Plasticity-Related Gene 3 (LPPR1) and Age at Diagnosis of Parkinson Disease E271

Volume 4, Number 5, October 2018 Neurology.org/NG

A peer-reviewed clinical and translational neurology open access journal

ARTICLE Plasticity-related gene 3 (LPPR1) and age at diagnosis of Parkinson disease e271

ARTICLE Increased KCNJ18 promoter activity as a mechanism in atypical normokalemic periodic paralysis e274

ARTICLE Protein network analysis reveals selectively vulnerable regions and biological processes in FTD e266

ARTICLE Bioenergetics in fi broblasts of Huntington disease patients are associated with age of onset e275 Academy Officers Neurology® is a registered trademark of the American Academy of Neurology (registration valid in the United States). Ralph L. Sacco, MD, MS, FAAN, President Neurology® Genetics (eISSN 2376-7839) is an open access journal published James C. Stevens, MD, FAAN, President Elect online for the American Academy of Neurology, 201 Chicago Avenue, Ann H. Tilton, MD, FAAN, Vice President Minneapolis, MN 55415, by Wolters Kluwer Health, Inc. at 14700 Citicorp Drive, Bldg. 3, Hagerstown, MD 21742. Business offices are located at Two Carlayne E. Jackson, MD, FAAN, Secretary Commerce Square, 2001 Market Street, Philadelphia, PA 19103. Production offices are located at 351 West Camden Street, Baltimore, MD 21201-2436. Janis M. Miyasaki, MD, MEd, FRCPC, FAAN, Treasurer © 2018 American Academy of Neurology. Terrence L. Cascino, MD, FAAN, Past President Neurology® Genetics is an official journal of the American Academy of Neurology. Journal website: Neurology.org/ng, AAN website: AAN.com Executive Office, American Academy of Neurology Copyright and Permission Information: Please go to the journal website (www.neurology.org/ng) and click the Permissions tab for the relevant Catherine M. Rydell, CAE, Executive Director/CEO article. Alternatively, send an email to [email protected]. 20l Chicago Ave General information about permissions can be found here: https://shop.lww.com/ journal-permission. Minneapolis, MN 55415 Disclaimer: Opinions expressed by the authors and advertisers are not Tel: 612-928-6000 necessarily those of the American Academy of Neurology, its affiliates, or of the Publisher. The American Academy of Neurology, its affiliates, and the Publisher disclaim any liability to any party for the accuracy, completeness, Editorial Office efficacy, or availability of the material contained in this publication (including drug dosages) or for any damages arising out of the use Patricia K. Baskin, MS, Executive Editor or non-use of any of the material contained in this publication. Kathleen M. Pieper, Senior Managing Editor, Neurology Advertising Sales Representatives: Wolters Kluwer, 333 Seventh Avenue, Lee Ann Kleffman, Managing Editor, Neurology: Genetics New York, NY 10001. Contacts: Eileen Henry, tel: 732-778-2261, fax: 973-215- 2485, [email protected] and in Europe: Avia Potashnik, Wolters Sharon L. Quimby, Managing Editor, Neurology® Clinical Practice Kluwer, tel: +44 207 981 0722; +44 7919 397 933 or e-mail: avia.potashnik@ Morgan S. Sorenson, Managing Editor, Neurology® Neuroimmunology & Neuroinflammation wolterskluwer.com. Careers & Events: Monique McLaughlin, Wolters Kluwer, Two Commerce Cynthia S. Abair, MA, Senior Graphics Editor Square, 2001 Market Street, Philadelphia, PA 19103, tel: 215-521-8468, fax: 215- Andrea R. Rahkola, Production Editor, Neurology 521-8801; [email protected]. Robert J. Witherow, Senior Editorial Associate Reprints: Meredith Edelman, Commercial Reprint Sales, Wolters Kluwer, Two Commerce Square, 2001 Market Street, Philadelphia, PA 19103, tel: 215-356-2721; Karen Skaja, Senior Editorial Associate [email protected]; [email protected]. Kaitlyn Aman Ramm, Editorial Assistant Special projects: US & Canada: Alan Moore, Wolters Kluwer, Two Kristen Swendsrud, Editorial Assistant Commerce Square, 2001 Market Street, Philadelphia, PA 19103, tel: 215-521-8638, [email protected]. International: Andrew Andrea Willgohs, Editorial Assistant Wible, Senior Manager, Rights, Licensing, and Partnerships, Wolters Kluwer; [email protected]. Publisher Wolters Kluwer Baltimore, MD

Publishing Staff Kim Jansen, Executive Publisher Jessica Heise, Production Team Leader, Neurology Journals Megen Miller, Production Editor Steve Rose, Editorial Assistant Stacy Drossner, Production Associate A peer-reviewed clinical and translational neurology open access journal Neurology.org/NG

Neurology® Genetics

Editor Stefan M. Pulst, MD, Dr med, FAAN Vision Neurology®: Genetics will be the premier peer- reviewed journal in the ﬁeld of neurogenetics. Deputy Editor Massimo Pandolfo, MD, FAAN Mission Neurology: Genetics will provide neurologists Associate Editors with outstanding original contributions that Alexandra Durr, MD, PhD elucidate the role of genetic and epigenetic Margherita Milone, MD, PhD variation in diseases and biological traits of Raymond P. Roos, MD, FAAN the central and peripheral nervous systems. Jeﬀery M. Vance, MD, PhD

Editorial Board Editorial Tel: 612-928-6400 Hilary Coon, PhD Giovanni Coppola, MD Inquiries Toll-free: 800-957-3182 (US) ChantalDepondt, MD, PhD Fax: 612-454-2748 Brent L. Fogel, MD, PhD, FAAN [email protected] AnthonyJ. Griswold, PhD Orhun H. Kantarci, MD Julie R. Korenberg, PhD, MD Stay facebook.com/NeurologyGenetics MargheritaMilone, MD, PhD Connected Davide Pareyson, MD twitter.com/greenjournal Shoji Tsuji, MD,PhD DinekeS. Verbeek,PhD youtube.com/user/NeurologyJournal David Viskochil, MD,PhD JulianeWinkelmann, MD Juan I. Young, PhD

Neurology® Journals

Editor-in-Chief Classification of Evidence Robert A. Gross, MD, PhD, FAAN Review Team Melissa J. Armstrong, MD Deputy Editor Richard L. Barbano, MD,PhD, FAAN Bradford B. Worrall, MD, MSc, FAAN RichardM.Dubinsky,MD,MPH,FAAN Jeffrey J. Fletcher, MD, MSc Section Editors Gary M. Franklin, MD, MPH, FAAN David S. Gloss II, MD,MPH&TM Biostatistics John J. Halperin, MD,FAAN Richard J. Kryscio, PhD Jason Lazarou, MSc, MD Christopher A. Beck, PhD Steven R. Messé, MD, FAAN Sue Leurgans, PhD Pushpa Narayanaswami, MBBS, DM, FAAN fi Classi cation of Evidence Evaluations Alex Rae-Grant, MD Gary S. Gronseth, MD, 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 TABLE OF CONTENTS Volume 4, Number 5, October 2018 Neurology.org/NG

e274 Increased KCNJ18 promoter activity as a mechanism in atypical normokalemic periodic paralysis M. Soufi, V. Ruppert, S. Rinn´e, T. Mueller, B. Kurt, G. Pilz, A. Maieron, R. Dodel, N. Decher, and J.R. Schaefer Open Access

e275 Bioenergetics in ﬁbroblasts of patients with Huntington disease is associated with age at onset S.L. Gardiner, C. Milanese, M.W. Boogaard, R.A.M. Buijsen, M. Hogenboom, R.A.C. Roos, P.G. Mastroberardino, W.M.C. van Roon-Mom, and N.A. Aziz Open Access

e276 Early-onset Parkinson disease caused by a mutation in CHCHD2 and mitochondrial dysfunction R.G. Lee, M. Sedghi, M. Salari, A.-M.J. Shearwood, M. Stentenbach, A. Kariminejad, H. Goullee, O. Rackham, N.G. Laing, H. Tajsharghi, and A. Filipovska Open Access

Clinical/Scientific Notes

e267 Homozygosity for SCN4A Arg1142Gln causes congenital myopathy with variable disease expression C.K. Sloth, F. Denti, N. Schmitt, B.H. Bentzen, C. Fagerberg, J. Vissing, and D. Gaist Open Access

e268 MT-CYB deletion in an encephalomyopathy with hyperintensity of middle cerebellar peduncles A. Chaussenot, C. Rouzier, K. Fragaki, S. Sacconi, S. Ait-El-Mkadem, V. Paquis-Flucklinger, and S. Bannwarth Articles Open Access e265 Genetic landscape of pediatric movement disorders e269 Deoxyguanosine kinase mutation producing and management implications juvenile-onset mitochondrial myopathy D. Cordeiro, G. Bullivant, K. Siriwardena, A. Evans, J. Kobayashi, F.N.U. Komal, P.M. Moretti, and A.I. Shaibani R.D. Cohn, and S. Mercimek-Andrews Open Access Open Access e270 CADASIL aﬀecting a black African man e266 Protein network analysis reveals selectively vulnerable L. Vlok and N. Brey regions and biological processes in FTD Open Access L.W. Bonham, N.Z.R. Steele, C.M. Karch, C. Manzoni, E.G. Geier, N. Wen, A. Ofori-Kuragu, P. Momeni, J. Hardy, Z.A. Miller, C.P. Hess, e272 DRESS after IV phenytoin associated with P. Lewis, B.L. Miller, W.W. Seeley, S.E. Baranzini, R.S. Desikan, R. Ferrari, and J.S. Yokoyama, on behalf of the International cytochrome P450 CYP2C9*3 homozygosity FTD-Genomics Consortium (IFGC) M.S. Nissen and C.P. Beier Open Access Open Access e271 Plasticity-related gene 3 (LPPR1) and age at diagnosis e273 AP4S1 splice-site mutation in a case of spastic of Parkinson disease paraplegia type 52 with polymicrogyria Z.D. Wallen, H. Chen, E.M. Hill-Burns, S.A. Factor, C.P. Zabetian, and S. Carmona, C. Marecos, M. Amorim, A.C. Ferreira, C. Conceição, H. Payami J. Br´as, S.T. Duarte, and R. Guerreiro Open Access Open Access TABLE OF CONTENTS Volume 4, Number 5, October 2018 Neurology.org/NG

Correction e277 Conﬁrming TDP2 mutation in spinocerebellar ataxia autosomal recessive 23 (SCAR23)

Cover image Universal lymphadenopathy visualized by PET-CT in a patient with CYP2C9*3 homozygosity who developed DRESS after intravenous administered phenytoin. Stylized by Kaitlyn Aman Ramm, Neurology Editorial Assistant. See e272 ARTICLE OPEN ACCESS Genetic landscape of pediatric movement disorders and management implications

Dawn Cordeiro, BN, RN, Garrett Bullivant, BA, Komudi Siriwardena, MD, Andrea Evans, MD, MSc, Correspondence Jeff Kobayashi, MD, Ronald D. Cohn, MD, and Saadet Mercimek-Andrews, MD, PhD Dr. Mercimek-Andrews [email protected] Neurol Genet 2018;4:e265. doi:10.1212/NXG.0000000000000265 Abstract Objective To identify underlying genetic causes in patients with pediatric movement disorders by genetic investigations.

Methods All patients with a movement disorder seen in a single Pediatric Genetic Movement Disorder Clinic were included in this retrospective cohort study. We reviewed electronic patient charts for clinical, neuroimaging, biochemical, and molecular genetic features. DNA samples were used for targeted direct sequencing, targeted next-generation sequencing, or whole exome sequencing.

Results There were 51 patients in the Pediatric Genetic Movement Disorder Clinic. Twenty-five patients had dystonia, 27 patients had ataxia, 7 patients had chorea-athetosis, 8 patients had tremor, and 7 patients had hyperkinetic movements. A genetic diagnosis was confirmed in 26 patients, including in 20 patients with ataxia and 6 patients with dystonia. Targeted next- generation sequencing panels confirmed a genetic diagnosis in 9 patients, and whole exome sequencing identified a genetic diagnosis in 14 patients.

Conclusions We report a genetic diagnosis in 26 (51%) patients with pediatric movement disorders seen in a single Pediatric Genetic Movement Disorder Clinic. A genetic diagnosis provided either disease-specific treatment or effected management in 10 patients with a genetic diagnosis, highlighting the importance of early and specific diagnosis.

From the Division of Clinical and Metabolic Genetics (D.C., G.B., R.D.C., S.M.-A.), Department of Pediatrics, Toronto, Ontario, Canada; Department of Medical Genetics (K.S.), University of Alberta, Edmonton, Canada; Department of Pediatrics (A.E., J.K., R.D.C., S.M.-A.), University of Toronto; the Emergency Medicine Division (A.E.), Department of Paediatrics, The Hospital for Sick Children; Division of Neurology (J.K.), Department of Paediatrics, The Hospital for Sick Children,; Genetics and Genome Biology Program (R.D.C., S.M.-A.), Research Institute, The Hospital for Sick Children; and Institute of Medical Sciences (S.M.-A.), University of Toronto, Toronto, Ontario, Canada.

Funding information and disclosures are provided at the end of the article. Full disclosure form information provided by the authors is available with the full text of this article at Neurology.org/NG.

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 © 2018 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary GLUT1 = glucose transporter 1; HMG = 3-hydroxy-3-methylglutaryl; MCT = medium chain triglyceride; PDHC = pyruvate dehydrogenase complex.

Pediatric movement disorders are usually part of complex of dystonia, of epilepsy, and of cellular energetic, and mito- neurodevelopmental disorders and are associated with global chondrial genome or whole exome sequencing (always per- developmental delay, cognitive dysfunction, behavioral dis- formed as trio, patient and both parents) according to various orders, and/or epilepsy. Pediatric movement disorders are clinical molecular genetics laboratories’ methods. All variants classified into 2 categories, including hyperkinetic and hypo- reported from the clinical laboratories were annotated kinetic movements.1,2 Pediatric hyperkinetic movements, using the Alamut database for predictions of pathogenicity, defined as unwanted and excess movements, include dystonia, cross-species conservation of nucleotides and amino acid chorea, athetosis, myoclonus, tremor, tics, and stereotypies sequences. We used the recommendations from mutation based on the definition and classification of The Taskforce on nomenclature (hgvs.org/mutomen) to name variants. Childhood Movement Disorders.1 Pediatric hypokinetic American College of Medical Genetics and Genomics movements, described as a decrease in the number of guidelines for variant classification for the molecular genetic movements, is called hypokinetic-rigid syndrome or parkin- result were used for interpretation.7 sonism.2 The underlying etiology of pediatric movement disorders can be acquired or hereditary. Management of We reviewed electronic patient charts for types of pediatric movement disorders requires multidisciplinary approach movement disorders, clinical features accompanying pediatric ranging from physiotherapy, pharmacologic treatment, to movement disorders, neuroimaging features, and biochemical – deep brain stimulation.3 5 investigations. We entered all information into an Excel database. We reported treatment outcome and management Recent advances in molecular genetic investigations, in- implications of a genetic diagnosis in patients with pediatric cluding targeted next-generation sequencing panels for dys- movement disorders. tonia or parkinsonism and whole exome and genome sequencing research studies, have discovered various novel Standard protocol approvals, registrations, genes causing pediatric movement disorders as part of and patient consents neurodevelopmental disorders. Recently, the Movement This study was approved by the Research Ethics Board at The Disorder Society Task Force for Nomenclature of Genetic Hospital for Sick Children (Approval#1000054997). Movement Disorders made recommendations for nomenclature of genetic movement disorders.6 Results We aim to determine prevalence of genetic diagnoses in patients with pediatric movement disorders using genetic There were 51 patients seen in this Pediatric Genetic Move- investigations in this retrospective cohort study. We also aim ment Disorder Clinic and included in this study. De- to identify disease-specific treatment or management effects mographics, clinical features, and genetic diagnosis of all based on the patients’ genetic diagnoses in pediatric move- patients are listed in table 1. Pediatric movement disorders ment disorders. included dystonia in 25 patients (49%) (generalized in 21 patients, focal in 3 patients, and hemidystonia in 1 patient), ataxia in 27 patients (53%), chorea-athetosis in 7 patients Methods (13%), tremor in 8 patients (16%), and hyperkinetic movements in 7 patients (13%). Twenty-three patients (45%) had This retrospective cohort study was performed in a single more than one movement disorder in combination. Global Pediatric Genetic Movement Disorder Clinic at an academic developmental delay or cognitive dysfunction was present in pediatric health science center from January 1, 2014 through 39 patients (76%). History of developmental regression was December, 2016. Inclusion criteria were as follows: (1) one of reported in 4 patients (8%). Epilepsy was present in 21 the pediatric movement disorders; (2) referred or being fol- patients (41%). lowed in this Pediatric Genetic Movement Disorder Clinic for investigations of underlying genetic causes (referrals from There was an identifiable underlying genetic cause in 26 pediatric neurology, pediatric epilepsy, developmental pedi- patients (51%). The most common movement disorder was atrics, and genetic clinics). ataxia (53%) and a genetic diagnosis was confirmed in 74% of those patients. The second most common movement DNA samples of all patients were used for genetic inves- disorder was dystonia (49%), and a genetic diagnosis was tigations, depending on their phenotype, including targeted confirmed in 24% of those patients. In 2 siblings, based on direct sequencing, targeted next-generation sequencing panel their phenotype and history of consanguinity in their

2 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Table 1 Clinical features and genetic diagnosis of patients

Patient/sex/age/consanguinity/ affected parents (reference) Diagnosis Movement disorders Other clinical features

1/M/9 y 11 mo/no/no GLUT1 D Ataxia GDD, epilepsy

2/M/dec. 4.5 y/no/no SURF1 Leigh disease Ataxia, tremor Gross motor delay, FTT

3/F/11.5 y/no/no PDHC D Ataxia GDD, microcephaly, FTT

4/F/15.5 y/yes/no8 DNAJC19 DCMPA Ataxia, dystonia GDD, DCMP, FTT

5/F/10 y 2 mo/no/no ND3 mitochondrial encephalopathy Dystonia Learning disability, right hemiparesis

6/F/9 y/no/no PDHC E3 D Ataxia Fatigue, hypotonia, retinopathy, ptosis

7/M/16 y 2 mo/no/no NCL type 2 Choreathetosis, parkinsonism Regression

8/M/11 y 3 mo/no/no NCL type 2 Ataxia Regression, epilepsy

9/F/11 y/yes/no HMG CoA synthase 2 D Ataxia, dyskinesia GDD, hypotonia,

10/M/7 y 11 mo/yes/no Riboflavin transporter 2 D Ataxia GDD, SN-HL, hypotonia

11/F/13 y 3 mo/no/no9 SCN2A EE Ataxia, hyperkinesia GDD, epilepsy, hypotonia

12/M/17 y/no/no9 Christianson syndrome Ataxia, tremor GDD, epilepsy

13/M/7.5 y/no/no9 STXBP1 EE Ataxia GDD, ASD, epilepsy

14/F/5 y 9 mo/no/no9 STXBP1 EE Ataxia, tremor GDD, epilepsy

15/F/7.5 y/no/no9 STXBP1 EE Ataxia GDD, epilepsy

16/F/10.5 y/no/no STXBP1 EE Ataxia, tremor GDD, epilepsy

17/M/8 y 10 mo/no/no CACNA1A EE Ataxia, tremor GDD, epilepsy

18*/M/8 y/yes/no AT Ataxia GDD, hypotonia

19*/F/9.5 y/yes/no AT Ataxia GDD, hypotonia

20/F/11 y 2 mo/no/no KCNA2 EE Ataxia, tremor GDD, epilepsy

21/M/6 y 4 mo/no/no CAMTA1 ataxia Ataxia GDD, hypotonia

22/F/16 y 9 mo/no/no ATP1A3 alternating hemiplegia Px dystonia, ataxia, GDD, epilepsy

23/M/6 y 9 mo/no/no MCT8 transporter D Dystonia GDD, hypotonia, epilepsy

24/M/5 y 2 mo/no/no10 BCAP31 encephalopathy Dystonia, choreathetosis GDD, SN-HL, hypotonia

25/F/8 y 8 mo/no/no CTNNB1 encephalopathy Dystonia GDD, microcephaly

26/F/5.5 y/yes/no SLC13A5 EE Hyperkinesia GDD, epilepsy, microcephaly

Abbreviations: AT = ataxia telangiectasia; dec = deceased; D = deficiency; DCMPA = dilated cardiomyopathy ataxia; dis = disease; GDD = global developmental delay; GLUT1 = glucose transporter 1; HMG = 3-hydroxy-3-methylglutaryl; NCL = neuronal ceroid lipofuscinosis; PDHC = pyruvate dehydrogenase complex; Px = paroxysmal; SN-HL = sensorineural hearing loss. parents, we suspected ataxia telangiectasia and confirmed the variants were found in The Single Nucleotide Polymorphism diagnosis by direct sequencing of ATM.Intheremaining24 Database (dbSNP) as polymorphisms. The novel variants patients, targeted next-generation sequencing panels (5 were moderately or highly conserved across species and patients for epilepsy; 3 patients for cellular energetic; 1 pa- reported to be disease causing in Mutation Taster and/or tient for Leigh disease) or whole exome sequencing (n = 14) deleterious in Sorting Intolerant From Tolerant prediction or mitochondrial genome sequencing (n = 1) confirmed the programs. In silico analysis results of all variants are listed in genetic diagnoses. Ten patients (38%) had autosomal re- supplemental e-table (table e-1, links.lww.com/NXG/A65). cessive, 11 patients (42%) had autosomal dominant de novo, 4 patients had X-linked inheritance pattern, and 1 patient We summarized all genetic causes of pediatric movement had mitochondrial inheritance pattern. We identified 26 disorders and the type of their confirmatory genetic test result variants in 21 genes in 26 patients, including 14 novel and 12 in figure 1. Ten patients had one of the inherited metabolic – known likely pathogenic variants.8 16 None of the novel disorders, including glucose transporter 1 (GLUT1)

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 3 Figure 1 All patients with pediatric genetic movement disorders, their genetic diagnoses, and type of genetic investigations

Abbreviations: CMP = cardiomyopathy; EE = epileptic encephalopathy; GDD = global developmental delay; NCL = neuronal ceroid lipofuscinosis; TNGS = targeted next generation sequencing; WES = whole exome sequencing.

deficiency (n = 1), neuronal ceroid lipofuscinosis type 2 (n = associated encephalopathy (n = 1), and SLC13A5-associated 2), SURF1-associated Leigh disease (n = 1), X-linked pyru- epileptic encephalopathy (n = 1). vate dehydrogenase complex (PDHC) deficiency (n = 1), dilated cardiomyopathy ataxia syndrome (3- In one patient with GLUT1 deficiency, despite seizure onset methylglutaconic aciduria) due to homozygous DNAJC19 in the first year of life, whole exome sequencing confirmed likely pathogenic variant (n = 1),8 mitochondrial encepha- the diagnosis at the age of 7 years 8 months. One patient lopathy due to mitochondrial ND3 likely pathogenic variant with neuronal ceroid lipofuscinosis type 2 presented with (n = 1), riboflavin transporter type 2 deficiency by homo- seizures and slowly progressive developmental regression zygous likely pathogenic variant in SLC52A2 (n = 1), mi- from the age of 5 years. His seizures were well controlled tochondrial 3-hydroxy-3-methylglutaryl (HMG) CoA with levetiracetam, which was discontinued 3 years later with synthase 2 deficiency (n = 1), PDHC E3 deficiency (n = 1). no seizure recurrence. He was diagnosed by whole exome In the remaining 16 patients, one of the neurogenetic dis- sequencing at the age of 11 years. In one patient with STXBP1- orders were identified, including STXBP1-associated epi- associated epileptic encephalopathy and in one patient with leptic encephalopathy (n = 4) (3 patients were reported CACNA1A-associated epileptic encephalopathy, whole exome previously9), SCN2A-associated epileptic encephalopathy sequencing was requested due to ataxia and tremor in addition (n = 1),9 SLC9A6-associated intellectual disability (Chris- to seizures and global developmental delay and confirmed the tianson syndrome) (n = 1),9 CACNA1A-associated epileptic diagnosis at the age of 9.5 and 8 years, respectively. One encephalopathy (n = 1), ataxia telangiectasia (n = 2), patient with SLC13A5-associated epileptic encephalopathy KCNA2-associated epileptic encephalopathy (n = 1), underwent targeted next-generation sequencing panel for ep- CAMTA1-associated cerebellar ataxia (n = 1), ATP1A3-as- ilepsy, which did not have this gene included at the time of the sociated alternating hemiplegia and dystonia (n = 1), genetic test. Currently, some of the commercially available Allan-Herndon-Dudley syndrome (MCT8-specificthyroid targeted next-generation sequencing panels for epilepsy in- hormone cell membrane transporter deficiency; n = 1), clude all those 5 genes and would have been identified those BCAP31-associated encephalopathy, deafness, dystonia, and genetic disorders without application of whole exome cerebral hypomyelination syndrome (n = 1),10 CTNNB1- sequencing.

4 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG All patients underwent metabolic investigations, including Treatment outcome of patients with genetic pediatric plasma amino acids, acylcarnitine profile, total and free car- movement disorders are listed in table 2. There were 10 nitine, homocysteine, urine organic acids, or urine guanidi- patients (38% of patients with a genetic diagnosis and 20% of noacetate and creatine to creatinine ratio (only in patients with all patients) with disease-specific or symptomatic treatment global developmental delay or cognitive dysfunction with no due to their confirmed genetic diagnosis. Because of late di- brain magnetic resonance spectroscopy). Eighteen patients agnosis and/or compliance problems, patients with GLUT1 (35% of all patients) underwent CSF metabolite measure- (n = 1) and riboflavin transporter type 2 (n = 1) deficiencies ments. There was no suspected diagnosis in any of those did not achieve normal neurodevelopmental outcome at the patients. In two patients with neuronal ceroid lipofuscinosis time of this study. A genetic diagnosis guided symptomatic type 2, blood dot spot tripeptidyl peptidase 1 activity, a non- medical treatment in 4 patients (2 with neuronal ceroid lip- invasive biochemical investigation, was measured after the ofuscinosis type 2; 1 with CACNA1A-associated epileptic genetic diagnosis of neuronal ceroid lipofuscinosis type 2. encephalopathy, and 1 with Allan-Herndon-Dudley syndrome [MCT8-specific thyroid hormone cell membrane Brain MRI was abnormal in 20 of 24 patients with a genetic transporter deficiency]). A genetic diagnosis guided pre- diagnosis (2 patients had no brain MRI). Abnormal brain ventative or sick day management or cancer screening for MRI of 8 patients is demonstrated in figure 2. Abnormal early detection in 4 patients (in HMG CoA synthase 2 de- brain MRI of 2 patients (patient 48 and patient 2410)were ficiency, ATP1A3-associated alternating hemiplegia of child- reported previously. hood and ataxia telangiectasia).

Figure 2 Specific brain MRI findings of 8 patients

(A) Brain MRI of patient 2 with Leigh disease shows increased T2 signal in the subthalamic nuclei and brain stem in axial image at the age of 3 years. (B) Brain MRI of patient 3 with PDHC deficiency shows dysgenesis of corpus callosum and in T1 sagittal image and dilated ventricles in T2 axial image at the age of 4 years. (C) Brain MRI of patient 5 with mitochondrial encephalopathy shows increased fluid-attenuated inversion recovery signal in bilateral putamen and left caudate head and body at the age of 6.5 years. (D) Brain MRI of patient 6 with PDHC E3 deficiency shows increased T2 signal in bilateral globus pallidi in axial image at the age of 3.5 years. (E) Brain MRI of patient 7 with neuronal ceroid lipofuscinosis type 2 shows increased T2 signal in cerebral white matter, cerebral atrophy and small thalami in axial image, and thin corpus callosum in T1 sagittal image at the age of 13 years. (F) Brain MRI of patient 8 with neuronal ceroid lipofuscinosis type 2 shows increased T2 white matter signal intensity and decreased T2 thalami signal intensity in axial image and cerebellar atrophy in T1 sagittal image at the age of 11 years. (G) Brain MRI of patient 9 with HMG CoA synthase 2 deficiency shows symmetrical increased signal intensity in putamen and caudate nucleus in T2 axial image at the age of 5 years. (H) Brain MRI of patient 20 with KCNA2-associated epileptic encephalopathy shows cerebellar atrophy in T1 sagittal image at the age of 10 years. Abbreviations: HMG = 3-hydroxy-3-methylglutaryl; PDHC = pyruvate dehydrogenase complex.

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 5 Table 2 Treatment and treatment outcome of patients

Patients/diagnosis Treatment Management implications Outcome

1/GLUT1 D 15% MCT ketogenic diet Resolved seizures, ataxia GDD, cognitive dysfunction for 1 year, no AED

7/NCL type 2 L-dopa/carbidopa (5.2 mg/kg/d) Improved parkinsonism Cognitive dysfunction

9/HMG CoA synthase 2 D Sick day management No further hypoglycemia GDD, hyperkinetic movements

10/Riboflavin transporter 2 D Riboflavin 65 mg/kg/d Independent steps GDD, ataxia, sensorineural hearing loss

17/12/CACNA1A EE Acetazolamide (10 mg/kg/d) Improved tremor, ataxia GDD, epilepsy

18/13/Ataxia telangiectasiaa None Cancer surveillance GDD, ataxia

19/14/Ataxia telangiectasiaa None Cancer surveillance GDD, ataxia

22/31/ATP1A3 alternating None Management of triggers GDD, ataxia, epilepsy hemiplegia of childhood

23/43/Allan-Herndon-Dudley syndrome L-dopa/carbidopa (7/kg/d) Improved dystonia GDD, spasticity, epilepsy

25/49/CTNNB1 encephalopathy L-dopa/carbidopa (5.2/kg/d) Mildly improved dystonia GDD, microcephaly, spasticity

Abbreviations: AED = antiepileptic drugs; D = deficiency; EE = epileptic encephalopathy; FTT = failure to thrive; GDD = global developmental delay; GLUT1 = glucose transporter 1; HMG = 3-hydroxy-3-methylglutaryl; MCT = medium chain triglycerides; N = normal; NCL = neuronal ceroid lipofuscinosis. a Siblings.

In the remaining 25 patients with no genetic diagnosis, 18 or deletion or duplication tests to confirm genetic diagnosis as patients had targeted next-generation sequencing panel for reported recently.17 A genetic diagnosis provided either dystonia, 1 patient had mitochondrial genome sequencing, disease-specific or symptomatic treatment in 38% of the and 10 patients had whole exome sequencing. Eight patients patients with pediatric movement disorders. Our study high- underwent both dystonia panel and whole exome sequencing. lights the importance of genetic investigations and effects of There was no genetic diagnosis in any of those patients. genetic diagnosis on the treatment and management of patients with pediatric movement disorders. Discussion To the best of our knowledge, there are 2 studies using targeted next-generation or whole exome sequencing to in- Our retrospective cohort study reports 51% prevalence of vestigate underlying genetic causes of movement disorders in genetic diagnoses in patients with pediatric movement dis- combined pediatric and adult patient populations.18,19 In the orders from a single Pediatric Genetic Movement Disorder first study, 61 patients with childhood and adult-onset dys- Clinic. There were 21 different genetic disorders combining tonia (age range, 1–73 years) were investigated for underlying inherited metabolic disorders (20% of all patients) and neu- genetic causes using targeted next-generation sequencing rogenetic disorders (31% of all patients) as underlying genetic dystonia panel of 94 genes. In that study, 9 of 61 patients causes of pediatric movement disorders. Only 4% of the (14.8%) had a genetic diagnosis, including PRKRA-associated patients were diagnosed by targeted direct sequencing based dystonia, tyrosine hydroxylase deficiency, glutaric aciduria on a clinical suspicion of ataxia telangiectasia. Targeted next- type I, Niemann-Pick type C, PRRT2-associated paroxysmal generation sequencing panels, including epilepsy, cellular en- kinesigenic dyskinesia (three patients), and Rett syndrome ergetic, and Leigh disease, confirmed a genetic diagnosis in (two patients).18 In the second study, 50 patients with 18% of the patients (9/51). Whole exome sequencing iden- movement disorders underwent whole exome sequencing, tified a genetic diagnosis in 27% of the patients (14/51). Mi- which targeted 151 genes, and the diagnostic yield was 20%. tochondrial genome sequencing confirmed a genetic diagnosis The type of movement disorders included hereditary spastic in 2% of the patients (1/51). Biochemical investigations did paraplegia (58.8%), cerebellar ataxia (23.5%), and dystonia not guide diagnoses in any of the patients. The most common (17.6%). Eighty percent of the patients with a genetic di- inheritance pattern was autosomal dominant or X-linked de agnosis had hereditary spastic paraplegia.19 The diagnostic novo inheritance. In pediatric movement disorders, we would yield of genetic investigations was higher in our retrospective recommend investigations in a stepwise approach for di- cohort study compared with the previous studies despite agnostic workup and depicted this approach in figure 3. Whole similar number of patients and genetic investigations. exome sequencing does not capture some of the intragenic rearrangements, whereas targeted next-generation sequencing In our retrospective cohort study, 41% of the patients had panels are designed with additional Sanger sequencing fill-ins epilepsy in addition to pediatric movement disorder. We

6 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Figure 3 Algorithm for diagnostic workup in pediatric movement disorders

confirmed a genetic diagnosis in 62% of these patients (13/21). or without seizures. Both patients were diagnosed more than 5 In a study, 400 patients with early-onset seizure disorders and/ years after their initial presentation by whole exome se- or severe developmental delay were investigated by targeted quencing. So far, less than 20 patients have been reported in next-generation sequencing of 46 epilepsy genes. Seventy-one the literature presenting with attenuated or juvenile onset – patients (18%) had a confirmed genetic diagnosis. Movement neuronal ceroid lipofuscinosis type 2.21 25 Pediatric move- disorders were reported in 4 patients (6%) with SCN2A-, ment disorders varied from isolated childhood onset pro- SCN8A-,andFOXG1-associated epilepsy.15 We think that we gressive ataxia22 to dystonia-parkinsonism.23 We think that will see increasing number of reports for overlapping pheno- history of developmental regression and progressive move- types of pediatric movement disorders and epilepsy in the ment disorder with or without seizures should prompt direct future because of application of genetic investigations and sequencing of TPP1 or measurement of tripeptidyl peptidase 1 identification of variable phenotypes for the same genotype. activity. Recently, intracerebroventricular enzyme replacement therapy has been approved for the treatment of neuronal Diagnostic yield of chromosomal microarray was reported in ceroid lipofuscinosis type 2.25 Early diagnosis and initiation of 25 patients with pediatric movement disorders, including this treatment will likely improve neurodevelopmental out- dystonia (n = 10), paroxysmal kinesigenic dyskinesia (n = 5), come of patients with neuronal ceroid lipofuscinosis type 2. tremor (n = 4), chorea (n = 3), myoclonus (n = 2), and paroxysmal non-kinesigenic dyskinesia (n = 1). Seven of 25 In our retrospective cohort study, basal ganglia or gray matter patients were diagnosed with a microdeletion (size ranged abnormalities with or without white matter changes were 160 kb to 3.27 Mb) as the potential cause of movement present in 67% of the patients (6/9) with inherited metabolic disorders. In 3 patients, microdeletion was inherited from an disorders. Whereas, 40% of the patients with neurogenetic affected parent.20 Of interest, chromosomal microarray was disorders had nonspecific white matter changes or delayed applied to 67% of the patients in our study cohort, and none of myelination. We think that, in patients with basal ganglia or the patients had any likely pathogenic copy number variants. gray matter changes in brain MRI, inherited metabolic dis- Despite the higher number of patients who underwent orders should be in the differential diagnosis. chromosomal microarray in our study compared with previous study, none of the patients with likely pathogenic copy Despite high prevalence of genetic diagnosis in our retro- number variants in chromosomal microarray raises the spective cohort study, we have some limitations, including: (1) question of whether chromosomal microarray should be used our clinic is not a referral clinic for patients with isolated as a first-line investigation in all patients with pediatric dystonia, chorea, myoclonus, tics, tremor, stereotypic move- movement disorders. We would recommend applying chro- ments, or transient movement disorders; (2) whole exome mosomal microarray to patients with dysmorphic features, sequencing was not a first-line genetic testing because of developmental disability, epilepsy, and movement disorders limited resources, despite the majority of the patients un- as complex neurodevelopmental disorders. derwent whole exome sequencing; (3) we will not be able to know, if the diagnostic yield would have been higher, if all Of interest, we report 2 new patients with attenuated neuronal patients would have undergone whole exome sequencing as ceroid lipofuscinosis type 2, who presented with childhood- first-line genetic testing; (4) we will not be able to know, if the onset developmental regression and movement disorder with diagnostic yield would have been lower, if all patients with

