lww nng nngcover 4 4 - genetics · academy officers ralphl.sacco,md,ms,faan,president...

Volume 4, Number 4, August 2018Neurology.org/NG

A peer-reviewed clinical and translational neurology open access journal

ARTICLE

ASFMR1 splice variant: A predictor of fragile X-associated tremor/ataxia syndrome e246

ARTICLE

Noncoding repeat expansions for ALS in Japan are associated with the ATXN8OS gene e252

ARTICLE

SCN11A Arg225Cys mutation causes nociceptive pain without detectable peripheral nerve pathology e255

ARTICLE

Longitudinal analysis of contrast acuity in Friedreich ataxia e250

Academy OfficersRalph L. Sacco, MD, MS, FAAN, PresidentJames C. Stevens, MD, FAAN, President ElectAnn H. Tilton, MD, FAAN, Vice PresidentCarlayne E. Jackson, MD, FAAN, SecretaryJanis M. Miyasaki, MD, MEd, FRCPC, FAAN, TreasurerTerrence L. Cascino, MD, FAAN, Past President

Executive Office, American Academy of NeurologyCatherine M. Rydell, CAE, Executive Director/CEO20l Chicago AveMinneapolis, MN 55415Tel: 612-928-6000

Editorial OfficePatricia K. Baskin, MS, Executive EditorKathleen M. Pieper, Senior Managing Editor, NeurologyLee Ann Kleffman, Managing Editor, Neurology: GeneticsSharon L. Quimby, Managing Editor, Neurology® Clinical PracticeMorgan S. Sorenson, Managing Editor,Neurology® Neuroimmunology & NeuroinflammationCynthia S. Abair, MA, Senior Graphics EditorAndrea R. Rahkola, Production Editor, NeurologyRobert J. Witherow, Senior Editorial AssociateKaren Skaja, Senior Editorial AssociateKaitlyn Aman Ramm, Editorial AssistantKristen Swendsrud, Editorial AssistantAndrea Willgohs, Editorial Assistant

PublisherWolters KluwerBaltimore, MD

Publishing StaffKim Jansen, Executive PublisherJessica Heise, Production Team Leader, Neurology JournalsMegen Miller, Production EditorSteve Rose, Editorial AssistantStacy Drossner, Production Associate

Neurology® is a registered trademark of the American Academy of Neurology(registration valid in the United States).

Neurology® Genetics (eISSN 2376-7839) is an open access journal publishedonline for the American Academy of Neurology, 201 Chicago Avenue,Minneapolis, MN 55415, by Wolters Kluwer Health, Inc. at 14700 CiticorpDrive, Bldg. 3, Hagerstown, MD 21742. Business offices are located at TwoCommerce Square, 2001 Market Street, Philadelphia, PA 19103. Productionoffices are located at 351 West Camden Street, Baltimore, MD 21201-2436.© 2018 American Academy of Neurology.

Neurology® Genetics is an official journal of the American Academy ofNeurology. Journal website: Neurology.org/ng, AAN website: AAN.com

Copyright and Permission Information: Please go to the journal website(www.neurology.org/ng) and click the Permissions tab for the relevantarticle. Alternatively, send an email to [email protected] information about permissions can be found here: https://shop.lww.com/journal-permission.

Disclaimer: Opinions expressed by the authors and advertisers are notnecessarily those of the American Academy of Neurology, its affiliates, or ofthe Publisher. The American Academy of Neurology, its affiliates, and thePublisher disclaim any liability to any party for the accuracy, completeness,efficacy, or availability of the material contained in this publication(including drug dosages) or for any damages arising out of the useor non-use of any of the material contained in this publication.

Advertising Sales Representatives: Wolters Kluwer, 333 Seventh Avenue,New York, NY 10001. Contacts: Eileen Henry, tel: 732-778-2261, fax: 973-215-2485, [email protected] and in Europe: Avia Potashnik, WoltersKluwer, tel: +44 207 981 0722; +44 7919 397 933 or e-mail: [email protected].

Careers & Events:Monique McLaughlin, Wolters Kluwer, Two CommerceSquare, 2001Market Street, Philadelphia, PA 19103, tel: 215-521-8468, fax: 215-521-8801; [email protected].

Reprints:Meredith Edelman, Commercial Reprint Sales, Wolters Kluwer, TwoCommerce Square, 2001Market Street, Philadelphia, PA 19103, tel: 215-356-2721;[email protected]; [email protected].

Special projects: US & Canada: Alan Moore, Wolters Kluwer, TwoCommerce Square, 2001 Market Street, Philadelphia, PA 19103, tel:215-521-8638, [email protected]. International: AndrewWible, Senior Manager, Rights, Licensing, and Partnerships, Wolters Kluwer;[email protected].

http://Neurology.org/ng

mailto:[email protected]

https://shop.lww.com/journal-permission

https://shop.lww.com/journal-permission









A peer-reviewed clinical and translational neurology open access journal Neurology.org/NG

Neurology® Genetics

EditorStefan M. Pulst, MD, Dr med, FAAN

Deputy EditorMassimo Pandolfo, MD, FAAN

Associate EditorsAlexandra Durr, MD, PhDMargherita Milone, MD, PhDRaymond P. Roos, MD, FAANJeffery M. Vance, MD, PhD

Editorial BoardHilaryCoon, PhDGiovanniCoppola,MDChantalDepondt,MD,PhDBrentL. Fogel,MD,PhD,FAANAnthony J.Griswold, PhDOrhunH.Kantarci,MDJulieR.Korenberg, PhD,MDMargheritaMilone,MD,PhDDavidePareyson,MDShojiTsuji,MD,PhDDinekeS. Verbeek, PhDDavidViskochil,MD,PhDJulianeWinkelmann,MDJuan I. Young, PhD

Neurology® Journals

Editor-in-ChiefRobert A. Gross, MD, PhD, FAAN

Deputy EditorBradford B. Worrall, MD, MSc, FAAN

Section Editors

BiostatisticsRichard J. Kryscio, PhDChristopher A. Beck, PhDSue Leurgans, PhD

Classification of Evidence EvaluationsGary S. Gronseth, MD, FAAN

PodcastsStacey L. Clardy, MD, PhDJeffrey B. Ratliff, MD, Deputy Podcast Editor

OmbudsmanDavid S. Knopman, MD, FAAN

Scientific Integrity AdvisorRobert B. Daroff, MD, FAAN

Classification of EvidenceReview Team

Melissa J. Armstrong,MDRichardL.Barbano,MD,PhD,FAANRichardM.Dubinsky,MD,MPH,FAANJeffrey J. Fletcher,MD,MScGaryM.Franklin,MD,MPH,FAANDavid S.Gloss II,MD,MPH&TMJohn J.Halperin,MD,FAANJasonLazarou,MSc,MDStevenR.Messe, MD, FAANPushpaNarayanaswami,MBBS,DM,

FAANAlexRae-Grant,MD

Vision Neurology®: Genetics will be the premier peer-reviewed journal in the field of neurogenetics.

Mission Neurology: Genetics will provide neurologistswith outstanding original contributions thatelucidate the role of genetic and epigeneticvariation in diseases and biological traits ofthe central and peripheral nervous systems.

EditorialInquiries

Tel: 612-928-6400Toll-free: 800-957-3182 (US)Fax: [email protected]

StayConnected

facebook.com/NeurologyGenetics

twitter.com/greenjournal

youtube.com/user/NeurologyJournal

http://neurology.org/ng


http://www.facebook.com/NeurologyGenetics

http://www.twitter.com/greenjournal

http://www.youtube.com/user/NeurologyJournal

TABLE OF CONTENTS Volume 4, Number 4, August 2018 Neurology.org/NG

Editorial

e247 These violent repeats have violent extendsJ. Couthouis and A.D. Gitler

Open Access Companion article, e252

Articles

e252 Noncoding repeat expansions for ALS in Japan areassociated with the ATXN8OS geneM. Hirano, M. Samukawa, C. Isono, K. Saigoh, Y. Nakamura, andS. Kusunoki

Open Access Editorial, e247

e246 ASFMR1 splice variant: A predictor of fragileX-associated tremor/ataxia syndromeP. Vittal, S. Pandya, K. Sharp, E. Berry-Kravis, L. Zhou, B. Ouyang,J. Jackson, and D.A. Hall

Open Access

e250 Longitudinal analysis of contrast acuity in FriedreichataxiaA.G. Hamedani, L.A. Hauser, S. Perlman, K. Mathews, G.R. Wilmot,T. Zesiewicz, S.H. Subramony, T. Ashizawa, M.B. Delatycki,A. Brocht, and D.R. Lynch

Open Access

e249 Population genealogy resource showsevidence of familial clustering for AlzheimerdiseaseL.A. Cannon-Albright, S. Dintelman, T. Maness, J. Cerny, A. Thomas,S. Backus, J.M. Farnham, C.C. Teerlink, J. Contreras,J.S.K. Kauwe, and L.J. Meyer

Open Access

e253 Impaired transmissibility of atypical prions fromgenetic CJDG114V

I. Cali, F. Mikhail, K. Qin, C. Gregory, A. Solanki, M.C. Martinez,L. Zhao, B. Appleby, P. Gambetti, E. Norstrom, andJ.A. Mastrianni

Open Access

e255 SCN11A Arg225Cys mutation causes nociceptivepain without detectable peripheral nervepathologyR. Castoro, M. Simmons, V. Ravi, D. Huang, C. Lee, J. Sergent,L. Zhou, and J. Li

Open Access

e256 Expanding the phenotype of de novoSLC25A4-linked mitochondrial disease toinclude mild myopathyM.S. King, K. Thompson, S. Hopton, L. He, E.R.S. Kunji,R.W. Taylor, and X.R. Ortiz-Gonzalez

Open Access

e254 Carey-Fineman-Ziter syndrome with mutationsin the myomaker gene and muscle fiberhypertrophyC. Hedberg-Oldfors, C. Lindberg, and A. Oldfors

Open Access

e251 Axon reflex–mediated vasodilation is reducedin proportion to disease severity inTTR-FAPI. Calero-Romero, M.R. Suter, B. Waeber, F. Feihl, andT. Kuntzer

Open Access

e257 Association study between multiple system atrophyand TREM2 p.R47HK. Ogaki, M.G. Heckman, S. Koga, Y.A. Martens, C. Labbe,O. Lorenzo-Betancor, R.L. Walton, A.I. Soto, E.R. Vargas,S. Fujioka, R.J. Uitti, J.A. van Gerpen, W.P. Cheshire, S.G. Younkin,Z.K. Wszolek, P.A. Low, W. Singer, G. Bu, D.W. Dickson, andO.A. Ross

Open Access


e262 Confirming TDP2 mutation in spinocerebellar ataxiaautosomal recessive 23 (SCAR23)G. Zagnoli-Vieira, F. Bruni, K. Thompson, L. He, S. Walker,A.P.M. de Brouwer, R. Taylor, D. Niyazov, and K.W. Caldecott

Open Access

Clinical/Scientific Notes

e248 Atypical Alexander disease with dystonia, retinopathy,and a brain mass mimicking astrocytomaK. Machol, J. Jankovic, D. Vijayakumar, L.C. Burrage, M. Jain,R.A. Lewis, G.N. Fuller, M. Xu, M. Penas-Prado, M.K. Gule-Monroe,J.A. Rosenfeld, R. Chen, C.M. Eng, Y. Yang, B.H. Lee,P.M. Moretti, Undiagnosed Diseases Network, and S.U. Dhar

Open Access

e258 De novo DNM1L mutation associated withmitochondrial epilepsy syndrome with feversensitivityE. Ladds, A. Whitney, E. Dombi, M. Hofer, G. Anand, V. Harrison,C. Fratter, J. Carver, I.A. Barbosa, M. Simpson, S. Jayawant, andJ. Poulton

Open Access

e259 GLRA1 mutation and long-term follow-up of thefirst hyperekplexia familyM. Paucar, J. Waldthaler, and P. Svenningsson

Open Access

e260 Case of late-onset Sandhoff disease due to a novelmutation in the HEXB geneA.R. Sung, P. Moretti, and A. Shaibani

Open Access

e261 Independent NF1 mutations underlie cafe-au-laitmacule development in a woman with segmentalNF1M.E. Freret, C. Anastasaki, and D.H. Gutmann

Open Access

e263 Novel ELOVL4 mutation associated witherythrokeratodermia and spinocerebellar ataxia(SCA 34)P.R. Bourque, J. Warman-Chardon, D.A. Lelli, L. LaBerge, C. Kirshen,S.H. Bradshaw, T. Hartley, and K.M. Boycott

Open Access

Cover imageAgarose gel electrophoretic analysis for the ATXN8OS gene in patients.Three patients had expansions of the CTA/CTG repeat as indicated,while controls had a normal repeat size. See “Noncoding repeatexpansions for ALS in Japan are associated with the ATXN8OS gene.”See e252

TABLE OF CONTENTS Volume 4, Number 4, August 2018 Neurology.org/NG


EDITORIAL OPEN ACCESS

These violent repeats have violent extendsJulien Couthouis, PhD, and Aaron D. Gitler, PhD

Neurol Genet 2018;4:e247. doi:10.1212/NXG.0000000000000247

Correspondence

Dr. Gitler

[email protected]

Amyotrophic lateral sclerosis (ALS) is a devastating neurodegenerative disease that causes pa-ralysis and death typically within 3 years of onset. The rapidly progressive loss ofmotor neurons inthe brain and spinal cord leads to muscle atrophy, causing weakness, muscle fasciculations, andspasticity. This interferes with normal movement, gait, speech, and swallowing. ALS progressesinexorably, causing paralysis and eventually death. There is currently no cure for ALS and only2 drugs have been approved, but these only offer limited benefit. More effective therapeuticstrategies are desperately needed.

Most ALS cases are sporadic, but around 10% have a family history. Although these familialcases are rarer than the sporadic ones, they have played an outsize role in contributing to insightinto the causes and mechanisms of ALS. This is because these familial cases can be analyzed bygenetic approaches to map and sequence genes that cause ALS. The first gene associated withfamilial ALS was SOD1.1 SOD1 mutations have been found in approximately 20% of familialcases, which equals approximately 2% of all ALS cases. It is believed that the ALS-causingmutations in SOD1 lead to a toxic gain of function of the protein. Clinical trials in humanpatients with ALS are currently under way to test the efficacy of an SOD1-lowering therapy,using antisense oligonucleotides.2 Although SOD1mutations in ALS are relatively rare, the pathfrom SOD1 gene discovery to targeted therapy illustrates the power of genetics in illuminatingthe causes of ALS, paving the path for the development of novel therapies. Owing to the explosionin new gene sequencing technologies, the number of ALS genes has been increasingexponentially1—each new gene offering the hope for a better understanding of ALS mechanisms.The identification of causative genes for ALS will also improve early diagnosis, which will un-doubtedly give therapies a better shot at success.

One of these newly discovered genes is now the most common genetic cause of ALS. Dis-covered in 2011, mutations in the C9ORF72 gene cause;40% of familial ALS cases and;10%of sporadic cases.1 The ALS-causing mutation in C9ORF72 is a hexanucleotide repeat ex-pansion located in a noncoding portion of the gene. Normally, the gene harbors only a handfulof GGGGCC repeats, but in individuals with C9ORF72 mutations, the number of repeatsincreases to hundreds or even thousands. Intense research efforts are under way to define howthese violent nucleotide repeat expansions in C9ORF72 cause ALS and to devise therapeuticstrategies. Of interest, the repeat expansion inC9ORF72 can be traced back to a single foundingevent, probably in Northern Europe.3 Despite being common in Europe and North America, itis extremely rare in Asian ALS populations.4 The paucity of C9ORF72 mutations in Asia raisesthe question of whether repeat expansions in other genes could bemore common causes of ALSin this population.

Beyond C9ORF72, repeat expansions in at least 2 other genes, NOP56 and ATXN2, have beenassociated with motor neuron disease, including ALS.5,6 Noncoding GGCCTG-repeatexpansions in the NOP56 gene cause spinocerebellar ataxia type 36, which can present withprominent motor neuron degeneration. Long trinucleotide CAG-repeat expansions in theATXN2 gene cause spinocerebellar ataxia type 2 (SCA2). Of interest, intermediate-length CAGrepeat expansions in the same gene are associated with ALS.6 Therapeutic strategies to targetATXN2 are being pursued for SCA2 and ALS.7,8

From the Department of Genetics, Stanford University School of Medicine, CA.

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 atNeurology.org/NG.

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 downloadingand sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

RELATED ARTICLE

Noncoding repeatexpansions for ALS inJapan are associated withthe ATXN8OS gene

Page e252

Copyright © 2018 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1

http://dx.doi.org/10.1212/NXG.0000000000000247


http://ng.neurology.org/lookup/doi/10.1212/NXG.0000000000000247

http://creativecommons.org/licenses/by-nc-nd/4.0/

In this issue of Neurology® Genetics, Hirano et al.9 analyzedthe length of the repeat expansion in 3 genes, ATXN8OS,C9ORF72 and NOP56, in 103 Japanese patients with ALS. Inthis cohort, 3 patients harbored a pathologic expansion in theATXN8OS gene. Although healthy controls harbored be-tween 15 and 50 CTA/CTG repeats, patients with ALS hadmore than 80 repeats. Since the authors found ATXN8OSrepeat expansions in 3% of patients with ALS, it suggests thatsuch mutations may be relatively common in Japan.

Nucleotide repeat expansions in ATXN8OS have been pre-viously connected with another neurodegenerative disease,spinocerebellar ataxia type 8 (SCA8). SCA8 is a slowly pro-gressive ataxia with disease onset from age 1 to the early 70sbut typically with an adult onset.10 As opposed to the rapidprogression observed in ALS, SCA8 progression is slow andspans decades, regardless of the initial age at onset; lifespan istypically not shortened. Initial symptoms are scanning dys-arthria with a characteristic drawn-out slowness of speech andgait instability. Various atypical clinical features have also beenidentified in patients carrying an ATXN8OS expansion, suchas parkinsonism, multiple-system atrophy, severe childhoodonset, oromandibular dystonia, dysphagia, respiratory muscleweakness, or seizure-like episodes.

The next steps are to extend the analysis of ATXN8OS ina larger Japanese ALS population to confirm these excitingfindings and to also test other Asian ALS patient populations(e.g., China and Korea). It will also be important to carefullyre-evaluate SCA8 cases and pedigrees to define the prevalenceof symptoms warranting a diagnosis of ALS. Looking to themechanism, future studies should focus on how the ATX-N8OS repeat expansions cause pathologies and how suchpathologies contribute to motor neuron degeneration.

The work of Hirano et al. is especially interesting in a clinicaldiagnostic standpoint. We now see a significant genetic overlapbetween neurodegenerative diseases such as spinocerebellar

ataxia, hereditary spastic paraplegia, ALS, and frontotemporaldementia. It is still unclear whether the clinical manifestationsof these mutations result from different pathologic mecha-nisms, different affected cell types, or different genetic back-grounds harboring a diverse set of genetic modifier genes. Itmay be beneficial to start developing and implementing broadmultigene test panels for neurodegenerative and neuromus-cular diseases, as is standard practice in cancer screening.

Study fundingNo targeted funding reported.

DisclosureJ. Couthouis has received research support from Target ALS.A.D. Gitler has been a consultant for Aquinnah Pharma-ceuticals, Prevail Therapeutics, and Third Rock Ventures andhas received research support from the NIH. Full disclosureform information provided by the authors is available with thefull text of this article at Neurology.org/NG.

References1. Brown RH Jr, Al-Chalabi A. Amyotrophic lateral sclerosis. N Engl J Med 2017;377:

1602.2. Miller TM, Pestronk A, David W, et al. An antisense oligonucleotide against

SOD1 delivered intrathecally for patients with SOD1 familial amyotrophic lateralsclerosis: a phase 1, randomised, first-in-man study. Lancet Neurol 2013;12:435–442.

3. Smith BN, Newhouse S, Shatunov A, et al. The C9ORF72 expansion mutation isa common cause of ALS+/-FTD in Europe and has a single founder. Eur J HumGenet2013;21:102–108.

4. Chen Y, Lin Z, Chen X, et al. Large C9orf72 repeat expansions are seen in Chinesepatients with sporadic amyotrophic lateral sclerosis. Neurobiol Aging 2016;38:217.e215–217.e222.

5. Miyazaki K, Yamashita T, Morimoto N, et al. Early and selective reduction of NOP56(Asidan) and RNA processing proteins in the motor neuron of ALS model mice.Neurol Res 2013;35:744–754.

6. Elden AC, Kim HJ, Hart MP, et al. Ataxin-2 intermediate-length polyglutamineexpansions are associated with increased risk for ALS. Nature 2010;466:1069–1075.

7. Becker LA, Huang B, Bieri G, et al. Therapeutic reduction of ataxin-2 extends lifespanand reduces pathology in TDP-43 mice. Nature 2017;544:367–371.

8. Scoles DR, Meera P, Schneider MD, et al. Antisense oligonucleotide therapy forspinocerebellar ataxia type 2. Nature 2017;544:362–366.

9. Hirano M. Noncoding repeat expansions for ALS in Japan are associated with theATXN8OS gene. Neurol Genet 2018;4:e252.

10. Ikeda Y, Shizuka M, Watanabe M, Okamoto K, Shoji M. Molecular and clinicalanalyses of spinocerebellar ataxia type 8 in Japan. Neurology 2000;54:950–955.

2 Neurology: Genetics | Volume 4, Number 4 | August 2018 Neurology.org/NG



ARTICLE OPEN ACCESS

Noncoding repeat expansions for ALS in Japanare associated with the ATXN8OS geneMakito Hirano, MD, PhD, Makoto Samukawa, MD, PhD, Chiharu Isono, BA, Kazumasa Saigoh, MD, PhD,

Yusaku Nakamura, MD, PhD, and Susumu Kusunoki, MD, PhD


Correspondence

Dr. Hirano

[email protected]

AbstractObjectiveTo assess the contribution of noncoding repeat expansions in Japanese patients with amyo-trophic lateral sclerosis (ALS).

MethodsSporadic ALS in Western countries is frequently associated with noncoding repeat expansionsin the C9ORF72 gene. Spinocerebellar ataxia type 8 (SCA8) is another noncoding repeatdisease caused by expanded CTA/CTG repeats in the ATXN8OS gene. Although the in-volvement of upper and lower motor neurons in SCA8 has been reported, a positive associationbetween SCA8 and ALS remains unestablished. Spinocerebellar ataxia type 36 is a recentlyidentified disease caused by noncoding repeat expansions in the NOP56 gene and is charac-terized by motor neuron involvement. We collected blood samples from 102 Japanese patientswith sporadic ALS and analyzed the ATXN8OS gene by the PCR–Sanger sequencing methodand the C9ORF72 and NOP56 genes by repeat-primed PCR assay.

ResultsThree patients with ALS (3%) had mutations in the ATXN8OS gene, whereas no patient hada mutation in the C9ORF72 or NOP56 gene. The mutation-positive patients were clinicallycharacterized by neck weakness or bulbar-predominant symptoms. None of our patients hadapparent cerebellar atrophy on MRI, but 2 had nonsymptomatic abnormalities in the whitematter or putamen.

ConclusionsOur finding reveals the importance of noncoding repeat expansions in Japanese patients withALS and extends the clinical phenotype of SCA8. Three percent seems small but is stillrelatively large for Japan, considering that the most commonly mutated genes, including theSOD1 and SQSTM1 genes, only account for 2%–3% of sporadic patients each.

From the Department of Neurology (M.H., M.S., K.S., and S.K.), Kindai University Faculty of Medicine, Osakasayama, Japan; and Department of Neurology (M.H., C.I., and Y.N.), KindaiUniversity Sakai Hospital, Japan.


The Article Processing Charge was funded by the authors.







Amyotrophic lateral sclerosis (ALS)mostly occurs sporadically.However, more than 10% of sporadic patients examined inJapan and in Western countries had mutations in various genescausative for familial ALS.1–3 Noncoding six-base (GGGGCC)repeat expansions in the C9ORF72 gene are causative forsporadic ALS and are frequent in Western countries (4%–21%of all sporadic ALS).2,3 By contrast, only small numbers ofsporadic patients in Japan had mutations in the C9ORF72 gene(<0.5%).4 Thus, the significance of noncoding repeat expan-sions in ALS has not been established in Japan.

Spinocerebellar ataxia type 8 (SCA8) is an autosomaldominant neurodegenerative disease, caused by noncodingCTA/CTG repeat expansions in the ATXN8OS (ATAXIN 8OPPOSITE STRAND) gene.5,6 The pathogenicity of theexpanded allele has been proven using a transgenic mousemodel.7 Many patients had pure cerebellar ataxia, whereassome had parkinsonism.5,6,8 Several studies have suggestedthe involvement of upper and lower motor neurons in SCA8,but a positive association between SCA8 and ALS remainsunestablished.9,10

Spinocerebellar ataxia type 36 (SCA36) is a recently identifieddisease, caused by noncoding six-base (GGCCTG) repeatexpansions in the NOP56 gene.11 Motor neuron involvementbecomes apparent during the disease course.

Identification of repeat abnormalities by current exome orwhole-genome analyses in ALS remains challenging, although effortsare underway.12 In this report, we analyzed the ATXN8OS,C9ORF72, and NOP56 genes in 102 Japanese patients withsporadic ALS and in 10 patients with mutations in various genescausative for ALS.

MethodsGenetic testingAll patients and controls were Japanese and were enrolled inthis study from the Kinki region in Japan between 2005 and2017. Blood samples were collected from 102 Japanesepatients with sporadic ALS who had no mutations in theSOD1, FUS, TARDBP, SQSTM1, VCP, OPTN, UBQLN2,ATXN1, ATXN2, SMN1, or AR genes.2,11,13,14 DNA wasextracted with a DNA extraction kit (Qiagen Inc, German-town, MD), and the region containing the CTA/CTG repeatof the ATXN8OS gene was amplified using PCR, as describedpreviously.5,15 The amplified products were purified with gelelectrophoresis and subjected to Sanger sequencing. We an-alyzed the C9ORF72 and NOP56 genes by repeat-primedPCR assay, as described previously.3,11 The normal number of

CTA/CTG repeats in the ATXN8OS gene ranges from 15 to50, whereas repeats of length 80 or more are pathogenic. Inseveral reports, expansions of more than 50 CTA/CTGrepeats, including intermediate expansions, were stated tocause ataxia at some point in the life; however, there were noclinical details.16,17 We also examined 10 patients who hadeither sporadic or familial ALS with mutations in the SOD1(n = 3, 1 man and 2 women), TARDBP (n = 1, 1 man), FUS(n = 1, 1 woman), VCP (n = 2, 1 man and 1 woman), orSQSTM1 (n = 3, 1 man and 2 women) genes.1,13,14 Whenmutations in the ATXN8OS, C9ORF72, or NOP56 gene weredetected, the ERBB4 and COQ2 genes and genes for SCA3,SCA6, SCA7, SCA12, SCA17, SCA31, and DRPLA were ad-ditionally examined.15,18 The patients had a diagnosis ofclinically definite, probable, or possible ALS, as defined in therevised El Escorial diagnostic criteria. Patients with sporadicALS included 68 men and 34 women with an age of 65 ± 12years (mean ± SD). The initial symptoms occurred in the limbmuscles in 76 patients, in the bulbar muscles in 20, in the neckmuscles in 3, in the respiratory muscles in 2, and in thefacial muscles in 1. A control group consisting of 200 appar-ently healthy controls (116men and 84women;mean age ± SD,71 ± 7 years) was also studied. We performed haplotypeanalyses in patients with ATXN8OS mutations using severalmicrosatellite markers, including YI18, D13S1296, JJ9, JJ10,and JJ12 surrounding the CTA/CTG repeats with distancesof 284, 57, 17, 20, and 52 kb, respectively.16 DNA from rel-atives of patients with ALS was not available. We additionallyexamined 2 ataxic sisters with SCA8 (aged 36 and 39 years) tocompare the haplotype data with those for ALS.

Standard protocol approvals, registrations,and patient consentsThis study was approved by our institutional review board. Allpatients provided written informed consent.

ResultsResults of genetic testing of our patientsGenetic testing revealed that 3 patients with ALS hadmutations in the ATXN8OS gene (figure 1), but none in theC9ORF72 gene or in the NOP56 gene. The clinical in-formation of our 3 patients with ALS and 1 reported Koreanpatient with an ALS-like phenotype is summarized in table 1,and the detailed clinical history of our patients is described asfollows. None of the controls had mutations in the ATX-N8OS, C9ORF72, or NOP56 genes. Controls had 26 ± 4repeats (mean ± SD), ranging from 18 to 32, in the ATX-N8OS gene. Double mutations in different genes werenot found in our cohort. Patient 1 had 223 repeats

GlossaryALS = amyotrophic lateral sclerosis; DTR = deep tendon reflex; PEG = percutaneous endoscopic gastrostomy; SCA =spinocerebellar ataxia.



(CTA11CTG212), and patient 2 had 170 repeats(CTA9CTG161). Patient 3 had 91 CTA/CTG repeats withan interruption (CTA19CTG3CTA1CTG68). Interruptionby CTA in a sequence of CTG repeats has previously beenreported, but its significance is unknown.17 The haplotypeanalyses identified apparently different haplotypes in thepatients with ALS, although their haplotypes were similar inpart. By contrast, ataxic sisters with SCA8 had the samehaplotype except for the numbers of CTG repeats in theATXN8OS gene (table 2). These results suggested that the 3patients with ALS seemed unrelated.

Clinical and imaging information for the 3patients with ATXN8OS-related ALS and 1patient with an ALS-like phenotypeThe results of MRI are summarized in figure 2, A-C. Thedetailed clinical histories of our patients are described below.

Patients 1A 56-year-old woman, with 223 repeats, who had dysarthriasince the age of 45 years visited a local hospital. She had nofamily history or previous history of neurologic disease. Thesymptoms progressed over 1 year. She visited the local

Table 1 Patients with ALS or an ALS-like phenotype associated with ATXN8OS mutations

Patient Patient 1 Patient 2 Patient 3 J Clin Neurol 2013;9:274

Sex F M F M

Age at examination (y) 56 68 76 65

Age at onset (y) 45 68 66 60

Age at PEG (y) 47 68 68 ND

Age at ventilator dependence (y) 48 68 (nasal) 71 (only night) ND

Initial symptoms Dysarthria Neck weakness Head drop Dysarthria

Rapid progression of dysphagia + + + +

Rapid progression of respiratory failure + + − +

Dementia − − − ND

Ataxia − − − ? (Unsteadiness)

Parkinsonism − − − ND

Family history − − − + (ataxia)

ATXN8OS repeat no. 223 170 91 86

Repeat configuration CTA11CTG212 CTA9CTG161 CTA19CTG3CTA1CTG68 ND

MRI abnormality T2-high in the white matter Not apparent T2-high in the putaminal rim Mild cerebellar atrophy

Note — Death at 68 — Vertical nystagmus

Abbreviation: ND, not described; PEG = Percutaneous Endoscopic Gastrostomy.

Table 2 Result of haplotype analyses in patients with mutations in the ATXN8OS gene

Distance from CTA/CTG repeat Repeat motif Pt1 Pt2 Pt3 Ataxia1 Ataxia2

YI18 (AY561223) 284 k GATA 12/13 10/10 11/11 10/11 10/11

D13S1296 (Z51394) 57 k CA 23/26 23/23 26/27 25/26 25/26

ATXN8OS 0 CTA 11/10 9/7 19/9 7/9 7/9

ATXN8OS 0 CTG 212/17 161/20 72/15 158/15 144/15

JJ9 (AY561232) 17 k GT 12/12 13/13 12/13 12/13 12/13

JJ10 (AY561233) 20 k CT 8/9 9/9 8/10 9/10 9/10

JJ12 (AY561234) 52 k GT 14/17 17/17 14/18 16/17 16/17

Abbreviations: Ataxia1 and Ataxia2 = sisters with cerebellar ataxia; Pt 1-3 = patients with amyotrophic lateral sclerosis.

Neurology.org/NG Neurology: Genetics | Volume 4, Number 4 | August 2018 3


hospital again at the age of 46 years, with atrophy of thetongue, an increased jaw jerk reflex, increased deep tendonreflexes (DTRs) in all 4 limbs, and positive plantar responses.She was then referred to our hospital. A needle electromyo-gram revealed neurogenic changes, with polyphasic or high-amplitude neuromuscular units with long durations in thetongue, left deltoid, and left tibialis anterior muscles. Duringthe following year, at the age of 47 years, dysarthria, dys-phagia, and neck flexor weakness became severe, but mildweakness of the limbs was seen only in the left upper ex-tremity. Fasciculation was observed in the tongue and all 4limbs. She was unable to perform a spirometry because ofweak lip seal. She underwent percutaneous endoscopic gas-trostomy (PEG) at the age of 47 years. She then showedprogressive respiratory failure and weakness and atrophy in all4 limbs. She underwent tracheostomy with mechanical ven-tilation at the age of 48 years. Her eyeball movement was notlimited and without nystagmus. At the age of 52 years, she wascompletely bedridden and moved only her fingers with verylimited ranges, communicating with a computer-based tool.Cerebellar ataxia, parkinsonism, sensory disturbance, or au-tonomic failure was not apparent during the disease course.The clinical diagnosis, according to the El Escorial diagnosticcriteria, fulfilled the criteria for definite ALS.

Patient 2A 68-year-old man with 170 repeats noticed weakness of theneck. Twomonths later, he visited our hospital at the age of 68years. He had no previous or family history of neurologicdiseases. Weight loss (10 kg in 2 months) was reported. Thefirst neurologic examination revealed very mild weakness ofneck flexion and flexion of the bilateral proximal upper limbs.No nystagmus was noted. A spirogram revealed that % vitalcapacity was 81.9%. A needle electromyogram revealed

neurogenic changes in the cranial nerve territory and upperand lower limbs, with fibrillation in the right tibialis anteriormuscle and fasciculation in the left trapezius and right deltoidmuscle. One month later, dysphagia, dysarthria, and re-spiratory failure became apparent and progressed very rapidly.Two months later, atrophy of the face and limb muscles be-came apparent. DTRs were normal in the atrophied upperlimbs, but increased in the lower limbs, with bilateral extensorplanter response. The clinical diagnosis, according to the ElEscorial diagnostic criteria, fulfilled the criteria for probableALS. Videofluorography showed moderate involvement oforal and pharyngeal phases, with penetration into the larynx.The patient then received a nasal ventilator and underwentPEG about 4 months after onset. Although weakness in the 4limbs remained mild (manual muscle testing 4/5), dysarthriaand respiratory failure progressed with reduced vital capacity(49.1%). He refused tracheostomy and died of respiratoryfailure at age 68 years, about 5 months after onset. An autopsywas not performed.

Patient 3A 76-year-old woman with 91 repeats noticed head droppingwhen walking at the age of 66 years. Ten months later (at theage of 67 years), she experienced dysphagia and weight loss,dropping from 78 to 50 kg in half a year. She had a history ofcomplete resection of breast cancer at age 67 years. She hadno previous or family history of neurologic diseases. Neu-rologic examinations revealed severe dysphagia with aspi-ration, dysarthria, facial weakness, and severe tongueatrophy. No nystagmus was noted. Mild weakness was ob-served in the neck and the right upper limb. DTRs werenormal in the bilateral upper limbs and in the right patella,with decreases in the left patella tendon and bilateral Achillestendon reflexes. Bilateral extensor planter responses werepositive. A needle electromyogram revealed neurogenicchanges in all the examined muscles, including the righttrapezius, biceps brachii, the first dorsal interosseous, rectusfemoris, tibialis anterior, and tongue muscles. Fibrillationpotentials were present in the tibialis anterior and the firstdorsal interosseous. At the age of 68 years, she began toreceive noninvasive positive pressure ventilation because ofreduced vital capacity (66.2%) and underwent PEG for tubefeeding. Cerebellar ataxia, parkinsonism, sensory distur-bance, and autonomic failure were not apparent. MRIrevealed no apparent cerebral atrophy, but with faint T2-hyperintense signals in the putaminal rim (not shown).Weakness of the upper limbs progressed slowly. She showedright vocal cord paralysis at the age of 69 years and then leftparalysis at age 70 years, while the airway remained open.She had aspiration pneumonia and underwent tracheostomyat the age of 71 years, 5 years after onset, but she becameindependent of mechanical ventilation during the day afterrecovery from pneumonia. She used a ventilator only duringthe night for safety. The diagnostic criteria fulfilled the cri-teria for possible ALS, but with atypically slow progression.MRI at the age of 75 years is shown in figure 2C. She livedalone, dealt with tube feeding and nighttime ventilation by

Figure 1 Agarose gel electrophoretic analysis for the ATX-N8OS gene in patients (Pt) 1–3

Three patients had expansions of the CTA/CTG repeat as indicated, whereascontrols (C) had a normal repeat size. M, 100 base pair (bp) marker.



herself, and walked independently at the age of 76 years,more than 10 years after onset.

DiscussionThis study found that 3 sporadic patients who fulfilled thediagnosis of ALS had mutations in the ATXN8OS gene.Haplotype analyses suggested that the patients seemed un-related. By contrast, no patient had a mutation in theC9ORF72 or NOP56 gene, which was consistent with theresults of other studies conducted in Japan.11,19 The ATX-N8OS gene has been viewed as causative for pure cerebellarataxia and parkinsonian disorders, both of which do not al-ways affect motor neurons as in ALS. However, a systematicreview reported that about half of patients with SCA8 hadhyperreflexia,20 an upper motor neuron sign. The phenotypeof the reported patient from Korea mimicked ALS with

involvement of both upper and lower motor neurons, al-though several atypical findings, such as mild atrophy of thecerebellum on MRI, vertical nystagmus, unsteady gait, anda family history of SCA8, precluded a diagnosis of ALS.9 Amore evident involvement of lower motor neurons in SCA8was described in an autopsied patient with atypical SCA8 andsubsequent motor neuron disease. Neuronal loss and gliosiswere found in the cranial motor nucleus with basophilicinclusions immunoreactive for TDP43, a protein causative offamilial ALS.10 These findings suggest that SCA8 occasionallyaffects upper and lower motor neurons, a prerequisite fora diagnosis of ALS.

We found that the predominant clinical symptoms of patientswith ATXN8OS-related ALS were neck weakness and bulbarsigns including dysarthria and dysphagia. Neck weakness wasa rare initial symptom in our cohort (only 3/102) and in

Figure 2 Imaging results for patients who had ALS with mutations in the ATXN8OS gene

(A.a–A.d) Results of MRI (Philips 1.5T, TE/TR = 100/4200 for axial T2-weighted image) of patient 1 (Pt 1) at the age of 47 years (2 years after onset). Atrophy of thecerebellum (A.a), pons (A.b), hippocampus (A.c, arrows), or frontal lobes (A.d) is not obvious; however, ischemic changes (A.d, arrows) in the white matter can beobserved. (B.a–B.d) Results of MRI (SIEMENS Symphony 1.5T, TE/TR = 83/3400 for axial T2-weighted image and TE/TR = 12/450 for coronal T1-weighted image) ofpatient 2 (Pt 2) at the age of 68 years (2 months after onset). Because this patient was not able to complete the MRI examination, the sequences, angles, andresolutions were limited. Atrophy of the cerebellum (B.a, B.b), pons (B.c), hippocampus (B.c, arrows), or frontal lobes (B.d) is not obvious. (C.a–C.d) Results of MRI(SIEMENS Symphony 1.5T, TE/TR = 89/4200 for axial T2-weighted image) of patient 3 (Pt 3) at the age of 75 years (9 years after onset). Atrophy of the cerebellum(C.a), pons (C.b), hippocampus (C.c, arrows), or frontal lobes (C.d) is not obvious; however, high intensities were observed in the putamen (arrowhead).



a reported study (2%).21 Of interest, neck weakness duringthe disease process was a sign closely associated with bulbarsymptoms.21 Bulbar-onset ALS, or ALS with bulbar signwithin 1 year of onset, has been associated with poor prog-nosis.4 As shown in table 1, 3 of the 4 patients indeed showedrapid progression. However, we found that 1 patient witha relatively small repeat size had severe and progressive bulbarsymptoms at the beginning but remained ambulant and in-dependent of mechanical ventilation during the day, morethan 10 years after onset. Thus, clinical severity may vary andmight be associated with repeat sizes, which should be con-firmed by future studies.

Two of our patients had apparently nonsymptomatic MRIabnormalities: patient 1 had T2-hyperintense lesions in thecerebral white matter, and patient 3 had T2-hyperintenselesions in the putaminal rim. The white matter lesions ob-served in patient 1 at the age of 47 years were a findingrelatively rare in individuals before the age of 50 years.However, a reported 22-year-old patient who had SCA8with 102 CTA/CTG repeats had T2-hyperintense lesionsin the periventricular white matter.22 Of interest, whitematter lesions have been reported in some patients withother noncoding repeat diseases such as C9ORF72-relatedALS23 and myotonic dystrophy type 1.24 These mightsuggest the importance of noncoding repeat expansions inglial degeneration. The T2-hyperintense lesions in theputaminal rim in patient 3 (without parkinsonism) wasa finding often seen in multiple system atrophy of parkin-sonian type, but also found in some control individuals.25

Such an MRI finding has not been described in SCA8.Although additional accumulation of patients will beneeded to draw firm conclusions, the MRI abnormalitiesmight be a clue to the associations of the ATXN8OS gene,as with nonsymptomatic atrophy in the hippocampus inVCP-related ALS.14,26

The observed finding that mutations in a single gene areassociated with multiple phenotypes, including ALS and spi-nocerebellar ataxia, has been described in several other genes.In the ATXN1 gene for spinocerebellar ataxia type 1 or theATXN2 gene for spinocerebellar ataxia type 2 (SCA2), in-termediate triplet repeat expansions or sometimes fullexpansions encoding polyglutamine proteins have been foundin patients with ALS.27–29 In addition, 2 independent reportsdescribed that ataxia and ALS phenotypes were present withinthe same family with ATXN2 mutations.29,30 The molecularmechanisms underlying ATXN1-related or ATXN2-relateddisorders may include altered clearance of cytosolic misfoldedproteins, a mechanism shared with that for general ALS.27,31

Very recent studies demonstrated that suppression of ATXN2expression by antisense oligonucleotides improved motorfunctions in SCA2 animal models32 and prolonged the life-span of TDP43-related ALSmodels.33 These findings supportthe idea that 2 different phenotypes may be at least partlyrelated to a similar pathomechanism, which can be a thera-peutic target.

