing cells. The presence of the effects of the different antibody isotypes. First, we tested the 10 μg of IgG4 from a Caspr1-positive patient, but not IgG1, impact of IgG4 and IgG1 fractions from Caspr1-positive patients abolished the interaction between CNTN1/Caspr1- and on the interaction of CNTN1/Caspr1 with Nfasc155 by a cell Nfasc155-expressing cells to the level of negative controls, thus aggregation assay (figure 3). Human embryonic kidney cells supporting a blocking function of anti-Caspr1 IgG4 antibodies. were transfected with Nfasc155 and mCherry and incubated for 2 hours with cells coexpressing CNTN1/Caspr1 and green To further confirm these findings, mouse sciatic nerve segments fluorescent protein (GFP). The percentage of cell clusters were incubated in vitro with 10 μgofIgG4orIgG1fractions showing contacts between red and green cells and the percent- from Caspr1-positive patients for 3 hours, and IgG deposition age of green and red cells per cell clusters were quantified. As was monitored. IgG4 but not IgG1 antibodies were found to a negative control, Nfasc155-mCherry-transfected cells were penetrate the paranodal regions (figure 4). These findings were incubated with cells expressing GFP only. In the absence of replicated in vivo by performing intraneural injections. Again, CNTN1/Caspr1, Nfasc155-expressing cells do not form signif- only IgG4 antibodies were found to penetrate the paranodal icant contact with GFP-expressing cells, and most clusters regions. Of interest, the level of antibody penetration across the

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN Figure 3 IgG4 to Caspr1 disrupt the interaction between Nfasc155 and CNTN1/Caspr1

(A–D) HEK cells transfected with CNTN1 and Caspr1 (green) or Nfasc155 (red) were incubated together for 2 hours in the presence of control IgG4 (B), anti- Caspr1 IgG1 (C), or anti-Caspr1 IgG4 (D). As negative controls, HEK cells transfected with Nfasc155 (red) were incubated with cells transfected with GFP (A). Anti-Caspr1 IgG4, but not anti-Caspr1 IgG1, abrogated the aggregation of Nfasc155-transfected cells with CNTN1/Caspr1. Scale bar: 10 μm. (E–F) The graphs represent the relative frequency of green cells per aggregates (n = 4 experiments for each condition). The percentage of cell clusters with contacts between green and red cells was quantified (F). The percentage of contacts was significantly decreased in the presence of anti-Caspr1 IgG4 (p < 0.005 by unpaired 2- tailed Student t tests and by one-way ANOVA, followed by Bonferroni post hoc tests). Bars represent mean and SEM. ANOVA = analysis of variance; Caspr1 = contactin-associated protein 1; CNTN1 = contactin-1; GFP = green fluorescent protein; HEK = human embryonic kidney; Nfasc155 = neurofascin-155. paranodal region was similar at 1 or 3 days after intraneural across the 2 testing centers in Pavia and Montpellier, in- injection, and IgG4 deposition was detected only at the border of dicating that these techniques, although not broadly available, the nodes of Ranvier (figure 4). No paranodes presented have a high reproducibility if performed in specialized centers. a complete invasion by IgG4 at 1 or 3 days postinjection. Our findings largely confirm previous observations by several groups showing a tight clinical serologic correlation of anti- Discussion bodies to nodal/paranodal proteins and clinical features of seropositive patients with CIDP. Antibodies to nodal/paranodal protein were found, albeit with a low frequency, in Italian patients with CIDP and are associ- The patients with anti-Nfasc155 IgG4 antibodies showed ated with target- and isotype-specific clinical features, as earlier onset, distal predominant lower limb weakness, and previously reported. The prevalence of anti-Nfasc155 and anti- CNTN1 antibodies was similar to that reported by previous gait disturbance. Tremor was also present, although only in 1 studies in European patients but was lower compared with that case this was disabling. One patient with such antibodies in Japanese patients.5,10 The discrepancy may be attributable to developed bilateral optic neuritis during steroid tapering 18 differences in inclusion criteria, ethnicities, and non- months after the onset of the neuropathy. standardized laboratory techniques. The frequency of anti- Caspr1 IgG4 antibodies was equal to that of antibodies against The patients with anti-CNTN1 antibodies were older and CNTN1, confirming that Caspr1 may also represent a relevant showed a subacute or rapidly progressive severe sensory and target of the immune response in Caucasian patients with motor neuropathy. Previous studies found that anti-CNTN1 CIDP. None of the healthy or pathologic controls tested, in- antibodies were associated with either predominant motor or fi cluding patients affected by other neuropathies, showed re- sensory impairment. Here, both sensory and motor bers were activity against Nfasc155, CNTN1, or Caspr1. These equally severely affected, and the patients became wheelchair maximum levels of specificity entail that the tests perform very dependent few months after the disease onset. Notably, 1 pa- well for the diagnosis of such CIDP subtypes. tient with anti-CNTN1 antibodies of IgG3 and IgG4 subclasses showed the contemporary occurrence of membranous glo- In addition, it is worth noting that the results of antibody merulonephritis. This association had been reported in 3 cases testing by ELISA and CBA were concordant in all samples with anti-CNTN1 antibodies,7,22,23 and either a direct damage

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 9 Figure 4 Anti-Caspr1 IgG4, but not IgG1, invades the paranodal regions

(A–C) Sciatic nerve fibers were incubated in vitro with purified control IgG4 (A), anti-Caspr1 IgG1 (B), or anti-Caspr1 IgG4 (C) for 3 hours and immunolabeled for IgG (green) and CNTN1 (red). (D–F) Sciatic nerves were fixed 1 or 3 days after intraneural injections of purified control IgG4 (D), anti-Caspr1 IgG4 (E–F), immunolabeled for CNTN1 (red), and human IgG (green). Note that only anti-Caspr1 IgG4 penetrated the paranodes. One or 3 days after injection, IgG4 antibodies were detected at the paranode borders (arrows). Images are representative of 3 independent experiments. Scale bar: 10 μm. Caspr1 = contactin-associated protein 1; CNTN1 = contactin-1.

or indirect damage, after the deposition of immune complexes, elongation of the node and loss of neurofascin and Caspr1 can be hypothesized. Of interest, a recent study showed that staining at paranodes, which were absent or less prominent in complement deposition may contribute to the pathophysiology seronegative CIDP cases or in the case with anti-Nfasc155 of anti-CNTN1-associated neuropathy, particularly in patients antibodies of an undetectable isotype. These data add to other with a predominance of the IgG3 subclass.24 recent histopathologic and neurophysiologic observations in patients with antibodies to nodal/paranodal proteins14,25 and Neuropathy in patients with anti-Caspr1 IgG4 antibodies was indicate that isotype determination appears to be crucial to also highly debilitating. Pain did not seem to be a clinical correctly identify such patients and to guide treatment. feature associated with the presence of anti-Caspr1 antibodies in our series, as opposed to the first report of 2 patients with The high frequency of anti-Caspr1 antibodies in our series Caspr1-associated inflammatory neuropathy.8 prompted us to further investigate their pathogenic effects. Albeit the experimental setting was limited to the examination As a novel observation of our study, we found that patients with of the effects of anti-Caspr1 autoantibodies from a single pa- anti-Nfasc155 or Caspr1 antibodies of the IgG1, IgG3, or un- tient, we found that anti-Caspr1 IgG4, but not anti-Caspr1 detectable IgG isotype did not show clinical features distinct IgG1, antibodies affect the interaction between the CNTN1/ from seronegative patients. Most importantly, patients with Caspr1 and Nfasc155 complex using a cell aggregation assay. anti-Nfasc155, CNTN1, or Caspr1 IgG4 antibodies, but not These findings are in line with the previously reported patho- patients with antibodies to Nfasc155 or Caspr1 of IgG1, IgG3, genic role of anti-CNTN1 IgG4 antibodies11,12 and suggest or undetectable IgG isotype, showed a significantly lower re- that IgG4 antibodies targeting the CNTN1/Caspr1 complex sponse rate to IVIG compared with seronegative patients. may have a function-blocking activity and disrupt the paranodal axo-glial contact. In keeping with this experimental evidence, Analysis of myelinated fibers in the skin biopsy of patients with our data also seem to indicate that anti-Caspr1 IgG4 antibodies anti-Nfasc155, CNTN1, or Caspr1 IgG4 antibodies showed can penetrate the paranodal regions. It is worth noting that the morphological changes of the nodes of Ranvier, including pathogenic mechanism of IgG4 antibodies in other diseases

10 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN implicates the disruption of cell adhesion protein complexes the NIH Rare Diseases Clinical Research Network (RDCRN) too, by inhibiting protein-protein interaction, and that thera- (U54NS065712). pies aiming at downregulating the humoral immune response, such as rituximab, showed some efficacy in anti-Nfasc155- and Disclosure anti-Caspr1 antibody-associated CIDP, as well as in other The authors report no disclosures. Go to Neurology.org/NN IgG4-related diseases, likely because of the depletion of the for full disclosures. Nfasc155- or Caspr1-reactive B cells.26 Publication history Despite the limitation entailed by the retrospective data Received by Neurology: Neuroimmunology & Neuroinflammation June collection, in the present series, immunosuppressive treat- 14, 2019. Accepted in final form October 3, 2019. ment, including rituximab, cyclophosphamide, and methotrexate, seemed effective in IVIg-resistant IgG4 seropositive patients if started early in the disease course. Of note, we did not detect a significant difference in the response to steroids Appendix Authors between patients with anti-paranodal IgG4 antibodies and Name Location Role Contribution seronegative patients, confirming steroids as an effective Andrea Department of Brain Author Designed and therapeutic option in CIDP cases, independent from their Cortese, MD, and Behavioral conceptualized the serologic status for anti-Nfasc155, CNTN1, or Caspr1 PhD Sciences, University of study, major role in Pavia, Pavia, Italy; the acquisition of antibodies. Low numbers of patients treated with plasma Department of data, analyzed the exchange, and the temporal overlap of plasma exchange Neuromuscular data, and drafted cycles with other treatments (case 12), prevented us from Disease, UCL Queen the manuscript for Square Institute of intellectual content drawing firm conclusions on its role in patients with anti- Neurology, London, paranodal IgG4 antibodies. United Kingdom Raffaella Neuroalgology Unit, Author Designed and Although patients with antibodies against the nodal/ Lombardi, IRCCS Fondazione conceptualized the MA Istituto Neurologico study, major role in paranodal component often, albeit not always, show addi- “Carlo Besta,” Milan, the acquisition of tional clinical features, CIDP phenotype was typical in all Italy data, analyzed the data, and drafted IgG4 and non-IgG4 seropositive patients according to the the manuscript for EFNS/PNS criteria, whereas none of the sera from atypical intellectual content

CIDP cases or control sera showed the presence of these Chiara Department of Author Major role in the antibodies. Together, this observation challenges the view Briani, MD Neurosciences, acquisition of data, University of Padova, analyzed the data, that the search for antibodies against the nodal/paranodal Padova, Italy and drafted the component may be limited to atypical CIDP cases. manuscript for intellectual content

Testing for the presence of antibodies against Nfasc155, Ilaria Neuroscience Author Major role in the CNTN1, and Caspr1, followed by IgG isotype determination Callegari, Consortium, University acquisition of data, MD of Pavia, Monza analyzed the data, in seropositive cases should be part of the diagnostic workup Policlinico and Pavia and drafted the in inﬂammatory neuropathies to improve diagnostic accuracy Mondino, Italy manuscript for intellectual content and guide treatment. Moreover, knowledge of the mechanism underlying these CIDP subtypes might shed light on the Luana Department of Author Major role in the Benedetti, Neuroscience, acquisition of data pathophysiology and help further understanding of this MD, PhD Rehabilitation, and revised the complex and heterogeneous disease. Ophthalmology, manuscript for Genetics, Maternal and intellectual content Child Health (DiNOGMI), Study funding University of Genova, Genova, Italy; IRCCS This work was supported by the Italian Ministry of Health Ospedale Policlinico San “Ricerca Corrente” 2017–2019 Grant (Grant code: Martino, Genova, Italy

RC1812C) to the IRCCS Mondino Foundation, GBS/CIDP Fiore Department of Author Major role in the Foundation Non-proﬁt Grant 2017 Grant code: 501(c)(3) to Manganelli, Neurosciences, acquisition of data MD Odontostomatological and revised the Fondazione IRCCS Istituto Neurologico Carlo Besta, and and Reproductive manuscript for Agence Nationale pour la Recherche (NECCIN; JD) and by Sciences, University of intellectual content the Association Française contre les Myopathies Naples “Federico II,” Naples, Italy (grant#21532; JD), the Italian Ministry of Health “Ricerca Finalizzata 2016 (Grant code: RF-2016-02361887) to Marco Fondazione Policlinico Author Major role in the Luigetti, MD, Universitario Agostino acquisition of data Humanitas Clinical and Research Center, IRCCS. Andrea PhD Gemelli-IRCCS. UOC and revised the Cortese is funded by MRC (MR/T001712/1) and received Neurologia, Universit`a manuscript for Cattolica del Sacro intellectual content funding from Wellcome Trust (204841/Z/16/Z) and the Cuore, Roma, Italy Inherited Neuropathy Consortium (INC), which is a part of Continued Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 11 Appendix (continued) Appendix (continued)

Name Location Role Contribution Name Location Role Contribution

Sergio Section of Neurology, Author Major role in the Matteo IRCCS Mondino Author Major role in the Ferrari, MD Department of acquisition of data Gastaldi, Foundation, Pavia, Italy acquisition of data Neuroscience, and revised the MD, PhD and revised the Biomedicine and manuscript for manuscript for Movement Sciences, intellectual content intellectual content University of Verona, Verona, Italy Elisa Vegezzi, Neuroscience Author Major role in the MD Consortium, University acquisition of data, Maurizio Department of Author Major role in the of Pavia, Monza analyzed the data, Clerici, MD Neurology and Stroke acquisition of data Policlinico and Pavia and revised the Unit, Ospedale di and revised the Mondino, Italy manuscript for Circolo/Fondazione manuscript for intellectual content Macchi, Varese, Italy intellectual content Enrico IRCCS Mondino Author Major role in the Girolama Department of Author Major role in the Alfonsi, MD Foundation, Pavia, Italy acquisition of data Alessandra Systems Medicine, acquisition of data and revised the Marfia, MD University of Rome Tor and revised the manuscript for Vergata, Rome, Italy manuscript for intellectual content intellectual content Angela IRCCS Mondino Author Major role in the Andrea Neurological Author Major role in the Berardinelli, Foundation, Pavia, Italy acquisition of data Rigamonti, Department, ASST acquisition of data MD and revised the MD Lecco and revised the manuscript for manuscript for intellectual content intellectual content Ludivine Referral Center for Author Major role in the Marinella Neurology Unit, Author Major role in the Kouton, MD Neuromuscular acquisition of data Carpo, MD, Ospedale Treviglio, acquisition of data Diseases and ALS, and revised the PhD ASST Bergamo Ovest, and revised the Timone University manuscript for Italy manuscript for Hospital, Marseille, intellectual content intellectual content France

Raffaella Department of Author Major role in the Constance Universit´ede Author Major role in the Fazio, MD Neurology, San acquisition of data Manso, MSc Bordeaux, acquisition of data Raffaele Scientific and revised the Interdisciplinary and revised the Institute, Milan, Italy manuscript for Institute for manuscript for intellectual content Neuroscience, intellectual content Bordeaux, France Massimo Department of Author Major role in the Corbo, MD Neurorehabilitation acquisition of data Claudia Humanitas Clinical and Author Major role in the Sciences, Casa Cura and revised the Giannotta, Research Center, acquisition of data Policlinico (CCP), Milan, manuscript for MA IRCCS, Milan, Italy and revised the Italy intellectual content manuscript for intellectual content Anna Department of Clinical Author Major role in the Mazzeo, MD and Experimental acquisition of data Pietro Humanitas Clinical and Author Major role in the Medicine, University of and revised the Doneddu, Research Center, acquisition of Messina, Messina, Italy manuscript for MD IRCCS, Milan, Italy data and revised intellectual the manuscript content for intellectual content Fabio Department of Author Major role in the Giannini, MD Medicine, Surgery and acquisition of data Patrizia Neuroalgology Unit, Author Major role in the Neurosciences, and revised the Dacci, MD, IRCCS Fondazione acquisition of data University of Siena, manuscript for PhD Istituto Neurologico and revised the Siena, Italy intellectual content “Carlo Besta,” Milan, manuscript Italy for intellectual Giuseppe IRCCS Mondino Author Major role in the content Cosentino, Foundation, acquisition of data MD, PhD Department of Brain and revised the Laura Neuroalgology Unit, Author Major role in the and Behavioral manuscript for Piccolo, MD IRCCS Fondazione acquisition of data Sciences, University of intellectual content Istituto Neurologico and revised the Pavia, Pavia, Italy “Carlo Besta,” Milan, manuscript for Italy intellectual content Elisabetta IRCCS Mondino Author Major role in the Zardini, MA Foundation, Pavia, Italy acquisition of data Marta Ruiz, Department of Author Major role in the and revised the MD Neurosciences, acquisition of data manuscript for University of Padova, and revised the intellectual content Padova, Italy manuscript for intellectual content Riccardo Neuroscience Author Major role in the Curro,` MD Consortium, University acquisition of data, Alessandro Department of Author Major role in the of Pavia, Monza analyzed the data, Salvalaggio, Neurosciences, acquisition of data Policlinico and Pavia and revised the MD University of Padova, and revised the Mondino, Italy manuscript for Padova, Italy manuscript for intellectual content intellectual content

12 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN Appendix (continued) Appendix (continued)

Name Location Role Contribution Name Location Role Contribution

Chiara De Department of Author Major role in the Angelo Department of Author Major role in the Michelis, MD Neuroscience, acquisition of data Schenone, Neuroscience, acquisition of data Rehabilitation, and revised the MD Rehabilitation, and revised the Ophthalmology, manuscript for Ophthalmology, manuscript for Genetics, Maternal intellectual content Genetics, Maternal and intellectual content and Child Health Child Health (DiNOGMI), (DiNOGMI), University of Genova, University of Genova, IRCCS Ospedale IRCCS Ospedale Policlinico San Martino, Policlinico San Martino, Genova, Italy Genova, Italy Enrico IRCCS Mondino Author Major role in the Emanuele Department of Author Major role in the Marchioni, Foundation, Pavia, Italy acquisition of data Spina, MD Neurosciences, acquisition of data MD and revised the Odontostomatological and revised the manuscript for and Reproductive manuscript for intellectual content Sciences, University of intellectual content Naples “Federico II,” Giuseppe Neuroalgology Unit, Author Designed and Naples, Italy Lauria, MD IRCCS Fondazione conceptualized the Istituto Neurologico study, major role in Antonietta Department of Author Major role in the “Carlo Besta,” Milan, the acquisition of Topa, MD Neurosciences, acquisition of data Italy; Department of data, analyzed the Odontostomatological and revised Biomedical and Clinical data, and drafted and Reproductive the manuscript Sciences “Luigi Sacco,” the manuscript for Sciences, University of for intellectual University of Milan, intellectual content Naples “Federico II,” content Milan, Italy Naples, Italy Eduardo Humanitas Clinical and Author Major role in the Giulia Centro Clinico NEMO Author Major role in Nobile- Research Center, acquisition of data Bisogni, MD Adulti, Roma, Italy the acquisition of Orazio, MD IRCCS, Milan, Italy and revised the data and revised manuscript for the manuscript intellectual content for intellectual content J´eromeˆ Institute for Author Designed and Devaux, PhD Neurosciences of conceptualized the Angela Institute of Neurology, Author Major role in Montpellier University, study, major role in Romano, MD Universit`a Cattolica del the acquisition of Hopital Saint Eloi, the acquisition of Sacro Cuore, Rome, data and revised Montpellier, France. data, analyzed the Italy the manuscript data, and drafted for intellectual the manuscript for content intellectual content

Sara Section of Neurology, Author Major role in Diego IRCCS Mondino Author Designed and Mariotto, Department of the acquisition Franciotta, Foundation, Pavia, Italy conceptualized the MD Neuroscience, of data and revised MD study, major role in Biomedicine the manuscript the acquisition of and Movement for intellectual data, analyzed the Sciences, University content data, and drafted of Verona, Verona, Italy the manuscript for intellectual content Giorgia Department Author Major role in Mataluni, of Systems the acquisition MD, PhD Medicine, University of of data and revised Rome Tor the manuscript Vergata, Rome, Italy for intellectual References content 1. Doneddu PE, Cocito D, Manganelli F, et al. Atypical CIDP: diagnostic criteria, progression and treatment response. Data from the Italian CIDP database. J Neurol Federica Department of Author Major role in the Neurosurg Psychiatry 2019;90:125–132. Cerri, MD, Neurology, San acquisition of data 2. Devaux JJ, Odaka M, Yuki N. Nodal proteins are target antigens in Guillain-Barré PhD Raffaele and revised the syndrome. J Peripher Nerv Syst 2012;17:62–71. Scientific Institute, manuscript 3. Querol L, Nogales-Gadea G, Rojas-Garcia R, et al. Antibodies to contactin-1 in Milan, Italy for intellectual chronic inflammatory demyelinating polyneuropathy. Ann Neurol 2013;73:370–380. content 4. Querol L, Nogales-Gadea G, Rojas-Garcia R, et al. Neurofascin IgG4 antibodies in CIDP associate with disabling tremor and poor response to IVIg. Neurology 2014;82:879–886. Claudia IRCCS Centro Author Major role in 5. Ogata H, Yamasaki R, Hiwatashi A, et al. Characterization of IgG4 anti-neurofascin Stancanelli, Neurolesi the acquisition of 155 antibody-positive polyneuropathy. Ann Clin Transl Neurol 2015;2:960–971. MD “Bonino Pulejo,” data and revised 6. Miura Y, Devaux JJ, Fukami Y, et al; CNTN1-CIDP Study Group. Contactin 1 IgG4 Messina, Italy the manuscript associates to chronic inflammatory demyelinating polyneuropathy with sensory for intellectual ataxia. Brain 2015;138:1484–1491. content 7. Doppler K, Werner C, Sommer C. Disruption of nodal architecture in skin biopsies of patients with demyelinating neuropathies. J Peripher Nerv Syst 2013;18:168–176. Mario Università Cattolica Author Major role in 8. Doppler K, Appeltshauser L, Villmann C, et al. Auto-antibodies to contactin- Sabatelli del Sacro Cuore; the acquisition of associated protein 1 (Caspr) in two patients with painful inflammatory neuropathy. Centro Clinico data and revised Brain 2016;139:2617–2630. NEMO Adulti, Roma, the manuscript for 9. Kadoya M, Kaida K, Koike H, et al. IgG4 anti-neurofascin155 antibodies in chronic Italy intellectual content inflammatory demyelinating polyradiculoneuropathy: clinical significance and diagnostic utility of a conventional assay. J Neuroimmunol 2016;301:16–22.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 13 10. Devaux JJ, Miura Y, Fukami Y, et al. Neurofascin-155 IgG4 in chronic inflammatory 19. Graham RC, Hughes RA. A modified peripheral neuropathy scale: the overall neu- demyelinating polyneuropathy. Neurology 2016;86:800–807. ropathy limitations scale. J Neurol Neurosurg Psychiatry 2006;77:973–976. 11. Manso C, Querol L, Mekaouche M, Illa I, Devaux JJ. Contactin-1 IgG4 antibodies 20. Lauria G, Hsieh ST, Johansson O, et al; European Federation of Neurological Soci- cause paranode dismantling and conduction defects. Brain 2016;139:1700–1712. eties; Peripheral Nerve Society. European Federation of Neurological Societies/ 12. Labasque M, Hivert B, Nogales-Gadea G, Querol L, Illa I, Faivre-Sarrailh C. Specific Peripheral Nerve Society Guideline on the use of skin biopsy in the diagnosis of small contactin N-glycans are implicated in neurofascin binding and autoimmune targeting fiber neuropathy. Report of a joint task force of the European Federation of Neu- in peripheral neuropathies. J Biol Chem 2014;289:7907–7918. rological Societies and the Peripheral Nerve Society. Eur J Neurol 2010;17:903–912, 13. Koike H, Kadoya M, Kaida KI, et al. Paranodal dissection in chronic inflammatory e44–49. demyelinating polyneuropathy with anti-neurofascin-155 and anti-contactin-1 anti- 21. Manso C, Querol L, Lleixà C, et al. Anti-neurofascin-155 IgG4 antibodies prevent bodies. J Neurol Neurosurg Psychiatry 2017;88:465–473. paranodal complex formation in vivo. J Clin Invest 2019;130:2222–2236. 14. Vallat JM, Yuki N, Sekiguchi K, et al. Paranodal lesions in chronic inflammatory 22. Taieb G, Le Quintrec M, Pialot A, et al. “Neuro-renal syndrome” related to anti- demyelinating polyneuropathy associated with anti-neurofascin 155 antibodies. contactin-1 antibodies. Muscle Nerve 2019;59:E19–E21. Neuromuscul Disord 2017;27:290–293. 23. Hashimoto Y, Ogata H, Yamasaki R, et al. Chronic inflammatory demyelinating 15. Ng JK, Malotka J, Kawakami N, et al. Neurofascin as a target for autoantibodies in polyneuropathy with concurrent membranous Nephropathy: an anti-paranode peripheral neuropathies. Neurology 2012;79:2241–2248. and podocyte protein antibody study and literature Survey. Front Neurol 2018;9: 16. Delmont E, Manso C, Querol L, et al. Autoantibodies to nodal isoforms of neurofascin 997. in chronic inflammatory demyelinating polyneuropathy. Brain 2017;140:1851–1858. 24. Appeltshauser L, Weishaupt A, Sommer C, Doppler K. Complement deposition 17. Joint Task Force of the EFNS and the PNS. European Federation of Neurological Societies/ induced by binding of anti-contactin-1 auto-antibodies is modified by immunoglo- Peripheral Nerve Society guideline on management of chronic inflammatory demyelinating bulins. Exp Neurol 2017;287:84–90. polyradiculoneuropathy: report of a joint task force of the European Federation of Neurological 25. Uncini A, Vallat JM. Autoimmune nodo-paranodopathies of peripheral nerve: the Societies and the Peripheral Nerve Society—first revision. J Peripher Nerv Syst 2010;15:1–9. concept is gaining ground. J Neurol Neurosurg Psychiatry 2018;89:627–635. 18. Cortese A, Devaux JJ, Zardini E, et al. Neurofascin-155 as a putative antigen in 26. Querol L, Rojas-Garc´ıa R, Diaz-Manera J, et al. Rituximab in treatment-resistant CIDP combined central and peripheral demyelination. Neurol Neuroimmunol Neuro- with antibodies against paranodal proteins. Neurol Neuroimmunol Neuroinflamm inflamm 2016;3:e238. doi: 10.1212/NXI.0000000000000238. 2015;2:e149. doi: 10.1212/NXI.0000000000000149.