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 7 various movement disorders would have been investigated for travel/speaker honoraria from Biomarin; serves on the edi- genetic causes. Despite all these limitations, we think that the torial board for the Journal of Pediatric Genetics; has received number of patients and genetic diagnosis confirmed in our industry funding for work on the prevalence of mucopoly- study population is still valuable to draw conclusions regarding saccharidosis in rheumatology and nerve conduction velocity the diagnostic yield of targeted next-generation sequencing or study patients. D. Cordeiro, G. Bullivant, A. Evans, and J. whole exome sequencing in pediatric movement disorders. Kobayashi report no disclosures. Full disclosure form information provided by the authors is available with the full Our retrospective cohort study reports high prevalence of text of this article at Neurology.org/NG. genetic diagnoses and 21 different genetic disorders showing genetic landscape of pediatric movement disorders. Targeted Received February 10, 2018. Accepted in final form May 8, 2018. next-generation sequencing panels and whole exome se- References quencing increased diagnostic yield to more than 40% in our 1. Sanger TD, Chen D, Fehlings DL, et al. Definition and classification of hyperkinetic study. A genetic diagnosis provided either disease-specificor movements in childhood. Mov Disord 2010;25:1538–1549. 2. Garcia-Cazorla A, Duarte ST. Parkinsonism and inborn errors of metabolism. J Inherit symptomatic therapy in 38% of the patients with a genetic Metab Dis 2014;37:627–642. diagnosis, highlighting the importance of genetic inves- 3. Kurian MA, Dale RC. Movement disorders presenting in childhood. Continuum tigations to confirm underlying genetic cause in patients with (Minneap Minn) 2016;22:1159–1185. 4. Koy A, Lin JP, Sanger TD, Marks WA, Mink JW, Timmermann L. Advances in pediatric movement disorders. management of movement disorders in children. Lancet Neurol 2016;15:719–735. 5. Lumsden DE, Kaminska M, Ashkan K, Selway R, Lin JP. Deep brain stimulation for childhood dystonia: is “where” as important as in “whom”? Eur J Paediatr Neurol Author contributions 2017;21:176–184. D. Cordeiro: data analysis, arrangement of clinical genetic 6. Marras C, Lang A, van de Warrenburg BP, et al. Nomenclature of genetic movement fi disorders: recommendations of the international Parkinson and movement disorder tests, drafting the manuscript, and critique of the nal version society task force. Mov Disord 2016;31:436–457. of the manuscript. G. Bullivant: application of in silico tools to 7. Richards S, Aziz N, Bale S, et al. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of all variants in various genes using database, review and critique medical genetics and genomics and the association for molecular pathology. Genet of the final version of the manuscript. K. Siriwardena, Med 2015;17:405–424. 8. Al Teneiji A, Siriwardena K, George K, Mital S, Mercimek-Mahmutoglu S. Pro- A. Evans, J. Kobayashi, R. D. Cohn: review and critique of the gressive cerebellar atrophy and a novel homozygous pathogenic DNAJC19 variant final version of the manuscript. S. Mercimek-Andrews: design, as a cause of dilated cardiomyopathy ataxia syndrome. Pediatr Neurol 2016;62: 58–61. execution, organization and conceptualization of the study, 9. Mercimek-Mahmutoglu S, Patel J, Cordeiro D, et al. Diagnostic yield of genetic analysis and interpretation of the data, and drafting and testing in epileptic encephalopathy in childhood. Epilepsia 2015;56:707–716. finalizing the manuscript for intellectual content. 10. Albanyan S, Al Teneiji A, Monfared N, Mercimek-Mahmutoglu S. BCAP31- associated encephalopathy and complex movement disorder mimicking mitochondrial encephalopathy. Am J Med Genet A 2017;173:1640–1643. Acknowledgment 11. Lee IC, El-Hattab AW, Wang J, Li FY, et al. SURF1-associated Leigh syndrome: a case series and novel mutations. Hum Mutat 2012;33:1192–1200. The authors would like to thank Dr. Michal Inbar-Feigenberg, 12. Imbard A, Boutron A, Vequaud C, et al. Molecular characterization of 82 patients with Dr. Julian Raiman, Dr. Robert Munn, Dr. Sandra Farrell, and pyruvate dehydrogenase complex deficiency. Structural implications of novel amino acid substitutions in E1 protein. Mol Genet Metab 2011;104:507–516. Dr. Cynthia Forster-Gibson for referring their patients. The 13. Taylor RW, Singh-Kler R, Hayes CM, Smith PE, Turnbull DM. Progressive mito- authors would like to thank all staff physicians, including chondrial disease resulting from a novel missense mutation in the mitochondrial DNA ND3 gene. Ann Neurol 2001;50:104–107. general pediatrics, neurology, and clinical genetics at The 14. Shaag A, Saada A, Berger I, et al. Molecular basis of lipoamide dehydrogenase de- Hospital for Sick Children who referred their patients for ficiency in Ashkenazi Jews. Am J Med Genet 1999;82:177–182. 15. Trump N, McTague A, Brittain H, et al. Improving diagnosis and broadening the diagnostic workup to our clinic. The authors would like to phenotypes in early-onset seizure and severe developmental delay disorders through thank Stacy Hewson for genetic counseling for the genetic test gene panel analysis. J Med Genet 2016;53:310–317. 16. Syrbe S, Hedrich UBS, Riesch E, et al. De novo loss- or gain-of-function mutations in results. The authors would like to thank Ashley Wilson for her KCNA2 cause epileptic encephalopathy. Nat Genet 2015;47:393–399. work on the Research Ethics Board application for this study. 17. Breen DP, Mercimek-Andrews S, Lang AE. Infantile-onset hand dystonia with intellectual disability: clues to ARX mutations. Neurology 2018;90:333–335. The authors would like to thank Khadine Wiltshire for 18. van Egmond ME, Lugtenberg CHA, Brouwer OF, et al. A post hoc study on gene formatting this manuscript according to journal’s requirements. panel analysis for the diagnosis of dystonia. Mov Disord 2017;32:569–575. 19. Neveling K, Feenstra I, Gilissen C, et al. A post-hoc comparison of the utility of sanger sequencing and exome sequencing for the diagnosis of heterogeneous diseases. Hum Study funding Mutat 2013;34:1721–1726. No targeted funding reported. 20. Dale RC, Grattan-Smith P, Nicholson M, Peters GB. Microdeletions detected using chromosome microarray in children with suspected genetic movement disorders: a single-centre study. Dev Med Child Neurol 2012;54:618–623. Disclosure 21. Saini AG, Sankhyan N, Singhi P. Chorea in late-infantile neuronal ceroid lipofuscinosis: an Atypical presentation. Pediatr Neurol 2016;60:75–78. K. Siriwardena has received funding for travel/speaker hon- 22. Dy ME, Sims KB, Friedman J. TPP1 deficiency: rare cause of isolated childhood-onset oraria from Biomarin; has received research support from progressive ataxia. Neurology 2015;85:1259–1261. 23. Di Giacopo R, Cianetti L, Caputo V, et al. Protracted late infantile ceroid lip- Biomarin. R. D. Cohn has received research funding from the ofuscinosis due to TPP1 mutations: clinical, molecular and biochemical character- NIH and the Muscular Dystrophy Association. S. Mercimek- ization in three sibs. J Neurol Sci 2015;356:65–71. fi 24. Kousi M, Lehesjoki AE, Mole SE. Update of the mutation spectrum and clinical Andrews has served on scienti c advisory boards of The correlations of over 360 mutations in eight genes that underlie the neuronal ceroid Hospital for Sick Children, Biomarin, Cost Effectiveness of lipofuscinoses. Hum Mutat 2012;33:42–63. 25. Elleder M, Dvoráková L, Stolnaja L, et al. Atypical CLN2 with later onset and pro- ERT, and the Homocysteine and Betaine Treatment for longed course: a neuropathologic study showing different sensitivity of neuronal Recordati Rare Diseases Canada; has received funding for subpopulations to TPP1 deficiency. Acta Neuropathol 2008;116:119–124.

8 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG ARTICLE OPEN ACCESS Protein network analysis reveals selectively vulnerable regions and biological processes in FTD

Luke W. Bonham, BS, Natasha Z.R. Steele, MPH, Celeste M. Karch, PhD, Claudia Manzoni, PhD, Correspondence Ethan G. Geier, PhD, Natalie Wen, Aaron Ofori-Kuragu, Parastoo Momeni, PhD, John Hardy, PhD, Dr. Yokoyama [email protected] Zachary A. Miller, MD, Christopher P. Hess, MD, PhD, Patrick Lewis, PhD, Bruce L. Miller, MD, William W. Seeley, MD, Sergio E. Baranzini, PhD, Rahul S. Desikan, MD PhD, Raffaele Ferrari, PhD, and Jennifer S. Yokoyama, PhD, on behalf of the International FTD-Genomics Consortium (IFGC)

Neurol Genet 2018;4:e266. doi:10.1212/NXG.0000000000000266 Abstract Objective The neuroanatomical proﬁle of behavioral variant frontotemporal dementia (bvFTD) suggests a common biological etiology of disease despite disparate pathologic causes; we investigated the genetic underpinnings of this selective regional vulnerability to identify new risk factors for bvFTD.

Methods We used recently developed analytical techniques designed to address the limitations of genome-wide association studies to generate a protein interaction network of 63 bvFTD risk genes. We characterized this network using gene expression data from healthy and diseased human brain tissue, evaluating regional network expression patterns across the lifespan as well as the cell types and biological processes most aﬀected in bvFTD.

Results We found that bvFTD network genes show enriched expression across the human lifespan in vulnerable neuronal populations, are implicated in cell signaling, cell cycle, immune function, and development, and are differentially expressed in pathologically confirmed frontotemporal lobar degeneration cases. Five of the genes highlighted by our differential expression analyses, BAIAP2, ERBB3, POU2F2, SMARCA2, and CDC37, appear to be novel bvFTD risk loci.

Conclusions Our findings suggest that the cumulative burden of common genetic variation in an interacting protein network expressed in specific brain regions across the lifespan may influence susceptibility to bvFTD.

From the Department of Neurology (L.W.B., N.Z.R.S., E.G.G., Z.A.M., B.L.M., W.W.S., J.S.Y.), Memory and Aging Center, University of California, San Francisco; Department of Psychiatry (C.M.K., N.W., A.O.-K.), Washington University, St. Louis, MO; School of Pharmacy (C.M., P.L.), University of Reading, Reading, UK; Laboratory of Neurogenetics (P.M.), Texas Tech University Health Science Center, Lubbock; Department of Molecular Neuroscience (C.M., J.H., P.L., R.F.), UCL Institute of Neurology, London, UK; Department of Radiology and Biomedical Imaging, Neuroradiology Section (C.P.H., R.S.D.), University of California, San Francisco; and Department of Neurology (S.E.B.), University of California, San Francisco.

Copyright © 2018 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary AD = Alzheimer disease; BP = biological process; CSEA = cell-type SEA; EGFR = epidermal growth factor; FDR = false- discovery rate; bvFTD = behavioral variant frontotemporal dementia; FTLD = frontotemporal lobar degeneration; GO = gene ontology; GWAS = genome-wide association study; IFGC = International FTD-Genomics Consortium; MAPK = mitogen- activated protein kinase; PD = Parkinson disease; PDGF = platelet-derived growth factor; PINBPA = protein interaction network-based pathway analysis; SEA = speciﬁc enrichment analysis; VEN = Von Economo neuron; W-PPI-NA = weighted protein-protein interaction network analysis.

Although defined by a clear clinical syndrome, behavioral Consortium (IFGC), which was composed of 1,377 bvFTD variant frontotemporal dementia (bvFTD) is the most path- cases and 4,308 controls with genotypes or imputed data at ologically diverse form of frontotemporal lobar degeneration 6,026,384 SNPs. Raw genetic data for each participant was not (FTLD). Despite variable pathology, we and others have available for analysis at the inception of this study. Cases and shown that a shared, selectively vulnerable brain network, controls within the cohort were diagnosed according to composed of disparate brain regions functionally connected published criteria for bvFTD7 by a trained neurologist or throughout the lifespan, degenerates in bvFTD.1 The fact that pathologic examination.8 different protein pathologies converge on a relatively consistent set of neuroanatomical regions suggests intrinsic mo- Gene network generation lecular properties may predispose this network to disease. We calculated gene-level significance values using Versatile 9 Nevertheless, despite extensive advances in our ability to Gene-based Association Study (VEGAS) and summary data clinically and pathologically diagnose bvFTD, little is known from a recent bvFTD genome-wide association study 8 about the protein networks (“nexopathies”)thatdrivethebi- (GWAS). VEGAS assigns SNPs to their respective autosomal ological processes (BPs) underlying this selective vulnerability.2,3 genes using their position on the University of California, Santa Cruz Genome Browser. To capture the effects of regulatory 10 New techniques enable evaluation of experimentally determined regions and SNPs in linkage disequilibrium, gene boundaries protein networks in disease by aggregating single-nucleotide were defined as 50 kb beyond the 59 and 39 untranslated polymorphism (SNP)-level risk metrics across an entire gene. regions of each gene. Empirical p-values for each gene were 5,9 This technique increases statistical power, overcomes locus calculated using Monte-Carlo simulations. Association heterogeneity likely to occur in clinical populations, and allows blocks were defined as groups of sequential genes with p <0.05. for better detection of multiple variants contributing to polygenic disease risk.4 Combining this information with existing protein We used protein interaction network-based pathway analysis interaction data has revealed new genetic risk loci and helped (PINBPA) to compute first-order interactions by filtering a ref- unravel the pathophysiology of complex diseases like multiple erence network containing 8,960 proteins and 27,724 interactions sclerosis.5 so that only the genes (and their protein products) with VEGAS p < 0.05 were retained as significant. We evaluated network strength In this study, we first generated networks of genes underlying using simulations, which assigned p-values to genes at random bvFTD risk to identify common pathophysiologic processes from the parent network to create a simulated distribution of underlying disease biology. Second, based on prior studies similarly sized networks. The empirical bvFTD network’schar- showing that the brain network affected in bvFTD is present and acteristics were compared against the simulated networks. functioning throughout life,6 we evaluated spatial and temporal risk gene expression patterns in human brain across the lifespan. To increase resolution on the candidate proteins isolated by Finally, we sought to validate risk network genes using brain the PINBPA method, we generated a second independent tissue from patients pathologically diagnosed with FTLD. network with recently developed weighted protein-protein interaction network analysis (W-PPI-NA)11 using the 63 bvFTD genes prioritized by PINBPA as seeds. Of note, the Methods genes prioritized by the PINBPA analysis (CALM2 and Standard protocol approvals, registrations, CALM3) were combined into one unique entry, CALM1, and patient consents through this step. We removed TrEMBL, non-protein inter- Written informed consent was obtained from all participants actors (e.g., chemicals), obsolete Entrez, and Entrez matching or their guardians at the site of sample collection, and relevant to multiple Swiss-Prot identifiers. institutional review boards approved all aspects of the study. Profiling of gene expression during Genome-wide association study development and across the human lifespan study participants To assess whether genes from the PINBPA networks showed This study used SNP level metadata from phase 1 FTD- enriched expression in human brain tissues of all ages and life GWAS data from the International FTD-Genomics stages, we used the R package ABAEnrichment and specific

2 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG enrichment analysis (SEA). ABAEnrichment performs ontol- (g:GOSt, biit.cs.ut.ee/gprofiler/15) (62 genes were used be- ogy gene set enrichment analyses for 26 brain regions ranging cause CALM2 and CALM3 were combined into one entry as in age from 8 postconception weeks to 40 years old from the previously described). A Fisher’s one-tailed test was used to BrainSpan data set distributed by the Allen Brain Institute. assess enrichment; the set counts and sizes threshold was Forty-five percent of participants were Caucasian, 33% were applied as a multiple test correction; statistical domain size African American, 10% were Hispanic, 5% were mixed was only annotated genes; and no hierarchical filtering was Caucasian/African American, and 5% had no ethnicity in- included. We then grouped enriched GO-BP terms into formation available. All participants were screened for neu- custom-made “semantic classes.” We removed general, thus rologic conditions and for large-scale chromosomal negligible, semantic classes, such as general, metabolism, abnormalities. Samples were grouped by age into 5 groups: enzymes, protein modification, and physiology. Semantic prenatal (all prenatal stages; n = 20), infant (0–2 years; n = 6), classes were further grouped by similarity in more general child (3–11 years; n = 6), adolescent (12–19 years; n = 4), and classes called functional blocks. adult (older than 19 years; n = 6). Additional demographic data are provided in table e-1 (links.lww.com/NXG/A72). All Gene expression in pathologically confirmed genes evaluated in enrichment analyses were expressed above FTLD cases the 10th percentile when ranked by expression. SEA allows for We analyzed gene expression data from 2 cohorts of patho- expression testing across 6 brain regions and 10 different logically confirmed FTLD along with Alzheimer disease (AD) developmental periods. Fisher’s exact test was used to assess and Parkinson disease (PD) cases compared with normal older statistical enrichment, and false-discovery rate (FDR) cor- adult controls. The first FTLD cohort contained 10 individuals rection was applied. with pathologically confirmed sporadic FTLD, 7 individuals with pathologically confirmed FTLD due to GRN mutations, Cell-type-specific gene expression profiling and 11 pathologically confirmed controls (GSE13162). The We used 2 RNA expression data sets examining several cell second contained 6 FTLD cases with primarily tau pathology types commonly found in the CNS, one to conduct de- (Constantinides group A), 6 FTLD cases with primarily tem- scriptive representations of the cells that most highly express poral atrophy and minimal tau pathology (Constantinides our genes of interest and the other to specifically test for group C1), and 5 normal controls (E-MEXP-2280). The AD enrichment. cohort included temporal cortex samples from 84 AD and 80 control brains (syn5550404).16 The PD cohort included sub- The first data set used gene expression from human temporal stantia nigra samples from 22 PD and 23 control brains lobe samples.12 For each gene implicated in the bvFTD net- (GSE7621).17 In linear regression analyses, we analyzed FTLD work, we queried the database and recorded the cell type cases together in a combined analysis and then performed (fetal astrocytes, mature astrocytes, neurons, oligoden- separate subgroup analyses. Following this, we analyzed the AD drocytes, microglia/macrophages, and endothelial cells) that and PD groups separately. In all analyses, we covaried for age, most highly expressed the gene of interest. The data set sex, and postmortem interval. represents expression across the human lifespan (18 postconception weeks [pcw] for fetal samples and 8–63 years old for non-fetal samples).12 Results The second data set used RNAseq data from the BrainSpan bvFTD GWAS data reveals a large protein atlas to characterize cell-specific expression patterns. Enrich- interaction network driven by ment for specific cell types was tested using cell-type SEA risk-associated alleles (CSEA),13 which employs expression cutoffs and tests for Using gene-based p-values for 17,463 genes derived from the cell-type-specific enrichment (CSEA employs methodology bvFTD GWAS cohort, we identified 475 association blocks similar to SEA). Cell types are determined using well- containing a total of 1,104 genes. We next generated PINBPA validated markers,14 and enrichment is tested using Fisher’s networks for the genes implicated in bvFTD. The bvFTD exact test with FDR correction. network contained 63 nodes and 72 edges. Permutation testing revealed that the bvFTD gene network was within the Biological pathways and processes analysis top 10th percentile for nodes and edges as compared with We classified the biological function of bvFTD network genes a sample of 1,000 randomly generated networks (figure 1 and using PANTHER (pantherdb.org) and Reactome (reactome. table e-2, links.lww.com/NXG/A73). org) to reduce the likelihood of spurious findings from any one source. Statistical tests for overrepresentation with Bon- We then sought to analyze the topological details of the 63 ferroni correction were performed using the PANTHER PINBPA-prioritized genes by generating protein interaction pathways annotation and Gene Ontology (GO) BPs data- networks with our W-PPI-NA to highlight key functional bases. We applied functional annotation analysis to the net- players within the identified gene set. After filtering and work built using the 62 PINBPA-prioritized genes as seeds by scoring the protein network, the result was composed of 1,913 performing GO BP enrichment analyses through g:Profiler nodes and 3,212 edges, where all but one node, PBX2, were

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 3 Figure 1 bvFTD PINBPA network

The bvFTD PINBPA network is shown. Nodes are color coded according to their gene-based p-value, with warmer colors indicating lower p-values and cooler colors representing p-values closer to 0.05. Node size represents closeness centrality, a measure of a node’s nearness to other nodes within a network. Edge thickness represents edge betweenness, a measure of the number of paths that go through an edge in a network.

interconnected (figure 2, A). We identified the inter- neurodevelopmental processes.18 Given this, we performed interactome hubs (IIHs; n = 10)—the proteins character- exploratory analyses using SEA and found that there were ized by the highest connection degree distribution (figure 2, B several nominally associated regions of gene expression en- and C)—and we used them to define the core of the network richment concentrated within the fetal period. At a specificity with the highest interconnectivity (figure 2, D). By comparing index probability level of 0.05, the amygdala (late fetal; p = the core of the network with randomly sampled parts of the 0.004), thalamus (early fetal; p = 0.006), hippocampus (late network, we verified that the IIHs-driven network is the most fetal; p = 0.007), and cortex (mid-late fetal; p = 0.014) showed densely connected. The core of the network was made of 52 nominal enrichment levels but did not withstand FDR cor- nodes (10 IIHs and their interactors) and 121 edges; these rection (table e-4, links.lww.com/NXG/A75). were strongly interconnected (average number of neighbors = 4.5), representing the proteins that keep the cohesion of bvFTD network genes are most highly expressed >16% of the initial seeds (n = 63; figure 2, C). in fetal astrocytes, mature astrocytes, microglia, and cortical layer 5b neurons bvFTD network genes show enriched Analysis of whole transcriptome data from specific cell types expression in bvFTD-relevant brain regions isolated from human brain tissue12,19 revealed that the bvFTD across the lifespan network genes clustered into 3 main brain cell groupings. The We next tested whether genes within the bvFTD network most commonly observed cell type was fetal astrocytes (37%), showed enriched or specific expression in distinct brain followed closely by neurons (35%), and microglia (20%). regions of healthy individuals using ABAEnrichment. Hyper- bvFTD network genes were present at a lower frequency in geometric tests revealed enriched expression for our bvFTD mature astrocytes (5%), oligodendrocytes (5%), and endo- network genes in neuroanatomical regions affected in bvFTD, thelial cells (5%) (figure e-1, links.lww.com/NXG/A79). including the dorsolateral prefrontal cortex, frontal cortex, Network genes represented only 3 of 100 and 11 of 500 of the and temporal cortex in most age groups (family-wise error top-expressed genes in each cell type, suggesting that the cell corrected p < 0.05 for each region, figure 3, A–D and table e-3, type enrichment was specific to bvFTD and not a function of links.lww.com/NXG/A74). detecting genes that were most highly expressed in these cell types. IL2 and SLA2 were not expressed at a level that allowed Our ABAEnrichment analyses suggested that enrichment was for reliable cell type determination. most common during early life (table e-3, links.lww.com/ NXG/A74). However, they did not provide information on Given our findings in the cell-type-specific expression data set, the role of gene expression within the early life stages when we next used CSEA to test for enrichment of our bvFTD gene expression patterns can vary temporally and influence network genes in specific cell types and populations. The

4 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Figure 2 Interactome analyses in bvFTD

(A) Protein network built through the W-PPI-NA pipeline around the 63 genes prioritized by the PINBPA protocol. The seeds are depicted in pink and their interactors in blue. (B) Inter-interactome degree distribution curve reporting the number of nodes (x-axis) able to bridge an N number of seeds (y-axis). The IIHs are the nodes marked by the rectangle. (C) List of IIHs reporting the number of seeds that they bridge and associated %. *UBC has been reported yet ignored as it is likely a false positive (as it may indicate unspecific binding of ubiquitin to proteins tagged for degradation). (D) Core of the network around the IIHs, which are depicted in pink. bvFTD network genes demonstrated enriched expression in macrophage colony-stimulating factor [GM-CSF]) (table layer 5b neuronal cells (p = 0.038; figure e-1, links.lww.com/ e-6, links.lww.com/NXG/A77). PANTHER analysis found NXG/A79 and table e-5, links.lww.com/NXG/A76); how- similar overrepresentation for interleukin signaling path- ever, this association did not withstand FDR correction. Von ways, EGFR signaling, and platelet-derived growth factor Economo neurons (VENs) and fork cells, which are among (PDGF) signaling. the earliest cells to degenerate in bvFTD, are located within layer 5b.20 We evaluated additional GO BPs via g:Profiler for the second protein network built by W-PPI-NA and focused on the Ontological analyses reveal interactomes of the 63 genes prioritized through the PINBPA overrepresentation of immune signaling protocol. This first functional enrichment was followed by in bvFTD risk genes a second iteration performed only on the core of the Functional (ontological) characterization of the 63 bvFTD network—i.e., defined by the IIHs and their interactors—thus network genes indicated overrepresentation of immune- containing the most densely connected proteins. Functional mediated pathways, cell signaling, cell cycle, and de- annotation of the latter indicated a list of semantic classes that velopment (table 1). Reactome analysis indicated the were a subset of the former: of interest, the subset terms greatest degree of overrepresentation in pathways known to (percentage of retention >15%) pointed to the following be involved with microglia-initiated inflammatory respon- functional blocks: (1) “development”; (2) “motility”; (3) ses, such as epidermal growth factor (EGFR) signaling, “protein modification”; and (4) “cell signaling” (figure e-2, mitogen-activated protein kinases (MAPKs), and links.lww.com/NXG/A80). These appeared as the common pro-inflammatory cytokines (IL-3, IL-5, granulocyte- functions characterizing the part of the protein network with

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 5 Figure 3 Regional gene expression enrichment across the lifespan

For each of the 5 age groups used in the ABAEnrichment analyses, we counted the number of times each of the 26 available brain regions showed enrichment of the bvFTD network genes. The superior temporal cortex was the most common region of enriched expression with 5 observations (i.e., in every age grouping). Generally, the FTD cohort showed enriched expression in both frontal (DFC, VFC, M1C, PFC, OFC) and temporal (A1C, ITC, STC, MFC) regions. (A and B) The number of times each region was associated is illustrated on a model brain. Brain regions were mapped onto the illustration using information provided in Bahl et al. 2017. When regional overlap was detected (e.g., IPC is contained within PCx or DFC; VFC and OFC are contained within PFC), the more specific region(s) was chosen for presentation. (C) A graphical depiction of the data shown in (A and B). The number of times each brain region was associated from the 5 age groupings is shown. (D) Detailed results are shown in a table format with age groupings as rows and brain regions as columns. Regions shaded in black were statistically associated while unshaded regions were not. A1C = primary auditory cortex; AMY = amygdaloid complex; CN = cerebral nuclei;CB= cerebellum; CBC = cerebellar cortex; DFC = dorsolateral prefrontal cortex; HIP = hippocampus; IPC = posteroventral (inferior) parietal cortex; ITC = infero- lateral temporal cortex; M1C = primary motor cortex; MFC = anterior (rostral) cingulate (medial prefrontal) cortex; OFC = orbital frontal cortex; PCx = parietal neocortex; PFC = prefrontal cortex; S1C = primary somatosensory cortex; STC = posterior (caudal) superior temporal cortex; STR = striatum; TCx = temporal neocortex; THA = thalamus; V1C = primary visual cortex; VFC = ventrolateral prefrontal cortex.

strongest cohesion among the initial seeds. In particular, the autosomal dominant (GRN) postmortem cases (GSD13162) functional block “development” contained semantic classes or sporadic cases with tau and non-tau pathology (E-MEXP- referring to cell differentiation and proliferation as well as glial 2280), we analyzed data in aggregate and by gene status or and neuronal cell development, while the functional block pathologic subtype (table e-7, links.lww.com/NXG/A78). In “cell signaling” contained semantic classes referring to the GSE13162, 19 bvFTD genes were differentially expressed fi previously mentioned MAPK, PDGF, and immune-speci c between cases vs controls (PFDR < 0.05). In E-MEXP-2280, ff signaling in addition to terms related to Fc (typically relevant 26 bvFTD genes were di erentially expressed (PFDR < 0.05). for immune cells), ERBB (relevant for both developmental Of these top genes, 9 (BAIAP2, CALD1, CDC37, ERBB3, and immune processes), growth factors, and hormone-based GSK3A, MAP3K5, POU2F2, SMARCA2, and TGFB1|1) signaling events. These findings confirmed the PANTHER showed dysregulation in both cohorts (table 2 and table e-7). and Reactome results, using a completely independent Several of these genes appear to be novel risk factors for approach. bvFTD (table 2). In AD cases, 7 of 9 bvFTD network genes were differentially expressed (table 2). In contrast, 0 of 9 net- bvFTD network genes are dysregulated work genes were differentially expressed in PD cases (table 2). in postmortem brain tissue To examine whether the risk-associated genes identified by the network analysis are differentially expressed in human Discussion disease, we assessed the expression patterns of our bvFTD genes in postmortem samples from 2 pathologically con- We identified a network of interacting bvFTD risk genes that firmed cohorts of FTLD cases along with AD16 and PD17 demonstrate enriched expression in bvFTD-affected brain cases compared with pathology-free controls. In the first regions and cell types across the human lifespan. Differential FTLD cohort, GSE13162, 61 bvFTD genes had expression expression analyses in pathologically confirmed cases of data available. In the second FTLD cohort, E-MEXP-2280, FTLD and controls showed that many of our network genes there were 58 bvFTD genes with expression data available. were dysregulated in pathologically confirmed FTLD, and 5 Because the 2 data sets represent a mix of sporadic and potentially novel bvFTD risk genes showed altered expression

6 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Table 1 Pathway analysis in bvFTD network genes

No. of genes reference, Biological candidate Fold pathway Pathway data set Candidate genes mapped enriched p Value Database

Immune system

− Interleukin signaling pathway 98, 5 IL-2, MYC, SLA2, SHC1, IRS2 17.26 1.83 × 10 3 PANTHER

− Interleukin 3, 5, and GM-CSF 250, 8 IL-2, RET, GDNF, SHC1, SHC3, IRS2, LYN, CBL 10.82 1.38 × 10 3 Reactome

Fc gamma receptor 90, 5 FCGR2A, HSP90AA1, BAIAP2, FCGR3A, VAV2 18.79 0.01 Reactome phagocytosis

GP VI mediated activation 55, 4 IL-2, SHC1, LYN, VAV2 24.60 0.04 Reactome cascade

Cell signaling

− EGF receptor signaling pathway 139, 6 SHC1, SHC3, CBL, ERBB4, ERBB3, MAP3K5 11.61 5.99 × 10 4 PANTHER

− 14, 4 SHC1, CBL, CDC37, HSP90AA1 96.65 1.19 × 10 4 Reactome

ERBB2 43, 4 CDC37, HSP90AA1, SHC1, PTPN12 31.47 0.02 Reactome

− ERBB4 315, 8 IL2, RET, ESR1, GDNF, SHC1, SHC3, IRS2, GSK3A 8.59 7.55 × 10 3 Reactome

− VEGF signaling 307, 9 IL2, RET, GDNF, SHC1, SHC3, IRS2, HSP90AA1, 9.92 5.52 × 10 4 Reactome BAIAP2, VAV2

− MAPK signaling 229, 7 IL2, RET, TYK2, GDNF, SHC1, SHC3, IRS2 10.34 9.55 × 10 3 Reactome

− FCERI-mediated MAPK 276, 8 IL2, RET, GDNF, SHC1, SHC3, IRS2, LYN, VAV2 9.80 2.87 × 10 3 Reactome activation

DAP12 signaling 330, 8 IL2, RET, GDNF, SHC1, SHC3, IRS2, GSK3A, VAV2 8.20 0.01 Reactome

Cell cycle

− Cyclin A CDK2-associated 70, 5 MYC, CDK2, CDC25B, CKS1B, MNAT1 24.16 4.06 × 10 3 Reactome events at S phase entry

− G1/S transition 139, 6 MYC, CKS1B, MNAT1, CDK2, CDC6, CDK6 14.6 6.73 × 10 3 Reactome

Development/ growth

− FGFR signaling 318, 8 IL2, RET, GDNF, SHC1, SHC3, IRS2, GSK3A, CBL 8.51 8.09 × 10 3 Reactome

− Signaling by SCF-KIT 311, 9 IL2, RET, GDNF, SHC1, SHC3, IRS2, GSK3A, LYN, 9.79 6.14 × 10 4 Reactome CBL

Developmental biology 806, 11 VAV2, LYN, HIST2H2BE, HSP90AA1, IRS2, SHC3, 4.62 0.04 Reactome SHC1, NCOR2M, GDNF, RET, IL2

Platelet-derived growth factor 149, 5 MYC, SHC1, SHC3, GSK3A, VAV2 11.35 0.01 PANTHER signaling pathway

Abbreviations: bvFTD = behavioral variant frontotemporal dementia; DAP = DNAX-activation protein; EGF = epidermal growth factor; FCERI = Fc epsilon receptor; FGFR = fibroblast growth factor receptor; GM-CSF = granulocyte-macrophage colony-stimulating factor; GP VI = Glycoprotein VI; MAPK = mitogen- activated protein kinase; SCF = stem cell factor; VEGF = vascular endothelial growth factor. Pathway analysis results are shown. For each broad biological pathway, specific pathways from PANTHER and Reactome databases are provided. In all analyses, the p-value presented has been adjusted using the Bonferroni technique, which is the default setting for both PANTHER and Reactome. See table e-6 (links.lww.com/NXG/A77) for additional details. in both pathologic cohorts. Taken together, our results pro- susceptibility to neurodegenerative diseases like bvFTD. Our vide evidence that the pathologic heterogeneity seen in analyses suggest that this bvFTD risk gene network may bvFTD may occur as the product of multiple dysregulated overlap with AD, a related disorder sharing genetic risk fac- cellular pathways that converge on specific brain networks tors23 and tau pathology24 that can also present with across the lifespan. Our findings also contribute to the mo- a bvFTD-like syndrome.25 Additional analyses will be useful lecular understanding of selective vulnerability in bvFTD and to elucidate the specificity of these findings to bvFTD com- suggest several potential mechanisms by which genetic vari- pared with AD and determine any shared mechanisms con- ation and gene expression abnormalities increase ferring risk across other neurodegenerative diseases.

Table 2 Differentially expressed genes enriched in both FTLD pathologic cohorts

Chen-Plotkin et al.21 Bronner et al.22 Allen et al.16 and Lesnick et al.17

Combined analyses Split by GRN genotype Combined analyses Split by pathologic diagnosis Disease comparisons

Gene GRN + GRN + GRN 2 GRN 2 Type Type Type Type PD PD Disease symbol Bronner Bronner Praw PFDR Praw PFDR Praw PFDR C1 Praw C1 PFDR APraw APFDR AD Praw AD PFDR Praw PFDR associations

− − − − − − BAIAP2a 6.28 × 10 3 0.03 0.56 0.65 2.35 × 10 4 2.43 × 10 3 7.48 × 10 3 0.05 0.02 0.45 0.08 0.32 2.75 × 10 6 8.41 × 10 5 0.08 0.51 AD,34 autism, ADHD35

− − − − − − − CALD1 5.84 × 10 4 0.01 3.64 × 10 4 7.51 × 10 3 4.31 × 10 3 0.02 8.78 × 10 3 0.05 0.19 0.69 0.03 0.32 1.60 × 10 7 1.12 × 10 5 0.05 0.36 ALS, AD, FTLD36

− − − ERBB3a 3.54 × 10 3 0.02 2.87 × 10 3 0.02 0.02 0.06 7.50 × 10 3 0.05 0.22 0.7 0.07 0.32 0.64 0.72 0.10 0.54 PD37

− − − − − − − GSK3A 3.20 × 10 4 9.05 × 10 3 0.01 0.04 6.15 × 10 4 5.08 × 10 3 2.67 × 10 3 0.04 0.19 0.69 0.04 0.32 1.43 × 10 3 8.53 × 10 3 0.22 0.75 FTLD,38 AD39

− − − − − − MAP3K5 2.59 × 10 3 0.02 9.20 × 10 3 0.03 0.03 0.09 3.97 × 10 3 0.04 0.17 0.69 2.77 × 10 3 0.14 2.77 × 10 6 8.44 × 10 5 0.42 0.86 ALS, FTLD, Huntington40

− − − − − − − − POU2F2a 8.51 × 10 4 0.01 0.17 0.27 7.24 × 10 5 1.43 × 10 3 5.68 × 10 5 4.29 × 10 3 0.01 0.45 2.65 × 10 3 0.14 1.07 × 10 3 6.98 × 10 3 0.18 0.68 —

− − − − SMARCA2a 2.63 × 10 3 0.02 0.32 0.43 1.32 × 10 4 1.93 × 10 3 8.90 × 10 3 0.05 0.24 0.7 0.06 0.32 0.41 0.52 0.10 0.55 Nicolaides-Braitser syndrome41

− − − − − − CDC37a 0.01 0.05 5.31 × 10 4 7.51 × 10 3 0.19 0.33 4.83 × 10 5 4.29 × 10 3 2.97 × 10 3 0.45 0.01 0.31 2.28 × 10 3 0.01 0.61 0.91 AD42

− − − − − TGFB1I1 8.84 × 10 3 0.04 1.37 × 10 3 0.01 0.08 0.18 6.79 × 10 3 0.05 0.11 0.68 5.09 × 10 3 0.19 2.85 × 10 3 0.01 0.87 0.96 AD,43 FTD44

Abbreviations: AD = Alzheimer disease; ADHD = attention-deficit/hyperactivity disorder; ALS = amyotrophic lateral sclerosis; FDR = false-discovery rate; FTD = frontotemporal dementia; FTLD = frontotemporal lobar degeneration; PD = Parkinson disease. Results are shown for differential expression analyses in 2 pathologically confirmed FTLD cohorts along with AD and PD for comparison. Nine genes were differentially expressed in the 2 cohorts: BAIAP2, CALD1, CDC37, ERBB3, GSK3A, MAP3K5, POU2F2, SMARCA2, CDC37, and TGFB1I1. For each public data source, the publication reference is provided. a Indicates to the best of our knowledge not previously reported in FTLD. By integrating GWAS with protein interaction data, gene impetus for further study of bvFTD risk gene expression expression data from across the lifespan, and postmortem patterns in these specific cell populations. gene expression data from confirmed FTLD cases, we highlight the importance of bvFTD risk genes during de- Limitations of this study include lack of a suitable case-control velopment, aging, and disease. Our network genes showed data set of sufficient size to confirm whether our network and enrichment in regions of the brain commonly affected in risk loci are associated with disease risk and lack of access to bvFTD such as the dorsolateral and ventrolateral prefrontal raw IFGC genotype data to perform subgroup analyses. Phase cortices, and inferior temporal cortex (figure 3). Because of III data of the IFGC (on over 2,000 new cases) will be re- the scarcity of pathologic samples available for enrichment leased in the near future and that cohort could represent analyses, some regions of the brain that have known roles in a suitable cohort for validation of these findings and further bvFTD, such as the cingulate and insula, were not available for characterization efforts. Our data-sharing agreement with the analysis. Of interest, our genes also showed enrichment in IFGC provided SNP-level GWAS results only and we did not brain regions typically spared in bvFTD patients such as su- have raw genetic data, which precludes potentially more perior temporal cortex, occipital cortex, and sensory cortex. powerful study designs, such as subsetting the cohort into Although the reason for this enrichment remains to be de- training and testing data sets. Beyond this, our network gen- termined, it is possible that spared regions contain similar cell eration methodology relies on already existing protein in- types as vulnerable ones but remain unaffected because they teraction data and thus could bias our ontological findings are functionally disconnected from pathologically affected toward more thoroughly characterized biological pathways regions in bvFTD and remain unexposed to pathologic and processes. An additional limitation of our gene expression aggregates that may disseminate through functionally con- analyses stems from Constantinidis diagnoses, which are not nected neural networks.26,27 Alternatively, and perhaps easily converted into more modern pathologic diagnoses. counterintuitively, consistently enriched expression (and thereby greater expression across the lifespan) of bvFTD In our analyses, we identified and bioinformatically charac- network genes in regions like cerebellum and superior tem- terized a gene network linked at the protein level that shows poral cortex could potentially provide protection against enriched expression across the human lifespan in brain areas neurodegenerative processes. In this framework, reduced ex- most affected by bvFTD and shows dysregulation in patho- pression of the same genes in other, less-enriched, brain logically diagnosed FTLD cases. These findings suggest that regions may result in vulnerability to disease pathology be- multiple distinct biological pathways are altered in sporadic cause less-enriched brain regions have less “reserve” expres- bvFTD, including signal transduction, cell cycle regulation, sion to rely on during aging, potentially reducing the immune/inflammation, and neurodevelopmental processes. threshold within which variable expression is tolerated by Together, these pathway enrichments raise an important and relevant biological systems. unanswered question regarding the temporal relationship between genetic variation and biological risk for disease. It is An exploratory analysis revealed that our bvFTD network plausible that a subset of risk variants promotes disease risk genes are most highly expressed in fetal astrocytes, neurons, specifically during development, while others promote risk and microglia/macrophages. The importance of fetal astro- through nuanced changes in biological function in aging. It is cytes is particularly interesting in the context of our pathway also possible that genetic variation may be helpful during findings as they are implicated in neurodevelopmental pro- development but detrimental during aging if activated in- cesses, complement signaling, and forming healthy neuronal appropriately (e.g., synaptic pruning by the innate immune synapses.28 Many of the identified bvFTD risk genes were system). Future longitudinal studies of at-risk individuals will most highly expressed in neuronal populations, which par- be required to fully address this question. Although no single ticipate in many of the cell signaling pathways highlighted in genetic variant used to generate our network directly causes our GO analyses.29 Neurons may also be the target of ab- bvFTD, the cumulative burden of multiple sub-GWAS- normal expression of cell-cycle regulation proteins, which associated variants within these pathways may alter the cel- have been shown in studies of neurodegenerative diseases lular landscape on which development and aging occurs, to result in apoptosis when activated.30,31 The third major thereby altering the amount of metabolic stress, inflammation, cell type expressing our genes of interest—microglia/ and apoptotic tendencies of particular cell types in defined macrophages—play an increasingly appreciated role in the brain networks. Our results support the utility of polygenic pathogenesis of bvFTD and other neurodegenerative dis- scores that incorporate the effects of multiple genetic loci for orders.32 Remarkably, our CSEA analyses highlighted layer 5b clinical risk assessment and prevention study enrichment, and cortical neurons as sites of bvFTD network gene expression genetic overlap with AD suggests shared underlying biology enrichment. Pathologic evidence in human brain tissue sug- across multiple neurodegenerative disorders. Understanding gests this cortical layer as a likely site of selectively vulnerable the temporal tropism of these risk factors on disease vulner- neuronal populations in bvFTD (VENs and fork cells).33 ability will also be critical for translating these results into While this finding was not significant after correction for therapeutic targets for clinical populations, especially for multiple testing, it is striking for its convergence on the site of processes such as metabolic stress and immune dysfunction bvFTD neurodegeneration in the cortex and provides a strong that are more amenable to intervention. If confirmed, our