Because noncoding repeat expansions may have some com-mon pathomechanisms, including formation of RNA foci andrepeat-associated non-ATG translation,34 a therapeutic ap-proach to C9ORF72-related ALS may also be applicable toATXN8OS-related ALS. In C9ORF72-related ALS, suppres-sion of abnormal transcription by antisense oligonucleotidesis an ongoing clinical project.35 A similar method was recentlyreported in SCA36.36 In another study, suppression of toxicityin an abnormal ATXN8OS transcript by a certain protein invivo exerted a therapeutic effect.37 Although the ATXN8OSgene, reported to have bidirectional transcripts, may havea more complex pathomechanism,7 suppression of at least 1pathologic pathway might help slow the disease process.

In this study, we found that 3 patients with ALS hadmutationsin the ATXN8OS gene. The coincidental occurrence of SCA8and ALS is possible because mutations in the aforementionedgene have been infrequently found in controls.5,6,38 However,the repeat sizes found in patients with ALS, 91-223 repeats,were not found in the reported control alleles in Japan (n =654).38 The relatively low prevalence of SCA8 (0.7/100,000)and ALS (5/100,000) in Japan suggests that their coincidentalcoexistence is unlikely to happen in the 3 presumably un-related patients. Three percent seems small, but is still rela-tively large for Japan, because the most commonly mutatedgenes, including the SOD1 and SQSTM1 genes, each accountfor 2%–3% of sporadic patients.13,19 Because ATXN8OSmutations were not found in general ALS in US,39 analysis ofcertain subtypes, such as cervical or bulbar-onset types, mightbe needed. Our results may improve the understanding of thepathomechanism of ALS and extend the clinical phenotype ofSCA8.

Author contributionsM. Hirano and K. Saigoh contributed to acquisition of thedata, analysis or interpretation of the data, and drafting orrevising the manuscript for intellectual content. M. Samukawaand C. Isono contributed to design or conceptualization of thestudy, acquisition of the data, and analysis or interpretation ofthe data. Y. Nakamura and S. Kusunoki contributed to designor conceptualization of the study, analysis or interpretation ofthe data, and drafting or revising the manuscript for in-tellectual content.

AcknowledgmentThe authors thank Dr. Abe of Department of Neurology,Okayama University for technical assistance for the analysis ofthe NOP56 gene.

Study fundingThis study was partly supported by Grants-in-Aid for Scien-tific Research from theMinistry of Education, Culture, Sports,Science and Technology of Japan (to MH16K09705).

DisclosureM. Hirano has received funding for speaker honoraria fromDaiichi Sankyo, Eisai, Otsuka Pharmaceutical, Sanofi, and



Shire; has received Grants-in-Aid from Japan’s Ministry ofEducation, Culture, Sports, Science and Technology; and hasreceived research support from Kindai University and theJapan ALS Association. M. Samukawa, C. Isono, K. Saigoh,and Y. Nakamura report no disclosures. S. Kusunoki has re-ceived funding for speaker honoraria from Teijin Pharma,Nihon Pharmaceuticals, Japan Blood Products Organization,Biogen, Novartis, Dainippon Sumitomo Pharma, KyowaKirin, Ono Pharmaceutical, Pfizer, Alexion, and Chugai;serves on the editorial boards of the Journal of Neuro-immunology, Clinical and Experimental Neurology, and theNeurology and Clinical Neuroscience; and has received researchsupport fromNovartis, Dainippon Sumitomo Pharma, Sanofi,Japan Blood Products Organization, Otsuka, Kyowa Kirin,Daiichi Sankyo, Eisai, Takeda, and Nihon Pharmaceuticals, aswell as from Japan’s Ministry of Education, Culture, Sports,Science and Technology and their Agency for Medical Re-search and Development. Full disclosure form informationprovided by the authors is available with the full text of thisarticle at Neurology.org/NG.

Received January 25, 2018. Accepted in final form May 7, 2018.

References1. Okazaki M, Suzuki H, Takahashi Y, et al. Novel mutation in the SOD1 gene in

a patient with early-onset, rapidly progressive amyotrophic lateral sclerosis. NeurolClin Neurosci 2017;5:189–191.

2. DeJesus-Hernandez M, Mackenzie IR, Boeve BF, et al. Expanded GGGGCC hex-anucleotide repeat in noncoding region of C9ORF72 causes chromosome 9p-linkedFTD and ALS. Neuron 2011;72:245–256.

3. Renton AE,Majounie E,Waite A, et al. A hexanucleotide repeat expansion in C9ORF72is the cause of chromosome 9p21-linked ALS-FTD. Neuron 2011;72:257–268.

4. Ishiura H, Takahashi Y, Mitsui J, et al. C9ORF72 repeat expansion in amyotrophiclateral sclerosis in the Kii peninsula of Japan. Arch Neurol 2012;69:1154–1158.

5. Koob MD, Moseley ML, Schut LJ, et al. An untranslated CTG expansion causesa novel form of spinocerebellar ataxia (SCA8). Nat Genet 1999;21:379–384.

6. Day JW, Schut LJ, Moseley ML, Durand AC, Ranum LP. Spinocerebellar ataxia type8: clinical features in a large family. Neurology 2000;55:649–657.

7. Moseley ML, Zu T, Ikeda Y, et al. Bidirectional expression of CUG and CAG ex-pansion transcripts and intranuclear polyglutamine inclusions in spinocerebellarataxia type 8. Nat Genet 2006;38:758–769.

8. Baba Y, Uitti RJ, Farrer MJ, Wszolek ZK. Sporadic SCA8 mutation resembling cor-ticobasal degeneration. Parkinsonism Relat Disord 2005;11:147–150.

9. Kim JS, Son TO, Youn J, Ki CS, Cho JW. Non-ataxic phenotypes of SCA8 mimickingamyotrophic lateral sclerosis and Parkinson disease. J Clin Neurol 2013;9:274–279.

10. Yokoyama T, Ishiyama M, Hasegawa K, Uchihara T, Yagishita S. Novel neuronalcytoplasmic inclusions in a patient carrying SCA8 expansion mutation. Neuropa-thology 2014;34:27–31.

11. Kobayashi H, Abe K, Matsuura T, et al. Expansion of intronic GGCCTG hex-anucleotide repeat in NOP56 causes SCA36, a type of spinocerebellar ataxia ac-companied by motor neuron involvement. Am J Hum Genet 2011;89:121–130.

12. Dolzhenko E, van Vugt J, Shaw RJ, et al. Detection of long repeat expansions fromPCR-free whole-genome sequence data. Genome Res 2017;27:1895–1903.

13. Hirano M, Nakamura Y, Saigoh K, et al. Mutations in the gene encoding p62 inJapanese patients with amyotrophic lateral sclerosis. Neurology 2013;80:458–463.

14. Hirano M, Nakamura Y, Saigoh K, et al. VCP gene analyses in Japanese patients withsporadic amyotrophic lateral sclerosis identify a newmutation. Neurobiol Aging 2015;36:e1601–e1606.

15. Isono C, Hirano M, Sakamoto H, Ueno S, Kusunoki S, Nakamura Y. Differentialprogression of dysphagia in heredity and sporadic ataxias involving multiple systems.Eur Neurol 2015;74:237–242.

16. Ikeda Y, Dalton JC, Moseley ML, et al. Spinocerebellar ataxia type 8: moleculargenetic comparisons and haplotype analysis of 37 families with ataxia. Am J HumGenet 2004;75:3–16.

17. Ayhan F, Ikeda Y, Dalton JC, Day JW, Ranum LPW. Spinocerebellar ataxia type 8. In:Pagon RA, Adam MP, Ardinger HH, editors. GeneReviews®. Seattle, WA: Universityof Washington; 1993.

18. Hirano M, Yamagishi Y, Yanagimoto S, Saigoh K, Nakamura Y, Kusunoki S. Timecourse of radiological imaging and variable interindividual symptoms in amyotrophiclateral sclerosis and frontotemporal dementia associated with p.Arg487His mutationin the VCP gene. Eur Neurol 2017;78:78–83.

19. Nakamura R, Sone J, Atsuta N, et al. Next-generation sequencing of 28 ALS-relatedgenes in a Japanese ALS cohort. Neurobiol Aging 2016;39:219.e1–219.e8.

20. Zeman A, Stone J, Porteous M, Burns E, Barron L, Warner J. Spinocerebellar ataxiatype 8 in Scotland: genetic and clinical features in seven unrelated cases and a reviewof published reports. J Neurol Neurosurg Psychiatry 2004;75:459–465.

21. Konno T, Shiga A, Tsujino A, et al. Japanese amyotrophic lateral sclerosis patientswith GGGGCC hexanucleotide repeat expansion in C9ORF72. J Neurol NeurosurgPsychiatry 2013;84:398–401.

22. Kumar N, Miller GM. White matter hyperintense lesions in genetically proven spi-nocerebellar ataxia 8. Clin Neurol Neurosurg 2008;110:65–68.

23. Oliveira Santos M, Caldeira I, Gromicho M, Pronto-Laborinho A, de Carvalho M.Brain white matter demyelinating lesions and amyotrophic lateral sclerosis ina patient with C9orf72 hexanucleotide repeat expansion. Mult Scler Relat Disord2017;17:1–4.

24. Hashimoto T, Tayama M, Miyazaki M, et al. Neuroimaging study of myotonic dys-trophy: I. Magnetic resonance imaging of the brain. Brain Dev 1995;17:24–27.

25. Tha KK, Terae S, Tsukahara A, et al. Hyperintense putaminal rim at 1.5 T: prevalencein normal subjects and distinguishing features from multiple system atrophy. BMCNeurol 2012;12:39.

26. Hirano M, Matsumura R, Nakamura Y, et al. Unexpectedly mild phenotype in an ataxicfamily with a two-base deletion in the APTX gene. J Neurol Sci 2017;378:75–79.

27. Elden AC, Kim HJ, Hart MP, et al. Ataxin-2 intermediate-length polyglutamineexpansions are associated with increased risk for ALS. Nature 2010;466:1069–1075.

28. Conforti FL, Spataro R, Sproviero W, et al. Ataxin-1 and ataxin-2 intermediate-lengthPolyQ expansions in amyotrophic lateral sclerosis. Neurology 2012;79:2315–2320.

29. Tazen S, Figueroa K, Kwan JY, et al. Amyotrophic lateral sclerosis and spinocerebellarataxia type 2 in a family with full CAG repeat expansions of ATXN2. JAMA Neurol2013;70:1302–1304.

30. Baumer D, East SZ, Tseu B, et al. FTLD-ALS of TDP-43 type and SCA2 ina family with a full ataxin-2 polyglutamine expansion. Acta Neuropathol 2014;128:597–604.

31. Park SH, Kukushkin Y, Gupta R, et al. PolyQ proteins interfere with nuclear degra-dation of cytosolic proteins by sequestering the Sis1p chaperone. Cell 2013;154:134–145.

32. Scoles DR, Meera P, Schneider MD, et al. Antisense oligonucleotide therapy forspinocerebellar ataxia type 2. Nature 2017;544:362–366.

33. Becker LA, Huang B, Bieri G, et al. Therapeutic reduction of ataxin-2 extends lifespanand reduces pathology in TDP-43 mice. Nature 2017;544:367–371.

34. Kearse MG, Todd PK. Repeat-associated non-AUG translation and its impact inneurodegenerative disease. Neurotherapeutics 2014;11:721–731.

35. Sareen D, O’Rourke JG, Meera P, et al. Targeting RNA foci in iPSC-derived motorneurons from ALS patients with a C9ORF72 repeat expansion. Sci Transl Med 2013;5:208ra149.

36. Matsuzono K, Imamura K, Murakami N, et al. Antisense oligonucleotides reduceRNA foci in spinocerebellar ataxia 36 patient iPSCs. Mol Ther Nucleic Acids 2017;8:211–219.

37. Tripathi BK, Surabhi S, Bhaskar PK, Mukherjee A, Mutsuddi M. The RNA bindingKH domain of Spoonbill depletes pathogenic non-coding spinocerebellar ataxia 8transcripts and suppresses neurodegeneration in Drosophila. Biochim Biophys Acta2016;1862:1732–1741.

38. Izumi Y, Maruyama H, Oda M, et al. SCA8 repeat expansion: large CTA/CTG repeatalleles are more common in ataxic patients, including those with SCA6. Am J HumGenet 2003;72:704–709.

39. FigleyMD, Thomas A, Gitler AD. Evaluating noncoding nucleotide repeat expansionsin amyotrophic lateral sclerosis. Neurobiol Aging 2014;35:e931–e934.




ARTICLE OPEN ACCESS

ASFMR1 splice variantA predictor of fragile X-associated tremor/ataxia syndrome

Padmaja Vittal, MD, MS, Shrikant Pandya, BS, Kevin Sharp, MS, Elizabeth Berry-Kravis, MD, PhD, Lili Zhou, MD,

Bichun Ouyang, PhD, Jonathan Jackson, BS, and Deborah A. Hall, MD, PhD


Correspondence

Dr. Hall

[email protected]

AbstractObjectiveTo explore the association of a splice variant of the antisense fragile X mental retardation 1(ASFMR1) gene, loss of fragile X mental retardation 1 (FMR1) AGG interspersions and FMR1CGG repeat size with manifestation, and severity of clinical symptoms of fragile X-associatedtremor/ataxia syndrome (FXTAS).

MethodsPremutation carriers (PMCs) with FXTAS, without FXTAS, and normal controls (NCs) hada neurologic evaluation and collection of skin and blood samples. Expression of ASFMR1transcript/splice variant 2 (ASFMR1-TV2), nonspliced ASFMR1, total ASFMR1, and FMR1messenger RNA were quantified and compared using analysis of variance. Least absoluteshrinkage and selection operator (LASSO) logistic regression and receiver operating charac-teristic analyses were performed.

ResultsPremutation men and women both with and without FXTAS had higher ASFMR1-TV2 levelscompared with NC men and women (n = 135,135, p < 0.0001), and ASFMR1-TV2 had gooddiscriminating power for FXTAS compared with NCs but not for FXTAS from PMC. Afteradjusting for age, loss of AGG, larger CGG repeat size (in men), and elevated ASFMR1-TV2level (in women) were strongly associated with FXTAS compared with NC and PMC(combined).

ConclusionsThis study found elevated levels of ASFMR1-TV2 and loss of AGG interruptions in both menand women with FXTAS. Future studies will be needed to determine whether these variablescan provide useful diagnostic or predictive information.

From the Department of Neurosciences (P.V.), Northwestern Medicine Regional Medical Group, Winfield, IL; the Department of Bioengineering (S.P.), University of Illinois at Chicago;the Department of Pediatrics, Neurological Sciences, Biochemistry (K.S., E.B.-K., L.Z., J.J.), and the Department of Neurological Sciences (B.O., D.A.H.), Rush University Medical Center,Chicago, IL.


The Article Processing Charge was paid for by Neurology: Genetics.







Fragile X-associated tremor/ataxia syndrome (FXTAS) is aninherited neurodegenerative disorder in adults older than 50years because of 55–200 (permutation range) CGG repeatexpansion in the 59 untranslated (UTR) region fragile Xmental retardation 1 (FMR1) gene causing adult-onset ataxia.1

Primary features include progressive action tremor and cer-ebellar ataxia, whereas associated findings include parkin-sonism, neuropathy, cognitive decline, and dysautonomia.2

The CGG repeat in the 59 UTR is unstable and its repeatlength in the normal population ranges from 6 to 55repeats.3,4 More than 200 repeats (full mutation) is associatedwith methylation and transcriptional silencing of FMR1 andconsequent absence or deficiency of fragile X mental re-tardation protein, resulting in fragile X syndrome5 (FXS). Theestimated prevalence of the FMR1 premutation is approxi-mately 1/430 males and 1/209 females.6 The molecularmechanism of FXTAS includes elevated levels of the ex-panded CGG repeat–containing messenger RNA (mRNA)transcript and abnormal repeat-associated non–AUG-initiated (RAN) translation products that lead to cellulartoxicity and neuronal intranuclear inclusions (including DNAdamage response proteins).7–9

Although there is a relationship in a large group analysis be-tween the age at onset/motor severity of FXTAS and CGGrepeat size10 with increasing risk and severity at higher repeatlengths, the CGG repeat size alone does not fully explain risk.Women have a protective normal X allele, and hence, thepropensity for developing FXTAS is different between menand women. In carrier women, a higher activation ratio (AR)(percentage of cells expressing the normal FMR1 gene) maylower the presence and severity of motor signs of FXTAS.10–12

It is probable, however, that there are additional secondarygenes, medical conditions, and environmental factors that af-fect the effects of the premutation expansion and contribute tothe risk of developing symptoms.

A second gene, antisense FMR1 (ASFMR1) overlaps a portionof the FMR1 gene including the CGG repeat sequence, istranscribed in the reverse direction from FMR1 with 5 splicevariants of unclear exact clinical impact.13 One of the variantspreviously described13 ASFMR1 splice variant 2 (ASFMR1transcript variant 2 [ASFMR1-TV2]) was further explored inthis study. Elevated levels of the expanded GCC repeat–containing ASFMR1 transcript are observed in premutationcarriers (PMCs),13 and small unpublished studies have sug-gested that certain splice variants of the ASFMR1 transcript

are associated with the presence of neurologic symptoms inPMC men with FXTAS. Within the CGG repeat element ofthe FMR1 gene, there are AGG trinucleotide interruptions(typically separated by 9–11 CGG repeats), which are knownto disrupt the otherwise pure CGG repeat motif. NormalFMR1 alleles typically possess 2 or 3 AGG interruptions;premutation alleles generally possess 2 or less interruptions,whereas larger premutation alleles tend to have fewer AGGinterruptions. The loss of AGG interruptions is believed toincrease the probability that large normal and small pre-mutation size alleles will be unstable on transmission andexpand to a full mutation allele on transmission from small ormoderate size alleles.14 The rationale for investigating AGGinterspersions is that pilot data from our group show thata lack of AGG interspersions is associated with neurologicsigns. In this study, we hypothesized that elevated specificsplice variants (ASFMR1-TV2) of the ASFMR1 transcript anda lack of AGG interspersions may play a role in FXTAS, andwe attempted to determine the correlation between thesefeatures in development of clinical symptoms in FXTAS.

MethodsParticipantsThis is a case-control study. Adult PMCs and healthy NCswere recruited from the Rush University Fragile X-AssociatedDisorders Program. Recruitment occurred from 3 studies in-vestigating neurologic phenotypes in PMCs and NCs15–17

through clinics, and fragile X community events. The pub-lished criteria for FXTAS were followed to classify as possible,probable, or definite FXTAS participants.3,4

Standard protocol approvals, registrations,and patient consentsThis study received approval from the Institutional ReviewBoard at the Rush University Medical Center. Written in-formed consent was obtained from all patients (or guardiansof patients) participating in the study. The family history ofFXS from participants (but not FXTAS or other neurologicconditions in the family) was obtained.

Neurologic testingEach subject was evaluated with a neurologic examination, theFXTAS Rating Scale (FXTAS-RS),18 and the Total Neu-ropathy Scale19 (modified to exclude those items that requirea nerve conduction study). The FXTAS-RS was developed tomeasure the salient features of the movement disorder seen in

GlossaryAR = activation ratio; ASFMR1 = antisense fragile-X mental retardation 1; ANOVA = analysis of variance; cDNA =complementary DNA; FMR1 = fragile X mental retardation 1; FXS = fragile X syndrome; FXTAS = fragile X-associated tremor/ataxia syndrome; FXTAS-RS = FXTAS Rating Scale; LASSO = least absolute shrinkage and selection operator; mRNA =messenger RNA;NC = normal control; PMC = premutation carrier; ROC = receiver operating characteristic; TV2 = transcriptvariant 2; UTR = untranslated.



FXTAS. The scale was constructed by combining 3 publishedrating scales commonly used to assess tremor (Clinical RatingScale for Tremor),20 ataxia (International Cooperative AtaxiaRating Scale),21 parkinsonism (Unified Parkinson’s DiseaseRating Scale),22 and a tandem gait item (Huntington’s Dis-ease Rating Scale).23

CGG and AGG determinationThe FMR1 PCR assay (Asuragen Inc, Austin, TX) was per-formed on blood samples from participants using a recentlydeveloped highly sensitive method, which detects all FMR1expansions and allows accurate quantification of allele-specificCGG repeat length,24,25 including identification of AGGinterspersions, using primers that flank the repeat and internalrepeat primers. Southern blot assay was performed on sam-ples from premutation women to determine the normalAR.26,27

Fibroblast collectionSkin biopsy was performed under sterile conditions and de-livered to the laboratory immediately for processing. The skinbiopsy was placed in culture media and later diced understerile conditions on a culture plate and then plated in T25flasks in CHANG-Amnio™-Medium (Irvine Scientific, SantaAna, CA), The flasks were incubated at 37°C with 5% carbondioxide. After cells grew out, longer-term cultures weremaintained in Corning™ cellgro™ Cell Culture Media–Dulbecco’s Modified Eagle’s Medium supplemented with10% fetal bovine serum (Gibco) and 1% Gibco® Antibiotic-Antimycotic (containing 10,000 units/mL of penicillin,10,000 μg/mL of streptomycin, and 25 μg/mL of GibcoAmphotericin B). Aliquots of all cell lines were frozen andkept at −80° for long-term use in future studies.

ASFMR1 and FMR1 messenger RNA assaysRNA was purified using the Qiagen RNeasy kit from blood orfibroblast samples. Samples were treated with DNase andscreened for contamination; then, complementary DNA(cDNA) synthesis was performed using the High CapacitycDNA Reverse Transcriptase Kit from Applied Biosystems,Foster City, CA. The defining characteristic of ASFMR1-TV2is the 84-nucleotide deletion near the beginning of the openreading frame and does include the repeat. The 84-nucleotidedeletion excises an AUG associated with a stronger Kozacksequence than the proline reading frame.13 Because the AUGfor the polyproline frame is retained and competition is re-duced, this potentially enhances the polyproline readingframe on ASFMR1-TV2.13 Relative expression of ASFMR1-TV2, nonspliced ASFMR1, total ASFMR1, and total FMR1mRNAwere quantified using the VIIA 7 real-time quantitativePCR system (Applied Biosystems). The results were calcu-lated using a “Delta-Delta CT” (DDCT) algorithm (firstDelta CT [DCT] = [target expression] − [Gus-B expression];second Delta CT [DDCT] = [DCT] − normal control [NC];fold increase of target vs NC = 2−(second delta CT) = 2−DDCT;fold increase = 2−DDCT) to compare relative expressions oftarget mRNA to the expression of the GUS-B control mRNA

and then to the expression of an NC sample (primers andprobes in table e-1, links.lww.com/NXG/A58).

Statistical analysisAnalysis of variance (ANOVA) or the Kruskal-Wallis test(where appropriate) was performed to compare the levels ofASFMR1-TV2, nonspliced ASFMR1, total ASFMR1, and totalFMR1 among NCs, PMCs without FXTAS (PMC), andPMCs with FXTAS (FXTAS). The sample size calculationwas based on ANOVA comparing NCs, PMC withoutFXTAS, and FXTAS. Pairwise comparisons were conductedwith Bonferroni adjustment for multiple comparisons. TheFXTAS-RS scores were compared among different clinicalphenotypes with ANOVA or the Kruskal-Wallis test. Receiveroperating characteristic (ROC) curve analysis was performedto evaluate the performance of ASFMR1-TV2 to discriminatepatients with FXTAS from others (NC and PMC). The fol-lowing guide was used for interpretation: excellent 0.9–1;good 0.8–0.9; fair 0.7–0.8; poor 0.6–0.7; and fail 0.5–0.6. TheSpearman correlation was used to examine the relationshipbetween the CGG repeat size and the ratio of the splicevariant and total ASFMR1mRNA (table e-2, links.lww.com/NXG/A58). Least absolute shrinkage and selection operator(LASSO) logistic regression was performed to examine theassociation of the splice variant, CGG repeat size, AGGinterruptions, nonspliced ASFMR1, total ASFMR1, FMR1,and family history of FXS with FXTAS adjusted for age.LASSO regression was used in this study to select thestrongest predictors. AR was included in the regression forwoman. Data for men and women were analyzed separatelybecause the penetrance of FXTAS is much higher in menthan in women. Missing data were excluded from theanalysis.

ResultsOne hundred thirty-five participants including 56men (mean,SD years; 65.2 ± 12.1 years) and 79 women (55.2 ± 15.5years) were recruited.

MenMen comprised 66% of those participants who met the cri-teria for FXTAS, 21% of PMCs without FXTAS and 45% ofNCs. Twenty-six participants whomet the criteria for FXTAS,with a mean age of 71 ± 8.4 years, had a median CGG repeatlength of 91.5 (interquartile range [IQR] 29) (table 1).Twelve PMC participants without FXTAS, with a mean age of61 ± 11.8 years, had a median CGG repeat length of 78.5(IQR 21). Eighteen NCs had a mean age of 59.7 ± 13.6 yearsand a median CGG repeat length of 30 (IQR 10). Men withFXTAS were older than PMC and NC, and the CGG repeatsize was higher in men with FXTAS than in PMC men andNC men. Nineteen (73.1%) men with FXTAS, 3 PMCs(25%), and 2 (11.8%) NC men had complete loss of AGGinterruptions. FXTAS-RS scores were higher in FXTAS par-ticipants (median [IQR]: 56.5 [42]) than in the other 2


http://links.lww.com/NXG/A58




groups (p < 0.0001). The total (modified) neuropathy scorewas 6 (IQR 5) in those with FXTAS, 4 (IQR 5) in PMCs and0 (IQR 0) in NCs. Eighty-four percent (p = 0.003) of menwith FXTAS had a family history of FXS. Participants whomet the diagnostic criteria for FXTAS included 2 with pos-sible, 10 with probable, and 14 with definite FXTAS. In men,there was a difference in all molecular variables (p < 0.01)among the 3 groups (table 2). Pairwise comparisons of themolecular results were only significant when comparingnormal vs FXTAS and normal vs PMC, but not when com-paring PMC vs FXTAS. ROC curve analysis indicated thatASFMR1-TV2 has a good discriminating power of FXTAS inmen (figure 1, A-F). The area under the curve (AUC) was0.80 (95%CI = 0.68, 0.92). LASSO logistic regression analysiswith the adjustment of age showed that loss of AGG (co-efficient = −0.55) has strongest relationship with FXTAS,followed by a larger CGG repeat size (coefficient = 0.47) andan elevated ASFMR1-TV2 level (coefficient = 0.21).

Coefficients for other independent variables including familyhistory of FXS, nonspliced ASFMR1, total ASFMR1, andFMR1 were shrunken to zero by the LASSO model, whichindicates that these variables did not have a strong relation-ship with FXTAS.

WomenThirteen participants who met the criteria for FXTAS, witha mean age of 69.8 ± 11 years, had an average CGG repeatlength of 92.2 ± 15.1 (table 3). Forty-four PMCs withoutFXTAS, with a mean age of 50.3 ± 14.4 years, had an averageCGG repeat length of 89.5 ± 19.6. Twenty-two NCs, witha mean age of 56.4 ± 14.5 years, had an average CGG repeatlength of 31.6 ± 4.7. Women with FXTAS were older thanPMC and NC, and the CGG repeat size was higher inwomen with FXTAS than in PMC women and NC women.Thirteen (100%) women with FXTAS, 32 PMCs (71%), and2 (9%) healthy women had complete loss of AGG

Table 1 Demographics in men (n = 56)

NC, N = 18 PMC, N = 12 FXTAS, N = 26 p Value

Age, mean (SD) 59.72 (13.64) 61 (11.81) 70.96 (8.38) 0.003

CGG, median (IQR) 30 (10) 78.5 (21) 91.5 (29) 0.15a

AGG, n (%) 0.0001

0 2 (11.76)1 3 (25) 19 (73.08)

≥1 15 (88.24) 9 (75) 7 (26.92)

FXTAS-RS, median (IQR) 5 (8) 7 (9.5) 56.5 (42) <0.0001

Neuropathy, median (IQR) 0 (0)7 4 (5)8 6 (5)9 0.002

Family history, n (%) 8 (44.44) 11 (91.67) 22 (84.62) 0.003

FXTAS diagnosis (%)

Possible 2 (7.69)

Probable 10 (38.46)

Definite 14 (53.85)

Abbreviations: FXTAS = fragile X-associated tremor/ataxia syndrome; FXTAS-RS = fragile X-associated tremor/ataxia syndrome rating scale; IQR = interquartilerange; NC = normal control; PMC = premutation carrier.Superscripts indicate the number of missing values.a Comparison between premutation and FXTAS.

Table 2 Molecular data in men (n = 56)

NC, n = 18 PMC, n = 12 FXTAS, n = 26 p Value

ASFMR1-TV2, median 1.11 (1.42) 11.35 (20.22) 13.63 (15.74) <0.0001

Nonspliced ASFMR1, mean 0.99 (0.59) 2.17 (1.54) 1.77 (1.22) 0.02

Total ASFMR1, mean 1.04 (0.78) 4.23 (3.26) 4.27 (2.91) 0.0002

Total FMR1, median 0.94 (0.44) 1.43 (0.46)1 1.21 (0.51)1 0.02

Abbreviations: ASFMR1-TV2 = antisense fragile-X mental retardation transcript variant 2; FMR1 = fragile X mental retardation 1; FXTAS = fragile X-associatedtremor/ataxia syndrome; NC = normal control; PMC = premutation carrier.



interruptions. FXTAS-RS scores were higher in FXTASparticipants (median [IQR]: 30 [26]) than in the other 2groups (p < 0.0001). The total (modified) Neuropathy Scalescore was 2 (IQR = 2) in those with FXTAS, 1 (IQR = 3) inPMC and 0 (IQR = 1) in NC women (p = 0.01). More than90% of women with FXTAS and PMCs had a family historyof FXS. Participants who met the diagnostic criteria for

FXTAS included 3 with possible, 4 with probable, and 6 withdefinite FXTAS. In women, ASFMR1-TV2 levels, totalASFMR1 levels, and FMR1 levels were different among thegroups (table 4). There were no differences in the non-spliced ASFMR1 levels among the groups. In women, pair-wise comparisons of the molecular results were significantwhen comparing between NC vs FXTAS and NC vs PMC,

Figure 1 ROC curve of ASFMR1-TV2

(A and B) ROC analysis indicated that ASFMR1-TV2 has a good discriminating power of FXTAS when comparing FXTAS with PMC and normal in men (A: AUC =0.80; 95% CI = 0.68, 0.92) and women (B: AUC = 0.80; 95% CI = 0.67, 0.9). (C and D) ROC analysis indicated that ASFMR1-TV2 did not have a good discriminatingpower when comparing FXTASwith PMC in bothmen (C) andwomen (D) (men: AUC = 0.53, 95% CI = 0.32, 0.74; women: AUC = 0.69, 95% CI = 0.53, 0.85). (E andF) ROC analysis indicated that ASFMR1-TV2 has a good discriminating power when comparing PMCwithout FXTAS with normal in bothmen (E) and women (F)(men: AUC = 0.95, 95%CI = 0.86, 1; women: AUC = 0.92, 95%CI = 0.85, 0.99). (A–F) The green line is the diagonal reference line, and the blue line is the empiricalROC curve comparing FXTAS with NCs and premutation carriers (A and B), FXTAS with PMC (C and D), and PMC without FXTAS with controls (E and F) (Thecloser the ROC curve gets to the reference line, the worse the test. A large value of the AUC indicates a good test. An area of 1 means a perfect test, and 0.5represents a worthless test). AUC = area under the curve; ASFMR1-TV2 = antisense fragile-X mental retardation-transcript variant 2; CI = confidence interval;FXTAS = fragile X-associated tremor/ataxia syndrome; NC = normal control; PMC = premutation carrier; ROC = receiver operating characteristic.



but were not when comparing PMC vs FXTAS. ASFMR1-TV2 showed a good discriminating power of FXTAS fromthe ROC curve analysis (figure 2, AUC = 0.8, 95% CI = 0.67,0.9). LASSO logistic regression, adjusting for age, showedthat the loss of AGG (coefficient = −0.5555) has thestrongest relationship with FXTAS, followed by ASFMR1-TV2 (coefficient = 0.4747) and total ASFMR1 (coefficient =0.2121). Other variables including family history of FXS,nonspliced ASFMR1, AR, CGG repeat size, and FMR1 didnot show a strong relationship with FXTAS from the LASSOregression model.

In men, we found a negative correlation between the CGGrepeat size and the ratio of the splice variant and totalASFMR1 mRNA among the NCs (correlation coefficient =−0.62, p = 0.04), but a positive correlation among the PMCs(correlation coefficient = 0.81, p < 0.0001). In women, there

was also a negative correlation among the controls (correla-tion coefficient = −0.74, p = 0.02), but no correlationwas found among the PMCs (CGG was adjusted for AR)(correlation coefficient = 0.08, p = 0.70).

DiscussionThe main purpose of this study was to identify molecularmarkers to help predict which PMC will get FXTAS becauseeven for male carriers, not all PMC will get the disease.14 Theexact clinical impact of ASFMR1 splice variants is unclear;however; it has been postulated that in addition to FMR1,elevated levels of the expanded GCC repeat containing theASFMR1 transcript are observed in PMCs andmay contributeto the variable phenotypes associated with the CGG repeatexpansion.14

Table 3 Demographics in women (n = 79)


Age, mean (SD) 56.36 (14.52) 50.29 (14.40) 69.77 (11.01) 0.0003

CGG, mean (SD) 31.59 (4.68) 89.52 (19.64) 92.15 (15.07) 0.66a

AR, mean (SD) NA 0.501 (0.23) 0.46 (0.31) 0.87a

AGG, n (%) <0.0001

0 2 (9.09) 32 (72.73) 13 (100)

≥1 20 (90.91) 12 (27.27) 0

FXTAS-RS, median (IQR) 3 (4) 7 (8) 30 (26) <0.0001

Neuropathy, median (IQR) 0 (1) 1 (3) 2 (2) 0.01

Family history, n (%) 5 (22.73) 40 (90.91) 12 (92.31) <0.0001

FXTAS diagnosis (%)

Possible 3 (23.08)

Probable 4 (30.77)

Definite 6 (46.15)

Abbreviations: AR = activation ratio; FXTAS = fragile X-associated tremor/ataxia syndrome; FXTAS-RS = fragile X-associated tremor/ataxia syndrome ratingscale; IQR = interquartile range; NC = normal control; PMC = premutation carrier.Superscripts indicate the number of missing values.a Comparison between premutation and FXTAS.

Table 4 Molecular results in women (n = 79)


ASFMR1-TV2, median 0.91 (0.67) 6.92 (10.29) 12.52 (29.37) <0.0001

Nonspliced ASFMR1, mean 0.78 (0.41) 0.93 (0.69) 1.3 (0.96) 0.1

Total ASFMR1, mean 0.93 (0.48) 2.42 (1.87) 3.52 (2.62) 0.0002

Total FMR1, median 0.89 (0.32) 1.05 (0.48) 1.22 (0.43) 0.01

Abbreviations: ASFMR1-TV2 = antisense fragile-X mental retardation transcript variant 2; FMR1 = fragile X mental retardation 1; FXTAS = fragile X-associatedtremor/ataxia syndrome; NC = normal control; PMC = premutation carrier.



This is the first study to evaluate ASFMR1 splice variant levelsin PMCs and FXTAS patients. The reason this study con-centrated on ASFMR1-TV2 (figure 2) was because smallunpublished studies suggested that this splice variant of theASFMR1 transcript might be associated with neurologicsymptoms in FXTAS men. This study reports elevated levelsof ASFMR1-TV2 in FXTAS and a loss of AGG interruptionsin men with FXTAS. In addition, this study also showed thatthe FXTAS-RS and Neuropathy Scale scores were higher inpatients with FXTAS compared with PMCs and NCs. Wewere not able to distinguish between PMCs and FXTAS withpairwise comparisons of ASFMR1-TV2, nonspliced ASFMR1,total ASFMR1, or total FMR1.

Although we found that ASFMR1-TV2 and lack of AGG weremore common in FXTAS, the power to discriminate betweenPMCs and FXTAS was not high enough to use alone di-agnostically or in a predictive manner. This may be becauseour sample sizes were too small overall. Some of the groups,such as the PMC men without FXTAS and the FXTASwomen group did not have many participants. Because thedisease is highly penetrant in men, it was difficult to recruitPMC men without FXTAS. Women with FXTAS were alsodifficult to recruit because there is a lower prevalence inwomen. The published criteria for FXTAS3,4 were followed,and because the women have milder features than men withFXTAS, many of them needed either the middle cerebellarpeduncle sign or brainstem white matter lesions to qualify forprobable FXTAS. Thus, the strict application of the diagnostic

criteria leads to the high number having the MCP sign ratherthan our population of FXTAS women being different fromthe populations that have been published in the past. Alter-natively, the patterns of relationships found may indicate thatregardless of the sample size, elevated RNA products from therepeat-containing gene are a phenomenon more related tojust the presence of a premutation, rather than reflectinga threshold of abnormal RNAs at which symptoms occur.RAN translation has been described in FXTAS as a potentialadditional mechanism of toxicity independent of the pre-sumed RNA toxicity. Measurement of RAN translationproducts could be more closely correlated with amount andon the time of the onset of symptoms. Additional studies maybe warranted to combine molecular and other biomarkers,such as examination, imaging, or laboratory markers, to in-crease the positive predictive value and the clinical utility.

Both AUG-initiated and repeat-associated non-AUG (RAN)translation have been shown to occur from ASFMR1 CCGrepeats.28 Associated transcripts are capable of forming strongsecondary structures in vitro, either RNA hairpins orG-quadruplexes, and are important to the process underlyingRAN translation. It is yet unknown if secondary structureslinked to ASFMR1-TV2 may also be linked to RAN trans-lation and whether toxicity occurs from the expression ofproteins generated from translation of this variant. Pilot workfrom our research group suggests an association between thelack of AGG interspersions in FMR1 PMCs and higher ratesof neurologic signs, which was confirmed in this study. It is

Figure 2 Antisense FMR transcript variant 2

Defining characteristic of ASFMR1-TV2. There is an 84-nt deletion at near the beginning of the ASFMR1 ORF that has an enhanced expression compared witha healthy control. We designed strand specific primers that target the presence and absence of the 84-nt sequence to determine the fold increase relative toGUS expression for ASFMR1-TV2 spliced and nonspliced assays. ASFMR1-TV2 = antisense fragile-X mental retardation transcript variant 2; FMR = fragile Xmentalretardation; nt = nucleotide; ORF = open reading frame.



recognized that the results regarding the lack of AGG inter-ruptions may represent instead a relationship with the longestpure CGG repeat size. The loss of AGG interruptions is seenin PMCs at a much higher rate than NCs because it makesthese repeats more unstable and hence more likely to expandmeiotically and (presumably) somatically in tissues. Similarly,one would presume that PMCs with fewer AGG interruptionswould have larger pure repeats. Then, these larger purerepeats are associated with enhanced transcription of theFMR1 and ASFMR1 loci because of changes in localchromatin.29,30 Additional studies are needed to determinewhich factor is responsible for the association.

It is important to find molecular markers that can serve asscreening tools and increase diagnostic certainty, not only tohelp families of patients at risk of disorder but also to enabledrug discovery, so timely treatment can be implemented.Both PMCs and FXTAS participants had elevated levels ofASFMR1 splice variant 2, and loss of AGG interruptions wasassociated with FXTAS in men. These markers may haveutility if combined with other predictive factors to betterpredict who will get FXTAS. Additional work also needs to bedone to determine the contribution to the mechanism ofCCG-mediated toxicity in PMCs.

Author contributionsP. Vittal: conception of the study, subject recruitment andneurologic evaluation, skin biopsy, and organizing and writingof the first draft of the manuscript. S. Pandya and K. Sharp:performance of RNA assays. J. Jackson: performance of gen-otyping and AGG analysis. L. Zhou: performance of geno-typing and AGG analysis, fibroblast cultures, and supervisionof laboratory assays. B. Ouyang: statistical analysis. E. Berry-Kravis: conception of the study, review and critique of themanuscript, and oversight of laboratory studies. D.A. Hall:conception of the study, subject recruitment, and review andcritique of the manuscript.

AcknowledgmentThe authors thank the American Brain Foundation and theNational Ataxia Foundation for their support of the ClinicalResearch Training Fellowship grant in Ataxia and TracyWaliczek and Evelyn Perez for their assistance during skinbiopsy and blood sample collection. They thank IntegratedDNA Technologies (Coralville, IA) for custom-designedprimers and probes. They also thank the participants of thisstudy.

Study fundingClinical Research Training Fellowship grant in Ataxia spon-sored by the American Brain Foundation and the NationalAtaxia Foundation.