14 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN ARTICLE OPEN ACCESS Neuromyelitis optica spectrum disorders with unevenly clustered attack occurrence

Tetsuya Akaishi, MD, PhD, Ichiro Nakashima, MD, PhD, Toshiyuki Takahashi, MD, PhD, Michiaki Abe, MD, PhD, Correspondence Tadashi Ishii, MD, PhD, and Masashi Aoki, MD, PhD Dr. Akaishi [email protected] Neurol Neuroimmunol Neuroinﬂamm 2020;7:e640. doi:10.1212/NXI.0000000000000640 Abstract Objective The aim of this study was to elucidate the characteristics of clinical attacks in neuromyelitis optica spectrum disorders (NMOSDs) with positive serum anti-aquaporin-4 antibody. Both the timing and sequential pattern of clinical types were analyzed.

Methods A total of 69 patients with NMOSD were enrolled in this study, all of whom were treated at a single university hospital. All data regarding the clinical attacks (including types and date) together with other clinical information were collected.

Results Analysis of clinical attacks from the enrolled patients showed that there were 2 distributional patterns of attack occurrence in each patient: (1) “clustered” occurrences, which occurred within 12 months from the previous attack, and (2) “nonclustered” intermittent occurrences, which occurred ≥12 months after the previous attack. These occurrences were regardless of the duration from the onset. During the “clustered” period, clinical attacks were more likely to show a similar clinical manifestation, such as optic neuritis or myelitis. After entering the “nonclustered” intermittent period, the relapses were of random clinical type, regardless of the previous clinical manifestation.

Conclusions Patients with NMOSD showed mixed periods of “clustered” occurrence with frequent attacks presenting with similar manifestations and “nonclustered” intermittent periods with sparse relapses. Approximately half of the relapses occurred during the “clustered” period within 12 months of the last clinical attack. Clinicians should pay special attention to whether the patients are presently in the “clustered” or “nonclustered” period to decide optimal relapse-preventive strategies.

From the Department of Neurology (T.A., T.T., M.A.), Tohoku University School of Medicine, Sendai, Japan; Department of Education and Support for Regional Medicine (T.A., M.A., T.I.), Tohoku University Hospital, Sendai, Japan; Department of Neurology (I.N.), Tohoku Medical and Pharmaceutical University, Sendai, Japan; and Department of Neurology (T.T.), National Hospital Organization Yonezawa National Hospital, Yonezawa, Japan.

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

Copyright © 2019 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary AQP4 = aquaporin-4; ARR = annualized relapse rate; NMOSD = neuromyelitis optica spectrum disorder; ON = optic neuritis.

Neuromyelitis optica spectrum disorder (NMOSD) is an au- additionally investigated the possible effect of uneven relapse toimmune disorder characterized by the presence of serum timing on the observed sequential pattern of clinical symptoms. anti-aquaporin-4 (AQP4) autoantibody, recurrent optic neuritis (ON), and myelitis in most cases.1,2 Unlike MS, neurologic For the analysis of the sequential pattern of clinical attacks, disability in patients with NMOSD typically accumulates with data from 50 of the 69 enrolled patients, who experienced each clinical episode.3,4 Thus, suppressing relapses is crucial in clinical relapse at least once by June 2019, were used. Clinical the management of NMOSD, and elucidating the character- attacks with ON and myelitis were counted, but those with istics and pathomechanisms of the relapses is essential.5 medullary lesions (i.e., hiccups/nausea/vomiting) or cerebral However, at present, little is known about the relapse timing lesions were excluded from the sequential analyses of clinical and sequential patterns of clinical attacks in patients with manifestations for accurate interpretation; the details of the NMOSD. As a result, it is challenging to evaluate the patients’ excluded data and effect of excluding them on the obtained level of clinical severity and disease activity in clinical trials or results are described in subsequent sections. when considering therapeutic options. For the analysis of the irregularity of relapse timing, data from In this study, we aimed to elucidate the chronological char- both the 50 patients with relapses and 19 patients without acteristics of clinical relapses in NMOSD by applying time relapses were used. Data on the elapsed time period (months) series analyses for each patient. We also evaluated the possible from the last clinical attack were collected from the 50 patients association between relapse timing and the sequential pattern with relapses, whereas those on the relapse-free period of clinical attacks. (months) from the last attack to June 2019 were collected from all 69 enrolled patients. Moreover, these data were separately evaluated based on the administration of oral Methods relapse-preventive therapies. All 69 enrolled patients were eventually treated with low-dose oral corticosteroid, with or Patient enrollment and variables evaluated without other maintenance treatments (11 patients with A total of 69 consecutive patients with serum anti-AQP4 azathioprine, 2 with mycophenolate mofetil, and 1 with antibody-positive NMOSD, who were treated in our univer- tacrolimus). Data of relapse timing during medication-free sity hospital and followed up for more than 10 years until June periods were collected from 36 patients who were not ad- fi 2019, were enrolled. All patients were con rmed to be serum ministered relapse-preventive therapies from the onset, positive for the anti-AQP4 antibody using a cell-based assay, whereas data during treatment with oral corticosteroid were which is described below. collected from all 69 enrolled patients.

The following demographic and clinical variables were col- Relapse frequency with no medication by the lected from the 69 enrolled patients: sex, onset age, disease time from clinical onset duration, data pertaining to the clinical attacks (types and date First, we estimated the mean annualized relapse rate (ARR), of each attack), and timing of the implementation of the which was calculated as the total number of relapses divided relapse-preventive therapy. by the total person-time of follow-up, in the following 2 time periods: 0–24 months from clinical onset (36 patients) and Serum anti-AQP4 antibody testing the following 24–48 months from onset (21 patients). We conducted a microscopic live cell-based assay for AQP4 6–8 based on previous reports. In brief, human M23-AQP4- The relapse-free period without relapse-preventive therapies expressing HEK293 cells were incubated with 1:16 diluted was then compared between the first 24 months from clinical serum samples and then stained with Alexa 488-conjugated onset (36 patients) and the next 24 months in patients secondary antibody. The positivity of the antibody was without relapses in the first 24 months (14 patients). The detected based on fluorescence emission using fluorescence purpose of this analysis was to compare the probability of microscopy. experiencing relapses in the near future between patients who have a history of relapses during the previous 24 months and Evaluated outcomes patients without relapses during the previous 24 months. The 2 major outcomes evaluated in this study were (1) the sequential pattern of the types of manifested clinical attacks Statistical analysis (i.e., ON, myelitis, medullary, and cerebral) and (2) time series For the comparisons of frequency using 2 × 2 tables, a χ2 test or analysis of relapse timings in each patient. After we confirmed Fisher exact test was implemented based on the sample size of the presence of irregularity in the relapse timing, we each cell. Comparisons of the frequency with k × 2 tables (k > 2)

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN were performed using a Cochran-Armitage trend test. The elapsed time periods from onset and from the last clinical attack Figure 1 The number of patients with ≥2 attacks based on for each clinical attack during the medication-free period were the rate of attacks with optic neuritis (ON) compared using the Mann-Whitney U test because both were nonnormally distributed. The relapse-free period without medication between 0–24 months and 24–48 months from onset was compared using Kaplan-Meier survival analysis. Statistical analyses were conducted using either SPSS Statistics Base 22 software (IBM, Armonk, NY) or MATLAB R2015a (Math- Works, Natick, MA).

Institutional Review Board This study was approved by the Institutional Review Board of Tohoku University Hospital (IRB No. 2010589). Written informed consent was obtained from all enrolled patients.

Data availability Qualified researchers may obtain deidentified clinical data used in the current study from the corresponding author on ON = optic neuritis reasonable requests, such as revalidation of the procedures and results. 43 patients with 3 or more clinical attacks are presented in figure 2. The presented clinical course in each patient suggests Result an unevenness in the density of relapse occurrence, irre- Clinical information from the enrolled patients spective of the duration from onset. Of the 69 enrolled patients, 66 (95.7%) were women and the remaining 3 (4.3%) were men. The mean and SD of the onset Based on this result, we evaluated the elapsed time from the last age was 42.7 ± 14.1 years. The months in which all clinical clinical attack (figure 3A) for each clinical episode of the enrolled episodes occurred (i.e., 342 attacks in total) showed no uneven patients. The relapse-free period from the last clinical attack in seasonal distribution. The mean ± SD disease duration in 2019 June 2019 is also shown in the histogram. As shown in these was 19.7 ± 10.7 years. Thirty-six of the 69 patients (52.2%) figures, clinical relapses in NMOSD were significantly likely to were not given long-term relapse-preventive treatments after densely cluster with a time period between attacks of <12 their first clinical attack (onset). As for the clinical manifesta- months. tion of the 342 clinical attacks, 110 were ON (32.2%), 203 were myelitis (59.4%), 19 were medullary lesions (5.6%), and 10 To exclude the possibility of the effect of treatments, we also were cerebral lesions (2.9%). The rate of myelitis based on age depicted the histograms of the time period between 2 tandem at the time of the attack was 46.3% for 11–30-year-olds, 52.3% episodes based on the presence of oral relapse-preventive ther- for 31–40-year-olds, 67.0% for 41–50-year-olds, and 62.9% for apy (i.e., low-dose oral corticosteroid in this study). Both the 51–80-year-olds, showing an increased frequency of myelitis time period between attacks without treatment (figure 3B) and with age (p = 0.0186, Cochran-Armitage trend test). with relapse-preventive treatment (figure 3C) reproduced a significantly skewed deviation with a peak at an intervening time Histogram of the ratio of optic neuritis period of <12 months. These histograms suggest that approxi- episodes in the enrolled patients mately half of the relapses in patients with NMOSD take place in If the enrolled patients show a bimodal distribution for the ratio of the “clustered” period within 12 months of the last attack, irre- ON episodes, such a distribution could affect the following anal- spective of the administration of relapse-preventive therapy. yses of the sequential pattern and unevenness of relapse timings. To exclude this possibility, we present a histogram of the ON ratio Relapse frequency with no medication based calculated in each of the 43 patients with 3 or more clinical attacks, on the time from onset including onset (figure 1). As shown in figure 1, the distribution of Next, the likeliness of experiencing relapses based on the time the ON ratios showed a monotonic decrease without bimodality. from the last clinical attack (i.e., 0–24 months vs 24–48 In other words, there was no peak of patients who exclusively months) among patients without relapse-preventive therapy presented with ON episodes without myelitis episodes. after the onset (n = 36) was evaluated. The estimated mean ARR in the “0–24 months from onset” period was 0.47 (95% Time series analysis for the timing of CI: 0.35–0.60; n = 36) and that in the “24–48 months from clinical attacks onset” period was 0.25 (95% CI: 0.12–0.38; n = 21), which Next, we evaluated the timing and density of clinical attacks. was significantly higher in the former group (p = 0.0259). The occurrence timing and clinical types of each attack in the Kaplan-Meier survival analysis to compare the relapse-free

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 3 Figure 2 Clinical course and relapses in each of the patients with ≥3 attacks (n = 43)

The presented clinical courses imply a clustered occurrence of attacks, irrespective of the duration from clinical onset.

period between the 2 time periods was performed. For this For reference, episodes of ≥2 attacks with medullary lesions analysis, data of the “0–24 months from onset” period was were seen in 2 patients (one patient with 5 attacks and an- obtained from all 36 patients and data of the “24–48 months other with 2 attacks), one of whom experienced tandemly from onset” period was obtained from the 14 patients who did repeated attacks with medullary lesions 5 times. Episodes of not experience a relapse during the first 24 months after the ≥2 attacks with cerebral lesions were seen in 1 patient, who onset. The obtained Kaplan-Meier curves (figure 4) revealed exclusively experienced attacks only with cerebral lesions 4 that the “0–24 months from onset” period showed a signifi- times during her entire clinical course. The abovementioned cantly shorter relapse-free period from onset with a lower conclusions about the sequential pattern of attack types in achievement rate of a relapse-free condition by the end of the NMOSD did not change even when we provisionally in- follow-up period, compared with the “24–48 months from corporated the episodes with brain lesions in the sequential onset” period (p = 0.0206, generalized Wilcoxon test). analyses after supposing them as attacks of ON or myelitis.

Sequential pattern of attack types by the time Discussion from the last attack Next, we studied the sequential pattern of attacks in each In this report, we demonstrated that patients with NMOSD patient. The sequential pattern of clinical relapse in the 50 showed a mixed clinical course of “clustered” periods (repeated patients with relapses is summarized in the upper half of table. relapses at <12 months from the last attack) and “nonclustered” For all 3 types of analyses by the number of sequential attacks, intermittent periods (repeated relapses at ≥12 months from the the latest preceding attack significantly affected the following last attack). Time series analysis of the relapse type in each clinical relapse (p = 0.0058 for 2 tandem episodes, p = 0.0015 patient showed that a similar clinical type of episode was likely for 3 tandem episodes, and p < 0.0001 for 4 tandem episodes). to be repeated during the “clustered” period, but not after entering the “nonclustered” intermittent period. In other Because patients with NMOSD showed a clustered pattern of words, if a patient with NMOSD does not experience relapses attack, irrespective of the duration from onset, we added time for more than a specific period of time, the patient can be series analyses for the 2 sequential clinical episodes after di- regarded as being in a “nonclustered” intermittent period, and viding the episodes into those during the “clustered” period the timing or clinical type of the following relapse cannot be and those during the “nonclustered” intermittent period. predicted by the preceding clinical attacks. These data from the sequentially repeated 2 clinical relapses for each time period between attacks are summarized in the The results of this study suggest that clinicians should pay spe- lower half of table. As presented, the last attack type was likely cific attention to whether their patients are presently in the to be repeated during the “clustered” period (p = 0.0203 for “clustered” period, with a high likelihood of relapse frequency, or analysis within a 6-month period, and p = 0.0006 for analysis in the “nonclustered” intermittent period, with a low likelihood within a 12-month period). However, the following type of of relapse frequency, because this could affect the appropriate clinical episode occurred completely randomly and was not type of relapse-preventive therapy in each patient at different affected by the last attack after the “nonclustered” intermittent time points. This likelihood of relapse appears to be independent period (p = 0.78 for analysis after a 12-month period, and p = of the duration from the onset or the administration of oral 1.00 for analysis after a 24-month period). immune suppressants (i.e., low-dose oral corticosteroid in this

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN Figure 3 Tendency of clustered attacks in NMOSD, irrespective of treatments

(A) Histograms of the months between 2 tandem attacks (black bars) and the months of relapse-free period by June 2019 (white bars) in all 69 enrolled patients. The center of the gray-colored diamond is the average and its width is the standard error (SE). (B) Histogram of the months between 2 tandem attacks during the medication-free period. (C) Histogram of the months between 2 tandem attacks during the treatment period with oral relapse-preventive therapy. NMOSD = neuromyelitis optica spectrum disorder.

study) and instead dependent solely on whether the patient is in A recent report also evaluated the usefulness of previous the “clustered” or “nonclustered” period. attacks for predicting the future clinical course.12 Although the report did not conclude on the importance of selecting Patients should be aggressively treated with adequate patients with a recent high disease activity for clinical trials, amounts of relapse-preventive therapies during the “clus- the data suggested that the relapse rate in the following 2 tered” period.9,10 Empirically, appropriate relapse-preventive years may be affected by the number of attacks in the past 2 treatments may be effective for shifting patients from the years. Considered together with the results of this study that “clustered” to the “nonclustered” period. Meanwhile, whether the relapse rates are significantly different between during the aggressive treatments should be continued during the the “clustered” and “nonclustered” intermittent periods, al- “nonclustered” intermittent period is not conclusive at pres- locating the patients without taking the present disease ac- ent.11 To elucidate the best therapeutic strategy during the tivity into consideration in clinical trials may cause biases in nonclustered intermittent period, further clinical studies will the achieved results. In some relatively old retrospective be needed. Nevertheless, a minimal level of relapse-preventive studies that assessed the effectiveness of relapse-preventive therapy should be continued even during the “nonclustered” therapies, such as rituximab, azathioprine, or mycophenolate intermittent period. mofetil, the number of attacks in the past 2 years was not

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 5 Figure 4 Kaplan-Meier survival analysis comparing relapse-free periods without medications between 0 and 24 months from onset and later

The analysis compared the relapse- free period without relapse-preventive therapies between 0 and 24 months from the onset (36 patients with no relapse-preventive therapy) and 24–48 months from the onset (14 patients without relapses in the first 24 months). The latter group represents the “nonclustered” intermittent period status. The former group showed a shorter relapse-free period and a higher relapse rate than the latter group (p = 0.0206, generalized Wilcoxon test).

fully taken into consideration at the participant enrollment For reference, in the clinical trials of the abovementioned 3 stage.13,14 As a result, the estimated effectiveness of the monoclonal antibodies (PREVENT study for eculizumab, studied relapse-preventive therapies may have been weak- SAkuraSky study for satralizumab, and N-MOmentum study ened because of the inclusion of some patients in the for inebilizumab), the number of clinical attacks in the past 2 “nonclustered” intermittent period. Meanwhile, in- years before the enrollment has been fully taken into con- tentionally including some patients during their “non- sideration in all of them, minimizing the risk of biases based clustered” intermittent period during enrollment for on whether participants were in the “clustered” or “non- a clinical trial might have some benefits because an impor- clustered” period during these clinical trials.15,20 tant question is whether patients in the “nonclustered” period without a recent history of attacks may benefit similarly Based on this study, the cutoff period for relapse-free status in from such additional relapse-preventive therapies. Subgroup such clinical studies would be somewhere between 12 months analysis after dividing the enrolled patients into “clustered” and 24 months. Because the peak relapse-free period from the and “nonclustered” groups at enrollment may answer this last clinical attack was <12 months, as shown in figure 3, a 12- question. month relapse-free status may be better than 12 months per se to consider switching the aggressiveness of relapse-preventive In the near future, several monoclonal antibody drugs are therapies. Meanwhile, approximately 20% of the patients had expected to be released in the market, such as eculizumab, their relapses between 12 and 24 months after the last attacks. – satralizumab, and inebilizumab.15 17 These monoclonal This fact may imply that a 24-month relapse-free status could antibodies have shown quite a high relapse-preventive effi- be safer than 12 months because that may cover significantly ciency but are very expensive to be continued throughout more relapses at up to 70% of all relapses. life.4,18,19 The results of this study may indicate that the usage of such expensive monoclonal antibodies could be limited to The analyses of this study were based on a cohort that was the “clustered” period and can be spared and switched to low- entirely of Asian ancestry. To determine whether the mixed dose oral corticosteroid or other oral immune suppressants time course of “clustered” vs “nonclustered” periods observed once the patients move into the “nonclustered” intermittent in this study is generalizable to Caucasian or African American period without relapses for more than 12 months. If such patients with NMOSD, whose clinical severity and disease a therapeutic standard is established with evidence, it will activity are higher than those of Asian patients, further clinical surely contribute to reducing the cost of medical expenses. studies will be helpful.21,22

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN In conclusion, patients with NMOSD showed a mixed Table Sequential pattern of clinical attacks in 50 patients clinical course of the “clustered” period with dense relap- with NMOSD with relapses ses and “nonclustered” intermittent period with sparse Next: Next: relapses. Approximately half of the relapses in these ON MY p Value patients occurred during the “clustered” period within 12

Previous clinical episode months of the last clinical attack. Clinicians should pay (n = 242 sequences) special attention to whether their patients are presently in

ON 34 49 p =0.0058(χ2) the relapse-dense period or intermittent period because this could inﬂuence the treatment. Furthermore, the ne- MY 38 121 cessity of relapse-preventive therapies during the in- Previous 2 clinical episodes termittent period should be established in future clinical (n = 202 sequences) studies. ON: 2 times 15 13 p = 0.0015 (Cochran- Study funding Armitage) No targeted funding. ON: 1 time, MY: 1 time 21 49

MY: 2 times 21 83 Disclosure T. Akaishi reports no disclosures. I. Nakashima received Previous 3 clinical episodes (n = 169 sequences) speaker honoraria and travel funding from Mitsubishi Tanabe Pharma, Biogen Japan, and Novartis Pharmaceut- ON: 3 times 12 2 p < 0.0001 (Cochran- icals and received research support from LSI Medience Armitage) Corporation. T. Takahashi, M. Abe, T. Ishii, and M. Aoki

ON: 2 times, MY: 1 time 11 18 report no disclosures. Go to Neurology.org/NN for full disclosures. ON: 1 time, MY: 2 times 13 43

MY: 3 times 13 57 Publication history Neurology: Neuroimmunology & Neuroinﬂammation Clustered occurrence period (<6 Received by July months from the last episode) 17, 2019. Accepted in ﬁnal form September 5, 2019.