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 9 findings represent a valuable step toward a genetic un- Center (NACC) Junior Investigator Award (R.S.D.), RSNA derstanding of selective vulnerability in neurodegenerative Resident/Fellow Grant (R.S.D.), Foundation of ASNR Alz- disease. heimer’s Imaging Grant (R.S.D.), and Alzheimer’s Society Grant 284 (R.F.). Additional support was provided by an Authors contributions MRC Programme grant (MR/N026004/1; J.H. and P.A.L.), L.W. Bonham: design and conceptualization of the study, NIA P01 AG1972403 (B.L.M.), NIA P50 AG023501 analysis and interpretation of the data, drafting and revising (B.L.M.), and a MRC New Investigator Research Grant the manuscript for intellectual content. N.Z.R. Steele: analysis (MR/L010933/1; P.A.L.). P.A. Lewis is a Parkinson’sUK and interpretation of the data, drafting and revising the research fellow (grant F1002). The funders had no role in manuscript for intellectual content. C.M. Karch: analysis and study design, data collection and analysis, decision to publish, interpretation of the data, revising the manuscript for in- or preparation of the manuscript. tellectual content. C. Manzoni: interpretation of the data, revising the manuscript for intellectual content. E.G. Geier: Disclosure interpretation of the data, revising the manuscript for in- L.W. Bonham has received research support from the Ra- tellectual content. N. Wen, interpretation of the data, revising diologic Society of North America. N.Z.R. Steele, C.M. Karch, the manuscript for intellectual content. A. Ofori-Kuragu: N. Wen, A. Ofori-Kuragu, P. Momeni report no disclosures. analysis and interpretation of the data, revising the manuscript C. Manzoni has received a bursary from NIH; serves on the for intellectual content. P. Momeni: design and conceptuali- editorial board of Frontiers; and has receives research support/ zation of the study, revising the manuscript for intellectual salary from the Medical Research Council. E.G. Geier holds content. J. Hardy: design and conceptualization of the study, a patent (pending) for Targeted platinum anticancer agents. J. revising the manuscript for intellectual content. Z.A. Miller: Hardy has served on scientific advisory boards for Eisai and design and conceptualization of the study, revising the man- Lilly; has been a consultant for Eisai and Ceracuity; and has uscript for intellectual content. C.P. Hess: design and con- received research support from the Medical Research Coun- ceptualization of the study, revising the manuscript for cil. Z.A. Miller holds patents for Dynamic and adjustable intellectual content. P. Lewis: design and conceptualization of portable X-ray filter grid and Dynamic and adjustable filter the study, revising the manuscript for intellectual content. B.L. grids; and has received research support from NIH. C.P. Hess Miller: design and conceptualization of the study, revising the serves on the editorial boards of the American Journal of manuscript for intellectual content. W.W. Seeley: design and Neuroradiology and PLoS One; has been a consultant for Im- conceptualization of the study, interpretation of the data, re- aging Endpoints; has received research support from General vising the manuscript for intellectual content. S.E. Baranzini: Electric and NIH; and has provided expert witness for various design and conceptualization of the study, interpretation of legal trials. P. Lewis has received travel funding from the the data, revising the manuscript for intellectual content. R.S. Biochemical Society; serves on the editorial boards of Biology Desikan: design and conceptualization of the study, in- Direct, PLoS One, and Frontiers in Neurodegenerative Disease; terpretation of the data, revising the manuscript for in- and has received research support from the Medical Research tellectual content. R. Ferrari: design and conceptualization of Council, the Biotechnology and Biological Sciences Research the study, analysis and interpretation of the data, revising the Council, Diamond Light Source, and Parkinson’s UK. B.L. manuscript for intellectual content. J.S. Yokoyama: design and Miller has served on the scientific advisory boards of the Tau conceptualization of the study, analysis and interpretation of Consortium, the John Douglas French Foundation, the Larry the data, drafting and revising the manuscript for intellectual L. Hillblom Foundation, the Consortium for Frontotemporal content. Dementia Research, the Global Brain Health Institute (GBHI), University of Washington ADRC, Stanford Uni- Acknowledgment versity ADRC, the Arizona Alzheimer’s Disease Center The authors thank the International FTD-Genomics Con- (ADC), the Massachusetts Alzheimer Disease Research sortium (IFGC) for providing phase I summary statistics data Center, and the International Society of FTD; has served on for these analyses. IFGC acknowledgements and full the editorial boards of Cambridge University Press, Guilford collaborator list are provided in appendix e-1 (links.lww. Publications Inc, Oxford University Press, Neurocase, Elsevier, com/NXG/A71). Up-To-Date, and Frontiers; receives publishing royalties from Cambridge University Press, Elsevier, Guilford Publications, Study funding and Oxford University Press; and has received research sup- Primary support for data analyses was provided by the Ra- port from Quest Diagnostics Incorporated, and NIA. W.W. diologic Society of North America (RSNA) RMS1741 Seeley serves on the editorial boards of Annals of Neurology, (L.W.B.), Larry L. Hillblom Foundation 2012-A-015-FEL and Acta Neuropathologica, and Neuroimage: Clinical; has been 2016-A-005-SUP (J.S.Y.), AFTD Susan Marcus Memorial a consultant for Bristol-Myers Squibb, Merck, and Biogen Fund Clinical Research Grant (J.S.Y.), NIA K01 AG049152 Idec; has received research support from NIH; and has pro- (J.S.Y.), Bluefield Project to Cure FTD (J.S.Y.), Tau Con- vided expert witness for a defendant. S.E. Baranzini has served sortium (J.S.Y.), NIA K01 AG046374 (C.M.K.), on scientific advisory boards for Novartis, EMD Serono, and U24DA041123 (R.S.D.), National Alzheimer’s Coordinating Sanofi-Aventis; has received gifts, travel funding, and speaker

10 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG honoraria from Novartis; serves on the editorial board of the 16. Allen M, Carrasquillo MM, Funk C, et al. Human whole genome genotype and transcriptome data for Alzheimer’s and other neurodegenerative diseases. Sci data Multiple Sclerosis Journal; holds a patent (pending) for a gene 2016;3:160089. expression signature that could identify patients at high risk of 17. Lesnick TG, Papapetropoulos S, Mash DC, et al. A genomic pathway approach to a complex disease: axon guidance and Parkinson disease. PLoS Genet 2007;3: developing multiple sclerosis; has been a consultant for 0984–0995. Novartis, EMD Serono, and TEVA Neuroscience; and has 18. Kang HJ, Kawasawa YI, Cheng F, et al. Spatio-temporal transcriptome of the human brain. Nature 2011;478:483–489. received research support from NIH, the Department of 19. Bennett ML, Bennett FC, Liddelow SA, et al. New tools for studying microglia in the Defense, and NMSS. R.S. Desikan has received research mouse and human CNS. Proc Natl Acad Sci 2016;113:E1738–E1746. support from NIH, the National Alzheimer’s Coordinating 20. Dijkstra AA, Lin LC, Nana AL, Gaus SE, Seeley WW. Von economo neurons and fork cells: a neurochemical signature linked to monoaminergic function. Cereb Cortex Center (NACC) Junior Investigator Award, RSNA, and the 2016;45:846–860. Foundation of ASNR. R. Ferrari has received travel funding/ 21. Chen-Plotkin AS, Geser F, Plotkin JB, et al. Variations in the progranulin gene affect global gene expression in frontotemporal lobar degeneration. Hum Mol Genet 2008; registration fee reimbursement from Sindem4Juniors and the 17:1349–1362. Alzheimer’s Association International Conference; and has 22. Bronner IF, Bochdanovits Z, Rizzu P, et al. Comprehensive mRNA expression profiling distinguishes tauopathies and identifies shared molecular pathways. PLOS ONE received research support from the Alzheimer’s Society. J.S. 2009;4:e6826. Yokoyama has served on the editorial boards of the Journal of 23. Ferrari R, Wang Y, Vandrovcova J, et al. Genetic architecture of sporadic fronto- ’ temporal dementia and overlap with Alzheimer’s and Parkinson’s diseases. J Neurol Alzheimer s Disease and Frontiers in Neurology; and has re- Neurosurg Psychiatry 2017;88:152–164. ceived research support from Quest Diagnostics, NIH, the 24. Iqbal K, Liu F, Gong CX. Tau and neurodegenerative disease: the story so far. Nat Rev Neurol 2015;12:15–27. Association for Frontotemporal Degeneration, the Larry L. 25. Sawyer RP, Rodriguez-Porcel F, Hagen M, Shatz R, Espay AJ. Diagnosing the frontal Hillblom Foundation, the Rainwater Charitable Foundation, variant of Alzheimer’s disease: a clinician’s yellow brick road. J Clin Mov Disord 2017;4:2. ’ 26. Brettschneider J, Del Tredici K, Lee VM, Trojanowski JQ. Spreading of pathology in the John Douglas French Alzheimer s Foundation, and the neurodegenerative diseases: a focus on human studies. Nat Rev Neurosci 2015;16: Bluefield Project to Cure FTD. Full disclosure form in- 109–120. 27. Cohen TJ, Lee VM, Trojanowski JQ. TDP-43 functions and pathogenic mechanisms formation provided by the authors is available with the full implicated in TDP-43 proteinopathies. Trends Mol Med 2011;17:659–667. text of this article at Neurology.org/NG. 28. Yang Y, Higashimori H, Morel L. Developmental maturation of astrocytes and pathogenesis of neurodevelopmental disorders. J Neurodev Disord 2013;5:22. 29. Ferrari R, Wang Y, Vandrovcova J, et al. Genetic architecture of sporadic fronto- Received January 29, 2018. Accepted in final form July 31, 2018. temporal dementia and overlap with Alzheimer’s and Parkinson’s diseases. J Neurol Neurosurg Psychiatry 2017;88:152–164. 30. Becker EB, Bonni A. Cell cycle regulation of neuronal apoptosis in development and References disease. ProgNeurobiol 2004;72:1–25. 1. Seeley WW, Crawford RK, Zhou J, Miller BL, Greicius MD. Neurodegenerative 31. Copani A, Caraci F, Hoozemans JJM, Calafiore M, Angela Sortino M, Nicoletti F. The diseases target large-scale human brain networks. Neuron 2009;62:42–52. nature of the cell cycle in neurons: focus on a “non-canonical” pathway of DNA 2. Warren JD, Rohrer JD, Schott JM, Fox NC, Hardy J, Rossor MN. Molecular nexopathies: replication causally related to death. Biochim Biophys Acta 2007;1772:409–412. a new paradigm of neurodegenerative disease. Trends Neurosci 2013:561–569. 32. Minami SS, Min SW, Krabbe G, et al. Progranulin protects against amyloid β de- 3. Zhou J, Gennatas ED, Kramer JH, Miller BL, Seeley WW. Predicting regional neu- position and toxicity in Alzheimer’s disease mouse models. Nat Med 2014;20: rodegeneration from the healthy brain functional connectome. Neuron 2012;73: 1157–1164. 1216–1227. 33. Santillo AF, Englund E. Greater loss of von Economo neurons than loss of layer II and 4. Boyle EA, Li YI, Pritchard JK. An expanded view of complex traits: from polygenic to III neurons in behavioral variant frontotemporal dementia. Am J Neurodegener Dis omnigenic. Cell 2017;169:1177–1186. 2014;3:64–71. 5. Baranzini S, Khankhanian P, Patsopoulos N, et al. Network-based multiple sclerosis 34. Chow F, Gong Y, Lippa CF. The potential role of insulin on the shank-postsynaptic pathway analysis with GWAS data from 15,000 cases and 30,000 controls. Am J Hum platform in neurodegenerative diseases involving cognition. Am J Alzheimer’s Dis Genet 2013;92:854–865. Other Demen 2014;29:303–310. 6. Wang C, Ji F, Hong Z, et al Disrupted salience network functional connectivity and 35. Kang J, Park H, Kim E. IRSp53/BAIAP2 in dendritic spine development, NMDA white-matter microstructure in persons at risk for psychosis: findings from the receptor regulation, and psychiatric disorders. Neuropharmacology 2016;100:27–39. LYRIKS study. Psychol Med 2016;46:2771–2783. 36. Narayanan RK, Mangelsdorf M, Panwar A, Butler TJ, Noakes PG, Wallace RH. 7. Rascovsky K, Hodges JR, Kipps CM, et al. Diagnostic criteria for the behavioral Identification of RNA bound to the TDP-43 ribonucleoprotein complex in the adult variant of frontotemporal dementia (bvFTD): current limitations and future direc- mouse brain. Amyotroph Lateral Scler Front Degener 2013;14:252–260. tions. Alzheimer Dis Assoc Disord 2007;21:S14–S18. 37. Lue LF, Schmitz CT, Snyder NL, et al. Converging mediators from immune and 8. Ferrari R, Hernandez DG, Nalls MA, et al. Frontotemporal dementia and its subtypes: trophic pathways to identify Parkinson disease dementia. Neurol Neuroimmunol a genome-wide association study. Lancet Neurol 2014;13:686–699. Neuroinflamm 2016;3:e193. 9. Liu JZ, McRae AF, Nyholt DR, et al. A versatile gene-based test for genome-wide 38. Hooper C, Killick R, Lovestone S. The GSK3 hypothesis of Alzheimer’s disease. association studies. Am J Hum Genet 2010;87:139–145. J Neurochem 2008;104:1433–1439. 10. Frazer KA, Ballinger DG, Cox DR; The International HapMap Corsortium. A 39. Trojanowski JQ, Duff K, Fillit H, et al. New directions for frontotemporal dementia second generation human haplotype map of over 3.1 million SNPs. Nature 2007;449: drug discovery. Alzheimers Dement 2008;4:89–93. 851–861. 40. Sekine Y, Takeda K, Ichijo H. The ASK1-MAP kinase signaling in ER stress and 11. Ferrari R, Lovering RC, Hardy J, Lewis PA, Manzoni C. Weighted protein in- neurodegenerative diseases. Curr Mol Med 2006;6:87–97. teraction network analysis of frontotemporal dementia. J Proteome Res 2017;16: 41. Sousa SB, Hennekam RC, Abdul-Rahman O, et al. Phenotype and genotype in 999–1013. Nicolaides-Baraitser syndrome. Am J Med Genet C Semin Med Genet 2014;166: 12. Zhang Y, Sloan SA, Clarke LE, et al. Purification and characterization of progenitor 302–314. and mature human astrocytes reveals transcriptional and functional differences with 42. Zhang YJ, Gendron TF, Xu YF, Ko LW, Yen SH, Petrucelli L. Phosphorylation mouse. Neuron 2016;89:37–53. regulates proteasomal-mediated degradation and solubility of TAR DNA binding 13. Sheng J, Su L, Xu Z, Chen G. Progranulin polymorphism rs5848 is associated with protein-43 C-terminal fragments. Mol Neurodegener 2010;5:1–13. increased risk of Alzheimer’s disease. Gene 2014;542:141–145. 43. Moradifard S, Hoseinbeyki M, Ganji SM, Minuchehr Z. Analysis of microRNA and 14. Schmidt EF, Warner-Schmidt JL, Otopalik BG, Pickett SB, Greengard P, Heintz N. gene expression profiles in Alzheimer’s disease: a meta-analysis approach. Sci Rep Identification of the cortical neurons that mediate antidepressant responses. Cell 2018;8:4767. 2012;149:1152–1163. 44. Dobson-Stone C, Luty AA, Thompson EM, et al. Frontotemporal dementia- 15. Reimand J, Arak T, Adler P, et al. g:Profiler: a web server for functional interpretation amyotrophic lateral sclerosis syndrome locus on chromosome 16p12.1-q12.2: genetic, of gene lists (2016 update). Nucleic Acids Res 2016;44:W83–W89. clinical and neuropathological analysis. Acta Neuropathol 2013;125:523–533.

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 11 ARTICLE OPEN ACCESS Plasticity-related gene 3 (LPPR1) and age at diagnosis of Parkinson disease

Zachary D. Wallen, MS, Honglei Chen, PhD, Erin M. Hill-Burns, PhD, Stewart A. Factor, DO, Correspondence Cyrus P. Zabetian, MD, MS, and Haydeh Payami, PhD Dr. Payami [email protected] Neurol Genet 2018;4:e271. doi:10.1212/NXG.0000000000000271 Abstract Objective To identify modiﬁers of age at diagnosis of Parkinson disease (PD).

Methods Genome-wide association study (GWAS) included 1,950 individuals with PD from the Neu- roGenetics Research Consortium (NGRC) study. Replication was conducted in the Parkin- ’ son s, Genes and Environment study, including 209 prevalent (PAGEP) and 517 incident (PAGEI) PD cases. Cox regression was used to test association with age at diagnosis. Indi- viduals without neurologic disease were used to rule out confounding. Gene-level analysis and functional annotation were conducted using Functional Mapping and Annotation of GWAS platform (FUMA).

Results The GWAS revealed 2 linked but seemingly independent association signals that mapped to LPPR1 on chromosome 9. LPPR1 was signiﬁcant in gene-based analysis (p = 1E-8). The top signal (rs17763929, hazard ratio [HR] = 1.88, p = 5E-8) replicated in PAGEP (HR = 1.87, p = 0.01) but not in PAGEI. The second signal (rs73656147) was robust with no evidence of heterogeneity (HR = 1.95, p = 3E-6 in NGRC; HR = 2.14, p = 1E-3 in PAGEP + PAGEI, and HR = 2.00, p = 9E-9 in meta-analysis of NGRC + PAGEP + PAGEI). The associations were with age at diagnosis, not confounded by age in patients or in the general population. The PD-associated regions included variants with Combined Annotation Dependent Depletion (CADD) scores = 10–19 (top 1%–10% most deleterious mutations in the genome), a missense with predicted destabilizing eﬀect on LPPR1, an expression quantitative trait locus (eQTL) for GRIN3A (false discovery rate [FDR] = 4E-4), and variants that overlap with enhancers in LPPR1 and interact with promoters of LPPR1 and 9 other brain-expressed genes (Hi-C FDR < 1E-6).

Conclusions Through association with age at diagnosis, we uncovered LPPR1 as a modiﬁer gene for PD. LPPR1 expression promotes neuronal regeneration after injury in animal models. Present data provide a strong foundation for mechanistic studies to test LPPR1 as a driver of response to damage and a therapeutic target for enhancing neuroregeneration and slowing disease progression.

From the Department of Neurology (Z.D.W., E.M.H.-B., H.P.), University of Alabama at Birmingham, Birmingham, AL; Department of Epidemiology and Biostatistics (H.C.), Michigan State University, East Lansing, MI; Department of Neurology (S.A.F.), Jean & Paul Amos Parkinson’s Disease and Movement Disorder Program, Emory University School of Medicine, Atlanta, GA; VA Puget Sound Health Care System and Department of Neurology (C.P.Z.), University of Washington, Seattle, WA; and Center for Genomic Medicine (H.P.), HudsonAlpha Institute for Biotechnology, Huntsville, AL.

Copyright © 2018 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary AIM = ancestry informative marker; FDR = false discovery rate; HR = hazard ratio; LD = linkage disequilibrium; MAF = minor allele frequency; MAP = moving average plot; NGRC = NeuroGenetics Research Consortium; PAGE = Parkinson’s, Genes and Environment; PC = principal component; PD = Parkinson disease.

The underlying neurodegenerative process that causes Par- clinics in Portland (OR), Seattle (WA), Albany (NY), and kinson disease (PD) begins decades before the disease is di- Atlanta (GA). Controls were spouses of patients or commu- agnosed.1 The current view is that following an initial insult nity volunteers, self-reported as being free of neurologic dis- (e.g., toxicity, trauma, or genetic), the disease starts with an ease. The eligibility criterion for cases was diagnosis of PD by asymptomatic phase of unknown duration, followed by de- a movement disorder specialist according to the UK Brain velopment of prodromal nonmotor symptoms such as con- Bank criteria.15 The eligibility criteria for controls were no stipation, anosmia, and sleep disorders. Years later, cardinal neurologic disease and genetically unrelated to patients. Age motor signs appear, at which point a diagnosis of PD is made. was defined as age at study entry. Age at diagnosis was Age at onset of motor signs, and therefore the age at diagnosis extracted from medical records or ascertained by self-report. of PD, is highly variable, ranging from teen ages to the 10th Age at onset of the first motor sign was obtained using a self- decade of life. The reason for this variation is unknown, and administered questionnaire. Age at onset and age at diagnosis understanding it will likely shed light on factors that affect the were highly correlated in the NGRC (r2 = 0.91, p < 2E-16). All rate of disease progression. participants were whites of European descent.13

There is substantial evidence that genetic factors play a major PAGE is a cross-sectional study nested in the longitudinal role in age at onset of motor signs and age at diagnosis of NIH-American Association of Retired Persons Diet and – PD.2 6 Genome-wide studies have identified numerous loci Health Study.14 Participants were enrolled in 1995–1997 that associate with the risk of developing PD,7 but the risk (irrespective of PD) via a food frequency questionnaire – factors do not explain the variation in age at onset.8 10 Three mailing16 and in the 2004–2006 follow-up visit were asked if loci have been nominated as modifiers of age at onset in they had been diagnosed with a major chronic disease in- familial PD.11,12 The present study was aimed at identifying cluding PD. Participants who had been diagnosed with PD genetic modifiers for common idiopathic PD. We hypothe- before enrollment (before 1998) were designated as prevalent fi sized that identi cation of the genetic basis to interindividual PD (PAGEP, N = 209), participants who were diagnosed variability in age at diagnosis will provide insights into the during follow-up (1998–2006) were designated as incident intrinsic mechanisms that determine the rate of deterioration PD (PAGEI, N = 517), and participants who did not have PD during preclinical disease. were designated as controls (N = 1,549). All participants in this study were non-Hispanic whites.

Methods Genotyping This study was a case-control GWAS, followed by replication NGRC participants were genotyped on Illumina Human- and functional annotation. Omni1-Quad v1-0 B array and Immunochip array. Genotypes and samples were filtered by call rate, minor allele frequency Standard protocol approvals, registrations, (MAF) < 0.01, Hardy-Weinberg, and cryptic relatedness, as and patient consents described before.13 Imputation was performed using IMPUTE The study was approved by the institutional review boards at v2.3.0,17 with the 1000G Phase3 integrated variant set all participating institutions. Written informed consent was (October 2014) as reference. Imputed single nucleotide obtained from all patients and controls for participation in the polymorphisms (SNPs) with info score < 0.9 or MAF < 0.01 study. were excluded. A total of 8.5 million SNPs (900,000 genotyped and 7.6 million imputed) were used in the analysis. Participants The study included 2 data sets. The NeuroGenetics Research PAGE participants were genotyped for rs73656147 (block 1) 13 Consortium (NGRC) data set was used for the discovery and rs17763929 (block 2). SNPs were chosen based on sta- GWAS, gene-based test, and functional annotations. The tistical significance and availability of predesigned validated 14 Parkinson’s, Genes and Environment (PAGE) study was TaqMan assay from Thermo Fisher (rs73656147 assay used for replication. Participants’ characteristics are shown in number = C__97534229_10; rs17763929 assay number = table 1 and figure e-1 (links.lww.com/NXG/A66). C__34297681_10).

NGRC is a case-control study of genetically unrelated par- Population structure ticipants, including 2000 PD cases and 1986 controls.13 Principal component (PC) analysis18 is used to infer population- Patients were enrolled sequentially from movement disorder specific genetic differences, which arise from ancestry differences

2 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Table 1 Data sets and participants’ characteristics

Discovery (NGRC) Replication (PAGE)

PD Controls PAGEP PAGEI Controls

N 2,000 1,986 209 517 1,549

Male/Female 1,346/654 769/1,217 164/45 396/121 1,213/336

Age at enrollment mean ± SD 67.3 ± 10.7 70.3 ± 14.1 62.6 ± 4.9 63.2 ± 4.9 63.4 ± 4.9

Age at follow-up mean ± SD NR NR 73.9 ± 4.9 74.5 ± 4.9 74.0 ± 4.9

N with age at onset data 1,999 NR 0 0 NR

Age at onset mean ± SD 58.3 ± 11.9 NR NA NA NR

N with age at diagnosis data 1,950 NR 209 517 NR

Age at diagnosis range 25–90 NR 42–72 53–81 NR

Age at diagnosis mean ± SD 60.4 ± 11.4 NR 59.9 ± 6.6 69.4 ± 5.4 NR

Abbreviations: NA = not available; NGRC = NeuroGenetics Research Consortium; NR = not relevant; PAGE = Parkinson’s, Genes and Environment. Participants were non–Hispanic whites and genetically unrelated. Data on the NGRC participants were collected at enrollment: patients already had the diagnosis of PD and controls were free of neurologic disease. NGRC participants were enrolled at 4 sites: Oregon, Washington, New York, and Georgia. Age at onset mean ± SD were as follows: Oregon = 56.6 ± 12.8, Washington = 58.7 ± 11.8, New York = 59.4 ± 11.5, and Georgia = 58.7 ± 11.1. Age at diagnosis mean ± SD were as follows: Oregon = 59.6 ± 11.7, Washington = 60.7 ± 11.6, New York = 60.9 ± 11.1, and Georgia = 60.3 ± 10.6. PAGE participants were originally enrolled in the longitudinal NIH-AARP diet study in 1995–1997. Their PD status was investigated in 2004–2006. Participants who had the diagnosis of PD before 1998 were classified as prevalent PD (PAGEP), participants who were diagnosed with PD during follow-up (between 1998 and 2006) were classified as incident PD (PAGEI), and participants who did not have PD were designated as controls. Because PAGE participants were of similar age at entry, the method of classifying the participants into prevalent vs incident cases inevitably assigned earlier ages at diagnosis to the prevalent group and later diagnoses to the incident group. in allele frequencies and can obscure genetic association studies used to visualize the chr9:103,865,000–104,055,000 region if not accounted for. NGRC PC analysis was conducted using (GWAS peak). Haploview v4.223 was used to generate linkage a pruned subset of 100K SNPs from the GWAS as previously disequilibrium (LD) plots of D9 and r2 for SNPs in the chr9: described.13 The top 3 PCs (effect sizes PC1 = 0.2%, PC2 = 103,865,000–104,055,000 region with GWAS p <1E-4.LD 0.06%, and PC3 = 0.06%) were included in the GWAS and between 2 SNPs was calculated using 1000G Phase3 v5 in adjusted for in all downstream analyses involving the NGRC. LDlink.24 Linear regression was used to estimate and test dif- The PAGE data sets used for replication did not have ancestry ferences in mean age at diagnosis (β). Conditional analysis was informative markers (AIMs); however, a subset of the partic- performed using coxph function in the survival v2.41 R pack- ipants (396 of 726 PD cases) was previously genotyped with the age. Moving average plots (MAPs) were generated using the Immunochip array. We conducted PC analysis using a pruned freqMAP v0.2 R package.25 set of 20K SNPs from the Immunochip array, using PLINK. Tests were conducted once using the full PAGE data set, with no Gene-based analysis was conducted using summary statistics PC adjustment, and again with a PAGE subset, adjusting for from the GWAS and LD from the 1000G Phase3 EUR to map ff PC1-3 (e ect sizes PC1 = 0.48%, PC2 = 0.20%, and PC3 = the GWAS SNPs to 18,985 protein-coding genes (hg19 build) 0.17%). NGRC and PAGE cluster with Europeans in the and to calculate gene-based p values, using MAGMA v1.06,26 fi 1000G_Phase_3 global data set ( gure e-2, links.lww.com/ as implemented in FUMA v1.3.0.20 Statistical significance was NXG/A67). set at Bonferroni-corrected p < 2.6E-6 (0.05/18,985).

Statistics Replication Discovery Cox regression (coxph function in the survival v2.41 R GWAS was conducted using PD cases only (1,950 NGRC package) was used to replicate the association of 2 SNPs with participants with known age at diagnosis). Association between age at diagnosis. We used the same model as the NGRC 8.5M SNPs and age at diagnosis was tested using Cox re- (additive genetic model, treating age at diagnosis as a quanti- gression in ProbABEL v0.5.0.,19 specifying an additive genetic tative trait). Because of the availability of PCs only in a subset model, treating age at diagnosis as a quantitative trait, and of PAGE, analyses were conducted twice: using the full PAGE adjusting for PC1-3. The statistical outcome of Cox regression data set without PC adjustment and using the subset that had was hazard ratios (HRs) and corresponding p values. Statistical AIMs and adjusting for PC1-3. PAGEI and PAGEP were significance was set at p < 5E-8. Manhattan plots and quantile- treated separately and were combined using meta-analysis quantile (QQ) plots were generated using FUMA v1.3.0.20 after testing for heterogeneity. If p of heterogeneity was <0.1, Genomic inflation factor (λ)wascalculatedusingthe the fixed-effect model was used. Meta-analysis was performed estlambda function in GenABEL v1.8 in R.21 LocusZoom22 was using the metagen function in the meta v4.8 R package.

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 3 Functional annotation whether and how allele frequencies vary by age in cases or in Functional annotation was conducted in FUMA v1.3.0,20 controls. Allele frequencies were plotted in a moving average using SNPs with GWAS p < 1E-6 and all variants in r2 ≥ 0.6 window as a function of age (figure e-3, links.lww.com/NXG/ with them, and included CADD analysis,27 eQTL mapping,28 A68). Starting at age 45 years, allele frequencies were the same 3D chromatin interaction mapping (Hi-C),29 annotation of in cases and controls. In controls, allele frequencies remained enhancers,30 tissue-specific expression of genes identified via the same across the age spectrum, whereas in cases, they Hi-C and eQTL mapping,28 and their age-specific expression decreased sharply and significantly by age and by age at di- in the brain (BrainSpan.org). The false discovery rate (FDR) agnosis. The effect was therefore in cases and not in controls. was used to correct for multiple testing. STRUM was used to Next, conditional analysis was conducted to tease age from predict the effect of a missense on the structural stability of age at diagnosis (table 2). The minor alleles of rs73656147 a protein.31 and rs17763929 were associated with age, as was expected, given their association with age at diagnosis. However, the Data availability association with age at diagnosis persisted when adjusted for NGRC genotype and phenotype data are available at dbGaP age, but the association with age was abolished when adjusted ncbi.nlm.nih.gov/gap accession number phs000196.v3.p1. for age at diagnosis. Hence, age at diagnosis was the driving force, and association with age was a by-product of the correlation. Results To gauge robustness of the association signals with age at GWAS diagnosis and to test for heterogeneity, we stratified the data In SNP-based GWAS, the most significant signal for associ- by 8 PD-relevant variables, tested the association of each SNP ation, at p = 5E-8, mapped to LPPR1 on chromosome 9q31.1 with age at diagnosis within each stratum, and compared the (figure 1, A and B). In the gene-based test, LPPR1 achieved results across strata for evidence of heterogeneity (table e-2, p = 1E-8, surpassing the genome-wide statistical significance links.lww.com/NXG/A70). The 8 categories of stratification threshold of p < 2.6E-6 (figure 1, C and D). The p values were were family history, sex, cigarette smoking, caffeine intake, not inflated (λ = 1.007 SNP based, λ = 1.04 gene based). nonsteroidal anti-inflammatory drugs use, recruitment site, Analysis of LD in the region revealed 2 haplotype blocks with Jewish heritage, and the European country of ancestral origin. seemingly independent signals for association (figure 1, E and The association signal for rs73656147 (block 1) was robust F). There was strong LD among SNPs in each block, but weak across all strata. rs17763929 (block 2) showed evidence of LD between the blocks (r2 ≤ 0.2) because of a recombination heterogeneity as a function of recruitment site and the Eu- hot spot between them (figure 1F). The 2 blocks were in ropean country of ancestral origin. Given these results, we a ;200 Kb region inside LPPR1. Block 1 consisted of 51 tested the association of the 2 SNPs with PCs. rs17763929 SNPs with MAF;0.01, which yielded HR = 2.02–1.88, with was associated with PC1 (p = 7E-6) and PC3 (p = 8E-3), and p = 9E-7 to 2E-5 for association with age at diagnosis. Block 2 rs73656147 was not (p > 0.05 for PC1-3), indicating the consisted of 39 SNPs with MAF;0.02, which yielded HR = presence of population structure in block 2 but not in block 1. 1.88–1.85, with p = 5E-8 to 7E-7. We chose 1 SNP to represent each block for replication: rs73656147 for block 1 Replication (MAF = 0.01, HR = 1.95, p = 3E-6) and rs17763929 for block In comparison to NGRC, which had a 65-year range for age at 2 (MAF = 0.02, HR = 1.88, p = 5E-8), both in Hardy- diagnosis, the PAGE data sets had a narrower range of less Weinberg (p > 0.3), with little correlation between them (r2 = than 30 years. Because PAGE participants were of similar age 0.2). Conditional analysis conducted to determine whether at study entry, the method of classifying the participants into the 2 blocks were tagging the same or different disease- prevalent PD (diagnosis before entry) vs incident PD (di- associated variants was inconclusive because although the agnosis after entry) inevitably assigned earlier ages at di- signals were weakened when adjusted for each other, neither agnosis to the prevalent group (PAGEP) and later diagnoses was abolished when conditioned on the other (table e-1, links. to the incident group (PAGEI). Mean age at diagnosis in lww.com/NXG/A69). PAGEP was 59.9 ± 6.6 years, which was similar to the NGRC (60.4 ± 11.4). PAGEI participants were on average 10 years There are 2 caveats in interpreting statistical evidence for older at diagnosis (69.4 ± 5.4, range 53–81 years). Given the association with age at diagnosis. First, age at diagnosis is disparity in the range and mean ages at diagnosis, we analyzed 2 correlated with age (r = 0.74, p < 2E-16), which can result in PAGEP and PAGEI separately. spurious conclusions if the driving force responsible for the association is not identified. Second, tests of age at diagnosis Association of rs73656147 (block 1) with age at diagnosis are conducted using patients only without the benefitof replicated robustly (table 3). There was no evidence of het- controls. For example, an SNP that appears to be associated erogeneity between PAGEI and PAGEP in the association of with earlier PD diagnosis may in fact be associated with an rs73656147 with age at diagnosis, although the signal was age-related event unrelated to PD. To interpret the statistical stronger in PAGEP than in PAGEI, which is not surprising, evidence for association with age at diagnosis, we examined given that the former is enriched in cases with earlier age at

4 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Figure 1 Results of genome-wide association study for age at diagnosis of PD

Genome-wide association was tested between 8.5 million SNPs and age at diagnosis in 1,950 PD cases from the NGRC, using the Cox hazard ratio regression method and adjusting for principal components (PC1-3). (A) Manhattan plot of SNP-based GWAS. Tallest peak, at p = 5E-8, was on chromosome 9q31.1. (B) QQ plot of SNP-based GWAS. The observed p values were not inflated (λ = 1.007). (C) Manhattan plot of gene-based GWAS. LPPR1 was at p = 1E-8. Statistical significance threshold was p < 2.6E-6, which is Bonferroni corrected for the 18,985 protein-coding genes tested. (D) QQ plot of gene-based GWAS. The observed p values were not inflated (λ = 1.04). (E) r2 (top panel) and D’ (bottom panel). Linkage disequilibrium (LD) across the SNPs that gave p < 1E-4 for association with age at diagnosis reveals 2 blocks represented by rs73656147 (left triangle) and rs17763929 (right triangle). (F) Magnified map of the associated region (chr9:103,865,000–104,055,000), showing that PD-associated SNPs map to LPPR1 and form 2 haplotype blocks separated by recombination hot spots (blue spikes). (G) Chromatin state of LPPR1 (Roadmap 111 Epigenomes), showing that active enhancers (yellow), transcription start site (red), and transcripts (green) of LPPR1 are seen only in stem cells and the brain and that the GWAS SNPs align with regulatory elements. ESC = embryonic stem cell; iPSC = induced pluripotent stem cell; TssA = active transcription start site (TSS); TssAFlnk = flanking active TSS; TxFlnk = transcription at gene 59 and 3’; Tx = strong transcription; TxWk = weak transcription; EnhG = genic enhancers; Enh = enhancers; ZNF/Rpts = zinc-finger genes and repeats; Het = heterochromatin; TssBiv = bivalent/poised TSS; BivFlnk = flanking bivalent TSS/enhancer; EnhBiv = bivalent enhancer; ReprPC = repressed polycomb; ReprPCWk = weak repressed polyComb; Quies = quiescent.