DisclosureP. Vittal has received funding for travel and/or speaker hon-oraria and served on speakers’ bureaus of Teva Pharmaceu-tical Industries Ltd and Adamas Pharmaceuticals Inc. and has

received research support from the American Brain Founda-tion and the National Ataxia Foundation. S. Pandya and K.Sharp report no disclosures. E. Berry-Kravis serves or hasserved on the scientific advisory boards of Vteese, BioMarin,Marinus, Fulcrum, Cydan, GW, and Ovid; holds a patentregarding a method for assay of CCHS-causing polyalaninerepeat expansion for diagnosis; receives or has received pub-lishing royalties for The Fragile X Tremor Ataxia Syndrome(FXTAS); consults or has consulted for Fulcrum andZynerba; and has received research support from Vtesse,Ovid, Neuren, NIH/NICHD, NIH/NCATS, CDC, NIH/NINDS, NIH/NIMH, and the John Merck Fund. L. Zhou, B.Ouyang, and J. Jackson report no disclosures. D.A. Hall servedor has served on the scientific advisory board of the NationalFragile X Foundation; serves or has served on the editorialboard of Neurology Today; consults or has consulted for theNIH; and has received research support from Pfizer, Neuro-crine, AbbVie, NIH, Shapiro Foundation, Anti-Aging Foun-dation, and Parkinson Foundation. Full disclosure forminformation provided by the authors is available with the fulltext of this article at Neurology.org/NG.

Received August 14, 2017. Accepted in final form April 11, 2018.

References1. Hagerman RJ, Leehey M, Heinrichs W, et al. Intention tremor, parkinsonism, and

generalized brain atrophy in male carriers of fragile X. Neurology 2001;57:127–130.

2. Hall DA, O’Keefe JA. Fragile x-associated tremor ataxia syndrome: the expandingclinical picture, pathophysiology, epidemiology, and update on treatment. TremorOther Hyperkinet Mov (N Y) 2012;2.

3. Jacquemont S, Hagerman RJ, Leehey M, et al. Fragile X premutation tremor/ataxiasyndrome: molecular, clinical, and neuroimaging correlates. Am J Hum Genet 2003;72:869–878.

4. Brunberg JA, Jacquemont S, Hagerman RJ, et al. Fragile X premutation carriers:characteristic MR imaging findings of adult male patients with progressive cerebellarand cognitive dysfunction. Am J Neuroradiol 2002;23:1757–1766.

5. Pieretti M, Zhang FP, Fu YH, et al. Absence of expression of the FMR-1 gene in fragileX syndrome. Cell 1991;66:817–822.

6. Tassone F, Iong KP, Tong TH, et al. FMR1 CGG allele size and prevalence ascer-tained through newborn screening in the United States. Genome Med 2012;4:100.

7. Tassone F, Beilina A, Carosi C, et al. Elevated FMR1mRNA in premutation carriers isdue to increased transcription. RNA 2007;13:555–562.

8. Todd PK, Oh SY, Krans A, et al. CGG repeat-associated translation mediates neu-rodegeneration in fragile X tremor ataxia syndrome. Neuron 2013;78:440–455.

9. Robin G, Lopez JR, Espinal GM, et al. Calcium dysregulation and Cdk5-ATMpathway involved in a mouse model of fragile X-associated tremor/ataxia syndrome.Hum Mol Genet 2017;26:2649–2666.

10. Leehey MA, Berry-Kravis E, Goetz CG, et al. FMR1 CGG repeat length predictsmotor dysfunction in FXTAS. Neurology 2008;70:1397–1402.

11. Berry-Kravis E, Potanos K, Weinberg D, et al. Penetrance of fragile X-associatedtremor/ataxia syndrome (FXTAS) in two sisters related to X-inactivation pattern.Ann Neurol 2005;57:144–147.

12. Hall DA, Robertson-Dick EE, O’Keefe JA, et al. Repeat size and X-inactivation in theclinical phenotype of fragile X premutation carrier sisters: a familial case series. NeurolGenet 2016;2:e45. doi: 10.1212/NXG.0000000000000045.

13. Ladd PD, Smith LE, Rabaia NA, et al. An antisense transcript spanning the CGGrepeat region of FMR1 is upregulated in premutation carriers but silenced in fullmutation individuals. Hum Mol Genet 2007;16:3174–3187.

14. Yrigollen CM, Durbin-Johnson B, Gane L, et al. AGG interruptions within the ma-ternal FMR1 gene reduce the risk of offspring with fragile X syndrome. Genet Med2012;14:729–736.

15. Hall D, Todorova-Koteva K, Pandya S, et al. Neurological and endocrine phenotypesof fragile X carrier women. Clin Genet 2016;89:60–67.

16. Berry-Kravis E, Goetz CG, Leehey MA, et al. Neuropathic features in fragile X pre-mutation carriers. Am J Med Genet A 2007;143A:19–26.

17. Jacquemont S, Hagerman RJ, Leehey MA, et al. Penetrance of the fragile X-associatedtremor/ataxia syndrome in a premutation carrier population. JAMA 2004;291:460–469.

18. Leehey M, Goetz C, Berry-Kravis E. Development of the FXTAS rating scale forquantitative motor analysis. Presented at the 10th International Congress of Par-kinson’s Disease and Movement Disorders; October 30, 2006; Kyoto, Japan.





19. Cornblath DR, Chaudhry V, Carter K, et al. Total neuropathy score: validation andreliability study. Neurology 1999;53:1660–1664.

20. Fahn S, Tolosa E, Marin C. Clinical rating scale for tremor. In: Jankovic J, Tolosa E,editors. Parkinson’s Disease and Movement Disorders. Baltimore, Munich: Urban &Schwartzenberg; 1987:225–234.

21. Trouillas P, Takayanagi T,HallettM, et al. International cooperative ataxia rating scale forpharmacological assessment of the cerebellar syndrome: the Ataxia NeuropharmacologyCommittee of the World Federation of Neurology. J Neurol Sci 1997;145:205–211.

22. Fahn S, Elton R; members of the UPDRS Development Committee. Unified Par-kinson’s disease rating scale. In: Fahn S, Marsden C, Calne D, Goldstein M, editors.Recent Development in Parkinson’s Disease. Florham Park, NJ: Macmillan HealthCare Information; 1987:153–164.

23. Siesling S, Zwinderman AH, van Vugt JP, et al. A shortened version of the motor sectionof the Unified Huntington’s Disease Rating Scale. Mov Disord 1997;12:229–234.

24. Chen L, Hadd A, Sah S, et al. An information-rich CGG repeat primed PCR thatdetects the full range of fragile X expanded alleles and minimizes the need for southernblot analysis. J Mol Diagn 2010;12:589–600.

25. Filipovic-Sadic S, Sah S, Chen L, et al. A novel FMR1 PCR method for the routinedetection of low abundance expanded alleles and full mutations in fragile X syndrome.Clin Chem 2010;56:399–408.

26. Godler DE, Tassone F, Loesch DZ, et al. Methylation of novel markers of fragile Xalleles is inversely correlated with FMRP expression and FMR1 activation ratio. HumMol Genet 2010;19:1618–1632.

27. Annemieke JMHV, Pieretti M, Sutcliffe JS, et al. Identification of a gene (FMR-1)containing a CGG repeat coincident with a breakpoint cluster region exhibiting lengthvariation in fragile X syndrome. Cell 1991;65:905–914.

28. Krans A, Kearse M, Todd PK. Repeat-associated non-AUG translation from antisenseCCG repeats in fragile X tremor/ataxia syndrome. Ann Neurol 2016;80:871–881.

29. Todd PK, Oh SY, Krans A, et al. Histone deacetylases suppress CGG repeat-inducedneurodegeneration via transcriptional silencing in models of fragile X tremor ataxiasyndrome. PLoS Genet 2010;6:e1001240.

30. Hoem G, Raske CR, Garcia-Arocena D, et al. CGG-repeat length threshold for FMR1RNA pathogenesis in a cellular model for FXTAS. Hum Mol Genet 2011;20:2161–2170.



ARTICLE OPEN ACCESS

Longitudinal analysis of contrast acuity inFriedreich ataxiaAli G. Hamedani, MD, MHS,* Lauren A. Hauser, MS, Susan Perlman, MD, Katherine Mathews, MD,

George R. Wilmot, MD, PhD, Theresa Zesiewicz, MD, S.H. Subramony, MD, Tetsuo Ashizawa, MD,

Martin B. Delatycki, MD, PhD, Alicia Brocht, MS, and David R. Lynch, MD, PhD


Correspondence

Dr. Hamedani

[email protected]

AbstractObjectiveTo determine the natural history of contrast acuity in Friedreich ataxia.

MethodsIn the Friedreich Ataxia–Clinical Outcome Measures Study, participants (n = 764) underwentbinocular high- and low-contrast visual acuity testing at annual study visits. Mixed-effects linearregression was used to model visual acuity as a function of time, with random intercepts andslopes to account for intraindividual correlation of repeated measurements. A time-varyingcovariate was used to adjust for diabetes, and interaction terms were used to assess for effectmodification by GAA repeat length, disease duration, and other variables.

ResultsAcross a median of 4.4 years of follow-up, visual acuity decreased significantly at 100% contrast(−0.37 letters/y, 95% confidence interval [CI]: −0.52 to −0.21), 2.5% contrast (−0.81 letters/year, 95% CI: −0.99 to −0.65), and 1.25% contrast (−1.12 letters/y, 95% CI: −1.29 to −0.96letters/year). There was a significant interaction between time and GAA repeat length such thatthe rate of decrease in visual acuity was greater for patients with higher GAA repeat lengths at2.5% contrast (p = 0.018) and 1.25% contrast (p = 0.043) but not 100% contrast. There was noeffect modification by age at onset after adjusting for GAA repeat length.

ConclusionsLow-contrast visual acuity decreases linearly over time in Friedreich ataxia, and the rate ofdecrease is greater at higher GAA repeat lengths. Contrast sensitivity has the potential to serveas a biomarker and surrogate outcome in future studies of Friedreich ataxia.

*Conducted statistical analysis (academically affiliated).

From the Department of Neurology (A.G.H., D.R.L.), University of Pennsylvania; Divisions of Neurology and Pediatrics (L.A.H., D.R.L.), Children’s Hospital of Philadelphia, PA;Department of Neurology (S.P.), University of California at Los Angeles; Departments of Neurology and Pediatrics (K.M.), University of Iowa; Department of Neurology (G.R.W.), EmoryUniversity, Atlanta, GA; Department of Neurology (T.Z.), University of South Florida, Tampa Bay; Department of Neurology (S.H.S.), University of Florida, Gainesville; Department ofNeurology (T.A.), Houston Methodist Hospital, TX; Murdoch Children’s Research Institute (M.B.D.), Melbourne, Victoria, Australia; and Department of Neurology (A.B.), University ofRochester, NY.


Supported by a grant from the Friedreich Ataxia Research Alliance.








Friedreich ataxia (FRDA) is the most common inheritedcause of ataxia.1 Optic neuropathy is a well-known but under-recognized clinical manifestation of FRDA.2–5 Although vi-sion loss is initially mild,6–8 it becomes more prominent as thedisease progresses, and there are numerous reports of patientswho experience dramatic subacute vision loss similar to Leberhereditary optic neuropathy.9–11

Low-contrast letter acuity is a well-established clinical corre-late of optic nerve disease across multiple neurologic dis-orders, and in FRDA, it has been found to be associated withpredictors of disease severity such as age and GAA repeatlength; non-neurologic measures of disease severity such asthe presence of diabetes, hearing loss, and cardiomyopathy;and neurologic measures of disease severity such as theFriedreich ataxia rating scale, timed 9-hole peg test (9HPT),and timed 25-foot walk (T25W).12,13 Because contrast acuityis easily and reliably measured, there is growing interest in itsuse as a clinical predictor of disease severity and surrogateoutcome in clinical trials for FRDA, particularly among moreaffected patients.

However, the use of contrast acuity as a biomarker in FRDA islimited by a lack of longitudinal data. To determine the nat-ural history of contrast acuity in these patients, we performeda longitudinal analysis of a large prospective cohort study ofpatients with FRDA. We hypothesized that contrast acuitywould decrease linearly over time and that the rate of decreasewould be greater for patients with longer GAA repeat lengthsand earlier age at onset.

MethodsStandard protocol approvals, registration, andpatient consentsInstitutional review board approval was obtained at eachparticipating site, and written informed consent was obtainedfrom all patients for participation in the study.

Study designThe Friedreich Ataxia–Clinical Outcome Measures Study(FA-COMS) is an ongoing prospective cohort study ofpatients with Friedreich ataxia being followed at one of 12clinical sites since 2001.14–16 Baseline genetic confirmationwas obtained via commercial or research testing, and in-formation regarding age at onset and other medical comor-bidities was collected. At each annual follow-up visit,participants underwent contrast letter acuity testing usingretro-illuminated charts. High-contrast acuity was tested bothmonocularly and binocularly. Low-contrast acuity was tested

binocularly using the 2.5% and 1.25% Low-Contrast SloanLetter Charts at a distance of 2 m (Precision Vision, LaSalle,IL) using a protocol developed to measure visual dysfunctionin MS and Friedreich ataxia,17 and the total number of letterscorrectly identified at each level of contrast was recorded.Visual acuity was measured by trained technicians, and par-ticipants wore their standard refractive correction for distance.

Statistical analysisStatistical analyses were performed using STATA version 12(College Station, TX). Multilevel linear regression modelswere used to model visual acuity at both high and low contrastas a function of time, with a random intercept and slope foreach participant to account for intraindividual correlation ofrepeated measurements. Regression models were adjusted fordiabetes using a time-varying covariate to capture both thosewho had diabetes at baseline and those who were diagnosedwith diabetes during the follow-up period. Modification of theeffect of time on visual acuity by GAA repeat length and age atonset was assessed using interaction terms. The duration offollow-up was highly variable, with many participants con-tributing only a single baseline visit and some contributing asmany as 12 years of follow-up. To assess for potential in-formation bias due to differential follow-up, linear regressionmodels were constructed for the follow-up duration asa function of baseline demographic characteristics.

ResultsAt the time of this analysis, the FA-COMS cohort consisted of764 patients, of whom 50.4% were female. The median age atonset was 11 years (interquartile range [IQR]: 7–17), witha median age at the baseline study visit of 23 years (IQR:15–34). The smaller GAA allele had a median length of666 repeats (IQR: 466–800), and the baseline prevalence ofdiabetes was 5%.

Of the 764 patients in the cohort, 332 participated in a singlestudy visit and thus did not contribute data to the longitudinalanalysis. The remaining 432 patients were followed up fora median of 5 visits over a median follow-up duration of 4.4years (IQR: 2.0–6.3). At baseline, mean binocular visualacuity at 100% contrast was 58.8 letters (IQR: 55–65) ofa maximum of 70, corresponding to a Snellen visual acuityequivalent of 20/20, but 7.6% of patients had 20/40 acuity orworse. This level of acuity generally qualifies as “visually im-paired” and would limit the ability to drive without restric-tions in almost all US states. At 2.5% contrast, the meannumber of letters was 31 letters (95% CI: 25–42), and at1.25%, it was 25 letters (95% CI: 14–33).

GlossaryCI = confidence interval; FA-COMS = Friedreich Ataxia–Clinical OutcomeMeasures Study; FRDA = Friedreich ataxia; IQR =interquartile range.



The mean difference in visual acuity between each follow-upvisit and the baseline study visit at 100%, 2.5%, and 1.25%contrast (figures 1–3), respectively, demonstrated a generallylinear pattern of decrease, particularly at low contrast. Indiabetes-adjustedmultilevel regressionmodels, the average rateof change in vision was −0.37 letters per year at 100% contrast(95% CI: −0.52 to −0.21), −0.81 letters per year at 2.5% con-trast (95% CI: −0.99 to −0.65), and −1.12 letters per year at1.25% contrast (−1.29 to −0.96). These estimates did notchange after adjusting for sex, age at onset, or GAA repeatlength. However, the addition of interaction terms did reveala significant interaction between the GAA repeat length and

time at low contrast but not high contrast. For each additionaltertile of GAA triplet repeat, the number of letters lost per yearincreased by 0.26 letters (95% CI: 0.05–0.47) at 2.5% contrastand 0.21 letters (95%CI: 0.006–0.41) at 1.25% contrast. Therewas no significant interaction between the age at onset andtime. Using these interaction terms, the predicted time to lossof 7 letters of contrast acuity stratified by tertiles of GAA repeatlength and age at onset is shown in tables 1 and 2.

At the final follow-up, mean binocular high-contrast acuitywas 57.2 letters, again corresponding to a Snellen equivalentof 20/20. However, the percentage of patients who saw 20/40

Figure 1 Difference in binocular 100% contrast acuity between baseline and follow-up visit

The sample size at each visit is indicated inparentheses.

Figure 2 Difference in binocular 2.5% contrast acuity between baseline and follow-up visit

The sample size at each visit is indicated in parentheses.



or worse increased to 11.9%. Of the 65 patients who were seenat 9 follow-up visits, 11 (16.9%) saw 20/40 or worse, and ofthe 12 patients who were seen at all 11 follow-up visits, 4(33.3%) saw 20/40 or worse.

Linear regressionmodels were also constructed for the follow-up duration as a function of baseline demographic variables toassess for information bias due to differential follow-up. Inunivariate analyses, individuals with a history of hypertrophiccardiomyopathy at baseline had an average of 0.90 less yearsof follow-up than those without (95% CI: 0.37–1.42), and thefollow-up duration decreased with increasing GAA repeatlength (0.25 less years per quintile of higher GAA repeatlength, 95% CI: 0.07–0.42). Otherwise, there was no associ-ation between age at onset, age at diagnosis, sex, diabetes, orscoliosis and follow-up duration.

DiscussionIn this large prospective cohort study of patients with FRDA,there is a significant essentially linear decrease in low-contrast

acuity over time that is greater in subjects with higher GAArepeat lengths. High-contrast acuity is also affected, but thisbecomes apparent only at longer follow-up intervals. Thisreduction in low-contrast acuity but sparing high-contrastacuity likely reflects subtle optic nerve dysfunction. Contrastsensitivity also relies on higher-order cortical visual in-tegration, and recent studies have found evidence of cognitiveimpairment in FRDA including with visuospatial reasoning, soit is possible that this reduction in low-contrast acuity ismultifactorial in etiology.18 However, cognitive impairment inFRDA is extremely mild and detectable only by detailedneuropsychological testing and thus is unlikely to affectcontrast sensitivity. The fact that low-contrast acuity dem-onstrates a predictable decrease over time that depends on aninternal predictor of disease severity (namely, GAA repeatlength) supports its use as a potential biomarker in futurestudies. As a biomarker, low-contrast acuity has the advantageof being easily and reliably measured. This may be particularlyrelevant for gene therapy, in which treatment delivery remainsa challenge, particularly for inherited CNS diseases. De-velopment and validation of vision-based outcomes open the

Figure 3 Difference in binocular 1.25% contrast acuity between baseline and follow-up visit

The sample size at each visit is indicated inparentheses.

Table 1 Estimated number of years to loss of 7 letters ofbinocular 2.5% contrast acuity stratified byminimum GAA repeat length and age at onset

Age atonset, y

MinimumGAA <510

MinimumGAA 512-733

MinimumGAA 738-1320

<8 22 8.6 5.3

9–15 20 8.2 5.2

>16 18 7.9 5.1

Table 2 Estimated number of years to loss of 7 letters ofbinocular 1.25% contrast acuity stratified byminimum GAA repeat length and age at onset

Age atonset, y

MinimumGAA <510

MinimumGAA512-733

MinimumGAA 738-1320

<8 8.9 6.6 5.3

9–15 7.7 6.0 4.9

>16 7.0 5.4 4.9



door to retinal gene therapy as a potential proof of concept inFriedreich ataxia and as a clinically relevant therapeuticintervention.

Despite being one of the largest cohorts of its kind, our studyexperienced sample size limitations, particularly at longerfollow-up intervals. Although this was primarily due to cen-soring (and possibly mortality, given the association betweenhypertrophic cardiomyopathy and shorter follow-up dura-tion) rather than loss to follow-up, this nevertheless intro-duces the potential for information bias. This becomesparticularly relevant when calculating the time to loss of 7letters, as our estimates of low-contrast acuity at 5–10 yearsare extrapolated from data collected primarily at 0–5 years.Furthermore, although 7 letters of low contrast represent 1line of Snellen acuity equivalent and has been validated asa clinically meaningful outcome in MS and other patientpopulations because it is associated with detectable reductionsin vision-related quality of life, it has not been specificallyvalidated in Friedreich ataxia. The degree of low-contrastacuity loss is similar in Friedreich ataxia compared with MS,14

but this does not guarantee a similar reduction in vision-related quality of life because long-standing impairment maygo unnoticed or be well compensated by early adaptivemechanisms. A recent survey of patients with FRDA and theirfamilies as part of a patient-focused drug developmentmeeting identified vision loss as a concern, particularly whencompounded on other sources of disability,19 and in this co-hort, a notable proportion of the population had symptomaticvision loss defined as 20/40 or worse, especially later in thedisease course. However, future studies of low-contrast acuityand vision-related quality of life are needed in Friedreichataxia if it is to be used as a biomarker or surrogate outcome.

Author contributionsA.G. Hamedani: conceived the study, performed primarystatistical analysis, and drafted the manuscript. L. Hauser:critically reviewed the manuscript. S. Perelman, K. Mathews,G.R. Wilmot, T. Zesiewicz, S.H. Subramony, T. Ashizawa,M.B. Delatycki, and A. Brocht: participated in initial datacollection and critically reviewed the manuscript. D.R.Lynch: conceived the study and critically reviewed themanuscript.

Study fundingT. Zesiewicz holds intellectual property for the use of nico-tinic agonists in cerebellar ataxias and nonataxic imbalance. T.Ashizawa’s spouse has equity in Bio-Path Holdings.

DisclosureA.G. Hamedani and L. Hauser report no disclosures. S. Per-elman receives research support from the Friedreich’s AtaxiaResearch Alliance and the National Ataxia Foundation;receives support for clinical trials from Biohaven Pharma-ceuticals, Horizon Pharma, Reata Pharmaceuticals, Retrotope,Takeda Pharmaceuticals, and Teva; and has received fundingfrom EryDel, ViroPharma/Shire, Edison, and Pfizer for past

clinical trial studies. K. Mathews receives support for clinicaltrials from Horizon Pharma; has served on the scientificadvisory boards of DSMB, Santhera, Sarepta, BMS, andMSA; has received research funding from the NIH, theCDC, and the Friedreich’s Ataxia Research Alliance; and hasreceived research support from PTC Therapeutics, SareptaTherapeutics, Eli Lilly, aTyr, Pfizer, FibroGen, Roche, andItalfarmaco for clinical trials. G.R. Wilmot receives researchsupport from the National Ataxia Foundation and Frie-dreich’s Ataxia Research Alliance and support for clinicaltrials from Reata Pharmaceuticals and Biohaven Pharma-ceuticals; has served on the Santhera Pharmaceuticals Sci-entific Advisory Board for Biohaven Pharmaceuticals; andhas received research support from Reata Pharmaceuticalsand Biohaven Pharmaceuticals. T. Zesiewicz receives re-search support from the Friedreich’s Ataxia Research Alli-ance and support for clinical trials from Retrotope, BiohavenPharmaceuticals, Reata Pharmaceuticals, Takeda Pharma-ceuticals, Pfizer, Sagene Pharmaceuticals, AbbVie, Cavion,Bristol-Myers Squibb, and Sage Therapeutics. She has re-ceived consulting fees from Steminent Inc; has served on theeditorial board of Neurodegenerative Disease Management;holds a patent for varenicline in the treatment of ataxia; andserved as a consultant for Steminent Inc. S.H. Subramonyreceives research support from the NIH, Friedreich’s AtaxiaResearch Alliance, National Ataxia Foundation, MuscularDystrophy Association, and Myotonic Dystrophy Associa-tion and support for clinical trials from Horizon Pharma,Reata Pharmaceuticals, Ionis Pharmaceuticals, and Accel-eron Pharma; has served on the editorial board of theHandbook of Clinical Neurology, 2011 edition; and hasreceived research support from the NIH and STTR. T.Ashizawa receives research support and serves on the sci-entific advisory boards of the NIH, National Ataxia Foun-dation, andMyotonic Dystrophy Association; holds a patentfor the DNA test for SCA10, PVA or PEG conjugates ofpeptides for epitope specific immunosuppression and for themethods and compositions involving nucleotide repeatdisorders; and has received research support from IonisPharmaceuticals, Biohaven Pharmaceuticals, Biogen,NINDS, the Weill Cornell Medical Colleges, and theMarigold Foundation. M.B. Delatycki receives researchsupport from the National Health and Medical ResearchCouncil, Friedreich’s Ataxia Research Alliance, and Frie-dreich Ataxia Research Association; has served on the edi-torial board of BMC Pediatric Genetics; and has receivedresearch funding from FARA USA and Australia. A. Brochtreports no financial disclosures. D.R. Lynch receives re-search support from the NIH, Friedreich’s Ataxia ResearchAlliance, and U.S. Food and Drug Administration; receivessupport for clinical trials from Reata Pharmaceuticals andHorizon Pharma; has received funding for travel fromFranklin and Marshall College; serves on the editorial boardof the Journal of Neurogenetics; holds a patent for the test foranti-NDMA receptor encephalitis; and has received researchfunding from Takeda Pharmaceuticals and ENTRADATherapeutics. Full disclosure form information provided by



the authors is available with the full text of this article atNeurology.org/NG.

Received December 23, 2017. Accepted in final form May 22, 2018.

References1. Campuzano V, Montermini L, Molto MD, et al. Friedreich’s ataxia: autosomal re-

cessive disease caused by an intronic GAA triplet repeat expansion. Science 1996;271:1423–1427.

2. Harding AE. Clinical features and classification of inherited ataxias. Adv Neurol 1993;61:1–14.

3. Epstein E, Farmer JM, Tsou A, et al. Health related quality of life measures inFriedreich ataxia. J Neurol Sci 2008;272:123–128.

4. Tsou AY, Paulsen EK, Lagedrost SJ, et al. Mortality in Friedreich ataxia. J Neurol Sci2011;307:46–49.

5. Newman N. Hereditary optic neuropathies. In: Miller N, Newman N, editors. Walsh& Hoyt: Clinical Neuro-Ophthalmology, Vol. 1, 5th ed. Baltimore: Williams &Wilkins; 1999:741–773.

6. Carroll WM, Kriss A, Baraitser M, Barrett G, Halliday AM. The incidence and natureof visual pathway involvement in Friedreich’s ataxia: A clinical and visual evokedpotential study of 22 patients. Brain 1980;103:413–434.

7. Fortuna F, Barboni P, Liguori R, et al. Visual system involvement in patients withFriedreich’s ataxia. Brain 2009;132:116–123.

8. Pinto F, Amantini A, de Scisciolo G, Scaioli V, Guidi L, Frosini R. Visual involvementin Friedreich’s ataxia: PERG and VEP study. Eur Neurol 1988;28:246–251.

9. Givre SJ, Wall M, Kardon RH. Visual loss and recovery in a patient with Friedreichataxia. J Neuroophthalmol 2000;20:229–233.

10. Diehl B, Lee MS, Reid JR, Nielsen CD, Natowicz MR. Atypical, perhaps under-recognized? An unusual phenotype of Friedreich ataxia. Neurogenetics 2010;11:261–265.

11. Porter N, Downes SM, Fratter C, Anslow P, Nemeth AH. Catastrophic visual loss ina patient with Friedreich ataxia. Arch Ophthalmol 2007;125:273–274.

12. Seyer LA, Galetta K, Wilson J, et al. Analysis of the visual system in Friedreich ataxia.J Neurol 2013;260:2362–2369.

13. Lynch DR, Farmer JM, Tsou AY, et al. Measuring Friedreich ataxia: complementaryfeatures of examination and performance measures. Neurology 2006;66:1711–1766.

14. Friedman LS, Farmer JM, Perlam S, et al. Measuring the rate of progression inFriedreich ataxia: implications for clinical trial design. Mov Disord 2010;25:426–432.

15. Patel M, Isaacs CJ, Seyer L, et al. Progression of Friedreich ataxia: quantitativecharacterization over 5 years. Ann Clin Transl Neurol 2016;3:684–694.

16. Baier ML, Cutter GR, Rudick RA, et al. Low-contrast letter acuity testing capturesvisual dysfunction in patients with multiple sclerosis. Neurology 2005;64:992–995.

17. Lynch DR, Farmer JM, Rochestie D, Balcer LJ. Contrast letter acuity as a measure ofvisual dysfunction in patients with Friedreich ataxia. J Neuroophthalmol 2002;22:270–274.

18. Nieto A, Correia R, de Nobrega E, Monton F, Barroso J. Cognition in late-onsetFriedreich ataxia. Cerebellum 2013;12:504–512.

19. Friedreich’s Ataxia Research Alliance et al. The Voice of the Patient: Summary reportresulting from an externally led Patient-Focused Drug Development Meeting,a parallel effort to the US Food and Drug Administration’s (FD’s) Patient-FocusedDrug Development Initiative. 2017.




ARTICLE OPEN ACCESS

Population genealogy resource shows evidence offamilial clustering for Alzheimer diseaseLisa Anne Cannon-Albright, PhD, Sue Dintelman, MS, Tim Maness, BS, Johni Cerny, BS, Alun Thomas, PhD,

Steven Backus, BS, James Michael Farnham, BS, Craig Carl Teerlink, PhD, Jorge Contreras, JD,

John S.K. Kauwe, PhD, and Laurence J. Meyer, MD, PhD


Correspondence

Dr. Cannon-Albright

[email protected]

AbstractObjectiveTo show the potential of a resource consisting of a genealogy of the US record linked toNational Veterans Health Administration (VHA) patient data for investigation of the geneticcontribution to health-related phenotypes, we present an analysis of familial clustering of VHApatients diagnosed with Alzheimer disease (AD).

MethodsPatients with AD were identified by the International Classification of Diseases code. The Ge-nealogical Index of Familiality method was used to compare the average relatedness of VHApatients with AD with expected relatedness. Relative risks for AD were estimated in first- tofifth- degree relatives of patients with AD using population rates for AD.

ResultsEvidence for significant excess relatedness and significantly elevated risks for AD in relatives wasobserved; multiple pedigrees with a significant excess of VHA patients with AD were identified.

ConclusionsThis analysis of AD shows the nascent power of the US Veterans Genealogy Resource, in earlystages, to provide evidence for familial clustering of multiple phenotypes, and shows the utilityof this VHA genealogic resource for future genetic studies.

From the Genetic Epidemiology Program (L.A.C.-A., A.T., S.B., J.M.F., C.C.T.), Department of Internal Medicine, University of Utah School of Medicine; George E. Wahlen Department ofVeterans Affairs Medical Center (L.A.C.-A., L.J.M.); Pleiades Software Development (S.D., T.M.), Inc, Salt Lake City; Lineages (J.C.), Draper; SJ Quinney College of Law (J.C.), University ofUtah; Department of Biology (J.S.K.K.), Brigham Young University, Provo; Department of Dermatology (L.J.M.), University of Utah School of Medicine, Salt Lake City; and Departmentof Veterans Affairs (L.J.M.), Washington DC.


The Article Processing Charge was funded by VHA.







The US Veterans Genealogy Project links a genealogy of theUnited States with medical data for Veterans who use theVeterans Health Administration (VHA) system. While still inits infancy, at a current size of 63 million individuals withgenealogy data linked to 810,632 VHA patients, it is alreadysufficiently large for investigation of familial clustering formany health-related phenotypes.1 With a resource that allowsidentification of individuals with a phenotype of interest, andfor which biological relationships are known, it is possible totest for an underlying genetic predisposition.2 This resourceprovides a unique opportunity to explore evidence for manyphenotypes and to identify a rich resource of extended high-risk pedigrees.

Using this unique VHA resource still under creation, wepresent analysis of close and distant relationships amongindividuals diagnosed with Alzheimer disease (AD). Thisanalysis of familial clustering in a population of US Veteransshows significant evidence for excess relatedness, significantlyelevated risks in relatives, and identifies multiple extendedhigh-risk pedigrees, confirming evidence supporting a geneticcontribution to AD and displaying the potential of a powerfulnew national resource for predisposition gene identification.

MethodsUS veterans genealogyGenealogic data for over 63 million individuals gathered frompublic sources have been linked to a US genealogy. This re-source is currently based on collected genealogy data that todate focused on Alaska, Arizona, Colorado, Idaho, Hawaii,Kansas, Montana, Nebraska, Nevada, New Mexico, NorthDakota, South Dakota, Oklahoma, Oregon, Utah, Wash-ington, Wyoming, and Massachusetts. Original sources usedto compile the genealogy data include vital records, churchrecords, censuses from 1850 to 1940, published genealogies,cemetery records, including gravestone inscriptions, familytrees shared publicly online, oral history, and a variety of othersources used by genealogists to compile family trees. Thedemographic data for over 11 million Veterans using the VHASystem was record-linked to this US genealogy data usingGenMergeDB (pleiades-software.com), which has been usedto create, and link records to, multiple genealogic resourcesfor decades.1 Over 810,000 VHA patients were record-linkedto a unique individual in the genealogy using name, birthdate,and relationship data. After record linking was accomplished,no individual identifying data were used. The most importantgenealogic data for individuals are that for ancestors. We se-lected those VHA patients who linked to good ancestral data

to allow more precise matching of controls; ancestral dataallow us to identify more distantly related individuals in thesame generation. Of the 810,632 VHA patients with linkedgenealogy data, 184,658 patients have genealogy data for atleast 8 of their immediate ancestors, including at least bothparents, all 4 grandparents, and at least 2 great grandparents(many patients have muchmore genealogy data). These VHApatients with at least 8 of their immediate ancestors wereanalyzed here.

Standard protocol approvals, registrations,and patient consentsAccess to health data for the (unidentified) VHA patients withlinked genealogy was approved by the University of Utah andSalt Lake Veterans Affairs Institutional Review Board, andapproval was obtained from an oversight committee for theVHA resource.

Although the construction of the US genealogy to date hasfocused on genealogy sources with life events in the Westernstates, and the resource represents less than 25% of the finalUS genealogy to be created, VHA patients born in every statehave been identified. Among the 810,632 Veterans who linkto the genealogy, there are patients identified in all the 18VHA Veterans Integrated Service Networks (VISNs orregions) across the United States. Of the 46% of the 810,632linked VHA patients who have VISN data available, the largestnumbers of linked VHA patients were in VISN 16 (SouthCentral: 29,907), VISN 8 (Sunshine: 28,299), VISN 23(Midwest: 25,618), and VISN 19 (Rocky Mountain: 23,291),and the smallest numbers of linked patients were in VISN 5(Capitol: 7,740), VISN 2 (Upstate New York: 7,813), andVISN 3 (New York, New Jersey: 8,876). Among the 810,632Veterans who link to the genealogy, there are 154,213 femalepatients (19%); among the 184,658 patients with good an-cestral data, 15% are female. A wide age range of VHA patientslinked to genealogy data, with birth years ranging from theearly 1900s to the 1990s. The birth year distribution of the810,632 VHA patients who linked to any genealogy differedslightly from that of the 184,658 VHA patients who haddeeper ancestral genealogy data. Among the 810,632 linkedVHA patients, 11% were born before 1911, compared with14% of the 184,658 VHA patients who linked to ancestraldata; 16% of all 810,632 VHA patients who linked to gene-alogy data were born in the 1960s to the 1990s, comparedwith 9% of the 184,658 VHA patients with ancestral data.These differences might be expected, given that males havehigher record linking rates than females because of fewername changes and that individuals born less recently can beexpected to have more descendants.

GlossaryAD = Alzheimer disease; CI = confidence interval;GIF = Genealogical Index of Familiality; ICD = International Classification ofDiseases; PTSD = posttraumatic stress disorder; RR = relative risk; TBI = traumatic brain injury; UPDB = Utah PopulationDatabase; VHA = Veterans Health Administration; VISN = Veterans Integrated Service Network.


http://www.pleiades-software.com


VHA patients diagnosed with ADThe VHA has used an electronic medical record system atmost VHA medical centers for inpatient and outpatient caresince 1994. These records provide a rich source of phenotypedata on the 11 million Veterans who use the system. In-ternational Classification of Diseases (ICD) Revision 9 codingwas used to identify patients with AD (331.0).

Genealogical Index of FamilialityThe Genealogical Index of Familiality (GIF) test is a well-established method for testing for excess relatedness. It wasdeveloped for use with the Utah Population Database(UPDB), the first US genealogic resource used in research,2,3

and has previously been used to establish evidence for manydisease phenotypes, e.g., all cancer,4,5 asthma mortality,6 ro-tator cuff disease,7 lumbar disc disease,8 Alzheimer mortality,9

and prostate cancer,10 among others. A similar method hasbeen used to establish evidence for familial clustering fora variety of medical conditions in the Icelandic Genealogyresource.11

The GIF statistic is a measure of the average pairwise re-latedness for a set of individuals, for example, all VHA patientswith AD. The pairwise relatedness is measured using theMalecot coefficient of kinship,12 which is computed fromgenealogy information to estimate genetic relatedness fora pair of individuals. The coefficient of kinship estimates theprobability that 2 alleles at a locus are identical by descent(inherited from a common ancestor) in a pair of individuals.All possible paths of relatedness are considered in the calcu-lation. Most pairwise relationships in a large population-basedgenealogy are genetic distance = 0 (unrelated). For relatedpairs, the genetic distance increases with genetic distance, forexample, for parent and offspring = 1, for siblings or forgrandparent/grandchild = 2, for avunculars = 3, for firstcousins = 4, for second cousins = 6, and, similarly, for moredistant relationships. The GIF statistic is multiplied by 105 forease of presentation.

The GIF test compares the average pairwise relatedness ofa group of individuals to the expected average relatedness,which is estimated for a group of similar individuals in thepopulation. The expected average pairwise relatedness fora set of VHA patients can be estimated for a randomly selectedset of matched controls for the cases from the population of allVHA patients with linked genealogy data; controls arematched to cases for sex and 5-year birth year cohort. Toestimate the mean expected pairwise relatedness, 1,000 sets ofmatched VHA controls were randomly selected and analyzed.The empirical significance of the GIF test was obtained bycomparing the case GIF statistic to the distribution of the1,000 control GIF statistics. This comparison of averagepairwise relatedness tests whether the VHA patients with ADhave significantly higher relatedness than expected in theVHA population. The GIF test does not allow determinationof whether the familial clustering observed is due to envi-ronmental factors, genetic factors, or some combination. To

consider whether familial clustering might primarily be due toshared exposures or behavior among close relatives, ratherthan shared genetics, the GIF method includes a distant re-latedness test (dGIF). The dGIF test is performed as for theGIF test, but all relationships closer than third degree areignored. Thus, the dGIF test ignores relationships most af-fected by shared environment or behavior and tests for thepresence of excess distant relatedness only. Significant evi-dence for excess distant relatedness is strongly suggestive ofa genetic contribution. The GIF statistic summarizes averagepairwise relatedness in a single measure.

The contribution to the GIF statistic can be quantified sep-arately for the different genetic distances observed amongpairs of cases and controls (figure 1). The genetic distancemeasure represents, for example, 1 for parent/offspring, 2 forsiblings or grandparent/grandchild, 3 for avunculars or simi-lar, 4 for first cousins or similar, 6 for second cousins orsimilar, and so forth.

RRs in relativesThe estimation of relative risk (RR) in relatives providesa more traditional mechanism for identifying evidence fora genetic contribution. A genetic contribution to a phenotypeis supported when both close and distant relatives show evi-dence of elevated risk. First-degree relatives include parents,siblings, or offspring; second-degree relatives are the first-degree relatives of first-degree relatives (e.g., uncle, grand-mother); third-degree relatives are the first-degree relatives ofsecond-degree relatives (e.g., first cousin, great grandchild),and so forth. RRs were estimated for first- to fifth-degreerelatives of VHA patients diagnosed with AD as follows; allrelatives considered were also VHA patients. All 184,658patients in the VHA genealogy with genealogy data includingat least 8 of 14 immediate ancestors were assigned to one of 67cohorts based on birth year (in 5 years groups) and sex. Thecohort-specific rate of AD was estimated as the number of ADcases in each cohort divided by the total number of linkedVHA patients in the cohort. Expected numbers of first-degreerelatives with AD were estimated by counting the number ofrelatives, all of whom were VHA patients with genealogy data,by cohort (without duplication), multiplying by the rate ofAD in each cohort, and summing over all cohorts. Observednumbers of AD cases among relatives were counted withoutduplication. RRs were estimated for each degree of relation-ship (= observed/expected AD patients); 95% confidenceintervals (CIs) for the RR were calculated using standardmethods.13

High-risk pedigreesTo identify high-risk pedigrees for AD, all relationshipsamong all VHA patients with AD were analyzed. Consider-ation of all ancestral vectors allowed identification of clustersof related cases; the nearest common ancestor was identifiedfor each independent cluster of related patients with AD. Nocompletely overlapping clusters were considered, but somecases appeared in more than 1 cluster (or pedigree). For



a given founder of a cluster of related patients, the number ofobserved AD cases among the descendants of the founderwho were VHA patients was counted. To estimate theexpected number of patients with AD among the descendants,all linked VHA patients among the descendants were countedby cohort; the number of linked VHA patients in each cohortwas multiplied by the cohort-specific rate for AD (estimatedas described above) and summed over all cohorts. A com-parison of the number of observed to expected linked ADcases among the descendants in each cluster (pedigree) wasmade; if a significant excess of AD patients was observed (p <0.05), the pedigree was termed high-risk.