Previous: ON 95p = 0.0203 (Fisher exact test) Appendix Authors Previous: MY 10 30 Name Location Role Contribution Clustered occurrence period (<12 months from the last episode) Tetsuya Tohoku University, Author Designed and Akaishi, MD, Sendai, Japan conceptualized the 2 Previous: ON 18 18 p =0.0006(χ PhD study, analyzed the test) data, and drafted the manuscript for Previous: MY 15 65 intellectual content

Nonclustered period (≥12 Ichiro Tohoku Medical Author Major role in the months from the last episode) Nakashima, and acquisition of data and MD, PhD Pharmaceutical revised the manuscript 2 Previous: ON 15 34 p =0.78(χ test) University, Sendai, for intellectual content Japan Previous: MY 27 55 Toshiyuki Tohoku University, Author Major role in the Nonclustered period (≥24 Takahashi, Sendai, Japan acquisition of data and months from the last episode) MD, PhD revised the manuscript for intellectual content Previous: ON 920p = 1.00 (Fisher exact test) Michiaki Tohoku University, Author Interpreted the data Abe, MD, Sendai, Japan and revised the Previous: MY 17 38 PhD manuscript for intellectual content Abbreviations: MY = myelitis; NMOSD = neuromyelitis optica spectrum disorder; ON = optic neuritis. Tadashi Tohoku University, Author Interpreted the data Upper = the clinical episode sequence was evaluated in each of the 50 Ishii, MD, Sendai, Japan and revised the patients with relapses. Clinical episodes with medullary (i.e., hiccups/nau- PhD manuscript for sea/vomiting) or cerebral lesions were not counted here. Clinical manifes- intellectual content tations of relapses tended to repeat and were significantly influenced by the previous relapse. Lower = sequential patterns of clinical relapses in the Masashi Tohoku University, Author Supervised the clustered occurrence period and in the nonclustered intermittent period Aoki, MD, Sendai, Japan research process and were evaluated. Attacks with similar clinical manifestation were likely to PhD revised the manuscript sequentially repeat during the clustered occurrence period, but a similar for intellectual tendency was not observed in the nonclustered intermittent period. content

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 7 References 12. Palace J, Lin DY, Zeng D, et al. Outcome prediction models in AQP4-IgG positive – 1. Wingerchuk DM, Banwell B, Bennett JL, et al. International consensus diagnostic neuromyelitis optica spectrum disorders. Brain 2019;142:1310 1323. criteria for neuromyelitis optica spectrum disorders. Neurology 2015;85:177–189. 13. Mealy MA, Wingerchuk DM, Palace J, Greenberg BM, Levy M. Comparison of 2. Wingerchuk DM, Lennon VA, Pittock SJ, Lucchinetti CF, Weinshenker BG. Revised relapse and treatment failure rates among patients with neuromyelitis optica: multi- ffi – diagnostic criteria for neuromyelitis optica. Neurology 2006;66:1485–1489. center study of treatment e cacy. JAMA Neurol 2014;71:324 330. 3. Pittock SJ, Lucchinetti CF. Neuromyelitis optica and the evolving spectrum of au- 14. Jacob A, Matiello M, Weinshenker BG, et al. Treatment of neuromyelitis optica with toimmune aquaporin-4 channelopathies: a decade later. Ann N Y Acad Sci 2016;1366: mycophenolate mofetil: retrospective analysis of 24 patients. Arch Neurol 2009;66: – 20–39. 1128 1133. 4. Trebst C, Jarius S, Berthele A, et al. Update on the diagnosis and treatment of 15. Pittock SJ, Berthele A, Fujihara K, et al. Eculizumab in aquaporin-4-positive neuro- neuromyelitis optica: recommendations of the Neuromyelitis Optica Study Group myelitis optica spectrum disorder. N Engl J Med 2019;381:614–625. (NEMOS). J Neurol 2014;261:1–16. 16. Borisow N, Mori M, Kuwabara S, Scheel M, Paul F. Diagnosis and treatment of NMO 5. Kleiter I, Gold R. Present and future therapies in neuromyelitis optica spectrum spectrum disorder and MOG-encephalomyelitis. Front Neurol 2018;9:888. disorders. Neurotherapeutics 2016;13:70–83. 17. Agius MA, Klodowska-Duda G, Maciejowski M, et al. Safety and tolerability of inebili- 6. Takahashi T, Fujihara K, Nakashima I, et al. Establishment of a new sensitive assay for zumab (MEDI-551), an anti-CD19 monoclonal antibody, in patients with relapsing anti-human aquaporin-4 antibody in neuromyelitis optica. Tohoku J Exp Med 2006; forms of multiple sclerosis: results from a phase 1 randomised, placebo-controlled, 210:307–313. escalating intravenous and subcutaneous dose study. Mult Scler 2019;25:235–245. 7. Takahashi T, Fujihara K, Nakashima I, et al. Anti-aquaporin-4 antibody is involved in 18. Akaishi T, Nakashima I. Efficiency of antibody therapy in demyelinating diseases. Int the pathogenesis of NMO: a study on antibody titre. Brain 2007;130:1235–1243. Immunol 2017;29:327–335. 8. Sato DK, Nakashima I, Takahashi T, et al. Aquaporin-4 antibody-positive cases be- 19. Lin J, Xue B, Li X, Xia J. Monoclonal antibody therapy for neuromyelitis optica yond current diagnostic criteria for NMO spectrum disorders. Neurology 2013;80: spectrum disorder: current and future. Int J Neurosci 2017;127:735–744. 2210–2216. 20. Cree BA, Bennett JL, Sheehan M, et al. Placebo-controlled study in neuromyelitis 9. Sherman E, Han MH. Acute and chronic management of neuromyelitis optica optica-Ethical and design considerations. Mult Scler 2016;22:862–872. spectrum disorder. Curr Treat Options Neurol 2015;17:48. 21. Mealy MA, Kessler RA, Rimler Z, et al. Mortality in neuromyelitis optica is strongly 10. Kessler RA, Mealy MA, Levy M. Treatment of neuromyelitis optica spectrum dis- associated with African ancestry. Neurol Neuroimmunol Neuroinflamm 2018;5:e468. order: acute, preventive, and symptomatic. Curr Treat Options Neurol 2016;18:2. doi: 10.1212/NXI.0000000000000468. 11. Kimbrough DJ, Fujihara K, Jacob A, et al. Treatment of neuromyelitis optica: Review 22. Kim SH, Mealy MA, Levy M, et al. Racial differences in neuromyelitis optica spectrum and recommendations. Mult Scler Relat Disord 2012;1:180–187. disorder. Neurology 2018;91:e2089–e2099.

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN ARTICLE OPEN ACCESS Neurologic complications of acute hepatitis E virus infection

Paolo Ripellino, MD, Emanuela Pasi, MSc, Giorgia Melli, PhD, MD, Claudio Staedler, MD, Monserrat Fraga, MD, Correspondence Darius Moradpour, MD, Roland Sahli, MSc, PhD, Vincent Aubert, MSc, PhD, Gladys Martinetti, MSc, Dr. Ripellino [email protected] Florian Bihl, MD, Enos Bernasconi, MD, Benedetta Terziroli Beretta-Piccoli, MD, Andreas Cerny, MD, Harry Roland Dalton, MD, PhD, Cinzia Zehnder, MSc, Barbara Mathis, MSc, Chiara Zecca, MD, Giulio Disanto, MD, PhD, Alain Kaelin-Lang, MD, PhD, and Claudio Gobbi, MD

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e643. doi:10.1212/NXI.0000000000000643

Abstract MORE ONLINE Videos Objective To assess the prevalence and clinical features of neurologic involvement in patients with acute hepatitis E virus (HEV) infection in Southern Switzerland.

Methods Among 1,940 consecutive patients investigated for acute hepatitis E, we identiﬁed 141 cases of acute of HEV infection (anti-HEV immunoglobulin M and immunoglobulin G both reactive and/or HEV RNA positive) between June 2014 and September 2017. Neurologic cases were followed up for 6 months. We compared patients with and without neurologic symptoms.

Results Neurologic symptoms occurred in 43 acute HEV cases (30.4%) and consisted of neuralgic amyotrophy (NA, n = 15, 10.6%) and myalgia (n = 28, 19.8%). All NA cases were immunocompetent. Men had higher odds (OR = 5.2, CI 1.12–24.0, p = 0.03) of developing NA after infection with HEV, and in 3 couples simultaneously infected with HEV, only men developed NA. Bilateral involvement of NA was predominant (2:1) and occurred only in men. Seven NA cases were viremic (all genotype 3), but HEV was undetectable in their CSF. In the acute phase of NA, 9 patients were treated with intravenous immunoglobulin and 4 with prednisone, reporting no side eﬀects and improvement in pain and strength. Myalgia occurred both without (n = 16) or with (n = 12) concomitant elevated serum creatinine kinase. Seven cases with myalgia in the shoulder girdle did not have muscle weakness (“forme fruste” of NA).

Conclusions Neurologic symptoms occurred in one-third of acute HEV infections and consisted of NA and myalgia. NA seems to occur more frequently in men infected by HEV and has a predominant (but not exclusive) bilateral involvement.

From the Department of Neurology (P.R., G.M., C.S., C.Z., G.D., A.K.-L., C.G.), Neurocenter of Southern Switzerland, Lugano, CH; Laboratory of Microbiology EOLAB (E.P., G.M.), Bellinzona, CH; Faculty of Biomedical Sciences, USI (G.M., C.Z., A.K.-L., C.G.), Lugano, CH; Division of Gastroenterology and Hepatology, Lausanne University Hospital (M.F., D.M.), Lausanne, CH; Institute of Microbiology, Lausanne University Hospital (R.S.), Lausanne, CH; Laboratory of Immunology, Lausanne University Hospital (V.A.), CH; Department of Hepatology, Hospital of Bellinzona (F.B.), CH; Division of Infectious Diseases (E.B.), Hospital of Lugano, CH; Epatocentro Ticino (B.T.B.-P., A.C.), Lugano, CH; Queens Park (H.R.D.), London, UK; Synlab Ticino (C.Z.), Bioggio, CH; and Unilabs Ticino (B.M.), Lugano, CH.

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

Copyright © 2019 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary ALT = alanine aminotransferase; ANA = antinuclear antibodies; CK = creatinine kinase; GBS = Guillain-Barr´e syndrome; HEV = hepatitis E virus; IFN = interferon; IVIg = intravenous immunoglobulin; NA = neuralgic amyotrophy.

Hepatitis E virus (HEV) is one of the most common causes of 2. To compare characteristics of patients with acute HEV acute hepatitis in the world,1 and its incidence is increasing.2 infection with and without neurologic involvement and In developed countries, infection is mainly due to ingestion of clinically characterize these neurologic manifestations contaminated raw meat of pork or game.1 In Switzerland, HEV seroprevalence is estimated to be approximately 20% in Inclusion criteria healthy blood donors, depending on age and geographical The diagnostic criteria for acute HEV infection were de- area, being the highest (35%) in Southern Switzerland.3 tection of 9,10

Most cases of HEV infection in humans are asymptomatic,1 1. serum anti-HEV immunoglobulin M and anti-HEV whereas some lead to acute hepatitis and only a minority to immunoglobulin G both reactive and/or extrahepatic manifestations.4,5 Intriguingly, the most com- 2. HEV RNA+ documented in serum and/or feces by mon extrahepatic features of HEV infection are neurologic reverse transcription polymerase chain reaction – diseases.6 8 These mainly include acute dysimmune neuropathies such as neuralgic amyotrophy (NA)9 and Guillain- Subjects with false-positive serologic results (low HEV IgM Barré syndrome (GBS),10 Bell palsy,11 myalgia/myositis,12 positivity, not confirmed by a repeated serologic test after 1 encephalitis,13 transverse myelitis,14 and, more recently, month) or with any other concomitant acute infections or myasthenia gravis.15 vaccinations and patients with any other causes of acute hepatitis were excluded. Recent cohort and case-control studies (mainly, but not exclusively from Europe) showed that 5%–11% of GBS Clinical data – cases10,16 18 and 10% of patients with NA9 were infected by Demographic (age and sex) and clinical data (immunosup- HEV at disease onset. Recent HEV infection was found in pression status, alanine aminotransferase [ALT], and bilirubin 2.4% of consecutive nontraumatic neurologic patients.19 On levels) were collected. All patients were screened for neurologic the other hand, a variable proportion of confirmed cases of symptoms and/or signs by the treating physician and referred acute HEV infection (between 5% and 22%, depending on the for neurologic evaluation at the Neurocenter of Southern geographical area and study type) have been reported to de- Switzerland, Lugano, on suspicion of neurologic involvement. – velop neurologic symptoms.20 22 Such patients underwent a full neurologic examination and had a series of investigations to exclude concomitant infections (hepatitis A virus, hepatitis B virus, hepatitis C virus, human Objectives immunodeficiency virus, cytomegalovirus, Epstein-Barr virus, Borrelia burgdorferi,andTreponema pallidum) and additional The aim of our study was to further explore the relationship tests as clinically appropriate (antinuclear antibodies [ANA], between HEV infection and neurologic complications in antidouble strand DNA antibodies, extractable nuclear antigen fi a well-de ned geographical area. antibodies, antineutrophil cytoplasmic antibodies screening, antiganglioside antibodies tests, lumbar puncture, brachial plexus, brain and spinal MRI, and electrodiagnostic studies) Methods and were all followed up for at least 6 months thereafter. Subjects fi 23 We prospectively studied consecutive adult patients di- NA was de ned by van Eijk et al. and GBS according to 24 fi agnosed with acute HEV infection between June 1, 2014, NINDS criteria. Myalgia was de ned as pain in the muscles and September 30, 2017. The study was conducted in present at rest, neither preceded by trauma or physical exer- Ticino, a region of Southern Switzerland with a population cise nor triggered by medication, with a minimal intensity of of approximately 350,000 inhabitants. During the study 5/10 on the visual analogue scale, independent of the creat- period, all laboratories in Ticino centralized the testing of inine kinase (CK) level. samples for HEV to the Laboratory of Microbiology EOLAB, Bellinzona. Virologic analyses Sera from patients with suspected HEV infection were tested Aims for HEV IgG and IgM with Dia.Pro kits (Diagnostic Biop- 1. To assess the prevalence of neurologic manifestations robes srl, Italy). Real-time quantitative HEV RNA reverse occurring in subjects with acute HEV infection living in transcriptase polymerase chain reaction and HEV genotype Ticino determination were performed according to Fraga et al.25

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN Statistical analysis with NA (n = 15, 10.6%) or myalgia (n = 28, 19.8%). Some of Continuous variables were described by median with inter- these cases have been reported in previous studies with dif- quartile ranges and categorical variables by counts and per- ferent aims.25,26 centages. Age, sex, immunosuppression status, and liver function parameters were tested for potential association with Age, sex, immunosuppression status, ALT, and bilirubin levels neurologic disease status using logistic regression with OR of the overall cohort at onset are summarized in table 1. The and 95% CI estimation. Statistical signiﬁcance was set at median age was 59 years, with a nonsigniﬁcant predominance 2-sided p values < 0.05. Data were analyzed with the R sta- of men (59% vs 41%). Compared with patients without tistics package (r-project.org/). neurologic manifestations, those with neurologic involvement were younger (OR = 0.97, CI 0.94–0.99, p = 0.01). NA oc- Ethical aspects curred more frequently in men (OR = 5.2, CI 1.12–24.0, p = The study was approved by the local Ethics Committee of 0.03), whereas myalgia occurred less frequently in men (OR = Ticino (CE 2932), and written informed consent was 0.36, CI 0.15–0.86, p = 0.02). We did not identify any dif- obtained from the participants in the study. Authorization has ferences in immunosuppression status, ALT, or serum bili- been obtained for disclosure of any recognizable person in rubin levels between patients with and without neurologic videos. symptoms.

Data availability Within the group with neurologic complications, we observed Anonymized data that are not published in the current study 3 couples (husband and wife) simultaneously infected by fi will be shared on request from any quali ed investigator. HEV, probably from a common food source. Although all husbands developed bilateral NA, their wives were asymptomatic (n = 2) or had transient myalgia (n = 1). Results Among 1,940 patients investigated for acute HEV infection, Neuralgic amyotrophy fi 150 cases were identified (figure 1). Nine were excluded be- NA cases appeared to cluster in rural areas ( gure 2) and not cause of false-positive serologic results (n = 5), concomitant in the city of Lugano, the most populated area (70,000 CMV infection (n = 3; one of these cases had GBS), or recent inhabitants). vaccination (n = 1; this patient had encephalomyelitis). The “ final analysis included 141 HEV IgM- and IgG-positive cases; Shoulder pain was often described as like a knife penetrating ” 35 (25%) were also HEV RNA positive, and all were HEV muscles and was severe enough to prompt patients to attend genotype 3 when sequencing was successful. the emergency ward even in the middle of the night. Neu- rologic involvement in HEV-NA was bilateral (although asymmetrical) in 10 men and predominantly located in the Reasons for HEV testing were acute hepatitis with (n = 55, 26 39%) or without (n = 34, 24%) jaundice, unexplained in- upper plexus. Scapular winging from long thoracic nerve creased liver enzyme resulting from screening of clinically damage was a recurrent feature (10/15, 67%) observed at 27 fi asymptomatic patients (n = 7, 5%), acute kidney failure (n = 2, disease onset ( gure 3A). In 3 cases, amyotrophy had a clear 1%), and neurologic symptoms (n = 43, 30.4%) consistent patchy distribution, suggesting that nerve damage occurred at the fascicular level23,28 (figure 3B).

Figure 1 Flowchart of the study design The laboratory and clinical data of the HEV-NA cases are summarized in table 2. None of the patients with HEV-NA were immunosuppressed at the time of infection. Lumbar puncture was performed in 6 of 10 NA cases with bilateral involvement. One patient with extended neurologic damage (involving the brachial plexus bilaterally, lumbosacral plexus, and trigeminal nerve) had an increased CSF cell count (55 cells/mcl) and protein level, whereas isolated albumin- cytologic dissociation was found in 2 other bilateral HEV- NA cases.

Viral HEV RNA gt-3 was detected in the serum of 7 of 8 NA cases tested within 1 week from onset. In 5 of these, concomitant CSF was collected on the same day, but none contained HEV RNA. Serum immunoﬁxation showed a monoclonal IgM peak (κ or λ) in 5 bilateral NA cases tested in the acute phase, which disappeared within a few weeks. ANA antibodies were positive at low titer (1/80–1/160) in 5 of 10 bilateral HEV-NA cases.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 3 Table 1 Study population characteristics

Neurologic cases (n = 43) Nonneurologic cases All (n = 141) NA (n = 15) Myalgia (n = 28) (n = 98)

Age (median) (IQR) 59 (48–70) 51 (45–61) 53 (43–62) 60 (49–73)

Male 83 (59%) 13 11 59

Female 58 (41%) 2 17 39

Immunosuppression 16 0 3 13

ALTa (median) (IQR) 175 (56–741) 240 (100–957) 88 (28–394) 211 (57–938)

Bilirubina (median) (IQR) 20 (13–53) 19 (14–35) 14 (11–17) 24 (14–74)

HEV RNA in serum (all gt3) 35/64 tested 7/12 7/12 24/40

Abbreviations: ALT = alanine aminotransferase (IU/L; normal < 50), bilirubin (μmol/L, normal <19); gt3 = genotype 3; HEV = hepatitis E virus; IQR = interquartile range; NA = neuralgic amyotrophy. a At disease presentation.

Serum antiganglioside antibodies (GM1, GM2, GD1a, GD1b, were normal, as previously reported by others.29 Needle elec- and GQ1b) and anti-myelin-associated glycoprotein were neg- tromyography showed in all cases in the acute phase a reduced ative in all bilateral HEV-NA cases. (neurogenic) recruitment pattern in aﬀected shoulder girdle muscles and also signs of acute denervation (ﬁbrillation Electrodiagnostic studies were performed within 1 month in all potentials and positive sharp waves) in 10 cases (66%), which HEV-NA cases. Motor and sensory nerve conduction studies disappeared within 6 months.

Figure 2 Location of Ticino in Switzerland and geographical distribution of HEV-NA cases

The residence of the 15 NA cases is approximated by blue dots. The population density in the area is shown in different colors in a red scale (legends on the left). Most of the cases were not located in urban but in rural areas (this map has been obtained from data of the Federal Office of Topography [Swisstopo]). HEV = hepatitis E virus; NA = neuralgic amyotrophy.

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN in the shoulder girdle (“forme fruste” of NA, n = 7). We Figure 3 Clinical picture of HEV-NA deﬁned “forme fruste” of patients with NA with severe, neuropathic shoulder pain (unilateral or bilateral) lasting not more than 72 hours, which was not followed by limb weakness. Three of these patients with “forme fruste” of NA had an increased CK level (up to 500 U/L), and 5 underwent electromyography studies which were normal. None of these patients received brachial plexus MRI or muscle biopsy.

Discussion The findings of our study suggest that neurologic symptoms in patients with acute HEV infection may be more common than previously estimated. Retrospective studies organized in hepatological settings described neurologic illness in 5%–7.5% of patients presenting with clinical or subclinical hepatitis.4,20,21 However, more recent findings from a prospective study performed in France22 showed that 22% of immunocompetent HEV cases had neurologic symptoms, (A) In bilateral HEV-NA cases (above), winged scapula (black arrows) is re- which is more in line with our findings. current and often asymmetrical. (B) In unilateral HEV-NA cases (below), amyotrophy is often very focal, affecting one muscle and sparing the neighboring muscle (left: involvement of the upper trapezius = green arrow, There are several possible reasons to explain why we observed deltoid spared; right: brachialis muscle affected = red arrow, biceps brachii spared). HEV = hepatitis E virus; NA = neuralgic amyotrophy. so many cases of neurologic injury associated with HEV. One possibility is that neurologic damage could be related to the neurotropism of the circulating HEV viral strains in Southern Two patients with unilateral HEV-NA did not receive any Switzerland. An alternative explanation could be that the local immunomodulatory treatment because they had minor weak- awareness about HEV infection prompted physicians to test ness. Regarding the remaining HEV-NA cases, they were for HEV cases with unexplained abnormally high serum treated within 10 days from onset either with intravenous im- aminotransferases, especially in case of pain. Finally, it is also munoglobulin (IVIg) 0.4 g/kg/d for 5 days in cases of bilateral likely that in previous studies, myalgia was not identified or involvement (n = 9) or with oral prednisone 60 mg/d × 7 days, classified as a neurologic symptom. This is especially true for 26 with tapering in 7 days in cases of unilateral weakness (n = 4), the “forme fruste” of NA, an entity that is currently over- as previously reported in a retrospective study.30 looked also by neurologists because of late (or absent) referral. All patients with HEV-NA receiving these immunomodulatory treatments reported both reduction in pain and im- None of our neurologic cases was jaundiced, and in most the provement in shoulder mobility and strength within 7–10 liver function test abnormality was modest (ALT < 1,000 IU/ days (video 1 for steroids, video 2 for IVIg). All patients L). In particular, NA appeared to be strongly associated with received supportive physiotherapy over the course of their HEV, occurring in 11% of acute HEV cases (especially in illness. Four cases received pregabalin 150 mg/d for 2–3 men) with a predominant (2:1) bilateral involvement. It is months to treat neuropathic pain. Regarding long-term re- noteworthy that all 12 bilateral NA cases diagnosed at the covery, 3 heavy manual workers with bilateral HEV-NA had to Neurocenter in Lugano in the past 3 years were men with stop their job for more than 6 months because of early fati- acute HEV infection. These findings confirm previous gability of affected arms. observations that the predominant clinical phenotype of HEV-associated NA is that of bilateral involvement of the Myalgia brachial plexus.26 Myalgia occurred in 28 cases, both with (n = 12) and without (n = 16) concomitant elevated (>200 U/L) serum CK. The It is currently unclear why after exposure to the same in- highest CK value in our patients was 4.100 U/L. The pain was fectious agent individuals develop different clinical manifes- often unbearable, was more intense in the night, and had clear tations. We identified sex as a possible risk factor because men neuropathic features characterized by “burning,”“squeezing,” with HEV had higher odds for NA. Our observation of 3 “electric shock,” and “lancinating” sensations (DN4 score couples infected simultaneously with HEV is similar to a case > 4).31 report32 more than 30 years ago, even before HEV discovery. Within an Italian family, 3 members simultaneously de- Muscle pain disappeared spontaneously after 2 or 3 weeks in veloped bilateral NA 1 week after a short fever, but the 2 all cases and was either diffuse (n = 21) or located selectively brothers developed the full picture of NA (with scapular

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 5 Table 2 Clinical features of HEV-related NA

HEV-NA Total n = 15 Bilateral n = 10 Unilateral n = 5

Age (median) (IQR) 51 (61–45) 58 49

Male 13 (86%) 10 3

Female 2 (14%) 0 2

Hospitalization 12 7 5

NRSa (median) at onset 8(9.5–8) 8 10

EMG spontaneous activity (fibrillations, positive waves) 10 (67%) 6 4 at 1 month

HEV RNA + (serum)b 77 0 All gt-3

HEV RNA+ (CSF)c 0 0 n.a.

Lumbar puncture 66 0

CSF WBC >10/mm3 1/6 1/6 n.a.

CSF protein increase 3/6 3/6 n.a.