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 5 Table 2 Association of LPPR1 variants with age and age at diagnosis is driven by age at diagnosis

Block 1 Block 2

rs73656147 rs17763929

Cox LR Cox LR

N HR p Value β [95% CI] HR p Value β [95% CI]

Ia. Association with age at diagnosis in cases 1,950 1.95 3E-6 −6.00 [−9.18 to −2.83] 1.88 5E-8 −5.65 [−8.20 to −3.11]

Ib. Association with age at diagnosis in 1,950 1.95 3E-6 −5.98 [−9.16 to −2.81] 1.88 6E-8 −5.61 [−8.16 to −3.07] cases adjusted for sex

II. Association with age in cases 2,000 1.48 5E-3 −4.19 [−7.1 to −1.3] 1.53 2E-4 −3.56 [−5.9 to −1.2]

III. Association with age in controls 1,986 0.83 0.08 2.34 [−0.6 to 5.2] 0.84 0.07 2.37 [−0.3 to 5.1]

IV. Association with age at diagnosis in 1,950 1.45 0.01 −2.30 [−3.9 to −0.7] 1.26 0.05 −2.11 [−3.4 to −0.8] cases adjusted for age

V. Association with age in cases adjusted 1,950 0.92 0.56 0.78 [−0.8 to 2.3] 0.99 0.96 0.68 [−0.6 to 1.9] for age at diagnosis

Abbreviations: CI = confidence interval; HR = hazard ratio; LR = linear regression; β = effect size on age at diagnosis (in years) per copy of minor allele. The associations were tested in the NGRC data set using Cox regression, and the effect sizes were estimated using linear regression (LR). HR is the age-for-age increase in the odds of event per copy of the minor allele, as estimated using Cox regression. β is the difference in years in age at diagnosis between carriers of 1 minor allele vs no minor allele, as estimated using linear regression. Age at diagnosis was the primary outcome of the study. Minor alleles of rs73656147 and rs17763929 were associated with higher HR and younger age at diagnosis (Ia). The association was not influenced by sex (Ib), which was expected because, unlike PD risk, which is significantly associated with sex (OR = 3.26, p < 2E-16), age at diagnosis is not associated with sex (HR = 0.99, p = 0.83). Minor alleles were also associated with younger ages in cases (II), but not in controls (III). Because age and age at diagnosis were correlated (r2 = 0.74, p < 2E-16), an association with one will show as an association with both. In conditional analysis, the association with age at diagnosis persisted when adjusted for age (IV), but the association with age was abolished when adjusted for age at diagnosis (V), suggesting that age at diagnosis was the driving force and association with age was a by-product of the correlation.

diagnosis. Nor was there evidence of heterogeneity between mapped to enhancers in the brain (table 4 and figure 1, G). PAGE and NGRC for the association of rs73656147 with age Eleven of the genes identified through Hi-C are expressed in at diagnosis. Meta-analysis yielded HR = 2.14, p = 1E-3 for the brain: LPPR1, SEC61B, MSANTD3-TMEFF1, TMEFF1, replication and HR = 2.00, p = 9E-9 for replication and dis- GALNT12, MURC, GRIN3A, NR4A3, ALG2, MRPL50, and covery. Mean difference in age at diagnosis per copy of ZNF189 (figure 2, B and C). The expression of LPPR1 in the rs73656147 minor allele was −6.0 (95% confidence interval: brain is the strongest in early prenatal stage and decreases with −9.18 to −2.83) years in the NGRC, −5.53 (−9.72 to −1.34) in developmental stage and increasing age (figure 2, C). − − − − PAGEP, 0.84 ( 4.22 to 2.55) in PAGEI, and 4.08 ( 7.45 to −0.70) in the meta-analysis of the 3 data sets. CADD analysis, a scoring system for deleteriousness of genetic variants, identified 5 SNPs in block 1 and 3 in block 2, Association of rs17763929 (block 2) with age at diagnosis with CADD = 10–19 (table 4), which places them among the fi showed signi cant heterogeneity between PAGEI and PAGEP top 10% (CADD > 10) to 1% (CADD > 20) of most dele- (table 3), as it had within the NGRC (table e-2, links.lww. terious mutations in the genome.27 rs41296085 (CADD = 18, com/NXG/A70). The association with rs17763929 repli- in block 1) is a missense (p.Ser12Ala) in exon 2, predicted to fi ΔΔ − cated in PAGEP but not in PAGEI. There was signi cant structurally destabilize the LPPR1 protein ( G= 1.2). The heterogeneity between PAGEI and NGRC, but not between remainder of the variants with high CADD scores are in PAGEP and NGRC. Meta-analysis of PAGEP and NGRC introns. eQTL analysis revealed an association between yielded HR = 1.88, p = 4E-9 for full PAGE data and HR = 1.95, rs117451395 (block 1) with expression levels of GRIN3A p = 3E-9 for the PAGE subsample adjusted for PC1-3. In- (FDR = 4E-4). ff cluding PAGEI with PAGEP and NGRC in a random-e ects meta-analysis diluted the effect size to HR = 1.53, p = 0.04. Mean difference in age at diagnosis per copy of rs17763929 Discussion minor allele was −5.65 (−8.20 to −3.11) years in the NGRC, −3.62 (−7.23 to −0.02) in PAGE , and 0.62 (−1.34 to 2.58) in There has been intense research on PD risk factors, which so P fi PAGEI. far has resulted in identi cation of numerous causative genes, 40 susceptibility loci, several environmental factors, and a few Functional annotation genes that interact with the environmental factors to increase Hi-C analysis showed significant (FDR < 1E-6) chromatin or reduce the risk of developing PD. In contrast, we know interaction between the PD-associated LPPR1 SNPs and little about factors that affect the rate of disease progression. promoters of LPPR1 and several genes on chromosome 9 In this study, we attempted to identify genetic modifiers of age (figure 2, A). Some of the SNPs that were significant in Hi-C at diagnosis, a reflection of rate of progression, using an

6 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Table 3 Replication

Age at diagnosis Block 1 rs73656147 Block 2 rs17763929

Data sets N PD cases Mean ± SD HR p Value HR p Value

NGRC (discovery) 1,950 60.4 ± 11.4 1.95 3E-6 1.88 5E-8

PAGEP (replication) 209 59.9 ± 6.6 2.88 7E-4 1.87 0.01

PAGEP with PC1-3 113 59.9 ± 6.8 2.17 0.05 3.03 4E-3

PAGEI (replication) 517 69.4 ± 5.4 1.62 0.07 1.04 0.41

PAGEI with PC1-3 283 69.2 ± 5.3 1.48 0.16 1.03 0.45

Meta-analysis A Heterogeneity rs73656147 Heterogeneity rs17763929

PAGEP and PAGEI ns 0.08 2.14 1E-3 1.34 0.31

NGRC and PAGEP ns ns 2.08 2E-8 1.88 4E-9

NGRC and PAGEI ns 0.01 1.90 9E-7 1.42 0.23

NGRC and PAGEP and PAGEI ns 0.02 2.00 9E-9 1.53 0.04

Meta-analysis B

PAGEP and PAGEI ns 0.02 1.73 0.07 1.67 0.34

NGRC and PAGEP ns ns 1.97 6E-7 1.95 3E-9

NGRC and PAGEI ns 0.02 1.89 2E-6 1.43 0.23

NGRC and PAGEP & PAGEI ns 0.02 1.91 5E-7 1.68 0.05

Abbreviations: HR = hazard ratio; NGRC = NeuroGenetics Research Consortium; ns = not statistically significant; PAGE = Parkinson’s, Genes and Environment; PC = principal component; PD = Parkinson disease. Two SNPs with signals for association with age at diagnosis of PD in the NGRC data set (discovery) were genotyped and tested for association with age at diagnosis of PD in the PAGE data set (replication). PAGE participants were designated as PAGEP if they were diagnosed before study entry or PAGEI if they were diagnosed during the study. Cox regression was used to test the association of SNP (additive model) with age at diagnosis (quantitative trait) and to calculate hazard ratios (HRs) and the corresponding significance (p). NGRC was adjusted for PC1-3 in GWAS and meta-analyses. Only a subset of PAGE had ancestry informative markers (AIMs) for which PC could be calculated; thus, results are shown for the full PAGE data set without PC adjustment and for the PAGE subsample with PC adjustment. p values are 2-sided for NGRC and one-sided for PAGE because of the directionality of the hypothesis being replicated. Meta-analysis A: NGRC (PC1-3 adjusted) and PAGE (all data without PC adjustment). Meta-analysis B: NGRC (PC1-3 adjusted) and PAGE (subset of data adjusted for PC1-3). rs73656147 replicated robustly with no evidence of heterogeneity across data sets. rs17763929 replicated in PAGEP and showed significant heterogeneity between PAGEI and PAGEP or NGRC. Meta-analysis was conducted using the fixed-effects model if there was no evidence for heterogeneity (p ≥ 0.1) and the random effects model if there was heterogeneity (p < 0.1).

unbiased genome-wide approach, followed by independent not fully capture the complexity of block 2. PAGEI participants fi replication, and functional annotation. being signi cantly older than NGRC and PAGEP participants – may also be a factor. LPPR1 promotes neuroregeneration,32 34 We uncovered evidence for association of genetic variants in but its expression diminishes with age to nearly undetectable neuronal plasticity-related gene 3 (LPPR1) with age at di- level by age 40 years (figure 2C). One can speculate that some agnosis of PD. Two signals of association were detected, each detrimental variants may not have an effect after a certain age representing a haplotype block of SNPs. The variants that when the gene is no longer expressed. were associated with earlier age at diagnosis had low allele frequencies (MAF = 0.01–0.02), as were the variants that Functional annotation of the PD-associated variants in LPPR1 were previously found for age at onset of familial PD.11 The revealed the presence of several variants with predicted del- low allele frequencies may be one reason why modifier genes eterious effects, including a missense that destabilizes the have been more difficult to detect than common variants that structure of LPPR1, a regulatory element that associates with associate with risk. expression levels of GRIN3A, and enhancers that interact with promoters of LPPR1 and several other genes in the brain. fi The association with block 1 replicated robustly in both PAGEP Some of the candidate genes that were identi ed via in- and PAGEI. Block 2 signal replicated in PAGEP but not in teraction with LPPR1 play key roles in pathways that are PAGEI. Block 2 has a complex LD structure, with evidence of implicated in PD, including GRIN3A (which encodes a sub- population substructure, which limits generalizability of results. unit of NMDA receptor involved in the glutamate-regulated Failure to capture a signal for block 2 in PAGEI may be because ion channels in the brain), SEC61B (protein transport appa- we had genotype on only 1 SNP in block 2 for PAGE, which did ratus of the endoplasmic reticulum membrane), MURC (Rho

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 7 Figure 2 Functionally significant genes

(A) 3D chromatin interaction (Hi-C) and eQTL analysis. Hi-C revealed significant interaction between GWAS variants in LPPR1 and 17 other genes on chromosome 9 (FDR < 1E-6, shown in orange). An SNP in LPPR1 was associated with the expression of GRIN3A (FDR = 4E-4, shown in green). (B) Tissue-specific expression of LPPR1, GRIN3A, and genes in Hi-C with LPPR1. Colors reflect average expression (log2 transformed) from highest (red) to lowest/absent (blue). (C) Age-specific expression of the genes in the brain. LPPR1 expression decreases with age.

kinase signaling), and MRPL50 (mitochondrial ribosomal we propose that LPPR1 is involved, not necessarily in the protein). cause of PD, rather in response to damage, and influences the efficacy of regeneration and the subsequent rate of de- LPPR1 is one of the 5 members of a brain-specific gene family terioration in preclinical PD. The actual cause of injury and that modulates neuronal plasticity during development, aging, neuronal death is not stipulated in this hypothesis; it could be – and after brain injury.32 34 LPPR1 is the strongest driver of head trauma, environmental toxins or genetic, but once the axonal outgrowth in the gene family. Studies in mice have initial damage is incurred, it is the efficacy of intrinsic mech- shown that after neuronal injury, overexpression of LPPR1 anisms of repair that determine the rate of disease pro- enhances axonal growth, improves motor behavior, and progression. Present findings provide a strong foundation for motes functional recovery.33,34 Extrapolating to our findings, mechanistic studies to investigate the role of LPPR1 in PD and

8 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Table 4 Functionally significant variants

Block GWAS SNP position:alleles GWAS pr2 eQTL CADD Hi-C/EnhBrain

1 rs77351585 9:103874925:C:T 2E-06 1 — 18 Hi-C/EnhBrain

1 rs73495940 9:103875807:G:C 9E-07 Lead —— Hi-C

1 rs150164200 9:103875896:A:C 2E-06 1 — 10.4 —

1 rs117583993a 9:103876647:G:A 3E-06 1 —— Hi-C/EnhBrain

1 rs148874623 9:103939117:A:C 9E-06 1 — 12.1 —

1 rs117451395 9:103941039:C:T 1E-05 1 GRIN3A — Hi-C

1 rs41296085 9:103947810:T:G 2E-05 1 — 18 (missense)b Hi-C

1 rs117900237 9:103959240:G:A 2E-05 1 — 10.5 Hi-C

2 rs17763929 9:103984900:A:G 5E-08 Lead —— Hi-C

2 rs61188842 9:103988006:C:T 8E-05 0.6 —— Hi-C/EnhBrain

2 rs117058418 9:104011717:T:C 2E-07 1 — 10.4 Hi-C

2 rs117314512 9:104014244:G:A 2E-07 1 — 12.4 Hi-C

2 rs149155028 9:104032402:TTC:T 1E-05 0.7 — 18.6 Hi-C

Functional annotation was conducted on SNPs with GWAS p < 1E-6 and SNPs that were in high LD with them (r2 > 0.6). Variants are shown if they are the lead SNPs (most significant) for the block, or an eQTL (FDR = 4E-4), or had a CADD score >10, or had both significant evidence for 3D chromatin interaction (Hi-C, FDR < 1E-6) and overlapped with an enhancer in the brain. Block 1 is a single block of SNPs in high LD. Block 2 has a complex LD structure with at least 3 subhaplotypes (figure e1-C, links.lww.com/NXG/A66). Variants are shown with their rs accession number, chromosome position and the 2 alleles (major: minor), GWAS p value for association with age at diagnosis of PD, and their correlation (r2) with the lead SNP of the block. eQTL: an SNP that is associated with gene expression, in this case, rs117451395, was associated with gene expression levels at GRIN3A (FDR = 4E-4). CADD: a predictive score for the deleteriousness of a variant. A CADD score of 10 usually means that the variant is among the top 10% of deleterious mutations in the genome. A CADD score of 20 puts the variant among the top 1% of deleterious mutations. Hi-C: SNPs with significant (FDR < 1E-6) evidence for interacting with the promoter region of LPPR1 or of another gene (figure 2 for the genes). Hi-C/EnhBrain: the subset of Hi-C SNPs that map to the enhancer regions of LPPR1 in the brain according to the Roadmap 111 epigenomes. a One SNP was shown to represent several variants in high LD (r2 > 0.9) with similar MAF, GWAS p value, and Hi-C/EnhBrain evidence. b This mutation yielded ΔΔG=−1.2, which predicts a destabilizing effect on the protein structure of LPPR1. determine its potential as a therapeutic target to impede data for replication (Parkinson, Genes, and Environment disease progression. (PAGE) study) was supported by the intramural research program of NIH National Institute of Environmental Health Author contributions Sciences grant Z01 ES101986. Funding agencies did not have Z.D. Wallen: statistical analysis, review, and critique of the a role in the design or execution of the study. manuscript. H. Chen: creation of the PAGE data set, acquisition of data, review, and critique of the manuscript. E.M. Disclosure Hill-Burns: assembly and QC of phenotype and genotype Z.D. Wallen, H. Chen, C.P. Zabetian, and E.M. Hill-Burns data, imputation, review, and critique of the manuscript. S.A. report no disclosures. S.A. Factor has received honoraria from Factor and C.P. Zabetian: creation of the NGRC data set, Neurocrine, Lundbeck, Teva, Avanir, Sunovion Pharmaceut- review, and critique of the manuscript. H. Payami: study icals, Adamas, and UCB; has received research support from concept, design and execution, creation of the NGRC data set, Ipsen, Medtronic, Teva, US WorldMeds, Sunovion Pharma- wrote the manuscript, and obtained funding. ceuticals, Solstice, Vaccinex, Voyager, the CHDI Foundation, the Michael J. Fox Foundation, and the NIH; and receives Study funding publishing royalties from Demos, Blackwell Futura, and This work was supported by the National Institute of Neuro- UpToDate. H. Payami has received research support from the logical Disorders and Stroke grant R01NS036960. Additional NIH and the University of Alabama at Birmingham. Full dis- support was provided by a Merit Review Award from the De- closure form information provided by the authors is available partment of Veterans Affairs grant 1I01BX000531; Office of with the full text of this article at Neurology.org/NG. Research & Development, Clinical Sciences Research & De- velopment Service, Department of Veteran Affairs; the Close to Received February 13, 2018. Accepted in final form June 11, 2018. the Cure Foundation; and the Sartain Lanier Family Founda- tion. Genome-wide array genotyping was conducted by the References 1. Obeso JA, Stamelou M, Goetz CG, et al. Past, present, and future of Parkinson’s Center for Inherited Disease Research, which is funded by the disease: a special essay on the 200th Anniversary of the Shaking Palsy. Mov Disord NIH grant HHSN268200782096C. Collection of samples and 2017;32:1264–1310.

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 9 2. Zareparsi S, Taylor TD, Harris EL, Payami H. Segregation analysis of Parkinson 18. Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D. Principal disease. Am J Med Genet 1998;80:410–417. components analysis corrects for stratification in genome-wide association studies. 3. Maher NE, Currie LJ, Lazzarini AM, et al. Segregation analysis of Parkinson disease Nat Genet 2006;38:904–909. revealing evidence for a major causative gene. Am J Med Genet 2002;109:191–197. 19. Aulchenko YS, Struchalin MV, van Duijn CM. ProbABEL package for genome-wide 4. McDonnell SK, Schaid DJ, Elbaz A, et al. Complex segregation analysis of Parkinson’s association analysis of imputed data. BMC Bioinformatics 2010;11:134. disease: the Mayo clinic family study. Ann Neurol 2006;59:788–795. 20. Watanabe K, Taskesen E, van Bochoven A, Posthuma D. Functional mapping and 5. Hamza TH, Payami H. The heritability of risk and age at onset of Parkinson’s disease annotation of genetic associations with FUMA. Nat Commun 2017;8:1826. after accounting for known genetic risk factors J Hum Genet 2010;55:241–243. 21. Aulchenko YS, Ripke S, Isaacs A, van Duijn CM. GenABEL: an R library for genome- – 6. Escott-Price V, Consortium IPsDG, Nalls MA, et al. Polygenic risk of Parkinson wide association analysis. Bioinformatics 2007;23:1294 1296. 22. Pruim RJ, Welch RP, Sanna S, et al. LocusZoom: regional visualization of genome- disease is correlated with disease age at onset. Ann Neurol 2015;77:582–591. wide association scan results. Bioinformatics 2010;26:2336–2337. 7. Chang D, Nalls MA, Hallgrimsdottir IB, et al. A meta-analysis of genome-wide as- 23. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and sociation studies identifies 17 new Parkinson’s disease risk loci. Nat Genet 2017;49: haplotype maps. Bioinformatics 2005;21:263–265. 1511–1516. ’ 24. Machiela MJ, Chanock SJ. LDlink: a web-based application for exploring population- 8. Nalls MA, Escott-Price V, Williams NM, et al. Genetic risk and age in Parkinson s specific haplotype structure and linking correlated alleles of possible functional var- – disease: continuum not stratum. Mov Disord 2015;30:850 854. iants. Bioinformatics 2015;31:3555–3557. ’ 9. Lill CM, Hansen J, Olsen JH, Binder H, Ritz B, Bertram L. Impact of Parkinson s 25. Payami H, Kay DM, Zabetian CP, Schellenberg GD, Factor SA, McCulloch CC. – disease risk loci on age at onset. Mov Disord 2015;30:847 850. Visualizing disease associations: graphic analysis of frequency distributions as a func- ’ 10. Pihlstrom L, Toft M. Cumulative genetic risk and age at onset in Parkinson s disease. tion of age using moving average plots (MAP) with application to Alzheimer’s and Mov Disord 2015;30:1712–1713. Parkinson’s disease. Genet Epidemiol 2010;34:92–99. 11. Hill-Burns EM, Ross OA, Wissemann WT, et al. Identification of genetic modifiers of 26. de Leeuw CA, Mooij JM, Heskes T, Posthuma D. MAGMA: generalized gene-set age-at-onset for familial Parkinson’s disease. Hum Mol Genet 2016;25:3849–3862. analysis of GWAS data. PLoS Comput Biol 2015;11:e1004219. 12. Trinh J, Gustavsson EK, Vilarino-Guell C, et al. DNM3 and genetic modifiers of age of 27. Kircher M, Witten DM, Jain P, O’Roak BJ, Cooper GM, Shendure J. A general onset in LRRK2 Gly2019Ser parkinsonism: a genome-wide linkage and association framework for estimating the relative pathogenicity of human genetic variants. Nat study. Lancet Neurol 2016;15:1248–1256. Genet 2014;46:310–315. 13. Hamza TH, Zabetian CP, Tenesa A, et al. Common genetic variation in the HLA 28. Consortium G. Human genomics. The Genotype-Tissue Expression (GTEx) pilot – region is associated with late-onset sporadic Parkinson’s disease. Nat Genet 2010;42: analysis: multitissue gene regulation in humans. Science 2015;348:648 660. 781–785. 29. van Berkum NL, Lieberman-Aiden E, Williams L, et al. Hi-C: a method to study the 14. Chen H, Huang X, Guo X, et al. Smoking duration, intensity, and risk of Parkinson three-dimensional architecture of genomes. J Vis Exp 2010:e-1869. – 30. Consortium RE, Kundaje A, Meuleman W, et al. Integrative analysis of 111 reference disease. Neurology 2010;74:878 884. – 15. Hughes AJ, Daniel SE, Kilford L, Lees AJ. Accuracy of clinical diagnosis of idiopathic human epigenomes. Nature 2015;518:317 330. 31. Quan L, Lv Q, Zhang Y. STRUM: structure-based prediction of protein stability Parkinson’s disease: a clinico-pathological study of 100 cases. J Neurol Neurosurg changes upon single-point mutation. Bioinformatics 2016;32:2936–2946. Psychiatry 1992;55:181–184. 32. Savaskan NE, Brauer AU, Nitsch R. Molecular cloning and expression regulation of PRG- 16. Schatzkin A, Subar AF, Thompson FE, et al. Design and serendipity in establishing 3, a new member of the plasticity-related gene family. Eur J Neurosci 2004;19:212–220. a large cohort with wide dietary intake distributions: the National Institutes of Health- 33. Broggini T, Schnell L, Ghoochani A, et al. Plasticity related gene 3 (PRG3) overcomes American Association of retired persons diet and Health study. Am J Epidemiol 2001; myelin-associated growth inhibition and promotes functional recovery after spinal 154:1119–1125. cord injury. Aging (Albany NY) 2016;8:2463–2487. 17. Howie BN, Donnelly P, Marchini J. A flexible and accurate genotype imputation 34. Fink KL, Lopez-Giraldez F, Kim IJ, Strittmatter SM, Cafferty WBJ. Identification of method for the next generation of genome-wide association studies. PLoS Genet intrinsic axon growth modulators for intact CNS neurons after injury. Cell Rep 2017; 2009;5:e1000529. 18:2687–2701.

10 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG ARTICLE OPEN ACCESS Increased KCNJ18 promoter activity as a mechanism in atypical normokalemic periodic paralysis

Muhidien Soufi, PhD, Volker Ruppert, PhD, Susanne Rinn´e, PhD, Tobias Mueller, MD, Bilgen Kurt, MD, Correspondence Guenter Pilz, PhD, Andreas Maieron, MD, Richard Dodel, MD, Niels Decher, PhD, and Juergen R. Schaefer, MD Dr. Schaefer [email protected] Neurol Genet 2018;4:e274. doi:10.1212/NXG.0000000000000274 Abstract Objective To identify the genetic basis of a patient with symptoms of normokalemic sporadic periodic paralysis (PP) and to study the eﬀect of KCNJ18 mutations.

Methods A candidate gene approach was used to identify causative gene mutations, using Sanger sequencing. KCNJ18 promoter activity was analyzed in transfected HEK293 cells with a luciferase assay, and functional analysis of Kir2.6 channels was performed with the two-electrode voltage- clamp technique.

Results Although we did not identify harmful mutations in SCN4A, CACNA1S, KCNJ2 and KCNE3, we detected a monoallelic four-fold variant in KCNJ18 (R39Q/R40H/A56E/I249V), together with a variant in the respective promoter of this channel (c.-542T/A). The exonic variants in Kir2.6 did not alter the channel function; however, luciferase assays revealed a 10-fold higher promoter activity of the c.-542A promoter construct, which is likely to cause a gain-of-function by increased expression of Kir2.6. We found that reducing extracellular K+ levels causes a paradoxical reduction in outward currents, similar to that described for other inward rectifying K+ channels. Thus, reducing the extracellular K+ levels might be a therapeutic strategy to antagonize the transcriptionally increased KCNJ18 currents. Consistently, treatment of the patient with K+ reducing drugs dramatically improved the health situation and prevented PP attacks.

Conclusions We show that a promoter defect in the KCNJ18 gene is likely to cause periodic paralysis, as the observed transcriptional upregulation will be linked to increased Kir2.6 function. This concept is further supported by our observation that most of the PP attacks in our patient disappeared on medical treatment with K+ reducing drugs.

From the Center for Undiagnosed and Rare Diseases (ZusE) (M.S., T.M., B.K., J.R.S.); Department of Cardiology (V.R.) and Department of Neurology (R.D.), University Hospital Giessen and Marburg; Vegetative Physiology (S.R., N.D.), Philipps-University Marburg, Institute of Physiology and Pathophysiology, & Marburg Center for Mind, Brain and Behavior, Marburg, Germany; Institute for Algebra (G.P.), Johannes Kepler University Linz; and Department of Gastroenterology (A.M.), Hospital Elisabethinen, Linz, Austria.

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 © 2018 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary ATS = Andersen-Tawil-Syndrome; EVS = Exome Variant Server; MAF = minor allele frequency; PP = periodic paralysis; PCR = polymerase chain reaction; SNV = single-nucleotide variation; SPP = sporadic PP; TEVC = two-electrode voltage- clamp; TPP = thyrotoxic PP.

Familiar forms of periodic paralysis (PP) belong to a group of genes of PP, as recommended by the guidelines of the Ger- skeletal muscle disorders characterized by episodes of muscle man Neurological Society (DGN). These included the fol- paralysis, accompanied by decreased (hypoPP) or increased lowing: SCN4A (NM_000334), CACNA1S (NM_000069), (hyperPP) serum potassium levels, which are associated with KCNJ2 (NM_000891,) KCNE3 (NM_005472), and KCNJ18 – mutations in the genes encoding the skeletal muscle isoforms (NM_001194958).10 12 Novel detected SNVs were further of the voltage-gated sodium channel SCN4A or the respective assessed for their appearance in a cohort of 100 controls. – calcium channel CACNA1S.1 3 Andersen-Tawil-Syndrome (ATS) that results from loss-of-function mutations in the Functional analysis of detected mutations inward rectifying Kir2.1 channel is characterized by a unique triad of symptoms, which include PP, cardiac arrhythmias, as Measurements of promoter activity in luciferase well as facial and skeletal malformations. In the skeletal assays muscle, the loss-of-function mutations in Kir2.1 are believed We analyzed a 766-bp fragment of the KCNJ18 promoter up- to cause a depolarization of the membrane potential.4 stream of exon 1 for promoter activity. Wild-type or mutant KCNJ18 promoter constructs were generated by cloning a 766- bp polymerase chain reaction (PCR) fragment from the The most common form of sporadic PP (SPP) is the thyro- ’ toxic PP (TPP), which is clinically similar to hypoPP or ATS. patient s genomic DNA containing KpnI/NcoI linkers into the TPP can result from mutations in the KCNJ18 gene, encoding corresponding sites of pGL3-Basic (Promega). Transient – Kir2.6 channels5 7 that are under transcriptional regulation of transfection assays in HEK293 cells were performed in six-fold 8 reactions using HeLafect (OZ Biosciences) with 0.5 μg KCNJ18 thyroid hormones. TPP-causing mutations in Kir2.6 include μ loss-of-function mutations, similar to that in Kir2.1 and ATS, pGL3 construct and 0.1 g internal control plasmid pcDNA3.1 but also gain-of-function mutations8 that are believed to hy- CMV-LacZ (Invitrogen). Cells were maintained in Dulbecco modified Eagle medium supplemented with 10% fetal bovine perpolarize the skeletal muscle, leading to a reduced excit- μ ability, as the membrane potential is too far from the action serum and penicillin/streptomycin (100 g/mL). Forty-eight potential threshold.3 hours after transfection, the cells were harvested and cellular luciferase expression was measured using the luciferase assay — While analyzing a case of sporadic normokalemic PP, we system (Promega) on a Glow-Max instrument (Promega). identified a novel disease-causing mechanism, with a heterozy- The assay was replicated 8 times in independent transfection gous variant in the KCNJ18 promoter that generates a strong experiments. Site-directed mutagenesis using the Gene Tailor increase in KCNJ18 transcription, which is likely to cause Site-Directed Mutagenesis System (Invitrogen) was used to validate functional mutations from luciferase assays. For this a hyperpolarization of the skeletal muscle membrane potentials, ff leading to inexcitability under normokalemic conditions. purpose, the deleterious e ect of the mutation was restored by introduction of the corresponding normal base into the mutant KCNJ18 promoter construct.

Methods Two-electrode voltage-clamp recordings of Kir2.6 Standard protocol approvals, registrations, channels in Xenopus oocytes and patient consents Ovarian lobes were dissected from mature Xenopus laevis Informed consent was provided by the patient and his rela- toads anesthetized with tricaine and treated with collagenase tives for genetic screening, as well as for publication of the (1 mg/mL, Worthington, type II) in OR2 solution (NaCl data in its present form. Controls (n = 100) without any 82.5 mM, KCl 2 mM, MgCl2 1 mM, HEPES 5 mM, pH 7.4) neurologic signs of paralysis were tested for the identiﬁed for 120 minutes. Isolated oocytes were stored at 18°C in 9 single-nucleotide variations (SNVs) of our index patient. ND96 recording solution (NaCl 96 mM, KCl 2 mM, CaCl2 The work was approved by the local ethics committee of the 1.8 mM, MgCl2 1 mM, HEPES 5 mM, pH 7.5) supplemented Philipps-University Marburg (study 10/03), and all proce- with Na-pyruvate (275 mg/L), theophylline (90 mg/L), and dures were in accordance with the Declaration of Helsinki of gentamicin (50 mg/L). Wild-type KCNJ18 and mutant 1975, as revised in 1996. channels were subcloned into an oocyte expression vector and cDNA was linearized. cRNA was synthesized with mMES- Screen for genomic mutations SAGE mMACHINE-Kit (Ambion). The quality of cRNA was For mutation analysis, we Sanger sequenced the genomic tested using gel electrophoresis. Oocytes were injected with DNA of the patient for mutations in all relevant candidate 50 nL (5 ng) of the wild-type KCNJ18 or a cRNA containing

2 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG the single-nucleotide variants (SNVs) of the patient. Standard fear of death. In these situations, the patient was always clear two-electrode voltage-clamp (TEVC) experiments were per- minded during these attacks. He was fully aware of the situ- formed at room temperature (21–22°C) with an Axoclamp ation but was unable to express himself verbally because of 900A amplifier, a Digidata 1440A, and pClamp10 software severe motoric dysarthria. Impairment of the muscles re- (Axon Instruments) 2 days after injection of oocytes with sponsible for speech is atypical for PP. Numerous EEGs, as cRNA. Microelectrodes were fabricated from glass pipettes well as EMGs, were without signs for epilepsy or myotonia, filled with 3-M KCl and had a resistance of 0.3–1.0 MΩ.To respectively. The family history was negative for any neuro- examine the effect of furosemide (10 μM, Sigma), currents logic disorders. The mother died at the age of 70 years from were elicited by a voltage step from a holding potential of stomach cancer, and the father died at the age of 77 years. The −100 to −150 mV, followed by a ramp to +40 mV in 0.8 brother of the patient died at the age of 61 years from stomach seconds (and a final voltage step to 0 mV, not illustrated). The and bone cancer. His children are both noncarriers of the blocker was applied in the ND96 recording solution. c.-542T/A mutation and showed no clinical symptoms.

Statistical analysis Despite various examinations by neurologists, psychiatrists, SigmaPlot/SigmaStat software (Systat Software) was used for internists, and general practitioners in the past decades, no statistical analyses. For luciferase expression analysis, the conclusive diagnosis was made. Repeated measurements of se- − continuous variables were expressed as mean and standard rum K+,Na+,Ca++,Cl, and glucose before and during these deviations. In TEVC experiments, data were expressed as attacks were found to be within the normal range. Treatment mean and SEM. Continuous variables were compared with trials with antidepressants, neuroleptic and anticonvulsant ther- the unpaired Student t test if they were normally distributed apy, as well as psychotherapeutic treatment, were unsuccessful. or the Mann-Whitney U test otherwise. For all tests, a p value ≤ 0.05 (*) was considered statistically significant and p values Detection of nonsynonymous SNVs in the ≤ 0.001 (***) as highly significant. SCN4A, CACNA1S, and KCNJ18 genes We consecutively screened the genes of the voltage-gated sodium channel type IV (SCN4A), the voltage-dependent Results L-type calcium channel 1S (CACNA1S), the potassium voltage-gated channel subfamily E regulatory subunit 3 Clinical case description (KCNE3), and the inwardly rectifying potassium channels J2 The patient described here is a 69-year-old man with more (KCNJ2) and J18 (KCNJ18) in the patient. DNA sequencing than a 50-year-long history of severe attacks of weakness in detected nine coding variants in the following genes: SCN4A, both, legs and arms, speech disturbances, as well as recurring CACNA1S, and KCNJ18 (table 1). Variation analysis at the extreme muscle tiredness mostly in the late afternoon. These Exome Variant Server (EVS) revealed a high minor allele attacks may lead to a tetraplegia sometimes associated with frequency (MAF) >5% and no deleterious effects of the

Table 1 Detected coding variants in the following genes: SCN4A, CACNA1S, and KCNJ18

Gene/SNV Nucleotide Amino acid SIFT Polyphen2 MAF

SCN4A

rs6504191 c.1570A>G p.S524G Tolerated (0.43) Benign (0.0) 0.06

rs2058194 c.4126A>G p.N1376D Tolerated (1.00) Benign (0.0) 0.44

CACNA1S

rs12742169 c.1373T>A p.L458H Tolerated (0.19) Benign (0.227) 0.13

rs12139527 c.5399T>C p.L1800S Deleterious—low confidence (0) Probably damaging (0.999) 0.23

KCNJ18

rs n.a. c.116G>A p.R39Q n.a. Benign (0.0) —

rs n.a. c.119G>A p.R40H n.a Benign (0.0) —

rs n.a. c.167C>A p.A56Q n.a Benign (0.0) —

rs n.a. c.576G>C p.Q192H n.a Benign (0.0) —

rs n.a. c.745A>G p.I249V n.a Benign (0.0) —

Abbreviations: MAF = minor allele frequency; n.a. = not available; SNV = single-nucleotide variation.

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 3 variants as predicted by SIFT and Polyphen2, indicating that Allelic variation in the KCNJ18 promoter these variants are naturally occurring polymorphisms. No Because the variants in the KCNJ18 channel did not affect EVS data were available for variants of KCNJ18 (table 1). mean current amplitudes, we assessed the promoter regions of However, the four-fold SNV in Kir2.6, which is located on SCN4A, CACNA1S, KCNJ2, and KCNE3 for SNVs based on a single allele of our patient, has been recently identified as published promoter sequence data from GeneCards.14 a natural variant, which apparently has no functional differ- However, no sequence information for the KCNJ18 promoter ence to wild-type channels.13 was available. Therefore, we designed and sequenced a 766-bp PCR fragment according to the mapped and published pro- The four-fold nonsynonymous SNV in the moter sequence of the KCNJ18 gene.8 No SNVs were found patient’s KCNJ18 gene does not alter the in the promoters of SCN4A, CACNA1S, KCNJ2, and KCNE3. electrophysiologic characteristics of Kir2.6 Sequencing analysis of the KCNJ18 promoter detected 13 The four-fold SNV in Kir2.6 R39Q/R40H/A56E/I249V heterozygous SNVs. To elucidate the overall frequencies of was previously studied under high extracellular potassium these SNVs, we sequenced the DNA of 200 control alleles and concentration (20 mM), so only inward currents could be detected a total of 14 polymorphic sites within the 766-bp analyzed and not the physiologically relevant outward KCNJ18 promoter fragment. However, the variant c.-542T>A 13 conductance and physiologic rectification pattern. was detected only in the genomic DNA of our patient (table Therefore, we recorded the variant under physiologic 2). A search for population diversity of detected variants in saline conditions by injecting similar amounts of cRNA in Ensembl and gnomAD revealed no frequency data. However, Xlaevisoocytes and recording currents 48 hours after we found entries by searching for small variations in dbSNP injection. In TEVC recordings, the current rectification based on data from NHLBI TOPMed (Trans-Omics for was indistinguishable between the two KCNJ18 variants Precision Medicine [TOPMed]) including >54,000 whole (figure 1, A and B). A representative current trace is il- genomes.15 Thirteen of our detected SNPs were common lustrated for the wild-type KCNJ18 (black) and the with high MAFs, whereas the c.-542T>A SNP was not listed. KCNJ18 SNVs (gray), respectively (figure 1A). The SNVs did not alter the rectification pattern or current ampli- Promoter variant c.-542A strongly increases tudes (figure 1, B and C), and the mean inward current KCNJ18 transcription amplitude analyzed at −140 mV was similar for the wild- We determined the functional relevance of the c.-542T>A type KCNJ18 and the SNVs (figure 1C), but no sub- KCNJ18 promoter SNV with luciferase assays in transfected stantial differences in physiologically relevant outward HEK293 cells. Compared with the wild-type promoter, the currents were observed (figure 1B). c.-542T/A promoter variant produced a tenfold higher

Figure 1 Voltage-clamp recording of wild-type KCNJ18 and the KCNJ18 SNVs expressed in Xenopus laevis oocytes

(A) Representative current traces of wild-type KCNJ18 channels (upper panel) or KCNJ18 channels containing the SNVs R39Q/R40H/A56E/I249V (lower panel). For these experiments, 5 ng of the respective cRNA was injected and the indicated voltage protocol was applied. Currents were elicited by a voltage step from a holding potential of −100 to −150 mV, followed by a ramp to +50 mV in 0.8 seconds. (B) Mean current-voltage relationship of wild-type KCNJ18 (black) or the KCNJ18 construct containing the SNVs of the patient (gray) n = 13. (C) Relative current amplitudes at −140 mV normalized to wild-type KCNJ18. Values are expressed as mean ± SEM. n.s. = no of significant difference; SNV = single-nucleotide variation.