ResultsIn the VHA resource, 4,117 Veterans with genealogy for atleast 8 immediate ancestors and who had an ICD-9 codeindicating AD were identified; 194 (5%) were female. Table 1

summarizes the results of the GIF analysis for AD. The GIFtest summary includes the sample size (n), GIF statistic forcases (case GIF), mean GIF statistic for 1,000 sets of controls(mean control GIF), empirical significance for comparison ofoverall GIF (empirical GIF p), distant GIF statistic for cases(case dGIF), mean dGIF statistic for controls (mean controldGIF), and empirical significance for comparison of dGIF(empirical dGIF p). The average pairwise relatedness for the4,117 patients with AD was higher than expected for the VHApatient population (p < 0.001). When relationships closerthan third degree (first cousins) were ignored, the averagepairwise relatedness of the patients with AD was still elevatedover expected relatedness (dGIF p < 0.001).

Figure 1 shows the contribution to the GIF statistic by thepairwise genetic distance for cases compared with averages forthe 1,000 sets of matched controls. The effect of some datacensoring based on the nature of the VHA data available canbe observed in figure 1. Data censoring is present because

Table 1 GIF analysis in the VHA genealogy resource

Disease (ICD-9 code) n

Mean

Empirical GIF p

Mean

Empirical dGIF pCase GIF Control GIF Case dGIF Control dGIF

AD (331.0) 4,117 0.23 0.15 <0.001 0.15 0.11 <0.001

Matched AD controls 4,117 0.15 0.15 0.449 0.10 0.11 0.720

Prostate cancer (185) 12,695 0.18 0.14 <0.001 0.12 0.10 0.003

Parkinson disease (332) 3,850 0.22 0.15 0.001 0.15 0.11 0.002

Random set of patients 5,000 0.09 0.13 1.000 0.06 0.09 1.000

Bold values indicate statistical significance.Abbreviations: AD = Alzheimer disease; GIF = Genealogical Index of Familiality; ICD = International Classification of Diseases.

Figure 1Contribution to the GIF statistic by pairwise genetic distance for cases comparedwith the average for 1,000 sets ofmatched controls



diagnostic data for VHA patients were available only from1994 to present. Because medical diagnosis was only availablefor slightly more than 20 years (1994–2017), and because ADis typically diagnosed in older ages, AD-affected relatives whoare in the same generation (e.g., first degree: siblings or thirddegree: cousins) are the most likely to be observed; affectedrelatives in different generations (which includes all second-and fourth-degree relatives, e.g., a grandparent and grand-child) are unlikely to be observed in this narrow window. Asthe resource increases in size and years of data, this censoringwill be lessened.

To validate that control matching and selection representedthe VHA population and to demonstrate the overall baselinerelatedness of VHA patients, we randomly selected 1 set ofmatched controls for the 4,117 patients with AD (termed“matched AD controls” in table 1); we treated this set ofrandomly selected VHA patients as a set of “cases” and per-formed GIF analysis to determine whether this single set ofcontrols differed from 1,000 sets of matched controls (se-lected to match the single set of matched AD controls). Thisoriginal set of controls for patients with AD did not differ inexpected relatedness from the set of 1,000 sets of matchedcontrols (GIF p = 0.449, dGIF p = 0.720, table 1).

Using the same methods, we also performed GIF analyses forVHA patients diagnosed with 2 other common phenotypes(prostate cancer and Parkinson disease) for purposes ofcomparison to AD and for comparison of results for thesephenotypes reported from other resources. Prostate cancercases were identified with ICD-9 code 185, and Parkinsondisease cases were identified with ICD-9 code 332. The overallGIF test and the distant dGIF test showed significant excessrelatedness for both phenotypes (table 1). These analysesconfirm previously published evidence of significant excessrelatedness for both close and distant relationships for these 2phenotypes from population-based genealogy resources inUtah,4,5,14 Iceland11 and Scandinavia.15 In addition, to dem-onstrate that not all sets of VHA patients show greater thanexpected excess relatedness, we randomly selected 5,000 VHApatients with genealogy data and no associated phenotype.Results for the GIF analysis of this random set of patientsindicate no excess relatedness (table 1).

RRs in relativesEstimated RRs for AD among relatives of patients with ADwho are also VHA patients are shown in table 2, which dis-plays degree relatedness, total number of relatives amonglinked VHA patients (n), observed number of relatives withAD who were VHA patients (obs), expected number of rel-atives with AD who were VHA patients (exp), RR, signifi-cance (p value), and 95%CI for the RR (95%CI). RRs for ADwere significantly elevated among first- (RR = 1.82) and fifth-degree relatives (RR = 1.22) of patients with AD who wereVHA patients and were elevated (RR = 1.06), but not sig-nificantly (p = 0.380), among third-degree relatives. RRresults for second- and fourth-degree relatives are affected by

data censoring issues, as discussed previously. This is apparentwhen, for example, the different types of first-degree relativesare considered separately; 558 of the 882 first-degree relativesidentified were siblings, whereas only 274 were children and50 were parents. Because only 5% of the VHA AD patientswere female, comparisons of effects by sex were not possible.For example, all the 41 affected sibling pairs observed werebrothers, and both of the observed affected children of af-fected parents were sons.

High-risk pedigreesTwo hundred forty-five high-risk AD pedigrees were identified(p < 0.05), with at least 2 and up to 117 related VHA ADpatients. Figure 2 shows an example high-risk AD pedigreeidentified in the VHA genealogy resource; only the descendinglines to the VHAADpatients are shown. The pedigree includes6 related VHA AD patients, only 1.2 AD cases were expectedamong the descendants who were VHA patients (p = 0.0012).The pedigree founder was born in Pennsylvania in the late1700s and has almost 6,000 descendants in the genealogy; 67descendants are VHA-linked patients.

DiscussionThe VHA Genealogy Resource is a unique resource thatcontinues to grow and improve. We used this partially con-structed US genealogy of over 63 million individuals linked toalmost 1 million patient records representing all VHA localareas to show the potential for genetic analyses, using AD,a complex disease, as an example. The Alzheimer’s Associa-tion (2015) has reported that AD is the sixth leading cause ofdeath in the United States. The annual cost of dementia in theUnited States has been estimated to be $215 billion in 2010and is expected to double by 2040.16 One study17 projected13.8 million people diagnosed with AD dementia by 2050 inthe United States. Although old age is the primary risk factorfor AD, a genetic contribution to AD predisposition is alsowell recognized.18 Mutations in AβPP, Presenilin 1, andPresenilin 2 have been implicated in familial or early-onsetAD; the APOE e4 allele is a major genetic risk factor for AD;and other genetic risk factors involving lipid metabolism andimmune function are recognized.19–22

Both traumatic brain injury (TBI) and posttraumatic stressdisorder (PTSD) have been linked to an increased risk of ADand other dementias, and both are “signature injuries” ofindividuals serving in the Iraq and Afghanistan conflicts.23,24

Although AD is recognized as an important public healthissue, the association of these military-related injuries with ADmakes it of particular importance among military healthissues. The association of TBI and PTSD with an increasedrisk of AD suggests that it may be valuable to identify thosemilitary personnel at high risk of AD and to develop inter-ventions that could limit the progression or onset of disease.The results of this analysis of the VHA population, in com-bination with similar studies, suggest that there is a geneticcontribution to AD and that predisposition can already be



recognized through knowledge of family history of AD. As ADpredisposition genes are identified, increased risk may also berecognized by screening.

Our analyses demonstrate significant evidence for excess fa-milial clustering of VHA patients with AD compared withexpected clustering in this population; significantly elevatedRRs for AD in both close and distant relatives were observed,and many pedigrees with a significant excess of AD cases havebeen identified. This combined evidence confirms a heritablecontribution to the observed familial clustering of AD. Thepedigrees we identified suggest a highly elevated risk amongsome families. Such pedigrees are the ideal starting point forwhole-genome sequence–based approaches to the identifi-cation and characterization of rare high-penetrance variants.

In addition to the evidence presented for excess relatedness ofindividuals diagnosed with AD, using the samemethods, we have

replicated published evidence for excess familial relatedness for 2other common phenotypes to generalize validation of the re-source. The 2 common disease phenotypes examined (prostatecancer and Parkinson disease) have previously been recognizedto have a heritable component in other populations. Analysis ofthe VHA genealogy resource confirmed significant evidence forexcess relatedness for both phenotypes; these results additionallyvalidate the US VHA Genealogy Resource in terms of dataquality and power for analysis of familial clustering.

There are limitations to this analysis. It is optimal to matchcontrols based on all characteristics that might affect recordlinking or that are associated with the phenotype examined;we know that birth year, sex, and birth state (Utah or not)affect the overall relatedness of individuals in the similarUPDB genealogic resource that represents the Utah pop-ulation.2 For this analysis of the VHA resource, matching wasperformed only for birth year and sex. Although the rate of AD

Table 2 Estimated RRs for AD among first-to fifth-degree relatives of patients with AD in the VHA genealogy resource

Degree relatedness n obs exp RR p Value 95% CI

First degree 882 45 24.7 1.82 0.0001 1.33–2.44

Sibling 558 41 20.6 1.99 4.7e25 1.43–2.70

Brother 527 41 20.1 2.04 2.8e25 1.46–2.77

Parent 50 2 1.9 1.05 1.000 0.13–3.78

Child 274 2 2.2 0.90 0.617 0.11–3.25

Son 231 2 2.2 0.93 0.637 0.11–3.37

Second degree 633 2 7.7 0.26 0.018 0.03–0.94

Grandparent 11 0 0.4

Grandchild 186 0 0.3 0.00 0.759

Father’s brother 28 0 1.1

Mother’s brother 27 1 1.2 0.86 0.674 0.02–4.77

Sister’s son 157 1 2.3 0.43 0.322

Brother’s son 133 0 1.6 0.00 0.213

Half-sibs 18 0 0.4 0.00 0.670

Father’s sister 2 0 0.1 0.00 0.931

Mother’s sister 3 0 0.03 0.00 0.967

Brother’s daughter 29 0 0.1 0.00 0.917

Sister’s daughter 39 0 0.2 0.00 0.820

Third degree 1,226 37 34.9 1.06 0.380 0.75–1.46

First cousins 984 37 33.4 1.11 0.289 0.78–1.53

Fourth degree 1,896 35 41.9 0.84 0.161 0.58–1.16

Fifth degree 4,517 152 124.6 1.22 0.016 1.03–1.43

Bold values indicate statistical significance.Abbreviations: AD = Alzheimer disease; CI = confidence interval; exp = expected number of relatives with AD who were patients with VHA; obs = observednumber of relatives with AD who were patients with VHA; RR = relative risk.



was similar (2%) in both the 810,632 VHA patients linking toany genealogy and the 184,658 VHA patients with at least 8 oftheir immediate ancestors, the VHA patients with at least 8 oftheir immediate ancestors had a slightly lower rate of femalesand represented a population born slightly later than all VHApatients who linked to genealogy. It is not clear what effectthese slight differences might have had on the results. His-torically, race data have not been stored for the majority ofdemographic records in the VHA system, and so was un-available. In future, as more data become available, we pro-pose to use data including birth state, VISN, occupationalexposures, rank, and socioeconomic status, for example, formatching.

Data censoring is also an issue for this resource. VHA caseswho fail to link to genealogy data are censored, as are di-agnoses made outside the VHA system, or before 1994. Thiscensoring of cases might affect the estimation of rates of AD;however, because rates were estimated for the entire linkedpopulation of VHA patients and were only used as relativecomparisons, this censoring is not expected to affect results ortests of hypotheses. The overall rate of AD among all VHApatients with linked genealogy data was 2.2% (4,117/184,658). AD rates ranged from ;5% among male patientsborn before 1925 decreasing to 0.1% in male patients born inthe 1960s, with rates of ;3% in females born before 1925decreasing to 0.1% in female patients born in the 1950s.

In addition, genealogy data may not always represent bi-ological relationships. However, such censoring is assumed tobe independent of phenotype and equally affects both casesand controls. Finally, the data set is limited to primarily males,individuals who are part of groups who have shared genealogicdata, and veterans who used the VHA system; this is true ofboth cases and controls, but may have affected results. Thefamilial clustering methods presented are very robust to datacensoring. Controls are VHA patients, matched for sex andbirth year, and are required to have linkage to genealogy dataof similar quality and quantity as cases. It is difficult to

conceive of a mechanism by which significant excess re-latedness would exhibit itself in the VHA resource in theabsence of any true heritable component. Because of cen-soring, it is much more likely that the evidence of excessrelatedness presented is conservatively estimated. The re-latedness analysis of a set of matched controls considered ascases and the results for the 5,000 random VHA patients withno selected phenotype both demonstrated that the analysismethod appropriately observes no evidence for familial clus-tering for these examples.

The methods presented are robust to misclassification ofcases. False negatives (missing identification of true VHA ADcases) could result in failing to observe evidence for excessrelatedness; this did not occur. False positives, even at a veryhigh rate, could only affect this familial clustering analysis ifthe assignment of the incorrect diagnosis of AD in a VHApatient occurred more often among close and distant relativesof AD cases than among all VHA patients. This is unlikely,given the US-wide coverage of VHA patients represented andthe distance of the genetic relationships providing evidence ofexcess clustering.

This VHA Genealogy Resource represents what we believe isalready the largest genealogy linked to phenotype data basedon its current size of over 63 million individuals and its linkageto medical data for over 810,000 VHA patients. This resourceis still under construction; we estimate that the eventual sizeof this US genealogy will exceed 300 million individuals, with40%–60% of the 11 million VHA patients with demographicdata linked to genealogy. The analysis of the clustering of ADpresented here is 1 example of the utility of the resource forgenetic studies. The utility of this resource includes (1)demonstration of evidence for a genetic contribution to pre-disposition to many health-related phenotypes not commonlyobserved in other populations; (2) identification and study ofhigh-risk pedigrees informative for predisposition gene iden-tification; (3) estimation of family history–based risk of anydisorder of interest, which may be widely applicable to the US

Figure 2 Example high-risk AD pedigree identified in the VHA resource

Male founder has 2 marriages as does male grandson of the founder’s first marriage; fully shaded are AD cases.



population; and (4) identification of both high- and low-riskindividuals for any phenotype of interest for epidemiologicstudies or clinical trials, among others. This resource may beuniquely powerful for analysis of phenotypes that are rarelyobserved and highly associated with military service and mayhave an underlying genetic contribution (e.g., PTSD).

Clearly, this resource will improve in 2 distinct ways. Expan-sion of the genealogy data to all states is in progress. Second,there is enormous potential to refine the phenotypes ana-lyzed. The use of ICD-9 diagnostic coding to identify indi-viduals with a phenotype is not optimal; such coding has otherpurposes than research and may misrepresent the phenotypein both directions (false positives and false negatives). Futureanalysis of this resource will dictate more refined phenotypedefinitions. The Electronic Medical Records and GenomicsNetwork has demonstrated the benefit of more robust anal-ysis using multiple components of the medical record.25 Useof natural language processing algorithms from text data inmedical notes may allow better identification of phenotypeson a large scale. The VHA is well positioned to take advantageof these methods, specifically using the VHA Informatics andComputing Infrastructure resource.

In conjunction with the recent Million Veterans Programsponsored by the VHA, which is collecting and storing DNAand demographic and risk data for 1 million VHA patients,extremely powerful genetic studies will soon be possible. Withthe future additions of genotypes, exposure data, and otherdata that are envisioned, this VHA genealogy/phenotype re-source will allow informative genetic studies that include re-lationship data on an extremely large scale, including gene byenvironment analyses, and analysis of other medical con-ditions related to service, which cannot be studied in mostpopulations.

An initial genetic analysis of AD has been presented usinga powerful new and growing national resource linking gene-alogy and medical data. AD was selected as the examplephenotype, given that it is a major health issue in the UnitedStates and the world and may be an important issue for theVHA in light of reported associations of AD with trauma. Theanalyses confirmed evidence for an inherited component toAD risk, identified a current resource of high-risk pedigreesthat could be used for predisposition gene/variant identifi-cation, and confirmed the power and utility of this VHA re-source for genetic studies of complex human disease.

Author contributionsL.A. Cannon-Albright: study concept, resource creation,study design, analysis, and manuscript preparation. S. Din-telman and T. Maness: genealogy data resource design andconstruction and record linking. J. Cerny: genealogy datacollection and organization. A. Thomas: refinement of anal-ysis methods and tools. S. Backus: creation of genealogy/phenotype resource and analysis tools. J.M. Farnham: re-finement of analysis tools. C.C. Teerlink, J. Contreras, and

J.S.K. Kauwe: critical refinement of the manuscript. L.J.Meyer: resource concept and manuscript preparation.

Study fundingThis material is based on work supported in part by the De-partment of Veterans Affairs, Veterans Health Administration,Office of Research and Development, and RF1 AG054052(J.S.K.K, PI).

DisclosureL.A. Cannon-Albright has received research support from theNIH, US Department of Veterans Affairs, US Department ofDefense, and Intermountain Research and Medical Founda-tion. S. Dintelman is or has been employed by PleiadesSoftware Development, Inc and has received research supportfrom the Department of Veterans Affairs. T. Maness is or hasbeen employed by Pleiades Software Development, Inc andhas received research support from the Department of Vet-erans Affairs and Veterans Health Administration Office ofResearch and Development. J. Cerny is or has been employedby Lineages, Inc. and has received research support from theDepartment of Veterans Affairs. A. Thomas reports no dis-closures. S. Backus has received research support from the USDepartment of Health and Human Services, Veterans AffairsSalt Lake City Health Care System, Army Medical ResearchAcquisition, and NIH. J.M. Farnham reports no disclosures.C.C. Teerlink has received research support from the NIH,Department of Veterans Affairs, and Department of Defense.J. Contreras has received funding for travel and/or speakerhonoraria from American University, Stanford University,Arizona State University, University of Texas, Intel Corp,Standards Engineering Society, Centre for InternationalGovernance Innovation, Harvard University, Tilburg Uni-versity (Netherlands), American Antitrust Institute, JindalGlobal University, Practicing Law Institute, University ofCalifornia Berkeley, University of Nevada Las Vegas, BrighamYoung University, National Law University—Delhi, India,Leiden University, Netherlands, National Law School Uni-versity, Hispanic National Bar Association, Creative Com-mons, Saint Louis University, Hunan University (China),RIETI (Japan), Japan Fair Trade Commission, and DukeUniversity; receives or has received publishing royalties re-lated to the Cambridge Handbook of Technical Standardi-zation Law, Vol. 1, Cambridge Univ. Press 2017; has served asa consultant for Internet Society/IETF, International SAEConsortium Ltd, Hillcrest Laboratories, Inc., Natl. Assn. ofState Securities Agents, IPBridge, William Hagmaier, BLUProducts, Wilson Electronics, Industry PharmacogenomicWorking Group; has received research support from the NIH,Arizona State University, University of Texas, InternationalDevelopment Research Centre (IDRC), Centre for In-ternational Governance Innovation, and Huntsman CancerInstitute and Foundation; and has participated in legal pro-ceedings for the Internal Revenue Service and IPBridge. J. S.K.Kauwe serves or has served on scientific advisory boards ofADx Health Care; has served on the editorial board of Alz-heimer’s & Dementia; holds patents related to systems, assays,



and methods for determining risk factors for Alzheimer’sdisease; and serves or has served as a consultant for GenomaLLC. Laurence L.J. Meyer reports no disclosures. Full dis-closure form information provided by the authors is availablewith the full text of this article at Neurology.org/NG.


References1. Cannon-Albright LA, Dintelman S, Maness T, Backus S, Thomas A, Meyer LJ.

Creation of a national resource with linked genealogy and phenotypic data: theVeterans Genealogy Project. Genet Med 2013;15:541–547.

2. Cannon-Albright LA. Utah family-based analysis: past, present and future. HumHered 2008;65:209–220.

3. Skolnick M. The Utah Geneological Database: a resource for genetic epidemiology.In: Banbury Report No 4: Cancer Incidence in Defined Populations. Cold SpringHarbor: Cold Spring Harbor Laboratories; 1980.

4. Cannon L, BishopDT, SkolnickMH,Hunt S, Lyon JL, Smart CR.Genetic epidemiologyof prostate cancer in the Utah Mormon genealogy. Cancer Surv 1982;1:48–69.

5. Cannon-Albright LA, Thomas A, Goldgar DE, et al. Familiality of cancer in Utah.Cancer Res 1994;54:2378–2385.

6. Teerlink CC,HegewaldMJ, Cannon-Albright LA. A genealogical assessment of heritablepredisposition to asthma mortality. Am J Respir Crit Care Med 2007;176:865–870.

7. Tashjian RZ, Farnham JM, Albright FS, Teerlink CC, Cannon-Albright LA. Evidencefor an inherited predisposition contributing to the risk for rotator cuff disease. J BoneJoint Surg Am 2009;91:1136–1142.

8. Patel AA, Spiker WR, Daubs M, Brodke D, Cannon-Albright LA. Evidence for aninherited predisposition to lumbar disc disease. J Bone Joint Surg Am 2011;93:225–229.

9. Kauwe JS, Ridge PG, Foster NL, Cannon-Albright LA. Strong evidence for a geneticcontribution to late-onset Alzheimer’s disease mortality: a population-based study.PLoS One 2013;8:e77087.

10. Nelson Q, Agarwal N, Stephenson R, Cannon-Albright LA. A population-basedanalysis of clustering identifies a strong genetic contribution to lethal prostate cancer.Front Genet 2013;4:152.

11. Sveinbjornsdottir S, Hicks AA, Jonsson T, et al. Familial aggregation of Parkinsons’sdisease in Iceland. NEJM 2000;343:1765–1770.

12. Malecot G. Les Mathematiques de l’heredite. Paris: Masson & Cie; 1948.13. Agresti A. Categorical Data Analysis. New York: Wiley; 1990.14. Savica R, Cannon-Albright LA, Pulst S. Familial aggregation of Parkinson disease in

Utah: a population-based analysis using death certificates. Neurol Genet 2016;2:e65.15. Hemminki K, Dong C. Familial prostate cancer from the family-cancer database. Eur J

Cancer 2000;36:229–234.16. Delavande A, Hurd MD, Martorell P, Langa KM. Dementia and out-of-pocket

spending on health care services. Alzheimers Dement 2013;9:19–29.17. Hebert LE, Weuve J, Scherr PA, Evans DA. Alzheimer disease in the United States

(2010–2050) estimated using the 2010 census. Neurology 2013;80:1778–1783.18. Ridge PG, Hoyt KB, Boehme K, et al. “Assessment of the genetic variance of late-onset

Alzheimer’s disease.” Neurobiol Aging 2016;41:213–220.19. Harold D, Abraham R, Hollingworth P, et al. “Genome-wide association study

identifies variants at CLU and PICALM associated with Alzheimer’s disease.” NatGenet 2009;41:1088–1093.

20. Hollingworth P, Harold D, Sims R, et al. Common variants at ABCA7, MS4A6A/MS4A4E, EPHA1, CD33 and CD2AP are associated with Alzheimer’s disease. NatGenet 2011;43:429–435.

21. Lambert JC, Ibrahim-Verbaas CA, Harold D, et al. “Meta-analysis of 74,046 indi-viduals identifies 11 new susceptibility loci for Alzheimer’s disease.” Nat Genet 2013;45:1452–1458.

22. Ridge PG, Ebbert MT, Kauwe JS. Genetics of Alzheimer’s disease. Biomed Res Int2013;2013:254954.

23. Yaffe K, Vittinghoff E, Lindquist K, et al. Posttraumatic stress disorder and risk ofdementia among US veterans. Arch Gen Psychiatry 2010;67:608–613.

24. Plassman BL, Havlik RJ, Steffens DC, et al. Documented head injury in early adulthoodand risk of Alzheimer’s disease and other dementias. Neurology 2000;55:1158–1166.

25. Gottesman O, Kuivaniemi H, Tromp G, et al. The electronic medical records andgenomics (eMERGE)Network; past, present, and future. GenetMed 2013;15:761–771.




ARTICLE OPEN ACCESS

Impaired transmissibility of atypical prions fromgenetic CJDG114V

Ignazio Cali, PhD, Fadi Mikhail, MD, Kefeng Qin, PhD, Crystal Gregory, MS, Ani Solanki, BS,

Manuel Camacho Martinez, PhD, Lili Zhao, MS, Brian Appleby, MD, Pierluigi Gambetti, MD,

Eric Norstrom, PhD, and James A. Mastrianni, MD, PhD


Correspondence

Dr. Mastrianni

[email protected]

AbstractObjectiveTo describe the clinicopathologic, molecular, and transmissible characteristics of genetic priondisease in a young man carrying the PRNP-G114V variant.

MethodsWe performed genetic, histologic, and molecular studies, combined with in vivo transmissionstudies and in vitro replication studies, to characterize this genetic prion disease.

ResultsA 24-year-old American man of Polish descent developed progressive dementia, aphasia, andataxia, leading to his death 5 years later. Histologic features included widespread spongiformdegeneration, gliosis, and infrequent PrP plaque-like deposits within the cerebellum andputamen, best classifying this as a Creutzfeldt-Jakob disease (CJD) subtype. Molecular typingof proteinase K-resistant PrP (resPrPSc) revealed a mixture of type 1 (;21 kDa) and type 2(;19 kDa) conformations with only 2, rather than the usual 3, PrPSc glycoforms. Brainhomogenates from the proband failed to transmit prion disease to transgenic Tg(HuPrP) micethat overexpress human PrP and are typically susceptible to sporadic and genetic forms of CJD.When subjected to protein misfolding cyclic amplification, the PrPSc type 2 (;19 kDa) wasselectively amplified.

ConclusionsThe features of genetic CJDG114V suggest that residue 114 within the highly conserved pal-indromic region (113-AGAAAAGA-120) plays an important role in prion conformation andpropagation.

From the Department of Pathology (I.C., M.C.M., P.G.), Case Western University, Cleveland, OH; Department of Neurology (K.Q., F.M., C.G., A.S., L.Z., J.A.M.), University of Chicago; andDepartment of Biological Sciences (E.N.), DePaul University, Chicago, IL.









Prion diseases are rare transmissible neurodegenerative dis-orders that result from the accumulation of a misfolded iso-form (PrPSc) of the cellular prion protein (PrPC).1 Severalsubtypes are recognized based on their clinical, histopatho-logic, and PrPSc conformational phenotype.2 PrPSc confor-mation is approximated by the electrophoretic gel migrationpattern of proteinase K (PK)-resistant PrPSc (resPrPSc).ResPrPSc from Creutzfeldt-Jakob disease (CJD) typicallydisplays an unglycosylated fraction that migrates at either;21 kDa (PrPSc type 1) or ;19 kDa (PrPSc type 2),3 alongwith 2 higher molecular weight monoglycosylated and digly-cosylated PrPSc fractions. Experimental transmission of spo-radic CJD (sCJD) and genetic CJD (gCJD) to transgenic(Tg) mice that express human (Hu) PrPC has been useful toprove transmissibility.4,5 The in vitro protein misfolding cyclicamplification (PMCA) assay has also been used to assesstransmission barriers and the propagation of PrPSc confor-mational subtypes between PrP species.6,7

Approximately 10% of prion diseases result from an autoso-mal dominant sequence change within the coding segment ofthe PRNP gene.2 A single nucleotide change within codon114, resulting in a valine (V) substitution of glycine (G), hasbeen reported in a family from China8 and Uruguay.9 Wereport an American case and his father. Several features ofgCJDG114V, including an early onset with possible incompletepenetrance, an atypical glycosylation pattern of PrPSc and,most strikingly, poor transmissibility to Tg(HuPrP) mice, allof which underscore a key role for residue 114 in prionconformation and propagation.

MethodsHistopathologyHistopathology and immunohistochemistry (IHC) offormalin-fixed, formic acid–treated brain tissue were per-formed, as described previously.10 Hematoxylin and eosinstaining was performed on 5-μm sections, whereas IHC to PrPand glial fibrillary acidic protein (GFAP) was performed on 8-μm sections. Each section was deparaffinized, rehydrated, andimmersed in tris-buffered saline with Tween 20 before in-cubation with the EnVision Flex Peroxidase Blocking Reagent(Dako). For PrP immunostaining, sections were incubatedwith hydrochloric acid (1.5 mmol/L) and microwaved for 15minutes, probed with monoclonal antibodies (mAb) 3F4 (1:1,000)11 or GFAP (1:20,000) (Sigma-Aldrich) for 60minutes, and then washed and incubated with Envision Flex/

horseradish polymerase polymer for 30 minutes (Dako) be-fore visualization of the immunoreactivity with Envision Flex3,39-diaminobenzidine tetrahydrochloride (Dako).

Molecular studiesWestern blot (WB) analysis was performed as previouslydescribed.10 Briefly, a 10% (wt/vol) brain homogenate (BH)was prepared in lysis buffer (100 mM NaCl, 10 mM EDTA,0.5%NP-40, 0.5% sodium deoxycholate, 10mMTris, pH 7.4)or in lysis buffer with 100 mM Tris (100 mM NaCl, 10 mMEDTA, 0.5% NP-40, 0.5% sodium deoxycholate, 100 mMTris, pH 8.0) and incubated for 1 hour at 37°C with 20 or100 μg/mL of PK (Sigma-Aldrich), followed by terminationof the reaction with 2 mM phenylmethylsulfonyl fluoride(Sigma-Aldrich). Samples were then diluted 1:1 with 2XLaemmli sample buffer (Bio-Rad) and denatured at 100°C for10 minutes prior loading onto 12%, 14%, or 15% poly-acrylamide gels. For experiments with PMCA, aliquots of“PMCA+” and “PMCA−” (see below) were incubated witheither 100 or 50 μg/ml PK, depending on whetherTg(HuPrP) or Tg(HuPrPGlyKO) was used as the substrate.Incubation with PK was performed at 40°C for 60 minuteswith shaking. After denaturation (at 100°C for 10 minutes),samples were precipitated with prechilled methanol (1:10dilution) sonicated, and loaded onto a 15% Tris-Glycineprecast gel (Bio-Rad). Proteins were transferred to poly-vinylidene difluoride membranes, probed with mAb 3F4 for 2hours at 1:40,000 dilution12 followed by incubation witha horseradish peroxidase–conjugated goat antimouse sec-ondary Ab for 1 hour at 1:3,000 dilution. After incubationwith a chemiluminescence substrate (Pierce or AmershamBiosciences), signal was recorded with a Bio-Rad XRS digitaldocument imager or visualized on Kodak films.

Protein misfolding cyclic amplificationBH from 2 distinct Tg mouse models were used as the sub-strate in PMCA. Tg(HuPrP) mice are hemizygous for humanPrPC, which expresses PrPC with methionine (M) at codon129 (i.e., PrPC-129M) on a PrPC null background (PrP−/−), at;8 × normal level of PrPC.13 The TgNN6h line, hereafterdesignated Tg(HuPrPGlyKO), expresses human PrPC with 2point mutations that eliminate 2 N-linked glycosylation siteson human PrPC (codons 181 and 197), in which asparagine(N) is replaced by glutamine (Q) (PrP-N181Q/N197Q).14

These mice are homozygous for the transgene that isexpressed at;0.6-fold that of normal mouse brain PrP levels.Brains harvested from 1.5-month-old healthy Tg mice wereperfused with 5mMEDTA in 1XDulbecco phosphate-buffered

GlossaryBH = brain homogenate; CJD = Creutzfeldt-Jakob disease; DPBS = Dulbecco phosphate-buffered saline; DWI = diffusion-weighted imaging; gCJD = genetic CJD; GFAP = glial fibrillary acidic protein; GSS = Gerstmann-Straussler-Scheinker; IBH =infectious BH; IHC = immunohistochemistry;MMSE = Mini-Mental State Examination; NBH = normal BH; PrPC = cellularprion protein; PK = proteinase K; PMCA = protein misfolding cyclic amplification; PTFE = polytetrafluoroethylene; sCJD =sporadic CJD; SD = spongiform degeneration; WB = Western blot.



saline (1X DPBS). Ten percent normal BH was prepared inconversion buffer (150 mM NaCl, 1% Triton X-100, 1XProtease Inhibitor Cocktail in 1X DPBS) and centrifuged at800 × g at 4°C for 1 minute. The collected supernatants(hereafter designated as NBH) were used as the “substrate”for the PMCA reaction. BH from the frontal cortex of theproband, used as “seed” (hereafter identified as infectious BHor IBH), was prepared following the same procedure used togenerate the NBH.

PMCA was performed as previously described with few mod-ifications.15 The NBH from Tg(HuPrP) or Tg(HuPrPGlyKO)and IBH from the proband or control cases of sCJD weremixed in a 9:1 ratio in a test tube, in the presence of 0.05%digitonin detergent (Sigma-Aldrich) and one 2.38 mm-diameter polytetrafluoroethylene bead (McMaster-Carr).One aliquot of the NBH-IBH mixture was subjected toPMCA (PMCA+), and one aliquot was stored at −80°C andused as the negative control (PMCA–). PMCA reactions werecarried out in a programmable sonicator (Misonix SonicatorS-4000; Misonix Inc, Farmingdale, NY) for a total of 96 cyclesof sonication. Each cycle consisted of 30 seconds of sonication(Amplitude 98, wattage 230–250) and 29.5 minutes of in-cubation at 37°C.

Standard protocol approvals, registrations,and patient consentsHuman-related studies were conducted under an approvedUniversity of Chicago Institutional Review Board protocol(#9713), and animal studies were performed under an ap-proved University of Chicago Institutional Animal Care andUse Committee approval of protocol (#70735).

PRNP gene analysisDNA was extracted from whole blood, and the entire codingsegment of the PRNP gene was amplified by PCR usingpreviously reported primers10 and bidirectional sequencingusing an automated sequencer.

Prion transmissionsFresh-frozen brain from the frontal cortex was homogenizedin 1X phosphate buffered saline to a final concentration of 1%,as previously described.16 Tg mice expressing human PrPC

with Met coding at residue 129 on a mouse PrP knock-outbackground (i.e., Tg[HuPrP-129M]Prnpo/o) were previouslyconstructed and described elsewhere.13 Mice were anes-thetized with a xylazine/ketamine mixture, the head fixed ina small animal stereotaxic instrument (Kopf) and in-tracerebrally inoculated with BH, directly through the scalp, asdescribed previously.13 Typical symptoms of prion disease inmice, including reduced spontaneous movement, scruffy coat,hunched back, and progressively unsteady gait, were moni-tored every other day. When symptomatic and near-terminal,mice were killed and their brains harvested, freshly hemi-sectioned, half frozen in dry ice, and half fixed in 2% formalin.All mouse studies were approved by the Institutional AnimalCare and Use Committee before study.

ResultsCase reportA 24-year-old right-handed white man with no known medicalhistory except for a mild learning disability developed generaland episodic anxiety, intermittent confusion, fatigue, memoryimpairment, and slurred speech;1 year before seeking medicalattention. Initial neurologic evaluation demonstrated mildhypomimia, finger agnosia, and fine bilateral hand tremor. Aneuropsychological battery revealed a full-scale IQ of 73, im-paired attention, concentration, memory, and mild-to-moderateexecutive dysfunction. An EEG showed paroxysmal diffuse 4–5Hz slow activity, with occasional sharp wave–like morphology.An extensive workup included routine laboratory tests for de-mentia, in addition to antithyroid antibodies, a paraneoplasticpanel that included NMDA and GAD65 antibodies, amino acidprofile, and heavy metal testing, all of which were negative. CSFcell count, glucose, protein, angiotensin converting enzyme level,VDRL, Lyme PCR, total tau, phosphorylated tau, Aβ42, and 14-3-3 protein were within normal range or negative, as was bac-terial, acid-fast bacilli, and fungal cultures. Brain MRI wasreported as normal.

Over the next year, he became more withdrawn, declined cog-nitively, developed progressive clumsiness, prominent word-finding difficulty, and slowed motor skills that led to 3 caraccidents. At the time, he had a coarse hand and head tremor,cogwheel rigidity of the limbs, stooped posture, and a mildlyspastic gait with impaired arm swing. He scored 20 of 30 on theMini-Mental State Examination.17 Fluorodeoxyglucose PETexhibited decreased but symmetric cortical metabolism, morepronounced in the posterior parietal, temporal, and occipitalregions, while thalamic uptake was similar to the cortex, andthere was a relative increase in metabolic activity in the basalganglia. Diffusion-weighted imaging (DWI) sequences on MRIdisplayed restricted diffusion throughout the cortical ribbon, andright, greater than left, cortical atrophy (figure 1, A and B). Overthe next 2 years, the patient developed global aphasia, choreiformmovements and myoclonus of the upper extremities, increasedaxial rigidity, and severe gait ataxia, ultimately confining him toa wheelchair and eventual death 5 years from the initial onset ofsymptoms, at age 29 years.

The proband’s father suffered a similar course, beginning at age32 years, with progressive cognitive decline and prominentaphasia that progressed to mutism, in addition to visual hallu-cinations, gait ataxia, postural instability, and generalized my-oclonus. The brainMRI report noted “diffuse cerebral atrophy,markedly out of proportion for age,” but films were not avail-able for review. He died at age 34 and underwent brain autopsy,although PRNP gene sequencing was not performed at thetime. Tissue was unavailable for genetic testing, but limitedfixed tissue was available for histologic staining.

Pathologic studiesWidespread spongiform degeneration (SD) and gliosis weremost severe in the frontal (figure 2A), temporal, and parietal



cotices, and least severe in the occipital lobe and hippocampus(figure 2, B and C). Severe SD was also present within theputamen (figure 2D), but not the adjacent globus pallidus,and within the medial nuclei of the thalamus, but not thelateral thalamus (not shown). The midbrain, cerebellum(figure 2, E and F), pons, and medulla were also devoid of SD

(not shown). IHC with anti-PrP mAb 3F4 revealed fine,diffuse PrP staining and occasionally larger granules withinthe cerebral cortex (figure 2, M–O) and neostriatum (figure2P), as well as plaque-like PrP deposits in the putamen andcerebellum (figure 2, P and Q). IHC for GFAP in frontal andtemporal lobes revealed intense reactive astrogliosis(figure 2R).

In a limited number of sections from the proband’s father, SDwas observed within the cerebral cortex and putamen (figure2, G, H, and J), with sparing of the hippocampus, midbrain,pons, medulla, and cerebellum (figure 2, I, K, and L).

Genetic studiesSequencing the entire coding segment of the PRNP gene ofthe proband revealed a heterozygous thymine substitution ofguanine at the second nucleotide of codon 114 (c.341G>T),resulting in a change in coding from glycine (G) to valine (V)(G114V) (figure 3A). The polymorphic codon 129, which isknown to confer susceptibility to prion disease, modify dis-ease phenotype, and influence the conformation of PrPSc, washomozygous for methionine (129MM) (figure 3). No otheralterations were detected. Although no genetic informationwas available from the proband’s father, based on the early ageat disease onset and similarity of the presentation, we assumedthat he carried the same PRNP-G114V variant. A pedigree wasconstructed, based on information provided by family members

Figure 1BrainMRI of the gCJDG114V proband during the 3rdyear of his disease

(A) DWI sequence shows restricted diffusion throughout the cortical ribbonand within the caudate nuclei. (B) FLAIR sequence shows diffuse corticalatrophy. DWI = diffusion-weighted imaging; FLAIR = fluid-attenuated in-version recovery.

Figure 2 Histopathology and immunohistochemistry of brain sections from the proband and proband’s father

(A–L) Hematoxylin and eosin (H&E) staining of the proband (A–F) and proband’s father (G–L). (A–D) Severe spongiformdegeneration (SD) affecting frontal lobe(A) and putamen (D), andmild SD in the occipital lobe (B) and CA1 region of the hippocampus (C). (E–F) No SD affecting themidbrain (E) and cerebellum (F). (G,H, J) Severe SD affecting the frontal lobe (G), andmild SD in the parietal lobe (H) and putamen (J). (I, K–L) No SD affecting the hippocampus (I), midbrain (K), andcerebellum (L). Arrows in B, C, and J indicate SD. (M–Q) PrP immunostaining of the proband. (M and N) Diffuse PrP staining affecting the frontal lobe (M) andsubiculum (N); the arrow in (N) indicates a cluster of larger PrP granules of different size. (O) Intense PrP staining at the interface between the dentate gyrusand Ammon horn (arrows); the inset depicts a high magnification of the area outlined by the dashed rectangle. (P and Q) Plaque-like PrP immunostaining inthe putamen (P) and cerebellum (Q); antibody: 3F4. (R) Intense gliosis (arrow) affecting the temporal cortex; antibody: GFAP.



(figure 3B). The paternal grandmother (II-E) was alive at 81,although the paternal grandfather reportedly died withoutdementia at age 84 years and none of the father’s 6 siblings,aged 40–59 years, developed a progressive neurologic dis-order. A paternal uncle died of cancer at age 57 years (III-G)and an aunt died 3 days after birth (IIIH). We sequenced thecoding segment of the PRNP gene of 5 unaffected familymembers, including the paternal grandmother (II-E), 3 ofhis father’s siblings (III-C, D, and E) and the proband’s sister(IV-B). None carried the G114V variant, and the poly-morphic codon 129 genotype was homozygous for Val(129VV) in 4 members (II-E, III-C, III-D, and IV-B) and129MV in 1 (III-E). Since the paternal grandmother (II-E)was 129VV and the G114V change is on the 129M allele, thegrandfather would be predicted to have a 129MV genotypeand be a silent carrier of G114V on the 129M allele. How-ever, surprisingly, the genotype of the proband’s aunt (III-E)was 129MV with a normal 129M allele.