ALTa (median) 240 (100–957) 579 178

ALPa (median) 126 (80–174) 126 128

Bilirubina (median) 19 (14–35) 52 13

GGTa (median) 144 (39–203) 178 55

ANA positivity 55 0

ENA positivity 0

Antiganglioside antibodies 0

IgM monoclonal peak in serum 55 0

IVIg 0.4 g/kg × 5d 98 1

Oral prednisone 60 mg/d × 7 d, tapering in 7 d 31 2

No immunomodulatory treatment 20 2

Abbreviations: ALKP = alkaline phosphatase; ALT = alanine aminotransferase; ANA = antinuclear antibodies; EMG = electromyography; ENA = extractable nuclear antigen; GGT = gamma glutamyl transferase; HEV = hepatitis E virus; IQR = interquartile range; IVIg = intravenous immunoglobulin; WBC = white blood count. Normal range: bilirubin < 19 μmol/L; ALT < 41 IU/L; ALKP = 40–129 IU/L; GGT < 61 IU/L; CSF analysis: WBC < 5; protein <0.40 g/L. a At disease presentation. b Tested within 7 days from disease onset. c CSF was collected the same day as the serum used for HEV RNA testing.

winging), whereas the mother suffered only from pain and When tested within 1 week from onset, 7 of 8 HEV-NA cases in had no weakness. Moreover, larger epidemiologic studies26 our cohort were viremic, as previously described.9,20,21,26 By have shown that NA occurs more commonly in men (ratio 2: contrast, HEV is much less commonly found in the CSF, and in 1), and also in a recent French study,22 5 of 6 HEV-NA cases the literature, there are only 2 reported NA cases where HEV were men. The reason(s) why HEV-associated NA is more RNA has been demonstrated in the CSF.37,38 None of the 5 common in men is unclear, but sex hormones are known to patients with NA who were viremic and received the lumbar influence innate33 and adaptive34,35 immune response during puncture on the same day of serum collection showed HEV viral infection. The immune response to virus is regulated by RNA in the CSF. Taken together, these data suggest that HEV is interferons (IFNs), and IFN treatment can induce bilateral cleared very early in the CSF and/or that neurologic injury is NA.36 mediated by mechanisms more complex than direct viral invasion, e.g., through immune-mediated mechanisms. Recent Regarding the host immune status, another factor that could studies showed that intrathecal synthesis of anti-HEV IgM can potentially influence clinical presentation, it is important to note occur in HEV-NA cases.38 Although none of our patients with that all our 15 HEV-NA cases were immunocompetent. NA or myalgia underwent a biopsy, it is tempting to speculate

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN that antibody/immunocomplexes deposition in the peripheral infection. The frequency and the clinical features of neurologic nerve and muscle tissues might contribute to neurologic damage. complications in immunosuppressed HEV-infected cases (e.g., Similar mechanisms have previously been shown to be operative transplant recipients) could be different from what we observed, in cryoglobulinemia39 and other extrahepatic complications re- yet 1 study showed that neurologic injury was less common in lated to HEV.40 Unfortunately, in our study, cryoglobulins were immunosuppressed patients.22 Finally, our observations should not tested, so this hypothesis remains speculative. be confirmed in different geographical regions. The clustered distribution and high number of NA cases observed might be According to our model of immune-mediated pathogenesis, explained by viral factors such as a special neurotropism. Of note, we treated patients with NA in the acute phase with IVIg or the viral strain most frequently isolated in Switzerland (3s) steroids. Both treatments were always discussed with the appears to be unique,41,42 but correlations between viral strains hepatologist and tailored for each patient based on safety and clinical symptoms are currently lacking, and larger collabo- reasons (mainly the viremic state) and individual comorbid- rative studies are underway in Europe (e.g., HEVnet, rivm.nl/en/ ities (e.g., diabetes). We chose IVIg for bilateral cases because hevnet) to address this issue. these patients were usually viremic (7/10), and we wanted to avoid worsening of viral replication/hepatitis by using high- In conclusion, neurologic symptoms occurred in approximately dose steroids. We instead preferred oral steroids for unilateral one-third of documented acute HEV cases and consisted of NA cases who were usually referred with a longer delay and were and myalgia. HEV-NA seems to occur more frequently in men never viremic (0/5). Treatment of patients with HEV-NA was and was bilateral in two-thirds of the cases. Larger prospective however beyond the scope of this study, and no firm con- epidemiologic studies on the potential role of HEV as a trigger clusions can be drawn from our observational study. How- of (neurologic) autoimmunity43 and randomized-controlled ever, it is worthwhile to note that all patients reported trials on the treatment of HEV-NA are warranted. improvement of neuropathic pain, shoulder girdle muscle strength, and scapular dyskinesia within 7–10 days (videos 1 Acknowledgment and 2) and that both IVIg and steroids were well tolerated. The authors are grateful to Ente Ospedaliero Cantonale, None of the viremic HEV-NA cases developed chronic HEV Neuromuscular Research Association of Basel, and Baasch- infection or experienced an increase in transaminase levels. Medicus Stiftung of Zurich for funding support; Klaus Toyka for helpful comments on the manuscript; Swiss Federal Office Regarding GBS, we excluded 2 cases of GBS who were reactive of Topography (Swisstopo) and Wabern for providing maps for both HEV IgM and IgG because they did not fulfill our and data of figure 2; and Mario Uhr (Synlab Ticino) for inclusion criteria. The first of these patients had concomitant collaborating on data collection. CMV infection, as reported by other authors,18 raising the question whether these observations were because of only Study funding crossreactivity or true co-occurrence of infections. Our second ABREOC 2016 (Ente Ospedaliero Cantonale funding) to PR case of HEV GBS was excluded only because the diagnosis was and CG; Neuromuscular Research Association Basel (NERAB) made 3 months after study termination. This was an 80-year- 2016 to PR; Baasch-Medicus Foundation 2018 to PR. old woman with an axonal form (AMSAN) of GBS. She developed acute tetraparesis, bilateral facial diplegia, and bulbar Disclosure symptoms with a need for tracheotomy and assisted ventilation All authors report no conflicts of interest to the present (GBS disability scale = 5). She received 2 cycles of intravenous project. Go to Neurology.org/NN for full disclosures. immunoglobulins and recovered completely after 6 months. Publication history Our clinical experience suggests that sometimes the clinical dif- Received by Neurology: Neuroimmunology & Neuroinflammation fi ferences between a bilateral form of NA and a pharyngeal- September 16, 2019. Accepted in nal form October 11, 2019. brachial-cervical variant of GBS can be modest, and some patients might be classified either as NA or GBS (see e.g., case 3 of ref. 18). Appendix Authors There are some limitations to our study. First is the issue of case Authors Location Role Contribution ascertainment bias because our study population represents only Paolo Neurocenter of Author Design and a limited proportion of all acute HEV-infected cases. Only Ripellino, Southern conceptualization of MD Switzerland, CH the study, major role a minority of patients with HEV develop symptoms or are in the acquisition of detected through an elevated serum aminotransferase level, and data, analysis of the data, and writing and most infections are never diagnosed. Moreover, because the revision of the referral to the Neurocenter was based on the treating physician’s manuscript decision, it cannot be excluded that neurologic cases with minor Emanuela Laboratory of Author Major role in the symptoms were missed. Second, our study population mostly Pasi, MSc Microbiology EOLAB, acquisition of data included immunocompetent patients in the acute phase of HEV CH and analysis of the data infection, and we did not focus on chronic cases of HEV Continued Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 7 Appendix (continued) Appendix (continued)

Authors Location Role Contribution Authors Location Role Contribution

Giorgia Neurocenter of Author Acquisition of data Cinzia Synlab Ticino, CH Author Major role in the Melli, MD, Southern and interpretation Zehnder, acquisition of data PhD Switzerland, CH; and revision of the MSc Faculty of biomedical manuscript for Sciences, USI, CH important intellectual Barbara Unilabs Ticino, CH Author Major role in the content Mathis, MSc acquisition of data

Claudio Neurocenter of Author Acquisition of data Chiara Neurocenter of Author Revision of the Staedler, Southern and interpretation Zecca, MD Southern manuscript for MD Switzerland, CH and revision of the Switzerland, CH important intellectual manuscript for content important intellectual content Giulio Neurocenter of Author Statistical analysis of Disanto, Southern data Monserrat Division of Author Data interpretation MD, PhD Switzerland, CH Fraga, MD Gastroenterology and revision of the and Hepatology, manuscript for Alain Neurocenter of Author Revision of the Lausanne University important intellectual Kaelin-Lang, Southern manuscript for Hospital, CH content MD, PhD Switzerland, CH; important intellectual Faculty of Biomedical content Darius Division of Author Data interpretation Sciences, USI, CH Moradpour, Gastroenterology and revision of the MD and Hepatology, manuscript for Claudio Neurocenter of Author Design and Lausanne University important intellectual Gobbi, MD Southern coordination of the Hospital, CH content Switzerland, CH; study, major role in Faculty of Biomedical the acquisition of Roland Institute of Author Major role in the Sciences, USI, CH data, and writing and Sahli, MSc, Microbiology, acquisition of data revision of the PhD Lausanne University and revision of the manuscript Hospital, CH manuscript for important intellectual content References Vincent Laboratory of Author Major role in the 1. Kamar N, Izopet J, Pavio N, et al. Hepatitis E virus infection. Nat Rev Dis Primers Aubert, Immunology, acquisition of data 2017;16:17086. MSc, PhD Lausanne University 2. Capai L, Falchi A, Charrel R. Meta-analysis of human IgG anti-HEV seroprevalence in Hospital, CH industrialized countries and a review of literature. Viruses 2019;11:E84. 3. Charrel C, Widmer N, Hotz M, et al. Current hepatitis E virus seroprevalence in Swiss Gladys Laboratory of Author Major role in the blood donors and apparent decline from 1997 to 2016. Euro Surveill 2018;23: Martinetti, Microbiology EOLAB, acquisition of data 1700616. MSc CH and analysis of the 4. Woolson KL, Forbes A, Vine L, et al. Extra-hepatic manifestations of autochthonous data hepatitis E infection. Aliment Pharmacol Ther 2014;40:1282–1291. 5. Kamar N, Marion O, Abravanel F, Izopet J, Dalton HR. Extrahepatic manifestations of Florian Bihl, Department of Author Major role in the hepatitis E virus. Liver Int 2016;36:467–472. MD Hepatology, Hospital acquisition of data 6. Dalton HR, Kamar N, van Eijk JJ, et al. Hepatitis E virus and neurological injury. Nat of Bellinzona, CH and revision of the Rev Neurol 2016;12:77–85. manuscript for 7. Mclean BN, Gulliver J, Dalton HR. Hepatitis E virus and neurological disorders. Pract important intellectual Neurol 2017;17:282–288. content 8. Dalton P, Norton B, van Eijk J, Dalton HR. Non-traumatic neurological injury and hepatitis E infection. Expert Rev Anti Infect Ther 2018;16:255–257. Enos Division of Infectious Author Data interpretation 9. van Eijk JJ, Madden RG, van der Eijk AA, et al. Neuralgic amyotrophy and hepatitis E Bernasconi, Diseases, Hospital of and revision of the virus infection. Neurology 2014;82:498–503. MD Lugano, CH manuscript for 10. van den Berg B, van der Eijk AA, Pas SD, et al. Guillain-Barré syndrome associated important intellectual with preceding hepatitis E virus infection. Neurology 2014;82:491–497. content 11. Jha AK, Nijhawan S, Nepalia S, Suchismita A. Association of Bell’s palsy with hepatitis E virus infection: a rare entity. J Clin Exp Hepatol 2012;2:88–90. Benedetta Epatocentro Ticino, Author Major role in the 12. Mengel AM, Stenzel W, Meisel A, Büning C. Hepatitis E-induced severe myositis. Terziroli- CH acquisition of data Muscle Nerve 2016;53:317–320. Beretta and revision of the 13. Kejariwal D, Roy S, Sarkar N. Seizure associated with acute hepatitis E. Neurology Piccoli, MD manuscript for 2001;57:1935. important intellectual 14. Sarkar P, Morgan C, Ijaz S. Transverse myelitis caused by hepatitis E: previously content undescribed in adults. BMJ Case Rep 2015;2015:bcr2014209031. 15. Wang L, Gao F, Lin G, et al. Association of hepatitis E virus infection and myasthenia Andreas Epatocentro Ticino, Author Data interpretation gravis: a pilot study. J Hepatol 2018;68:1318–1320. Cerny, MD CH and revision of the 16. Geurtsvankessel CH, Islam Z, Mohammad QD, Jacobs BC, Endtz HP, Osterhaus AD. manuscript for Hepatitis E and Guillain-Barré syndrome. Clin Infect Dis 2013;57:1369–1370. important intellectual 17. Fukae J, Tsugawa J, Ouma S, Umezu T, Kusunoki S, Tsuboi Y. Guillain–Barré and content Miller Fisher syndromes in patients with anti-hepatitis E virus antibody: a hospital- based survey in Japan. Neurol Sci 2016;37:1849–1851. Harry Queens Park, Author Data interpretation 18. Stevens O, Claeys KG, Poesen K, Saegeman V, Van Damme P. Diagnostic challenges Dalton, MD, London, UK and revision of the and clinical characteristics of hepatitis E virus-associated Guillain-Barré syndrome. PhD manuscript for JAMA Neurol 2017;1:26–33. important intellectual 19. Dalton HR, van Eijk JJJ, Cintas P, et al. Hepatitis E virus infection and acute non- content traumatic neurological injury: a prospective multicentre study. J Hepatol 2017;67: 925–932.

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN 20. Kamar N, Bendall RP, Peron JM, et al. Hepatitis E virus and neurologic disorders. 32. Martinelli P, Pazzaglia P, Marchiori L, Lugaresi E. Simultaneous occurrence of neu- Emerg Infect Dis 2011;17:173–179. ralgic amyotrophy in three members of one family. Eur Neurol 1980;19:316–319. 21. Perrin HB, Cintas P, Abravanel F, et al. Neurologic disorders in immunocompetent 33. Kovats S. Estrogen receptors regulate innate immune cells and signaling pathways. patients with autochthonous acute hepatitis E. Emerg Infect Dis 2015;21:1928–1934. Cell Immunol 2015;294:63–69. 22. Abravanel F, Pique J, Couturier E, et al. Acute hepatitis E in French patients and 34. Park HJ, Choi JM. Sex-specific regulation of immune responses by PPARs. Exp Mol neurological manifestations. J Infect 2018;77:220–226. Med 2017;49:e364. 23. Van Eijk JJJ, Groothuis JT, Van Alfen N, et al. Neuralgic amyotrophy: an update on 35. Nussinovitch U, Shoenfeld Y. The role of gender and organ specific autoimmunity. diagnosis, pathophysiology, and treatment. Muscle Nerve 2016;53:337–350. Autoimmun Rev 2012;11:377–385. 24. Asbury AK, Cornblath DR. Assessment of current diagnostic criteria for Guillain- 36. Bernsen PL, Wong Chung RE, Vingerhoets HM, Janssen JT. Bilateral neuralgic Barré syndrome. Ann Neurol 1990;27:S21–S24. amyotrophy induced by interferon treatment. Arch Neurol 1988;45:449–451. 25. Fraga M, Doerig C, Moulin H, et al. Hepatitis E virus as a cause of acute hepatitis 37. Silva M, Wicki B, Tsouni P, et al. Hepatitis E virus infection as a direct cause of acquired in Switzerland. Liver Int 2018;38:619–626. neuralgic amyotrophy. Muscle Nerve 2016;54:325–327. 26. Van Eijk JJJ, Dalton HR, Ripellino P, et al. Clinical phenotype and outcome of 38. Fritz M, Berger B, Schemmerer M, et al. Pathological cerebrospinal fluid findings in hepatitis e virus-associated neuralgic amyotrophy. Neurology 2017;89:909–917. patients with neuralgic amyotrophy and acute hepatitis E virus infection. J Infect Dis 27. van Alfen N, van Engelen BG. The clinical spectrum of neuralgic amyotrophy in 246 2018;217:1897–1901. cases. Brain 2006;129:438–450. 39. Ramos-Casals M, Stone JH, Cid MC, Bosch X. The cryoglobulinaemias. Lancet 2012; 28. Watson BV, Rose-Innes A, Engstrom JW, Brown JD. Isolated brachialis wasting: an 379:348–360. unusual presentation of neuralgic amyotrophy. Muscle Nerve 2001;24:1699–1702. 40. Marion O, Abravanel F, Del Bello A, et al. Hepatitis E virus-associated cry- 29. van Alfen N, Huisman WJ, Overeem S, van Engelen BG, Zwarts MJ. Sensory nerve oglobulinemia in solid-organ-transplant recipients. Liver Int 2018;38:2178–2189. conduction studies in neuralgic amyotrophy. Am J Phys Med Rehabil 2009;88: 41. Kubacki J, Fraefel C, Jermini M, et al. Complete genome sequences of two Swiss 941–946. hepatitis E virus isolates from human stool and raw pork sausage. Genome Announc 30. van Eijk JJJ, van Alfen N, Berrevoets M, van der Wilt GJ, Pillen S, van Engelen BGM. 2017;5:e00888–17. Evaluation of prednisolone treatment in the acute phase of neuralgic amyotrophy: an 42. Sahli R, Fraga M, Semela D, Moradpour D, Gouttenoire J. Rabbit HEV in immu- observational study. J Neurol Neurosurg Psychiatry 2009;80:1120–1124. nosuppressed patients with hepatitis E acquired in Switzerland. J Hepatol 2019;70: 31. Bouhassira D, Attal N, Alchaar H, et al. Comparison of pain syndromes associated 1023–1025. with nervous or somatic lesions and development of a new neuropathic pain di- 43. Terziroli Beretta-Piccoli B, Ripellino P, Gobbi C, et al. Autoimmune liver disease agnostic questionnaire (DN4). Pain 2005;114:29–36. serology in acute hepatitis E virus infection. J Autoimmun 2018;94:1–6.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 9 ARTICLE OPEN ACCESS SHP2 inhibitor protects AChRs from eﬀects of myasthenia gravis MuSK antibody

Saif Huda, MD, DPhil,* Michelangelo Cao, MD, PhD,* Anna De Rosa, MD, Mark Woodhall, PhD, Correspondence Pedro M. Rodriguez Cruz, MD, DPhil, Judith Cossins, DPhil, Michelangelo Maestri, MD, Roberta Ricciardi, MD, Dr. Vincent [email protected] Amelia Evoli, MD, David Beeson, PhD, and Angela Vincent, MD, FRCPath, FRS

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e645. doi:10.1212/NXI.0000000000000645

Abstract RELATED ARTICLE Objective Editorial fi To determine whether an SRC homology 2 domain-containing phosphotyrosine phosphatase 2 Treating muscle-speci c (SHP2) inhibitor would increase muscle-specific kinase (MuSK) phosphorylation and override kinase myasthenia gravis the inhibitory effect of MuSK-antibodies (Abs). from the inside out Page e646 Methods The effect of the SHP2 inhibitor NSC-87877 on MuSK phosphorylation and AChR clustering was tested in C2C12 myotubes with 31 MuSK-myasthenia gravis (MG) sera and purified MuSK-MG IgG4 preparations.

Results In the absence of MuSK-MG Abs, NSC-87877 increased MuSK phosphorylation and the number of AChR clusters in C2C12 myotubes in vitro and in DOK7-overexpressing C2C12 myotubes that form spontaneous AChR clusters. In the presence of MuSK-MG sera, the AChR clusters were reduced, as expected, but NSC-87877 was able to protect or restore the clusters. Two puriﬁed MuSK-MG IgG4 preparations inhibited both MuSK phosphorylation and AChR cluster formation, and in both, clusters were restored with NSC-87877.

Conclusions Stimulating the agrin-LRP4-MuSK-DOK7 AChR clustering pathway with NSC-87877, or other drugs, could represent a novel therapeutic approach for MuSK-MG and could potentially improve other NMJ disorders with reduced AChR numbers or disrupted NMJs.

*Joint first authors—these authors contributed equally to the manuscript.

From the Department of Clinical Neurosciences (S.H., M.C., M.W., P.M.R.C., J.C., D.B., A.V.), Weatherall Institute of Molecular Medicine and Nuffield, University of Oxford, UK; Department of Clinical and Experimental Medicine (A.D.R., M.M., R.R.), Neurology Unit, Pisa; and Department of Neuroscience (A.E.), Catholic University, Rome, Italy.

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 Medical Research Council/Oxford University. 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 © 2019 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary AChR = acetylcholine receptor; CBA = cell-based assay; DOK7 = downstream of kinase 7; FACS = fluorescence-activated cell sorting; KO = knockout; LRP4 = low-density lipoprotein receptor-related protein 4; MG = myasthenia gravis; MuSK = muscle- specific kinase; NMJ = neuromuscular junction; SHP2 = Src homology 2 domain-containing phosphotyrosine phosphatase 2; SK = specific kinase.

Myasthenia gravis (MG) is an autoimmune disease of the this mechanism, increasing MuSK phosphorylation could neuromuscular junction (NMJ). In a variable percentage of represent a potential therapeutic strategy for the development patients, muscle-specific kinase antibodies (MuSK-Abs) are of specific treatments. present.1 In humans, symptoms occur particularly in cranial, bulbar, and respiratory muscles with frequent respiratory Among several regulators of the AChR clustering pathway, the crises.2 Although immunomodulatory treatments can be SRC (Rous sarcoma gene) homology 2 domain-containing beneficial in the longer term, there is an unmet need for cheap, phosphotyrosine phosphatase 2 (SHP2) is a phosphatase that fast, and effective treatments. reduces MuSK phosphorylation. Importantly, for the purposes of this study, NSC-87877, an SHP2 inhibitor, has been shown MuSK is normally activated following binding of agrin, secreted to enhance agrin-induced and agrin-independent AChR clus- from the motor nerve terminal, to low-density lipoprotein tering in vitro.7 We tested NSC-87877 for its ability to increase receptor-related protein 4 (LRP4). LRP4 then binds to MuSK MuSK phosphorylation and to reverse or prevent the effects of leading to its dimerization and to auto- and trans- MuSK-Abs in 2 in vitro models. phosphorylation in the MuSK juxtamembrane region.3 The subsequent recruitment of intracellular downstream of kinase 7 (DOK7) further stimulates MuSK phosphorylation, causing Methods activation of a phosphorylation cascade that ultimately leads to clusters of acetylcholine receptor (AChR) anchored by in- Materials tracellular rapsyn on the postsynaptic membrane of the NMJ.4 Serum samples were collected, with informed consent, from 31 patients with typical MuSK-MG symptoms and immu- MuSK-Abs are predominantly of the IgG4 subclass and in- notherapy responses. MuSK-Ab titers of patients’ sera and hibit the interaction between LRP4 and MuSK, preventing purified immunoglobulin G (IgG) fractions were determined MuSK phosphorylation and AChR clustering.5,6 Considering by radioimmunoassay and cell-based assay (CBA).8,9 Sera