4 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Table 2 Detected single-nucleotide variations (SNVs) in the KCNJ18 promoter (n = 100 controls)a

Promoter SNV Genotype (n) Index patient c.-405 C/A (M) C/C (16) C/A (84) A/A (0) C/A c.-452 C/T (Y) C/C (3) C/T (97) T/T (0) C/T c.-542 T/A (W) T/T (100) T/A (0) A/A (0) T/A c.-578 G/A (R) G/G (87) G/A (13) A/A (0) G/A c.-585 G/A (R) G/G (77) G/A (19) A/A (4) G/A c.-645 C/T (Y) C/C (54) C/T (46) T/T (0) C/T c.-740 C/G (S) C/C (62) C/G (38) G/G (0) C/G c.-793 C/T (Y) C/C (35) C/T (58) T/T (7) C/T c.-835 G/A (R) G/G (37) G/A (63) A/A (0) G/A c.-859 A/G (R) A/A (23) A/G (77) G/G (0) A/G c.-861 G/C (S) G/G (50) G/C (50) C/C (0) G/G c.-903 G/A (R) G/G (0) G/A (100) A/A (0) G/A c.-1009 G/T (K) G/G (0) G/T (100) T/T (0) G/T c.-1053 G/A (R) G/G (13) G/A (87) A/A (0) G/A a Positions are numbered according to the mapped and published KCNJ18 promoter sequence from Ryan et al.8 The mutated base found in the patient is indicated in bold.

luciferase activity (p < 0.001). This effect was completely restored to normal wild-type promoter activity, when we in- Figure 2 Luciferase assays of transiently transfected HEK 293 cells with reporter constructs containing the troduced the corresponding wild-type base T at position c.- KCNJ18 542 in the mutant promoter (figure 2). These results show 766-bp wild-type or c.-542 T/A mutant promoter that the c.-542A KCNJ18 promoter SNV increases KCNJ18 transcription tremendously and thus might be the key to the phenotype of SPP in our patient. KCNJ18 currents have an extracellular potassium dependency Inward rectifier K+ (Kir) channels play an important role in maintaining stable resting membrane potentials, thus controlling excitability, shaping the initial depolarization, and the final repolarization of action potentials in many cell types. It has been known for several decades that Kir-mediated outward currents increase paradoxically with increasing extra- + 16 cellular potassium (K o) concentration. However, to our knowledge, it is not known whether KCNJ18 also has this paradoxical activation by extracellular potassium. Thus, we studied the extracellular potassium dependency of wild-type KCNJ18 and the SNVs using TEVC experiments and sys- + tematically changing the K o concentration. Note that in the oocyte expression system, because of a lower overall osmo- + Activation of the promoter-driven luciferase in HEK293 cells by wild-type larity, the physiologic intracellular K is 96 mM and the ex- and mutated constructs, as well as in mutated constructs after intracellular K+ is reflected by 2 mM K+ . When K+ was troduction of wild-type base T at position c.-542 in the mutant construct. An o o approximately 10-fold increased luciferase activation could be detected only gradually increased between 0 and 4 mM, the outward current in the c.-542A construct. Experiments were performed in a six-fold approach, and mean and standard deviations are shown. *** indicates t = increased, despite a reduction in the electrochemical gradient 16.983, p < 0.001 vs wild-type KCNJ18 and t = 17.855, p < 0.001 vs c.-542T, (figure 3A). This effect was similarly observed for wild-type respectively. KCNJ18 (left) and the KCNJ18 containing four-fold SNV

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 5 Figure 3 Voltage-clamp recording of the wild-type KCNJ18 (left) and the KCNJ18 carrying the patient’s SNVs (right) expressed in Xenopus laevis oocytes

(A) Different extracellular K+ concentrations of 0–4 mM and (B) under normokalemic, hypokalemic (50%), and no potassium conditions, analyzed in the voltage range of −40 to −80 mV. Data are presented as mean ± SEM. *** = p < 0.001; ns = not significant; SNV = single-nucleotide variation.

(right). In the physiologic range of the resting membrane 4.1 mmol/L. A few hours later, he had a paralysis attack and potential of the skeletal muscle, the outward current was the potassium level was at this point 4.2 mmol/L. Three days clearly reduced under hypokalemic compared with normo- after the initiation of furosemide 20 mg per day, the patient kalemic conditions (figure 3B). Figure 3 illustrates a strong felt much better, had no attacks, and was able to walk long reduction of the outward currents at −80 and −70 mV for the distances. The potassium level in the following days was 3.4, wild-type KCNJ18 (left) and the SNVs (right) when extra- 3.1, and 3.0 mmol/L, respectively. The chloride levels were cellular K+ was reduced. The current reduction at 50% hy- within the normal range. Since the diagnosis, the patient was + pokalemic conditions (K o = 1 mM instead of 2 mM) and no able to lower his potassium level under tight control to a mean + extracellular potassium (K o = 0 mM) conditions was 23% level of 3.5 mmol/L (SD ±0.46; n = 234). He felt clearly better and 46% for wild-type KCNJ18 or 22% and 44% for the on days with potassium levels below normal and had no severe KCNJ18 containing the SNVs, respectively. Thus, reducing paralysis attacks for more than 2 years. + the K o concentration can be used to antagonize the increased KCNJ18 currents caused by the gain-of-function in the pro- Furosemide does not alter KCNJ18 currents moter of the channel. Since the treatment of the patient with furosemide ame- liorated symptoms of PP, we analyzed whether this drug Clinical treatment of the patient in relation to may directly antagonize the KCNJ18 gain-of-function experimental findings resulting from the increased transcription. Furosemide We studied the Na+ and K+ levels in more detail, and our 10 μM was applied to oocytes expressing either wild-type patient recorded the electrolyte content of his food closely. KCNJ18 or the SNVs of our patient. In TEVC recordings, From this information, in agreement with our KCNJ18 pro- neither wild-type KCNJ18 nor the KCNJ18 containing the moter findings, we decided to test the effect of lowering po- SNVs was blocked by furosemide (figure 4), supporting an tassium in vivo by the administration of furosemide (starting indirect effect of furosemide by reducing serum potassium with 20 mg per day) during a short inpatient stay in our levels, which antagonizes the Kir2.6 gain-of-function in our hospital. Our patient was admitted with a potassium level of patient.

6 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG PP by influencing Kir2.6 current levels in the skeletal muscle. Figure 4 Voltage-clamp recordings of KCNJ18 carrying the Kir channels play a key role in the pathogenesis of PP, as they ’ Xenopus laevis patient s SNVs expressed in set the resting membrane potentials, thus controlling excit- oocytes under control conditions and after ap- ability. Increased Kir2.6 expression is likely to hyperpolarize plication of 10 μM furosemide the membrane potentials in the skeletal muscle cells, leading to a reduced excitability. It is well known that the outward currents of Kir channels are paradoxically increased when + 18 extracellular Ko concentration is elevated. However, it was not known whether Kir2.6 channels also have this paradoxical activation by extracellular potassium. Both gain-of-function mutations and loss-of-function mutations can cause TPP. For gain-of-function mutations, it was proposed that a hyperpolarization by increased IKir currents leads to a hyperpolarization and difficulties to reach the action potential threshold, ultimately resulting in an altered muscle excitability, weakness, and/or paralyses.8 Therefore, we tested the extracellular potassium dependence of Kir2.6 with and without the four- fold nonsynonymous SNVs. Our data obtained in vitro (Kir2.6 measurements) and in vivo (effects of furosemide by lowering plasma K+ and dietary control of potassium intake) SNV = single-nucleotide variation. show that reducing IKir currents is in our patient protective against further attacks, as it should antagonize the hyperpolarization caused by increased Kir2.6 transcription and IKir Discussion currents. Kir2.6 channels are believed to act as a “brake” or suppressor of inward rectifier currents. Kir2.6 is largely In this report, we describe a novel pathophysiologic mecha- retained in the ER where it can bind and sequester Kir2.1 and nism for sporadic normokalemic PP in a patient by a gain-of- Kir2.2 to reduce the surface expression of inward rectifier K+ function variant in the promoter of the KCNJ18 gene. We channels. Therefore, the increased expression of Kir2.6 might identified a variation in position c.-542 of the KCNJ18 pro- reduce Kir2.x currents.19 This model would be consistent with moter, which dramatically (>tenfold) increases KCNJ18 the loss-of-function defects of Kir2.x channels in Andersen- transcription. The c.-542T>A KCNJ18 SNV was not found in Tawil syndrome. However, low extracellular K+ levels are 200 control alleles of healthy individuals and is also not believed to reduce Kir2.x outward currents in muscle cells. reported in several databases.15,17 We detected no other pu- Thus, our approach of keeping plasma K+ levels low would tative genetic defects in PP-associated ion channels or their rather aggravate symptoms than being beneficial, if the subunits, indicating that the KCNJ18 promoter variant is the disease-causing mechanism would already include a Kir2.x most likely disease-relevant mutation. However, transfection reduction–mediated depolarization. Moreover, the pheno- experiments were performed only in HEK293 cells; it would type of our patient is rather atypical and does not match with be beneficial to confirm KCNJ18 promoter activity in a muscle an ATS-like PP, making a similar disease-causing mechanism biopsy of the patient. But because of the strong results of the less likely. An alternative explanation why the patient im- cell culture experiments and because the patient is stable and proved on furosemide would be the fact that inhibition of the − feels much better under the current medication, we have NKCC cotransporter prevents Cl influx into muscle cells and − refrained from a biopsy. KCNJ18 was described as an inwardly thereby lowers myoplasmic [Cl ], which is protective against rectifying potassium channel in the skeletal muscle, and attacks of PP.20 However, although this is clearly a beneficial mutations in this gene were identified in patients with TPP.8 mechanism for PP attacks with depolarized membrane This study also reported that the KCNJ18 gene contains a 59- potentials, the therapeutical effect is in our case more likely upstream thyroid hormone response element at position related to the furosemide-mediated lowering of the plasma c.-594, which enhances transcription by thyroid hormones. K+ level, which antagonizes the Kir2.6 gain-of-function. This The authors proposed that mutations in the gene of Kir2.6 in is supported by our observation that the K+ plasma levels, combination with transcriptional upregulation of the KCNJ18 which we monitored, correlate with the occurrence of PP gene by T3 thyroid hormones can cause TPP, predisposing attacks and the fact that the incidence of the attacks can be patients to muscle weakness and paralysis. The authors titrated not only by furosemide but also by the dietary of reported that TPP might involve as a secondary cause of potassium intake. increased KCNJ18 promoter activity; our study describes for the first time a mutation in the KCNJ18 promotor as a primary Perhaps the most important regulatory mechanism that is cause of an atypical form of PP.8 Considering the effects of involved in the pathogenesis of TPP is the transcriptional thyroid hormone–dependent KCNJ18 transcription in TPP, it regulation of the KCNJ18 gene through the action of thyroid becomes evident that KCNJ18 promoter variants may cause hormones via a thyroid hormone response element at

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 7 position c.-594 in the KCNJ18 promoter. Because our supervision, writing, drafting, and revising the manuscript for c.-542T>A mutation is not localized in this thyroid response intellectual content. element, we searched for other transcription factor binding sites in the KCNJ18 promoter. In silico analysis for poten- Acknowledgment tial transcription factor binding sites21 revealed that the The authors are grateful for the participation and support of c.-542T>A variation replaces an androgen receptor binding the patient and his family to perform this study. site (AR) by an estrogen receptor element (ER). In PP, the attacks increase in frequency and severity during puberty, Study funding adolescence, and early adulthood. This may be caused by Dr. Reinfried Pohl Foundation provided financial support. a change in hormone levels. Regulatory variants that affect clinical phenotypes have been identified in more than 700 genes so far, and most of these variations are located in Disclosure fi proximal and distal promoter elements that map within 1 kb M. Sou , V. Ruppert, and S. Rinné report no disclosures. T. of the transcription start site.22,23 The impact of SNVs in Mueller is an employee of Rhön Klinikum AG. B. Kurt, G. promoters and enhancer elements on voltage-gated channel Pilz, and A. Maieron report no disclosures. R. Dodel has fi expression has been already shown for the SCN5A/SCN10A served on scienti c advisory boards of Octapharma and Lilly; – fi locus in arrhythmia susceptibility and Brugada syndrome.24 26 has received travel funding/speaker honoraria from P zer, Therefore, our findings support that in PP, promoter muta- Lundbeck, Merz, Solvay, AstraZeneca, CSL Behring, Octa- tions can be disease causative by altering the current densities pharma, Orion Pharma, UCB, Teva, and BMBF; receives of ion channels. publishing royalties from Thieme Verlag Stuttgart; has been a consultant for Octapharma and Lilly; has received research There has been a controversial debate to consider KCNJ18 as support from BMBF, JPND (BMBF), UKGM Inhouse grants, a PP gene because of its large homology with KCNJ12, and Behring Röntgen Stiftung, the Michael J. Fox Foundation, conflicting results about the pathogenicity of rare variants in Deutsche Parkinson-Vereinigung, the International Parkin- these genes for the diagnosis of PP exist, as shown in a recently son Foundation, Faber Stiftung, the Movement Disorders published work.27 We detected several promoter SNVs and Society, Hector-Stiftung, and Alzheimer Forschungsinitiative; functionally tested their impact in transfection experiments. receives revenue from patents (which are held by the Uni- Consistent with our findings, a recently published article has versity of Marburg); and holds the following patents: Method shown that RYR-1 mutations can cause atypical PP pheno- of detecting progression of a neurodementing disease; types in patients in whom mutations in SCN4A, CACNA1S, Method of treatment of neurodementing diseases using iso- and KCNJ2 have been excluded.28 Here, we did not screen the lated, monoclonal, human, anti-b-amyloid antibody; Use of RYR-1 gene for mutations. Nevertheless, from the clinical tetracyclines as neuro-protective agents and for the treatment ’ symptoms, our patient can be distinguished from the RYR-1 of Parkinson s disease and related disorders; Human beta- ’ phenotype as he lacks symptoms of exertional myalgia, amyloid antibody and use thereof for treatment of Alzheimer s cramps, or ophthalmoplegia. Given the increasing use of disease; Human monoclonal anti-amyloid-beta antibodies; whole-exome sequencing approaches to identify disease- Synthetische Liganden für humane anti-aß-Antikörper; Di- ’ causing genes and mechanisms, this first description of agnosis, prophylaxis and therapy of Alzheimer s disease and promoter-induced dysregulation of KCNJ18 in PP highlights other neurodementing disorders; Naturally occurring auto- the need to analyze nonexonic genomic patients’ DNA with antibodies against alpha-synuclein that inhibit the aggregation atypical phenotypes of PP. However, because we report on and cytotoxicity of alpha-synuclein; and Verfahren, insbe- a unique case so far, further studies are needed to proof sondere enzyme-linked immunosorbent assay (ELISA), zum our—as we think—intriguing concept. in vitro Nachweis von Amyloid beta Autoantikörpern, Mik- rotiterplatte und Testkit. N. Decher reports no disclosures. fi fi Author contributions J.R. Schaefer serves as a scienti c advisor for MSD, Sano - M. Soufi and V. Ruppert: genetic analysis and cell culture Genzyme, and Amgen; received lecture fees from MSD, fi experiments, writing and drafting the manuscript, and study Sano -Genzyme, Novartis Academy, Synlab Academy, and concept. S. Rinné: voltage clamp experiment data collection Berlin-Chemie; serves on the editorial board of Der Internist and analysis, revising the manuscript for content, and data (Springer Nature Verlag); and has received research support interpretation. T. Mueller, B. Kurt, and A. Maieron: clinical from the Dr. Reinfried Pohl Foundation. Full disclosure form examination of the patient, revising the manuscript for con- information provided by the authors is available with the full tent, and data interpretation. G. Pilz: mathematical analysis, text of this article at Neurology.org/NG. revising the manuscript for content, and data interpretation. Received March 13, 2018. Accepted in final form July 2, 2018. R. Dodel: clinical examination of the patient, writing and revising the manuscript for content, and data interpretation. References N. Decher: voltage clamp data collection and analysis, 1. Lehmann-Horn F, Jurkat-Rott K. Voltage-gated ion channels and hereditary disease. Physiol Rev 1999;79:1317–1372. writing, drafting and revising the manuscript for intellectual 2. Burge JA, Hanna MG. Novel insights into the pathomechanisms of skeletal muscle content, and study concept. J.R. Schaefer: study concept and channelopathies. Curr Neurol Neurosci Rep 2012;12:62–69.

8 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG 3. Tricarico D, Camerino DC. Recent advances in the pathogenesis and drug action in 17. Ensembl. Available at: ensembl.org/Homo_sapiens/Gene/Summary?db=core; periodic paralyses and related channelopathies. Front Pharmacol 2011;2:8. g=ENSG00000260458;r=17:21692523-21704612;t=ENST00000567955 Accessed 4. Plaster NM, Tawil R, Tristani-Firouzi M, et al. Mutations in Kir2.1 cause the de- April 23, 2018. velopmental and episodic electrical phenotypes of Andersen’s syndrome. Cell 2001; 18. Jurkat-Rott K, Weber MA, Fauler M, et al. K+-dependent paradoxical membrane 105:511–519. depolarization and Na+ overload, major and reversible contributors to weakness by 5. Falhammar H, Thoren M, Calissendorff J. Thyrotoxic periodic paralysis: clinical and ion channel leaks. Proc Natl Acad Sci U S A 2009;106:4036–4041. molecular aspects. Endocrine 2013;43:274–284. 19. Dassau L, Conti LR, Radeke CM, et al. Kir2.6 regulates the surface expression of 6. Cheng CJ, Lin SH, Lo YF, et al. Identification and functional characterization of Kir2.6 Kir2.x inward rectifier potassium channels. J Biol Chem 2011;286:9526–9541. mutations associated with non-familial hypokalemic periodic paralysis. J Biol Chem 20. Wu F, Mi W, Cannon SC. Beneficial effects of bumetanide in a CaV1.1-R528H mouse 2011;286:27425–27435. model of hypokalaemic periodic paralysis. Brain 2013;136:3766–3774. 7. Zheng J, Liang Z, Hou Y, et al. A novel Kir2.6 mutation associated with hypokalemic 21. Messeguer X, Escudero R, FarréD,Núñez O, Mart´ınez J, Albà MM. PROMO: periodic paralysis. Clin Neurophysiol 2016;127:2503–2508. detection of known transcription regulatory elements using species-tailored searches. 8. Ryan DP, da Silva MR, Soong TW, et al. Mutations in potassium channel Kir2.6 cause Bioinformatics 2002;18:333–334. susceptibility to thyrotoxic hypokalemic periodic paralysis. Cell 2010;140:88–98. 22. Epstein DJ. Cis-regulatory mutations in human disease. Brief Funct Genomic Pro- 9. Soufi M, Sattler AM, Maerz W, et al. A new but frequent mutation of apoB-100-apoB teomic 2009;8:310–316. His3543Tyr. Atherosclerosis 2004;174:11–16. 23. De GM, Viprakasit V, Hughes JR, et al. A regulatory SNP causes a human 10. DGN-Leitlinien. Available at: dgn.org/leitlinien/3457-030-055-myotone-dystro- genetic disease by creating a new transcriptional promoter. Science 2006;312: phien-nicht-dystrophe-myotonien-und-periodische-paralysen-2017#definition. 1215–1217. Accessed April 23, 2018. 24. Bezzina CR, Shimizu W, Yang P, et al. Common sodium channel promoter haplotype 11. Jurkat-Rott K, Lehmann-Horn F. Genotype-phenotype correlation and therapeutic in Asian subjects underlies variability in cardiac conduction. Circulation 2008;113: rationale in hyperkalemic periodic paralysis. Neurotherapeutics 2007;4:216–224. 338–344. 12. GHR. Available at: ghr.nlm.nih.gov/search?query=periodic+paralysis. Accessed April 25. van den Boogaard M, Barnett P, Christoffels VM. From GWAS to function: genetic 23, 2018. variation in sodium channel gene enhancer influences electrical patterning. Trends 13. Paninka RM, Mazzotti DR, Kizys MM, et al. Whole genome and exome sequencing Cardiovasc Med 2014;24:99–104. realignment supports the assignment of KCNJ12, KCNJ17, and KCNJ18 paralogous 26. Park JK, Martin LJ, Zhang X, Jegga AG, Benson DW. Genetic variants in SCN5A genes in thyrotoxic periodic paralysis locus: functional characterization of two poly- promoter are associated with arrhythmia phenotype severity in patients with het- morphic Kir2.6 isoforms. Mol Genet Genomics 2016;291:1535–1544. erozygous loss-of-function mutation. Heart Rhythm 2012;9:1090–1096. 14. GeneCards. Available at: genecards.org/cgi-bin/carddisp.pl?gene=KCNJ18. Accessed 27. Kuhn M, Jurkat-Rott K, Lehmann-Horn F. Rare KCNJ18 variants do not explain April 23, 2018. hypokalaemic periodic paralysis in 263 unrelated patients. J Neurol Neurosurg Psy- 15. dbSNP. Available at: ncbi.nlm.nih.gov/snp/?term=kcnj18. Accessed April 23, 2018. chiatry 2016;87:49–52. 16. Leech CA, Stanfield PR. Inward rectification in frog skeletal muscle fibres and its de- 28. Matthews E, Neuwirth C, Jaffer F, et al. Atypical periodic paralysis and myalgia: pendence on membrane potential and external potassium. J Physiol 1981;319:295–309. a novel RYR1 phenotype. Neurology 2018;90:e412–e418.

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 9 ARTICLE OPEN ACCESS Bioenergetics in ﬁbroblasts of patients with Huntington disease are associated with age at onset

Sarah L. Gardiner, MD,* Chiara Milanese, PhD, Merel W. Boogaard, BSc, Ronald A.M. Buijsen, PhD, Correspondence Marye Hogenboom, MSc, Raymund A.C. Roos, MD, PhD, Pier G. Mastroberardino, PhD, Ms. Gardiner [email protected] Willeke M.C. van Roon-Mom, PhD, and N. Ahmad Aziz, MD, PhD*

Neurol Genet 2018;4:e275. doi:10.1212/NXG.0000000000000275 Abstract Objective We aimed to assess whether diﬀerences in energy metabolism in ﬁbroblast cell lines derived from patients with Huntington disease were associated with age at onset independent of the cytosine-adenine-guanine (CAG) repeat number in the mutant allele.

Methods For this study, we selected 9 pairs of patients with Huntington disease matched for mutant CAG repeat size and sex, but with a difference of at least 10 years in age at onset, using the Leiden Huntington disease database. From skin biopsies, we isolated fibroblasts in which we (1) quantified the ATP concentration before and after a hydrogen-peroxide challenge and (2) measured mitochondrial respiration and glycolysis in real time, using the Seahorse XF Extra- cellular Flux Analyzer XF24.

Results The ATP concentration in fibroblasts was significantly lower in patients with Huntington disease with an earlier age at onset, independent of calendar age and disease duration. Maximal respiration, spare capacity, and respiration dependent on complex II activity, and indices of mitochondrial respiration were significantly lower in patients with Huntington disease with an earlier age at onset, again independent of calendar age and disease duration.

Conclusions A less efficient bioenergetics profile was found in fibroblast cells from patients with Huntington disease with an earlier age at onset independent of mutant CAG repeat size. Thus, differences in bioenergetics could explain part of the residual variation in age at onset in Huntington disease.

*Completed the statistical analysis.

From the Department of Neurology (S.L.G., M.H., R.A.C.R., N.A.A.), Department of Human Genetic (S.L.G., R.A.M.B., W.M.C.v.R.-M.), and Department of Clinical Genetics (M.W.B.), Leiden University Medical Centre, Leiden; Department of Molecular Genetics (C.M., P.G.M.), Erasmus Medical Centre, Rotterdam, The Netherlands; and German Center for Neurodegenerative Diseases (DZNE) (N.A.A.), Bonn, Germany.

Copyright © 2018 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary CAG = cytosine-adenine-guanine; ECAR = extracellular acidiﬁcation rate; OCR = oxygen consumption rate.

Huntington disease, a devastating neurodegenerative disor- case temporary cessation of medication use (for about 7 times der, is caused by an elongated cytosine-adenine-guanine the drug’s half-life) was not possible before sampling. (CAG) repeat sequence in exon 1 of the huntingtin gene – (HTT).1 3 The length of the CAG repeat sequence accounts Standard protocol approvals, registrations, for 50%–70% of the variation in age at onset, leaving a sub- and patient consents stantial amount of unexplained variation, which could be at- The study protocol was approved by the local ethics com- tributed to genetic and environmental modifiers.4,5 In mittee, and written informed consent was obtained from all addition to progressive motor disturbances, neuropsychiatric participants. symptoms, and cognitive decline, patients with Huntington disease and premanifest mutation carriers suffer from un- Sampling and cell culture intended weight loss.6,7 Recently, we demonstrated that this During a regular visit to our outpatient clinic, we obtained weight loss was associated with a faster rate of disease pro- phenotypic data and a small skin sample (i.e., 3 mm diameter) gression independent of CAG repeat number.8 Various evi- from the upper thigh of each participant via a punch biopsy. dence indeed indicates that disturbances in energy Fibroblasts from the skin samples were cultured in minimal metabolism and mitochondrial defects play a role in Hun- essential medium (Gibco #10370-07) supplemented with 15% – tington disease pathology.9 11 Furthermore, mitochondrial heat-inactivated fetal bovine serum (Gibco #10270106), 1% metabolism was found to be impaired in peripheral tissues of penicillin-streptomycin (10,000 U/mL, Gibco #15140122) and patients with Huntington disease, including lymphoblastoid 1% GlutaMAX supplement (Gibco #35050061), and stored in – fi cell lines and skin fibroblasts.12 15 In support of the associa- ahumidi ed incubator at 37°C with 5% CO2. For the experi- fi tion between mitochondrial defects and Huntington disease ments described below, the broblasts were grown up to symptomology, Huntington disease characteristics emerged a maximum of 15 passages and harvested by trypsinization with in humans after accidental exposure to 3-nitropropionic acid trypsin-EDTA (0.05%, Gibco #25300054) at 37°C. (3-NP), a mitochondrial toxin that selectively inhibits the – ATP concentration under oxidative stress activity of mitochondrial complex II.16 19 However, to what To assess whether bioenergetics differences in fibroblasts’ extent differences in energy metabolism are associated with response to oxidative stress between the matched pairs were the onset of Huntington disease symptoms independent of present, we quantified the ATP concentration in the fibro- the CAG repeat number is unknown. Therefore, the aim of blasts after subjecting them to 0.5 mM H O for 0, 5, 10, and our study was to investigate whether differences in energy 2 2 15 minutes. In solid 96-well plates, we plated 12 replicates per metabolism were present in fibroblast cell lines from patients cell line (i.e., 3 replicates per exposure period) with 30,000 with Huntington disease with identical CAG repeat sizes but cells in 100 μL regular culturing medium per well. Per plate, a large difference in age at onset. we included 14 blank wells (i.e., containing no cells) to which different ATP standard concentrations would later be added. Afterward, we incubated the plate for a minimum of 2 hours in Methods fi a humidi ed incubator at 37°C with 5% CO2 to allow the cells Participants to attach. From the Leiden Huntington disease database that contained data on 356 patients with Huntington disease, we selected 9 We applied oxidative stress to the fibroblasts by replacing the pairs of patients with Huntington disease older than 18 years, regular culturing medium with a medium containing 0.5 mM matched for sex and CAG repeat length, but with a large H2O2 and 5% fetal bovine serum for the appointed stress ff fi 20 ff di erence in age at onset, as de ned by an expert neurologist periods. To block the e ect of H2O2 at the end of this (R.A.C.R.) based on motor, psychiatric, and/or cognitive period, we added catalase from bovine liver (50 U/mL me- symptoms. As a cutoff, we used a difference in age at onset of dium, Sigma-Aldrich #C9322-1G)20 and replaced the 0.5 mM μ at least 10 years between each pair of patients. Aside from the H2O2 medium with 250 L per well of the regular culturing age at onset derived from the date of clinical Huntington medium. disease, we also noted the age at onset as estimated by the rater based on patient information and the age at onset of We quantified the ATP concentration of the fibroblasts per well different Huntington disease symptoms (table 1). Exclusion with the Luminescent ATP detection Assay Kit (Abcam criteria were presence of inflammatory diseases, an active in- #ab113849) and added different concentrations of the ATP fectious disease, and the use of anti-inflammatory or immu- standard (i.e., 0, 10, 100, 1,000, 10,000, 100,000, and 1,000,000 nosuppressive drugs (e.g., non-steroidal anti-inflammatory nM) provided by the kit to the blank wells. Luminescence was drugs, and corticosteroids) or antioxidants (e.g., vitamin C) in assessed with Perkin Elmer Multimode Plate Reader, Victor

2 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Table 1 Age at onset of Huntington disease symptoms

Difference in age at symptom CAG Age at Difference in age Age at symptom onset repeat Calendar Disease clinical at clinical onset estimated by estimated by Symptom Couple number Sex age, y duration, y diagnosis, y diagnosis, y rater, ya rater, ya at onset

1 41 F 55.0 8.6 46.4 24.7 46.4 21.7 m

41 F 73.8 2.8 71.1 68.1 m

2 41 M 58.2 4.1 54.1 10.1 50.1 — m, p

41 M 67.4 3.2 64.3 — m

3 42 F 52.9 7.8 45.1 28.2 45.1 26.2 p

42 F 78.9 5.6 73.3 71.3 m

4 42 F 42.3 6.4 35.9 18.0 34.9 14.0 m

42 F 59.8 5.9 53.9 48.9 m, p

5 42 M 43.5 6.6 36.9 23.9 35.9 21.9 p

42 M 61.8 1.0 60.8 57.8 m, p

6 42 M 43.0 20.8 22.3 27.1 20.3 28.1 p

42 M 52.0 2.7 49.3 48.3 m, c

7 45 M 43.5 2.2 41.3 15.2 34.3 12.2 p

45 M 58.1 1.6 56.5 46.5 m

8 46 F 39.3 3.9 35.4 11.6 34.4 9.6 m, p, c

46 F 49.9 2.9 47.0 44.0 c

9 46 M 49.0 16.0 33.0 14.0 33.0 13.0 p, c

46 M 51.9 4.9 47.0 46.0 m

Average, y 54.5 5.9 19.2 18.3

Abbreviations: c = cognitive symptom; CAG = cytosine-adenine-guanine; F = female; M = male; m = motor symptom; p = psychiatric symptom. a A medical doctor estimates the time at which the symptoms of the patient started based on the information provided by the patient and the patient’s family.

X3. We quantified the ATP concentration per well by creating 3-NP, complex II inhibitor), and 1 μM antimycin A (complex III an ATP standard curve with a corresponding equation based on inhibitor). Fibroblasts were plated 1 day in advance of the ex- the luminescence per ATP standard. Using this standard curve periment at 60,000 cells per well in a XF24 cell culture micro- equation, we determined the ATP concentration according to plate. This density resulted in confluent cultures in which cell the average luminescence of the 3 wells per time point per cell growth was blocked because of contact inhibition, avoiding po- line. tential biases because of different growth rates between fibroblast cell lines.21 On the day of the experiment, the regular fibroblast Mitochondrial respiration and glycolysis medium was removed, and the cells were washed twice with XF Using the Seahorse XF Extracellular Flux Analyzer XF24, we assay medium at 37°C, supplemented with 5 mM glucose and could measure mitochondrial respiration and glycolysis simul- 1 mM sodium pyruvate, and the medium was buffered at pH 7.4. taneously and in real time in our fibroblast cell lines. Respiration Subsequently, 675 μL of the XF assay medium was added to each was measured as the oxygen consumption rate (OCR), and well, and the cells were incubated for 60 minutes in a 37°C fi glycolysis was measured as the extracellular acidi cation rate incubator without CO2 to allow the cells to equilibrate to the (ECAR). In addition, the Seahorse allowed for the injection of 4 new medium. toxins during the experimental run and could monitor their effects over time. In succession, we injected the following tox- During the experiment, 4 measurements were taken at base- ins: 1 μM oligomycin (ATP-synthase inhibitor); 1 μMcarbonyl, line for both OCR and ECAR. Afterward, 3 measurements cyanide-4-(trifluoromethoxy) phenylhydrazone (or cabonyl were taken after every toxin injection. From these values, we cyanide-p-trifluoromethox-yphenyl-hydrazon [FCCP], oxidative calculated 6 OCR parameters and 2 ECAR parameters. Basal phosphorylation uncoupler); 1M 3-nitropropionic acid (or respiration was defined as the average OCR values at baseline.

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 3 The average of the 3 OCR values after oligomycin injection Results was defined as the OCR due to proton leak. Maximal respiration was calculated by taking the average OCR of the 3 Differences in age at onset measurements after FCCP injection. The average OCR The age at onset in the 18 selected patients with Huntington value after injection of 3-NP was defined as “respiration disease ranged between 22 and 73 years, and the average after 3-NP injection,” and the average respiration after difference in age at onset between patient pairs was 19.2 years. antimycin A injection was defined as “non-mitochondrial The patient pairs had CAG repeat sequences between 41 and respiration.” From these OCR parameters, another 3 46 repeats, and 5 of the 9 pairs were men. Furthermore, the parameters were calculated: respiration dedicated to ATP symptoms at onset of disease varied per patient between production (defined as basal respiration minus proton motor, psychiatric, and cognitive symptoms (table 1). leak), spare capacity (defined as maximal respiration minus basal respiration), and respiration dependent on complex II ATP concentrations were lower in the skin activity (defined as maximal respiration minus respiration fibroblasts of patients with Huntington after 3-NP injection). Basal glycolysis was defined as the disease with an earlier age at onset fi average ECAR of the 4 baseline ECAR measurements, and In all broblast cell lines, the ATP concentration decreased the increase in glycolysis after blocking ATP synthase was over the time exposed to 0.5 mM H2O2. The decrease was fi fi calculated by subtracting the basal glycolysis from the av- exponential ( gure 1 and gure e-1, links.lww.com/NXG/ erage ECAR after FCCP injection. To ensure that differ- A84). Therefore, we used the natural logarithmic transform ences in the absolute averages were not due to variations in of the ATP concentration (lnATP) as the target variable in fi basal respiration and glycolysis, we also calculated the mi- the analysis. We found that lnATP was signi cantly lower in tochondrial measurements as percentages of basal respira- patients with an earlier age at onset compared with the fi tion and glycolysis. matched patients with a later age at onset ( gure 1 and table 2). This difference was present at baseline in 7 of the 9 couples and continued to be evident throughout the period Statistical analysis the cells were subjected to oxidative stress in 7 couples To account for both the correlation within matched pairs, as (figure e-1). Because the difference in age at onset as esti- well as the correlation due to serial measurements in time mated by the rater was missing in couple number 2, causing during each trial, we applied generalized linear mixed-effects thedataofthiscoupletobelessreliable(table1),weper- models to analyze the results of the bioenergetics experi- formed a sensitivity analysis by excluding these patients. ments. We set the calculated average ATP concentrations Exclusion of these patients from the analysis did not mate- after oxidative stress of every cell line per time point as the rially alter our results (table 2). target variable. Because of an exponential association between the ATP concentrations and time, we used the natural logarithmic transform of the index variable as the target variable. Group (i.e., earlier age at onset vs later age at onset), disease duration, calendar age at the time of biopsy, and Figure 1 Average ATP concentration in the skin fibroblasts of patients with Huntington disease exposed to time of exposure to H O were included as fixed effects. 2 2 oxidative stress over time Furthermore, we included a random intercept for each patient pair, as well as a random slope for time of exposure to H2O2, to adequately account for both the matching between each pair of patients and the correlated measurements on each individual during the experiments. For mitochondrial respiration and glycolysis, we analyzed every calculated functional index separately. For each functional index, we defined the absolute OCR and ECAR or the relative OCR and ECAR as the target variables and constructed models as described above. For all models, we used an unstructured random effect covariance matrix and robust estimations of covariance, which result in consistent parameter estimates even if model assumptions are violated. All models were checked both graphically and analytically. All tests were 2-tailed, and the threshold for statistical significance (i.e., α) was set at 0.05. All analyses were performed in SPSS version In all fibroblast cell lines, the average ATP concentration decreased expo- nentially as the time exposed to oxidative stress increased. At every time 23.0 (IBM SPSS Statistics for Windows, IBM Corp). point, the group of patients with an earlier age at onset had a lower ATP concentration compared with the group of patients with a later age at onset. Data availability p < 0.001*** indicates the significant effect of group (i.e., earlier or later age at onset) on the ATP concentration determined using linear mixed-effects Additional data will be made available at the request of other models. Error bars indicate ±SD. investigators.

4 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Table 2 Association between the logarithmic transform of the average ATP concentration and earlier vs later age at onset in HD

Analysis Fixed effects β-coefficienta SE t p Value 95% CI

All cases Group (earlier) −0.709 0.177 −4.00 <0.001*** −1.063 −0.355

Disease duration 0.016 0.025 0.66 0.511 −0.033 0.066

Calendar age 0.000 0.010 −0.04 0.969 −0.021 0.020

Time −0.102 0.008 −13.01 <0.001*** −0.118 −0.087

Sensitivity analysisb Group (earlier) −0.642 0.255 −2.52 0.015* −1.153 −0.132

Disease duration 0.013 0.027 0.50 0.619 −0.040 0.067

Calendar age 0.000 0.011 0.03 0.978 −0.021 0.021

Time −0.101 0.009 −11.56 <0.001*** −0.119 −0.084

Abbreviation: CI = confidence interval. *p < 0.05, ***p < 0.001. a This column indicates the change in logarithmic transform of the ATP concentration. b The analysis excluding patient pair number 2 (table 1).