Typing of PrPSc

The PrPSc conformation enciphers the clinicopathologicphenotype of prion disease.16,18 The gel mobility of resPrPSc

is an indirect measure of the PrPSc conformation. Althougheasily detected by WB within each brain region tested, the

level of resPrPSc varied from highest in the parietal cortex tolowest in the cerebellum (figure 4A). In contrast to the typicalresPrPSc of sCJD, resPrPSc from gCJDG114V lacked thediglycosylated band, exhibiting only prominent mono-glycosylated, and less prominent unglycosylated, resPrPSc

bands. In addition, compared with the ;21 kDa gel mobilityof unglycosylated resPrPSc from all other brain regions, initialWB of resPrPSc from the frontal lobe showed a gel mobility of;19 kDa, matching that of PrPSc type 2 (figure 4A). Fol-lowing multiple extended electrophoretic runs, a doublet wasrevealed, supporting the coexistence of PrPSc types 1 and 2 inthe frontal cortex (figure 4B).19 These findings were in-dependently replicated at the University of Chicago and CaseWestern Reserve University.

Transmission studiesPrevious work shows that sequence homology between PrPSc

and PrPC at critical sites within PrP, especially at the poly-morphic residue 129, is known to affect the efficiency oftransmission.5,20 As such, we used Tg(HuPrP-129M) micethat express human PrP with Met at residue 129, to match the129 coding of the proband. Surprisingly, after nearly 700 daysfollowing intracerebral inoculation with BH from the pro-band, no clinical signs of disease or histopathologic changes

Figure 3 Gene sequencing and pedigree analysis

(A) Chromatogram segments of the PRNP genesequences surrounding the codon 114 sequencealteration (*) and the polymorphic codon 129. Aguanine (G) to thymine (T) change at the secondnucleotide of codon 114, resulting in a valine (V)substitution for glycine (G) at residue 114 of PrP.The polymorphic codon 129 was homozygous formethionine (Met). (B) Pedigree of the gCJDG114V

family. The proband (IV-A) and his father (III-B)exhibited progressive dementia at young ages.Sequencing of the PRNP gene of the proband (ar-row) revealed the G114V variant. DNA from theproband’s father was not available. No other familymembers were reportedly affected. Sequencingwas performed on individuals marked with an as-terisk, although none carried the G114V variant.The polymorphic codon 129 status is representedwithin each symbol. V = valine, M = methionine.Numbers below symbols represent the age atdeath or at the time of genetic testing. Square =male; circle = female; slash = deceased; filled sym-bol = affected individual.



(not shown) were evident, whereas control Tg(HuPrP-129M) mice inoculated with BH prepared from a case ofsCJDMM1 yielded efficient transmission after 161.7 ± 4.0days (figure 4C). WB prepared from inoculated mice revealedno resPrPSc from mice inoculated with gCJDG114V comparedwith robust levels of resPrPSc type 1 in mice inoculated withsCJDMM1 (figure 4D).

Protein misfolding cyclic amplificationWe wondered whether propagation of PrPSc fromgCJDG114V could be demonstrated using PMCA becausethis technique allows for a more controlled setting thatmight enhance prion propagation. BH from healthyTg(HuPrP-129M) and Tg(HuPrPGlyKO) mice was used asthe substrate in separate reactions with PrPSc from theproband as the seed. The proband’s resPrPSc prepared inconversion buffer and not subjected to PMCA confirmedthe presence, under these conditions, of mixed PrPSc types 1and 2 lacking the diglycosylated fraction (figure 5A). PMCAreactions using Tg(HuPrP) as substrate amplified PrPSc

type 2, but not type 1. Only monoglycosylated and non-glycosylated bands were present, confirming the absence ofa diglycosylated seed. Preferential PrPSc type 2 amplificationwas also observed using BH from Tg(HuPrPGlyKO) mice

that express a mutated form of human PrP that lacksN-linked sugar chains at residues 181 and 197 (figure 5A).As the control, PrPSc from sCJDMM1, with all glycoforms,was robustly amplified using Tg(HuPrP) brain as the PrPC

substrate (figure 5B).

DiscussionWe report the clinicopathologic, molecular, and transmissibleproperties of gCJDG114V in an American-born individual ofPolish descent, with no known linkage to either of the 2previously reported families from China and Uruguay.8,9 Al-though the proband and his father had a generally similarcourse that included progressive dementia and aphasia, dis-ease onset in the proband was marked by behavioral change,primarily anxiety attacks, followed by extrapyramidal features,myoclonus, and generalized ataxia. A similar course, withbehavioral changes, was described for most of the 5 affectedmembers of the Uruguayan family, although ataxia was nota feature. Ataxia was also not described in the Chinese family,although aphasia was reported. DWI imaging of our probandrevealed widespread cortical and caudate hyperintensities,although CSF 14-3-3 protein was negative. A similar pattern

Figure 4 Assessment of PrPSc type and transmissibility of gCJDG114V to Tg(HuPrP) mice

(A) WB prepared from brain homogenates (BHs) of the proband and control patient with sCJDMM1 probed with mAb 3F4. Equal total protein loads compareseveral brain regions with a case of sCJDMM1 before (top) and after (bottom) proteinase K (PK) digestion to reveal PK-resistant PrPSc (resPrPSc). Theelectrophoretic profile of untreated samples from gCJDG114V and sCJDMM1 appeared similar, although after digestion, all samples from the proband lackedthe diglycosylated band. The parietal lobe carried the highest level of resPrPSc. (B) WB showing coexisting resPrPSc types 1 and 2 in the frontal cortex (FC; lane3), and type 1 in the occipital cortex (OC; lane 4) of gCJDG114V, compared with resPrPSc type 1 (lane 1) and type 2 (lane 2) from sCJDMM1 and sCJDMM2,respectively. (C) Survival times (days) of Tg(HuPrP-129M)mice following intracerebral inoculation of 30μL of 1% frontal cortex BH from the gCJDG114V probandcompared with sCJDMM1 (n = 6 mice per group). One mouse inoculated with gCJDG114V was killed 600 days after inoculation because of signs of aging,although the remainder were healthy andwithout symptoms after nearly 700 days when they were killed. All Tg(HuPrP-129M)mice inoculatedwith sCJDMM1died with signs of prion disease (rough coat, hunched back, and unsteady gait) at 161.7 ± 4.0 days. (D) WB of representative mouse brain samples confirmingthe absence of resPrPSc in mice inoculated with gCJDG114V compared with the high level of resPrPSc in mice inoculated with sCJDMM1.



of widespread DWI hyperintensity and negative CSF 14-3-3protein was noted in the Chinese family proband, althoughneither of these studies were provided in the Uruguayanfamily report.

The early onset of our proband and his father compare wellwith onsets ranging from 18 to 45 years in the earlierreported families. Despite the early onset of disease, whichsuggests high penetrance, 45- and 61-year-old asymptomaticcarriers of the G114V variant of PRNP were reported in theUruguayan and Chinese families, which suggests incompleteor variable penetrance. Although our proband and his fatherwere the only 2 affected family members, the G114V variantwas not detected in other members sampled, which limitsour ability to comment on penetrance. Based on the codon129 genotypes, we suggest that the sequence alteration couldhave either arisen de novo in the paternal grandfather’sgermline or that a nonpaternity event skewed the genetics ofthis pedigree. However, the typically early onset and the lackof disease in previous generations raise questions whetheranticipation or, more likely, other genetic or environmentalfactors play a significant role in disease manifestation linkedto this genetic variant.

The pathogenicity of the G114V variant is supported by itsposition within the highly conserved palindromic segmentof PrP (113-AGAAAAGA-120) that features directly in theformation of PrPSc and the PrPSc-PrPC complex that leadsto propagation of PrPSc.21 Although the G114V and A117Vvariants lie within this palindrome, 3 other pathogenicalterations (G131V, S132I, and A133V) lie slightly down-stream, but still within the hydrophobic core thatextends from residue 113 to 134. Curiously, each of thesepromotes a Gerstmann-Straussler-Scheinker (GSS) disease

phenotype, defined by the presence of multicentric PrPamyloid plaques isolated to the cerebellum or diffuselydistributed throughout the cortex and deep nuclei.22–27

Globular PrP deposits were present within the cerebellumand putamen of the proband, but these were sparse andwere not detected in his father or the 2 previously reportedgCJDG114V families.8,9

The gel mobility of resPrPSc is an indirect measure of PrPSc

conformation, which differs among the major subtypes ofprion disease.2,28 The gel migration pattern of CJD-derivedresPrPSc typically includes 3 glycoforms that range from;19or 21 to 30 kDa, whereas GSS-associated resPrPSc has a singlenonglycosylated band of 7–16 kDa, representing N- andC-terminally cleaved PrPSc. The resPrPSc associated withA133V has not been reported, but that from A117V,29,30

S132I,22 and G131V24 conforms to the GSS pattern. Thatfrom our proband migrates more characteristically as CJD,despite the absence of the diglycosylated fraction. Of interest,a similar pattern has been reported with the T183A,31

V180I,32 and P105S10 variants and in a single case of sCJD.33

Although the sequence changes at residues 180 and 183 coulddirectly interfere with nearby N-linked glycosylation sites atresidue 181, the distal location of residues 105 and 114, andthe occurrence of this resPrPSc type in a sCJD case, suggeststhat the conformation of this rare PrPSc type favors selectionof the unglycosylated and monoglycosylated PrPSc isoforms.Although we also found a mixture of ;19- and ;21-kDaresPrPSc in our proband, this feature has been reported in asmany as ;40% of sCJD cases with the 129MM genotype19

and in rare cases of gCJDE200K 34 and gCJDD178N.35 Overall,the clinicopathologic and molecular typing characteristicsbest suggest that the PRNP-G114V variant should be cate-gorized as a CJD subtype.

Figure 5 Amplification of PrPSc by protein misfolding cyclic amplification (PMCA)

Brain homogenates prepared from thefrontal cortex of gCJDG114V and sCJDMM1were used to seed the conversion of PrPC

into PrPSc. (A) After PMCA (+), the ;19-kDaPK-resistant PrPSc (resPrPSc) fraction (*)from gCJDG114V, but not ;21-kDa resPrPSc

(**), was amplified using PrPC from theTg(HuPrP) or Tg(HuPrPGlyKO) mouse brain.Sporadic CJDMM1 and sCJDMM2 controlswere loaded for typing purposes. (B) AfterPMCA (+), PrPSc from sCJDMM1 was effi-ciently amplified using PrPC fromTg(HuPrP) as the substrate. PMCA (−) refersto the same samples that were not sub-jected to PMCA reaction. The mAb 3F4 wasused to probe immunoblots. Similarresults were obtained from 3 independentPMCA reactions.



The lack of transmission of gCJDG114V to TgHuPrP miceimplies that the sequence change at residue 114 introducesa barrier to transmission. Efficient prion transmission has beenshown to depend on the balance of two critical determinants:(1) sequence homology between PrPSc and PrPC at key sites ofassociation within the PrPSc-PrPC interface and (2) the abilityof PrPC to acquire the conformation of PrPSc.We can speculatethat the sequence alteration at residue 114 that sits within thehighly conserved palindrome represents a mismatch at an as-sociation site within the PrPSc-PrPC interface or the confor-mation of PrPSc induced by the G114V variant falls outside theconformational repertoire that PrPC in TgHuPrP mice canadopt. Although data are limited regarding the transmissibilityof disease resulting from other variants within the palindromicregion, we have not had success transmitting GSSA117V toTg(HuPrP) mice (unpublished observations). Whether thisrelates to a universal feature of GSS to poorly transmit diseasecompared with CJD36,37 or the sequence mismatch at residue117, was addressed by 1 group that successfully transmittedGSSA117V to a Tg mouse expressing the homologous sequencechange.38 In addition, a more recent study found that bankvoles are susceptible to GSS prions, suggesting that the se-quence of bank vole PrPC can adopt the conformation of PrPSc

in GSSA117V despite the absence of homology at residue 117between PrPSc and PrPC.39 Thus, the conformation of PrPSc

linked to the PRNP-G114V variantmay, in fact, be the principalbarrier to transmission. The finding that PMCA amplificationof this conformation was possible suggests that conditions ofthis artificial system are more favorable for the PrPSc-PrPC

association that leads to propagation of PrPSc. However, de-spite the partially positive results, a conformational barrier toprion propagation was still evident, based on the selectiveamplification of the 19-kDa resPrPSc fraction. Recent worksuggests that in the presence of 2 conformational subtypes ofPrPSc, PMCA favors amplification of the least stable con-former,14 which is suggested to be the 19-kDa PrPSc fraction inthis case, based on its limited presence within the frontal lobes,compared with the widespread presence of the 21-kDaconformer.

Our findings demonstrate that the G114V variant of PRNPpromotes a CJD clinicopathologic phenotype linked to anatypical PrPSc conformation that lacks the diglycosylatedfraction and is poorly transmissible to Tg(HuPrP) mice. Al-though the G114V variant of PRNP appears to be associatedwith variable penetrance that is not yet well characterized, itslocation within the highly conserved palindromic segment ofthe hydrophobic core of PrP and the associated atypicalconformation of PrPSc are likely to contribute to the uniqueproperties of gCJDG114V prions that we observed. Overall,these findings highlight the importance of this residue and thepalindrome in prion conformation and propagation and asa potential target of therapy.

Author contributionsI. Cali: histologic preparations, data interpretation, and gen-eration of figures and text for manuscript preparation. K. Qin:

preparation of Western blots, interpretation of data, and as-sistance with text preparation. F. Mikhail: patient data ac-quisition, preparation of figures, and manuscript writing.A. Solanki: breeding andmaintenance of mice and inoculationand monitoring of mice for transmission studies. M.C.Martinez: performed PMCA. L. Zhao: mouse screening andgene sequencing of family members and transgenic mice.C. Gregory: preparation of Western blots, collection of bloodsamples from subjects, preparation of inoculums, and in-oculation of mice for transmission studies. P. Gambetti:histologic studies, data interpretation, and manuscript prep-aration. E. Norstorm: preparation of figures, data assessment,and manuscript writing. J.A. Mastrianni: design of the study,interpretation of data, and manuscript preparation. B. Appleby:tissue samples and personnel resources.

Study fundingThis work was supported by an endowment from the BrainResearch Foundation (J.A.M.), NIH P01AI106705, NIH5R01NS083687, and Charles S. Britton Fund (P.G.); TheNational Prion Disease Pathology Surveillance Center.

DisclosureI. Cali serves as an editorial board member of theWorld Journalof Neurology, BioMed International, and the Asian Journal ofNeuroscience. K. Qin reports no disclosures. F. Mikhail has re-ceived research funding from the NRSA T32 Ruth KirschsteinTraining Grant. C. Gregory, A. Solanki, M.C. Martinez, and L.Zhao report no disclosures. P. Gambetti has served on thescientific advisory board of Ferring Pharmaceuticals; holdsa patent for Sant Cruz BioTechnology (A2006-03237); and hasreceived research funding from the NIH. E. Norstrom reportsno disclosures. J.A. Mastrianni has served on the scientific ad-visory board of the Alzheimer’s Association; has consulted forCVS Caremark and the Federal Trade Commission; has re-ceived commercial research funding fromEli Lilly and the CMSand Alzheimer’s Association; and has received research fundingfrom the Brain Research Foundation. Full disclosure form in-formation provided by the authors is available with the full textof this article at Neurology.org/NG.

Received February 9, 2018. Accepted in final form May 14, 2018.

References1. Prusiner SB. Prions (Les Prix Nobel Lecture). In: Frangsmyr T, editor. Les Prix

Nobel. Stockholm: Almqvist & Wiksell International; 1998:268–323.2. Brown K, Mastrianni JA. The prion diseases. J Geriatr Psychiatry Neurol 2010;23:

277–298.3. Parchi P, Castellani R, Capellari S, et al. Molecular basis of phenotypic variability in

sporadic Creutzfeldt-Jakob disease. Ann Neurol 1996;39:767–778.4. Watts JC, Prusiner SB. Mouse models for studying the formation and propagation of

prions. J Biol Chem 2014;289:19841–19849.5. Mallik S, Yang W, Norstrom EM, Mastrianni JA. Live cell fluorescence resonance

energy transfer predicts an altered molecular association of heterologous PrPSc withPrPC. J Biol Chem 2010;285:8967–8975.

6. Castilla J, Gonzalez-Romero D, Saa P, Morales R, De Castro J, Soto C. Crossing thespecies barrier by PrP(Sc) replication in vitro generates unique infectious prions. Cell2008;134:757–768.

7. Jones M, Peden AH, Head MW, Ironside JW. The application of in vitro cell-freeconversion systems to human prion diseases. Acta Neuropathol 2011;121:135–143.

8. Ye J, Han J, Shi Q, et al. Human prion disease with a G114V mutation andepidemiological studies in a Chinese family: a case series. J Med Case Rep 2008;2:331.




9. Rodriguez MM, Peoc’h K, Haik S, et al. A novel mutation (G114V) in the prionprotein gene in a family with inherited prion disease. Neurology 2005;64:1455–1457.

10. Tunnell E, Wollman R, Mallik S, Cortes CJ, Dearmond SJ, Mastrianni JA. A novelPRNP-P105S mutation associated with atypical prion disease and a rare PrPSc con-formation. Neurology 2008;71:1431–1438.

11. Kascsak RJ, Rubenstein R, Merz PA, et al. Mouse polyclonal andmonoclonal antibodyto scrapie-associated fibril proteins. J Virol 1987;61:3688–3693.

12. Zou WQ, Puoti G, Xiao X, et al. Variably protease-sensitive prionopathy: a newsporadic disease of the prion protein. Ann Neurol 2010;68:162–172.

13. Telling GC, Scott M, Mastrianni J, et al. Prion propagation in mice expressing humanand chimeric PrP transgenes implicates the interaction of cellular PrP with anotherprotein. Cell 1995;83:79–90.

14. Haldiman T, Kim C, Cohen Y, et al. Co-existence of distinct prion types enablesconformational evolution of human PrPSc by competitive selection. J Biol Chem2013;288:29846–29861.

15. Castilla J, Saa P, Morales R, Abid K, Maundrell K, Soto C. Protein misfolding cyclicamplification for diagnosis and prion propagation studies. Methods Enzymol 2006;412:3–21.

16. Mastrianni JA, Nixon R, Layzer R, et al. Prion protein conformation in a patient withsporadic fatal insomnia. N Engl J Med 1999;340:1630–1638.

17. Folstein M, Folstein S, McHugh PR. “Mini-mental status”: a practical method forgrading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12:189–198.

18. Telling GC, Parchi P, DeArmond SJ, et al. Evidence for the conformation of thepathologic isoform of the prion protein enciphering and propagating prion diversity.Science 1996;274:2079–2082.

19. Cali I, Castellani R, Alshekhlee A, et al. Co-existence of scrapie prion protein types 1and 2 in sporadic Creutzfeldt-Jakob disease: its effect on the phenotype and prion-type characteristics. Brain 2009;132:2643–2658.

20. Wadsworth JD, Asante EA, Desbruslais M, et al. Human prion protein with valine 129prevents expression of variant CJD phenotype. Science 2004;306:1793–1796.

21. Norstrom EM, Mastrianni JA. The AGAAAAGA palindrome in PrP is required togenerate a productive PrPSc-PrPC complex that leads to prion propagation. J BiolChem 2005;280:27236–27243.

22. Hilton DA, Head MW, Singh VK, Bishop M, Ironside JW. Familial prion disease witha novel serine to isoleucine mutation at codon 132 of prion protein gene (PRNP).Neuropathol Appl Neurobiol 2009;35:111–115.

23. Panegyres PK, Toufexis K, Kakulas BA, et al. A new PRNP mutation (G131V)associated with Gerstmann-Straussler-Scheinker disease. Arch Neurol 2001;58:1899–1902.

24. Jansen C, Parchi P, Capellari S, et al. A second case of Gerstmann-Straussler-Scheinker disease linked to the G131V mutation in the prion protein gene in a Dutchpatient. J Neuropathol Exp Neurol 2011;70:698–702.

25. Kovacs GG, Ertsey C, Majtenyi C, et al. Inherited prion disease with A117V mutationof the prion protein gene: a novel Hungarian family. J Neurol Neurosurg Psychiatry2001;70:802–805.

26. Mastrianni JA, Curtis MT, Oberholtzer JC, et al. Prion disease (PrP-A117V) pre-senting with ataxia instead of dementia. Neurology 1995;45:2042–2050.

27. Mohr M, Tranchant C, Steinmetz G, Floquet J, Grignon Y, Warter JM. Gerstmann-Straussler-Scheinker disease and the French-Alsatian A117V variant. Clin Exp Pathol1999;47:161–175.

28. Parchi P, Giese A, Capellari S, et al. Classification of sporadic Creutzfeldt-Jakobdisease based onmolecular and phenotypic analysis of 300 subjects. AnnNeurol 1999;46:224–233.

29. Yang W, Cook J, Rassbach B, Lemus A, DeArmond SJ, Mastrianni JA. A new trans-genic mouse model of Gerstmann-Straussler-Scheinker syndrome caused by theA117V mutation of PRNP. J Neurosci 2009;29:10072–10080.

30. Tagliavini F, Lievens PM, Tranchant C, et al. A 7-kDa prion protein (PrP) fragment,an integral component of the PrP region required for infectivity, is the major amyloidprotein in Gerstmann-Straussler-Scheinker disease A117V. J Biol Chem 2001;276:6009–6015.

31. Grasbon-Frodl E, Lorenz H, Mann U, Nitsch RM, Windl O, Kretzschmar HA. Loss ofglycosylation associated with the T183A mutation in human prion disease. ActaNeuropathol (Berl) 2004;108:476–484.

32. Chasseigneaux S, Haik S, Laffont-Proust I, et al. V180I mutation of the prion proteingene associated with atypical PrPSc glycosylation. Neurosci Lett 2006;408:165–169.

33. Zanusso G, Polo A, Farinazzo A, et al. Novel prion protein conformation and gly-cotype in Creutzfeldt-Jakob disease. Arch Neurol 2007;64:595–599.

34. Kovacs GG, Seguin J, Quadrio I, et al. Genetic Creutzfeldt-Jakob disease associatedwith the E200K mutation: characterization of a complex proteinopathy. Acta Neu-ropathol 2011;121:39–57.

35. Haık S, Peoc’h K, Brandel JP, et al. Striking PrPsc heterogeneity in inherited priondiseases with the D178N mutation. Ann Neurol 2004;56:909–910. author reply10–11.

36. Brown P, Gibbs CJ Jr, Rodgers-Johnson P, et al. Human spongiform encephalopathy:the National Institutes of Health series of 300 cases of experimentally transmitteddisease. Ann Neurol 1994;35:513–529.

37. Tateishi J, Kitamoto T, Hoque MZ, Furukawa H. Experimental transmission ofCreutzfeldt-Jakob disease and related diseases to rodents. Neurology 1996;46:532–537.

38. Asante EA, Linehan JM, Smidak M, et al. Inherited prion disease A117V is not simplya proteinopathy but produces prions transmissible to transgenic mice expressinghomologous prion protein. PLoS Pathog 2013;9:e1003643.

39. Pirisinu L, Di Bari MA, D’Agostino C, et al. Gerstmann-Straussler-Scheinker diseasesubtypes efficiently transmit in bank voles as genuine prion diseases. Sci Rep 2016;6:20443.



ARTICLE OPEN ACCESS

SCN11A Arg225Cys mutation causes nociceptivepain without detectable peripheral nervepathologyRyan Castoro, DO, MS, Megan Simmons, MS, Vignesh Ravi, BS, Derek Huang, MD, Christopher Lee, MD,

John Sergent, MD, Lan Zhou, MD, PhD, and Jun Li, MD, PhD


Correspondence

Dr. Li

[email protected]

AbstractObjectiveThe SCN11A gene encodes the NaV1.9 sodium channel found exclusively in peripheral noci-ceptive neurons.

MethodsAll enrolled participants were evaluated clinically by electrophysiologic studies, DNA se-quencing, and punch skin biopsies.

ResultsAll affected family members are afflicted by episodes of pain. Pain was predominantly noci-ceptive, but not neuropathic in nature, which led a diagnosis of fibromyalgia in some patients.All patients had normal findings in nerve conduction studies for detecting large nerve fiberneuropathies and skin biopsies for detecting small nerve fiber pathology.

ConclusionsUnlike those patients with missense mutations in SCN11A, small fiber sensory neuropathy, andneuropathic pain, the Arg225Cys SCN11A in the present study causes predominantly noci-ceptive pain with minimal features of neuropathic pain and undetectable pathophysiologicchanges of peripheral neuropathy. This finding is consistent with dysfunction of nociceptiveneurons. In addition, since nociceptive pain in patients has led to the diagnosis of fibromyalgia,this justifies a future search of mutations of SCN11A in patients with additional pain phenotypessuch as fibromyalgia to expand the clinical spectrum beyond painful small fiber sensoryneuropathy.

From the Department of Physical Medicine and Rehabilitation (R.C.), Vanderbilt University Medical Center; Department of Neurology (M.S., V.R., D.H., C.L., J.L.), Center for HumanGenetic Research, and Vanderbilt Brain Institute, Vanderbilt University Medical Center; Division of Rheumatology (J.S.), Department of Medicine, Vanderbilt University MedicalCenter, Nashville, TN; and Department of Neurology (L.Z.), University of Texas Southwestern Medical Center, Dallas, TX.


The Article Processing Charge was funded by NIH grant.







Mutations in genes encoding voltage-gated sodium channels(Nav) have been instrumental to the understanding ofmechanisms underlying pain. This has mainly involved 3types of Nav—Nav1.7, Nav1.8, and Nav1.9. The key proteinelement, α-subunit, of these Nav is encoded by SCN9A,SCN10A, and SCN11A genes,1–3 respectively.

Genotype and phenotype associations have not been clear inmutations of SCN11A. Three patients with L811P andL1302F missense mutations develop congenital insensitivityto pain.4,5 Other missense mutations (Ile381Thr, Lys419Asn,Ala582Thr, Ala681Asp, Ala842Pro, Leu1158Pro, andPhe1689Leu) were associated with painful sensory neuropa-thy.4 These 7 mutations were not verified through cose-gregation analysis. Finally, 4 additional missense mutations(Arg225Cys, Ala808Gly, Arg222His, and Arg222Ser) wereidentified in 4 large Asian families and an Argentina family,6–8

each cosegregated with family members affected by episodicpain syndrome (EPS). However, EPS was not systematicallycharacterized; in particular, it is unknown whether thesepatients have pathologic or electrophysiologic changes inperipheral nerves.

The Arg225Cys mutation in SCN11A is predicted to affectthe S4 voltage sensor domain. This domain consists of anα-helix with positively charged amino acids such as arginineand lysine that allow for ion exchange. Patch-clamp studieshave demonstrated increased conductance for mutantArg225Cys, with no effect on activation or inactivationpotentials when compared with wild type.9

In this study, we report a family with the Arg225Cys mis-sense mutation in SCN11A that exhibits no clinical andpathophysiologic evidence of peripheral neuropathy, butnociceptive pain. This disassociation suggests dysfunctionalnociceptive neurons by the mutation.

MethodsPatientsThe affected proband (0001) was initially evaluated at theVanderbilt Rheumatology Clinic for fibromyalgia. She wasthen referred to the Vanderbilt Charcot-Marie-Tooth (CMT)clinic. The remaining 1 unaffected and 5 affected familymembers were enrolled and evaluated at the Clinical ResearchCenter at Vanderbilt University Medical Center (figure 1A).This study was approved by the local institutional reviewboard (IRB). A written consent form was obtained from allparticipants.

In addition to medical history and neurologic examination,the CMT neuropathy score10 was obtained from all patients.The score comprised sensory/motor symptoms and physicalfindings in limb sensation/muscle strength. The electro-physiologic portion of the score was omitted to produce theCMT Examination Score (CMTES) and will be describedseparately. The CMTES ranges from 0 to 28, with higherscores indicating an increase in disease severity.

Nerve conduction studiesNerve conduction study (NCS) data were acquired usingconventional methods.11 For motor nerves, the distal stimu-lation distances for motor conduction studies were 7 cm in thearms and 9 cm in the legs. For the sensory nerves, the

Figure 1 Family pedigree and DNA sequencing results

(A) Family pedigree: filled shapes represent affected members, and unfilledshapes represent unaffected members. The arrow points to the proband.(B) Representative sanger sequencing traces were from an affected and anunaffected participant. The box outlies the mutated nucleotide.

GlossaryALS = amyotrophic lateral sclerosis; CMT = Charcot-Marie-Tooth; CMTES = CMT Examination Score; CMTNS = CMTneuropathy score; DP = diabetic polyneuropathy; EPS = episodic pain syndrome; IRB = institutional review board; NCS =nerve conduction study; RAPS = Rheumatoid Arthritis Pain Scale.



stimulation distance was 14 cm for median, ulnar, and suralnerves but 10 cm for the radial nerves.

DNA sequencingThe proband’s DNA was initially studied by targeted gene-panel next-generation sequencing, a commercial diagnosticservice provided by Medical Neurogenetics (Atlanta, GA).The panel (NGS400) covers 39 pain-related genes. Its tech-nique was similar to what was described.12

DNA from blood cells was extracted from all other familymembers using a commercial kit (#A1620; Promega, Madi-son, WI). DNA was then sequenced by Sanger sequencing forthe Arg225Cys mutation. Primers were designed as 59-TCGATTTCACCTTGGAGGTC-39 and 59-CAGTTAG-CACAGTGCCTGGA-39. PCR products were sequenced inGenHunter Corporation (Nashville, TN).

Human skin biopsy and intraepidermal nervefiber densityThis technique has been detailed before.13 In brief, punch skinbiopsies (3 mm in diameter) were obtained from the distal legand distal and proximal thigh of all participants. The tissueswere fixed in fixatives (4% paraformaldehyde). The tissueswere shipped to the coauthor’s (L.Z.) laboratory for pro-cessing. Epidermal nerve fibers were stained with antibodiesagainst PGP9.5 and counted under light microscopy. Theexperimenter who processed and counted the skin biopsieswas blinded to the patient’s ID and genotypic information.

Pain questionnairesTo characterize the nature of pain in the affected familymembers, we have identified 2 pain scales in the literature.The Rheumatoid Arthritis Pain Scale was developed to eval-uate nociceptive pain in patients with rheumatoid arthritis.14

Because of the features and the responsiveness of pain toNSAIDs in our patients, this pain scale was considered to besuitable. We omitted those questions specific for arthritis orjoint problems. The painDETECT scale was developed toevaluate neuropathic pain and was able to distinguish neu-ropathic pain from nociceptive pain in previously publishedstudies.15–17 We reworded the questions to improve theirclarity for patients. The questionnaire is detailed in supple-mentary materials (links.lww.com/NXG/A60).

The pain scale was first tested in patients with amyotrophiclateral sclerosis (ALS) and diabetic polyneuropathy (DP).The study was conducted on consecutively identified patientswith DP and ALS at the Vanderbilt Neurology Clinic overa 12-month period. Participants were excluded if they wereyounger than 18 years, if they had a coexisting medical con-dition that may cause significant pain such as severe cervical orlumbar spine disease, rheumatoid arthritis, or osteoarthritis,or if they answered “no” for the first question. We contactedpatients by telephone to administer the scale prospectivelyafter verbal consent was obtained. The scale was then appliedto the affected family members with the mutation in SCN11A.

The neuropathic and nociceptive subscores were comparedusing the Wilcoxon signed-rank test. It should be clarified thatthis study is not aimed at validating new pain scales, but atusing the previously published scales to characterize the painfeatures quantitatively.

Standard protocol approvals, registrations,and patient consentsThis project was approved by the Vanderbilt University IRB.All patients consented to participation in this study.

ResultsPatients present with pain that has episodic,length-dependent, and inflammatory features

Clinical phenotypesPatient II-5This is a 29-year-woman with a chief complaint of chronicpain. She came to the Vanderbilt Rheumatology clinic forevaluation of fibromyalgia. Pain is mainly concentrated in feetbut only 1 place at a time. She describes it as “deep, aching”pain. It typically lasts from 15 to 30 minutes. It starts mildlyand then peaks. Affected areas may be warm during theseepisodes but range of motion, strength, and sensation areintact and showed no erythema or swelling. Pain occurreddaily during her childhood, but now twice per week. Pain ismostly at night and may wake her up from sleep but some-times during the day. She rates pain at 6–7 of 10. It is oftenworsened with weather changes but could occur withoutobvious reason. Ibuprofen helps but cannot prevent therecurrence. After she was seen in the Vanderbilt CMT clinic,she discontinued ibuprofen for 12 months but took gaba-pentin that failed to control pain. Ibuprofen was restarted.

Symptoms disappeared when she was pregnant but returned2 months after childbirth. She denied a history of diabetes,renal disease, vitamin deficiency, or HIV infection. She isa practicing optometrist with no history of cognitivedysfunction.

On neurologic examination, she was found to have normalcranial nerve functions. Her muscle strength was 5 on theMedical Research Council scale in all muscle groups. Sensa-tion was intact to light touch, pinprick, and vibration. Deeptendon reflexes were normal. Her CMTES was 0.

A complete laboratory evaluation for systemic rheumatologicdisorders was performed. It revealed a mildly elevatederythrocyte sedimentation rate of 32 (range 0–20), otherwisenormal for antinuclear antibodies and creatine kinase.

Patient I-1The proband’s aunt is a 53-year-old woman. She had a normaldevelopmental history. Her symptoms started in early child-hood. Pain occurred in her bilateral knees, ankles, and toesand was about 3–4 times per week. She describes pain as dull/




aching type. It was made worse by exertion and weatherchanges and made better by taking nonsteroidal anti-inflammatory drugs (NSAIDs). It typically lasts 20–30minutes.She denied numbness or tingling associated with her pain. Ofinterest, at the age of;20 years, her symptoms including painresolved spontaneously. She denied fainting, sweating, ororthostasis associated with the pain episodes. Her neurologicexamination was normal. The CMTES was 0 at her clinic visit.The remaining 4 affected members in table 1 showed symp-toms similar to the 2 cases detailed above.

Taken together, these 6 affected members demonstrated thefollowing key phenotypic features: (1) pain is episodic and fitswith nociceptive or inflammatory pain, but not neuropathicpain, in nature; (2) sites of pain are length dependent, and all6 patients have pain mainly in feet and distal legs; (3)NSAIDs, but not gabapentin, are effective in all 6 patients,supporting inflammatory pain; and (4) all patients have noabnormal physical findings.

DNA analysis shows a missense mutation ofArg225Cys in the SCN11A geneThe proband’s DNA was initially studied by targeted next-generation sequencing (Medical Neurogenetics, Atlanta).This test covered 39 pain-related genes (mnglabs.com/). Itdetected a missense mutation of p.Arg225Cys in SCN11A.The same mutation has been reported to be pathogenic forEPS in 2 Chinese families.6

We verified the Arg225Cys mutation using Sanger sequencingin our laboratory (figure 1B). Arg225Cys was found only inaffected family members (0001, 0100, 0101, 0103, 1000, and1001) but absent in a nonaffected family member (0102).

The mutation was evaluated by 2 servers—PolyPhen-2 andSorting Intolerant fromTolerant (SIFT).18 Both PolyPhen-2 andSIFT predicted that Arg225Cys is pathogenic. In addition, theArg225Cys allele was absent in 107,784 chromosomes of controlpopulation in the ExAC database (exac.broadinstitute.org/).

Laboratory tests showed no pathophysiologicabnormalities in both small and larger nervefibers of patients affected by the diseaseMissense mutations in Nav1.7, Nav1.8, and Nav1.9 have allbeen associated with painful sensory neuropathy with puresmall nerve fibers affected or small fibers predominantlydamaged.4 To further clarify this issue, we performed NCS inall our participants (table 2). There was no evidence of pe-ripheral neuropathy in all studied participants.

NCS mainly tests larger diameter nerve fibers but is in-sensitive to the changes in small diameter nerve fibers.19 Wethus performed skin biopsies to quantify intraepidermal nervefiber density. This test targets nonmyelinated sensory nervefibers and is highly sensitive in diagnosing small fiber sensoryneuropathy.13 All studied participants showed normal EDNF(table 3). Therefore, there is no evidence of small and larger

Table 1 Clinical phenotype in patients with the Arg225Cys mutation in SCN11A

IdentificationaAge/Sex Frequencyb Factorc

Duration,min Nature Severityd Npe Ncf Localization Medication

II-5 29/F 2/wk PAg 10–20 Deep ache 4 0 4 Feet andhands

Ibuprofen andacetaminophen

II-1 30/M 2/mo PA andWCh

30 Deep ache, mildburning, andmild tingling

5 2 17 Limbs Ibuprofen andnaproxen

II-2 28/M 2/wk PA andWC

20–30 Deep andthrobbing ache

5 0 16 Feet, legs,and hands

Ibuprofen andnaproxen

II-3i 27/F NA NA NA NA 0 0 0 NA NA

II-4 22/F 1–2/d PA andWC

10–30 Deep ache 4.5 0 4 Feet, legs,and hands

Ibuprofen

I-1j 53/F NA PA andWC

NA NA 0 0 0 NA NA

I-2 51/F 2/mo PA andWC

45–60 Cold, gnawingache, and sharp

4 0 13 Limbs Ibuprofen andacetaminophen

Abbreviation: NA = not applicable.a Identification used in the pedigree.b Pain frequency.c Factors exacerbating or triggering pain.d Visual analog scale (1–10).e Neuropathic score.f Nociceptive score.g Prolonged activities.h Weather change.i This participant had no mutation found in SCN11A and thus exhibited no symptom.j This patient’s symptoms resolved spontaneously in her 20s, thus exhibited no phenotype at the clinical visit.


https://mnglabs.com/

http://exac.broadinstitute.org/


fiber peripheral neuropathy in our patients with the Arg225-Cys mutation.

Evaluation through pain questionnaires is alsosupportive of a nociceptive or inflammatorypain, not neuropathic painTo systematically characterize pain in patients with theArg225Cys mutation, we used previously published scalesthat were designed to detect nociceptive and neuropathicpain.14,15 ALS primarily affects motor nerves. It is mainlyassociated with nociceptive pain. By contrast, patients withdiabetic neuropathy are well known to develop neuropathicpain.20We therefore tested the scale to determine whether thescale can show the difference of neuropathic vs nociceptivepain in these 2 populations of patients. The distribution ofdemographics was similar between DP and ALS groups (tablee-1 and figure e-1, links.lww.com/NXG/A60). Patients withDP showed significantly more neuropathic pain than noci-ceptive pain (p = 0.01). By contrast, participants with ALSshowed more nociceptive pain (p = 0.02).

We then administered the scale to patients with theArg225Cys mutation (table 1). Patients with the Arg225Cysmutation showed almost exclusively nociceptive pain but no

or minimal neuropathic pain (9.0 ± 7.2 nociceptive vs 0.3 ±0.82 neuropathic; p = 0.015). Note that participant #1000 hasnot had pain since her 20s due to spontaneous remission. Shewas removed from the statistical comparison between neu-ropathic and nociceptive pain.

DiscussionOur study has identified a Caucasian family afflicted bya missense mutation of Arg225Cys in SCN11A that has notbeen described in North America except in Asian families.6 Inconsidering the cosegregation of the mutation with affectedfamily members in previously reported families6–8 plus thefamily reported in the present study, the Arg225Cys is clearlypathogenic. However, we found no pathophysiologic evi-dence of peripheral neuropathy in our patients.

All affected patients exhibit features of nociceptive pain that istypically seen in patients with rheumatoid arthritis or otherinflammatory joint diseases but does not fit with neuropathicpain. In addition, patients are all responsive to NSAIDs butnot drugs against neuropathic pain. These observations sup-port inflammatory stimuli that either evoked or exacerbated

Table 2 Quantification of epidermal nerve fiber density in patients with the Arg225Cys mutation in SCN11A

Identificationa Age Distal leg Distal thigh Proximal thigh

Normsb 20–29 11.3 ± 5.9 15.5 ± 5.9 13.2 ± 6.1

II-5 29 18.1 19.1 26.4

II-1 30 13.2 13.5 13.8

II-2 28 8.9 14.8 14.8

II-3 27 13.6 17.5 17.9

II-4 22 13.6 16.3 20.4

Norms 50–59 14.4 ± 5.7 21.4 ± 14.1 17.9 ± 11.2

II-1 53 10.2 14.4 16.5

II-2 51 11.4 20.4 22.3

a Identification used in the pedigree.b Normative values from previously published data. The unit is “nerve fibers/mm surface length of skin.”

Table 3 Nerve conduction studies

Identification

Sensory NCS Motor NCS

DL (ms)/Amp (mV)/CV (m/s) DL (ms)/Amp (mV)/CV (m/s)

Median Ulnar Sural Median Ulnar Peroneal

Normsa 3.5/22.0/50.0 3.5/10.0/50.0 4.4/6.0/40.0 4.4/4.0/49.0 3.3/6.0/49.0 6.5/2.0/44.0

I-1 2.9/36.0/50.0 3.2/17.9/58.0 2.4/9.1/52.0 2.9/7.0/51.0 3.3/8.6/51.6 3.5/2.9/50.0

I-2 3.1/47.0/61.0 4.0/18.7/54.0 3.4/8.3/47.0 2.7/6.7/49.0 3.2/8.1/49.0 3.1/3.2/55.0

Abbreviations: Amp = amplitude; CV = conduction velocity; DL = distal latency; NCS = nerve conduction study.a Lower limit of normal.




pain along with the dysfunctional nociceptive neurons by themutation.