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN were heated, dialyzed, and sterile filtered before use. IgG control sera or 0.5 nM MuSK-Ab of each purified preparation fractions were purified from plasmapheresis of 2 additional (IgG or IgG4) as illustrated in the figures. patients with MuSK-MG using protein G sepharose and an IgG4 affinity matrix.5,8 Effectiveness of IgG subclass purifica- MuSK phosphorylation tion was tested with CBA. Briefly, MuSK-transfected human MuSK phosphorylation was assessed by Western blotting on embryonic kidney cells were incubated with the different IgG immunoprecipitates from the C2C12 myotubes. They were subclasses (1:20) and probed with anti-human-IgG1, -IgG2, starved of fetal calf or horse serum for 3 hours, incubated in -IgG3, and -IgG4 monoclonal mouse antibodies (1:50) (I2513, medium alone or medium plus NSC-87877 (1–1,000 μM; I5635, I7260, and I7385, Sigma). After fixing with 3% formal- usually 100 μM) for 40 minutes at 37°C, lysed in cold lysis dehyde, cells were stained with Alexa Fluor 488 goat anti- buffer (10 mM Tris-HCl, 1 mM EDTA, 100 mM, NaCl, 1% mouse IgG (1:200) (A32723, Invitrogen) and images captured Triton X-100, 1× protease [P8340, Sigma] and phosphatase using the Olympus IX71 fluorescence microscope with Simple [78420, Thermo Fisher] inhibitor cocktails), and centrifuged. PCI (Digital Pixel). The SHP2 inhibitor NSC-87877 (#2613) The supernatant was immunoprecipitated with anti-MuSK- was obtained from Tocris Bioscience, Bristol, United Kingdom. Ab (AF562, Bio-Techne, Minneapolis, MN) using Protein G Dynabeads (10004D, Thermo Fisher) or directly with C2C12 myotube cultures and AChR patients’ MuSK-Abs for experiments with purified IgG and cluster analysis IgG4. Immunoprecipitated proteins were eluted into sodium C2C12 mouse myoblasts (91031101, ATCC) were main- dodecyl sulphate sample buffer, and the Western blots probed tained and differentiated into myotubes after 5–6daysindif- with anti-phosphotyrosine antibody 1:1,000 (4G10, Upstate ferentiation medium as previously reported (Dulbecco’s Biotechnology, MA). The nitrocellulose was then stripped modified Eagle medium [DMEM] with 2% fetal calf serum/ and reprobed with AF562. Band densitometry was performed horse serum).9 MuSK-MG sera with a broad range of MuSK- using ImageJ software. MuSK phosphorylation levels were Abs (nM) were chosen according to availability. MuSK-MG normalized to levels of immunoprecipitated MuSK. sera (1:10, 1:30, and 1:90) or purified MuSK-Ab subclasses (0.5 nM) were applied to myotubes for 30 minutes and then Statistical analysis incubated overnight with agrin 1:800 (short rat form, pro- Statistical tests were performed with GraphPad Prism 6 as fi ducing approximately 50% of maximum AChR clusters) in the indicated in the gure legends. Error bars represent the ff presence and absence of NSC-87877 100 μM. AChR clusters standard error of the mean unless otherwise stated. Di er- p were then labeled with Alexa Fluor 594 α-bungarotoxin (1: ences with values <0.05 were considered statistically fi 1,000) (B13422, Invitrogen) and fixed in 3% formaldehyde. signi cant. Twenty fields selected with bright field were analyzed for 8–11 Standard protocol approvals, registrations, number and cluster length (>3 μm) using ImageJ software. and patient consents The Musk “knockout” (KO) C2C12s were generated with the The study of archived patient samples was approved by the clustered regularly interspaced short palindromic repeats Oxfordshire Research Ethics Committee A (07 Q160X/28). 12 fl (CRISPR)-Cas9 system using standard methods. Brie y, Data availability Musk guide oligonucleotides (Integrated DNA Technologies) The authors confirm that all the data supporting the findings were designed and cloned into a modified plasmid pX335-U6- of this study will be made available on reasonable request to Chimeric_BB-CBh-hSpCas9n-D10A (42335, Addgene). the corresponding author. Guide A oligonucleotide sequences were A-forward: 59- CACCGCATTCTCCCGGATGCTGTAG-39 and A-reverse: 59-AAACCTACAGCATCCGGGAGAATGC-39.Guide Results B oligonucleotide sequences were B-forward: 59- CACCGCTCCTCACCATTCTGAGCG-39 and B-reverse: SHP2 inhibition increased MuSK 59-AAACCGCTCAGAATGGTGAGGAGC-39. Myoblasts phosphorylation and induced agrin- were electroporated with 10 μg of each plasmid using the independent AChR clusters Neon Transfection System (Life Technologies), selected with First, we studied the effect of SHP2 inhibition on MuSK Geneticin Antibiotic (10131035, Life Technologies), and phosphorylation and on AChR clustering in the absence of cloned in 96-well plates using fluorescence-activated cell agrin. We exposed C2C12 myotubes to increasing concen- sorting. Clones were screened by Sanger sequencing trations of NSC-87877 (1–400 μM) for 40 minutes. MuSK after PCR of genomic DNA using the primers 59- was immunoprecipitated by commercial anti-MuSK AF562. TGGTGCTTTGGTTATGGAGCC-39 and 59-GAG- Phosphorylation was identified in the immunoprecipitates GAGGGGTCTAAGGCTTG-39. Musk KO generation was by Western blotting with an anti-tyrosine phosphorylation confirmed by Western blot. antibody and MuSK expression by incubating the stripped membranes with anti-MuSK AF562. AChR clusters were DOK7-overexpressing myoblasts were prepared as previously measured on parallel myotube cultures by labeling with described.10,11 The myotubes were exposed to MuSK-Ab or fluorescent alpha-bungarotoxin after 12 hours. The ratio of

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 3 phosphorylated MuSK to total MuSK increased substantially expressed as % of those in the presence of healthy control (each normalized to that in DMEM alone) with maximal effect serum dilutions with agrin but without the SHP2 inhibitor. at 100 μM; at higher concentrations, phosphorylation de- MuSK-MG sera reduced the number and size of AChR creased (figure 1A). C2C12 myotubes, in the absence of agrin, clusters with most effect at 1:10 dilution as expected and, at express few spontaneous clusters (<10 per field; figure 1B each serum concentration, NSC-87877 (100 μM) increased DMEM), but after incubation with agrin, the numbers in- the number and size of AChR clusters (figure 3, A.a and A.b). creased up to 40/field. As previously shown,7 NSC-87877 at There was a modest correlation between the number of 100 μM, without agrin, increased the numbers of AChR clus- clusters at 1:30 serum dilution and MuSK-Ab radio- ters similar to those achieved with agrin alone (figure 1B upper immunoprecipitation titers, with and without NSC-87877 panel). Again, the optimal concentration was 100 μM, which (figure 3B), with similar results or trends at 1:10 and 1:90 was used for all experiments thereafter. As expected, neither (data not shown). MuSK nor phosphorylated MuSK was detectable in the Musk KO myotubes (figure 1B, lower panel). Mean results of dif- When SHP2 inhibitor and agrin were added 6 hours after the ferent experiments are shown in figures 1, C and D. MuSK-Abs to see whether the inhibitor could recluster the AChRs following dispersal, the numbers of clusters were again To ensure that 40 minutes was appropriate for the experi- increased at each serum dilution (figure 3C). ments, a time course was performed. This showed that MuSK phosphorylation increased over time and was clearly On the DOK7-overexpressing myotubes, where the clusters detectable at 40 and 90 minutes for agrin, NSC-87877, or are well established, MuSK-MG sera reduced the number of both NSC-87877 and agrin. MuSK phosphorylation was AChR clusters but to a lesser extent than the same sera on modestly increased when both were given together com- agrin-induced AChR clusters. At 1:30 serum dilutions, for pared with NSC-87877 alone (p <0.05at409, p <0.01at instance, the number of agrin-induced clusters was 27.9% ± 909)(figure 1E), and the submaximal 40 minutes was chosen 9.3% (n = 10) of control, whereas on DOK7-overexpressing for future experiments looking at changes in MuSK myotubes, the number of clusters was 69.6 ± 10.9%, n = 10; p phosphorylation. = 0.01) of control-treated myotubes. Nevertheless, NSC- 87877 increased the number of clusters (figure 3D) at 1:10 C2C12 myotubes overexpressing DOK7 develop stable and 1:30 dilutions. At 1:90 dilution, the sera did not signifi- AChR clusters in the absence of agrin.11 MuSK phosphory- cantly reduce the number of clusters, and NSC-87877 had no lation was clearly present in the DOK7-overexpressing cells effect. without added agrin and was further increased by NSC-87877 (figure 2A). Moreover, the addition of NSC-87877 to the SHP2 inhibition increases AChR clusters in DOK7-overexpressing myotubes increased the number of IgG4-treated C2C12 myotubes AChR clusters that mirrored the increased phosphorylation Finally, to demonstrate that the effects were due (at least (figure 2B). mainly) to IgG4 MuSK-Abs, purified MuSK-Ab IgGs and IgG4 subfractions were prepared from plasmapheresis Patients with MuSK-MG material of 2 additional patients with typical MuSK-MG. To investigate whether SHP2 inhibition was able to protect For myotubes not exposed to patient antibodies, MuSK was against the effects of MuSK-Abs, 31 MuSK-MG sera (female: immunoprecipitated by the anti-MuSK AF562 antibody male, 9:1; median age at onset 42 years [range 18–70 years]) from the cell lysates (to avoid MuSK phosphorylation with were studied. All patients except 3 were on immunotherapies, concomitant use of the commercial antibody). For myo- mainly prednisolone and/or azathioprine, but all had pre- tubes exposed to patient antibodies, to ensure that only dominantly bulbar disease (as indicated by b grades), and 8 phosphorylation of the antibody-bound MuSK was still had severe involvement (Myasthenia Gravis Foundation assessed, MuSK was immunoprecipitated by the patient of America 3b or 4b, see table). MuSK-Ab levels measured by IgG. To adjust for this difference in the precipitation radioimmunoprecipitation were typically variable technique, the phosphorylation signals were always nor- (0.25–24.26 nM). CBAs showed that, as expected, IgG4 malized to the MuSK protein signal. An example immu- MuSK-Abs were present and predominated over IgG1, 2, or 3 noblot is shown in figure 4, A.a and results of 3–6 (data not shown). experiments in figure 4, A.b. Both total MuSK-MG IgG and IgG4 fractions reduced MuSK phosphorylation, and SHP2 inhibition alleviates the effect of phosphorylation was enhanced by NSC-87877 (figure 4, MuSK-Abs on agrin-induced and A.a and A.b). agrin-independent AChR clusters Three dilutions (1:10, 1:30, and 1:90) of 21 MuSK-Abs and Figure 4, B.a and B.b show the effects on AChR clusters. 2 control sera were incubated with the myotubes for 30 Again, both total IgG and IgG4 preparations substantially minutes before adding agrin with or without NSC-87877; 12 reduced agrin-induced AChR clustering, as previously hours later, the numbers of AChR clusters were measured reported.8 Moreover, in each case, NSC-87877 protected the (see time line in figure 3A). In each case, the results were myotube clusters from the inhibitory effects.

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN Figure 1 SHP2 inhibition by NSC-87877 induces MuSK phosphorylation and AChR clustering

(A) Dose-response curve of the effects of SHP2 inhibition on AChR phosphorylation. Phosphorylation increased progressively reaching a maximal effect at 100 μM and decreasing at higher concentrations. (B) Clustering of AChRs (labeled with Alexa Fluor 594 α-bungarotoxin) in C2C12 myotubes after 12 hours incubation with DMEM, agrin, or 100 μM NSC-87877. Both agrin and NSC-87877 increased AChR clusters in native C2C12 myotubes but not in MuSK knockout cells. The optimal NSC-87877 concentration was 100 μM, and cluster numbers decreased at higher concentrations. (C and D) The pooled results of 3–6 independent experiments measuring MuSK phosphorylation (C) and AChR clusters (D) are shown. (E) Time course of MuSK phosphorylation in C2C12 myotubes after exposure to agrin (1:800), NSC-87877 (100 μM), and agrin/NSC-87877. In each case, a similar time course was observed with maximum effect between 40 and 90 minutes. Phosphorylation and MuSK blots shown for NSC-87877 alone. All images are 20× magnification. Scale bar represents 50 μm. Mean + standard error of the mean are shown. Results were analyzed by 2-sided t tests. AChR = acetylcholine receptor; MuSK = muscle-specific kinase; SHP2 = SRC homology 2 domain-containing phosphotyrosine phosphatase 2.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 5 Figure 2 Effect of SHP2 inhibition in DOK7-overexpressing C2C12s

(A) MuSK phosphorylation was already high in DOK7-overexpressing C2C12s (data not shown) but was increased further by NSC-87877. (B) Equally DOK7- overexpressing myotubes showed AChR clusters that were further increased by NSC-87877. All images are 20× magnification. Scale bar represents 50 μm. Results were analyzed by 2-sided t tests. DOK7 = downstream of kinase 7; MuSK = muscle-specific kinase; SHP = SRC homology 2 domain-containing phosphotyrosine phosphatase 2.

Discussion AChR clustering could lead to tumor development as many of them, such as ErbB2 and mitogen-activated protein kinase/ MuSK-MG can be a life-threatening disease that responds extracellular-signal-regulated-kinase,13,14 play an important poorly to symptomatic drugs such as cholinesterase inhib- role in cell growth. By contrast, Src homology 2 domain- itors and often requires high doses of steroids and other containing phosphatase 2 (SHP2), encoded by the PTPN11 immunosuppressive drugs for long-term treatment. Patients gene, is a potentially safer target. Under physiologic con- may have myasthenic crises with respiratory and bulbar in- ditions, through a negative-feedback loop, SHP2 reduces volvement that often require urgent treatment in intensive MuSK phosphorylation, providing homeostatic control of care. The development of a therapy that acts specifically on AChR clusters at the NMJ. SHP2 also enhances cell replica- the mechanisms of the disease could achieve quick control of tion inducing different pathways such as Ras/Erk, PI3K, ser- symptoms and in the longer term might help reduce the ff ine/threonine kinase 1, and janus kinase 2-signal transducer potential side e ects from immunosuppressive therapies. 15–17 Thus, targeting the intracellular regulation of the AChR and activator of transcription ; thus, SHP2 inhibition clustering pathway by inhibition of the specific phosphatase, should not increase the risk of tumorigenesis. Indeed, phar- SHP2, as we show here, could represent a novel and specific macologic inhibition of SHP2 has already demonstrated effi 18 therapeutic strategy for MuSK-MG. The results also draw cacy in the treatment of a mouse glioma xenograft model. attention to the potential relevance of targeting intracellular pathways in diseases caused by extracellular antibodies or The selective inhibition of the SHP2 phosphatase that con- other mechanisms. trols MuSK phosphorylation by NSC-87877 was shown previously to increase AChR clustering in C2C12 myotubes7; There are several proteins downstream of MuSK that could be here, we confirm that this works through increasing phos- potential pharmacologic targets, including Abl tyrosine kina- phorylation, which was clearly seen irrespective of the pres- ses and guanosine triphosphatases of the Rho family. How- ence of agrin or MuSK-Abs. Although NSC-87877 is known ever, increasing the activity of these regulators to enhance to act on both SHP1 and SHP2, only SHP2 has been detected

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN Table Clinical details of patients

Patient Age at onset, sex MGFA onset, sampling Radioimmunoassay titer Immunotherapies at sampling

M081 67, F 3b, 1 4.87 Aza 100 mg, Pred 12.5 mg

M083 19, F 2b, 2b 6.88 Pred 25/0 mg

M085 30, F 3b, 3b 12.72 None

M087 61, M 4b, 1 0.63 Pred 50 mg

M089 59, F 3b, 3b 10.08 Cycl 100 mg, Pred 15 mg, IVIG

M092 28, F 5, 1 0.41 Pred 30 mg

M093 63, F 4b, 3b 12.72 Pred 12.5 mg

M097 24, F 2b, 1 12.73 Pred 40 mg

M098 70, F 4b, 2b 0.25 Pred 12.5 mg

M103 42, M 3b, 0 3.42 Pred 25 mg

M105 67, F -, 0 0.73 Pred 20 mg

M110 19, F 3b, 2b 9.25 Aza 100 mg, Pred 12.5 mg

M116 32, F 3b, 0 9.65 Pred 17.5 mg

M125 36, F 5, 2b 3.84 Pred 35 mg

M126 57, F 3b, 3b 7.97 Aza 100 mg, Pred 12.5 mg,

M128 50, F 4b, 0 17.47 Pred 12.5 mg

M129 37, M 2b, 1 4.15 Pred 15 mg

M133 23, F 3b, 0 6.54 Pred 12.5

M147 33, F 3b, 2b 1.24 Pred 25 mg

M149 36, F -, 1 6.88 Aza, Pred

M154 50, F 3b, 1 6.88 Pred 12.5 mg, IVIG

M159 28, F 4b, 4b 24.96 Pred 50 mg

M165 49, F 5, 2b 5.27 Aza 100 mg, Pred 62.5 mg

M166 29, F 2b, 0 2.75 Pred 12.5 mg

M167 46, F 2b, 3b 9.45 None

M168 49, F 2b, 0 3.66 Pred 10 mg

M173 42, F 3b, 3b 8.57 Pred 50 mg

M175 44, F 3b, 2b 4.86 Pred 17.5 mg

M176 18, F 3b, 0 4.69 Pred 37.5 mg, PLEX

M180 47, F 3b, 2b 12.72 Aza 100 mg, Pred 5 mg

M181 56, F 3b, 4b 4.37 None

Abbreviations: Aza = azathioprine; b = bulbar predominant disease; Cycl = cyclosporin; IVIG = intravenous immunoglobulins; MGFA = Myasthenia Gravis Foundation of America Scale; PLEX = plasma exchange; Pred = prednisolone. Serum dilution giving a score of 1 (range 0–4 representing increasing MuSK-Abs).9 All patients were studied at the University of Pisa. in C2C12 myotubes,7,19 and, reassuringly, conditional MuSK-Abs are mainly IgG4 subclass and are known to in- knockout of SHP2 in skeletal muscle did not compromise the hibit MuSK activation and phosphorylation by blocking the formation and maintenance of the NMJ, suggesting that interaction between LRP4 and MuSK.5,8 The inhibitory pharmacologic inhibition of SHP2 should be well tolerated in eﬀects of MuSK-Abs on AChR clustering were present to skeletal muscle,20 although other SHP2-speciﬁc inhibitors varying degrees in all sera studied, correlating broadly with should also be studied. MuSK-Ab titers. NSC-87877 improved phosphorylation and

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 7 Figure 3 SHP2 inhibition reverses effects of MuSK-Abs on AChR clusters in C2C12s and DOK7-overexpressing C2C12s

(A) The treatment protocols are shown above the results. MuSK-Ab positive (n = 21) or healthy control sera (n = 2) were incubated with myotubes for 30 minutes, followed by agrin (1:800) with or without NSC-87877 (100 μM). After 12 hours, the AChR clusters were analyzed and expressed as a percentage of control serum results. The number (A.a) of the AChR clusters without NSC-87877 was clearly reduced by the MuSK-Abs (to 14%, 35%, and 60% at dilutions 1:10, 1:30, and 1:90, red columns); the size of the clusters was similarly reduced (A.b). In each case, the effects were partially or completely reversed by NSC-87877 (green columns). (B) Inverse correlation between the number of AChR clusters at 1:30 serum dilution and MuSK-Ab titers (nM) with or without NSC-87877.(C) Using a different protocol, AChR clusters were similarly reduced, and there was partial or complete protection by NSC-87877. (D) The effects of the different serum dilutions on AChR clusters and the protection by NSC-87877 were also seen in the DOK7-overexpressing C2C12s, although to a lesser extent. In A, C, and D, comparisons at each different serum concentration were analyzed by 2-sided t tests. For B, linear regression was computed by GraphPad Prism. The scatter plots show results of individual sera, with mean ± SDs. AChR = acetylcholine receptor; MuSK = muscle-specific kinase; SHP2 = SRC homology 2 domain- containing phosphotyrosine phosphatase 2.

8 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN Figure 4 IgG and IgG4 subclass MuSK-Abs reduce AChR clusters, which are restored by NSC-87877

Agrin-stimulated myotubes were exposed for 40 minutes to MuSK IgG or IgG4 subfraction (adjusted to 0.5 nM MuSK-Ab) purified from 2 patients with MuSK- MG (the total number of experiments was IgG n = 5; IgG4 n = 6). (A.a) Example of phosphorylation blots. MuSK was immunoprecipitated by anti-MuSK AF562 or MuSK-MG IgG fractions. Pooled results (A.b) show that total IgG and IgG4 inhibited MuSK phosphorylation, and this was increased substantially by NSC-87877 (100 μM). (B.a) Examples of AChR clusters. Pooled results (B.b) show that total IgG and Ig4 MuSK-Ab fractions inhibited the formation of AChR clusters, whichin each case were increased by NSC-87877. Two-sided t tests were used for comparisons at each IgG concentration. AChR = acetylcholine receptor; MG = Myasthenia gravis; MuSK = muscle-specific kinase.

AChR clustering at each concentration of the sera tested. MuSK phosphorylation still remains the main effect of MuSK Moreover, when the phosphorylation and AChR clustering antibodies, as confirmed by our results with purified total IgG experiments were conducted comparing total IgG and the IgG4 and by the effectiveness of the inhibition of SHP2. subfraction from 2 patients, both reduced MuSK phosphorylation and AChR clustering, and SHP2 inhibition by NSC- There were some limitations to the experimental procedure. 87877 was able to reverse their pathologic effects. Of interest, For the phosphorylation experiments, the patient antibody- the inhibitory effect of IgG4 on AChR clustering does not precipitated MuSK was less than the MuSK precipitated by depend on its monovalency and divalent monoclonal anti- commercial antibody from the control cultures. There were bodies to MuSK also inhibit clustering.21,22 These recent 2 reasons for this. First, we chose a concentration of patient findings suggest that MuSK antibodies could have additional MuSK antibody (0.5 nM) that was submaximal for in- pathogenic mechanisms, other than the direct inhibition of hibition of clustering, making sure that any effects on clus- agrin/MuSK interaction, which contribute to the disruption of tering were clearly evident, but we needed to ensure that we the AChR clustering pathway. Nevertheless, the inhibition of only measured the phosphorylation in those MuSK

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 9 molecules that were bound by patient IgG. Second, the commercial antibody was applied to the cell lysate from the Appendix Authors

myotubes rather than to the live cells and thus could also Name Location Role Contribution precipitate MuSK that was not on the surface. To adjust for these diﬀerences in MuSK immunoprecipitation, the phos- Saif Huda, University Author Designed and conceptualized MD, DPhil of Oxford, the study, analyzed the data, phorylation was normalized to the MuSK protein signal for UK and drafted the manuscript each lane. Further in vitro studies examining the intracellular for intellectual content molecular mechanisms need to be performed, and the pro- Michalengelo University Author Designed and conceptualized tective eﬀects of SHP2 inhibition against MuSK antibodies Cao, MD, PhD of Oxford, the study, analyzed the data, UK and drafted the manuscript demonstrated in vivo. for intellectual content

Anna De Rosa, University Author Major role in acquisition of Most novel therapies in MG depend on monoclonal anti- MD of Pisa, data and revised the bodies targeting the immune system, but given the un- Pisa manuscript

derstanding of the mechanisms involved in the MuSK Mark University Author Helped perform CBAs for pathway, faster pharmacologic treatments should be ex- Woodhall, of Oxford, MuSK antibody subclasses plored. Recently, a monoclonal MuSK antibody was shown PhD UK to stimulate MuSK phosphorylation and to preserve NMJs Pedro M. University Author Provided MuSK knockout cells in a motor neuron disease model.23 Using new or repur- Rodriguez of Oxford, Cruz, MD, UK posed drugs to enhance MuSK phosphorylation, or other DPhil downstream components of the signaling pathway, could be Judith University Author Supervision and laboratory beneﬁcial in MuSK-Ab myasthenia and perhaps in other Cossins, DPhil of Oxford, support disorders of neuromuscular transmission where loss of UK AChRs or disruption of the NMJ structure underlies the Michelangelo University Author Major role in acquisition of muscle weakness. Maestri, MD of Pisa, data and revised the Pisa manuscript

Conflict of interest Roberta University Author Major role in acquisition of Ricciardi, MD of Pisa, data and revised the The University of Oxford hold a patent for use of MuSK for Pisa manuscript antibody assays, licensed to Athena Diagnostics. AV receives ﬂ Amelia Evoli, Catholic Author Major role in acquisition of a proportion of royalties. No other reported con icts of interest. MD University, data and revised the Rome manuscript

Acknowledgment David Beeson, University Author Supervision and laboratory The authors are grateful to Dr. Jonathan Cheung and Dr. PhD of Oxford, support Dominic Waithe (University of Oxford) for developing UK ImageJ macro used for automated counting of AChR clusters. Angela University Author Designed and conceptualized They are very grateful to the Watney Trust/Myaware/ Vincent, MD, of Oxford, the study, analyzed the data, FRCPath, FRS UK drafted the manuscript for National Institutes of Health Research Oxford Biomedical intellectual content, Research Centre (A.V., S.H.), the European Academy of supervision, and laboratory support Neurology (M.C.), and the Medical Research Council (MR/ M006824; J.C., D.B.) for their support. References Study funding 1. Hoch W, McConville J, Helms S, Newsom-Davis J, Melms A, Vincent A. Auto- The authors are very grateful to the Watney Trust/Myaware/ antibodies to the receptor tyrosine kinase MuSK in patients with myasthenia gravis without acetylcholine receptor antibodies. Nat Med 2001;7:365–368. National Institute of Health Research Oxford Biomedical 2. Evoli A, Tonali PA, Padua L, et al. Clinical correlates with anti-MuSK-Abs in gen- Research Centre (A.V., S.H.), the European Academy of eralized seronegative myasthenia gravis. Brain 2003;126:2304–2311. 3. Hopf C, Hoch W. Dimerization of the muscle-specific kinase induces tyrosine Neurology (M.C.), and the Medical Research Council (MR/ phosphorylation of acetylcholine receptors and their aggregation on the surface of M006824; J.C., D.B.) for their support. myotubes. J Biol Chem 1998;273:6467–6473. 4. Wu H, Xiong WC, Mei L. To build a synapse: signaling pathways in neuromuscular junction assembly. Development 2010;137:1017–1033. Disclosure 5. Huijbers MG, Zhang W, Klooster R, et al. MuSK IgG4 autoantibodies cause myas- The University of Oxford holds a patent for use of MuSK for thenia gravis by inhibiting binding between MuSK and Lrp4. Proc Natl Acad Sci U.S.A 2013;110:20783–20788. antibody assays, licensed to Athena Diagnostics. AV receives 6. Koneczny I, Stevens JA, De Rosa A, et al. IgG4 autoantibodies against muscle-specific a proportion of royalties. M.C., A.R., M.W., P.R.C., J.C., M.M., kinase undergo Fab-arm exchange in myasthenia gravis patients. J Autoimmun 2016; 77:104–115. R.R., A.E., and D.B. report no disclosures. Go to Neurology. 7. Zhao XT, Qian YK, Chan AWS, Madhavan R, Peng HB. Regulation of ACh receptor org/NN for full disclosures. clustering by the tyrosine phosphatase Shp2. Dev Neurobiol 2007;67:1789–1801. 8. Koneczny I, Cossins J, Waters P, Beeson D, Vincent A. MuSK myasthenia gravis IgG4 disrupts the interaction of LRP4 with MuSK but both IgG4 and IgG1-3 can disperse Publication history preformed agrin-independent AChR clusters. PLoS One 2013;8:e80695. fi Neurology: Neuroimmunology & Neuroinflammation 9. Huda S, Waters P, Woodhall M, et al. IgG-speci c cell-based assay detects potentially Received by August pathogenic MuSK-Abs in seronegative MG. Neurol Neuroimmunol Neuroinflamm 15, 2019. Accepted in final form October 8, 2019. 2017;4:e357. doi:10.1212/NXI.0000000000000357.