Mitochondrial indices were lower in the skin adjustment for disease duration and calendar age (figure 2A-I, fibroblasts of patients with Huntington and figure e-3). The effects did not change after excluding disease with an earlier age at onset couple 2 in the sensitivity analysis (table 4). The average OCR and the ECAR per Huntington disease patient group during the experiment are presented in figure Of interest, disease duration and calendar age were also signifi- e-2, links.lww.com/NXG/A84. Neither the absolute mito- cantly associated with several indices of mitochondrial bio- chondrial parameters nor the parameters expressed as per- energetics. Longer disease duration was accompanied by lower centage of basal respiration differed significantly per group absolute values of basal respiration, maximal respiration, before correction for disease duration and calendar age (figure respiration after 3-NP injection, respiration dependent on e-3). However, after correction, we found the absolute OCR complex II activity, and glycolysis after blocking ATP synthase. averages to be significantly different between the 2 groups in 4 Furthermore, patients with a higher calendar age had signifi- of the 8 estimated parameters (table 3). The average maximal cantly lower levels of maximal respiration, respiration dedicated respiration was lower in the group of patients with Hun- to ATP production, spare capacity, respiration dependent on tington disease with an earlier age at onset. All 9 patients with complex II activity, basal glycolysis, and glycolysis after blocking an earlier age at onset had a lower maximal respiration com- ATP synthase. In addition, the parameters maximal respiration, pared with the matched patients with a later age at onset. spare capacity, and respiration dependent on complex II activity Similarly, in 8 of the 9 couples, the spare capacity was lower in were also significantly lower with a higher calendar age when the earlier age at onset group. Furthermore, the respiration expressed as percentages of the basal respiration (table 3). dedicated to ATP production (in 6 couples) and the respiration dependent on complex II activity (in 8 couples) were Discussion significantly lower in the group of patients with Huntington disease with an earlier age at onset. The average basal respi- We found lower ATP concentrations in the skin fibroblasts of ration was not markedly different between the 2 groups. Av- patients with Huntington disease with an earlier age at onset erage OCR as a percentage of the basal respiration differed in compared with those with a later age at onset, independent of 3 parameters between the groups, including maximal respi- CAG repeat number, sex, calendar age, and disease duration. ration (in 8 couples), spare capacity (in 8 couples), and res- In addition, we demonstrated that the fibroblasts of patients piration dependent on complex II activity (in 9 couples). No with Huntington disease with an earlier age at onset exhibited difference in average respiration dedicated to ATP production lower mitochondrial respiration indices, including lower was found when this variable was defined as a percentage of maximal respiration, spare capacity, and respiration de- basal respiration. The glycolysis parameters did not differ pendent on complex II activity in an absolute sense, as well as between the groups, although the difference in basal glycolysis relative to their basal respiration levels. Furthermore, we showed a trend toward statistical significance in which the found that disease duration and age at biopsy were also sig- group of patients with Huntington disease with an earlier age nificantly associated with several parameters of mitochondrial at onset had a lower average basal ECAR. To illustrate these bioenergetics. Although mitochondrial defects have been ex- effects, we plotted the unadjusted values, as well as the pre- tensively documented before in Huntington disease, to our dicted values, calculated using the model estimates after knowledge, we are the first to demonstrate that differences in

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 5 Table 3 Association between the calculated mitochondrial stress test parameters and age at onset in Huntington disease

Measurement Output Variable β-coefficienta SE t p Value 95% CI

Basal respiration OCR Group (earlier) −4.520 18.249 −0.25 0.808 −43.661 34.621

Disease −1.799 0.650 0.02 0.015 −3.193 −0.405 duration

Calendar age 0.194 0.566 0.74 0.737 −1.020 1.409

Proton leak OCR Group (earlier) −14.298 11.319 −1.26 0.227 −38.575 9.979

Disease 0.143 0.472 0.30 0.766 −0.869 1.156 duration

Calendar age −0.343 0.435 −0.79 0.443 −1.275 0.589

%OCR Group (earlier) −6.784 8.820 −0.77 0.455 −25.701 12.133

Disease 0.807 0.475 1.70 0.112 −0.213 1.826 duration

Calendar age −0.174 0.378 −0.46 0.652 −0.984 0.636

Maximal respiration OCR Group (earlier) −54.176 17.454 −3.10 0.008** −91.610 −16.741

Disease −3.317 0.822 −4.04 0.001** −5.080 −1.554 duration

Calendar age −2.118 0.818 −2.59 0.021* −3.872 −0.365

%OCR Group (earlier) −46.730 20.325 −2.30 0.037* −90.321 −3.138

Disease −0.430 0.930 −0.46 0.651 −2.426 1.565 duration

Calendar age −2.577 0.949 −2.72 0.017* −4.613 −0.542

Respiration after 3-NP injection OCR Group (earlier) −17.226 9.145 −1.88 0.081 −36.839 2.387

Disease −1.277 0.421 −3.03 0.009** −2.180 −0.374 duration

Calendar age −0.529 0.259 −2.05 0.060 −1.084 0.025

%OCR Group (earlier) −5.133 8.749 −0.59 0.567 −23.898 13.632

Disease −0.527 0.459 −1.15 0.269 −1.511 0.456 duration

Calendar age −0.453 0.565 −0.80 0.436 −1.664 0.758

Nonmitochondrial respiration OCR Group (earlier) −12.355 9.515 −1.30 0.215 −32.763 8.053

Disease 0.699 0.434 1.61 0.129 −0.231 1.629 duration

Calendar age −0.139 0.386 −0.36 0.723 −0.968 0.689

%OCR Group (earlier) −0.607 8.521 −0.07 0.944 −18.881 17.668

Disease 0.933 0.503 1.85 0.085 −0.146 2.013 duration

Calendar age 0.292 0.344 0.85 0.411 −0.446 1.030

Respiration dedicated to ATP production OCR Group (earlier) −38.106 11.243 −3.39 0.004** −62.219 −13.992

Disease 0.936 0.655 0.1.43 0.175 −0.468 2.341 duration

Calendar age −1.818 0.450 −4.04 0.001** −2.783 −0.852

%OCR Group (earlier) 6.784 8.820 0.46 0.455 −12.133 25.701

Disease −0.807 0.475 −1.70 0.112 −1.826 0.213 duration

Calendar age 0.174 0.378 0.46 0.652 −0.636 0.984

Continued 6 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Table 3 Association between the calculated mitochondrial stress test parameters and age at onset in Huntington disease (continued)

Measurement Output Variable β-coefficienta SE t p Value 95% CI

Spare capacity OCR Group (earlier) −50.884 9.894 −5.14 <0.001*** −72.104 −29.663

Disease −1.482 0.700 −2.12 0.053 −2.984 0.019 duration

Calendar age −2.387 0.685 −3.48 0.004** −3.856 −0.918

%OCR Group (earlier) −46.730 20.325 −2.30 0.037* −90.321 −3.138

Disease −0.430 0.930 −0.46 0.651 −2.426 1.565 duration

Calendar age −2.577 0.949 −2.72 0.017* −4.613 −0.542

Respiration dependent on complex II OCR Group (earlier) −36.950 13.397 −2.76 0.015* −65.685 −8.216 activity

Disease −2.039 0.598 −3.41 0.004** −3.323 −0.756 duration

Calendar age −1.589 0.597 −2.66 0.019* −2.868 −0.309

%OCR Group (earlier) −31.122 12.292 −2.53 0.024* −57.486 −4.759

Disease −0.737 0.415 −1.78 0.097 −1.627 0.153 duration

Calendar age −1.683 0.520 −3.24 0.006** −2.797 −0.569

Basal glycolysis ECAR Group (earlier) −5.404 2.763 −1.96 0.071 −11.331 0.523

Disease −0.330 0.193 −1.71 0.110 −0.745 0.085 duration

Calendar age −0.447 0.129 −3.46 0.004** −0.724 −0.170

Glycolysis after blocking ATP-synthase ECAR Group (earlier) 0.614 3.302 0.19 0.855 −6.468 7.697

Disease −0.643 0.153 −4.20 0.001** −0.971 −0.314 duration

Calendar age −0.187 0.079 −2.38 0.032* −0.356 −0.019

%ECAR Group (earlier) 48.604 25.707 1.891 0.080 −6.532 103.739

Disease −0.010 2.250 −0.004 0.997 −4.835 4.816 duration

Calendar age 4.988 2.202 2.266 0.040* 0.267 9.710

Abbreviations: CI = confidence interval; ECAR = extracellular acidification rate; %ECAR = ECAR as a percentage of basal ECAR; OCR = oxygen consumption rate; %OCR = OCR as a percentage of basal OCR. *p < 0.05, **p < 0.01, ***p < 0.001. a This column indicates the change in OCR, ECAR, %OCR, or %ECAR.

bioenergetics were associated with age at onset among mitochondrial respiration. Maximal respiration and mito- patients with Huntington disease, and thereby, could be an chondrial spare capacity are measures of the ability of mito- important target for future therapeutic interventions. chondria to react to increased energy demands and are critical – for neuronal survival.22 24 These measures were lower in The fact that the ATP concentration was lower in patients with patients with an earlier age at onset, suggesting that neurons Huntington disease with an earlier age at onset suggests that in these patients could be more vulnerable to damage and the production of ATP in the cells of these patients may be death, thus causing Huntington disease symptoms to start at impaired to a greater extent compared with their counterparts a younger age. Of interest, a previous study showed a correla- or that the cells of patients with a later age at onset carry tion between decreased levels of spare capacity and increased a mechanism that protects their ATP metabolism. The pres- reactive oxygen species and mitochondrial DNA lesions in ence of a potential problem with ATP production was sup- mutant Huntington disease striatal immortalized neuronal ported by comparable diﬀerences in various indices of cells.15 Respiration dependent on complex II activity was also

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 7 Figure 2 Adjusted estimates of the functional indices calculated from the mitochondrial stress test per age group at symptom onset in Huntington disease

There was no difference in basal respiration between the 2 groups (A). The absolute mitochondrial respiration dedicated to ATP production was significantly lower in the group of patients with Huntington disease with an earlier age at onset (B). Both the absolute maximal respiration and the maximal respiration as a percentage of the basal respiration were significantly lower in the group of patients with Huntington disease with an earlier age at onset (C and D). The absolute spare capacity and the spare capacity as a percentage of the basal respiration were significantly lower in the group of patients with Huntington disease with an earlier age at onset (E and F). The absolute respiration dependent on complex II activity and the respiration dependent on complex II activity as a percentage of the basal respiration were also significantly lower in the group of patients with Huntington disease with an earlier age at onset (G and H). The basal glycolysis did not differ significantly between the 2 groups (I). Adjusted estimates = values adjusted for disease duration and calendar age at the time of biopsy. Error bars indicate ±SD. *p-value < 0.05. **p-value < 0.01. ***p-value < 0.001. ECAR = extracellular acidification rate; OCR = oxygen consumption rate.

8 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Table 4 Sensitivity analysis of the association between the mitochondrial stress test parameters and age at onset in Huntington disease

Measurementa Output Variable β-coefficientb SE t p Value 95% CI

Maximal respiration OCR Group (earlier) −56.810 24.570 −2.31 0.039* −110.343 −3.277

Disease −2.938 1.079 −2.72 0.018* −5.288 −0.588 duration

Calendar age −2.815 0.568 −4.96 <0.001*** −4.053 −1.577

%OCR Group (earlier) −56.330 24.401 −2.31 0.040* −109.495 −3.164

Disease 0.018 0.965 0.02 0.985 −2.084 2.120 duration

Calendar age −2.903 1.061 −2.74 0.018* −5.215 −0.591

Respiration dedicated to ATP production OCR Group (earlier) −31.600 9.310 −3.39 0.005** −51.884 −11.315

Disease 0.619 0.506 1.22 0.245 −0.483 1.721 duration

Calendar age −1.684 0.396 −4.25 0.001** −2.547 −0.821

%OCR Group (earlier) 15.130 7.065 2.14 0.053 −0.263 30.523

Disease −1.186 0.328 −3.61 0.004** −1.901 −0.471 duration

Calendar age 0.482 0.312 1.55 0.148 −0.197 1.162

Spare capacity OCR Group (earlier) −56.929 11.504 −4.95 <0.001*** −81.995 −31.863

Disease −1.066 0.619 −1.72 0.111 −2.414 0.283 duration

Calendar age −2.825 0.556 −5.08 <0.001*** −4.037 −1.613

%OCR Group (earlier) −56.330 24.401 −2.31 0.040* −109.495 −3.164

Disease 0.018 0.965 0.02 0.985 −2.084 2.120 duration

Calendar age −2.903 1.061 −2.74 0.018* −5.215 −0.591

Respiration dependent on complex II OCR Group (earlier) −39.637 17.613 −2.25 0.044* −78.012 −1.261 activity

Disease −1.773 0.790 −2.24 0.045* −3.495 −0.051 duration

Calendar age −2.015 0.483 −4.17 0.001** −3.067 −0.962

%OCR Group (earlier) −35.371 14.785 −2.39 0.034* −67.584 −3.157

Disease −0.522 0.465 −1.12 0.284 −1.535 0.492 duration

Calendar age −1.866 0.559 −3.34 0.006** −3.083 −0.649

Abbreviations: CI = confidence interval; ECAR = extracellular acidification rate; %ECAR = ECAR as a percentage of basal ECAR; OCR = oxygen consumption rate; %OCR = OCR as a percentage of basal OCR. *p < 0.05, **p < 0.01, ***p < 0.001. a Analyses performed with exclusion of couple number 2. b This column indicates the change in OCR, ECAR, %OCR, or %ECAR.

lower in patients with Huntington disease with an earlier age at causes the loss of complex II is unknown. Possible hypotheses onset. Diﬀerent studies showed that in postmortem samples of include the direct association of mutant HTT with the mito- the striatum and cerebral cortex of patients with Huntington chondrial membrane, causing decreased import of subunits disease, complex II of the mitochondrial electron transport into the mitochondria, increased degradation, or abnormal chain displayed reduced activity, which was associated with assembly.13,29,30 Furthermore, mutant HTT has been shown – diminished expression of 2 complex II subunits.25 29 The exact to increase cellular oxidative stress, which was associated with mechanism as to how the mutated huntingtin protein (HTT) impaired activity of complex II in a yeast model of Huntington

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 9 – disease.31 33 Our results here suggest that the reduced ex- reliable. Last, we performed several statistical analyses without pression of complex II is more pronounced in patients with an applying a correction for multiple testing, given our relatively earlier age at onset independent of CAG repeat length, which small sample size, as well as the fact that bioenergetics could perhaps be due to a more potent association of mutant assessments are highly correlated and, thus, cannot be HTT with the mitochondrial membrane or a higher vulner- regarded as independent. Nevertheless, to obtain results with ability to oxidative stress. a higher level of confidence, future research investigating similar parameters should include a larger number of patients. Candidate studies found that a single nucleotide polymorphism in PPARGC1A, encoding the mitochondrial regulator peroxi- We demonstrated an association between age at onset and the some proliferator-activated receptor gamma coactivator-1α bioenergetics profile in patients with Huntington disease. Thus, (PGC-1α), affected the age at onset in 3 European Huntington differences in bioenergetics could explain part of the residual disease cohorts, suggesting that our findings might have a ge- variation in age at onset among patients with Huntington netic origin.34,35 Moreover, in striata from patients with Hun- disease, whereas therapies aimed at enhancing mitochondrial tington disease and different rodent Huntington disease models, function may delay disease onset. However, further research the expression of PGC-1α mRNA was reduced, and upregu- into the mechanisms mediating the association between bio- lating this expression caused prevention of striatal neuronal at- energetics and age at onset are needed to develop novel ther- rophy, improvement of motor deficits, and protection against apies aimed at delaying symptom onset and disease progression mitochondrial dysfunction and cell death, implying that in Huntington disease. PGC-1α expression may modify mutant HTT–induced mito- – chondrial toxicity.36 38 Together with our results, these findings Author contributions indicate that differences in PGC-1α expression and the conse- S.L. Gardiner and N.A. Aziz contributed to the conception quential variations in bioenergetics profile may result in addi- and design of the study and to the acquisition and analysis of tionalvariationinageatonsetinHuntingtondisease. data. C. Milanese and W.M.C. van Roon-Mom offered crucial supervision during the experiments. All authors contributed We found that as patients suffered from Huntington disease for to drafting the text and preparing the figures. a longer period of time, several indices of mitochondrial respiration were significantly lower, independent of calendar age. Acknowledgment This finding is in line with the fact that Huntington disease The authors cordially thank all patients and caregivers involved pathogenesis is known to involve mitochondrial deficits.39,40 in the experiments for their valuable time and efforts. The Therefore, a reasonable derivative is that these mitochondrial authors also thank Ms. L.J. Schipper, M.D., for assistance in deficits become worse as the disease progresses. Mitochondrial taking skin biopsies. function is also known to decrease with age.41 In accordance, we found that as the calendar age of patients increased, several Study funding parameters of mitochondrial function decreased. The aging NAA is supported by a VENI-grant (#91615080) from the process itself is unlikely to have accounted for our main finding Netherlands Organization for Scientific Research and a Marie (i.e., a better bioenergetics profile in late-onset patients), given Sklodowska-Curie Individual Fellowship grant from the Eu- that patients with a later age at onset were older at the time of ropean Union (Horizon 2020, #701130). biopsy (table 1). Disclosure Our study had several limitations. First, although Huntington S. Gardiner, C. Milanese, M.W. Boogaard, R.A.M. Buijsen, and disease is primarily a neurologic disease, we were able to find M. Hogenboom report no disclosures. R.A.C. Roos is a con- differences between patients with Huntington disease with an sultant for uniQure. P.G. Mastroberardino serves on the edi- early onset and late onset of symptoms in skin fibroblasts. The torial boards of Neurobiology of Disease, Frontiers in Cellular fact that we could acquire relevant results in these relatively Neuroscience,andCell Death and Disease; is a consultant for the easy to obtain cells is intriguing and supports the use of fi- Neurological Institute; and has received research support from broblast cell models in the research of neurodegenerative Dorpmans-Wigmans Stichting. W.M.C. van Roon-Mom has diseases. However, these cells are not the primary cells of served on the medical advisory board of the ADCA Patients’ interest. Therefore, repeating similar experiments in a neuro- Association; is the main inventor on 2 published patent nal cell model, such as striatal neurons derived from induced applications (WO2012/018257 and WO2015/053624) re- pluripotent stem cells, is warranted. Second, establishing the garding exon skipping approaches for neurodegenerative dis- age at onset in Huntington disease is subjective.42 In order to eases; and is coinventor on 1 published patent application on achieve additional certainty concerning this age, we noted the exon skipping approaches for SCA3 (WO 2017/053781). N.A. age at Huntington disease diagnosis, as well as the age esti- Aziz has received research support from the Netherlands Or- mated by the rater, which were analogous in 8 of the 9 cou- ganization for Scientific Research (NWO), the European ples. In addition, our results did not change after excluding the Union, and the European Huntington’s Disease Network. Full couple in which the rater-estimated age at onset was missing disclosure form information provided by the authors is available for 1 patient, illustrating that our results are robust and with the full text of this article at Neurology.org/NG.

10 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Received May 4, 2018. Accepted in final form August 8, 2018. 22. Kumar MJ, Nicholls DG, Andersen JK. Oxidative alpha-ketoglutarate dehydrogenase inhibition via subtle elevations in monoamine oxidase B levels results in loss of spare respiratory capacity: implications for Parkinson’s disease. J Biol Chem 2003;278: References 46432–46439. ’ 1. Bruyn G. Huntington s chorea: historical, clinical and laboratory synopsis. In: Vinken P, 23. Vesce S, Jekabsons MB, Johnson-Cadwell LI, Nicholls DG. Acute glutathione de- – Bruyn G, editors. Handbook of Clinical Neurology. Amsterdam: Elsevier; 1968:298 378. pletion restricts mitochondrial ATP export in cerebellar granule neurons. J Biol Chem 2. A novel gene containing a trinucleotide repeat that is expanded and unstable on 2005;280:38720–38728. ’ ’ Huntington s disease chromosomes. The Huntington s Disease Collaborative Re- 24. Yadava N, Nicholls DG. Spare respiratory capacity rather than oxidative stress reg- – search Group. Cell 1993;72:971 983. ulates glutamate excitotoxicity after partial respiratory inhibition of mitochondrial 3. Langbehn DR, Hayden MR, Paulsen JS. CAG-repeat length and the age of onset in complex I with rotenone. J Neurosci 2007;27:7310–7317. Huntington disease (HD): a review and validation study of statistical approaches. Am – 25. Gu M, Gash MT, Mann VM, Javoy-Agid F, Cooper JM, Schapira AH. Mitochondrial J Med Genet B Neuropsychiatr Genet 2010;153B:397 408. defect in Huntington’s disease caudate nucleus. Ann Neurol 1996;39:385–389. 4. Wexler NS, Lorimer J, Porter J, et al. Venezuelan kindreds reveal that genetic and 26. Brennan WA Jr, Bird ED, Aprille JR. Regional mitochondrial respiratory activity in environmental factors modulate Huntington’s disease age of onset. Proc Natl Acad Sci Huntington’s disease brain. J Neurochem 1985;44:1948–1950. USA 2004;101:3498–3503. 27. Browne SE, Bowling AC, MacGarvey U, et al. Oxidative damage and metabolic 5. Gusella JF, MacDonald ME, Lee JM. Genetic modifiers of Huntington’s disease. Mov dysfunction in Huntington’s disease: selective vulnerability of the basal ganglia. Ann Disord 2014;29:1359–1365. Neurol 1997;41:646–653. 6. Aziz NA, van der Burg JM, Landwehrmeyer GB, Brundin P, Stijnen T, Roos RA. 28. Tabrizi SJ, Cleeter MW, Xuereb J, Taanman JW, Cooper JM, Schapira AH. Bio- Weight loss in Huntington disease increases with higher CAG repeat number. chemical abnormalities and excitotoxicity in Huntington’s disease brain. Ann Neurol Neurology 2008;71:1506–1513. 1999;45:25–32. 7. Mochel F, Charles P, Seguin F, et al. Early energy deficit in Huntington disease: 29. Benchoua A, Trioulier Y, Zala D, et al. Involvement of mitochondrial complex II identification of a plasma biomarker traceable during disease progression. PLoS One defects in neuronal death produced by N-terminus fragment of mutated huntingtin. 2007;2:e647. Mol Biol Cell 2006;17:1652–1663. 8. van der Burg JMM, Gardiner SL, Ludolph AC, Landwehrmeyer GB, Roos RAC, Aziz 30. Choo YS, Johnson GV, MacDonald M, Detloff PJ, Lesort M. Mutant huntingtin NA. Body weight is a robust predictor of clinical progression in Huntington disease. – directly increases susceptibility of mitochondria to the calcium-induced permeability Ann Neurol 2017;82:479 483. – 9. Feigin A, Leenders KL, Moeller JR, et al. Metabolic network abnormalities in early transition and cytochrome c release. Hum Mol Genet 2004;13:1407 1420. ’ – 31. Browne SE, Ferrante RJ, Beal MF. Oxidative stress in Huntington’s disease. Brain Huntington s disease: an [(18)F]FDG PET study. J Nucl Med 2001;42:1591 1595. – 10. Ciarmiello A, Cannella M, Lastoria S, et al. Brain white-matter volume loss and Pathol 1999;9:147 163. glucose hypometabolism precede the clinical symptoms of Huntington’s disease. 32. Bogdanov MB, Andreassen OA, Dedeoglu A, Ferrante RJ, Beal MF. Increased oxi- ’ J Nucl Med 2006;47:215–222. dative damage to DNA in a transgenic mouse model of Huntington s disease. – 11. Koroshetz WJ, Jenkins BG, Rosen BR, Beal MF. Energy metabolism defects in J Neurochem 2001;79:1246 1249. Huntington’s disease and effects of coenzyme Q10. Ann Neurol 1997;41:160–165. 33. Solans A, Zambrano A, Rodriguez M, Barrientos A. Cytotoxicity of a mutant hun- 12. Sawa A, Wiegand GW, Cooper J, et al. Increased apoptosis of Huntington disease tingtin fragment in yeast involves early alterations in mitochondrial OXPHOS com- – lymphoblasts associated with repeat length-dependent mitochondrial depolarization. plexes II and III. Hum Mol Genet 2006;15:3063 3081. Nat Med 1999;5:1194–1198. 34. Che HV, Metzger S, Portal E, Deyle C, Riess O, Nguyen HP. Localization of sequence fl ff 13. Panov AV, Gutekunst CA, Leavitt BR, et al. Early mitochondrial calcium defects in Hun- variations in PGC-1alpha in uence their modifying e ect in Huntington disease. Mol tington’sdiseaseareadirecteffect of polyglutamines. Nat Neurosci 2002;5:731–736. Neurodegener 2011;6:1. fi 14. Seong IS, Ivanova E, Lee JM, et al. HD CAG repeat implicates a dominant property of 35. Taherzadeh-Fard E, Saft C, Andrich J, Wieczorek S, Arning L. PGC-1alpha as modi er huntingtin in mitochondrial energy metabolism. Hum Mol Genet 2005;14:2871–2880. of onset age in Huntington disease. Mol Neurodegener 2009;4:10. 15. Siddiqui A, Rivera-Sanchez S, Castro Mdel R, et al. Mitochondrial DNA damage is 36. Cui L, Jeong H, Borovecki F, Parkhurst CN, Tanese N, Krainc D. Transcriptional associated with reduced mitochondrial bioenergetics in Huntington’s disease. Free repression of PGC-1alpha by mutant huntingtin leads to mitochondrial dysfunction – Radic Biol Med 2012;53:1478–1488. and neurodegeneration. Cell 2006;127:59 69. 16. He F, Zhang S, Qian F, Zhang C. Delayed dystonia with striatal CT lucencies induced 37. Chaturvedi RK, Calingasan NY, Yang L, Hennessey T, Johri A, Beal MF. Impairment by a mycotoxin (3-nitropropionic acid). Neurology 1995;45:2178–2183. of PGC-1alpha expression, neuropathology and hepatic steatosis in a transgenic 17. Ludolph AC, He F, Spencer PS, Hammerstad J, Sabri M. 3-Nitropropionic acid- mouse model of Huntington’s disease following chronic energy deprivation. Hum exogenous animal neurotoxin and possible human striatal toxin. Can J Neurol Sci Mol Genet 2010;19:3190–3205. 1991;18:492–498. 38. Hathorn T, Snyder-Keller A, Messer A. Nicotinamide improves motor deficits and 18. Brouillet E, Jenkins BG, Hyman BT, et al. Age-dependent vulnerability of the striatum upregulates PGC-1alpha and BDNF gene expression in a mouse model of Hun- to the mitochondrial toxin 3-nitropropionic acid. J Neurochem 1993;60:356–359. tington’s disease. Neurobiol Dis 2011;41:43–50. 19. Brouillet E, Conde F, Beal MF, Hantraye P. Replicating Huntington’s disease phe- 39. Bossy-Wetzel E, Petrilli A, Knott AB. Mutant huntingtin and mitochondrial dys- notype in experimental animals. Prog Neurobiol 1999;59:427–468. function. Trends Neurosci 2008;31:609–616. 20. Mocali A, Caldini R, Chevanne M, Paoletti F. Induction, effects, and quantification of 40. Chen CM. Mitochondrial dysfunction, metabolic deficits, and increased oxidative sublethal oxidative stress by hydrogen peroxide on cultured human fibroblasts. Exp stress in Huntington’s disease. Chang Gung Med J 2011;34:135–152. Cel Res 1995;216:388–395. 41. Bratic A, Larsson NG. The role of mitochondria in aging. J Clin Invest 2013;123: 21. Ambrosi G, Ghezzi C, Sepe S, et al. Bioenergetic and proteolytic defects in fibroblasts 951–957. from patients with sporadic Parkinson’s disease. Biochim Biophys Acta 2014;1842: 42. Orth M, Schwenke C. Age-at-onset in Huntington disease. PLoS Curr 2011;3: 1385–1394. RRN1258.

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 11 ARTICLE OPEN ACCESS Early-onset Parkinson disease caused by a mutation in CHCHD2 and mitochondrial dysfunction

Richard G. Lee, BSc (Hons),* Maryam Sedghi, MSc,* Mehri Salari, MD, Anne-Marie J. Shearwood, MSc, Correspondence Maike Stentenbach, BSc, Ariana Kariminejad, MD, Hayley Goullee, BSc (Hons), Oliver Rackham, PhD, Dr. Tajsharghi [email protected] Nigel G. Laing, PhD, Homa Tajsharghi, PhD,§ and Aleksandra Filipovska, PhD§ or Dr. Filipovska [email protected] Neurol Genet 2018;4:e276. doi:10.1212/NXG.0000000000000276 Abstract Objective Our goal was to identify the gene(s) associated with an early-onset form of Parkinson disease (PD) and the molecular defects associated with this mutation.

Methods We combined whole-exome sequencing and functional genomics to identify the genes associated with early-onset PD. We used ﬂuorescence microscopy, cell, and mitochondrial biology measurements to identify the molecular defects resulting from the identiﬁed mutation.

Results Here, we report an association of a homozygous variant in CHCHD2, encoding coiled-coil- helix-coiled-coil-helix domain containing protein 2, a mitochondrial protein of unknown function, with an early-onset form of PD in a 26-year-old Caucasian woman. The CHCHD2 mutation in PD patient ﬁbroblasts causes fragmentation of the mitochondrial reticular morphology and results in reduced oxidative phosphorylation at complex I and complex IV. Although patient cells could maintain a proton motive force, reactive oxygen species production was increased, which correlated with an increased metabolic rate.

Conclusions Our ﬁndings implicate CHCHD2 in the pathogenesis of recessive early-onset PD, expanding the repertoire of mitochondrial proteins that play a direct role in this disease.

*These authors contributed equally to the study and share first authorship.

§These authors share senior authorship and are co-corresponding authors.

From the Centre for Medical Research (R.G.L., A.-M.J.S., M. Stentenbach, H.G., O.R., N.G.L., H.T., A.F.), University of Western Australia and the Harry Perkins Institute for Medical Research, Nedlands, Western Australia, Australia; Department of Genetics (M. Sedghi), University of Isfahan, Isfahan; Functional Neurosurgery Research Center (M. Salari), Shohada Tajrish Neurosurgical Center of Excellence, Shahid Beheshti University of Medical Sciences, Tehran, Iran; Kariminejad-Najmabadi Pathology and Genetics Center (A.K.), Tehran, Iran; School of Molecular Sciences (O.R., A.F.), The University of Western Australia, Crawley; Department of Diagnostic Genomics (N.G.L.), PathWest, QEII Medical Centre, Nedlands, Western Australia, Australia; and Division Biomedicine and Public Health (H.T.), School of Health and Education, University of Skovde, Sweden.

The Article Processing Charge was funded by the Australian Research Council. 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 © 2018 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary BCA = Bicinchoninic acid; DHE = dihydroethidium; ETC = electron transport chain; MICOS = mitochondrial contact site and cristae organizing system; NGS = next-generation sequencing; OXPHOS = oxidative phosphorylation; PBS = phosphate- buﬀered saline; PD = Parkinson disease; RCR = respiratory control ratio; ROS = reactive oxygen species; WES = whole-exome sequencing.

Parkinson disease (PD, MIM168600) is the most common Cell culture and transfections movement disorder of aging and the second-most com- Dermal ﬁbroblast cultures from the patient were established mon neurodegenerative disease, after Alzheimer disease by standard protocols, after written informed consent was (MIM104300).1,2 PD has been associated with mutations in obtained. Cultured ﬁbroblasts from a healthy age-matched multiple genes including PARK2, PINK1, PARK7, ATP13A2, individual were used as a control. Detailed methods are pro- – SCNA, LRRK2, VPS35, EIF4G1, DNAJC13, and CHCHD2.3 6 vided in supplementary information.

Recent whole-exome sequencing (WES) of Japanese patients Mitochondrial isolation and cell lysis with autosomal dominant or sporadic PD identified a hetero- Mitochondria were isolated from fibroblasts as previously zygous CHCHD2 missense change (p.Thr61Ile) in a family described.8 Cell lysates were prepared using a buffer con- with autosomal dominant late-onset PD.5 Subsequently, the taining the following: 150 mM NaCl, 0.1% (vol/vol) Triton identical variant was identified in a Chinese family with auto- X-100, and 50 mM Tris-HCl (pH 8.0). Protein concentra- somal dominant PD.6 Additional CHCHD2 variants, including tion was determined using a bicinchoninic acid (BCA) p.Ala32Thr, p.Pro34Leu, and p.Ile80Val, have been described assay. in 4 western European familial patients with PD.7 The patho- mechanism of the CHCHD2 variants in these studies is, how- Fluorescence microscopy ever, unclear as is the physiologic role of CHCHD2. Detailed methods are described in the supplementary material, links.lww.com/NXG/A85. Here, we identify a new CHCHD2 variant in a patient with autosomal recessive early-onset PD and establish the patho- Long-range PCR and mitochondrial DNA copy genic mechanism of this variant. We have shown that the number quantitative PCR mutation results in a fragmented mitochondrial morphology Detailed methods are described in the supplementary mate- and reduced electron transport chain (ETC) activity. rial. The sequences of all primers used for this study are detailed in table e-1, links.lww.com/NXG/A85. Methods SDS-PAGE and immunoblotting Mitochondria (25 μg) isolated from control and patient Standard protocol approvals, registrations, fibroblasts were separated on 4%–12% Bis-Tris gels (Invi- and patient consents trogen) and transferred onto a polyvinylidene difluoride The study was approved by the ethical standards of the rel- (PVDF) membrane (Bio-Rad). All antibodies used are evant institutional review board, the Ethics Review Com- detailed in the supplementary information. mittee in the Gothenburg Region (Dn1: 842-14), and the Human Research Ethics Committee of the University of Mitochondrial protein synthesis Western Australia. Informed consent was obtained from De novo mitochondrial protein synthesis was analyzed in patients included in this study after appropriate genetic fibroblasts using 35S-radiolabeling of mitochondrially enco- counseling. Blood samples were obtained from patients, their ded proteins in the presence of emetine as previously de- parents, and siblings. scribed.9 The cells were suspended in phosphate-buffered saline (PBS) and 20 μg of protein was resolved on a 12.5% Clinical evaluation SDS-PAGE gel, and radiolabelled proteins were visualized Medical history was obtained, and physical examination was on film. performed as part of routine clinical workup. Respiration Genetic analysis Respiration was measured in fibroblasts as previously de- Next-generation sequencing (NGS) whole-exome and/or tar- scribed.10 The full methods are detailed in supplementary geted neuromuscular panel sequencing (WES or neuromus- information. cular sub-exomic sequencing [NSES]) was performed on the patient’sDNA.Confirmatory bidirectional Sanger sequencing Cell function measurements was performed in patients with PD and their family members JC-1, mitochondrial mass, dihydroethidium (DHE), and MTS (see e-Methods, links.lww.com/NXG/A85). assay full methods are detailed in supplementary information.

2 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Statistical analysis Mutation Database and ClinVar, and the most recent literature All data are reported as mean ± SEM. Statistical differences allowed the identification of homozygous CHCHD2 (chro- were determined using a two-tailed Student t test. JC-1, DHE, mosome 7) and TOP1MT (chromosome 8) variants. A novel mitochondrial mass, and MTS data are expressed as a percent homozygous missense mutation in exon 2 of CHCHD2 of average control fibroblast values for the respective media (c.211G>C) (p.Ala71Pro) was identified. The size of the ho- treatment. mozygous region covering CHCHD2 variant on chromosome 7 was 158Mb. In addition, a homozygous missense variant of Data availability TOP1MT (c.661G>A, ENST00000523676, transcript ID: Study data for the primary analyses presented in this report NM_001258447.1), changing aspartic acid to asparagine are available upon reasonable request from the corresponding (p.Asp221Asn), was identified. No rare, likely pathogenic and senior author. heterozygous or homozygous variants in the PD-associated genes, including ATP13A2, CHCHD10, DNAJC13, EIF4G1, LRRK2, PARK2, PARK7, PARK15, PINK1, SNCA,andVPS35, Results were identified in the exome sequencing data. The appearance of CHCHD2 and TOP1MT variants was examined in available Clinical characteristics of the patient family members by sequencing analysis. Sanger sequencing A 30-year-old right-handed Caucasian woman (V:2) was born confirmed segregation of both variants compatible with a re- to healthy consanguineous parents aged 55 and 59 years cessive pattern of inheritance. The unaffected parents (IV:1, IV: fi fi ( gure 1A). There was no signi cant family history, and she 2) were both heterozygous for the CHCHD2 variant, and the denied a history of other diseases or past drug use. The clinical healthy sister (V:1) was not a carrier (figure 1B). Both parents presentations were consistent with PD. Symptoms developed and the sister were heterozygous carriers of the TOP1MT rapidly at age 26 years including resting tremor mostly af- variant (figure 1C). The c.211G>C, (p.Ala71Pro) (CHCHD2) fecting her left arm and bradykinesia. Neurologic examination variant was excluded in East and South Asian, European, Af- at age 28 years showed general bradykinesia, rigidity, resting rican, Latino, and Ashkenazi Jewish Allele Frequency Com- tremor in both hands and legs, but most prominent in the left munities (AFC), in the ExAC database, the Genome side, with superimposed action-induced myoclonus. She also Aggregation Database (gnomAD), and the 1000 Genome showed hypomimic face and mydriatic pupils with sluggish database. The TOP1MT variant (c.661G>A, p.Asp221Asn, reaction to the light, indicating autonomic dysfunction. There rs143769145) was a rare heterozygous variant reported in were no pyramidal or cerebellar signs. On fundoscopic ex- the AFC (frequency 0.025%), the ExAC (frequency amination, Kayser-Fleischer-like rings were negative. Neither 0.026%), and in the gnomAD (frequency 0.023%) data- sphincter dysfunction nor orthostatic hypotension was pres- bases; the variant was identified in the homozygous state. In ent. Laboratory evaluation including thyroid function, liver silico combined annotation dependent depletion (CADD) function, serum and urine copper, and ceruloplasmin was analysis of the missense variants revealed high deleterious unremarkable. Brain MRI was also unremarkable. The patient scores (CHCHD2 [p.Ala71Pro]: 32.000 and TOP1MT showed several clinical features of mitochondrial dysfunction, [p.Asp221Asn]: 26.000). In silico analysis predicted both such as neuropsychological disturbance, dysautonomia, and CHCHD2 and TOP1MT substitutions to be potentially myoclonus. Low-dose levodopa was started with dramatic disease causing (MutationTaster, mutationtaster.org/). The response, but after 1 month, she had severe dyskinesia, and 2 amino acid residues affected are highly conserved across fl after 3 months, motor uctuation, anxiety, and insomnia be- species (figure 1D). The CHCHD2-substituted residue is came the main issues and she had dopamine dysregulation located within the central conserved hydrophobic domain, syndrome. After 6 months, psychiatric problems, including the transmembrane element of the CHCHD2 protein aggression, depression, and impulsiveness, were additional (figure 1E). In addition, the entire coding sequence of symptoms. At follow-up at age 30 years, PD symptoms had CHCHD2 was analyzed in 2 affected individuals of 2 large not progressed and became stable. However, she showed families with familial early-onset PD of Iranian origin. The signs of cognitive decline, and orthostatic hypotention added CHCHD2 p.Ala71Pro substitution was not identified, and to her symptoms. we did not detect any other variant in this gene.