Nociceptive pain in these patients was often described as anaching type pain, which is commonly associated with muscu-loskeletal or joint pain. During their pain episodes, symptomswere commonly described as originating unilaterally with pointtenderness and would slowly, in a matter of minutes to hours,ascend proximally in the affected limb. In the lower extremity,pain was described as ascending to the mid-lateral thigh;however, 2 patients were affected as proximal as the hip.

Of note, 1 patient (II-1) described their pain as havinga burning and tingling sensation on their questionnaire, whichwas mild (1 on scale of 1–5). However, these were not thefeatures in other patients. Otherwise, his pain description waslargely nociceptive in features. Thus, if there were any neu-ropathic features of pain, it was minimal.

Of interest, in our family, the most common exacerbatingfactors for pain was exertion/physical activity and weatherchanges, both of which have been associated with increasedlocalized inflammation. In line with this issue, SCN11Aknockout mice showed impairment of somatic inflammatorypain behavior21,22 Together, these observations also suggestthat pain by the Arg225Cys mutation may involve partic-ipations of other cell types such as inflammatory cells.

The exact pathogenic mechanisms between neurons, glia andinflammatory cells in relation to pain are unknown. Pain mayarise from the distal part of peripheral nerves since pain showsa clear length-dependent pattern in all studied patients withthe Arg225Cys mutation.

Missense mutations in Nav have been associated with painfulsensory neuropathies.1,4,23 This is well expected for Nav1.7and Nav1.8, given the known biology of the 2 types of Nav.However, this association is not well explained in patientswith mutations in Nav1.9 since Nav1.9 is exclusively expressedin nociceptive neurons. Indeed, our patients rarely presentedwith features of neuropathic pain. Some patients with theArg225Cys mutation have even been diagnosed with fibro-myalgia. Furthermore, bothNCS and skin biopsies showed noevidence of peripheral neuropathy. Therefore, the EPS phe-notype with nociceptive pain observed in the present study iswell in line with the known biology of Nav1.9. Clinicians mayadd EPS into their differential diagnosis when dealing withcases presenting no peripheral neuropathy but nociceptivepain. This observation also redirects our attention to addi-tional pain phenotypes that may relate to the genetic alter-ations in SCN11A.

Here, we have described EPS in a Caucasian family with theArg225Cys mutation in SCN11A. Detailed characterization ofthe family enables us to reveal pain likely arising from dys-functional nociceptive neurons of peripheral nerves that lacksdetectable pathologic alterations in large and small nerve

fibers. This pain phenotype is also consistent with the hy-perexcitability observed in dorsal root ganglion neuronsexpressing the mutant Nav1.9.

6 This phenotype is clearlydistinct from neuropathic pain observed in patients withpainful small fiber sensory neuropathy and other missensemutations in Nav.

Author contributionsR. Castoro: participated in conceptualization of the study,data acquisition, analyzed and interpreted data, and draftedthe manuscript. M. Simmons, L. Zhou, V. Ravi, D. Huang,C. Lee, and J. Sergent: acquisition of data. J. Li: conceptual-ized the study, analyzed and interpreted data, drafted themanuscript, and supervised the study.

Study fundingThis research is supported by grants from the NINDS(R01NS066927), the Muscular Dystrophy Association, andthe National Center for Advancing Translational Sciences(UL1TR000445).

DisclosureR. Castoro, M. Simmons, V. Ravi, and D. Huang report nodisclosures. C. Lee has served on the commercial advisoryboard of CSL Behring. J. Sergent has received research sup-port from the NIH. L. Zhou reports no disclosures. J. Li hasserved on the scientific advisory boards of the MuscularDystrophy Association and the Charcot-Marie-Tooth Asso-ciation and serves on the editorial boards of the Journal of thePeripheral Nervous System, Experimental Neurology, Neurology:Neuroimmunology & Neuroinflammation, and Neural Re-generation Research. Full disclosure form information providedby the authors is available with the full text of this article atNeurology.org/NG.


References1. Dib-Hajj SD, Rush AM, Cummins TR, et al. Gain-of-function mutation in Nav1.7 in

familial erythromelalgia induces bursting of sensory neurons. Brain 2005;128:1847–1854.

2. Cox JJ, Reimann F, Nicholas AK, et al. An SCN9A channelopathy causes congenitalinability to experience pain. Nature 2006;444:894–898.

3. Faber CG, Lauria G, Merkies IS, et al. Gain-of-function Nav1.8 mutations in painfulneuropathy. Proc Natl Acad Sci USA 2012;109:19444–19449.

4. Huang J, Han C, Estacion M, et al. Gain-of-function mutations in sodium channelNav1.9 in painful neuropathy. Brain 2014;137:1627–1642.

5. Leipold E, Hanson-Kahn A, Frick M, et al. Cold-aggravated pain in humans caused bya hyperactive NaV1.9 channel mutant. Nat Commun 2015;6:10049.

6. Zhang XY, Wen J, Yang W, et al. Gain-of-Function mutations in SCN11A causefamilial episodic pain. Am J Hum Genet 2013;93:957–966.

7. Okuda H, Noguchi A, Kobayashi H, et al. Infantile pain episodes associated with novelNav1.9 mutations in familial episodic pain syndrome in Japanese families. PLoS One2016;11:e0154827.

8. Han C, Yang Y, Te Morsche RH, et al. Familial gain-of-function Nav1.9 mutation ina painful channelopathy. J Neurol Neurosurg Psychiatry 2017;88:233–240.

9. Leng XR, Qi XH, Zhou YT, Wang YP. Gain-of-function mutation p.Arg225Cys inSCN11A causes familial episodic pain and contributes to essential tremor. J HumGenet 2017;62:641–646.

10. Murphy SM, Herrmann DN, McDermott MP, et al. Reliability of the CMT neu-ropathy score (second version) in Charcot-Marie-Tooth disease. J Peripher Nerv Syst2011;16:191–198.

11. Li J, Krajewski K, Shy ME, Lewis RA. Hereditary neuropathy with liability to pressurepalsy: the electrophysiology fits the name. Neurology 2002;58:1769–1773.

12. Hu B, Arpag S, Zuchner S, Li J. A novel missense mutation of CMT2P alters tran-scription machinery. Ann Neurol 2016;80:834–845.




13. Lauria G, Lombardi R, Borgna M, et al. Intraepidermal nerve fiber density in rat footpad: neuropathologic-neurophysiologic correlation. J Peripher Nerv Syst 2005;10:202–208.

14. Anderson DL. Development of an instrument to measure pain in rheumatoid arthritis:rheumatoid Arthritis Pain Scale (RAPS). Arthritis Rheum 2001;45:317–323.

15. Freynhagen R, Baron R, Gockel U, Tolle TR. painDETECT: a new screeningquestionnaire to identify neuropathic components in patients with back pain. CurrMed Res Opin 2006;22:1911–1920.

16. Fishbain DA, Cole B, Lewis JE, Gao J. What is the evidence that neuropathic pain ispresent in chronic low back pain and soft tissue syndromes? An evidence-basedstructured review. Pain Med 2014;15:4–15.

17. Ahmed S, Magan T, Vargas M, Harrison A, Sofat N. Use of the painDETECT tool inrheumatoid arthritis suggests neuropathic and sensitization components in painreporting. J Pain Res 2014;7:579–588.

18. Lopes MC, Joyce C, Ritchie GR, et al. A combined functional annotation score fornon-synonymous variants. Hum Hered 2012;73:47–51.

19. Aaij R, Adeva B, AdinolfiM, et al. Search for long-lived particles decaying to jet pairs.Eur Phys J C Part Fields 2015;75:152.

20. Bouhassira D, LetanouxM,Hartemann A. Chronic pain with neuropathic characteristicsin diabetic patients: a French cross-sectional study. PLoS One 2013;8:e74195.

21. Priest BT, Murphy BA, Lindia JA, et al. Contribution of the tetrodotoxin-resistantvoltage-gated sodium channel NaV1.9 to sensory transmission and nociceptive be-havior. Proc Natl Acad Sci USA 2005;102:9382–9387.

22. Amaya F, Wang H, Costigan M, et al. The voltage-gated sodium channel Na(v)1.9 isan effector of peripheral inflammatory pain hypersensitivity. J Neurosci 2006;26:12852–12860.

23. Dib-Hajj SD, Cummins TR, Black JA, Waxman SG. Sodium channels in normal andpathological pain. Annu Rev Neurosci 2010;33:325–347.



ARTICLE OPEN ACCESS

Expanding the phenotype of de novo SLC25A4-linked mitochondrial disease to include mildmyopathyMartin S. King, PhD,* Kyle Thompson, PhD,* Sila Hopton, BSc, Langping He, PhD, Edmund R.S. Kunji, PhD,

Robert W. Taylor, PhD, FRCPath, and Xilma R. Ortiz-Gonzalez, MD, PhD


Correspondence

Dr. Ortiz-Gonzalez

[email protected]

AbstractObjectiveTo determine the disease relevance of a novel de novo dominant variant in the SLC25A4 gene,encoding the muscle mitochondrial adenosine diphosphate (ADP)/adenosine triphosphate(ATP) carrier, identified in a child presenting with a previously unreported phenotype of mildchildhood-onset myopathy.

MethodsImmunohistochemical and western blot analysis of the patient’s muscle tissue were used toassay for the evidence of mitochondrial myopathy and for complex I–V protein levels. Todetermine the effect of a putative pathogenic p.Lys33Gln variant on ADP/ATP transport, themutant protein was expressed in Lactococcus lactis and its transport activity was assessed withfused membrane vesicles.

ResultsOur data demonstrate that the heterozygous c.97A>T (p.Lys33Gln) SLC25A4 variant isassociated with classic muscle biopsy findings of mitochondrial myopathy (cytochrome coxidase [COX]-deficient and ragged blue fibers), significantly impaired ADP/ATP transport inLactococcus lactis and decreased complex I, III, and IV protein levels in patient’s skeletal muscle.Nonetheless, the expression levels of the total ADP/ATP carrier (AAC) content in the musclebiopsy was largely unaffected.

ConclusionsThis report further expands the clinical phenotype of de novo dominant SLC25A4mutations toa childhood-onset, mild skeletal myopathy, without evidence of previously reported clinicalfeatures associated with SLC25A4-associated disease, such as cardiomyopathy, encephalopathyor ophthalmoplegia. The most likely reason for the milder disease phenotype is that the overallAAC expression levels were not affected, meaning that expression of the wild-type allele andother isoforms may in part have compensated for the impaired mutant variant.

*These authors contributed equally to the manuscript.

From the Medical Research Council Mitochondrial Biology Unit (M.S.K., E.R.S.K.), University of Cambridge, Wellcome Trust/MRC Building, Cambridge Biomedical Campus, UK;Wellcome Centre for Mitochondrial Research (K.T., S.H., L.H., R.W.D.), Institute of Neuroscience, Newcastle University, UK; and Department of Neurology (X.R.O.), Perelman School ofMedicine, Division of Neurology and Center for Mitochondrial and Epigenomic Medicine, The Children’s Hospital of Philadelphia, University of Pennsylvania.


The Article Processing Charge was funded by the Wellcome Centre for Mitochondrial Research.

This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (CC BY), which permits unrestricted use, distribution, and reproduction in anymedium, provided the original work is properly cited.





http://creativecommons.org/licenses/by/4.0/

SLC25A4 (ANT1 and AAC1) gene mutations cause an in-triguing spectrum of human disease, with dominant mutationsfirst reported in adults with progressive external oph-thalmoplegia (PEO),1 whereas recessive loss-of-functionmutations cause cardiomyopathy and skeletal myopathy.2,3

Furthermore, SLC25A4 de novo dominant mutations canpresent in neonates with lactic acidosis, severe hypotonia, andrespiratory failure.4 Here, we report a patient who presentedat age 2 years with mild weakness and hypotonia and wasfound to have a novel de novo heterozygous SLC25A4 variant(c.97A>T;p.Lys33Gln).

The SLC25A4 gene encodes the mitochondrial AAC1, whichimports ADP into the mitochondrion and exports ATP.5

Humans have 4 AAC isoforms, with AAC1 being specific tothe heart, skeletal muscle, and brain.6 Previous functionalstudies correlating the rate of ADP/ATP exchange withphenotype severity have only partially solved the puzzle. Indominant-acting mutations, the rate of transport does seem tocorrelate with clinical severity, with mutations associated withPEO having higher residual ADP/ATP transport rates com-pared with de novo mutations associated with severe neonataldisease.4 However, loss-of-function recessive mutationspresent later than de novo cases with a predominant cardiacphenotype, despite in vitro functional studies showing es-sentially no measurable ADP/ATP exchange4

Here, we report that the de novo dominant variantp.Lys33Gln in SLC25A4 is clinically associated with mildmyopathy despite significant mitochondrial pathology inmuscle biopsy and functional data showing abolished ADP/ATP exchange in vitro. This case expands the known phe-notype of SLC25A4 disease to include childhood-onset mildskeletal myopathy without evidence of cardiac, brain, orextraocular muscle involvement.

MethodsCase presentationA 2-year old girl presented to the neurology clinic at theChildren’s Hospital of Philadelphia (CHOP) for hypotoniaand mild gross motor delays. Neurologic examination atpresentation was only remarkable for hypotonia and a 1-handed Gower maneuver, suggestive of mild weakness.Family history was unremarkable. Laboratory workup foundelevated creatine kinase (CK) (616 U/L, normal 60–305) andlactic acidosis (3.76 mM, normal 0.8–2.0) levels; therefore,muscle biopsy and subsequent genetic testing were pursued.Clinical testing for nuclear mitochondrial disease genes and

full mitochondrial DNA (mtDNA) sequencing (in blood andmuscle) were obtained via the next-generation sequencingpanel.

Standard protocol approvals, registrations,and patient consentsInformed consent was obtained from the child’s parents toenroll in a human subject’s research protocol approved by theInstitutional Review Board at CHOP.

Human muscle immunohistochemistryand analysisStandard histologic and histochemical analyses were per-formed on 10-μm transversely oriented muscle cryosections.Quadruple immunofluorescence analysis of NDUFB8(complex I) and COXI (complex IV)7 and western blotanalysis3 were performed as previously reported. mtDNAcopy number assessment in muscle was undertaken asdescribed.8

Functional studies of p.Lys33Gln variant inLactococcus lactisThe SLC25A4 gene was cloned into the L. lactis expressionvector pNZ8048 by established procedures,9 and thep.Lys33Gln variant was introduced and confirmed by se-quencing. Growth of L. lactis, membrane isolation, vesiclepreparation, transport assays, and western blot analysis wereperformed as reported.4,9

ResultsPatient resultsHistopathologic assessment of skeletal muscle from the pa-tient demonstrated a mosaic pattern of cytochrome c oxidase(COX) deficiency (figure 1A). Quadruple immunofluores-cence analysis confirmed a mitochondrial defect involvingboth complexes I and IV with some fibers exhibiting normalprotein expression (figure 1B). MtDNA copy number wasdecreased to approximately 40% in patient muscle comparedwith age-matched controls. Western blot analysis showeda slight decrease in AAC protein levels, associated with moremarkedly decreased steady-state protein levels of componentsof respiratory complexes (CI, CIII, and CIV) in patientskeletal muscle (figure 1C).

Genetic testing revealed a heterozygous SLC25A4 variant(c.97A>C, p.Lys33Gln) that was confirmed to have arisen denovo following parental testing. Serial cardiac evaluationsincluding ECG, echocardiogram, and Holter monitoring wereunremarkable from diagnosis at age 2 years to current age of 8

Glossary:AAC = ADP/ATP carrier; ADP = adenosine diphosphate; ATP = adenosine triphosphate; COX = cytochrome c oxidase;CHOP = Children’s Hospital of Philadelphia; CK = creatine kinase; mtDNA = mitochondrial DNA; OXPHOS = oxidativephosphorylation; PEO = progressive external ophthalmoplegia.



years. Neurologic evaluations remain stable, only remarkablefor mild proximal weakness, hyperCKemia, and lactic acidosis,with normal extraocular movements and no cognitiveabnormalities.

Assessment of SLC25A4 p.Lys33Glnvariant functionThe transport mechanism of SLC25A4 involves the disrup-tion and formation of the matrix and cytoplasmic salt bridgenetwork in an alternative way (figure 2A).5 Residue Lys33 inSLC25A4, which is conserved among AAC from fungi, plants,and metazoans (figure 2B), forms a salt bridge with theconserved Asp232 in the matrix network (figure 2, C and D).The p.Lys33Gln mutation would eliminate this interaction, asglutamine is a neutral amino acid residue and too short toform a hydrogen bond (figure 2E). Below the salt bridge isGln37 that forms a highly conserved glutamine brace10 (figure2, B–D), which would also be disrupted by the mutation(figure 2E).

We introduced the p.Lys33Gln mutation into the humanSLC25A4 sequence and expressed it L. lactis membranes. Theuptake of radio-labeled ADP in exchange for loaded ADP was

measured for an empty vector control (figure 3A), wild-typeSLC25A4 (figure 3B), and p.Lys33Gln, using the specific in-hibitor carboxyatractyloside as control (figure 3C). Themutantprotein is expressed to approximately the same levels as wild-type in lactococcal membranes, suggesting that the mutationdoes not affect biogenesis, protein folding, or targeting tothe membrane in this expression system (figure 3D). Thep.Lys33Glnmutant was not able to transport ADP (figure 3E).

DiscussionPreviously, we have functionally characterized the effect of 9pathogenic variants in human SLC25A4 using L. lactis, de-termining residual transport activity compared with wild-typeprotein.4 Broadly speaking, residual transport activities of themutants segregate with the associated clinical phenotype;mutations associated with adult-onset dominant PEO displayhigher residual transport activities (24%–56%) than muta-tions associated with recessive disease, which were effectivelynonfunctional.4 There was also a correlation between theseverity of the clinical phenotype and residual transport ac-tivity of previously documented dominant mutations, with

Figure 1 The de novo p.Lys33Gln mutation leads to OXPHOS defect in muscle

(A) Histopathologic analysis of patient skeletal muscle sections showing hematoxylin and eosin (H&E) staining (far left), COX histochemistry (middle left), SDHhistochemistry (middle right), and sequential COX-SDH histochemistry (far right). Scale bar = 100 μm. (B) Respiratory chain profile following quadruple oxidativephosphorylation immunofluorescence analysis of cryosectionedmuscle from the index case, confirming the presence of fibers lacking complex I (NDUFB8) andcomplex IV (COXI) protein. Each dot represents the measurement from an individual muscle fiber, color coded according to its mitochondrial mass (blue-low,normal-beige, high-orange, and very high–red). Gray dashed lines indicate SD limits for the classification of fibers. Lines next to x- and y-axes represent the levels(SDs from the average of control fibers after normalization to porin/VDAC1 levels; _z= Z-score, see Methods section of Rocha et al. 2015 for full description ofstatistics7) ofNDUFB8andCOX1, respectively (beige=normal [>−3], light beige= intermediatepositive [−3 to−4.5], light purple = intermediate negative [−4.5 to−6],and purple = deficient [<−6]). Bold dotted lines indicate the mean expression level observed in respiratory normal fibers. (C) Western blot analysis of AAC andOXPHOS complex subunits on control andpatient skeletalmuscle samples. AAC= ADP/ATPcarrier; ADP= adenosinediphosphate; ATP = adenosine triphosphate;COX = cytochrome c oxidase; GADPH = Glyceraldehyde 3-phosphate dehydrogenase; OXPHOS = oxidative phosphorylation; SDH = succinate dehydrogenase.



Figure 2 The p.Lys33Gln mutation eliminates a conserved salt bridge interaction of the matrix network

(A)Transport cycleof themitochondrialAACSLC25A4.Thedisruptionandformationof salt bridgesbetweenpositively (blue)andnegatively (red) chargedresiduesof thecytoplasmic andmatrix networks, top and bottom respectively, change the access of the substrates to the central substrate binding site, indicated by a hexagon, fromthe intermembranespace (i) andmatrix (m) sideof themembrane. The importedADP is shown ingreensphere representationand theexportedATP incyan. (B)Aminoacid sequence alignment of AAC1 sequences of fungi, plants, and metazoan, showing that Lys33, Gln37, and Asp232 are highly conserved amino acid residues. (C)Lateral view of the human ADP/ATP carrier from the membrane, showing the residues of the matrix and cytoplasmic networks (blue and red sticks) and substratebinding site (green sticks, hexagon). ADP (light blue ball and stick) and the glutamine brace (light green stick) are also shown. Residue Lys33 that ismutated is shown inyellow. (D) Cytoplasmic view of the carrier showing only the residues of the matrix salt bridge network of SLC25A4 (blue and red sticks). (E) As (D), except for thep.Lys33Glnmutation, which is shown inmagenta. The ionic interactions (black dash lines) are indicated with plus andminus signs. Themodel of human SLC25A4wasgenerated in SwissModel, using the structure of the closely related bovinemitochondrial AAC as template (PDB file: 1OKC). Adapted fromFigure 3 in reference 4. AAC =ADP/ATP carrier; ADP = adenosine diphosphate; ATP = adenosine triphosphate.



mutations associated with inherited PEO showing higheractivity than mutations associated with de novo severe neo-natal presentation.3

It remains unclear exactly why the presently documented casewith the de novo dominant c.97A>T(p.Lys33Gln) SLC25A4variant presents with a much milder clinical phenotype thanthe other de novo cases3 despite the p.Lys33Gln mutantprotein revealing essentially null ADP/ATP transport activity(figure 3E). This case further supports that residual ADP/ATP transport activity of mutant AAC1 is only part of thepuzzle. In vivo, there are 3 other AAC isoforms with varyingexpression levels in different tissues. Also, since the de novovariants are heterozygous, the relative expression ratios ofwild-type vs mutant AAC1, as well as other isoforms, will havean effect on the overall transport capacity. Their relative ex-pression levels of carriers cannot be assessed from messengerRNA levels, as there are many poorly characterized steps in-volved in their biogenesis and turnover, and the available anti-bodies are unable to distinguish between the various isoforms.

Despite these issues, there is a clear correlation between theseverity of OXPHOS dysfunction in skeletal muscle and theclinical phenotype between the current and previouslyreported de novo mutations. Patients with p.Arg80His orp.Arg235Gly mutations, showing severe neonatal pre-sentations, had markedly decreased levels of total AAC(<30%),3 whereas there is only a slight decrease in the currentpatient (figure 1C). Similarly, the mosaic pattern of OXPHOSdeficiency (figure 1, A and B) is not seen in previouslydocumented de novo cases, where steady state protein levelsof various OXPHOS components were completely

undetectable.3 These data suggest that the clinical severity ofde novo dominant mutations in SLC25A4 could be explainedby the relative expression of wild-type and mutant alleles andthe expression of other AAC isoforms, rather than thetransport activity of the mutant variant alone.

This report expands the clinical phenotype for SLC25A4-associated mitochondrial disease, with the mildest childhood-onset presentation to date.

Author contributionsX.R. Ortiz-Gonzalez, R.W. Taylor, and E.R.S. Kunji: studyconcept and design and study supervision. R.W. Taylor andE.R.S. Kunji: study funding. M.S. King, K. Thompson, S.Hopton, and L. He: experimental data acquisition and anal-ysis. M.S. King and K. Thompson: statistical analysis for ex-perimental data. X.R. Ortiz-Gonzalez: clinical dataacquisition. X.R. Ortiz-Gonzalez, R.W. Taylor, E.R.S. Kunji,M.S. King, and K. Thompson: data interpretation and criticalrevision of the manuscript for important intellectual content.

Study fundingR.W. Taylor is supported by the Wellcome Centre for Mito-chondrial Research (203105/Z/16/Z), the Medical ResearchCouncil (MRC) Centre for Translational Research in Neuro-muscular Disease, Mitochondrial Disease Patient Cohort (UK)(G0800674), the Lily Foundation, the UK NIHR BiomedicalResearch Centre for Ageing and Age-related disease award tothe Newcastle upon Tyne Foundation Hospitals NHS Trust,the MRC/EPSRC Molecular Pathology Node, and the UKNHS Highly Specialised Service for Rare Mitochondrial Dis-orders of Adults and Children. The research of M.S. King and

Figure 3 Transport activity of human SLC25A4 and SLC25A4 p.Lys33Gln

Transport of [14C]-labeled ADP into vesicles of L. lactis membranes expressing (A) an empty vector control, (B) SLC25A4, or (C) SLC25A4 p.Lys33Gln in thepresence (black symbols) or absence (white symbols) of 20 μM carboxyatractyloside. Transport was initiated by the addition of 5 μM [14C]-ADP and wasterminated by filtration and washing at the indicated time intervals. The data are represented by the average and SD of 4 assays. (D) Expression levels of wildtype and p.Lys33Gln determined by western blot analysis. (E) Transport rate of the empty vector control, human SLC25A4, and p.Lys33Gln, corrected forbackground binding. AAC = ADP/ATP carrier; ADP = adenosine diphosphate; ATP = adenosine triphosphate.



E.R.S. Kunji was funded by the MRC programme grant MC_UU_00015/1. X.R. Ortiz-Gonzalez is supported by the RobertWood Johnson Foundation Harold Amos Faculty De-velopment Award and the NIH (5K12NS049453-08).

DisclosuresK. Thompson, S. Hopton, L. He, E.R.S. Kunji, M.S. King, andR.W. Taylor report no disclosures. X.R. Ortiz-Gonzalez has re-ceived funding for travel/speaker honoraria from BioMarin andhas received research support from the NINDS and the RobertWood Johnson Foundation. Full disclosure form informationprovided by the authors is available with the full text of this articleat Neurology.org/NG.


References1. Kaukonen J, Juselius JK, Tiranti V, et al. Role of adenine nucleotide translocator 1 in

mtDNA maintenance. Science 2000;289:782–785.

2. Palmieri L, Alberio S, Pisano I, et al. Complete loss-of-function of the heart/muscle-specific adenine nucleotide translocator is associated with mitochondrial myopathyand cardiomyopathy. Hum Mol Genet 2005;14:3079–3088.

3. Strauss KA, Dubiner L, Simon M, et al. Severity of cardiomyopathy associated withadenine nucleotide translocator-1 deficiency correlates with mtDNA haplogroup.Proc Natl Acad Sci USA 2013;110:3453–3458.

4. Thompson K, Majd H, Dallabona C, et al. Recurrent de novo dominant mutations inSLC25A4 cause severe early-onset mitochondrial disease and loss of mitochondrialDNA copy number. Am J Hum Genet 2016;99:860–876.

5. Kunji ER, Aleksandrova A, King MS, et al. The transport mechanism of the mito-chondrial ADP/ATP carrier. Biochim Biophys Acta 2016;1863:2379–2393.

6. Dolce V, Scarcia P, Iacopetta D, Palmieri F. A fourth ADP/ATP carrier isoform inman: identification, bacterial expression, functional characterization and tissue dis-tribution. FEBS Lett 2005;579:633–637.

7. Rocha MC, Grady JP, Grunewald A, et al. A novel immunofluorescent assay toinvestigate oxidative phosphorylation deficiency in mitochondrial myopathy: un-derstanding mechanisms and improving diagnosis. Sci Rep 2015;5:15037.

8. Blakely E, He L, Gardner JL, et al. Novel mutations in the TK2 gene associated withfatal mitochondrial DNA depletion myopathy. Neuromuscul Disord 2008;18:557–560.

9. King MS, Boes C, Kunji ER. Membrane protein expression in Lactococcus lactis.Methods Enzymol 2015;556:77–97.

10. Ruprecht JJ, Hellawell AM, Harding M, Crichton PG, Mccoy AJ, Kunji ERS.Structures of yeast mitochondrial ADP/ATP carriers support a domain-basedalternating-access transport mechanism. Proc Natl Acad Sci USA 2014;111:E426–E434.




ARTICLE OPEN ACCESS

Carey-Fineman-Ziter syndrome with mutationsin the myomaker gene and muscle fiberhypertrophyCarola Hedberg-Oldfors, PhD, Christopher Lindberg, MD, PhD, and Anders Oldfors, MD, PhD


Correspondence

Dr. Hedberg-Oldfors

[email protected]

AbstractObjectiveTo describe the long-term clinical follow-up in 3 siblings with Carey-Fineman-Ziter syndrome(CFZS), a form of congenital myopathy with a novel mutation in the myomaker gene(MYMK).

MethodsWe performed clinical investigations, repeat muscle biopsy in 2 of the siblings at ages rangingfrom 11 months to 18 years, and whole-genome sequencing.

ResultsAll the siblings had a marked and characteristic facial weakness and variable dysmorphicfeatures affecting the face, hands, and feet, and short stature. They had experienced musclehypotonia and generalized muscle weakness since early childhood. The muscle biopsiesrevealed, as the only major abnormality at all ages, a marked hypertrophy of both type 1 andtype 2 fibers with more than twice the diameter of that in age-matched controls. Geneticanalysis revealed biallelic mutations in theMYMK gene, a novel c.235T>C; p.(Trp79Arg), andthe previously described c.271C>A; p.(Pro91Thr).

ConclusionsOur study expands the genetic and clinical spectrum of MYMK mutations and CFZS. Themarked muscle fiber hypertrophy identified from early childhood, despite apparently normalmuscle bulk, indicates that defective fusion of myoblasts during embryonic muscle de-velopment results in a reduced number of muscle fibers with compensatory hypertrophy andmuscle weakness.

From the Department of Pathology and Genetics (C.H.-O., A.O.), Sahlgrenska Academy, University of Gothenburg, and Department of Neurology (C.L.), Neuromuscular Center,Sahlgrenska University Hospital, Gothenburg, Sweden.


The Article Processing Charge was funded by the Swedish Research Council.







Carey-Fineman-Ziter syndrome (CFZS, MIM 254940) is anautosomal recessive inherited disorder. Clinically, patients aredescribed as having nonprogressive congenital myopathy withmarked facial weakness, together with other clinical attributessuch as Moebius and Pierre Robin sequence, facial abnor-malities, and growth delay. Recently, autosomal recessivemutations in the gene myomaker (MYMK/TMEM8C) werefound to be associated with CFZS.1

MYMK is a plasma transmembrane protein and is necessaryfor the fusion of mononuclear myoblasts to multinucleate

myocytes in the skeletal muscle.2–6 In this article, we describea family with 3 siblings affected with CFZS due to biallelicmutations in MYMK.

MethodsPatientsThis Swedish family had 3 affected children with healthyunrelated parents (figure 1L). A summary of results from theclinical investigations is given in table.

Figure 1 Clinical features and pedigree

(A–D) Facial photographs of individuals II:1 and II:3 demonstrating facial weakness, slight ptosis, broad nasal tip, and in II:1, also slight retrognathia and somedegree of epicanthus. (E–G and M) All 3 siblings had small hands with brachydactyly and short tapering fingers, with the thumbs situated proximally on thehands. (K) Individual II:1 also had camptodactyly. (H–J) They had small feet with short toes and sandal gap deformity. (I, J, and M) Individuals II:2 and II:3appeared puffy on the dorsal aspects of the hands and feet. (L) Pedigree.

GlossaryCFZS = Carey-Fineman-Ziter syndrome; MYMK = myomaker.



All siblings presented at birth with hypotonia and feedingdifficulties. They started to walk at around 15 months of age,but they used the Gower maneuver to come to an uprightposition. Individual II:1 was operated for cryptorchidism at 6weeks of age. All siblings had Achilles tendon contractures,but only individual II:1 had been operated with lengthening ofboth Achilles tendons at the age of 9 years. They werehypermobile in most joints, including the elbows. All usedglasses, individual II:1 since 13 years of age due to hyperopia,

II:2 since 6 years of age, and II.3 since 26 years of age due tomyopia. Individual II.2 had a hearing impairment involvinglow-frequency tones and started to use a hearing aid at 9 yearsof age. All had normal cognitive function.

At the time of genetic diagnosis in 2017, they all exhibited severefacial weakness together with dysmorphic features (figure 1A–KandM, table). They showed slight bilateral ptosis, individual II:1had slight limitation of eye abduction, and individuals II:1 and II:

Table Clinical findings

Individual II:1 II:2 II:3

Sex Male Female Female

Age at last examination (y) 37 31 28

Descent Swedish Swedish Swedish

Height (cm)a 167 158 160

Phenotype

Facial weakness Severe Severe Severe

Broad nasal tip Yes Yes Yes

Retrognathia Slight Marked Absent

Dysarthria Moderate Severe Moderate

Tongue/hypoglossia Small and short Weakness Weakness (slight)

Palatal weakness Yes Yes Yes (slightly)

Pectoralis hypoplasia Yes No No

Scoliosis No Yes (15°–25°) No

Muscle weakness

Neck flexors None Severe Severe

Hands Moderate Severe None

Proximal upper extremities Slight Slight None

Proximal lower extremities Slight Slight Slight

Distal lower extremities Slight Slight Slight

Laboratory testing

EMG NA Myopathic (age 17 y) NA

Forced vital capacity(L/% predicted)

2.15/49 2.1/66b 2.9/88

Electrocardiography Normal (age 36 y) AV block grade II (age 30 y) Normal (age 28 y)

Echocardiography Normal (age 36 y) Normal (age 28 y) Normal (age 20 y)

Creatine kinase (IU/L) 840 (<294) 1,200 (<210) 294 (<210)

Muscle pathology Hypertrophy of type 1 and 2 fibers Hypertrophy of type 1 and 2 fibers NA

DNA analysis MYMK Compound heterozygous; c.235T>C;c.271C>A

Compound heterozygous; c.235T>C;c.271C>A

Compound heterozygous; c.235T>C;c.271C>A

Predicted protein change p.(Trp79Arg); p.(Pro91Thr) p.(Trp79Arg); p.(Pro91Thr) p.(Trp79Arg); p.(Pro91Thr)

Abbreviations: MYMK = myomaker; NA, not available.a Parental height: mother 163 cm and father 183 cm.b Technical difficulties.



2 had a slight defect in elevation of the eyes. Individual II:3 hada bifid uvula and an exostosis in the midline of the palate.

Standard protocol approvals, registrations,and patient consentsThe study complied with the Declaration of Helsinki, andinformed consent was obtained from the patients.

Morphological analysisOpen skeletal muscle biopsies from the vastus lateralis of thequadricepsmuscle were performed. Specimens were snap-frozenin liquid propane chilled with liquid nitrogen for cryostatsectioning and histochemistry.7 Morphometric analyses oftype 1 and type 2 muscle fibers were performed on adenosinetriphosphatase–stained sections (pH 9.4).

Molecular genetic analysisWhole-genome sequencing was performed on genomic DNAfrom individual II:2 using the TruSeq PCR free library prepara-tion kit, and the Illumina HiSeq X platform was used forsequencing (Illumina, San Diego, CA). The paired-end readswere aligned to the reference genome (hg19) using the CLCBiomedical Genomics workbench (Qiagen, Hilden, Germany).Data were analyzed using Ingenuity Variant Analysis(ingenuity.com/products/variant-analysis) (Qiagen). Weperformed a search for compound heterozygous or homozygous

variants that were predicted to be damaging using SIFT (sortingintolerant from tolerant algorithm), Mutation Taster, andPolyPhen2, analyzing whether the variants affected con-served amino acids and whether they are uncommon inthe population (using 1000 Genomes [1000genomes.org/],NHLBI Exome Sequencing Project [evs.gs.washington.edu/EVS/], and the Genome Aggregation Database [gnomAD][gnomad.broadinstitute.org/]), to reduce the number of var-iants. Sanger sequencing was used for confirmation and analysisof mutations identified in individuals I:2, II:1, and II:3.

ResultsMorphological analysisMuscle biopsy at the ages of 4.4 years and 6.3 years in in-dividual II:1 and at the ages of 11 months and 18 years inindividual II:2 showed (on all occasions) marked hypertrophyof both type 1 and type 2 fibers (figure 2, A–C). The medianmuscle fiber diameter was more than twice as large as in age-matched controls (figure 2D). A few internalized nuclei wereobserved in individual II:2 at 18 years of age, and at that age,there were also some irregularities in the intermyofibrillarnetwork as revealed by Nicotinamide adenine dinucleotide(NADH)-tetrazolium reductase staining (figure 2, B–C).There was no increase in interstitial connective tissue, and nonecrotic or regenerating muscle fibers were observed.

Figure 2 Muscle histology and morphometric analysis demonstrating muscle fiber hypertrophy, occasional internalizednuclei and slight irregularity in the muscle fiber intermyofibrillar network

(A) Muscle fiber hypertrophy affecting both type 1 fibers(dark) and type 2 fibers (unstained) (adenosine triphos-phatase, pH 4.3). (B) Muscle fiber hypertrophy and someinternalized nuclei (hematoxylin and eosin). (C) Muscle fi-ber hypertrophy and patchy irregularities of the inter-myofibrillar network (NADH-tetrazolium reductase). (D)Results from morphometric analysis. The lesser innerdiameters (micrometers) of type 1 and type 2 fibers of 100adjacent muscle fibers were measured in 4 differentmuscle biopsies from individuals II:1 and II:2 at differentages. Measurement of the lesser inner diameter avoids therisk of errors caused by oblique sections andwas also usedin the previously published measurements on controlsthat were used as references in this study.7 The results aregiven as medians, quartiles, and range. The median is atthe border between the red and green boxes. The top ofthe green box is the 75th percentile of the sample, and thebottom of the red box is the 25th percentile. Normal meandiameters in age-matched controls are indicated withasterisks. Child controls included both males and females,and adult controls included females only. 1 = type 1 fibers;2 = type 2 fibers.


http://www.ingenuity.com/products/variant-analysis

http://www.1000genomes.org/

http://evs.gs.washington.edu/EVS/

http://evs.gs.washington.edu/EVS/

http://gnomad.broadinstitute.org/


Molecular genetic analysisIndividuals II:1, II:2, and II:3 were all biallelic for2 missense mutations in the MYMK gene (TMEM8C)(figure 3C). One was a missense mutation (c.235T>C;p.(Trp79Arg)) not previously described. The secondmutation was a previously described missense mutation

(c.271C>A; p.(Pro91Thr)) associated with CFZS (figure3, A–D).1

Other candidate genes associated with myopathies accordingto the NMD Gene Table 2017 (musclegenetable.fr/) wereexcluded.8 Sanger sequencing confirmed both mutations in all

Figure 3 Molecular genetics

(A) Illustration showing all the pathogenic mutations identified in myomaker (MYMK). The novel mutation c.235T>C (M1) and the previously describedmutation c.271C>A (M2) identified in the 3 siblings in this study are indicated (NM_001080483.2). (B) The 2D structure of MYMK showing the location ofmutations (illustration adapted from reference 1). The p.Trp79Arg (M1) is located in one of the transmembrane domains and changes the large, nonpolartryptophan (Trp/W) residue at position 79 to the large, positively charged arginine (Arg/R) residue―thus creating a shift in polarity from nonpolar to positivelycharged. (C) Pedigree of the family. Filled squares and circles indicate individuals with Carey-Finman-Ziter syndrome. Asterisk indicates the individualanalyzed by whole-genome sequencing. (D) Illustration showing the evolutionary conservation of amino acids, the p.Trp79Arg (M1) with a conservationphyloP p value of 5.834E−5. The blue bars in the upper part show the residues that were found to be mutated.


http://www.musclegenetable.fr/


3 siblings, and the mother was a heterozygous carrier of themutation c.271C>A; p.(Pro91Thr).

DiscussionHere, we have described 3 siblings with CFZS due tobiallelic missense mutations in MYMK, adding to the re-cent description of 8 individuals with MYMK mutations.1

Our cases show the same clinical spectrum of muscleweakness, where profound facial muscle weakness is theclinical hallmark of the disease. Severe neck flexor weak-ness, short stature, small hands and feet, and hyperlaxity ofmost joints in combination with distal contractures areadditional clues to the correct diagnosis. Restrictive lungdisease was described in the previous series of patients, andin 2 of our patients, a limitation in forced vital capacity wasfound, but to date, none of them have shown any clinicalsymptoms of hypoventilation. One of our patients hadhearing impairment, which has also been described pre-viously in CFZS.9

All siblings in this report were biallelic forMYMKmissensemutations: p.(Pro91Thr) and p.(Trp79Arg). Thep.(Pro91Thr) mutation has been identified in 4 of 5 ap-parently unrelated families already described.1 This mu-tation has also been identified in 328 of 276,516 alleles, butonly in heterozygous individuals in the gnomAD database.The second mutation, p.(Trp79Arg), was not present ingnomAD, is predicted by in silico programs to be delete-rious, and alters a phylogenetically highly conservedresidue.

Our patients had had muscle hypotonia and weakness sinceearly childhood and had been considered to have a congen-ital myopathy. The 4 muscle biopsies performed in 2 of theindividuals at various ages, with the earliest biopsy per-formed at 11 months of age and the latest at 18 years,revealed a remarkable hypertrophy of both type 1 and type 2muscle fibers as the only important pathology at all timepoints. Results from 1 muscle biopsy in a previous reportdemonstrated marked muscle fiber hypertrophy, but mainlyaffecting the type 2 fibers.1 This marked hypertrophy with-out any apparent increase in muscle bulk indicates a re-duction in the number of muscle fibers to approximately onequarter of normal in our patients. One may postulate thatmuscle fiber hypertrophy, also seen in other neuromusculardiseases such as spinal muscular atrophy,10 compensates forthe loss of functioning muscle fibers. At present, the mark-edly reduced number of muscle fibers seen in our CFZSpatients from early childhood onward remains unexplained.However, as MYMK is essential for the fusion of mono-nuclear myoblasts to multinucleate myocytes during em-bryonic muscle development,4,11 one may speculate thata reduced MYMK function due to mutations inMYMK mayplay a role. NewbornMYMK−/−mice have profound muscleweakness, lack multinucleated muscle fibers, and die before

postnatal day 7, suggesting that the lack of MYMK is in-compatible with life.4 This suggests that the mutationsidentified in patients with CFZS in humans give rise topartially functioning MYMK protein, which has also beensupported by functional studies.1 The p.(Pro91Thr) muta-tion has some residual function and is therefore compatiblewith life in combination with null alleles.1 The individualsdescribed in this report carry the p.(Pro91Thr) mutation on1 allele, and most likely, the combination with the mutationon the other allele, p.(Trp79Arg), leads to a severely reducedfunction of the MYMK protein.