10 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN 10. Beeson D, Higuchi O, Palace J, et al. Dok-7 mutations underlie a neuromuscular 17. Frankson R, Yu ZH, Bai Y, et al. Therapeutic targeting of oncogenic tyrosine phos- junction synaptopathy. Science 2006;313:1975–1978. phatases. Cancer Res 2017;77:5701–5705. 11. Cossins J, Liu WW, Belaya K, et al. The spectrum of mutations that underlie the 18. Bunda S, Burrell K, Heir P, et al. Inhibition of SHP2-mediated dephosphorylation of neuromuscular junction synaptopathy in DOK7 congenital myasthenic syndrome. Ras suppresses oncogenesis. Nat Commun 2015;6:1–12. Hum Mol Genet 2012;21:3765–3775. 19. Chen L, Sunn SS, Yip ML, et al. Discovery of a novel Shp2 protein tyrosine phos- 12. Ran FA, Hsu PD, Wright J, et al. Genome engineering using the CRISPR-Cas9 phatase inhibitor. Mol Pharmacol 2006;70:562–570. system. Nat Protoc 2013;8:2281–2308. 20. Dong XP, Li XM, Gao TM, et al. Shp2 is dispensable in the formation and mainte- 13. Leah LN, Glathe S, Vaisman N, et al. The ErbB-2/HER2 oncoprotein of human nance of the neuromuscular junction. Neurosignals 2006;15:53–63. carcinomas may function solely as a shared coreceptor for multiple stroma-derived 21. Huijbers MG, Vergoossen DL, Fillié-Grijpma YE, et al. MuSK myasthenia gravis growth factors. Proc Natl Acad Sci 1999;96:4995–5000. monoclonal antibodies: valency dictates pathogenicity. Neurol Neuroimmunol 14. Steelman L, Chappell WH, Abrams SL, et al. Roles of the Raf/MEK/ERK and PI3K/ Neuroinflamm 2019;21:e547. doi:10.1212/NXI.0000000000000547. PTEN/Akt/mTOR pathways in controlling growth and sensitivity to therapy- 22. Takata K, Stathopoulos P, Cao M, et al. Characterization of pathogenic monoclonal implications for cancer and aging. Aging 2011;3:192–222. autoantibodies derived from muscle-specific kinase myasthenia gravis patients. JCI 15. Chan RJ, Feng GS. PTPN11 is the first identified proto-oncogene that encodes Insight 2019;4:e127167. a tyrosine phosphatase. Blood 2007;109:862–867. 23. Cantor S, Zhang W, Delestrée N, Remédio L, Mentis GZ, Burden SJ. Preserving 16. Matozaki T, Murata Y, Saito Y, Okazawa H, Ohnishi H. Protein tyrosine phosphatase SHP- neuromuscular synapses in ALS by stimulating MuSK with a therapeutic agonist 2: a proto-oncogene product that promotes Ras activation. Cancer Sci 2009;100:1786–1793. antibody. Elife 2018;7:e34375.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 11 CLINICAL/SCIENTIFIC NOTES OPEN ACCESS A case of idiopathic multicentric Castleman disease in an alemtuzumab-treated patient with MS

Leoni Rolfes, MD,* Steffen Pfeuffer, MD, Tobias Ruck, MD, Susanne Windhagen, MD, Ilske Oschlies, MD, Correspondence Hermann-Joseph Pavenst¨adt, MD, Linus Angenendt, MD, Heinz Wiendl, MD, Julia Kr¨amer, MD,* and Dr. Rolfes [email protected] Sven G. Meuth, MD, PhD*

Neurol Neuroimmunol Neuroinﬂamm 2020;7:e638. doi:10.1212/NXI.0000000000000638

Alemtuzumab is an eﬃcacious therapy for active relapsing-remitting MS (RRMS), but its use is complicated by the potential development of secondary autoimmunity.1 Recent data from phase 3 extension studies conﬁrm thyroid autoimmunity as the most abundant entity of secondary autoimmunity found in up to 30%–40% of treated patients, with most events mild or moderate in severity. Data also show low rates of previously known autoimmune phenomena, such as immune thrombocytopenia and nephropathy.2 However, further entities including sarcoidosis, vitiligo, and hemophagocytic lymphohistiocytosis (HLH) have been described in real-world cohorts.3 Here, we present a case of another lymphoproliferative syndrome, namely idiopathic multicentric Castleman disease (iMCD), after alemtuzumab treatment.

Case report A 30-year-old Caucasian man was ﬁrst diagnosed with RRMS in 2003. After initial interferon-β1a and natalizumab (61 infusions) treatment, he received 2 courses of alemtuzumab (December 2014 and December 2015) because of persistent disease activity. Infusions were administered without complications, and disease activity stabilized over time. The patient developed Graves’ disease 24 months after alemtuzumab initiation.

In April 2018, he presented to his local hospital with fatigue, fever, dyspnea, and thoracic and abdominal pain. Clinical examination revealed generalized lymphadenopathy, hepatosplenomegaly, and edema. Sonography indicated pleural and pericardial eﬀusion and ascites suggestive of polyserositis. He developed pancytopenia and showed elevated serum C-reactive protein (CRP, ﬁgure, A and B) and hypoalbuminemia (2.24 g/dL). A mesenteric lymph node (LN) biopsy was performed and was deemed inconclusive.

He was referred to our clinic. Repeated serologic testing for autoimmune, paraneoplastic, and infectious causes remained inconspicuous, especially regarding antinuclear antibodies and herpesvirus (including human herpesvirus (HHV)-8 and HIV PCR in peripheral blood). Whole-body 18F-FDG PET/CT, bronchoscopy, and gastro-/colonoscopy provided no further explanations. Persistent disseminated lymphadenopathy, polyserositis, and organomegaly were detected through CT (figure, F–H). An axillary LN biopsy was performed 3 weeks after the initial presentation, but changes were nonspecific despite extensive immunohistochemistry and clonality analysis. Bone marrow puncture revealed increased megakaryocytosis (figure, I).

*These authors contributed equally to the manuscript.

From the Clinic of Neurology with Institute of Translational Neurology (L.R., S.P., T.R., H.W., J.K., S.G.M.), University Hospital Munster;¨ Department of Neurology and Neurological Rehabilitation (S.W.), Hospital Osnabruck,¨ Am Finkenhugel¨ 1; Department of Pathology (I.O.), Hematopathology Section and Lymph Node Registry, University-Hospital Schleswig- Holstein, Campus Kiel; Department of General Internal Medicine (H.-J.P.), Nephrology, Hypertensiology, and Rheumatology, University Hospital Munster;¨ and Department of He- matology (L.A.), Hemostaseology, Oncology, and Pneumology, University Hospital Munster,¨ Germany.

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

Copyright © 2019 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Figure Diagnostic work-up and treatment regimes in an alemtuzumab-treated patient with RRMS developing idiopathic multicentric Castleman disease

(A–E) Charts show the course of CRP, hemoglobin, platelets, white blood cell count, and blood urea nitrogen since the initial presentation during hospitalization. Time points and length of different treatment regimens are outlined in (A) (♦: IV methylprednisolone 250 mg, cumulative dose 2,750 mg; •: immunoadsorption; *: IVIG single dose, 1 mg/kg). Arrows represent platelet (B) and erythrocyte concentrate (C) transfusions and hemodialysis (E). (F) Abdominal CT scan outlining extensive abdominal (red arrows) and inguinal (yellow arrows) lymphadenopathy. (G–H) Axial thoracic and abdominal CT scan indicating polyserositis with pleural (G: yellow arrows) and pericardial (G: red arrows) effusions and ascites (F: yellow arrows) as well as hepatosplenomegaly (F: red arrows). (I) H&E staining of bone marrow puncture displaying megakaryocytosis. (J–K) Mediastinal lymph node biopsy consistent with iMCD plasma cell type. (J) Haematoxylin and eosin stain displaying regressive germinal centers (blue arrows), with small vessels reaching into germinal centers (“lollipop vessels,” black arrows). (K) Giemsa staining displaying interfollicular proliferation of plasma cells (black arrows).

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN After immunosuppressive therapy with high-dose IV meth- risk management plan, especially beyond 48 months after the ylprednisolone (IVMPS, figure, A♦) and immunoadsorption last alemtuzumab infusion.5 Furthermore, differentiation of (IA, figure, A•), inflammatory activity decreased and blood HLH and iMCD is crucial because the latter can be treated with cell counts rose again. Pleural and ascites punctures had to be the monoclonal anti-IL6 antibody siltuximab.5 performed less frequently. In conclusion, the current case highlights iMCD as another However, 28 days after symptom onset, the patient rapidly rare but potentially life-threatening secondary autoimmune redeteriorated with increasing serum CRP-levels and re- event after alemtuzumab treatment. Until risk management emerging cytopenia, especially thrombocytopenia refractory plans have been adjusted to include these new disorders, to IV immunoglobulins (figure, A and B*). Because of in- clinicians should perform a thorough clinical follow-up (e.g., trathoracic bleeding, he required multiple erythrocyte and screening of B symptoms, examination of LNs) and, in sus- platelet transfusions and an emergency thoracotomy. picious cases, seek early interdisciplinary support. Thereafter, the patient was referred to the intensive care unit because of renal failure (predominantly caused by systemic Study funding inflammation (procalcitonin 18 ng/mL) and third spacing) This study was financially supported by the Competence and underwent continuous hemodialysis under which blood Network Multiple Sclerosis (01GI1603D, PROGRAMMS to cell counts, effusions, edema, and retention parameters T. Ruck, H. Wiendl and S.G. Meuth). This study was also slowly recovered (figure, A–E). A LN extracted during supported by Genzyme Therapeutics Ltd., United Kingdom thoracotomy revealed findings typical for Castleman disease, (“Alemtuzumab in autoimmune inflammatory neuro- including hyperplasia, germinal center regression, plasma degeneration: mechanisms of action and neuroprotective po- cell expansion, and hyalinized vessels (figure, J and K). tential” [ALAIN01] to H. Wiendl, S.G.Meuth, and T. Ruck). HHV-8 immunohistochemical staining was negative. Ascites interleukin (IL)-6 levels were strongly increased (4,030 Disclosure pg/mL, reference: <3 pg/mL, otherwise published value: L. Rolfes received travel reimbursements from Merck Serono 4,440 pg/mL4). Consequently, a diagnosis of iMCD with and Sanofi Genzyme. S. Pfeuffer received travel reimbursements TAFRO (Thrombocytopenia, Anasarca, MyeloFibrosis, from Sanofi Genzyme and Merck Serono; received honoraria Renal failure, Organomegaly) subtype (typical histopatho- for lecturing from Sanofi Genzyme, Biogen, and Mylan logic findings, HHV-8 negative immunohistochemistry, Healthcare; and received research support from Merck Serono, thrombocytopenia (figure, B), fever, organomegaly, absence DiaMed, and the German Multiple Sclerosis Society. T. Ruck of hypergammaglobulinemia (0.8 g/dL), lymphadenopathy, received travel expenses and financial research support from and hyperplasia of megakaryocytes) was established.5 Genzyme and Novartis and received honoraria for lecturing from Roche, Merck, Genzyme, Biogen, and Teva. S. Wind- Repeated IVMPS and IA treatment (figure, A♦,•) induced hagen, I. Oschlies, H.-J. Pavenstädt, and L. Angenendt report no ongoing remission. The patient fully recovered and continues disclosures. H. Wiendl received compensation for serving on to receive a small dose of oral prednisolone (5 mg/d). He is Scientific Advisory Boards/Steering Committees for Bayer regularly screened (every 3 months) for re-emerging disease HealthCare, Biogen Idec, Sanofi Genzyme, Merck Serono, and activity and development of secondary malignancies.6 Novartis; received speaker honoraria and travel support from Bayer Vital GmbH, Bayer Schering AG, Biogen, CSL Behring, EMD Serono, Fresenius Medical Care, Genzyme, Merck Discussion Serono, OmniaMed, Novartis, and Sanofi-Aventis; received compensation as a consultant from Biogen Idec, Merck Serono, Development of secondary autoimmunity after alemtuzumab Novartis, Roche, and Sanofi Genzyme;andalsoreceivedre- treatment is a well-known phenomenon and has been search support from Bayer HealthCare, Bayer Vital, Biogen addressed by a strict risk management program screening Idec, Merck Serono, Novartis, Sanofi Genzyme, Sanofi US, and especially for thrombocytopenia, thyroid, and kidney auto- Teva. J. Krämer received honoraria for lecturing from Biogen, immunity. Although the incidence of these disorders might Novartis,Mylan,andTevaandfinancial research support from have been overestimated according to postmarketing studies,2 Sanofi Genzyme. S.G. Meuth received honoraria for lecturing a variety of other rare entities have been described in the past and travel expenses for attending meetings from Almirall, 2 years. These reports include both organ-restricted diseases Amicus Therapeutics Germany, Bayer HealthCare, Biogen, and systemic conditions.3 Two additional iMCD cases have Celgene, DiaMed, Genzyme, MedDay Pharmaceuticals, Merck been described, with disease onset at 31 and 41 months after Serono, Novartis, Novo Nordisk, Ono Pharma, Roche, Sanofi- alemtuzumab initiation.6,7 Aventis, Chugai Pharma, QuintilesIMS, and Teva and his research is funded by the German Ministry for Education and One feature shared by HLH and iMCD is their association with Research (BMBF), Deutsche Forschungsgemeinschaft (DFG), different malignancies which can occur before, concurrent with, Else Kröner Fresenius Foundation, German Academic Ex- or after iMCD diagnosis (with an incidence of up to 19%). change Service, Hertie Foundation, Interdisciplinary Center for These require careful consideration when adjusting the patient’s Clinical Studies (IZKF) Muenster, German Foundation

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 3 Neurology, Almirall, Amicus Therapeutics Germany, Biogen, DiaMed, Fresenius Medical Care, Genzyme, Merck Serono, Appendix (continued)

Novartis, Ono Pharma, Roche, and Teva. Go to Neurology.org/ Name Location Role Contribution NN for full disclosures. Linus University Author Critical revision of the Angenendt, Hospital manuscript for intellectual Publication history MD Munster,¨ content Received by Neurology: Neuroimmunology & Neuroinﬂammation June Munster¨ 14, 2019. Accepted in ﬁnal form September 6, 2019. Heinz University Author Critical revision of the Wiendl, MD Hospital manuscript for intellectual Munster,¨ content Munster¨ Appendix Authors Julia University Author Study concept and design Name Location Role Contribution Kramer,¨ MD Hospital and critical revision of the Munster,¨ manuscript for intellectual Leoni University Author Study concept and design, Munster¨ content Rolfes, MD Hospital acquisition and Munster,¨ interpretation of data, and Sven G. University Author Study concept and design Munster¨ drafted the manuscript Meuth, MD, Hospital and critical revision of the PhD Munster,¨ manuscript for intellectual Steffen University Author Acquisition and Munster¨ content Pfeuffer, Hospital interpretation of data and MD Munster,¨ drafted the manuscript Munster¨

Tobias Ruck, University Author Acquisition of data and MD Hospital critical revision of the References Munster,¨ manuscript for intellectual 1. Coles AJ, Twyman CL, Arnold DL, et al. Alemtuzumab for patients with relapsing Munster¨ content multiple sclerosis after disease-modifying therapy: a randomised controlled phase 3 trial. Lancet 2012;380:1829–1839. Susanne Hospital Author Critical revision of the 2. Coles AJ, Cohen JA, Fox EJ, et al. Alemtuzumab CARE-MS II 5-year follow-up: Windhagen, Osnabruck,¨ manuscript for intellectual efficacy and safety findings. Neurology 2017;89:1117–1126. MD Osnabruck¨ content 3. Saarela M, Senthil K, Jones J, et al. Hemophagocytic lymphohistiocytosis in 2 patients with multiple sclerosis treated with alemtuzumab. Neurology 2018;90: Ilske University- Author Acquisition of data and 849–851. Oschlies, Hospital critical revision of the 4. Kubokawa I, Yachie A, Hayakawa A, et al. The first report of adolescent TAFRO MD Schleswig- manuscript for intellectual syndrome, a unique clinicopathologic variant of multicentric Castleman’s disease. Holstein, Kiel content BMC Pediatr 2014;14:139. 5. Liu AY, Nabel CS, Finkelman BS, et al. Idiopathic multicentric Castleman’s disease: Hermann- University Author Critical revision of the a systematic literature review. Lancet Haematol 2016;3:e163–e175. Joseph Hospital manuscript for intellectual 6. Somerfield J, Hill-Cawthorne GA, Lin A, et al. A novel strategy to reduce immuno- Pavenstadt,¨ Munster,¨ content genicity of biological therapies. J Immunol 2010;185:763–768. MD Munster¨ 7. Boddu PC, Hassan A, Mauer A, Singh DA, Weisenberg ES. Rare case of multicentric castleman disease secondary to alemtuzumab therapy. Int J Tumor Ther 2015;4:1–4.

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN CLINICAL/SCIENTIFIC NOTES OPEN ACCESS West Nile virus-associated vasculitis and intracranial hemorrhage

Adil Harroud, MD, Ahmad Almutlaq, MBBS, David Pellerin, MD, MSc, Dan Paz, MD, Grant J. Linnell, DO, and Correspondence Daniel Gendron, MD Dr. Harroud [email protected] Neurol Neuroimmunol Neuroinﬂamm 2020;7:e641. doi:10.1212/NXI.0000000000000641

West Nile virus (WNV) is a neurotropic flavivirus and the leading cause of mosquito-borne diseases in North America. In approximately 20% of cases, infection leads to a self-limited febrile illness and, in less than 1%, to a neuroinvasive disease often manifesting as meningoencephalitis with or without acute flaccid paralysis.1 Herein, we report the first case of WNV- associated CNS vasculitis and intracranial hemorrhage, thus expanding the spectrum of WNV infection.

Case report A 73-year-old woman was admitted in late summer to the Montreal Neurological Institute and Hospital in Montreal, Canada, with a 2-day history of acute confusion and high fever. She had a remote history of renal cell and breast carcinomas, both in complete remission and oﬀ treatment. Initial neurologic examination was remarkable for encephalopathy and right extensor plantar response. Antimicrobials for suspected meningoencephalitis were empirically started. Initial head CT scan was unremarkable. CSF examination revealed a lymphocytic pleocytosis (28 white cells/μL with 59% lymphocytes), no red cells, protein of 0.84 mg/dL, and glucose of 53.6 mg/dL. Antimicrobials were stopped the next day after negative CSF cultures and herpes simplex virus PCR. On day 6, the patient developed a markedly decreased level of consciousness and aspiration pneumonia requiring intubation. Brain MRI showed extensive and conﬂuent leukoencephalopathy and interval appearance of bilateral convexity subarachnoid hemorrhage (SAH).

On day 8, the patient was found to have generalized myoclonus. Continuous EEG monitoring showed severe slowing but no epileptic activity. The patient was extubated a few days later and had persistent attention deficit, fluctuating dysphasia, and right extensor plantar response. Repeat brain MRI with angiography demonstrated interval enlargement of convexity SAHs (figure, A) and tapered stenosis of the left anterior cerebral artery A2 segments. A digital subtraction angiography (DSA) confirmed the findings on MRA (figure, B). Follow-up brain MRI with vessel wall imaging a month later showed a new small left frontal parasagittal intracerebral hemorrhage (ICH) in the absence of cortical or deep-seated microbleeds and intramural concentric enhancement of the wall of arteries in the circle of Willis (figure, C and D).

During hospitalization, WNV serology resulted positive for immunoglobulin (Ig) M and IgG in serum and CSF (performed on day 9 and repeated on day 41). A positive plaque reduction neutralization test (PRNT, titers 1:80 and 1:40 on days 9 and 41, respectively) conﬁrmed the diagnosis of WNV neuroinvasive disease. An extensive additional workup ruled out coexisting autoimmune, paraneoplastic, and systemic vasculitis. Of note, varicella zoster virus CSF PCR

From the Department of Neurology and Neurosurgery (A.H., A.A., D. Pellerin, D.G.), Montreal Neurological Hospital and Institute, McGill University, Montreal, Quebec, Canada; Centre for Clinical Epidemiology (A.H.), Department of Epidemiology, Lady Davis Institute for Medical Research, Jewish General Hospital, Montreal, Quebec, Canada; National Neuroscience Institute (A.A.), King Fahad Medical City, Riyadh, Saudi Arabia; and Department of Radiology (D. Paz, G.J.L.), Montreal Neurological Hospital and Institute, McGill University, Montreal, Quebec, Canada.

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

(A) Axial susceptibility-weighted image showing prominent blooming (arrows) in the sulci of the bilateral posterior parietal regions at the ver- tex, frontal lobes, and sylvian fis- sures, indicating subarachnoid hemorrhage. (B) DSA showing mild tapered stenosis in the proximal extent of the A2 segment of the left anterior cerebral artery (oval circle) and mild focal ectasia of left angular artery (rectangle). (C) Follow-up MRI demonstrating a new small left frontal intracerebral hemorrhage (arrow) on an axial susceptibility-weighted image. (D) Concentric gadolinium enhancement of the supraclinoid segment of the internal carotid arteries and proximal middle cerebral arteries bilaterally (arrows), seen on magnetic resonance vessel wall imaging. (E) Follow-up DSA showing interval resolution of the vascular changes previously noted in the A2 segment of the left anterior cerebral artery (oval circle) and left angular artery (rectangle). DSA = digital subtraction angiography.

and serologies for syphilis and HIV were negative. Whole- vascular irregularities suggestive of a vasculopathy in a pa- body PET/CT showed no evidence of recurrent cancer. The tient with positive WNV serology.3 In our case, intracranial patient improved with supportive treatment including neu- vessel wall MRI was consistent with arterial inflammation, rointensive care monitoring and IV hydration. Cortico- further supporting that the vasculopathy was secondary to steroids and other previously tried therapeutic agents for WNV neuroinvasive disease. There is in vitro evidence that WNV (e.g., IV immunoglobulin, ribavirin, and interferon- WNV may gain entry to the CNS and infect neurons partly alpha)1 were considered but not administered, given the lack through invasion of endothelial cells,1,5 providing a poten- of evidence for efficacy and the patient’s continuing clinical tial explanatory mechanism for the development of vascu- improvement. Repeat CSF examination on day 41 revealed litis in this setting. We hypothesize that the intracranial improved pleocytosis (7 white cells/μL), and repeat DSA 3 hemorrhages observed in this case were caused by rupture days later showed resolution of the previously noted left A2 of small vasculitic vessels. This is analogous to the obser- tapering (figure, E). Radiologic monitoring showed resolu- vations of retinal hemorrhages with WNV retinal vasculitis.6 tion of the intracranial hemorrhages and vascular stenosis Similarly, vasculitis accompanied by cerebral hemorrhage over the ensuing months. She did not complain of overt has been documented with other viral encephalitides, such headache during the course of the illness, although this was as herpesviruses.7 limited by the initial encephalopathy. On discharge, the patient was able to walk without support but suffered from In conclusion, this case highlights the need for clinicians to be residual cognitive deficits. vigilant for vasculitic complications, including intracranial hemorrhage, in cases of WNV infection. Similarly, WNV neuroinvasive disease should be considered in the differential Discussion diagnosis of infectious cerebral vasculitis and atypical hemorrhagic stroke. This case illustrates some of the typical manifestations of WNV neuroinvasive disease, including encephalitis with lymphocytic pleocytosis and myoclonus. The diagnosis was Study funding confirmed based on positive IgM serologies and PRNT in No targeted funding reported. serum and CSF. In addition, this case showcases the unique complication of SAH and ICH associated with WNV in- Disclosure fection. Although rare cases of ischemic stroke in the setting of The authors report no disclosures. Go to Neurology.org/NN – WNV have been described,2 4 intracranial hemorrhage had for full disclosures. not been previously reported to our knowledge. Publication history We also describe the rare occurrence of WNV-associated Received by Neurology: Neuroimmunology & Neuroinflammation CNS vasculitis. A single report previously documented July 18, 2019. Accepted in final form October 4, 2019.