Genetic findings Clinical status of confirmed carriers Data from NGS of DNA from 1 affected (V:2) and 1 unaffected Cosegregation studies confirmed that the parents were car- family member (V:1) were analyzed. No likely pathogenic riers of CHCHD2 and TOP1MT variants, and the unaffected mutations were identified in the genes included in the targeted sister was heterozygous for the TOP1MT variant. Given that panel of the 336 neuromuscular disease genes. WES was per- the previously reported CHCHD2 variants are mostly asso- formed on the patient and her unaffected sister. No rare, likely ciated with late-onset dominant PD, carrier parents were ex- pathogenic heterozygous variants in the PD-associated genes amined by a neurologist. They showed no evidence of were identified. The filtering strategy of initially concentrating neurologic or movement disorder by history. However, the on homozygous coding variants in known neurogenetic disease parents are still younger than the average age of PD onset genes, selected based on variant databases Human Genome (<60).11

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 3 Figure 1 Pedigree and genetic findings

(A) Pedigree of the family. The affected individual (V:2) is represented with a shaded symbol. (B) Sanger sequence analysis demonstrates the presenceof homozygous variants in CHCHD2 and (C) TOP1MT in the patient and the segregation of the variants. (D) Multiple sequence alignment confirms that the p.Ala71Pro (CHCHD2) and p.Asp221Asn (TOP1MT) substitutions affect evolutionarily conserved residues (shaded). (E) Mutated residues and CHCHD2 protein structure: previously identified heterozygous missense variants associated with familial PD (black arrows) and the currently identified homozygous variant (red arrow) are all located in exon 2. The CHCHD2 p.Ala71Pro amino acid substitution is located within the central conserved hydrophobic domain at the transmembrane element.

4 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Reduction of CHCHD2 but not TOP1MT causes medium that stimulates a dependence on aerobic metabolism the PD pathology compared with cells grown in glucose media, which can rely on To investigate the molecular mechanisms behind this form of anaerobic metabolism. Growth in galactose exacerbated differ- early-onset PD, dermal fibroblasts were established from the ences in mitochondrial morphology seen in patient cells (figure patient and compared with controls established from a healthy 2D). Our experiments excluded the mtDNA abnormalities that individual. We investigated the protein abundance by immu- wouldbelinkedtoTOP1MTdysfunctionandidentified changes noblotting in patient and control fibroblasts to determine in mitochondrial morphology that is related to the proposed which variant was pathogenic. We observed a small reduction function and membrane association identified for CHCHD2 as in TOP1MT levels, whereas the CHCHD2 reduction was summarized in figure e-1, links.lww.com/NXG/A85. more pronounced (figure 2A). As TOP1MT is the top- oisomerase that catalyzes supercolied mtDNA relaxation,12 we The role of other CHCHD family proteins in the mito- investigated mtDNA stability and abundance. Long-range PCR chondrial contact site and cristae organizing system showed no mtDNA fragmentation and changes in mtDNA size (MICOS) complex14,15 lead us to investigate how reduced in patient cells (figure 2B). Furthermore, qPCR analysis showed CHCHD2 expression affects MICOS complex subunits. Al- no significant difference in mtDNA copy number, indicating though levels of the intermembrane protein CHCHD3 were that the TOP1MT variant has negligible effects on mtDNA reduced in patient cells, levels of the lipid binding proteins structure (figure 2, B and C). APOO and APOOL were unchanged (figure 2E). We con- clude that although CHCHD2 does not have a direct role in CHCHD2 has been shown to regulate mitochondrial morphol- MICOS complex function or stability, the CHCHD2 variant ogy in Drosophila melanogaster.13 Therefore, we investigated mi- results in reduction of the membrane-associated protein tochondrial morphology in patient cells using MitoTracker CHCHD3. Orange. In glucose medium, we identified differences in morphology and fusion between patient and control cells, character- Reduced CHCHD2 expression results in ized by a reduced reticular network, fragmentation of filamentous OXPHOS dysfunction mitochondria, and accumulation of granular bodies around the Impaired oxidative phosphorylation (OXPHOS) complex nucleus (figure 2D). We cultured both fibroblasts in galactose formation and function has been identified in PD and

Figure 2 CHCHD2 but not TOP1MT contributes to PD pathogenesis

(A) SDS-PAGE and immunoblotting were performed on 25 μg of isolated mitochondria from control and patient fibroblasts. (B) Mito- chondrial DNA integrity was examined using long-range PCR of control and patient fibroblast DNA. DNA was separated on a 1% agarose gel stained with ethidium bromide. (C) Mitochondrial DNA copy number was examined using quantitative PCR of control and fibroblast DNA. (D) Mitochondrial morphology of control and patient fibroblasts grown in either high glucose or galactose media was examined using MitoTracker Or- ange and visualized by fluorescence microscopy. The scale bar represents 10 μm. Arrows indicate refractile granules characteristic of network fragmentation. (E) Twenty-five micrograms of isolated mitochondria from control and patient fibroblasts was analyzed for mitochondrial contact site and cristae organizing system (MICOS) protein levels by immunoblotting. SDHA was used as a loading control.

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 5 drug-induced parkinsonism.16 Therefore, we investigated production (figure 3D). This indicates that reductions in OXPHOS complex subunit levels by immunoblotting. We OXPHOS activity are likely due to impaired mitochondrial found reduced levels of complex I, IV, and V subunits, morphology and electron transport between complexes. whereas complexes II and III showed negligible reductions in patient cells (figure 3A). De novo mitochondrial trans- To test whether CHCHD2 dysfunction resulted in PD, we lation examined using 35S-methionine showed no difference performed rescue experiments by expressing the wild-type in mitochondrial translation rates between control and pa- CHCHD2 or TOP1MT proteins in the control and patient tient cells (figure 3B). This indicates that the CHCHD2 cells (figure 4). The fragmented mitochondrial network was mutation causes changes in OXPHOS subunit abundance restored to filamentous, reticular appearance (figure 4A), through reduced protein stability and morphologic changes, and respiration was restored to levels comparable to the not impaired protein synthesis. control fibroblast when rescued by expression of the wild- type CHCHD2 protein, but not by expression of the wild- Next, we investigated ETC activity by measuring mitochon- type TOP1MT protein (figure 4B). These experiments drial respiration at each complex (figure 3C). We show re- validated the causative role of the CHCHD2 mutation in PD duced oxygen consumption at complexes I and IV but not pathology. complexes II and III, suggesting that CHCHD2 may regulate electron transfer between these complexes, as previously Mutation in CHCHD2 causes increased suggested.13,17 We also examined the respiratory control ratio oxidative stress (RCR) by measuring oxygen consumption with succinate and Next, we investigated the mitochondrial membrane potential Dψ fi rotenone under ATP-generating conditions (phosphorylating ( m) using JC-1. There were no signi cant changes in the Dψ state 3) and non-ATP generating conditions (non- m in patient cells grown in glucose or galactose, indicating fi fi Dψ fi phosphorylating state 4). We identi ed signi cant decreases that the proton motive force maintains the m ( gure 5A). in the RCR in patient cells grown in glucose and galactose, The decrease in membrane potential was greater in patient which is characteristic of an uncoupling of the ETC and ATP cells treated with FCCP, indicating a reduction in maximal

Figure 3 Patient cells show impaired OXPHOS complex levels and function

(A) SDS-PAGE and immunoblotting were performed on 25 μg of isolated mitochondria from control and patient fibroblasts. (B) Mitochondrial protein synthesis was examined in control and patient fibroblasts using 35S radiolabeling of mitochondrial translation products. Twenty micrograms of cell lysate was separated on a 12.5% SDS-PAGE gel and visualized by autoradiography. (C) Oxygen consumption of specific respiratory complexes was measured in control and patient fibroblasts using an OROBOROS respirometer. (D) Phos- phorylated (state 3) and non- phosphorylated (state 4) respiration was measured in the presence of 10 mM succinate/0.5 μM rotenone in digitonin-permeabilized control and patient cells to determine the respiratory control ratio under both glucose and galactose media conditions. Data are presented as mean ± SEM and were analyzed using a Student’s t test. *p < 0.05, **p < 0.01.

6 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Figure 4 CHCHD2 but not TOP1MT expression rescues molecular defects

(A) Mitochondrial morphology of control and patient fibroblasts, in the presence or absence of either wild-type CHCHD2 or wild-type TOP1MT, was examined using MitoTracker Orange and visualized by fluorescence microscopy. The scale bar represents 10 μm. Arrows indicate refractile granules characteristic of network fragmentation. (B) Oxygen consumption of respiratory complexes was measured in control, patient fibroblasts, and patient fibroblasts expressing either wild-type CHCHD2 or wild-type TOP1MT using an OROBOROS respirometer. Data are presented as mean ± SEM and were analyzed using a Student’s t test. *p < 0.05, **p < 0.01. respiratory capacity. This reduction was greater under galac- Discussion tose conditions (figure 5, A and B). The reduced stability of OXPHOS subunits and decreased respiratory capacity were Here, we have identified homozygous CHCHD2 and TOP1MT independent of mitochondrial mass (figure 5C). variants in a Caucasian woman presenting with characteristic features of PD at 26 years. Both healthy parents were hetero- Next, we investigated reactive oxygen species (ROS) pro- zygous CHCHD2 and TOP1MT variant carriers, but not the duction.18 Patient cells showed significantly increased su- unaffected sister. The likely pathogenicity of the variants was peroxide levels when grown in glucose, which was further supported by the results of the prediction tools PolyPhen-2, exacerbated when cells were grown in galactose (figure SIFT, and CADD, where the TOP1MT variant (c.661G>A, 5D). As ROS have been shown to affect metabolic rate and p.Asp221Asn, rs143769145) was identified at a low frequency viability, we investigated the metabolic consequences of of 0.023%–0.025% in the Genome Aggregation Databases with increased ROS levels. Patient cells showed a mild meta- no homozygotes. The mutated residues of both CHCHD2 and bolic increase under glucose conditions and a further in- TOP1MT are highly conserved during evolution. Taken together, crease under galactose conditions, indicating that the our finding suggests that both CHCHD2 and TOP1MT might be increase in ROS is not inherently cytotoxic, but is sufficient causative genes of recessive early-onset PD. Mitochondrial dys- to modulate the metabolic rate in skin fibroblasts function has been shown to be a key factor in PD pathogenesis in (figure 5E). both animal models and patients, with disease-associated genes

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 7 Figure 5 Patient cells demonstrate increased oxidative stress

(A) JC-1 assay was used to examine the strength of the proton motive force in both control and patient fibroblasts under high glucose (A) and galactose (B) media conditions both in the absence of FCCP (+Dψm) and the presence of FCCP (−Dψm) (n = 6). (C) Mitochondrial mass was measured in control and patient fibroblasts using nonyl acridine orange (NAO) under high glucose conditions (n = 6). (D) Reactive oxygen species (ROS) levels were measured in control and patient fibroblasts using dihydroethidium (DHE) under both high glucose and galactose media conditions (n = 6). (E) Metabolic activity was measured in control and patient fibroblasts using an MTS assay (n = 6). All data are presented as mean ± SEM and were analyzed using a Student’s t test. *p < 0.05, ***p < 0.001.

PINK1 and CHCHD2 being involved in mitochondrial was excluded because expression of this protein did not rescue – function.13,19 21 Recently, however, the association of mitochondrial morphology and function. CHCHD2 with PD has displayed mixed results in different – populations,6,7,19,22 25 suggesting that it is infrequent and Changes in mitochondrial morphology have been previously ethnic specific. Here, we have used cultured patient fibroblasts linked with PD-characteristic OXPHOS deficiencies.16,20,21,27,28 to establish the pathogenic mechanisms of the identified Reductions in complexes I and IV subunits were consistent with CHCHD2 and TOP1MT variants and expand the repertoire reduced activities of these complexes in patient cells. Although of genes implicated in the pathology of PD. reduced complex IV activity has been previously reported in PD, there is much less evidence for impaired complex IV activity, Homozygous variants of both CHCHD2 and TOP1MT affected contributing to PD development.27 The reduction in complex IV protein levels, albeit more pronounced for CHCHD2. In patient formation is consistent with studies finding an interaction fibroblasts, the reduced CHCHD2 expression caused mito- between CHCHD2 and the cytochrome c–binding protein chondrial network fragmentation when cells were forced to rely MICS1.13,29 The reduction in complex activity and function was on oxidative phosphorylation. This is consistent with previous consistent with changes in the mitochondrial membrane and studies that implicated CHCHD2 in mitochondrial morphology indicated an uncoupling of the ETC and impaired electron and its importance for energy production.13,26 The morphology transfer through the ETC,30,31 as seen previously.17 It is possible and respiration defects in the patient fibroblasts were rescued by that reduced CHCHD2 expression causes impaired MICS1 expressing wild-type CHCHD2, providing evidence that the function, therefore impairing movement of electrons from CHCHD2 mutation contributes to the PD pathology. Reduced complex III to IV by cytochrome c, as previously suggested.17 TOP1MT expression causes increased negative supercoiling of 26 Dψ fi fi mtDNA, resulting in reduced mtDNA replication. However, The m was not signi cantly altered in patient broblasts the lack of changes in mtDNA integrity or copy number in despite inefficient electron transfer and reduced respiratory patient fibroblasts indicated that the TOP1MT variant had capacity. The greater difference in uncoupled state shown anegligibleeffect on TOP1MT function. Furthermore, the under galactose conditions indicates that when patient cells contribution of the TOP1MT variant to the PD pathogenesis are made to rely on aerobic energy sources, they can

8 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG compensate for abnormalities in ETC coupling and the H. Goullee by NHMRC EU Collaborative Grant electron leak through increasing the electrons flowing APP1055295, O. Rackham by a Cancer Council WA Research through the OXPHOS complexes. Although increasing elec- Fellowship, and A. Filipovska is supported by NHMRC tron flow may allow cells to maintain the proton motive force, Senior Research Fellowship (APP1005030) and NHMRC and it has negative consequences through ROS formation. Australian Research Council Discovery projects (to A. Filipovska and O. Rackham). The funders had no role in the The increased ROS levels were not cytotoxic and, in contrast, design of the study and collection, analysis, decision to publish, patient fibroblasts demonstrated increased metabolic rate. interpretation of data, or preparation of the manuscript. Although mild increases in ROS levels have been shown to increase metabolic rate and proliferation in certain cells, such Disclosure – as fibroblasts, neuronal cells show reduced viability.32 35 The R.G. Lee, M. Sedghi, M. Salari, A.-M.J. Shearwood, mitochondrial toxicity of oxidized dopamine derivatives, in M. Stentenbach, A. Kariminejad, H. Goullee, and O. Rack- combination with the inherently low mitochondrial mass of ham report no disclosures. N.G. Laing has received a speaker DA neurons, may potentiate mitochondrial dysfunction honorarium from the World Muscle Society; has received resulting from reduced CHCHD2 expression and cause travel funding from the Asian Oceanian Myology Centre, neuronal cell death.33 As dermal fibroblasts may not fully Sanofi, and the Ottawa Neuromuscular Meeting; serves on emulate the characteristics of DA neurons, further in- the editorial board of Neuromuscular Disorders;receives vestigation is needed to understand how increased ROS levels publishing royalties for The Sarcomere and Skeletal Muscle alter metabolic function in DA neurons and how they po- Disease, Springer Science and Business Media, Landes tentiate mitochondrial dysfunction resulting from reduced Bioscience, 2008; and has received research support from CHCHD2 expression. theAustralianNationalHealthandMedicalResearch Council, the US Muscular Dystrophy Association, Associa- Recent findings demonstrate that mutations in CHCHD2 and tion Francaise contre les Myopathies, Foundation Building its paralogue CHCHD10 are associated with multiple neuro- Strength for Nemaline Myopathy, Motor Neuron Disease – degenerative disorders.25,36 38 The existence of a complex Research Institute of Australia, Western Australian Health containing CHCHD2 and CHCHD10 may explain shared and Medical Research Infrastructure Fund, and the Perpet- features between disorders.39,40 In addition, CHCHD10 ual Foundation. H. Tajsharghi and A. Filipovska report no knockdown and knockout models show altered respiratory disclosures. Full disclosure form information provided by activity and OXPHOS subunit levels.39,40 Further in- the authors is available with the full text of this article at vestigation is required to fully understand how this complex Neurology.org/NG. regulates mitochondrial and neurologic functions. Received February 28, 2018. Accepted in final form June 18, 2018. Author contributions A. Filipovska and H. Tajsharghi: study concept and design. R.G. References 1. Liu G, Bao X, Jiang Y, et al. Identifying the association between Alzheimer’s disease Lee, M. Sedghi, M. Salari, A.-M.J. Shearwood, M. Stentenbach, and Parkinson’s disease using genome-wide association studies and protein-protein A. Kariminejad, H. Goullee, and O. Rackham: acquisition of interaction network. Mol Neurobiol 2015;52:1629–1636. 2. Mhyre TR, Boyd JT, Hamill RW, Maguire-Zeiss KA. Parkinson’s disease. Subcell data.R.G.Lee,M.Sedghi,M.Salari,A.-M.J.Shearwood, Biochem 2012;65:389–455. M. Stentenbach, A. Kariminejad, H. Goullee, O. Rackham, 3. Thomas B, Beal MF. Parkinson’s disease. Hum Mol Genet 2007;16: R183–R194. H. Tajsharghi, and A. Filipovska: analysis and interpretation. 4. Lill CM. Genetics of Parkinson’s disease. Mol Cell Probes 2016;30:386–396. R.G. Lee, A.-M.J. Shearwood, H Goullee, and H. Tajsharghi: 5. Funayama M, Ohe K, Amo T, et al. CHCHD2 mutations in autosomal dominant late- statistical analysis. R.G. Lee, O. Rackham, N.G. Laing, onset Parkinson’s disease: a genome-wide linkage and sequencing study. Lancet Neurol 2015;14:274–282. H. Tajsharghi, and A. Filipovska: critical revision of the 6. Shi CH, Mao CY, Zhang SY, et al. CHCHD2 gene mutations in familial and sporadic manuscript for important intellectual content. O. Rackham, Parkinson’s disease. Neurobiol Aging 2016;38:217.e9–217.e13. 7. Jansen IE, Bras JM, Lesage S, et al. CHCHD2 and Parkinson’s disease. Lancet Neurol N.G. Laing, H. Tajsharghi, and A. Filipovska: study supervision. 2015;14:678–679. 8. Rackham O, Davies SMK, Shearwood AMJ, Hamilton KL, Whelan J, Filipovska A. Pentatricopeptide repeat domain protein 1 lowers the levels of mitochondrial leucine Acknowledgment tRNAs in cells. Nucleic Acids Res 2009;37:5859–5867. The authors thank the family members who provided samples 9. Davies SMK, Rackham O, Shearwood AMJ, et al. Pentatricopeptide repeat domain protein 3 associates with the mitochondrial small ribosomal subunit and regulates and clinical information for this study. The authors thank Dr translation. FEBS Lett 2009;583:1853–1858. Maryam Gholami for obtaining skin biopsy. 10. Sanchez MIGL, Mercer TR, Davies SMK, et al. RNA processing in human mitochondria. Cell Cycle Georget Tex 2011;10:2904–2916. 11. Lesage S, Brice A. Parkinson’s disease: from monogenic forms to genetic susceptibility Study funding factors. Hum Mol Genet 2009;18:R48–R59. 12. Zhang H, Zhang YW, Yasukawa T, Dalla Rosa I, Khiati S, Pommier Y. Increased The study was supported by grants from the European negative supercoiling of mtDNA in TOP1mt knockout mice and presence of top- Union’s Seventh Framework Programme for research, tech- oisomerases IIα and IIβ in vertebrate mitochondria. Nucleic Acids Res 2014;42: 7259–7267. nological development and demonstration under grant 13. Meng H, Yamashita C, Shiba-Fukushima K, et al. Loss of Parkinson’s disease- agreement no. 608473 (H. Tajsharghi) and the Swedish Re- associated protein CHCHD2 affects mitochondrial crista structure and destabilizes cytochrome c. Nat Commun 2017;8:15500. search Council (H. Tajsharghi). N.G. Laing is supported by 14. Ding C, Wu Z, Huang L, et al. Mitofilin and CHCHD6 physically interact with Sam50 NHMRC Principal Research Fellowship (APP1117510), to sustain cristae structure. Sci Rep 2015;5:16064.

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 9 15. Li H, Ruan Y, Zhang K, et al. Mic60/Mitofilin determines MICOS assembly essential 28. JasonCannon R, Tapias VM, Na HM, Honick AS, Drolet RE, Greenamyre JT. A highly for mitochondrial dynamics and mtDNA nucleoid organization. Cell Death Differ reproducible rotenone model of Parkinson’s disease. Neurobiol Dis 2009;34:279–290. 2016;23:380–392. 29. Oka T, Sayano T, Tamai S, et al. Identification of a novel protein MICS1 that is 16. Richardson JR, Caudle WM, Guillot TS, et al. Obligatory role for complex I inhibition involved in maintenance of mitochondrial morphology and apoptotic release of cy- in the dopaminergic neurotoxicity of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine tochrome c. Mol Biol Cell 2008;19:2597–2608. (MPTP). Toxicol Sci 2007;95:196–204. 30. Brand MD, Nicholls DG. Assessing mitochondrial dysfunction in cells. Biochem J 17. Aras S, Bai M, Lee I, Springett R, Hüttemann M, Grossman LI. MNRR1 (formerly 2011;435:297–312. CHCHD2) is a bi-organellar regulator of mitochondrial metabolism. Mitochondrion 31. Jastroch M, Divakaruni AS, Mookerjee S, Treberg JR, Brand MD. Mitochondrial 2015;20:43–51. proton and electron leaks. Essays Biochem 2010;47:53–67. 18. Murphy MP. How mitochondria produce reactive oxygen species. Biochem J 2009; 32. Hwang O. Role of oxidative stress in Parkinson’s disease. Exp Neurobiol 2013;22: 417:1. 11–17. 19. Liu Z, Guo J, Li K, et al. Mutation analysis of CHCHD2 gene in Chinese familial 33. Valencia A, Morán J. Reactive oxygen species induce different cell death mechanisms Parkinson’s disease. Neurobiol Aging 2015;36:3117.e7–3117.e8. in cultured neurons. Free Radic Biol Med 2004;36:1112–1125. 20. Morais VA, Haddad D, Craessaerts K, et al. PINK1 loss-of-function mutations affect 34. Emdadul Haque M, Asanuma M, Higashi Y, Miyazaki I, Tanaka K, Ogawa N. mitochondrial complex I activity via NdufA10 ubiquinone uncoupling. Science 2014; Apoptosis-inducing neurotoxicity of dopamine and its metabolites via reactive qui- 344:203–207. none generation in neuroblastoma cells. Biochim Biophys Acta 2003;1619:39–52. 21. Morais VA, Verstreken P, Roethig A, et al. Parkinson’s disease mutations in PINK1 35. Liang CL, Wang TT, Luby-Phelps K, German DC. Mitochondria mass is low in result in decreased Complex I activity and deficient synaptic function. EMBO Mol mouse substantia nigra dopamine neurons: implications for Parkinson’s disease. Exp Med 2009;1:99–111. Neurol 2007;203:370–380. 22. Liu G, Li K. CHCHD2 and Parkinson’s disease. Lancet Neurol 2015;14:679–680. 36. Bannwarth S, Ait-El-Mkadem S, Chaussenot A, et al. A mitochondrial origin for 23. Foo JN, Liu J, Tan E-K. CHCHD2 and Parkinson’s disease. Lancet Neurol 2015;14: frontotemporal dementia and amyotrophic lateral sclerosis through CHCHD10 in- 681–682. volvement. Brain J Neurol 2014;137:2329–2345. 24. IqbalZ,ToftM.CHCHD2andParkinson’s disease. Lancet Neurol 2015;14: 37. Johnson JO, Glynn SM, Gibbs JR, et al. Mutations in the CHCHD10 gene are a common 680–681. cause of familial amyotrophic lateral sclerosis. Brain J Neurol 2014;137:e311. 25. Rubino E, Brusa L, Zhang M, et al. Genetic analysis of CHCHD2 and 38. Müller K, Andersen PM, Hübers A, et al. Two novel mutations in conserved codons CHCHD10 in Italian patients with Parkinson’s disease. Neurobiol Aging 2017; indicate that CHCHD10 is a gene associated with motor neuron disease. Brain J 53:193.e7–193.e8. Neurol 2014;137:e309. 26. Rambold AS, Kostelecky B, Elia N, Lippincott-Schwartz J. Tubular network formation 39. Straub IR, Janer A, Weraarpachai W, et al. Loss of CHCHD10–CHCHD2 complexes protects mitochondria from autophagosomal degradation during nutrient starvation. required for respiration underlies the pathogenicity of a CHCHD10 mutation in ALS. Proc Natl Acad Sci USA 2011;108:10190–10195. Hum Mol Genet 2018;27:178–189. 27. Benecke R, Strümper P, Weiss H. Electron transfer complexes I and IV of platelets are 40. Burstein SR, Valsecchi F, Kawamata H, et al. In vitro and in vivo studies of the ALS- abnormal in Parkinson’s disease but normal in Parkinson-plus syndromes. Brain J FTLD protein CHCHD10 reveal novel mitochondrial topology and protein inter- Neurol 1993;116(pt 6):1451–1463. actions. Hum Mol Genet 2018;27:160–177.

10 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG CLINICAL/SCIENTIFIC NOTES OPEN ACCESS Homozygosity for SCN4A Arg1142Gln causes congenital myopathy with variable disease expression

Christine K. Sloth, MD, Federico Denti, PhD, Nicole Schmitt, PhD, Bo Hjorth Bentzen, PhD, Correspondence Christina Fagerberg, MD, John Vissing, DMSci, and David Gaist, PhD Prof. Gaist [email protected] Neurol Genet 2018;4:e267. doi:10.1212/NXG.0000000000000267

Congenital myopathy has recently been associated with biallelic pathogenic variants in the 1,2 SCN4A gene that encodes the voltage-dependent sodium channel NaV1.4. In 13 previously reported cases,1,2 7 died in utero or shortly after birth. The 6 survivors showed features consistent with “classical” congenital myopathy. Here, we report 2 new familial cases with variable phenotype.

Written informed consent was obtained from both patients. Permission for the study was granted by the Danish Data Protection Agency.

The index patient, an 18-year-old woman, born as the second of 2 children to consanguineous parents, reported weakness and dyspnea from early childhood. Pregnancy and birth were normal, but postpartum, she had dysphagia and was tube fed. Early motor milestones were delayed, with independent ambulation achieved at age 2.25 years and inability to lift her head from a supine position until age 3 years. She experienced improvement in her condition throughout childhood, but still had reduced walking distance (2.5 km), diﬃculty lifting heavy objects, and experienced patella luxations.

On examination, the patient is 150 cm tall and has a dolichocephalic head shape and elbow hypermobility (ﬁgure 1A). Strength testing showed diﬀuse muscle force reduction at Medical Research Council (MRC) grade 4, with no distal/proximal gradient, and axial weakness. Spirometry showed normal forced vital capacity (FVC) (88%) and forced expiratory volume (FEV1) (96%).

Creatine kinase (CK) levels and neurophysiologic findings were normal. Replacement of muscle by fat on MRI was pronounced in gluteus maximus and hamstring muscles (figure 1B). Muscle biopsy, at age 4 years, displayed myopathic features with varying fiber size, increased number of internalized nuclei, atrophic fibers, and endomysial fibrosis and fat infiltration.

Next-generation sequencing revealed homozygosity for a previously described missense variant in SCN4A (NM_000334.4: c.3425G>A(p.Arg.1142Gln)),2 conﬁrmed through Sanger sequencing, and was deemed to be pathogenic by 6 prediction tools. Parents were heterozygous for the variant.

The 22-year-old sister of the index patient was also homozygous for Arg1142Gln. She had milder muscular complaints than her sister, which included diﬃculties lifting her head when lying down, exertional shortness of breath, and poor cycling capacity since childhood. She had elbow joint hypermobility like her sister. Her motor milestones were normal. Strength testing

From the Department of Neurology (C.K.S., D.G.), Odense University Hospital; and Department of Clinical Research (C.K.S., D.G.), Faculty of Health Sciences, University of Southern Denmark, Odense; Department of Biomedical Sciences (F.D., N.S., B.H.B.), Faculty of Health and Medical Sciences, University of Copenhagen; Department of Clinical Genetics (C.F.), Odense University Hospital; and Copenhagen Neuromuscular Center, Department of Neurology (J.V.), Rigshospitalet, University of Copenhagen, Denmark.

Featuring the dococephalic head form in the index patient (left image) and hypermobility of the elbows in both sisters (middle and right image) (A). T1-weighted muscle MRI images of the index patient show severe fatty infiltration and atrophy of gluteus medius (arrows, left image) and adductor magnus and, to a lesser degree, the hamstrings bilaterally (right image) (B).

Figure 2 The mutation R1142Q causes loss-of-function of NaV1.4 current

(A) Representative current traces for NaV1.4 WT and R1142Q recorded from transiently transfected HEK293 cells. (B) I/V relationship for the peak current density for NaV1.4 WT (n = 9) and R1142Q (n = 8). Voltage protocol shown as inset. (C) Steady-state inactivation and activation curves for NaV1.4 WT (black) and R1142Q (gray). Voltage protocol shown as inset. (D) Comparison of activation V50 values between NaV1.4 WT (black) and R1142Q (gray). *p <0.05;**p < 0.006; ****p < 0.0001.

2 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG showed reduced neck flexion (MRC 4+), shoulder abduction J. Vissing and D. Gaist: design, revised the manuscript, (MRC4+), and hip flexion (MRC 4+). CK, lung function interpreted clinical data, and drafted figures. tests, and MRI of thigh muscles were normal. Study funding Functional assessment of the Arg1142Gln (R1142Q) variant in This study did not receive external funding. human embryonic kidney 293 (HEK293) cells revealed partial loss-of-function effects (figure 2), as previously reported in Disclosure 1,2 SCN4A-related congenital myopathy. NaV1.4 R1142Q peak C.K. Sloth, F. Denti, N. Schmitt, B.H. Bentzen, and current density was significantly lower than wild type (WT) C. Fagerberg report no disclosures. J. Vissing has served on (WT: 106.4 ± 12.4 pA/pF, R1142Q: −39.0 ± 6.4 pA/pF at the scientific advisory boards of Sanofi Genzyme, aTyr 5mV,figure 2, A and B), and the voltage dependence of channel Pharma, Ultragenyx Pharmaceuticals, Santhera Pharmaceut- activation was significantly changed (figure 2, C and D). icals, Sarepta Therapeutics, Audentes Therapeutics, Novo Nordisk, Alexion Pharmaceuticals, and Stealth BT; has re- Variants in SCN4A were originally linked to congenital ceived travel funding and speaker honoraria from Sanofi myasthenia,3,4 but recently, also to severe fetal hypokinesia and Genzyme, Ultragenyx Pharmaceuticals, Santhera Pharma- early lethality1 and to sudden infant death syndrome.5 A ceuticals, and aTyr Pharma; serves on the editorial boards of strikingly milder phenotype of “classical” congenital myopathy Neuromuscular Disorders and the Journal of Neuromuscular has been described in 6 patients with SCN4A variants in a re- Diseases; has been a consultant for Sanofi Genzyme, Ultra- cessive pattern, only 3 of whom were adults (aged 18–35 years genyx Pharmaceuticals, Santhera Pharmaceuticals, and aTyr old).1,2 Our 18-year-old index patient exhibited a phenotype Pharma; and has received research support from the Danish similar to that previously reported,1 while her 20-year-old sister Medical Research Council, the University of Copenhagen, the was only marginally affected. Our index patient’s characteristic Augustinus Foundation, the NOVO Nordic Foundation, and muscle MRI findings were similar to 4 other patients with the Lundbeck Foundation. D. Gaist has received honoraria from SCN4A mutations, including 2 brothers, compound hetero- AstraZeneca (Sweden) for participation as a coinvestigator in zygous for c.3425G>A(p.Arg1142Gln) and another missense a research project and has received research support from the variant c.1123T>C (p.Cys375Arg).1,2 The brothers, unlike our Danish Cancer Society. Full disclosure form information pro- patients, had elongated faces, ptosis, facial weakness, scoliosis, vided by the authors is available with the full text of this article at and elevated CK.2 We speculate whether homozygosity for the Neurology.org/NG. p.Arg1142Gln variant conferred the milder phenotype observed in our patients. The present report expands our Received June 26, 2018. Accepted in final form July 26, 2018. knowledge regarding SCN4A-related congenital myopathy in References adulthood and underscores that the phenotype of this disorder 1. Zaharieva IT, Thor MG, Oates EC, et al. Loss-of-function mutations in SCN4A cause may vary considerably, even within members of the same severe foetal hypokinesia or “classical” congenital myopathy. Brain J Neurol 2016;139: ff 6 674–691. family, as in other recessive channelopathies a ecting muscles. 2. Gonorazky HD, Marshall CR, Al-Murshed M, et al. Congenital myopathy with “corona” fibres, selective muscle atrophy, and craniosynostosis associated with novel Author contributions recessive mutations in SCN4A. Neuromuscul Disord NMD 2017;27:574–580. 3. Tsujino A, Maertens C, Ohno K, et al. Myasthenic syndrome caused by mutation of C.K. Sloth: drafted the manuscript and performed adminis- the SCN4A sodium channel. Proc Natl Acad Sci USA 2003;100:7377–7382. trative work. F. Denti: tested the effect of the SCN4A muta- 4. Arnold WD, Feldman DH, Ramirez S, et al. Defective fast inactivation recovery of Nav ff 1.4 in congenital myasthenic syndrome. Ann Neurol 2015;77:840–850. tion in a cell line. N. Schmitt: tested the e ect of the SCN4A 5. Männikkö R, Wong L, Tester DJ, et al. Dysfunction of NaV1.4, a skeletal muscle mutation in a cell line, interpreted data, revised the manu- voltage-gated sodium channel, in sudden infant death syndrome: a case-control study. script, and drafted figures. B.H. Bentzen: tested the effect of Lancet 2018;391:1483–1492. 6. Sahin I, Erdem HB, Tan H, Tatar A. Becker’s myotonia: novel mutations and clinical the SCN4A mutation in a cell line. C. Fagerberg: in charge of variability in patients born to consanguineous parents. Acta Neurol Belg Epub 2018 DNA-testing of the 2 sisters and revised the manuscript. Feb 26.

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 3 CLINICAL/SCIENTIFIC NOTES OPEN ACCESS MT-CYB deletion in an encephalomyopathy with hyperintensity of middle cerebellar peduncles

Annabelle Chaussenot, MD, C´ecile Rouzier, MD, PhD, Konstantina Fragaki, PhD, Sabrina Sacconi, MD, PhD, Correspondence Samira Ait-El-Mkadem, PhD, V´eronique Paquis-Flucklinger, MD, PhD,* and Sylvie Bannwarth, PhD* Dr. Bannwarth [email protected] Neurol Genet 2018;4:e268. doi:10.1212/NXG.0000000000000268 or Dr. Paquis-Flucklinger [email protected]

The oxidative phosphorylation system, involved in cellular adenosine triphosphate production, is composed of 5 complexes (complexes I–V). The complex III catalyzes the transfer of electrons from ubiquinol to cytochrome c and contains 11 subunits among which only cytochrome b is encoded by a mitochondrial gene: MT-CYB. Nonsense, missense, and frameshift mutations in MT-CYB have been reported in patients mainly presenting with an exercise intolerance – associated with encephalomyopathy, cardiomyopathy, or multisystemic disorder.1 4 We report an in-frame 21-base pair (bp) deletion in MT-CYB responsible for a severe clinical presentation with an atypical brain MRI of mitochondrial diseases.

Case report The patient was a 43-year-old woman who presented with recurrent episodes of coma with lactic acidosis since her infancy. Crises were reversible but the symptoms slowly progressed over time leading to cognitive decline, and spastic tetraparesis (walk with bilateral help) with muscle weakness predominantly aﬀecting the proximal muscles of the upper limbs. Visual disturbances started during infancy with cataracts, and the patient became blind at 18 years of age. At age 43, the ophthalmologic examination revealed retinitis pigmentosa and glaucoma. The patient had isolated sinus tachycardia without cardiomyopathy. She presented choreiform movements involving the distal limbs and the tongue. Brain MRI showed bilateral symmetrical T2 hyperintensity of middle cerebellar peduncles with mild cerebellar atrophy and severe cortical and subcortical cerebral atrophy (ﬁgure, A). Lactic acid was elevated (2.62 mmoL/L; normal value, 0.50–2.20 mmoL/L) and the patient also had high levels of serum creatine phosphokinase (387 U/L; normal value, 20–167 U/L). There was no familial history. The patient died suddenly, at age 45, of unknown etiology.

Muscle biopsy was performed after written informed consent was obtained. Histochemical analysis revealed the presence of typical features of mitochondrial myopathy with numerous ragged red fibres. Spectrophotometric analysis of respiratory chain enzyme activities showed fi complex I and combined segment I + III and II + III defects ( gure, B). The Coenzyme Q10 content, measured by liquid chromatography coupled with tandem mass spectrometry detection, was normal. We studied the entire mitochondrial genome from muscle using next generation sequencing (NGS), and we identified a heteroplasmic mitochondrial DNA (mtDNA) deletion affecting the MT-CYB gene. We quantified the mtDNA deletion level from NGS experiments (figure, C) and found a mutation load of 72%. This in-frame deletion is 21-bp long and encompasses nucleotide positions 15385–15405, causing the loss of 7 amino acids (His-Ser-Asp- Lys-Ile-Thr-Phe, p.His215_Phe221del) in the protein. Sanger sequencing confirmed the presence of the deletion in patient’s muscle but we did not find deleted molecules in blood (figure, D). PCR amplification of a short fragment of 255-bp (nucleotide positions 15238–15493) confirmed the

*These authors should be regarded as co-corresponding authors.

From the Universit´eCoteˆ d’Azur, CHU de Nice, Inserm, CNRS, IRCAN, France.