Our study provides additional evidence for the associationbetween CFZS and MYMK mutation. The marked musclefiber hypertrophy despite the normal or reduced muscle bulkidentified from early childhood indicates that defective fusionof myoblasts during embryonic muscle development results inreduced numbers of muscle fibers, with consequent hyper-trophy and muscle weakness.

Author contributionsC.Hedberg-Oldfors performed the genetic analysis, interpretedthe data, and drafted and revised the manuscript. C. Lindbergperformed the clinical examinations, interpreted the data andrevised the manuscript. A. Oldfors performed the muscle pa-thology analysis and drafted and revised the manuscript.

AcknowledgmentThe authors thank Brith Leidvik for technical assistance andthe patients for their support.

Study fundingThis study was supported by the Research Fund for Neuro-muscular Disorders in West Sweden and by the SwedishResearch Council (project no. 2012-02014).

DisclosureC. Hedberg-Oldfors reports no disclosures. C. Lindberg hasreceived research support from the following foundations:Sweden Muskelfonden and Neuroforbundet. A. Oldfors serveson the editorial board of Neuromuscular Disorders; receivespublishing royalties fromElsevier publishing for “Muscle Biopsy:A Practical Approach”; and has received research funding fromthe Swedish Research Council and the World Muscle Society.Full disclosure form information provided by the authors isavailable with the full text of this article at Neurology.org/NG.


References1. Di Gioia SA, Connors S, Matsunami N, et al. A defect in myoblast fusion underlies

Carey-Fineman-Ziter syndrome. Nat Commun 2017;8:16077.2. Gamage DG, Leikina E, Quinn ME, Ratinov A, Chernomordik LV, Millay DP.

Insights into the localization and function of myomaker during myoblast fusion. J BiolChem 2017;292:17272–17289.

3. Goh Q, Millay DP. Requirement of myomaker-mediated stem cell fusion for skeletalmuscle hypertrophy. Elife 2017;6:e20007.

4. Millay DP, O’Rourke JR, Sutherland LB, et al. Myomaker is a membrane activator ofmyoblast fusion and muscle formation. Nature 2013;499:301–305.

5. Millay DP, Sutherland LB, Bassel-Duby R, Olson EN. Myomaker is essential formuscle regeneration. Genes Dev 2014;28:1641–1646.




6. Mitani Y, Vagnozzi RJ, Millay DP. In vivo myomaker-mediated heterologous fusionand nuclear reprogramming. FASEB J 2017;31:400–411.

7. Dubowitz V, SewryCA,Oldfors A.Muscle Biopsy: A Practical Approach. Philadelphia, PA:Elsevier; 2013:1–592.

8. Kaplan JC, Hamroun D, Rivier F, Bonne G. The 2017 version of thegene table of neuromuscular disorders. Neuromuscul Disord 2016;26:895–929.

9. Carey JC. The Carey-Fineman-Ziter syndrome: follow-up of the original siblings andcomments on pathogenesis. Am J Med Genet A 2004;127A:294–297.

10. Kingma DW, Feeback DL, Marks WA, Bobele GB, Leech RW, Brumback RA. Se-lective type II muscle fiber hypertrophy in severe infantile spinal muscular atrophy.J Child Neurol 1991;6:329–334.

11. Chen EH, Olson EN. Unveiling the mechanisms of cell-cell fusion. Science 2005;308:369–373.



ARTICLE OPEN ACCESS

Axon reflex–mediated vasodilation is reduced inproportion to disease severity in TTR-FAPIrene Calero-Romero, MD, Marc R. Suter, MD, Bernard Waeber, MD, Francois Feihl, MD,*

and Thierry Kuntzer, MD*


Correspondence

Dr. Kuntzer

[email protected]

AbstractObjectiveTo evaluate the area of the vascular flare in familial amyloid polyneuropathy (FAP).

MethodsHealthy controls and patients with genetically confirmed FAP were prospectively examined, onthe upper and lower limbs, for thermal sensitivity (Medoc TSA-II thermal analyzer) and foraxon reflex-mediated flare. The latter was induced by iontophoresis of histamine on the forearmand leg on 2 different visits. We used laser Doppler imaging (LDI) to measure the flare area(LDIflare).

ResultsSix patients had FAP of variable severity; 1 had generalized analgesia secondary to leprosy (usedas a positive control). The median Neurological Impairment Score–Lower Limbs (NIS-LLs)was 6 (0–27). The warmth detection thresholds in the feet were higher in patients (median 43°,interquartile range 39.0°–47.6°) compared with controls (37.4°, 35.3°–39.2°), indicating smallfiber impairment. On the leg, LDIflare was smaller in the patients on 2 consecutive visits(controls: median 13.0 and 13.3 cm2, IQR 9.7–22.8 and 8.3–16.9; patients 6.9 and 8.0 cm2,2.6–10.8 and 6.4–12.1; p = 0.011). LDIflare on the leg was correlated with NIS-LL (Spearmanrank correlation 0.73, p = 0.09 on the first visit; Spearman rank correlation 0.85, p = 0.03 on thesecond visit).

ConclusionsOur study underscores that histamine-induced axon reflex–mediated vascular flare on the leg isreduced in proportion to disease severity in patients with FAP.

*Co-senior last authors.

From the Division of Clinical Pathophysiology (I.C.-R., B.W., F.F.), Department of Anaesthesiology, Pain Centre (M.R.S.), and Nerve-Muscle Unit, Neurology Service, Department ofClinical Neurosciences (T.K.), Lausanne University Hospital (CHUV), University of Lausanne, Switzerland.









Familial amyloid polyneuropathies (FAPs) are a group of life-threatening multisystem disorders transmitted as an autoso-mal dominant trait. Nerve lesions are induced by deposits ofamyloid fibrils most commonly due to a transthyretin (TTR)gene mutation. The most frequently identified cause is theTTR Val30Met mutation.1,2 TTR-FAP typically causes earlylength-dependent small fiber polyneuropathy with alterationof temperature and pain sensations including neuropathicpain. This neurologic deficit typically points to involvement ofunmyelinated and small myelinated fibers.2,3 Diagnosis ofC-fiber impairment is challenging; however, the axon reflexcan be assessed by objective approaches.4,5 The axon reflex isa physiologic phenomenon caused by the excitation of dermalC-fibers. Nerve impulses invade peripheral branches anti-dromically causing vascular dilatation, the so-called axon re-flex flare reaction.4 For a long time, this flare reaction wasconsidered unreliable as a clinical tool, making thermotestingthe only recognized biomarker despite its subjectivity.1

However, recent studies using laser Doppler imaging (LDI)have demonstrated a correlation between the reduced flaresize in small fiber neuropathies6 and nerve fiber density in skinbiopsies, which is also reduced.5,7

Our aims were to establish (1) whether the surface area ofneurogenic flare triggered by a standard stimulus could bereliably and reproducibly measured by LDI in healthy controlsand patients with TTR-FAP, (2) whether this method coulddiscriminate between the 2 groups of participants, and (3)whether the surface area of the flare in patients was reduced inproportion to TTR-FAP disease severity.

MethodsStandard protocol approvals, registrations,and patient consentsThis project was approved by the Swiss Ethics Committee onresearch involving humans. All participants provided writteninformed consent.

ParticipantsSeven patients and 10 healthy controls were recruited pro-spectively for this study between spring and fall 2015. Allcontrols were examined to ascertain their good health statusincluding absence of skin disorders.

Patients were recruited from our own registry based on 2inclusion criteria: the presence of neuropathic symptoms(small fiber with or without large fiber involvement) and an

amyloid TTR (ATTR) mutation confirmed genetically.Patients with liver transplantation could be included. Allpatients underwent systematic interview and examination bythe same neurologist (T.K.) and had a thorough laboratoryand imaging workup. Clinical symptoms were assessed by theNeurological Symptom Score (NSS) that includes motor,sensory, and autonomic items.8 The examination performedin a temperature-controlled room included testing musclestrength, tendon reflexes, and sensory function testing in-cluding light touch, pinprick, vibration, joint position sense,and thermal sensation to calculate the Neurological Impair-ment Score–Lower Limbs (NIS-LLs).9 ATTR patients wereclassified according to the Clinical Staging of TTR-FAP, thePolyNeuropathy Disability (PND) score, and the Portugueseseverity classification of TTR-FAP.1

Investigation techniquesThe included patients underwent standard electrophysiologicrecordings of median, sural, and fibular nerve conduction onboth sides. Nerve conduction studies were considered ab-normal if either nerve conduction velocity or amplitude of thesensory nerve action potentials or compound motor actionpotentials were abnormal (normal values from our de-partment were used).

Quantitative sensitivity testingThermotesting was performed in a temperature-controlledroom using a warm and cold perception threshold TSA-IINeuroSensory Analyzer (Medoc Advanced Medical Systems,Durham, NC). The established method of limits algorithm asreported from the German Research Network on Neuro-pathic Pain was also used.10 In short, a thermode of 9 cm2 wasattached to the skin over the dorsum of the hand and footbilaterally, and temperature ramps of 1°C/s were appliedstarting from 32°C. Participants were asked to press a buttonwhen they felt a cold or warm sensation. The investigatormemorized instructions so that she could present them ina uniform fashion to each participant.

Laser Doppler imagingThe spatial distribution of skin blood flow (SkBF) wasrecorded with a LDI (software version 5.01; Moor Instru-ments, Axminster, United Kingdom) using a scanning laserbeam at a wavelength of 633 nm, as previously described.11,12

Histamine induced axon reflexLDI was used to quantitate the flare response induced by theiontophoresis of histamine (LDIflare). This response is neu-rologically mediated as it is blocked by surface anesthesia, as

GlossaryATTR = amyloid transthyretin; AUC = area under the curve; FAP = familial amyloid polyneuropathy; LDI = laser Dopplerimaging; NIS-LL = Neurological Impairment Score—Lower Limbs; NSS = Neurological Symptom Score; PND =polyneuropathy disability; QST = quantitative sensitivity testing; ROC = receiver operating characteristic; SkBF = skin bloodflow; TTR = transthyretin.



previously shown,13 and verified by us in preliminary experi-ments in healthy controls (data not shown).

Methodological detailsA 3.5 × 4.5 cm zone of skin devoid of visible superficial veinswas chosen on the volar face of the forearm, or on the lateralside of the calf, from which a baseline LDI scan was obtained.A ring-shaped chamber made of rubber, with an internal di-ameter of 0.8 mm fitted with a copper electrode, was thenfixed to the center of the zone (using double-sided tape) andfilled to the rim with 1% histamine in deionized water. Areference ECG electrode was placed similarly, nearby. Bothelectrodes were connected to a current source (MIC-1eiontophoresis controller; Moor Instruments). A direct currentof either 20 or 100 μA was applied for 60 seconds. As hista-mine in solution is negatively charged, the iontophoresischamber was on the positive end of the controller. Oncecurrent delivery was terminated, the chamber was removed,the skin dried, and the area directly exposed to the histaminesolution was masked with a circular piece of opaque tape. Fiveconsecutive LDI scans of the zone were then acquired ata frequency of one every 2 minutes. For each of them, thespatial extension of SkBF increase (flare area) was de-termined. The maximal flare area expressed in cm2 was con-sidered as quantification of the axon reflex within the probedskin zone. Each recorded image consisted of approximately20,000 pixels, each with an attached flux value. Within thepost-iontophoresis scans, any pixel whose value exceededa specific threshold was considered part of the flare area. Thechosen threshold was the 99th percentile of all pixel values inthe baseline image.

Because of its potential confounding effect, surface tempera-ture within the chosen skin zone was systematically recordedwith a thermocouple in the minute before iontophoresis.Heat-conducting paste was used to ensure proper thermalcontact between the skin and the thermocouple.

ProtocolEach participant was examined during 3 successive visitsseparated by at least 3 days. Visits 1 and 2 were for recordingthe histamine flare at various locations, whereas quantitativesensitivity testing (QST) was performed on visit 3. The sameinvestigator (the first author, I.C.-R.) performed all exami-nations between 8 and 12 AM, in a quiet temperature-controlled room (22°C–24°C), with the participant lyingcomfortably on a hospital bed.

On visit 1, 4 pairs of skin zones were chosen, 1 on each forearmand 1 on each calf. Zones within each pair were at least 10 cmapart and were both allocated to receive histamine iontopho-resis, with a current intensity of 20 μA on 1 zone and 100 μA onthe other. On the upper and lower limbs, treated zones at bothintensities were placed as symmetrically as possible and theirlocation recorded on transparent film. A total of 8 histamineflares were thus recorded: 2 intensities on each of the 4 limbs.The order of rotation between limbs and for intensities on

limbs was randomized. The whole procedure including setup,acquisition of baseline, iontophoresis, acquisition of post-iontophoresis images required approximately 30 minutes,making for a total visit duration of 4 hours.

For each participant, visit 1 protocol was repeated identicallyon visit 2, so as to assess the intraindividual variability. Inparticular, the recording zones were repositioned as exactly aspossible at the same spot (made possible by the aforemen-tioned recording on transparent film), with intensity alloca-tion and rotation order identical to those of visit 1.

Data analysisHistamine flare and local skin temperature data were evaluatedstatistically with mixed-model analysis of variance, using thextmixed procedure in Stata version 12 (Stata Corp LP).Responses obtained at the same iontophoretic current intensityat symmetrical locations were considered as replicates, and datafrom the upper and lower limbs were analyzed separately. Thus,the model comprised participant group (control or patient),current intensity (20 or 100 μA), visit number (1 or 2), and all2- and 3-way interactions. In a second step, interactions withp values > 0.1 were removed. The p values reported in theresults are from this more parsimonious model.

The same general method was used to evaluate each of the 4outcome measures generated by QST testing (detection andpain thresholds for warm and cold stimulations). In this case,the model was simpler because there was only 1 visit and thatstimulus intensity was not a factor.

The relationship between histamine flare and clinical severityscore (NIS-LL) was assessed with Spearman rank correlationcoefficient. Finally, we compared the ability of histamine flareand QST data to discriminate between groups, using receiveroperating characteristic (ROC) curve analysis. Logistic re-gression was performed with group membership as the out-come variable and either histamine flare at a specific intensityor specific QST responses as predictors (procedure logistic,followed by the lroc command in Stata). The area under theROC curve (AUC) was calculated. The AUC can range from0.5 to 1, with the former value indicating no, and the latter,perfect discriminating ability.

ResultsWe recruited 6 patients, all women, with FAP of variableseverity (table 1). All had a confirmed mutation in the TTRgene: Val30Met in 5 cases and Glu89Lys in one. As a positivecontrol (table 1), we included a seventh male patient withgeneralized analgesia secondary to lepromatous leprosy. Hisindividual data indicated below is excluded from all statisticalcalculations (see Patients’ characteristics). By “patient group”(or “patients”), we therefore mean only the 6 patients withFAP. Of the 10 healthy controls, 5 were women and theremaining were men. Their median age was 50 years (range



20–64). For comparison, the median age of the 6 FAPpatients was also 50 years (range 37–58).

Patients’ characteristicsAll underwent extensive blood workup and imaging studiesand excluded metabolic, immunologic, or infectious causes ofsmall fiber neuropathy, in particular, diabetes, neurodegen-erative diseases, and paraneoplastic syndromes.

Table 1 contains a detailed description of our patients.Patients were overweight with a median body mass index of29.4 kg/m2 (range 28.5–31.9) corresponding to the range ofthe controls (mean 29.8 kg/m2: 27.5–31.7). The medianduration of FAP was 12 years. Echocardiography was abnor-mal in P04 and P06 only. Nerve conduction studies revealedcarpal tunnel syndrome in 4 patients (P01, P02, P04, andP06) and reduced or absent sural/fibular nerve potentials in 3(P01, P03, and P04). Four patients had undergone livertransplantation. Among them, the patient with the GLu89Lysmutation (P04) also had a heart transplant, as publishedpreviously.14 Clinical manifestations were scored according toNSS, a well-disseminated score used to quantify symptoms inpatients with neuropathies of various causes, and all hadneuropathic symptoms; neurologic examination was quanti-fied using scores dedicated to patients with FAP (the so-calledNIS-LL). These scores were correlated with clinical FAP se-verity and with signs of large nerve fiber neuropathy on ex-amination. As it can be deduced from NSS and NIS-LL data,patients P02, P05, and P06 were mildly affected having onlya small nerve fiber neuropathy. Patients P01, P03, and P04had the highest NSS and NIS-LL scores and were the mostaffected having concomitant small and large nerve fiber neu-ropathy. These 3 patients had abnormal sensory nerve actionpotentials. Concerning the clinical staging of the disease, all 6patients were at stage I of the Clinical Staging of TTR-FAP,a stage used to start therapeutic options in the clinical setting.Patients were heterogeneously distributed according to thePND score (median 1 of 4) or to the Portuguese severityclassification of TTR-FAP (median 2 of 6). The patient withleprosy was the most severely affected.

ThermotestingThermotesting results from the hands and feet in the 2 groupsare shown in table 2. Patients differed statistically from con-trols on the hands and on the feet, with respect to detectionthreshold for both warm and cold. Differences were moremarked on the feet, however, indicating length-dependentsmall fiber impairment in the patients.

Histamine flareFigure 1 shows the area of the flare induced by the 2 ionto-phoretic doses of histamine, measured by LDI on the forearmsand legs on visits 1 and 2, for each study participant. Regardingstatistical analysis, the only interaction terms that came outmarginally significant when fitting the full model was betweenthe group and the iontophoretic dose (p = 0.07, below thepredefined level of 0.1, see Methods). Therefore, the statisticalTa

ble

1Individual

patientch

arac

teristics

Patients

Age

PNPdura

tion(age

oftransp

lant)

Nerv

eco

nduction

TTR

muta

tion

NSS

NIS-LL

FAP

staging

PND

score

Portugu

ese

classification

CTS

Sura

l/Fibular

Tota

lMoto

rSe

nso

ryAuto

nomic

Tota

lWeakness

Reflexe

sSe

nsa

tion

P01

3710

(31)

Yes

AbN

Val30

Met

40

22

60

06

11

2

P02

5010

Yes

NVal30

Met

20

20

20

02

11

1

P03

5114

(40)

No

AbN

Val30

Met

52

21

122

46

12

3

P04

559(49)

Yes

AbN

Glu89

Lys

40

22

60

24

11

2

P05

3714

(26)

No

NVal30

Met

20

20

00

00

10

1

P06

5815

Yes

NVal30

Met

40

40

120

84

12

2

Positive

control

535

Yes

AbN

Leprosy

103

52

278

316

—3

—

Abbreviations:CTS

=ca

rpal

tunnel

syndrome;

FAP=familial

amyloid

polyneu

ropathy;

NIS-LL=Neu

rologica

lImpairm

entS

core–Lo

wer

Limbs;NSS

=Neu

rologica

lSym

ptom

Score

(sen

sory

man

ifes

tationsincludeneu

ropathic

pain);PND=polyneu

ropathydisab

ility;P

NP=polyneu

ropathy;

TTR=tran

sthyretin.



results reported here are from a model comprising simple maineffects and this specific interaction. The dose-response effectwas clearly visible and statistically highly significant on bothvisits (p < 0.001). On the forearm, the histamine flare area did

not differ, at either dose, between the 2 groups (p = 0.30). Bycontrast, on the leg, this response was less intense in patients, atleast for the higher dose of histamine, on both visit 1 (controls:median 13.0 cm2, interquartile range [IQR] 9.7–22.8; patients:median 6.9 cm2, IQR 2.6–10.8; p = 0.011 vs controls) and visit2 (controls: median 13.3 cm2, IQR 8.3–16.9; patients: median8.0 cm2, IQR 6.4–12.1; p = 0.011).

The measurements of local skin temperature (T_skin) per-formed before each iontophoresis are detailed in table e-1(links.lww.com/NXG/A61). On average, T_skin appearedslightly but significantly lower in patients compared withcontrols. Statistical analysis was repeated while adding thispossible confounder to the model with essentially identicalresults: the dose-response effect remained (p < 0.001); thedifference in the flare area between patients and controls wasstill significant on the leg at a current dose of 100 μA (p =0.019), but not on the forearm at either dose (p = 0.47).

In patients, the area of histamine flare on the leg was inverselycorrelated with the clinical status, as shown in figure 2. P01,P03, and P04 were the most affected patients, having con-comitant small and large nerve fiber neuropathy and abnormalsensory nerve action potentials. This correlation held upwhether values for the histamine flare area were taken fromvisit 1 (Spearman R2 0.73, p = 0.09) or visit 2 (R2 0.85, p =0.03). Indeed, flare areas on both visits were nearly identical in4 of the 6 patients, although further apart in the other 2.

The intraindividual variation in the histamine flare area fromvisit 1 to visit 2 is shown more completely in figure 3 (whichfor the sake of brevity only displays data obtained from the

Table 2 Results of QST in the control and patient groups

Controls PatientspValueMedian IQR Median IQR

Hand

Detection,warm

34.4 33.4–36.1 36.1 35.1–40.5 0.053

Detection,cold

31.2 30.2–31.3 30.4 28.8–30.7 0.025

Pain, warm 44.9 42.9–47.7 43.5 40.1–48.2 ns

Pain, cold 14.2 7.5–20.2 8.6 3.2–10.5 ns

Foot

Detection,warm

37.4 35.3–39.2 43.0 39.0–47.6 0.003

Detection,cold

28.5 27.3–29.8 26.3 4.0–28.3 0.040

Pain, warm 45.7 44.6–46.6 48.6 46.5–50.0 ns

Pain, cold 7.7 2.7–21.9 2.1 0.0–8.8 ns

Abbreviations: IQR = interquartile range; QST = quantitative sensitivitytesting.Measurements made on the left and right sides were considered as repli-cates: averaged at the individual level and then summarized for the 2 groupsas indicated. All values are absolute temperature in °C.p Values refer to the comparison of patients with controls.

Figure 1 Histamine flare in patients and controls

Measurements made on the left and right sides of the same person were considered replicates, averaged and plotted as shown, for each study participant.The box plots indicate themedian, interquartile range, and extreme values. *p < 0.05 patients vs controls. **p < 0.01 iontophoretic dose of 100 vs 20 μA. Thered line corresponds to the positive control patient (table 1).




highest iontophoretic current dose) in the form of Bland andAltman plots. These plots reveal negligible bias (none ofwhich reached statistical significance), but relatively widelimits of agreement.

As evaluated from ROC curve analysis (figure 4), the ability ofhistamine flare to discriminate between patients and controlswas similar on visit 1 (AUC 0.81) and visit 2 (0.77). Fur-thermore, the performances of histamine flare and QSTtesting were comparable (AUC for QST testing, 0.78).

DiscussionBased on our previous experience in quantifying spatial dis-tribution of SkBF with LDI,12,15,16 we present here the

findings from a prospective study comparing LDIflare inpatients with FAP and controls. Our results indicate that axonreflex-mediated neurogenic cutaneous vasodilation in re-sponse to histamine measured by LDI (1) is easily elicited inboth groups of participants, (2) is only moderately re-producible at the intraindividual level, (3) despite this latterlimitation appears clearly reduced in patients according toa length-dependent pattern, and (4) shows an ability similarto QST testing for discriminating between patients andcontrols.

The reproducibility of our results was assessed by intra-individual variation in the histamine flare area from visit 1 tovisit 2 (figure 3), revealing negligible bias with no statisticalsignificance, but relatively wide limits of agreement. This re-sult is in accordance with what is already known in LDIflarestudies.17,18

We showed a length-dependent pattern in our patients withFAP: the histamine flare response on the leg was less intensein patients than in controls, at least for the higher dose ofhistamine, whereas it did not differ, at either dose, between the2 groups on the forearm. This result is consistent with QSTtesting: in patients, the detection thresholds for cold andwarm were clearly more abnormal in the feet compared withthe hands (table 2). This is in line with observations reportedby others in patients with FAP despite the subjectivity of QSTtesting.1

Among the tests for small fiber function,3 QST is considereda gold standard tool.1,19 In terms of ability to discriminatebetween patients and healthy controls, LDIflare has beenvariously reported as inferior to QST in diabetic neuropa-thies20 or equivalent in neuropathies of mixed etiologies.21 Inkeeping with this latter study, ROC curve analysis in our caseshowed a similar performance of histamine flare and QST(figure 4). The discrepancy between studies could reflectmethodological differences. Alternatively, it may be related to

Figure 2 Relationship between Neurologic Severity Score and the area of histamine flare on the leg

R2: Spearman rank correlation coefficient. Data pointslabeled according to the patient ID (table 1). NIS-LL =Neurological Impairment Score–Lower Limbs; PC =positive control.

Figure 3Bland andAltmanplots of histamine flare areas onthe leg, recorded on visits 1 and 2

Data from flares induced with the 100 μA iontophoretic current on the leg.Full horizontal lines: average difference across participants (bias). Dashedlines: limits of agreement. Numerical values for bias and limits of agreementshown on each plot, expressed in cm2.



different sample sizes and case mix. Our patients had pe-ripheral neuropathies of homogeneous etiology. However,their small number precludes any meaningful calculation ofcutoff points for estimating the respective sensitivities andspecificities of LDIflare and QST. A larger study is required tothat effect, keeping in mind that the ultimate clinical useful-ness of these tests would depend on their negative/positivepredictive values. The latter, in addition to specificity andsensitivity, are in turn determined by the prevalence of thecondition under investigation.

An important result of our study is the correlation betweenthe LDIflare area and NIS-LL, a score recently found to bea valid and reliable measure of TTR-FAP severity.17 Thisa new correlation being up to now only demonstrated bynerve biomarkers obtained from skin biopsies.22 This is validfor our patients independently of clinical small or large nervefiber involvement or absent/reduced sensory nerve actionpotentials on nerve conduction studies.

The current study has limitations. Considering FAP hetero-geneity,1 our small patient sample may not be representative.However, our data provide a rationale for the use of reflex-mediated neurogenic cutaneous vasodilatation in response tohistamine in a larger scale. This reflex measured by LDI is anobjective and quantitative measure that can potentially serveas a useful biomarker, assessing TTR-FAP severity in

epidemiologic surveys and therapeutic trials, even at an earlydisease stage.

Author contributionsI. Calero-Romero: acquisition of data, analysis or interpretationof data, and revision of the manuscript. M. R. Suter: analysis ofdata and revision of the manuscript. B. Waeber: revision of themanuscript and study supervision. F. Feihl: study concept anddesign, acquisition of data, study supervision, analysis and in-terpretation of data, statistical analysis, and drafting and re-vision of the manuscript. T. Kuntzer: study concept or design,acquisition of data, study supervision, analysis and in-terpretation of data, and drafting and revision of themanuscript.

AcknowledgmentThe authors are grateful to all patients, relatives, and othercontrols for accepting to participate in the study. They thankDr. Melanie Price and Dr. Pinelopi Tsouni for critical readingof the manuscript.


DisclosureI. Calero-Romero reports no disclosures. M.R. Suter hasserved on the scientific advisory boards of Indivior and Pfizer;

Figure 4 Discrimination of control and patient groups with histamine flare compared with quantitative sensitivity testing(QST)

Receiver operating characteristic (ROC) curves derivedfrom logistic regression analysis, with group mem-bership as the dependent variable. Predictors in themodel were those on which the groups differed sta-tistically, i.e., histamine flare area on the legs at a cur-rent intensity of 100 μA (figure 2) or detectionthreshold for warm and cold on the hands and feet(table 2).



serves on the editorial board of Revue Medicale Suisse; and hasreceived research support from Pfizer, IASP, and the Euro-pean Society of Anaesthesiology. B. Waeber and F. Feihl re-port no disclosures. T. Kuntzer has served on the scientificadvisory board of the French Peripheral Society and has re-ceived funding for travel/speaker honoraria from CSL Behr-ing. Full disclosure form information provided by the authorsis available with the full text of this article at Neurology.org/NG.


References1. Ando Y, Coelho T, Berk JL, et al. Guideline of transthyretin-related hereditary am-

yloidosis for clinicians. Orphanet J Rare Dis 2013;8:31.2. Plante-Bordeneuve V, Said G. Familial amyloid polyneuropathy. Lancet Neurol 2011;

10:1086–1097.3. Hoeijmakers JG, Faber CG, Lauria G, Merkies IS, Waxman SG. Small-fibre

neuropathies–advances in diagnosis, pathophysiology and management. Nat RevNeurol 2012;8:369–379.

4. Chiu IM, von Hehn CA, Woolf CJ. Neurogenic inflammation and the peripheralnervous system in host defense and immunopathology. Nat Neurosci 2012;15:1063–1067.

5. Bickel A, Heyer G, Senger C, et al. C-fiber axon reflex flare size correlates withepidermal nerve fiber density in human skin biopsies. J Peripher Nerv Syst 2009;14:294–299.

6. Kubasch ML, Kubasch AS, Torres Pacheco J, et al. Laser doppler assessment ofvasomotor axon reflex responsiveness to evaluate neurovascular function. FrontNeurol 2017;8:370.

7. Namer B, Pfeffer S, Handwerker HO, Schmelz M, Bickel A. Axon reflex flare andquantitative sudomotor axon reflex contribute in the diagnosis of small fiber neu-ropathy. Muscle Nerve 2013;47:357–363.

8. Dyck PJ, Hughes RAC, O’Brien PC. Quantitating overall neuropathic symptoms,impairments, and outcomes. In: Peripheral Neuropathy, 4th ed: Elsevier; 2005:1031–1051.

9. Bril V. NIS-LL: the primary measurement scale for clinical trial endpoints in diabeticperipheral neuropathy. Eur Neurol 1999;41(suppl 1):8–13.

10. Magerl W, Krumova EK, Baron R, Tolle T, Treede RD, Maier C. Reference data forquantitative sensory testing (QST): refined stratification for age and a novel methodfor statistical comparison of group data. Pain 2010;151:598–605.

11. Golay S, Haeberli C, Delachaux A, et al. Local heating of human skin causes hyperemiawithout mediation by muscarinic cholinergic receptors or prostanoids. J Appl Physiol2004;97:1781–1786.

12. Frantz J, Engelberger RP, Liaudet L, Mazzolai L, Waeber B, Feihl F. Desensitization ofthermal hyperemia in the skin is reproducible. Microcirculation 2012;19:78–85.

13. Harper EI, Beck JS, Spence VA. Effect of topically applied local anaesthesia on his-tamine flare in man measured by laser doppler velocimetry. Agents Actions 1989;28:192–197.

14. Niederhauser J, Lobrinus JA, Ochsner F, et al. Successful heart and liver trans-plantation in a Swiss patient with Glu89Lys transthyretin amyloidosis. Trans-plantation 2011;91:e40–e42.

15. Engelberger RP, Pittet YK, Henry H, et al. Acute endotoxemia inhibits microvascularnitric oxide-dependent vasodilation in humans. Shock 2011;35:28–34.

16. Ciplak M, Pasche A, Heim A, et al. The vasodilatory response of skin microcirculationto local heating is subject to desensitization. Microcirculation 2009;16:265–275.

17. Vas PR, Rayman G. The rate of decline in small fibre function assessed using axonreflex-mediated neurogenic vasodilatation and the importance of age related centilevalues to improve the detection of clinical neuropathy. PLoS One 2013;8:e69920.

18. Green AQ, Krishnan ST, Rayman G. C-fiber function assessed by the laser dopplerimager flare technique and acetylcholine iontophoresis. Muscle Nerve 2009;40:985–991.

19. Devigili G, Tugnoli V, Penza P, et al. The diagnostic criteria for small fibre neurop-athy: from symptoms to neuropathology. Brain 2008;131:1912–1925.

20. Gibbons CH, Freeman R, Veves A. Diabetic neuropathy: a cross-sectional study of therelationships among tests of neurophysiology. Diabetes Care 2010;33:2629–2634.

21. Abraham A, Alabdali M, Alsulaiman A, et al. Laser doppler flare imaging and quan-titative thermal thresholds testing performance in small and mixed fiber neuropathies.PLoS One 2016;11:e0165731.

22. Ebenezer GJ, Liu Y, Judge DP, et al. Cutaneous nerve biomarkers in transthyretinfamilial amyloid polyneuropathy. Ann Neurol 2017;82:44–56.





ARTICLE OPEN ACCESS

Association study between multiple systematrophy and TREM2 p.R47HKotaro Ogaki, MD, PhD, Michael G. Heckman, MS, Shunsuke Koga, MD, PhD, Yuka A. Martens, PhD,

Catherine Labbe, PhD, Oswaldo Lorenzo-Betancor, MD, PhD, Ronald L. Walton, BSc, Alexandra I. Soto, BSc,

Emily R. Vargas, MPH, Shinsuke Fujioka, MD, Ryan J. Uitti, MD, Jay A. van Gerpen, MD, William P. Cheshire, MD,

Steven G. Younkin, MD, PhD, Zbigniew K. Wszolek, MD, Phillip A. Low, MD, Wolfgang Singer, MD,

Guojun Bu, PhD, Dennis W. Dickson, MD, and Owen A. Ross, PhD


Correspondence

Dr. Ross

[email protected]

AbstractObjectiveThe triggering receptor expressed on myeloid cells 2 (TREM2) p.R47H substitution(rs75932628) is a risk factor for Alzheimer disease (AD) but has not been well studied inrelation to the risk of multiple system atrophy (MSA); the aim of this study was to evaluate theassociation between the TREM2 p.R47H variant and the risk of MSA.

MethodsA total of 168 patients with pathologically confirmedMSA, 89 patients with clinically diagnosedMSA, and 1,695 controls were included. TREM2 p.R47H was genotyped and assessed forassociation with MSA. Positive results in the Taqman genotyping assay were confirmed bySanger sequencing. The primary comparison involved patients with pathologically confirmedMSA and controls due to the definitive MSA diagnosis in the pathologically confirmed series.

ResultsWe identified TREM2 p.R47H in 3 patients with pathologically confirmed MSA (1.79%), 1patient with clinically diagnosedMSA (1.12%), and 7 controls (0.41%). Minimal AD pathologywas observed for the pathologically confirmed MSA p.R47H carriers. For the primary com-parison of patients with pathologically confirmed MSA and controls, risk of disease was sig-nificantly higher for p.R47H carriers (odds ratio [OR]: 4.39, p = 0.033). When supplementingthe 168 pathologically confirmed patients with the 89 clinically diagnosed and examining thecombined MSA series, the association with TREM2 p.R47H remained significant (OR: 3.81, p= 0.034).

ConclusionsOur preliminary results suggest that the TREM2 p.R47H substitution may be a risk factor forMSA, implying a link to neuroinflammatory processes, especially microglial activation. Vali-dation of this finding will be important, given our relatively small sample size; meta-analyticapproaches will be needed to better define the role of this variant in MSA.

From the Department of Neuroscience (K.O., S.K., Y.A.M, C.L., O.L.-B., R.L.W., A.I.S., S.F., S.G.Y., G.B., D.W.D., O.A.R.), Mayo Clinic, Jacksonville, FL; Department of Neurology (K.O.),Juntendo University Shizuoka Hospital, Izunokunishi, Shizuoka, Japan; Division of Biomedical Statistics and Informatics (M.G.H., E.R.V), Mayo Clinic, Jacksonville, FL; Department ofNeurology (R.J.U., J.A.v.G., W.P.C., Z.K.W.), Mayo Clinic, Jacksonville, FL; Department of Neurology (P.A.L., W.S.), Mayo Clinic, Rochester, MN; Mayo Graduate School (O.A.R.),Neurobiology of Disease, Jacksonville, FL; and Department of Clinical Genomics (O.A.R.), Jacksonville, FL.









Multiple system atrophy (MSA) is a rare, rapidly progressing,adult-onset neurodegenerative disease that is clinically charac-terized by parkinsonism, autonomic failure, pyramidal symp-toms, and cerebellar ataxia and pathologically defined by thepresence of α-synuclein-positive glial cytoplasmic inclusions.1

Uncovering genetic risk factors for MSA has proven difficultbecause of the challenge of accumulating a reasonable samplesize and potential for clinical misdiagnosis. Nonetheless, studieshave nominated several genes that may be associated with therisk of MSA.1,2 The first genome-wide association study of MSAwas recently completed, and although no genome-wide signifi-cant associations were identified, suggestive associations wereidentified involving variants in MAPT, FBXO47, ELOVL7, andEDN1.3 Clearly, much remains to be understood regarding thegenetics of MSA, and replication of previously nominated MSAsusceptibility loci remains crucial.

A relatively rare substitution p.R47H (rs75932628) in thetriggering receptor expressed on myeloid cells 2 (TREM2)protein is now a well-established risk factor for Alzheimerdisease (AD) and has been suggested to play a role in otherneurodegenerative disorders such as frontotemporal de-mentia (FTD) and Parkinson disease (PD).4 Of interest,mutations in the TREM2 gene were originally reported asa genetic cause of Nasu-Hakola disease, a rare leukodystrophycharacterized by progressive presenile dementia and bonecysts.4 Given the important role of microglia in neuro-inflammation, we evaluated the association between theTREM2 p.R47H variant and the risk of MSA using a pre-dominantly neuropathologically confirmed MSA series.

MethodsStudy participantsThis study included 168 patients with pathologically con-firmed MSA, 89 patients with clinically diagnosed MSA, and1,695 healthy controls. The pathologically confirmed MSAseries consisted of all cases obtained from the Mayo Clinicbrain bank for neurodegenerative disorders in Jacksonville,FL, between 1998 and 2017. Pathologically confirmed MSAwas diagnosed by a single neuropathologist (D.W.D.) aspreviously described.5 Immunohistochemistry for α-synuclein(NACP, 1:3,000 rabbit polyclonal) was performed on sec-tions of the basal forebrain, striatum, midbrain, pons, medulla,and cerebellum to establish a neuropathologic diagnosis. MSAcases were pathologically subclassified as MSA with pre-dominant striatonigral involvement (MSA-P), MSA withpredominant olivopontocerebellar involvement (MSA-C),and MSA with equally severe involvement of striatonigral and

olivopontocerebellar systems (MSA-mixed).5 One case wasnot assigned pathologic subtype because only the pons wasavailable for histology.

The clinically diagnosed patients (N = 81 probableMSA, N = 8possible MSA) were seen at the Mayo Clinic in Rochester, MN(N = 35), or Jacksonville, FL (N = 54), where diagnosis wasmade according to the criteria of Gilman et al.6 Clinically di-agnosed MSA cases were free of notable cognitive symptomsthat might suggest the presence of an overlapping dementia orother neurodegenerative conditions. Pathologic assessmentand clinical diagnoses were made without the knowledge of theTREM2 p.R47H genotype. Controls were also from the MayoClinic in Rochester, MN (N = 991), or Jacksonville, FL (N =704). Controls were free of neurologic symptoms and withouta family history of movement disorders. All participants wereunrelated non-Hispanic Caucasians. The primary comparisonof TREM2 p.R47H involved the patients with pathologicallyconfirmed MSA and controls. In secondary analysis, we com-bined the patients with pathologically confirmed MSA andpatients with clinically diagnosed MSA and compared thiscombined MSA series with the controls. Table 1 shows char-acteristics of patients with MSA and controls.

Standard protocol approvals, registrations,and patient consentsTheMayo Clinic IRB approved the use of human participantsin this study.

Genetic analysisGenomic DNA was extracted from brain tissue or peripheralblood monocytes using standard protocols. Genotyping ofTREM2 exon 2 variant rs75932628 (NM_018965.3:c.140G>A, p.R47H) was performed using a custom TaqManAllelic Discrimination Assay on an ABI 7900HT Fast Real-Time PCR system (Applied Bio-systems, Foster City, CA)according to the manufacturer’s instructions (sequencesavailable upon request). Genotype calls were made using SDS2.2.2 software. Positive or ambiguous results in the TaqManassay were also confirmed and resolved via Sanger sequencing.There was no evidence of a departure from Hardy-Weinbergequilibrium in controls (p = 0.93).

Statistical analysisWe evaluated the association between TREM2 p.R47H andMSA using logistic regressionmodels. Unadjustedmodels werefirst examined. Subsequently, although it is uncertain whetheradjusting for any variable is reasonable, given the rare nature ofthis variant, we did also adjust our logistic regressionmodels forage and sex in a sensitivity analysis. Odds ratios (ORs) and 95%

GlossaryAD = Alzheimer disease; FTD = frontotemporal dementia;MSA = multiple system atrophy;MSA-C =MSA with predominantolivopontocerebellar involvement; MSA-P = MSA with predominant striatonigral involvement; OR = odds ratio; PD =Parkinson disease; TREM2 = triggering receptor expressed on myeloid cells 2.



confidence intervals were estimated. p ≤ 0.05 was consideredstatistically significant. All statistical analyses were performedusing SAS (version 9.4; SAS Institute, Inc., Cary, NC).