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN References Appendix Authors 1. Suthar MS, Diamond MS, Gale M Jr. West Nile virus infection and immunity. Nat Rev Microbiol 2013;11:115–128. Name Location Role Contribution 2. Alexander JJ, Lasky AS, Graf WD. Stroke associated with central nervous system vasculitis after West Nile virus infection. J Child Neurol 2006;21:623–625. Adil McGill Author Conception and design, 3. Lowe LH, Morello FP, Jackson MA, Lasky A. Application of transcranial Doppler Harroud, University, drafting the manuscript, and sonography in children with acute neurologic events due to primary cerebral and West – MD Montreal, review of the literature Nile vasculitis. AJNR Am J Neuroradiol 2005;26:1698 1701. Canada 4. Zafar S, Dash D, Chachere M, Cowart J, Kass J. West Nile virus infection associated with central nervous system vasculitis and strokes. Neurology 2012;78 (suppl 3): Ahmad McGill Author Conception and design, P03.264. Almutlaq, University, drafting the manuscript, and 5. Maximova OA, Pletnev AG. Flaviviruses and the central nervous system: revisiting – MBBS Montreal, review of the literature neuropathological concepts. Annu Rev Virol 2018;5:255 272. Canada 6. Garg S, Jampol LM, Wilson JF, Batlle IR, Buettner H. Ischemic and hemorrhagic retinal vasculitis associated with West Nile virus infection. Retina 2006;26:365–367. David McGill Author Revised the manuscript for 7. Zepper P, Wunderlich S, Forschler A, Nadas K, Hemmer B, Sellner J. Pearls & Oy- Pellerin, University, intellectual content sters: cerebral HSV-2 vasculitis presenting as hemorrhagic stroke followed by mul- – MD, MSc Montreal, tifocal ischemia. Neurology 2012;78:e12 e15. Canada

Dan Paz, McGill Author Revised the manuscript for MD University, intellectual content Montreal, Canada

Grant J. McGill Author Revised the manuscript for Linnell, University, intellectual content and DO Montreal, approved the final version for Canada publication

Daniel McGill Author Revised the manuscript for Gendron, University, intellectual content and MD Montreal, approved the final version for Canada publication

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 3 CLINICAL/SCIENTIFIC NOTES OPEN ACCESS An inﬂammatory milieu Optic perineuritis, retroperitoneal ﬁbrosis, and giant cell arteritis

Doria M. Gold, MD and Steven L. Galetta, MD Correspondence Dr. Gold Neurol Neuroimmunol Neuroinﬂamm 2020;7:e642. doi:10.1212/NXI.0000000000000642 [email protected]

Case summary Optic nerve perineuritis targets the optic nerve sheath; it is idiopathic or a manifestation of systemic inflammatory diseases such as myelin oligodendrocyte glycoprotein (MOG) antibody syndrome, sarcoidosis, granulomatosis with polyangiitis, IgG4-related disease, or giant cell arteritis (GCA).1 Radiographically, there is optic nerve sheath enhancement and, occasionally, orbital fat “streaking.”1 Biopsies have identified dural sheath lymphocytic infiltrate, perineural fibrous tissue, granulomas, or evidence of small-vessel vasculitis.1 Retroperitoneal fibrosis (RPF) is defined by abdominal organ fibrosis. Usually idiopathic, it may be secondary to medication, radiation, or systemic inflammatory conditions.2 Pathologically, its hallmark is fibrosis and infiltration of macrophages, B-lymphocytes, and T-lymphocytes.2 Often considered when older patients present with headache or ischemic optic neuropathy, GCA is the most common large- vessel vasculitis.3 It has been proposed that GCA results from a T-cell-mediated inflammatory cascade, causing vascular intimal hyperplasia.3 GCA has rarely been reported in association with RPF4,5 or perineuritis.6,7 We present a patient initially diagnosed with idiopathic RPF treated with immunosuppression, who subsequently developed optic perineuritis followed by GCA.

A 71-year-old man presented to the emergency department with urosepsis and renal insufficiency. He was treated with intravenous (IV) antibiotics for positive urine and blood cultures, but his kidney function remained abnormal. Renal ultrasound revealed bilateral hydronephrosis, and MRI abdomen/pelvis showed a retroperitoneal mass. Biopsy demonstrated patchy fibrosis and soft tissue lymphoplasmacytic infiltrate. IgG antibody staining was negative. He was treated with mycophenolate for 2 years.

Six years later, he presented for having had 5 days of left eye vision loss that remained stable from onset. He endorsed jaw soreness but denied diplopia. Acuity was 20/25 in each eye. He had a trace left afferent pupillary defect, and color plates were performed more slowly on the left. Optic discs were without swelling or pallor. He had normal ocular motility. Humphrey visual fields revealed few nonspecific points of depression in each eye and an inferonasal defect in the left eye that corresponded with his complaint. He had no temporal artery thickening or tenderness. He was sent to the emergency department for blood work and IV steroids for presumed GCA. Erythrocyte sedimentation rate (20 mm/h) and platelets (281,000/mm3)wereinthenormal range, and C-reactive protein was mildly elevated (13 mg/L). IV methylprednisolone was initiated. Chest X-ray was normal. MRI/magnetic resonance angiography of the brain demonstrated mild microvascular ischemic changes and volume loss with normal vasculature. Orbital MRI showed optic nerve sheath enhancement and orbital fat stranding (figure, A and C).

Laboratory workup revealed elevated HgA1c (7.9%), serum glucose (239 mg/dL), and subclinical hypothyroidism (thyroid stimulating hormone 5.980 mIU/L). In addition, there were negative Lyme titers, serum paraneoplastic panel, aquaporin-4 receptor, MOG, and antineutrophil cytoplasmic antibodies. Serum IgG4 levels were not elevated. CSF examination showed red blood cells

From the Department of Neurology (D.M.G.) and Departments of Neurology and Ophthalmology (S.L.G.), NYU Langone Medical Center, New York.

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 NYU Langone Medical Center Department of Neurology. 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.

Coronal (A) and axial (C) MRI orbits at initial presentation demonstrating optic perineuritis (red arrows) and fat stranding (blue arrowheads) left greater than the right. Axial MRI orbits after 6 months demonstrating bilateral thickening and enhancement of the ophthalmic arteries (blue arrows) and improved optic perineuritis (red arrows) (B and D).

3, white blood cells 3, elevated protein (74 mg/dL), elevated improve. In the cases of RPF described with GCA, patients had glucose(144mg/dL),negativecytology,flow cytometry, cul- a more fulminant presentation. Perhaps our patient’s immuno- ture, venereal disease research laboratory test (to evaluate for suppression for RPF with mycophenolate, which suppresses syphilis), and Lyme antibodies. Serum immunoglobulins, CSF B-lymphocyte and T-lymphocyte proliferation, dampened the IgG index, and oligoclonal bands were not checked. later cell-mediated response. Fortunately, our patient had an indolent course despite tissue and arterial inflammatory changes. Corticosteroids were tapered over 5 months. One month To our knowledge, the conditions RPF, perineuritis, and GCA after complete discontinuation, he endorsed a new left have not been documented in a single patient nor after pro- temporal headache. He denied visual symptoms, jaw longed immunosuppression. claudication, or muscle weakness. Examination was improved with acuity 20/20 OU. Repeat MRI brain/orbits Study funding revealed bilateral ophthalmic and superficial temporal ar- No targeted funding reported. tery thickening and enhancement, compatible with GCA. There was near-complete resolution of the left optic per- Disclosure ineural enhancement (figure, B and D). Temporal artery D.M. Gold and S.L. Galetta report no disclosures. Go to biopsy confirmed the diagnosis with lymphocytes, mac- Neurology.org/NN for full disclosures. rophages, granulomatous inflammation, and no IgG4- positive cells. High-dose oral prednisone (1 mg/kg) treatment Publication history was initiated and followed by a slow taper, which maintained Received by Neurology: Neuroimmunology & Neuroinflammation resolution of headaches. Repeat imaging showed decreased September 23, 2019. Accepted in final form October 23, 2019. ophthalmic and temporal artery enhancement and resolved optic nerve sheath inflammation. There was no aortic involvement on MRA chest. Appendix Authors Our patient manifested with 3 sequential inflammatory dis- Name Location Role Contribution orders: RPF, optic perineuritis, and GCA. Clinically, each was Doria M. NYU Langone Author Drafting/revising the mild. Although an umbrella of GCA-related vasculitis may be Gold, MD Medical Center, manuscript, data acquisition, suggested, his course supports a less specific systemic in- New York and study concept or design flammatory process uniquely expressed at 3 different time Steven L. NYU Langone Author Study concept or design, data Galetta, Medical Center, acquisition, and revised the points. Although the exact pathogenesis of perineuritis, RPF, MD New York manuscript for intellectual and GCA typically do not overlap, each condition results from content cell-mediated immunologic overactivation, resulting in fibrosis and granulomatous inflammation. References In the few reported cases of perineuritis and GCA, most had 1. Purvin V, Kawasaki A, Jacobson DM. Optic perineuritis: clinical and radiographic bilateral perineural involvement on neuroimaging. In addition, features. Arch Ophthalmol 2001;119:1299–1306. 2. Vaglio A, Maritati F. Idiopathic retroperitoneal fibrosis. J Am Soc Nephrol 2016;27: examination often demonstrated optic neuropathy or ocular 1880–1889. motility impairment.6,7 Unlike our patient who rapidly respon- 3. Watanabe R, Goronzy JJ, Berry G, Liao YJ, Weyand CM. Giant cell arteritis: from pathogenesis to therapeutic management. Curr Treat Opt Rheumatol 2016;2: ded to steroids, the optic nerve dysfunction typically did not 126–137.

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN 4. Protopsaltis I, Sotiropoulos A, Foteinos A, et al. Chronic periaortitis (retroperi- 6. Pappolla A, Silveira F, Norscini J, Miquelini L, Patrucco L. Bilateral optic per- toneal fibrosis) concurrent with giant cell arteritis: a case report. J Med Case Rep ineuritis as initial presentation of giant cell arteritis. Neurologist 2019;24: 2014;8:167. 26–28. 5. Tolle PA, Kesten F, Daikeler T. Giant cell arteritis followed by idiopathic retroperi- 7. Liu TY, Miller NR. Giant cell arteritis presenting as unilateral anterior ischemic optic toneal fibrosis in the same patient—an unexpected positron emission tomography neuropathy associated with bilateral optic nerve sheath enhancement on magnetic finding. Rheumatology (Oxford) 2012;51:1549. resonance imaging. J Neuroophthalmol 2015;35:360–363.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 3 VIEWS & REVIEWS OPEN ACCESS An argument for broad use of high eﬃcacy treatments in early multiple sclerosis

James M. Stankiewicz, MD, and Howard L. Weiner, MD Correspondence Dr. Stankiewicz Neurol Neuroimmunol Neuroinflamm 2020;7:e636. doi:10.1212/NXI.0000000000000636 [email protected] Abstract Two different treatment paradigms are most often used in multiple sclerosis (MS). An escalation or induction approach is considered when treating a patient early in the disease course. An escalator prioritizes safety, whereas an inducer would favor efficacy. Our understanding of MS pathophysiology has evolved with novel in vivo and in vitro observations. The treatment landscape has also shifted significantly with the approval of over 10 new medications over the past decade alone. Here, we re-examine the treatment approach in light of these recent developments. We believe that recent work suggests that early prediction of the disease course is fraught, the amount of damage to the brain that MS causes is underappreciated, and its impact on patient function oftentimes is underestimated. These concerns, coupled with the recent availability of agents that allow a better therapeutic effect without compromising safety, lead us to believe that initiating higher efficacy treatments early is the best way to achieve the best possible long-term outcomes for people with MS.

From the Department of Neurology, Brigham and Women’s Hospital, Partners MS Center, Harvard Medical School, Boston, MA.

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

Copyright © 2019 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary DTI =diffusion tensor imaging; EDSS = Expanded Disability Status Scale; HETA = highly effective treatment early approach; IFN = interferon; MTR = magnetization transfer ratio; NARCOMS = North American Research Committee on Multiple Sclerosis; NEDA = no evidence of disease activity; NFL = neurofilament light; PCORI = Patient-Centered Outcomes Research Institute; PML = progressive multifocal leukoencephalopathy; SC = subcutaneous.

MS is a chronic autoimmune disorder resulting in accumu- progress to an Expanded Disability Status Scale (EDSS) lated damage to the brain, spinal cord, and optic nerves that score of 6 (requires unilateral assistance). However, the can lead to significant neurologic disability. Two and a half reported absolute coefficients of magnitude for the effect are decades after the first drug approval for MS, no consensus has underwhelming with male sex increasing the chance of been achieved regarding which treatments should be most reaching a DSS of 6 by 0.17, age at onset of 0.03, motor often used with practices varying widely1 and recently pub- symptoms at onset of 0.21, and limb ataxia/balance symp- lished international consensus guidelines differing.2,3 toms at onset of 0.29. In a separate cohort, Tremlett et al.5 reported similar weak effects with male sex contributing a 1.1 Two potential treatment frameworks have been articulated. hazard relative to female sex and motor onset of symptoms Some advocate an escalation approach, positing that a pre- with a hazard ratio of 1 to probability of reaching an EDSS of mium should be placed on safety. The argument continues 6. Other large cohort studies do not present the effect on the that if during prospective monitoring a patient experiences overall risk of achieving EDSS in a readily applicable way. A breakthrough disease as evinced by relapses, MRI change, or study examining the effect of African American race on disability, then strong consideration should be given to EDSS progression failed to show an effect.6 Cohort studies switching to an agent with increased efficacy. In this way, examining whether demographic factors influence the risk of patients who do not require stronger medicines are spared developing progressive disease report similar weak effects. potential side effects and safety concerns. A second model is Another cohort reported that sex and symptom type at an “induction” or, better named, a highly effective treatment presentation do not influence the risk of conversion to early approach (HETA). Proponents of the HETA contend secondary progressive MS (SPMS), whereas age at disease that a neurologist’s current ability at initial presentation to onset has little influence (hazard ratio 1.02) although sta- predict the long-term outcome is limited. They also argue that tistically significant.7 A study of over 8,000 patients found a neurologist’s ability to prospectively determine whether that the type of attack (monofocal vs polyfocal) was not ongoing damage to the nervous system is occurring is limited. predictive of conversion to SPMS, whereas the magnitude of In addition, HETA advocates view the risk profile of some the effect for sex (−1.24 for women) and onset age (−0.93) highly efficacious drugs as not materially worse than less ef- was small although statistically significant.8 ficacious treatments. A number of thoughtful articles considering whether higher Conventional clinical imaging can efficacy agents should be used early have been published. Nevertheless, we believe it is time to re-examine these miss or underestimate frameworks in more detail, given the accumulating research ongoing damage on pathology, imaging, and disease course and in light of the evolving treatment landscape. Here, we elaborate on why we When “normal-appearing white matter” is examined with im- believe a HETA is warranted for most patients with MS early munohistochemical techniques, more damaged axons are seen 9 in disease (age <40 years old). in patients with MS than in normal controls. As the MRI magnet strength increases, better fit is obtained with clinical measures, and even at the highest available strengths, correla- Our ability to predict disease course at tions are fair at best.10 Studies comparing 1.5T and 7T MRI for onset is limited detecting white matter and cortical lesions identify a markedly increased yield at 7T.11 In addition, the above-referenced study Clinicians oftentimes discuss how they use factors that have by Sinnecker et al. reported that every T2 hyperintensity vi- been identified in the literature as carrying a poor prognosis sualized in the brain was hypointense on magnetization pre- to guide the choice of high- vs low-efficacy medication start. pared rapid acquisition with gradient echo at 7T, suggesting Commonly cited demographic factors for poorer outcomes that all MS lesions are in fact “black holes.” Even then, 7T include older age at onset, male sex, race, and motor or imaging is an imperfect tool. Pitt et al.12 reported that 7T MRI cerebellar presentation. In fact, the relationship between visualized 82%–93% of cortical lesions apparent at pathology. these factors and prognosis in longitudinal patient cohorts is Measures of whole-brain atrophy and gray matter atrophy fair at best. Weinshenker et al. 4 reported a significant effect correlate cross-sectionally better with clinical disability than of the above-cited factors on whether patients were liable to conventional measures such as T2 lesion volume, T1

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN hypointensities, and gadolinium enhancing lesions,13 but depression in 35%, fatigue in 72%, and tremor/coordination such measures are not available in routine clinical practice. in 51%.21 Novel imaging techniques such as diffusion tensor imaging (DTI), magnetization transfer ratio (MTR), and 23Na MRI A recent survey of NARCOMS participants with a mean age imaging all demonstrate that more damage is present than is of 55 years found that 58% of patients were not working, with visualized on conventional MRI. Imaging of diffusion of 48.5% of those surveyed received disability benefits. Of those water along axonal tracts (DTI) with sophisticated pro- employed, 27% reported missing 6 days or more of work cessing has suggested that the contribution of white matter a year. Moderate to severe cognitive impairment, fatigue, and abnormalities to cognitive impairment is determined by hand function problems were associated with both disability damage in the otherwise normal-appearing white matter and and absenteeism.22 that this association is in part independent of gray matter volume atrophy and lesion load.14 23Na MRI imaging allows Of adults 65 years or older, 15% of patients without MS direct visualization of ongoing cellular dysfunction and cell reported using an assistive device, whereas 81% of patients death. This is found to be abnormal in both lesioned areas with MS required an assistive device in the NARCOMS co- of the brain and normal-appearing white matter.15 A study of hort, a 6-fold difference.23,24 45 patients with newly diagnosed relapsing-remitting MS found a decrease of whole-brain cerebral viscoelasticity relative to matched healthy volunteers and suggested a more widespread disturbance of tissue integrity than expected Long-term follow-up studies of from the few visible T2 lesions.16 A measure of macromo- patients on platform agents reveal the lecular integrity, MTR is reduced in normal-appearing white risks of undertreatment and gray matter from patients even at the earliest clinical stages of the disease17 with reduced MTR correlated with An extension of the pivotal glatiramer acetate trial found that disability18 and cognitive impairment.19 of patients who were maintained on treatment for an average of 13.6 years, 35% were likely to develop secondary progressive disease. When considering disability progression, 59% progressed to at least an EDSS of 4, with 18% worsening MS is rarely benign over the long term to an EDSS of 6, and 3% reaching an EDSS of 8.25 A follow-up when dysfunction is study of patients enrolled in the interferon (IFN) β-1a sub- carefully interrogated cutaneous (SC) pivotal trial26 reported that of those receiving IFN β-1a SC 44mcg regularly for at least 7 years, 23.9% Many MS clinicians will discuss patients who have been progressed to an EDSS of 4, 19.7% to an EDSS of 6, 12% to an maintained off treatment or on low efficacy treatment that EDSS of 6.5, and 6.1% to an EDSS of 7, with a mean increase have done “well.” One might wonder, however, what qualifies in EDSS in both treated groups (22 and 44 μg) of 1.1 and as well? Would this patient be in the same condition if they did mean EDSS score of 3.5 at the last visit. When considering the not have MS? As clinicians, how hard are we looking? long-term follow-up group as a whole (22 and 44 μg doses), 19.7% were found to have converted to SPMS. Of patients A recent report from Tallantyre et al.20 found that in a care- taking IFN β-1b assessed at the 16-year pivotal phase III fol- fully examined untreated population with disease duration low-up,27 45.8% had reached an EDSS of 6. These lackluster >15 years, 2.9% of patients had “benign” disease as defined by results are despite these studies being subject to corruption by an EDSS <3, no significant fatigue, mood disturbance, cog- survivorship bias so that patients who are not doing well drop nitive impairment, or disrupted employment. Indeed, the out of the treated group or might not be assessed in the follow- absolute number of patients identified is instructive with only up, potentially leading these studies to report better outcomes 9 patients in a cohort of 1,049 patients included for analysis than those that truly exist. Some would interpret these results classified as truly benign. as positive and that those included who have suffered from long-standing neurologic illness are “well enough,” given that Recent patient survey work and disability claims reveal sig- they have had the disease for some time. We do not believe nificant disability over time. In the North American Research that this is acceptable and take as our goal that patients track in Committee on Multiple Sclerosis (NARCOMS) database of a similar way to individuals not diagnosed with MS. One patients mostly treated at 15 years after diagnosis, only 13% might also question how accurately the EDSS reflects true of patients reported no or mild symptoms. Moderate or disability, given only fair inter-rater and intra-rater reliability at greater disability was reported by >40% of patients in hand lower levels of the scale and less sensitivity at higher levels function, vision, cognition, bowel/bladder function, spas- with ambulatory dysfunction primarily determining the level ticity, pain, fatigue, and coordination. At 30 years, moderate in the upper ranges.28 or greater disability was reported for mobility in 69%, hand function in 60%, vision in 47%, cognition in 50%, bowel/ More efficacious MS drugs, which better control inflammatory bladderfunctionin70%,spasticityin65%,painin64%, disease, are not a panacea. Studies of even the most effective

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 3 agents available for MS indicate that no evidence of disease The challenges of patient preference activity (NEDA) is achieved in only 47.7% of ocrelizumab- treated patients29 when followed up for 96 weeks and 29.5% of and adherence, therapeutic inertia, natalizumab-treated patients followed up for 2 years.30 So, even and insurance coverage with best possible efficacy, an argument can be lodged that our treatments are not fully adequate. Other barriers exist to HETA implementation. Patient preference is a fundamental part of medication selection. The literature indicates that patient preferences may not be Short-term comparison studies in line with an appropriate treatment approach. Patients with longer disease duration prefer more efficacious therapies42 demonstrate superiority in reduction and tend to underestimate therapy risks and overestimate of relapses, MRI change, and disability benefits.43 This approach may be counter to disease biology, which likely benefits from the use of higher efficacy agents progression for some agents earlier in the disease. Good quality evidence suggests that ocrelizumab and alem- – tuzumab are more effective than IFN β-1a SC31 33 and fin- In addition, adherence and persistence remain a challenge with fi golimod is more effective than IFN β-1a IM34 in a selected a retrospective claims analysis nding that only between 52% and patient group. There is also MRI evidence that dimethyl fu- 62% of patients had injectable medication in their possession 44 marate better attenuates new MRI lesions than glatiramer 80% of the time. Adherence is better but also not ideal with oral fi acetate, although clinical outcomes over the 2-year trial were treatments with a claims database analysis nding that 80% fi mixed.35 We have been cautioned by statisticians not to medication possession was seen in 98.2% of patients on ngo- 45 compare across trials, but these head-to-head comparisons are limod and 87.8% on dimethyl fumarate. It is already a challenge informative. to treat patients adequately even if a patient is fully adherent.