(A) Brain MRI. Axial T2-weighted (A.a) and coronal T2-weighted fast spin-echo (FSE) (A.c) MRI showing bilateral symmetrical hyperintensity of middle cerebellar peduncles. Axial T2-weighted (A.b) and coronal T2-weighted FSE (A.c) showing severe cortical and subcortical cerebral atrophy and mild cerebellar atrophy. (B) Mitochondrial respiratory enzyme activities of muscle from the patient and controls. (C.a) Read coverage analysis from nucleotide 15380 to 15409 showing the base read depth of each nucleotide. (C.b) Sequence from nucleotide 15380 with deduced amino acid sequence showing the 21-bp deletion. (D.a) Sanger sequence chromatogram from patient’s muscle. (D.b) PCR products separated on a 2% agarose gel obtained with primers 15238-forward (59-CTGAG- GAGGCTACTCAGTAG-39) and 15493-reverse (59-GAGGTCTGGTGAGAATAGTGT-39) that encompass the 21-bp deletion. PCR was performed on genomic DNA extracted from patient’s tissues (blood, lane 2; muscle, lane 3) and blood of a control (lane 4). The wild-type and deleted amplicons are 255- and 234-bp, respectively. Lane 1: molecular weight marker 1 kb Plus (ThermoFisher Scientific), lane 5: no template negative PCR control. (E) Western-blot analysis from equal amounts (10 μg) of muscle homogenates of patient (P) and control (C) using oxidative phosphorylation system antibodies cocktail detecting NDUFB8 (complex I), SDHB (complex II), UQCRC2 (complex III), COXII (complex IV), ATP5A (complex V), and anti-GAPDH antibody for loading control. (F) Blue native polyacrylamide gel analysis from muscle homogenates of patient (P) and control (C). Equal amounts (15 μg) of proteins were separated on a 4%–13% acrylamide gradient gel and electroblotted onto a polyvinylidene fluoride membrane prior to incubation with specific antibodies against GRIM19 (subunit of complex I), SDHA (subunit of complex II), UQCRC2 (subunit of complex III), MTCO1 (subunit of complex IV), and ATP5A (subunit of complex V).

high level of deleted mtDNA in muscle and the absence of polyacrylamide gel assay revealed assembly defects of both deletion in blood (figure, D). Consistent with the pathogenicity complexes I and III (figure, F). The deletion removes a β-strand of this deletion, Western blot analysis showed a reduction in of the cytochrome b subunit and this structural modification is complex III protein levels in muscle (figure, E) and blue native- likely responsible for complex III assembly deficiency.

2 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Discussion critical revision of the manuscript for intellectual content. V. Paquis-Flucklinger: study concept and design, critical revision Mutations in the MT-CYB gene are most commonly re- of the manuscript. S. Bannwarth: study concept and design, sponsible for myopathy and exercise intolerance associated acquisition, analysis and interpretation of data. with complex III deficiency. Only a few MT-CYB deletions have been described, and this is the fourth report of an in-frame Acknowledgment 2,4,5 intragenic deletion of this gene. This in-frame 21-bp de- The authors thank Gaëlle Augé, Charlotte Cochaud, and Sandra letion is associated with a severe multisystemic disorder but Foustoul for technical help. without prominent exercise intolerance, reminding the absence of genotype-phenotype correlation in mtDNA diseases. Of in- Study funding terest, the brain MRI shows a bilateral symmetrical T2 hyper- Study funded by the French Ministry of Health and the As- intensity of middle cerebellar peduncles, which is an atypical sociation Française contre les Myopathies (AFM). neuroimaging finding in mitochondrial diseases. This specific MRI sign is usually associated with adrenoleukodystrophy or Disclosure fragile X-associated tremor ataxia syndrome. Mitochondrial A. Chaussenot, C. Rouzier, and K. Fragaki report no dis- 6 dysfunction has been already described in FXTAS, and our closures. S. Sacconi has received speaker honoraria from observation raises the question of the possible role of mito- Sanofi Genzyme, Alnylam, and Biogen. S. Ait-El-Mkadem chondria in brain damage associated with fragile X premutation. reports no disclosures. V. Paquis-Flucklinger has received research support from the French Ministry of Health and Several previous studies reported MT-CYB mutations asso- the Association Française contre les Myopathies (AFM). ciated with deficiency of both complex I and complex III. This S. Bannwarth reports no disclosures. Full disclosure form observation is due to a structural dependence among these 2 information provided by the authors is available with the full complexes, a fully assembled complex III being needed for the text of this article at Neurology.org/NG. stability and the assembly of complex I.7 Our observation is consistent with these data because we found a complex I Received March 16, 2018. Accepted in final form June 21, 2018. activity deficiency with a complex III activity in low values in muscle, whereas complex III levels were highly decreased with References 1. Keightley JA, Anitori R, Burton MD, Quan F, Buist NR, Kennaway NG. Mitochon- an assembly defect of both complexes I and III. drial encephalomyopathy and complex III deficiency associated with a stop-codon mutation in the cytochrome b gene. Am J Hum Genet 2000;67:1400–1410. 2. Carossa V, Ghelli A, Tropeano CV, et al. A novel in-frame 18-bp microdeletion in The MT-CYB deletion is found at a high percentage in our MT-CYB causes a multisystem disorder with prominent exercise intolerance. Hum patient’s muscle and absent in leukocytes as it has been fre- Mutat 2014;35:954–958. fi 3. Emmanuele V, Sotiriou E, Rios PG, et al. A novel mutation in the mitochondrial DNA quently reported. Our ndings highlight the interest of both cytochrome b gene (MTCYB) in a patient with mitochondrial encephalomyopathy, muscle biopsy and NGS to detect small deletions responsible lactic acidosis, and strokelike episodes syndrome. J Child Neurol 2013;28:236–242. ff for mtDNA disease. They also expand the neuroimaging and 4. Mori M, Goldstein J, Young SP, Bossen EH, Sho ner J, Koeberl DD. Complex III deficiency due to an in-frame MT-CYB deletion presenting as ketotic hypoglycemia clinical spectrum associated with MT-CYB mutations. and lactic acidosis. Mol Genet Metab Rep 2015;4:39–41. 5. Rana M, de Coo I, Diaz F, Smeets H, Moraes CT. An out-of-frame cytochrome b gene deletion from a patient with parkinsonism is associated with impaired Author contributions complex III assembly and an increase in free radical production. Ann Neurol 2000; A. Chaussenot: clinical investigation, analysis of data, and 48:774–781. 6. Rizzo G, Pizza F, Scaglione C, et al. A case of fragile X premutation tremor/ataxia critical revision of the manuscript. C. Rouzier: analysis of NGS syndrome with evidence of mitochondrial dysfunction. Mov Disord 2006;21: data and critical revision of the manuscript. K. Fragaki: analysis 1541–1542. 7. Ac´ın-Pérez R, Bayona-Bafaluy MP, Fernández-Silva P, et al. Respiratory complex III is of biochemical experiments and critical revision of the man- required to maintain complex I in mammalian mitochondria. Mol Cell 2004;13: uscript. S. Sacconi: clinical investigation. S. Ait-El-Mkadem: 805–815.

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 3 CLINICAL/SCIENTIFIC NOTES OPEN ACCESS Deoxyguanosine kinase mutation producing juvenile-onset mitochondrial myopathy

F.N.U. Komal, MD, Paolo M. Moretti, MD, and Aziz I. Shaibani, MD Correspondence Dr. Komal Neurol Genet 2018;4:e269. doi:10.1212/NXG.0000000000000269 [email protected]

Mitochondrial myopathies are associated with mutations of both mitochondrial and nuclear DNA. Deoxyguanosine kinase (DGUOK) is a nuclear gene responsible for maintaining the mitochondrial deoxynucleotide pool required for replication and maintenance of mitochondrial DNA (mtDNA). Homozygous or compound heterozygous DGUOK mutations are associated with decreased activity of the mtDNA-encoded respiratory chain complexes, mtDNA – depletions1 4 and deletions.5 We describe a case of juvenile-onset mitochondrial myopathy associated with DGUOK-related multiple mtDNA deletions.

Case report A 38-year-old woman presented with painless slowly progressive symmetrical proximal muscle weakness since she was 8 years old. On initial presentation, she reported difficulty climbing, holding her arms upright, running, and competing with her classmates. Her weakness was largely attributed to her being overweight. For 1 year before presentation, she had also developed diplopia in far vision. However, no primary eye disease was found. She denied shortness of breath, exercise-induced muscle cramps, hearing impairment, palpitations, seizures, and speech difficulty. Her medical history was significant for obsessive-compulsive disorder and Raynaud phenomena. Her family history included her mother and a brother with big calf muscles who were positive for proximal muscle weakness.

Her neurologic examination was notable for limited eye abduction bilaterally and mild right lid ptosis. Other cranial nerves were intact. She had mild calf hypertrophy and scapular winging. She had mild proximal muscle weakness graded on the Oxford scale as 4/5 in deltoids, 4/5 in supraspinatus muscles, 3/5 in hip ﬂexors, and 5/5 elsewhere. Ankle reﬂexes were diminished with no gait abnormality. Cardiac and ear examinations were normal.

Her creatine kinase was elevated to 503 U/L. MRI of the pelvis and thighs without contrast was performed, which showed symmetric fatty infiltration of pelvic and thigh muscles with reduced muscle bulk. EMG did not reveal a myogenic or neurogenic pattern. Biopsy of the left biceps was consistent with a mitochondrial myopathy, with many ragged red and blue fibers, moderate variation in fiber morphology, and many cytochrome oxidase–negative and succinate dehydrogenase–positive fibers (figure, A).

Next-generation sequencing and deletion/duplication analysis of 319 nuclear genes using the blood sample revealed 2 mutations in the DGUOK gene: c.195G>A and c.462T>A in exon 2 and 4, respectively. Parental data were not available to confirm cis or trans configuration. Mitochondrial genome sequencing and deletion analysis was performed using the muscle sample, which identified 3 large deletions of the mitochondrial genome: 10.7, 12.6, and 9.6 kb. The sum total of heteroplasmy of these deletions was estimated to be less than 15 percent as it was confirmed by sequencing and not by array comparative genomic hybridization.

From the University of Texas Health Science Center (F.N.U.K.), Houston; University of Utah (P.M.M.); and Baylor College of Medicine (A.I.S.), Houston, TX.

Copyright © 2018 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 onset of symptoms.3 Moreover, compared with cases with Figure COX:succinate dehydrogenase (SHD) stain and decreased mtDNA content, only a few cases with mtDNA deoxyguanosine kinase (DGUOK) primary deletions producing myopathy have been reported. Six such sequence cases were reported in 2012, in which age at onset of limb or extraocular weakness ranged from 20 to 69 years, and the majority (5) were older than 40 years.5 In contrast, our case presented with proximal muscle weakness at age 8 years and remained free of liver disease. Our participant harbors 2 mutations in the DGUOK gene, which have both been known to occur only in trans to other pathogenic variants, suggesting a trans configuration in our patient. The c.195G>A mutation is predicted to produce a p.Trp65Ter nonsense pathogenic − variant. It has a frequency of 8.12 × 10 6 in a control population of normal individuals sequenced by next-generation sequencing (NGS).7 It has been reported with neonatal hepatocerebral disease in trans to another truncating mutation.4. The second variant, c.462T>A, is predicted to result in the Asn154Lys substitution, with a significantly higher − frequency of 1.34 × 10 4 in normal controls.7 It has been reported in cases of adult-onset PEO in trans with other pathogenic variants.5 The asparagine at position 154 (figure, B) is highly conserved in evolution from Drosophila melanogaster to humans, suggesting an important role in maintaining protein structure or function.

Our case expands the phenotypic spectrum of DGUOK mutations and highlights the importance of NGS in children (A) Muscle biopsy showing many cytochrome oxidase–negative/SHD-positive (blue stained) fibers (×100 magnification). (B) Primary sequence of and adults to timely diagnose mitochondrial myopathy. The DGUOK showing the position of asparagine (N) in different species including markedly slow progression of symptoms observed in our case Drosophila humans (Hs), mouse (Mm), zebrafish (Dr), and (Dm). may have contributed to the delay in diagnosis. Longitudinal studies are needed to further investigate the course and predict the outcomes in patients harboring DGUOK mutations. The patient was diagnosed with DGUOK-related autosomal recessive multiple mitochondrial deletion syndrome, producing Author contributions proximal muscle weakness and external ophthalmoparesis. She F.N.U. Komal: drafting the manuscript. P.M. Moretti: critical was counseled about exercise, dietary, and supportive measures. revision of the manuscript for important intellectual content and preparation of the figure. A.I. Shaibani: acquisition of data Discussion and study supervision. Mitochondrial myopathy is a disease of skeletal muscles, with Study funding or without CNS involvement, caused by defective mito- This was not an industry-supported study. All authors have chondrial metabolism. It is produced by defects in nuclear or reported no financial conflicts of interest. There was no in- mtDNA. Several nuclear genes are responsible for replication vestigational or off-label use. and maintenance of mtDNA including POLG, POLG2, 6 C10ORF2, TYMP, TK2, RR2MB, and DGUOK. Defects in Disclosure ff these genes a ect mtDNA content (number of copies) or This was not an industry-supported study. All authors have cause mtDNA deletions. reported no financial conflicts of interest. There was no investigational or off-label use. F.N.U. Komal reports no Loss-of-function mutations in DGUOK are associated with disclosures. P.M. Moretti has received research support from autosomal recessive inheritance of 3 main phenotypes: the NIH, the New Jersey Commission on Brain Injury Re- mtDNA depletion syndrome-3; noncirrhotic portal hyper- search, and the Department of Veterans Affairs. A.I. Shaibani tension; and autosomal recessive progressive external oph- receives publishing royalties from Oxford University Press. thalmoplegia (PEO) with mtDNA deletions. Mutations in Full disclosure form information provided by the authors is DGUOK have largely been described in mtDNA depletion available with the full text of this article at Neurology.org/NG. syndromes, producing a multisystem illness or isolated liver disease.1,2,4 Myopathy has rarely been observed with early Received February 26, 2018. Accepted in final form June 26, 2018.

2 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG References 4. Dimmock D, Zhang Q, Dionisi-Vici C, et al. Clinical and molecular features of 1. El-Hattab AW, Scaglia F. Mitochondrial DNA depletion syndromes: review and mitochondrial DNA depletion due to mutations in deoxyguanosine kinase. Hum – updates of genetic basis, manifestations, and therapeutic options. Neurotherapeutics Mutat 2008;29:330 331. 2013;10:186–198. 5. Ronchi D, Garone C, Bordoni A, et al. Next-generation sequencing reveals DGUOK 2. Freisinger P, F¨utterer N, Lankes E, et al. Hepatocerebral mitochondrial DNA mutations in adult patients with mitochondrial DNA multiple deletions. Brain 2012; depletion syndrome caused by deoxyguanosine kinase (DGUOK) mutations. Arch 135:3404–3415. Neurol 2006;63:1129–1134. 6. Copeland WC. Defects in mitochondrial DNA replication and human disease. Crit 3. Buchaklian AH, Helbling D, Ware SM, Dimmock DP. Recessive deoxyguanosine Rev Biochem Mol Biol 2012;47:64–74. kinase deﬁciency causes juvenile onset mitochondrial myopathy. Mol Genet Metab 7. Lek M, Karczewski KJ, Minikel EV, et al. Analysis of protein-coding genetic variation 2012;107:92–94. in 60,706 humans. Nature 2016;536:285.

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 3 CLINICAL/SCIENTIFIC NOTES OPEN ACCESS CADASIL aﬀecting a black African man

Louis Vlok, MD, and Naeem Brey, MD Correspondence Dr. Vlok Neurol Genet 2018;4:e270. doi:10.1212/NXG.0000000000000270 [email protected]

Cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) is an important genetic cause of strokes and vascular dementia.1 The disease is characterized by 5 main symptoms: migraines, subcortical ischemic events, mood disturbances, apathy, and cognitive impairment.

The causative mutation lies in the Notch3 gene on chromosome 19. The gene encodes for the NOTCH3 protein, which is expressed in vascular smooth muscle cells, and mutations lead to the development of arterial stenosis.1

CADASIL has been described mostly in Caucasian families, but has been diagnosed in- frequently in Arabian2 and Asian families.3 Only 1 case report on an African American family has been previously published.4 There are no cases documenting indigenous black sub-Saharan African patients.

Case report The patient is a 36-year-old black Zimbabwean man. He awoke one morning with right-sided hemiparesis. There was no history of a prior stroke-like episode. The patient denied social risk factors and reported occasional migraine-like headaches that started in his late twenties. There were no other signiﬁcant medical or surgical comorbidities.

He recalled that a paternal uncle previously had a stroke, while another paternal uncle was admitted to a psychiatric hospital and died of an unknown cause. He also noted that his paternal grandfather died of a stroke in his sixties. Unfortunately, most of his family was displaced because of the political unrest in Zimbabwe, and he had lost contact with them.

On admission, the patient had normal vital signs and no evidence of cardiovascular disease. Neurologic examination showed intact language function. Examination of the motor system revealed a right hemiparesis, with symmetrical brisk reﬂexes. Sensation was normal, as was coordination.

Laboratory tests, including full blood count, lipogram, fasting glucose, HIV serology, syphilis serology, antinuclear factor, and erythrocyte sedimentation rate were all normal. ECG, carotid Doppler, and transthoracic echocardiogram were normal.

MRI detected T2 bilateral deep periventricular white matter hyperintensities. These lesions were especially confluent in the centrum semiovale (figure). There were also bilateral subcortical lesions in the anterior temporal lobes and diffuse cerebral atrophy.

Given a history of migraines and a history of young-onset stroke in the patient’s family, a diagnosis of CADASIL was considered.

From the Department of Neurology, Tygerberg Academic Hospital, University of Stellenbosch, Cape Town, South Africa.

The Article Processing Charge was waived at the discretion of the Editor. 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 © 2018 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Figure T2 FLAIR images depict subcortical lesions in the anterior temporal lobes and deep white matter with confluent centrum semiovale hyperintensities

Molecular genetic analysis of the Notch3 gene was per- In the correct clinical context, along with suggestive MRI formed. This revealed a heterozygous missense mutation, ﬁndings, CADASIL could be considered in a young-onset which resulted in a substitution of a cysteine for serine stroke patient, regardless of his or her ancestry. amino acid at codon 183 within exon 4. The patient’sfamily could not be tested as they remained in Zimbabwe at the Author contributions time of diagnosis. L. Vlok: writing of the initial draft of the manuscript, design, and conceptualization of the study. N. Brey: critical revision of the manuscript for intellectual concept.

Discussion Study funding Since the discovery of the Notch3 mutation in 1993, many No targeted funding received. patients with CADASIL have been identified worldwide. It is rare to identify a CADASIL mutation in a black indigenous Disclosure African patient. However, the disease progression and imag- L. Vlok is a registrar in neurology at the University of ing findings are typical, and the specific mutation has been Stellenbosch (receives bursary from the Mediclinic private previously identified in other confirmed cases.5 hospital group). N. Brey reports no disclosures. Full disclosure form information provided by the authors is available Our patient is originally from the eastern highlands of Zim- with the full text of this article at Neurology.org/NG. babwe. This mountainous region range forms a natural border fi with neighboring Mozambique, which was previously a Por- Received April 10, 2018. Accepted in nal form June 18, 2018. tuguese colony. Although speculative, it is possible that the References mutation could have been introduced by a Portuguese im- 1. Chabriat H, Joutel A, Dichgans M, Tournier-Lasserve E, Bousser MG. Cadasil. Lancet migrant, accounting for how this typically Caucasian mutation Neurol 2009;8:643–653. fi 2. Bohlega S, Al Shubili A, Edris A, et al. CADASIL in Arabs: clinical and genetic findings. was found in this patient. In addition, the speci c mutation BMC Med Genet 2007;8:67. (c. 547 T > A; p. Cys183Ser) has also been described in other 3. Tan ZX, Li FF, Qu YY, Liu J, et al. Identification of a known mutation in Notch 3 in 6 familiar CADASIL in China. PLoS One 2012;7:e36590. European families. 4. Lee SL, Meng H, Elmadhoun O, et al. Cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy affecting an African American man: identification of a novel 15–base pair NOTCH3 duplication. Arch Neurol 2011;68: De novo mutations have been described in syndromic 1584–1586. patients without a significant family history. Furthermore, 5. Rutten J, Haan J, Terwindt G, Van Duinen S, Boon E. Interpretation of NOTCH3 mutations in the diagnosis of CADASIL. Expert Rev Mol Diagn 2014;14:593–603. CADASIL-causing mutations cluster in regions of sequence 6. Dichgans M, Ludwig H, Müller-Höcker J, Messerschmidt A, Gasser T. Small in-frame 7 diversity and low evolutionary conservation. Thus, the deletions and missense mutations in CADASIL: 3D models predict misfolding of Notch3 EGF-like repeat domains. Eur J Hum Genet 2000;8:280–285. possibility exists that the mutation may have arisen 7. Donahue CP, Kosik KS. Distribution pattern of Notch3 mutations suggests a gain-of- spontaneously. function mechanism for CADASIL. Genomics 2004;83:59–65.

2 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG CLINICAL/SCIENTIFIC NOTES OPEN ACCESS DRESS after IV phenytoin associated with cytochrome P450 CYP2C9*3 homozygosity

Mette S. Nissen, MD, and Christoph P. Beier, MD Correspondence Dr. Beier Neurol Genet 2018;4:e272. doi:10.1212/NXG.0000000000000272 [email protected]

Phenytoin is a first-generation antiepileptic drug, which is used in the treatment of focal seizures and as standard of care for patients with benzodiazepine-refractory status epilepticus. Elimi- nation of phenytoin occurs primarily via CYP enzyme–dependent hepatic clearance. The first step is the para-hydroxylation from active phenytoin to inactive hydroxy-phenytoin, which is dependent on CYP2C9 and, to a lesser extent, on CYP2C19. A considerable disparity in CYP2C9 alleles exists, and the frequency varies among different ancestry groups, the CYP2C9*3 allele being less frequent in Caucasians and more frequent in Asian populations.1,2 Homozygosity for the CYP2C9*3 allele leads to a significant reduction in enzyme activity, resulting in increased plasma levels of phenytoin.3

Severe cutaneous adverse drug reactions (SCARs) such as Steven Johnson syndrome, toxic epidermal necrolysis, or drug reaction with eosinophilia and systemic symptoms (DRESS) are known complications of phenytoin treatment. The pharmacogenetic basis of SCARs due to phenytoin treatment is not yet fully understood. Certain human leukocyte antigen (HLA) subtypes (e.g., HLA-B*13:01, HLA-B*56:02/04) are associated in some Asian populations with an increased risk of developing SCARs, however, the association is less pronounced as compared to carbamazepine.4 In 2014, CYP2C9 polymorphism (CYP2C9*3) was ﬁrst described as a new and highly signiﬁcantly associated nonimmunologic genetic risk factor for SCARs in a Japanese population.4

Case report Here, we report a case of DRESS after IV administered phenytoin in a patient with CYP2C9*3 homozygosity. The 54-year-old, non-consanguineous, female refugee from northern Iran (and born in Pakistan) presented to the neurologic department at the Odense University Hospital (Denmark) with a first-time bilateral tonic-clonic seizure, with subsequent series of focal seizures with impaired consciousness, fulfilling status epilepticus criteria. Treatment with 10 mg of IV diazepam remained ineffective. Fosphenytoin (20 mg phenytoin-equivalents/kg) was administered IV, and the seizures were terminated. As per protocol, the patient received a maintenance dose of phenytoin 100 mg q8h. MRI of the brain revealed sequela after traumatic brain injury in the left frontal lobe as likely epileptic focus; routine blood tests were normal. Two days after admission, the patient developed nystagmus, ataxia, vertigo, nausea, and universal pruritus. Blood tests revealed plasma phenytoin levels equal to phenytoin poisoning. Despite initial reduction and termination of the maintenance dose of phenytoin, plasma levels continued to increase in the following days, reaching a maximum of 142 μM(35.8μg/mL) 4 days after the first administration. At day 8, the patient developed SCARs that fulfilled the RegiSCAR criteria for DRESS.5 She developed fever (39.5°C), hypotension, acute rash (figure, A), leukocytosis, lever abnormalities (alanine transaminase > 400 U/L), and lymphadenopathy. PET-CT 10 days after the start of symptoms revealed universal lymphadenopathy (figure, B), resembling lymphoma. Liver biopsy showed inflammation compatible with hepatotoxicity. Skin biopsy showed inflammation, but with no significant eosinophilia. The patient’s antiepileptic treatment was changed to levetiracetam.

From the Department of Neurology (M.S.N., C.P.B.), Odense University Hospital; and Department of Clinical Research (C.P.B.), University of Southern Denmark, Odense, Denmark.

The Article Processing Charge was funded by Free Research Grants, Odense University Hospital. 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.

(A) The patient developed severe rash on the entire body (image taken at day 24). (B) Universal lymphadenopathy visualized by PET-CT at day 18. (C) PET-CT 1 month after the hospital stay showed complete recovery of lymphadenopathy (58 days after admission).

After 2 weeks, the symptoms slowly started to remit as the Author contributions plasma levels of phenytoin normalized, and after 4 weeks, the M.S. Nissen and C.P. Beier treated the patient and wrote and patient was fully recovered. A control PET-CT was performed ﬁnally approved the manuscript. 6 weeks after debut of the symptoms to rule out lymphoma; the lymphadenopathy had completely resolved (ﬁgure, C). Ge- Study funding netic testing revealed homozygosity for the P450 CYP2C9*3 The study was supported by the Region of Southern Denmark allele. The HLA types were HLA-A 30:68 and HLA-B 14:15. to C.P.B. (17/18517).

Disclosure Discussion M.S. Nissen reports no disclosures. C.P. Beier served on the scientific advisory boards of UCB and Eisai; received speaker Plasma phenytoin levels are greatly affected by CYP2C9 honoraria from UCB and Eisai; and received research support polymorphisms. It is estimated that roughly 5% of the Scan- from UCB, Eisai, Novartis, Pfizer, and the Region of Southern dinavian population are heterozygous for CYP2C9*3 allele;2 Denmark. Full disclosure form information provided by however, the prevalence is substantially higher in India and in the authors is available with the full text of this article at the Middle East, where up to 16% carry a CYP2C9*3 allele, Neurology.org/NG. resulting in a higher proportion of homozygous patients (up to 1 4,6 7 1.6% in India ). In several Asian but not in Caucasian Received April 30, 2018. Accepted in final form July 26, 2018. populations, the CYP2C9*3 allele was associated with SCARs such as DRESS. The ethnic differences may be due to im- References ff 1. Gaikwad T, Ghosh K, Shetty S. VKORC1 and CYP2C9 genotype distribution in munologic di erences that are also seen in analyses focusing on – 4,6,7 ’ Asian countries. Thromb Res 2014;134:537 544. HLA subtypes associated with SCARs. Our patient shis- 2. Pedersen RS, Verstuyft C, Becquemont L, Jaillon P, Brosen K. Cytochrome P4502C9 tory illustrates the impressive systemic and cutaneous side (CYP2C9) genotypes in a Nordic population in Denmark. Basic Clin Pharmacol ff Toxicol 2004;94:151–152. e ects of phenytoin associated with CYP2C9*3 homozygosity 3. Silvado CE, Terra VC, Twardowschy CA. CYP2C9 polymorphisms in epilepsy: in this Asian patient. It stresses the importance of being aware influence on phenytoin treatment. Pharmacogenomics Pers Med 2018;11: 51–58. of interracial genetic variability that can result in substantially 4. Chung WH, Chang WC, Lee YS, et al. Genetic variants associated with phenytoin- increased risk of intoxication and greater risk of developing related severe cutaneous adverse reactions. JAMA 2014;312:525–534. severe side effects such as SCARs. For practical purposes, ge- 5. Kardaun SH, Sekula P, Valeyrie-Allanore L, et al. Drug reaction with eosinophilia and systemic symptoms (DRESS): an original multisystem adverse drug reaction. Results netic screening before IV phenytoin treatment is not feasible. from the prospective RegiSCAR study. Br J Dermatol 2013;169:1071–1080. Increasing phenytoin plasma concentrations despite dose 6. Tassaneeyakul W, Prabmeechai N, Sukasem C, et al. Associations between HLA class I and cytochrome P450 2C9 genetic polymorphisms and phenytoin-related severe reductions, as seen in this patient, has to prompt precaution cutaneous adverse reactions in a Thai population. Pharmacogenet Genomics 2016;26: and eventually rapid change of the antiepileptic treatment. 225–234. 7. McCormack M, Urban TJ, Shianna KV, et al. Genome-wide mapping for clinically In patients with established homozygosity for the P450 relevant predictors of lamotrigine- and phenytoin-induced hypersensitivity reactions. CYP2C9*3 allele, phenytoin treatment should be avoided. Pharmacogenomics 2012;13:399–405.

2 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG CLINICAL/SCIENTIFIC NOTES OPEN ACCESS AP4S1 splice-site mutation in a case of spastic paraplegia type 52 with polymicrogyria

Susana Carmona, PhD, Clara Marecos, MD, Marta Amorim, MD, Ana C. Ferreira, MD, Carla Conceição, MD, Correspondence Jos´eBr´as, PhD, Sofia T. Duarte, MD, PhD,* and Rita Guerreiro, PhD* Dr. Guerreiro [email protected] Neurol Genet 2018;4:e273. doi:10.1212/NXG.0000000000000273

Hereditary spastic paraplegias (HSPs) are a group of rare inherited neurodegenerative disorders that result from primary retrograde dysfunction of the long descending ﬁbers of the corticospinal tract, causing lower limb spasticity and muscular weakness. This group of diseases has a heterogeneous clinical presentation. An extensive list of associated genes, diﬀerent inheritance patterns, and ages at onset have been reported in HSPs.1 Spastic paraplegia type 52 (SPG52) is an autosomal recessive disease caused by AP4S1 mutations. The disease is characterized by neonatal hypotonia that progresses to hypertonia and spasticity in early childhood, developmental delay, mental retardation, and poor or absent speech. Febrile or afebrile seizures – may also occur.2 4

Clinical case presentation We report the case of a Portuguese 2-year-old boy born to healthy nonconsanguineous parents after a full-term gestation with intrauterine growth restriction after week 37. During the first months of life, the patient presented poor weight gain, hyperammonemia with elevation of glutamine and ornithine, low citrulline, and negative orotic acid. Weight recovery and nor- malization of amino acid profile were observed after protein restriction and remained normal after reintroduction of normal diet. Genetic study of urea cycle disorders (NAGS, CPS, and OTC) was negative. Around 9 months of age, global developmental delay, hypotonia, and strabismus were evident. Brain MRI with spectroscopy (performed at 10 months) showed delayed myelination/hypomyelination associated with a posterior perisylvian polymicrogyria, thinning of the corpus callosum, dilation and dysmorphia of the ventricles, and enlargement of the subarachnoid frontotemporal space (figure A). Spectroscopy suggested a possible discrete reduction of N-acetylaspartate. EEG showed a slight intermittent lentification in the left temporal region. At 15 months of age, the patient had 1 afebrile episode of status epilepticus. Two previous shorter episodes with fever had also occurred. Levetiracetam was started. No regression of psychomotor development after seizure was observed, and the patient has been evolving gradually with improvement of axial hypotonia. He says a few simple words, responds to his name, and has some nonverbal communication. The most recent neurologic evaluation revealed an alteration of the muscle tone (hypertonia) in the left lower limb and pyramidal signs in both legs.

Exome sequencing of the proband and parents was performed as described in Supplementary Material (links.lww.com/NXG/A86) and revealed the homozygous AP4S1 splice site NM_001128126.2:c.294+1G>T r.619_687del variant in the proband, present in the heterozygous state in the parents (ﬁgure B). The variant was located in a 2.4-Mb homozygous region

*These authors contributed equally to the manuscript.

From the Department of Molecular Neuroscience (S.C., J.B., R.G.), UCL Institute of Neurology, University College London, United Kingdom; Paediatric Neurology Department (C.M., S.T.D.), Hospital Dona Estefânia, Centro Hospitalar de Lisboa Central; Genetics Department (M.A.), Hospital Dona Estefânia, Centro Hospitalar de Lisboa Central; Reference Center of Inherited Metabolic Diseases (A.C.F.), Centro Hospitalar de Lisboa Central; Neuroradiology Department (C.C.), Hospital Dona Estefânia, Centro Hospitalar de Lisboa Central, Lisbon, Portugal; UK Dementia Research Institute (J.B., R.G.), University College London, United Kingdom; and Department of Medical Sciences (J.B., R.G.), Institute of Biomedicine, iBiMED, University of Aveiro, Portugal.

Copyright © 2018 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Figure Patient’s brain imaging, pedigree, and consequences of AP4S1 c.294+1G>T variant

(A) Left image: sagittal T1 weighted image showing thin corpus callosum. Central and right images: axial T2 weighted images showing delayed myelination, bilateral posterior perisylvian polymicrogyria, dysmorphic and enlarged ventricles, and enlargement of subarachnoid space. (B) Family pedigree. The proband presents the splice-site variant chr14:g.31542180G>T NM_001128126.2:c.294+1G>T in the homozygous state, and both parents are heterozygous for the variant. +: c.294+1G>T allele; −: wild-type allele. (C) AP4S1 transcript size of the homozygous patient, both heterozygous parents, and the wild-type individual. A shorter transcript is produced in the presence of the variant. Each band of the marker ladder represents 100 bp (band size from gel bottom to top: 100, 200, 300, 400, 500, 600, 700, 800, 900, and 1000 bp). (D) Alignment of the wild-type protein sequence (NP_001121598) to the mutated protein. The alignment was performed with Clustal Omega. The variant leads to the loss of amino acids 76–98. C− = negative control; F = father; M = marker ladder; Mo = mother; mut = mutated protein; P = patient; WT = wild-type.

of chromosome 14. This variant is extremely rare in the ventricles and thinning of the corpus callosum.7 Similar population, with only 1 heterozygous individual present in changes have also been seen in the patient described here, the Genome Aggregation Database. In silico analysis predicted together with febrile and afebrile seizures. When exome se- the loss of the donor splice site of exon 4. A transcript size quencing was performed and analyzed, the patient did not analysis and Sanger sequencing of cDNA confirmed the show hypertonia in the lower limbs. However, as reported in presence of a shorter transcript skipping exon 4 associated other patients, this clinical entity may progress from hypotonic with the variant (figure C). As a consequence, the polypeptide to hypertonic status. The most recent neurologic evaluation of 23 amino acids (76 a.a.–98 a.a.) encoded by exon 4 is lost revealed the presence of hypertonia in the left leg, associated (figure D). with pyramidal signs, suggesting the possibility of future development of a spastic paraparesis, typical of this disease.

Discussion Here, we report a case of SPG52 associated with posterior perisylvian polymicrogyria, unexplained transitory hyper- AP4S1 encodes the small subunit of the adaptor protein ammonemia, and absence of facial dysmorphisms, which complex-4 (AP4 complex). This complex is recruited to the suggest an expansion of the disease phenotype. trans-Golgi network, where it mediates vesicle traﬃcking to endosomes or basolateral plasma membrane in a clathrin- independent manner.5 Mutations in the 4 subunits of the Author contributions complex have been associated with similar autosomal recessive S. Carmona: study concept and design; acquisition, analysis phenotypes mainly characterized by spastic tetraplegia.6 The and interpretation of data; and writing of the manuscript. mutation found in our patient leads to the loss of exon 4, with C. Marecos, M. Amorim, A.C. Ferreira, and C. Conceição: predicted important consequences to the protein structure and acquisition of patient data and critical revision of the manu- the AP4 complex function. Anatomical changes similar to script for intellectual content. J. Br´as: study concept and those observed in patients have been reported in an AP-4 design; analysis and interpretation of data; and critical re- complex knockout mouse model: enlargement of the lateral vision of the manuscript for intellectual content. S.T. Duarte:

2 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG acquisition of patient data; analysis and interpretation of data; Journal of Parkinson’s Disease and has received research sup- and writing and critical revision of the manuscript for in- port from the Alzheimer’s Society. Full disclosure form intellectual content. R. Guerreiro: study concept and design; formation provided by the authors is available with the full analysis and interpretation of data; study supervision; and text of this article at Neurology.org/NG. writing and critical revision of the manuscript for important intellectual content. Received February 20, 2018. Accepted in final form July 31, 2018. References Acknowledgment 1. de Souza PVS, de Rezende Pinto WBV, de Rezende Batistella GN, Bortholin T, The authors acknowledge the family who participated in this Oliveira ASB. Hereditary spastic paraplegia: clinical and genetic hallmarks. Cerebel- lum 2017;16:525–551. study. 2. Abou Jamra R, Philippe O, Raas-Rothschild A, et al. Adaptor protein complex 4 deficiency causes severe autosomal-recessive intellectual disability, progressive spastic Study funding paraplegia, shy character, and short stature. Am J Hum Genet 2011;88:788–795. ’ 3. Hardies K, May P, Djémié T, et al. Recessive loss-of-function mutations in AP4S1 cause This work was partially funded by the Alzheimer s Society. mild fever-sensitive seizures, developmental delay and spastic paraplegia through loss of AP-4 complex assembly. Hum Mol Genet 2015;24:2218–2227. Disclosure 4. Tessa A, Battini R, Rubegni A, et al. Identification of mutations in AP4S1/SPG52 through next generation sequencing in three families. Eur J Neurol 2016;23:1580–1587. S. Carmona, C. Marecos, M. Amorim, A.C. Ferreira, and 5. Park SY, Guo X. Adaptor protein complexes and intracellular transport. Biosci Rep C. Conceição report no disclosures. J. Brás serves on the 2014;34:e00123. ’ 6. Tüysüz B, Bilguvar K, Koçer N, et al. Autosomal recessive spastic tetraplegia caused by editorial board of the Journal of Parkinson s Disease and has AP4M1 and AP4B1 gene mutation: expansion of the facial and neuroimaging features. received research support from the Alzheimer’s Society. Am J Med Genet A 2014;164A:1677–1685. S. Temudo Duarte has received travel funding from Nutricia. 7. Ivankovic D, López-Doménech G, Drew J, Tooze SA. AP-4 Mediated ATG9A Sorting Underlies Axonal and Autophagosome Biogenesis Defects in a Mouse Model of AP-4 R. Guerreiro serves on the editorial boards of ScienceMatters, Deficiency Syndrome. bioRxiv. Epub 2017. Available at: biorxiv.org/content/early/ the American Journal of Neurodegenerative Disease, and the 2017/12/16/235101.abstract. Accessed January 31, 2018.

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 3 CORRECTION Conﬁrming TDP2 mutation in spinocerebellar ataxia autosomal recessive 23 (SCAR23) Neurol Genet 2018;4:e277. doi:10.1212/NXG.0000000000000277

In the Article "Conﬁrming TDP2 mutation in spinocerebellar ataxia autosomal recessive 23 (SCAR23)" by G. Zagnoli-Vieira et al.1, there is an error in the seventh author’s name, which should have read “Robert W. Taylor” rather than “Robert Taylor” as originally published. The authors regret the error.

REFERENCE 1 Zagnoli-Viera G, Bruni F, Thompson K, et al. Conﬁrming TDP2 mutation in spinocerebellar ataxia autosomal recessive 23 (SCAR23). Neurol Genet 2018;4:e262.

Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 1