Health insurance is another potential hurdle. Over 6% of the The reported safety profiles of some patients do not receive treatment because of financial concerns and insurance barriers.46 Governmental and private insurers highly effective agents do not currently struggle to construct policies without consensus from the MS materially differ from lower community regarding treatment. Developed policies can prevent ffi patient access to highly effective therapies by requiring step edits e cacy agents before approval. These policies may do patients harm, leading to The risk of developing progressive multifocal leukoence- a delayed start of the most appropriate treatments. Recent phalopathy (PML) in a John Cunningham virus antibody- understandings and trial data raise a concern that insurers that ff negative natalizumab-treated patient is estimated at 1: restrict patients with MS access to highly e ective immuno- fi 10,00036 and may be further reduced with extended interval therapies may not only be damaging their long-term nancial dosing.37 Natalizumab is otherwise generally well tolerated interest but also acting unethically. with a paucity of other potential side effects. In a population with rheumatoid arthritis, the risk of rituximab with long-term use is estimated to be 1:30,000.38 Some groups have con- Early intervention might substantively ducted retrospective analysis on the long-term safety of rit- alter disease course and prevent uximab in their patients with MS and found few adverse events.39,40 The safety profile of ocrelizumab is being defined irreversible progression, whereas later with time, but based on the molecular structure, one might treatment might not confer expect a similar profile to that of rituximab. So far, a single case much benefit of PML has been reported that can be directly attributed to ocrelizumab monotherapy. The imbalance of breast cancer Scalfari et al.47 reported that relapses which occur in the first 2 seen in trials requires further monitoring in the future, but years influence the disease course, whereas relapses that occur a long-term cohort treated with ocrelizumab does not vary in after this do not. A nationwide Danish registry found that cancer risk from risk estimates in a general population and MS patients left untreated for 2 years after MS symptom onset non–ocrelizumab-treated group.41 Although ostensibly high- reached an EDSS of 6 faster than those treated and trended efficacy agents’ safety may be in line with that seen in lower toward experiencing an earlier death.48 A meta-analysis of MS efficacy agents, further definition of the safety profile with clinical trials found a larger reduction in relapses and MRI time will be paramount. It seems short sighted, however, to damage in a younger patient group, with a lack of difference allow greater CNS tissue loss now because of worry about between treated and untreated patients seen in patients older drug side effects not currently manifest, especially given a lack than age 40.5 years.49 A population-based cohort in the United of definite mechanistic rationale to support safety concerns. Kingdom found that patients (n = 104) treated initially with

4 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN highly effective therapy had a lower mean change in EDSS at 5 been reported in patients who meet NEDA criteria.59 As such, years and a higher median time to sustained accumulation of a neurologist looking for overt signs of drug failure as manifest disability than those treated with moderately effective therapies by clinical or MRI change may be ill advised. We believe a re- (n = 488). This effect persisted after adjustment for covariates cent study helped expose the conceptual fallacy of switching including age and sex.50 after a patient experienced an “efficacy failure” on their current disease-modifying agent. De Flon et al.60 switched patients Once inflammatory nervous system damage occurs, other considered “stable” on their current injectable agent and found downstream effects might be seen including mitochondrial that after the switch, these patients developed fewer MRI dysfunction, oxidative damage from iron deposition, micro- lesions and had lower neurofilament light (NFL) levels. glial activation, and glutamate toxicity, which might cause additional neuronal damage and propagate a feed-forward loop.51 Damage to the nervous system early in the disease is The way forward oftentimes well compensated for initially but becomes ap- NFL holds promise as a potentially more sensitive way to parent after reserve is lost, typically decades after disease monitor neuronal damage. As such, it might be used to illu- onset.52 Keegan et al.53 reported that progressive lower ex- minate the question of early use of high-efficacy therapies vs tremity weakness can be traced to critically placed lesions in escalation approaches. Recently, the UCSF-Epic cohort the corticospinal tract, most often in the spinal cord. Once reported that at the 5-year follow-up, those treated with highly patients reach a DSS of 4, progression to cane dependence effective therapies experienced greater decreases in serum (DSS of 6) occurs independent of relapses.54 NFL compared with those on platform therapies, with a significant interaction found between NFL and EDSS.61 This Indeed, a rational interpretation of the published literature biomarker may also have utility in clinic as an adjunct to might be to use a HETA in patients younger than 40 years of conventional MRI and may even have prognostic value at age with a de-escalation strategy used in an older patient disease onset. However, further standardization and re- group who experience immune senescence and who could be finement of serum neurofilament testing is necessary before at more risk of infections and other potential complications. neurofilament can enter routine clinical use.62 If the working theories described above are correct, then, one Two Patient-Centered Outcomes Research Institute- would expect that more effectively preventing inflammation sponsored prospective studies will further explore the early would result in better protection from irreversible dis- benefit/risk profile of using more effective agents earlier vs an ability. This seems to be the case. The MSBase consortium escalation approach. Both populations will be deeply charac- recently analyzed a cohort of 1,555 patients with MS and terized beyond their primary endpoint. On the other hand, found that individuals treated with higher efficacy agents de- blinding is limited and the populations are quite heteroge- veloped secondary progressive disease later.55 neous. As such, it might be difficult to know how these results might apply to a specific patient in clinic given limited subsets at different ages and ranges of disease activity. Assuming that The perils of escalation enrollment goes as planned (a potential challenge given the current treatment landscape and fixed beliefs amongst clini- Neurologists choosing an escalation path face challenges. cians), these studies are not scheduled to report results until Therapeutic inertia may contribute to the loss of brain tissue 2022. One might reasonably question how much these studies in patients with MS. Harding et al.50 found a median delay of will add to double-blinded randomized placebo-controlled 2.4 years in patients who escalated disease-modifying agents. trials, which employed more rigorous methodologies. Another issue is a period of vulnerability that may exist in treatment transitions. Drugs may be sequenced too closely One might even ask, given the confluence of current evidence, together, temporally resulting in harmful effects. A more whether we need additional trial evidence to begin a HETA frequent harm is loss of nervous system tissue related to initiation in younger patients. Rheumatologists as a group de- a therapeutic gap during treatment transitions, with potential cided to begin more efficacious agents earlier based on a strong relapses, MRI activity, or disability accrual as a result.56 correlation between disease duration and chances of achieving Neurologists choosing more efficacious agents will less often remission or low disease activity. This adjustment improved the need to transition medications and expose their patients less outcomes achieved in patients with rheumatoid arthritis.63 to these potential dangers. In addition, a neurologist waiting for “efficacy failure” may be Conclusion engaged in a frivolous exercise. Our current ability to monitor disease using clinical and radiologic measures is fair at best. An examination of relevant pathology and MRI studies sug- Studies evaluating the prognostic value of NEDA find that it is gest that MS is a chronically progressive neurologic condition poorly predictive of the outcome over the longer term.57,58 rather than one characterized by episodic disease. These lines Ongoing brain structural damage (as measured by DTI) has of inquiry also suggest that we underappreciate the amount of

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 5 brain and spinal cord damage our patients experience over References time. We also lack the ability to effectively predict at initial 1. Marziniak M, Ghorab K, Kozubski W, et al. Variations in multiple sclerosis practice presentation who will do well and who will not over the long within Europe—is it time for a new treatment guideline? Mult Scler Relat Disord 2016;8:35–44. term. Even careful early prospective monitoring of disease 2. Rae-Grant A, Day GS, Marrie RA, et al. Comprehensive systematic review summary: activity may not adequately predict the outcome. It should be disease-modifying therapies for adults with multiple sclerosis. Neurology 2018;90: 789–800. our goal for patients that they adhere as closely to normal 3. Montalban X, Gold R, Thompson AJ, et al. ECTRIMS/EAN Guideline on the aging processes as possible. Current data suggest that we are pharmacological treatment of people with multiple sclerosis. Mult Scler 2018;24: 96–120. oftentimes not meeting this goal. In addition, recent devel- 4. Weinshenker BG, Rice GP, Noseworthy JH, Carriere W, Baskerville J, Ebers GC. The opments in the MS treatment landscape have left us with natural history of multiple sclerosis: a geographically based study: 3. MULTIVARI- some medications that recast the balance so that the potential ATE analysis OF predictive FACTORS and MODELS of outcome. Brain 1991;114: 1045–1056. harm to an MS patient’s nervous system outpaces by far po- 5. Tremlett H, Paty D, Devonshire V. Disability progression in multiple sclerosis is tential concerns about serious side effects. Given these factors, slower than previously reported. Neurology 2006;66:172–177. 6. Weinstock-Guttman B, Jacobs LD, Brownscheidle CM, et al. Multiple sclerosis we strongly believe neurologists and policy makers can better characteristics in African American patients in the New York state multiple sclerosis keep patients with MS on a course toward long-term healthy consortium. Mult Scler 2003;9:293–298. 7. Koch M, Uyttenboogaart M, van Harten A, De Keyser J. Factors associated with the aging by using a HETA stratagem. risk of secondary progression in multiple sclerosis. Mult Scler 2008;14:799–803. Available at: ncbi.nlm.nih.gov/pubmed/18573840. Accessed July 24, 2019. 8. Manouchehrinia A, Zhu F, Piani-Meier D, et al. Predicting risk of secondary progression in multiple sclerosis: a nomogram. Mult Scler 2019;25:1102–1112. Acknowledgment 9. Trapp BD, Peterson J, Ransohoff RM, Rudick R, Mörk S, Bö L. Axonal transection in The authors would like to thank Nardine Moawde for editing the lesions of multiple sclerosis. N Engl J Med 1998;338:278–285. 10. Nielsen AS, Kinkel RP, Madigan N, Tinelli E, Benner T, Mainero C. Contribution of the manuscript for nonintellectual content. cortical lesion subtypes at 7T MRI to physical and cognitive performance in MS. Neurology 2013;81:641–649. Study funding 11. Kollia K, Maderwald S, Putzki N, et al. First clinical study on ultra-high-field MR imaging in patients with multiple sclerosis: comparison of 1.5T and 7T. Am J Neu- Supported in part by the National MS Society and the Water roradiol 2009;30:699–702. Cove Charitable Foundation. 12. Pitt D, Boster A, Pei W, et al. Imaging cortical lesions in multiple sclerosis with ultra- high-field magnetic resonance imaging. Arch Neurol 2010;67:812–818. 13. Bermel RA, Bakshi R. The measurement and clinical relevance of brain atrophy in Disclosure multiple sclerosis. Lancet Neurol 2006;5:158–170. 14. Bodini B, Cercignani M, Khaleeli Z, et al. Corpus callosum damage predicts disability J.M. Stankiewicz has received consulting fees from Biogen Idec, progression and cognitive dysfunction in primary-progressive MS after five years. Novartis, EMD Serono, Hoffman-La Roche, Celgene, and Hum Brain Mapp 2013;34:1163–1172. Sanofi Genzyme. H.L. Weiner has served on the advisory boards 15. Inglese M, Madelin G, Oesingmann N, et al. Brain tissue sodium concentration in multiple sclerosis: a sodium imaging study at 3 tesla. Brain 2010;133:847–857. of The Guthy-Jackson Charitable Foundation, Teva, Biogen, 16. Wuerfel J, Paul F, Beierbach B, et al. MR-elastography reveals degradation of tissue Novartis, Sanofi-Aventis, Tilos, C-Bridge Capital, Genentech, integrity in multiple sclerosis. Neuroimage 2010;49:2520–2525. 17. Filippi M, Rocca MA. Magnetization transfer magnetic resonance imaging of the Genzyme, vTv Therapeutics, and MedDay; consulted for Bio- brain, spinal cord, and optic nerve. Neurotherapeutics 2007;4:401–413. dextris,Biogen,Novartis,Serono,Teva,Sanofi Genzyme, Tilos, 18. Ranjeva JP, Audoin B, Au Duong MV, et al. Local tissue damage assessed with statistical mapping analysis of brain magnetization transfer ratio: relationship with Tiziana Life Sciences, vTv Therapeutics, MedDay, Genentech, functional status of patients in the earliest stage of multiple sclerosis. AJNR Am J and CBridge Capital; and received research support from EMD Neuroradiol 2005;26:119–127. 19. Amato MP, Portaccio E, Stromillo ML, et al. Cognitive assessment and quantitative Serono,GoogleLifeSciences,theNIH,andtheNMSS.Goto magnetic resonance metrics can help to identify benign multiple sclerosis. Neurology Neurology.org/NN for full disclosures. Funding information is 2008;71:632–638. provided at the end of the article. 20. Tallantyre EC, Major PC, Atherton MJ, et al. How common is truly benign MS in a UK population? J Neurol Neurosurg Psychiatry 2019;90:522–528. 21. Kister I, Bacon TE, Chamot E, et al. Natural history of multiple sclerosis symptoms. Publication history Int J MS Care 2013;15:146–158. Neurology: Neuroimmunology & Neuroinflammation 22. Salter A, Thomas N, Tyry T, Cutter G, Marrie RA. Employment and absenteeism in Received by working-age persons with multiple sclerosis. J Med Econ 2017;20:493–502. August 28, 2019. Accepted in final form October 1, 2019. 23. Kister I, Chamot E, Salter AR, Cutter GR, Bacon TE, Herbert J. Disability in multiple sclerosis: a reference for patients and clinicians. Neurology 2013;80:1018–1024. 24. Stephen H. Mobility device use in the United States. Disability statistics report 14 Appendix Authors MF01/PC03 Plus Postage [online]. Available at: dsc.ucsf.edu. Accessed August 1, 2019. 25. Ford C, Goodman AD, Johnson K, et al. Continuous long-term immunomodulatory Name Location Role Contribution therapy in relapsing multiple sclerosis: results from the 15-year analysis of the US prospective open-label study of glatiramer acetate. Mult Scler 2010;16:342–350. James M Department of Author Design or 26. Kappos L, Traboulsee A, Constantinescu C, et al. Long-term subcutaneous interferon Stankiewicz, Neurology, Brigham conceptualization of ’ beta-1a therapy in patients with relapsing-remitting MS. Neurology 2006;67: MD and Women s the review and – Hospital; Partners drafting or revising 944 953. 27. Ebers GC, Traboulsee A, Li D, et al. Analysis of clinical outcomes according to original Multiple Sclerosis the manuscript for Center; Harvard intellectual content treatment groups 16 years after the pivotal IFNB-1b trial. J Neurol Neurosurg Psy- – Medical School chiatry 2010;81:907 912. 28. Meyer-Moock S, Feng YS, Maeurer M, Dippel FW, Kohlmann T. Systematic literature Howard L Department of Author Design or review and validity evaluation of the Expanded Disability Status Scale (EDSS) and the Weiner, MD Neurology, Brigham conceptualization of Multiple Sclerosis Functional Composite (MSFC) in patients with multiple sclerosis. and Women’s the review BMC Neurol 2014;14:58. Hospital; Partners 29. Havrdová E, Arnold DL, Bar-Or A, et al. No evidence of disease activity (NEDA) Multiple Sclerosis analysis by epochs in patients with relapsing multiple sclerosis treated with ocreli- Center; Harvard zumab vs interferon beta-1a. Mult Scler J Exp Transl Clin 2018;4:2055217318760642. Medical School 30. Havrdova E, Galetta S, Hutchinson M, et al. Effect of natalizumab on clinical and radiological disease activity in multiple sclerosis: a retrospective analysis of the

6 Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 Neurology.org/NN natalizumab safety and efficacy in relapsing-remitting multiple sclerosis (AFFIRM) 47. Scalfari A, Neuhaus A, Daumer M, Muraro PA, Ebers GC. Onset of secondary pro- study. Lancet Neurol 2009;8:254–260. gressive phase and long-term evolution of multiple sclerosis. J Neurol Neurosurg 31. Hauser SL, Bar-Or A, Comi G, et al. Ocrelizumab versus interferon beta-1a in re- Psychiatry 2014;85:67–75. lapsing multiple sclerosis. N Engl J Med 2017;376:221–234. 48. Chalmer TA, Baggesen LM, Nørgaard M, Koch-Henriksen N, Magyari M, Sorensen 32. Coles AJ, Twyman CL, Arnold DL, et al. Alemtuzumab for patients with relapsing PS. Early versus later treatment start in multiple sclerosis: a register-based cohort multiple sclerosis after disease-modifying therapy: a randomised controlled phase 3 study. Eur J Neurol 2018;25:1262–e110. trial. Lancet 2012;380:1829–1839. 49. Weideman AM, Tapia-Maltos MA, Johnson K, Greenwood M, Bielekova B. Meta- 33. Cohen JA, Coles AJ, Arnold DL, et al. Alemtuzumab versus interferon beta 1a as first- analysis of the age-dependent efficacy of multiple sclerosis treatments. Front Neurol line treatment for patients with relapsing-remitting multiple sclerosis: a randomised 2017;8:577. controlled phase 3 trial. Lancet 2012;380:1819–1828. 50. Harding K, Williams O, Willis M, et al. Clinical outcomes of escalation vs early 34. Cohen JA, Barkhof F, Comi G, et al. Oral fingolimod or intramuscular interferon for intensive disease-modifying therapy in patients with multiple sclerosis. JAMA Neurol relapsing multiple sclerosis. N Engl J Med.2010;362:402–415. 2019;76:536–541. 35. Fox RJ, Miller DH, Phillips JT, et al. Placebo-controlled phase 3 study of oral BG-12 or 51. Correale J, Gaitán MI, Ysrraelit MC, Fiol MP. Progressive multiple sclerosis: from glatiramer in multiple sclerosis. N Engl J Med 2012;367:1087–1097. pathogenic mechanisms to treatment. Brain 2017;140:527–546. 36. Ho PR, Koendgen H, Campbell N, Haddock B, Richman S, Chang I. Risk of 52. Krieger SC, Cook K, De Nino S, Fletcher M. The topographical model of multiple natalizumab-associated progressive multifocal leukoencephalopathy in patients with sclerosis. Neurol Neuroimmunol Neuroinflamm 2016;3:e279. doi:10.1212/NXI. multiple sclerosis: a retrospective analysis of data from four clinical studies. Lancet 0000000000000279. Neurol 2017;16:925–933. 53. Keegan BM, Kaufmann TJ, Weinshenker BG, et al. Progressive motor impairment 37. Zhovtis Ryerson L, Frohman TC, Foley J, et al. Extended interval dosing of natali- from a critically located lesion in highly restricted CNS-demyelinating disease. Mult zumab in multiple sclerosis. J Neurol Neurosurg Psychiatry 2016;87:885–889. Scler 2018;24:1445–1452. 38. Berger JR, Malik V, Lacey S, Brunetta P, Lehane PB. Progressive multifocal leu- 54. Confavreux C, Vukusic S, Moreau T, Adeleine P. Relapses and progression of dis- koencephalopathy in rituximab-treated rheumatic diseases: a rare event. J Neurovirol ability in multiple sclerosis. N Engl J Med 2000;343:1430–1438. 2018;24:323–331. 55. Brown JWL, Coles A, Horakova D, et al. Association of initial disease-modifying 39. Barra ME, Soni D, Vo KH, Chitnis T, Stankiewicz JM. Experience with long-term therapy with later conversion to secondary progressive multiple sclerosis. JAMA 2019; rituximab use in a multiple sclerosis clinic. Mult Scler J Exp Transl Clin 2016;2: 321:175–187. 2055217316672100. 56. Cohen M, Maillart E, Tourbah A, et al. Switching from natalizumab to fingolimod in 40. Salzer J, Svenningsson R, Alping P, et al. Rituximab in multiple sclerosis. A retro- multiple sclerosis: a French prospective study. JAMA Neurol 2014;71:436–441. spective observational study on safety and efficacy. Neurology 2016;87:2074–2081. 57. Cree BAC, Gourraud PA, Oksenberg JR, et al. Long-term evolution of multiple 41. Juanatey A, Blanco-Garcia L, Tellez N. Ocrelizumab: its efficacy and safety in multiple sclerosis disability in the treatment era. Ann Neurol 2016;80:499–510. sclerosis [in Spanish]. Rev Neurol 2018;66:423–433. Available at: ncbi.nlm.nih.gov/ 58. Rotstein DL, Healy BC, Malik MT, Chitnis T, Weiner HL. Evaluation of no evidence pubmed/29897610. Accessed July 24, 2019. of disease activity in a 7-year longitudinal multiple sclerosis cohort. JAMA Neurol 42. Bottomley C, Lloyd A, Bennett G, Adlard N. A discrete choice experiment to de- 2015;72:152–158. termine UK patient preference for attributes of disease modifying treatments in 59. Harel A, Sperling D, Petracca M, et al. Brain microstructural injury occurs in patients Multiple Sclerosis. J Med Econ 2017;20:863–870. with RRMS despite “no evidence of disease activity”. J Neurol Neurosurg Psychiatry 43. Reen GK, Silber E, Langdon DW. Multiple sclerosis patients’ understanding and 2018;89:977–982. preferences for risks and benefits of disease-modifying drugs: a systematic review. 60. de Flon P, Gunnarsson M, Laurell K, et al. Reduced inflammation in relapsing- J Neurol Sci 2017;375:107–122. remitting multiple sclerosis after therapy switch to rituximab. Neurology 2016;87: 44. Halpern R, Agarwal S, Dembek C, Borton L, Lopez-Bresnahan M. Comparison of 141–147. adherence and persistence among multiple sclerosis patients treated with disease- 61. Cantó E, Barro C, Zhao C, et al. Association between serum neurofilament light chain modifying therapies: a retrospective administrative claims analysis. Patient Prefer levels and long-term disease course among patients with multiple sclerosis followed Adherence 2011;5:73–84. up for 12 years. JAMA Neurol Epub 2019 Aug 12. 45. Johnson KM, Zhou H, Lin F, Ko JJ, Herrera V. Real-world adherence and persistence 62. Leppert D, Kuhle J. Blood neurofilament light chain at the doorstep of clinical ap- to oral disease-modifying therapies in multiple sclerosis patients over 1 year. J Manag plication. Neurol Neuroimmunol Neuroinflamm 2019;6:e599. doi: 10.1212/NXI. Care Spec Pharm 2017;23:844–852. 0000000000000599. 46. Wang G, Marrie RA, Salter AR, et al. Health insurance affects the use of disease- 63. Burmester GR, Pope JE. Novel treatment strategies in rheumatoid arthritis. Lancet modifying therapy in multiple sclerosis. Neurology 2016;87:365–374. 2017;389:2338–2348.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 7, Number 1 | January 2020 7 CORRECTION Identification of circulating MOG-specific B cells in patients with MOG antibodies Neurol Neuroimmunol Neuroinflamm 2020;7:e647. doi:10.1212/NXI.0000000000000647

In the article “Identification of circulating MOG-specific B cells in patients with MOG antibodies” by Winklmeier et al.,1 first published online October 14, 2019, on the right side panel in figure 2a, the labelling of patient 15 should not have been included while the labelling of patient 26b should have been included. The corrected figure appears below. The publisher and the authors regret the error.

Figure

Reference 1. Winklmeier S, Schlüter M, Spadaro M, et al. Identification of circulating MOG-specific B cells in patients with MOG antibodies. Neurol Neuroimmunol Neuroinflamm 2019;6:e625. doi: 10.1212/NXI.0000000000000625.

Copyright © 2019 American Academy of Neurology 1 Copyright © 2019 American Academy of Neurology. Unauthorized reproduction of this article is prohibited. CORRECTION Cholecalciferol in relapsing-remitting MS: A randomized clinical trial (CHOLINE) Neurol Neuroimmunol Neuroinﬂamm 2020;7:e648. doi:10.1212/NXI.0000000000000648

In the article “Cholecalciferol in relapsing-remitting MS: A randomized clinical trial (CHOLINE)” by Camu et al.,1 ﬁrst published online August 6, 2019, a coinvestigator appendix should have been included, and it is printed below. The authors regret the error.

Coinvestigators are listed at links.lww.com/NXI/A168.

Reference 1. Camu W, Lehert P, Pierrot-Deseilligny C, et al. Cholecalciferol in relapsing-remitting MS: a randomized clinical trial (CHOLINE). Neurol Neuroimmunol Neuroinﬂamm 2019;6:e597. doi: 10.1212/NXI.0000000000000597.

Copyright © 2019 American Academy of Neurology 1 Copyright © 2019 American Academy of Neurology. Unauthorized reproduction of this article is prohibited. CORRECTION IL6 receptor358Ala variant and trans-signaling are disease modiﬁers in amyotrophic lateral sclerosis Neurol Neuroimmunol Neuroinﬂamm 2020;7:e650. doi:10.1212/NXI.0000000000000650

In the article “IL6 receptor358Ala variant and trans-signaling are disease modifiers in amyotrophic lateral sclerosis” by Wosiski-Kuhn et al.,1 first published online October 14, 2019, the y-axis label in figure 1C should be “Serum sIL6R (ng/mL).” In figure 1E, the y-axis label should be “Serum sgp130 (ng/mL).” In figure 2A, the y-axis label should be “Serum sIL6R (ng/mL).” The editorial office regrets the errors.

Reference 1. Wosiski-Kuhn M, Robinson M, Strupe J, et al. IL6 receptor358Ala variant and trans-signaling are disease modiﬁers in amyotrophic lateral sclerosis. Neurol Neuroimmunol Neuroinﬂamm 2019;6:e631. doi: 10.1212/NXI.0000000000000631.

Copyright © 2019 American Academy of Neurology 1 Copyright © 2019 American Academy of Neurology. Unauthorized reproduction of this article is prohibited. CORRECTION A case of idiopathic multicentric Castleman disease in an alemtuzumab-treated patient with MS Neurol Neuroimmunol Neuroinﬂamm 2020;7:e657. doi:10.1212/NXI.0000000000000657

In the Clinical/Scientific Note “A case of idiopathic multicentric Castleman disease in an alemtuzumab-treated patient with MS” by Rolfes et al.,1 first published online November 8, 2019, the label in figure 1F should read “Abdominal CT scan outlining extensive abdominal (yellow arrows) and inguinal (red arrows) lymphadenopathy.” The authors regret the error.

Reference 1. Rolfes L, Pfeuﬀer S, Ruck T, et al. A case of idiopathic multicentric Castleman disease in an alemtuzumab-treated patient with MS. Neurol Neuroimmunol Neuroinﬂamm 2020;7:e638. doi: 10.1212/NXI.0000000000000638.