ResultsAn evaluation of the association between TREM2 p.R47H andMSA is shown in table 2. The p.R47H variant was observed in 3patients with pathologically confirmedMSA (1.79%). Of these 3p.R47H carriers, neuropathologic examination was available for2 and revealed MSA subtypes of MSA-P and MSA-C, Braakneurofibrillary tangle stages of I and III, and Thal amyloid phasesof 0 and 0. For the 89 patients with clinically diagnosed MSA,there was 1 p.R47H carrier (1.12%, from Mayo Jacksonville),and this patient’s MSA subtype was probable MSA-P. Therewere 7 carriers of p.R47H in the controls (0.41%, N = 4 MayoJacksonville, N = 3 Mayo Rochester), corresponding to a minorallele frequency of 0.21%, which is similar to the frequencyreported in the ExAC database (Cambridge, MA [exac.broad-institute.org]) for this population (0.26%).7 For the primarycomparison of patients with pathologically confirmed MSA andcontrols, risk ofMSAwas significantly higher for p.R47H carriers(OR = 4.39, p = 0.033). When combining the pathologicallyconfirmed and clinically diagnosed MSA series, the risk of MSA(vs controls) was also significantly higher (OR = 3.81, p =0.034). In a sensitivity analysis adjusting for age and sex, thefindings regarding the patients with pathologically confirmedMSA (OR= 3.55, p= 0.076) and the combinedMSA series (OR= 3.17, p = 0.078) remained relatively consistent though weak-ened slightly and were no longer quite significant.

DiscussionTREM2 p.R47H is a well-established risk factor for AD4; how-ever, results have been mixed for FTD and PD,4 while studiesinvolving amyotrophic lateral sclerosis and dementia with Lewybodies mostly point to a lack of association.4 Of interest, theresults of our current study suggest that the TREM2 p.R47Hvariant may be associated with an increased risk of MSA. Witha notably higher frequency in our combined MSA series (1.6%)compared with controls (0.4%), we observed a nearly 4-foldincreased risk ofMSA in p.R47H carriers. It is important that thisfinding was strongest in our pathologically confirmed MSA se-ries. The absence of notable AD pathology in pathologicallyconfirmed MSA p.R47H carriers indicates that the observedassociation was not driven by such pathology. The associationwas attenuated slightly when adjusting for age and sex; however,those results should not be overinterpreted because it is un-certain whether attempting any adjustment is reasonable, giventhe low frequency of p.R47H. These preliminary findings suggestthat TREM2 p.R47H may play a role in susceptibility to MSAand that further study is warranted.

To the best of our knowledge, only 1 association study be-tween TREM2 p.R47H andMSA has been conducted to date.In a study of 407 Chinese patients with clinically diagnosedMSA and 869 controls, the p.R47H substitution was observedin 1 MSA patient and no controls (p = 0.14).8 This negativeresult is not surprising because TREM2 p.R47H is extremelyrare in East Asians.

Table 1 Characteristics of patients withMSA and controls

Variable

Patients withpathologicallyconfirmedMSA (N = 168)

PatientswithclinicallydiagnosedMSA(N = 89)

Controls(N = 1,695)

Age at onset in patientswith MSA or age atblood draw in controls(y)

58 (39, 88) 64 (46, 83) 76 (30, 88)

Age at death (y) 66 (47, 91) NA NA

Sex

Male 96 (57.1%) 53 (59.6%) 786(46.4%)

Female 72 (42.9%) 36 (40.4%) 909(53.6%)

MSA subtype

MSA-P 73 (43.5%) 62 (69.7%) NA

MSA-C 27 (16.1%) 26 (29.2%) NA

Mixeda 67 (39.9%) 1 (1.1%) NA

Unclassifiedb 1 (0.6%) 0 (0.0%) NA

Braak NFT stage

0 29 (17.8%) NA NA

I 61 (37.4%) NA NA

II 49 (30.1%) NA NA

III 16 (9.8%) NA NA

IV 5 (3.1%) NA NA

V 2 (1.2%) NA NA

VI 1 (0.6%) NA NA

Thal amyloid phase

0 88 (55.7%) NA NA

1 33 (20.9%) NA NA

2 18 (11.4%) NA NA

3 9 (5.7%) NA NA

4 8 (5.1%) NA NA

5 2 (1.3%) NA NA

Abbreviations: MSA = multiple system atrophy; MSA-C = MSA with pre-dominant olivopontocerebellar involvement; MSA-P = MSA with predominantstriatonigral involvement; NA = not applicable.The samplemedian (minimum,maximum) is given for age. Information wasunavailable regarding age at onset (N = 49), Braak NFT stage (N = 5), and Thalamyloid phase (N = 10) for the patients with pathologically confirmed MSA.a MSA with equally severe involvement of striatonigral and olivopontocer-ebellar systems was classified as MSA-mixed.5b One patient withMSAwas considered to be “unclassified” because only thepons was available for histologic assessment; therefore, we could not di-agnose this patient as MSA-C or MSA-mixed.


http://exac.broadinstitute.org

http://exac.broadinstitute.org


How the TREM2 p.R47H variant could increase risk of MSAis an important topic for further study. MSA is regarded as anoligodendrogliopathy, but the mechanism of demyelinationremains unclear. While demyelination depends primarily onoligodendrocytes, microglia also contribute to myelinationand myelin homeostasis.9 In the CNS, TREM2 is pre-dominantly expressed in microglia, and it has been shown toplay critical roles in phagocytic clearance of apoptotic cellsand disease-associated molecules and modulating microglialimmune response.10 A recent study showed that TREM2 isrequired for myelin debris removal and remyelination aftercuprizone-induced demyelination.11 We hypothesize thatTREM2 p.R47H, a loss-of-function mutation, leads to im-paired clearance of myelin debris and aberrant microglial ac-tivation, which result in increased risk of MSA.

Our results provide evidence that the TREM2 p.R47H sub-stitution may be a genetic susceptibility factor for MSA. Ourstudy is limited by the relatively small number of patients withpathologically confirmedMSA, and therefore, validation will becrucial. Recently, the clinical diagnostic accuracy of MSA wasreported to be 62%, further underscoring the importance ofusing neuropathologically confirmed MSA cases in researchstudies.12 We hope that these initial promising results will formthe basis for future examinations of TREM2 p.R47H in MSA.

Author contributionsK. Ogaki: drafting/revising the manuscript; study concept anddesign; and interpretation of data. M. G. Heckman andS. Koga: drafting/revising the manuscript and interpretationof data. Y.A. Martens: drafting/revising the manuscript.C. Labbe and O. Lorenzo-Betancor: revising the manuscript.R.L. Walton, A.I. Soto, and E.R. Vargas: revising the manu-script and interpretation of data. S. Fujioka: revising themanuscript. R.J. Uitti, J. Van Gerpen, and W.P. Cheshire: re-vising the manuscript and providing material for this study.S.G. Younkin: revising the manuscript. Z.K. Wszolek: revisingthe manuscript, providing material and funding for this study,and IRB committee approval. P.A. Low and W. Singer: revising

the manuscript and providing material for this study. G. Bu:revising the manuscript. D.W. Dickson: drafting/revisingthe manuscript, providing material and funding for this study,and IRB committee approval. O.A. Ross: drafting/revising themanuscript; study concept and design; and interpretationof data.

AcknowledgmentThe authors thank all those who have contributed to ourresearch, particularly the patients and families who donatedBrain and DNA samples for this work.

Study fundingK. Ogaki is supported by a research grant from the NAITOFoundation, JSPS KAKENHI Grant Number JP17K14966and a grant from Institute for Environmental and Gender-Specific Medicine, Juntendo University. R.J. Uitti, O.A. Ross,Z.K. Wszolek, and D.W. Dickson are partially supported by theNIH/NINDS P50 NS072187. O.A. Ross is supported by theNINDS R01# NS078086, U54 NS100693, DOD W81XWH-17-1-0249, and in part by the Michael J. Fox Foundation forParkinson’s Research. C. Labbe is the recipient of a F.R.S.Q.fellowship. Z.K. Wszolek receives research support fromNIH/NIA (primary), and NIH/NINDS (secondary)1U01AG045390-01A1, Mayo Clinic Center for RegenerativeMedicine, Mayo Clinic Neuroscience Focused Research Team(Cecilia and Dan Carmichael Family Foundation, and theJames C. and Sarah K. Kennedy Fund for NeurodegenerativeDisease Research atMayoClinic in Florida), and is also partiallysupported by a gift from Carl Edward Bolch, Jr, and Susan BassBolch, The Sol Goldman Charitable Trust, and Donald G. andJodi P. Heeringa. P.A. Low receives research support from theCure MSA Foundation. Samples included in this study wereclinical patients or brain donors to the brain bank atMayo Clinicin Jacksonville, which is supported by the CurePSP Society forProgressive Supranuclear Palsy and Udall Center for Excellencein Parkinson Research (P50 NS072187), APDA Center forAdvanced Research, an Alzheimer’s disease Research Centergrant (NIA P50 AG16574), and NINDS R01 NS078086.

Table 2 Association between TREM2 p.R47H and MSA

Disease group NNo. (%) with TREM2p.R47H

Unadjusted analysis Adjusting for age and sex

OR (95% CI) p Value OR (95% CI) p Value

Controls 1,695 7 (0.41) 1.00 (reference) NA 1.00 (reference) NA

Patients withpathologicallyconfirmed MSA

168 3 (1.79) 4.39 (1.12–17.12) 0.033 3.55 (0.88–14.38) 0.076

All patients withMSA (pathologicallyconfirmed andclinically diagnosed)

257 4 (1.56) 3.81 (1.11–13.12) 0.034 3.17 (0.88–11.45) 0.078

Abbreviations: CI = confidence interval; MSA = multiple system atrophy; NA = not applicable; OR = odds ratio; TREM2 = triggering receptor expressed onmyeloid cells.ORs, 95% CIs, and p values result from logistic regression models.For age, the age that was adjusted for was age at death for patients with pathologically confirmedMSA, age atMSA onset for patients with clinically diagnosedMSA, and age at blood draw for controls.



DisclosureK. Ogaki receives JSPS KAKENHI Grant NumberJP17K14966 and a grant from Institute for Environmental &Gender-Specific Medicine, Juntendo University. M.G.Heckman serves as the editorial board member of Parkin-sonism & Related Disorders. S. Koga and Y.A. Martens reportno disclosures. C. Labbe receives a Fonds de recherche duQuebec-Sante postdoctoral fellowship. O. Lorenzo-Betancor, R.L. Walton, A.I. Soto, and E. Vargas report nodisclosures. S. Fujioka receives JSPS KAKENHI GrantNumber JP15K19501 and is an editorial board member ofParkinsonism & Related Disorders. R.J. Uitti serves as an as-sociate editor of Neurology. J.Van Gerpen report no dis-closures. W.P. Cheshire serves in the editorial boards ofParkinsonism&Related Disorders and Autonomic Neuroscienceand as an associate editor of Clinical Autonomic Research.S.G. Younkin reports no disclosures. Z.K. Wszolek receivesresearch support from P50 NS072187, NIH/NIA (primary),and NIH/NINDS (secondary) 1U01AG045390-01A1, MayoClinic Center for Regenerative Medicine, Mayo Clinic Neu-roscience Focused Research Team (Cecilia and Dan Carmi-chael Family Foundation, and the James C. and Sarah K.Kennedy Fund for Neurodegenerative Disease Research atMayo Clinic in Florida), the gift from Carl Edward Bolch, Jr.,and Susan Bass Bolch, The Sol Goldman Charitable Trust, andDonald G. and Jodi P. Heeringa; serves as coeditor-in-chief ofParkinsonism & Related Disorders and an associate editor of theEuropean Journal of Neurology; is on the editorial boards ofNeurologia i Neurochirurgia Polska, European Journal of Clinicaland Experimental Medicine, Clinical and Experimental MedicalLetters, and Wiadomosci Lekarskie; holds and has contractualrights for receipt of future royalty payments from patents for “ANovel Polynucleotide Involved in Heritable Parkinson’s Dis-ease”; and receives royalties from publishing Parkinsonism &Related Disorders (Elsevier, 2016, 2017) and the EuropeanJournal of Neurology (Wiley Blackwell, 2016, 2017). P.A. Lowreceives research support from the NIH (P01 NS44233, U54NS065736, R01 NS092625, and UL1 TR000135), FDA (R01FD004789), Cure MSA Foundation, and Mayo Funds and

serves as a clinical editor of Autonomic Neuroscience. W. Singerreports no disclosures. G. Bu received support fromP50AG016574, RF1AG051504, R01AG027924, R01AG035355,R01AG046205, and P01NS074969 and a grant from the CureAlzheimer’s fund. D.W. Dickson receives support from P50-AG016574, P50-NS072187, P01-AG003949, and CurePSP:Foundation for PSP|CBD and Related Disorders. D.W. Dicksonis an editorial board member of Acta Neuropathologica, Annalsof Neurology, Brain, Brain Pathology, and Neuropathology andis editor-in-chief of the American Journal of NeurodegenerativeDisease and International Journal of Clinical and ExperimentalPathology. O.A. Ross received support from R01 NS078086,P50-NS072187, and U54 NS100693 and the Michael J.Fox Foundation. O.A. Ross is an editorial board member ofthe American Journal of Neurodegenerative Disease,Neurologia iNeurochirurgia Polska, and Molecular Neurodegeneration. Fulldisclosure form information provided by the authors isavailable with the full text of this article at Neurology.org/NG.


References1. Ahmed Z, Asi YT, Sailer A, et al. The neuropathology, pathophysiology and genetics

of multiple system atrophy. Neuropathol Appl Neurobiol 2012;38:4–24.2. Scholz SW, Bras J. Genetics underlying atypical parkinsonism and related neurode-

generative disorders. Int J Mol Sci 2015;16:24629–24655.3. Sailer A, Scholz SW, Nalls MA, et al. A genome-wide association study in multiple

system atrophy. Neurology 2016;87:1591–1598.4. Jay TR, von Saucken VE, Landreth GE. TREM2 in neurodegenerative diseases. Mol

Neurodegener 2017;12:56.5. Koga S, Parks A, Uitti RJ, et al. Profile of cognitive impairment and underlying

pathology in multiple system atrophy. Mov Disord 2017;32:405–413.6. Gilman S, Wenning GK, Low PA, et al. Second consensus statement on the diagnosis

of multiple system atrophy. Neurology 2008;71:670–676.7. Lek M, Karczewski KJ, Minikel EV, et al. Analysis of protein-coding genetic variation

in 60,706 humans. Nature 2016;536:285–291.8. Chen Y, Chen X, Guo X, et al. Assessment of TREM2 rs75932628 association with

Parkinson’s disease and multiple system atrophy in a Chinese population. Neurol Sci2015;36:1903–1906.

9. Simons M, Lyons DA. Axonal selection and myelin sheath generation in the centralnervous system. Curr Opin Cel Biol 2013;25:512–519.

10. Painter MM, Atagi Y, Liu CC, et al. TREM2 in CNS homeostasis and neurodegen-erative disease. Mol Neurodegener 2015;10:43.

11. Poliani PL, Wang Y, Fontana E, et al. TREM2 sustains microglial expansion duringaging and response to demyelination. J Clin Invest 2015;125:2161–2170.

12. Koga S, Aoki N, Uitti RJ, et al. When DLB, PD, and PSP masquerade as MSA: anautopsy study of 134 patients. Neurology 2015;85:404–412.




ARTICLE OPEN ACCESS

Confirming TDP2 mutation in spinocerebellarataxia autosomal recessive 23 (SCAR23)Guido Zagnoli-Vieira, BSc, Francesco Bruni, PhD,* Kyle Thompson, PhD, Langping He, MD, PhD,

Sarah Walker, PhD, Arjan P.M. de Brouwer, PhD, Robert Taylor, PhD, FRCPath, Dmitriy Niyazov, MD,

and Keith W. Caldecott, PhD


Correspondence

Dr. Caldecott

[email protected]

AbstractObjectiveTo address the relationship between mutations in the DNA strand break repair protein tyrosylDNA phosphodiesterase 2 (TDP2) and spinocerebellar ataxia autosomal recessive 23(SCAR23) and to characterize the cellular phenotype of primary fibroblasts from this disease.

MethodsWe have used exome sequencing, Sanger sequencing, gene editing and cell biology, bio-chemistry, and subcellular mitochondrial analyses for this study.

ResultsWe have identified a patient in the United States with SCAR23 harboring the same homozy-gous TDP2mutation as previously reported in 3 Irish siblings (c.425+1G>A). The current andIrish patients share the same disease haplotype, but the current patient lacks a homozygousvariant present in the Irish siblings in the closely linked gene ZNF193, eliminating this asa contributor to the disease. The current patient also displays symptoms consistent withmitochondrial dysfunction, although levels of mitochondrial function in patient primary skinfibroblasts are normal. However, we demonstrate an inability in patient primary fibroblasts torapidly repair topoisomerase-induced DNA double-strand breaks (DSBs) in the nucleus andprofound hypersensitivity to this type of DNA damage.

ConclusionsThese data confirm theTDP2mutation as causative for SCAR23 and highlight the link betweendefects in nuclear DNA DSB repair, developmental delay, epilepsy, and ataxia.

*Current address: Department of Biosciences Biotechnologies and Biopharmaceutics, University of Bari Aldo Moro, Bari, Italy.

From the Genome Damage and Stability Centre (G.Z-V., K.W.C.), University of Sussex, Falmer, Brighton, United Kingdom; Wellcome Centre for Mitochondrial Research (F.B., K.T., L.H.,R.T.), Institute of Neuroscience, Newcastle University, Tyne, United Kingdom; Sussex Drug Discovery Centre (S.W.), University of Sussex, Falmer, Brighton, United Kingdom;Department of Human Genetics (A.P.M.d.B.), Donders Institute for Brain, Cognition and Behaviour, Radboud University Medical Center, Nijmegen, The Netherlands; and MedicalGenetics (A.P.M.d.B., D.N.), Ochsner Health Center for Children, New Orleans, LA.









DNA is under constant threat from attack by endogenous andexogenous electrophilic molecules,1 and DNA topoisomeraseenzymes can introduce DNA breaks as abortive intermediatesof their activity.2–4 Topoisomerase “poisons” such as etopo-side inhibit the ligation activity of topoisomerase 2 (TOP2),thereby promoting the formation of abortive DNA double-strand break (DSB) intermediates that require DSB repair.DSBs are repaired in cells by either homologousrecombination–mediated repair or by nonhomologous endjoining (NHEJ).2 The repair of TOP2-induced DSBs byNHEJ involves the enzyme tyrosyl DNA phosphodiesterase 2(TDP2), which removes trapped topoisomerase peptide fromthe 59-termini at the DSB and thereby allows the DNA ends tobe ligated.3–5 The loss of TDP2 in mouse results in reducedexpression of >100 genes in the brain,6 and TDP2mutation inhumans has been associated with intellectual disability, seiz-ures, and ataxia,6 a disease now denoted as spinocerebellarataxia, autosomal recessive 23 (SCAR23). To date, our un-derstanding of SCAR23 has been limited by the availability ofonly 3 Irish siblings with a mutation in TDP2 and by the lackof availability of fibroblast cell lines from these patients formolecular and cellular characterization. Here, we haveaddressed these limitations and identified an SCAR23 patientin the United States with the same homozygous TDP2 mu-tation as present in the Irish siblings, confirming the associ-ation of this disease with mutated TDP2. In addition, we havecharacterized at the molecular and cellular level primary pa-tient fibroblasts from the current SCAR23 patient.

MethodsPatient case reportThe patient is currently a 6-year-old caucasian boy fromthe United States presenting with developmental delay,microcephaly, and failure to thrive. His mother and fatherare aged 29 years and 32 years, respectively, and there is noprevious family history of related disease, but there ispossible consanguinity (3rd cousins). The patient startedwalking at the age of 14 months and talking at 3 years, hishearing was reportedly normal, but he did not have audi-tory brainstem response. The patient is easily fatigued, andhis parents reported that he eats excessively, becomes ir-ritable, and falls asleep sometimes for days or up to 2weeks. He is now below the fifth percentile for weight and5% for height, with a body mass index at <5%. He hasa G-tube and is followed by a dietitian and receives nutri-tional liquid supplements (PediaSure; Abbott, Lake Forest,IL). He has a history of constipation and is followed up bya gastroenterologist. He has a 12-year-old brother who is

also easily fatigued and a 14-year-old maternal half-sisterwith attention deficit hyperactivity disorder (ADHD).

The patient has difficulty keeping balance and has an ataxicgait. He has been tested several times for abnormalities inblood, urine, and CSF, and by MRI and EEG, and only thelatter is abnormal. Seizures began at the age of 5 months, andhis EEGs show increased slowing and occasional spikes in theright posterior quadrant, and during drowsiness, he had 2generalized bursts of polyspike and slow-wave activity ata frequency of 4–5 Hz. The patient’s 180K oligoarray, verylong chain fatty acids, carbohydrate-deficient transferrin, CSFlactate, and neurotransmitters and plasma amino acids werenormal.

The patient exhibited several phenotypes consistent with mi-tochondrial dysfunction such as hypotonia, low energy, fati-gability, hypersomnia, failure to thrive, short stature,constipation, neutropenia, hyponatremia, cardiac arrhythmia,and gastrointestinal dysmotility. Consistent with this, electrontransport chain (ETC) studies on muscle biopsy at age 1 yearshowed a severe reduction of complex I + III and II + IIIactivity, which satisfied the major Walker criteria after cor-rection for increased citrate synthase activity. CoQ10 de-ficiency was suggested based on the complex I + III and II + IIIdeficiency, and his metabolic tests were abnormal with thehigh lactate and high lactate/pyruvate ratio. The patient wasplaced on ubiquinol, carnitine, and leucovorin, and anecdot-ally, he responded well (particularly to liposomal ubiquinol,the active or reduced form of CoQ10) in terms of energy anddevelopmental progress. He is reportedly no longer sleepy andlethargic and has a stable gait, but is “still behind” in languageand cognitive skills and is in special education. He still hasidiopathic fevers but has not been admitted to the hospital for15 months. He does not have as many infections but still hasgastric dysmotility. His seizures have reduced in frequency.

Standard protocol approvals, registrations, andpatient consentsWe confirm that we have received approval from an in-stitutional ethics standards committee for this work, and wehave written informed consent for research from the guardianof the patient for participation in this study.

Exome sequencing and haplotype analysisWhole-exome sequencing (WES) (trio study) of the patient wasperformed by GeneDx using the Agilent Clinical ResearchExome kit to target the exonic regions and flanking splicejunctions of the genome. These targeted regions were sequencedsimultaneously by massively parallel (NextGen, Irvine, CA)

GlossaryDSB = double-strand break; ETC = electron transport chain; FCS = fetal calf serum; NHEJ = nonhomologous end joining;SCAR23 = spinocerebellar ataxia, autosomal recessive 23; TDP2 = tyrosyl DNA phosphodiesterase 2; WCE = whole-cellextract; WES = Whole-exome sequencing.



sequencing on an Illumina HiSeq sequencing system with100bp paired-end reads. Bidirectional sequence was assembled,aligned to reference gene sequences based on human genomebuild GRCh37/UCSC hg19, and analyzed for sequence var-iants in the selected genes or regions of interest using a custom-developed analysis tool (XomeAnalyzer). Capillary sequencingwas used to confirm all potentially pathogenic variants identi-fied in this individual. Sequence alterations were reportedaccording to the nomenclature guidelines of the HumanGenome Variation Society. The WES identified a homozygoussplice site mutation (c.425+1G>A) in the TDP2 gene. Forcomparison of the current patient with the Irish pedigreepreviously reported,6 we conducted haplotype analysis. Var-iants were considered for homozygosity if they were (1) cov-ered by at least 4 reads or more, (2) present in 80% of all readsor more, (3) designated as a substitution, (4) uniquely posi-tioned in the human genome, and (5) present in the exomedata of both individuals. Homozygous regions were de-termined using a sliding window, accepting 2 or less homozy-gous variants per 10 variants assessed.

Mitochondrial preparation and subcellularfractionationMitochondria were prepared as described previously,7 with fewmodifications. HeLa cells and fibroblasts (control and patient)were harvested, resuspended in homogenization buffer (HB[0.6 M mannitol, 10 mM Tris-HCl pH 7.4, 1 mM (ethyleneglycol-bis(β-aminoethyl ether)-N,N,N9,N9-tetraacetic acid)(EGTA), 0.1% bovine serum albumin (BSA) (wt/vol]), andsubjected to differential centrifugation. Mitochondria were pel-leted at 11.000g for 10 minutes at 4°C and resuspended in HB;the postmitochondrial supernatant was retained after centrifu-gation (“post-mito spin”). For submitochondrial fraction prepa-ration, HeLa cell mitochondria (300 μg) were treated with 1.6 μgof proteinase K on ice for 30minutes, followed by the addition of5 mM phenylmethanesulfonyl fluoride (PMSF). This fractionwas pelleted at 11.000g for 10minutes at 4°C and resuspended inHB. Mitoplasts were obtained by resuspending PK-treated mi-tochondria in 9 volumes of 10 mM Tris-HCl (pH 7.4) andtreated with PK, as described earlier. Inner mitochondrialmembrane proteins were extracted in the presence of 100 mMNa2CO3, followed by centrifugation at 100.000g for 15 minutesat 4°C. Proteins (30 μg) from each fraction were loaded onto12% SDS-PAGE gel, transferred to the polyvinylidene difluoride(PVDF) membrane, and analyzed by immunoblotting usingprimary antibodies to apoptosis inducing factor (AIF) (NEB),eIF4E (Cell Signalling), EF-Tu (custom made), NDUFB8(Mitosciences), and TDP2 (see Western Blotting, below).

Cell culture and vectorsHuman A549 cells were grown in Dulbecco Modified EagleMedium (Gibco, ThermoFisher, Waltham, MA) containing10% fetal calf serum (FCS), 2 mM glutamine, penicillin (100units/mL), and streptomycin (100 μg/mL). Human fibroblastswere grown in Minimum Essential Media (Gibco) containing15% FCS, 2 mM glutamine, penicillin (100 units/mL), andstreptomycin (100 μg/mL). All cells were grown at 5% CO2 at

37°C. TDP2-mutated patient primary human fibroblasts wereestablished from a patient’s skin biopsy and were denoted 850-BR. The control human fibroblast cell line 1-BR.3 (denotedhere for simplicity as 1-BR) was previously derived from anunrelated normal individual and has been described pre-viously.8 For complementation experiments, we used 1-BRcells that were immortalized previously with hTERT (denotedas 1-BR hTERT) and a derivative of 850-BR immortalized inthe current study by retroviral-mediated hTERT expressionand selected in a medium containing 1 μg/ml puromycin(denoted as 850-BR hTERT). For complementation withhuman TDP2, 850-BR hTERT cells were transfected with ei-ther empty eGFP-N1 vector or eGFP-N1 construct encodingGFP-tagged human TDP2 (denoted as TDP2-GFP-N1) andstable transfectants selected for 21 days by growth in a mediumcontaining 0.5mg/mLG418 (Gibco, ThermoFisher,Waltham,MA). TDP2-GFP-N1 was generated by PCR amplificationof the human TDP2 ORF using the primers TDP2_FW(59-AAAGAATTCATGGAGTTGGGGAGTTGCCTG-39)and TDP2_RV (59-AAAGGATCCAATATTATATCTAAGTTGCACAGAAGACC-39) and subcloning the PCRproduct into the EcoRI/BamHI sites of eGFP-N1.

CRISPR/Cas9 gene editingA549 TDP2−/− cells were created as previously reported.9 Inbrief, we used a 17-bp (minus the PAM) RNA sequencetargeting TDP2 exon 4 (59-GTAGAAATATCACATCT-39),which was selected using the tool E-CRISP (e-crisp.org/E-CRISP/) and cloned into the guide RNA vector #41824(AddGene). The TDP2 guide construct was cotransfectedwith hCas9 expressed from plasmid #41815 (AddGene) usingAmaxa Nucleofector plataform (Lonza, Basel, Switzerland)Kit T program X-001. Transfected cells were enriched byselection in 1 mg/mL G418 (Thermofisher) for 5 days beforeisolation of single clones and screening for loss of TDP2expression by Western blotting.

Sanger sequencingDNA was extracted from 850-BR cells using the DNeasyBlood & Tissue kit (Qiagen, Manchester, UK). PCRreactions used Phusion HF-DNA Polymerase (NEB) andthe following primers: TDP2_FW GCCAGTGTT-GACCTAACCAATGAAGA; TDP2_RV CTGTAGAAA-TATCACATCTGGGCTGTACC; ZSCAN9_FWATGAAGTAACCAAGACTGAGGACAGAGAG; and ZSCAN9_RVAGACCAGCTCAGCCACTGTGTGGATCT. PCR productswere purified before sequencing using a QIAquick PCR purifi-cation kit (Qiagen).

Western blottingAnti-TDP2 antibody was used at 1:5000 in Western blottingand has been described previously.10 Anti-Actin (SigmaA4700) was used at 1:2000. Western blot assessment ofOXPHOS components in patient primary fibroblasts wasconducted as described previously,11 using antibodies againstNDUFB8 (Abcam cat# ab110242), SDHA (Abcam cat#ab14715), UQCRC2 (Abcam cat# ab14745), COXI (Abcam


http://www.e-crisp.org/E-CRISP/

http://www.e-crisp.org/E-CRISP/


cat# ab14705), ATP5A (Abcam cat# ab14748), SDHB(Abcam cat# ab14714), COX II (Abcam cat# 110258), andVDAC1 (Abcam cat# ab14734).

Clonogenic survival assays1-BR and 850-BR primary fibroblasts cells were plated ontofeeder layers (see below) and 3 hours later treated with indicatedconcentrations of etoposide for 21 days to allow the formation ofmacroscopic colonies, which were rinsed in phosphate bufferedsaline (PBS) and fixed/stained in 70% ethanol/1% methyleneblue. A549 and TDP2−/− A549 were plated 4 hours beforetreatment with the indicated concentrations of etoposide for 12days and stained as described earlier. The surviving fraction ateach dose was calculated by dividing the average number of

colonies (>50 cells) in treated dishes by the average number inuntreated dishes. For feeder layers, 1-BR cells were irradiated (35Gy) and plated 24 hours before use at 5 × 104 cells/10 cm dish.

Tyrosyl DNA phosphodiesterase assaysWhole-cell extract (WCE) was prepared by resuspension of1-BR or 850-BR primary fibroblast cell pellets (1×106cells) in100μL lysis buffer (40 mM Tris/HCl pH 7.5, 100 mM NaCl,0.1% Tween-20, 1 mM DTT, 1 mM PMSF, 1x EDTA freeprotease cocktail inhibitor), followed by 30 minutes of in-cubation on ice and mild sonication. The WCE was clarifiedby centrifugation for 10 minutes at 4°C at 16000g ina microfuge and the protein concentration quantified usingthe bicinchoninic acid (BCA) assay reagent (ThermoFisher).

Figure 1 TDP2 splice site mutation in an individual from the United States with SCAR23

(A) Pedigree analysis. White symbols:wild-type TDP2. Black symbols: ho-mozygosity for the TDP2 splice sitemutation c.425+1G>A (IVS3+1G>A) inthe proband (“1”). Dotted symbols,heterozygosity for c.425+1G>A (IVS3+1G>A). (B) Sanger sequencing of theproband, demonstrating (left) thehomozygous TDP2 mutation c.425+1G>A (IVS3+1G>A) and (right) the ab-sence of the ZSCAN9 variantc914A>G. Patient sequences areshown at the bottom, and referencesequences are shown at the top. The2 nucleotides relevant to the muta-tions are underlined. (C) Pathologicfeatures of the Irish and US patients.(D) Haplotypes of Irish patients(51670; Irish pedigree) and the cur-rent patient from the United States(1583786) carrying the homozygousc.425+1G>A variant in TDP2. Only thevariants bordering and directly withinthe homozygous regions and witha minor allele frequency of 5% areshown. Variants are indicated by theiraccession number in the dbSNP da-tabase, and the position of c.425+1G>A is in bold. The minimal over-lapping region is delimited byrs9396886 and rs749338 and is 15.1Mb in size. Black bars represent thehomozygous haplotype as inferredfrom exome sequencing data. TDP2 =tyrosyl DNA phosphodiesterase 2.



Clarified WCE (15μg total protein) was incubated with 40nM TDP2 substrate (Cy5-59Tyrosine-ssDNA19-BHQ) orTDP1 substrate (BHQ-ssDNA13-39Tyrosine-Cy5) in re-action buffer (50 mM Tris/HCl pH8.0, 10 mM MgCl2,80 mM KCl, and 1 mM DTT, 0.05% Tween-20) in a totalvolume of 6μL at room temperature, and Cy5 fluorescencewas measured at 640 nm at the indicated time intervals ona BMG PHERAstar plate reader.

DSB repair assaysCells were grown on coverslips until confluent and thentreated for 30 minutes with 25μM etoposide or irradiatedwith x-rays (2 Gy). After treatment, cells were rinsed andfixed for 10 minutes in PBS containing 4% para-formaldehyde at the indicated time points. Cells werepermeabilized (20 minutes in PBS-0.2% Triton X-100),

blocked (1 hour in PBS-5% BSA), and incubated with anti-γH2AX (Millipore, 05-636, 1:2500) and anti-CENP-F(Abcam, ab5, 1:2500) antibodies for 3 hours in PBS con-taining 5% BSA. Cells were then washed (3 × 5 minutes inPBS containing 0.1%Tween-20), incubated for 1h with thecorresponding Alexa Fluor conjugated secondary antibody(1:1000, 5% BSA), and washed again as described earlier.Finally, cells were counterstained with DAPI (Sigma,Gillingham, UK) and mounted in VECTASHIELD (VectorLabs, Peterborough, UK). Images were acquired on anautomated wide-field microscopy Olympus ScanR system(motorized IX83 microscope) with ScanR Image Acquisi-tion and Analysis Software, 20×/0.45 (LUCPLFLN20×PH) dry objectives, and Hamamatsu ORCA-R2 digitalCCD camera C10600. For the analysis of G0/G1 cells inpatient primary fibroblasts, only CENP-F-negative cells

Figure 2 Greatly reduced TDP2 protein and activity in 850-BR patient fibroblasts

(A) TDP2 and actin protein levels in WCE (20 μg total protein) from 1-BR normal and 850-BR patient primary fibroblasts, as measured by immunoblotting. (B)TDP1 and TDP2 biochemical assay. Hydrolysis of a 39-phosphotyrosyl (TDP1 assay) or 59-phosphotyrosyl (TDP2 assay) bond releases the associated Cy5fluorophore and alleviates quenching by the BHQ located on the opposite oligonucleotide terminus, resulting in elevated fluorescence at 640 nm. (C) Real-time (0–30minutes)measurements of 59-tyrosyl DNA phosphodiesterase activity (TDP2 assay) in reaction buffer (as a negative control) orWCE (15μg protein)from 1-BR normal or 850-BR patient primary fibroblasts, using the single-stranded oligonucleotide (40 nM) with a 59-phosphotyrosyl-linked Cy5 fluorophoreand 39-BHQ. (D) Real-time measurements (0–30 minutes) of 39-tyrosyl DNA phosphodiesterase (TDP1 assay) conducted as above but using a 39-phospho-tyrosyl-linked Cy5 fluorophore and 59-BHQ. Data are mean of 3 independent experiments ±SEM. Statistical significance was determined by two way analysisof variance (ANOVA). ***p < 0.001. BHQ = black hole quencher; TDP2 = tyrosyl DNA phosphodiesterase 2; WCE = whole-cell extract.



were scored. For A549 and A549 TDP2−/−, cells were gatedto the G1 population according to the DAPI profile ofScanR Image Analysis Software. For patient fibroblastcomplementation experiments, cells were gated accord-ingly to eGFP expression on ScanR Analysis Software, inaddition to the G1 population according to the DAPIcontent.

ResultsWe recently described mutations in TDP2 in three Irishpatients from the same family with intellectual disability,seizures, and ataxia, a disease now denoted as spinocerebellarataxia 23 (SCAR23)6. Here, we describe a 6-year-old patientin the United States with very similar pathology includingdevelopmental delay, epilepsy, and ataxia and in whom weidentified by whole-exome and Sanger sequencing possessesthe same homozygous splice site mutation in TDP2 (c.425+1G>A) (figure 1, A–C). Whether there is a connection be-tween the current patient and the Irish family is not clear, buta comparison of the WES data from the two families revealedthat the two apparently unrelated affected individuals ofwhom their exome was sequenced share the same homozy-gous haplotype of ;15.1 Mb (figure 1D). It is important tonote that in contrast to the Irish patients, however, the currentpatient lacks the mutation c.914A>G (p.His305Arg) inZNF193, a zinc finger protein of unknown function (figure1B, right). We previously concluded that this variant was not

disease causing in the Irish patients,6 and the absence of thisvariant in the current patient confirms that this and also thatTDP2 mutation is the cause of SCAR23.

To enable additional molecular and cellular analyses ofSCAR23, we generated primary fibroblasts (denoted 850-BR)from a skin biopsy kindly provided by the current patient.Western blotting failed to detect TDP2 protein in the 850-BRpatient fibroblasts (figure 2A), consistent with the lack ofdetectable TDP2 protein previously reported in lympho-blastoid cells from the Irish patients.6 The absence of de-tectable TDP2 protein in the lymphoblastoid cell lines wasexplained by the impact of the TDP2 mutation on splicing,which resulted in nonsense-mediated decay and greatly re-duced (<20%) levels of TDP2 mRNA.6 To confirm the im-pact of the splice site mutation on TDP2 activity, we useda highly sensitive tyrosyl DNA phosphodiesterase bio-chemical assay. This assay uses a single-stranded oligonucle-otide substrate in which hydrolytic release of a Cy5-labeledtyrosine moiety present on one terminus of the oligonucle-otide results in increased fluorescence due to evasion ofa black hole quencher present on the opposite terminus(figure 2B). By situating the fluorescent-labeled tyrosine oneither the 39 or 59 terminus, this assay can detect TDP1 orTDP2 activity, respectively. Although whole-cell extract fromnormal 1-BR human fibroblasts exhibited robust 59-tyrosylDNA phosphodiesterase activity, 850-BR patient fibroblastsdid not (figure 2C). The small amount of activity observed in

Figure 3 Reduced DSB repair in TDP2-mutated patient fibroblasts and A549 cells after topoisomerase 2-induced DNAdamage

(A) DSBs were measured by γH2AXimmunostaining in normal 1-BR andpatient 850-BR primary fibroblasts be-fore and after treatment for 30 minuteswith DMSO vehicle or 25 μM etoposide(“0”), followed by subsequent in-cubation in a drug-free medium for theindicated repair periods. (B) DSBs weremeasured as above, in wild-type A549cells and in A549 cells in which TDP2wasdeleted by CRISPR/Cas9 gene editing(TDP2−/− A549). The level of TDP2 andactin (loading control) in the cell linesused for these experiments is shown byWestern blotting on the right. (C) DSBswere measured as above in hTERT-im-mortalized 1-BR fibroblasts and 850-BRfibroblasts and in the latter cells aftertransfection with the empty GFP vectoror vector encoding recombinant humanTDP2-GFP. The level of TDP2 and actin(loading control) in the cell lines used forthese experiments is shown byWesternblotting on the right. Data are mean ±SEMof 3 independent experiments, andstatistically significant differences weredetermined by the t test (ns = not sig-nificant, *p < 0.05; **p < 0.01; ***p <0.001). DSB = double-strand break;GFP = green fluorescent protein; TDP2 =tyrosyl DNA phosphodiesterase 2.



850-BR cell extract, compared with reactions supplementedwith reaction buffer alone, reflects nucleases in the cell extractrather than residual TDP2 activity because similar results wereobserved when we used cell extract from human cells in whichTDP2 was deleted by CRISPR/Cas9 (unpublished observa-tions). The lack of TDP2 activity in patient cells did notreflect differences in the technical quality of cell extracts be-cause the level of TDP1 activity in these extracts was similar tothat in normal human 1-BR fibroblasts (figure 2D). Weconclude from these data that the TDP2 splice site mutationgreatly reduces and most likely ablates both TDP2 proteinand activity in 850-BR patient fibroblasts.

Next, we addressed the impact of the TDP2 splice site mu-tation on nuclear DSB repair rates in 850-BR patient fibro-blasts using immunofluorescent detection of γH2AX as anindirect measure of DSBs.12 Although similar levels of nuclearγH2AX foci were present in normal and patient fibroblastsimmediately after treatment with etoposide for 30 minutes,the levels of these γH2AX foci decreased far more slowly inthe patient fibroblasts during subsequent incubation in a drug-free medium (figure 3A). However, DSB repair was com-pleted in patient cells within 8–24 hours, consistent with theestablished existence of alternative, nuclease-dependent

mechanisms for repair of TOP2-induced DSBs in humancells.4,13,14 Similar results were observed in human A549 cellsin which TDP2 was mutated by CRISPR/Cas9 gene editing,confirming the importance of TDP2 for repair of TOP2-induced DSBs (figure 3B). More importantly, complemen-tation of 850-BR cells with expression construct encodingrecombinant human TDP2 restored normal rates of nuclearDSB repair, confirming that the DNA repair defect in thepatient fibroblasts was the result of theTDP2mutation (figure3C). This defect in DSB repair was accompanied by cellularhypersensitivity to TOP2-induced DSBs because both 850-BR patient fibroblasts and TDP2−/− A549 cells were hyper-sensitive to etoposide in clonogenic survival assays (figure 4, Aand B). In contrast to etoposide-induced DSBs, both DSBrepair rates and levels of cell survival were normal in 850-BRpatient fibroblasts after treatment with ionizing radiation(figure 4, C and D), confirming the specificity of the DNArepair defect in the patient’s cells for DNA breaks induced byTOP2.

Finally, because analyses of muscle biopsy from the patient(see the case report) suggested the presence of possibledefects in the ETC, we examined 850-BR patient fibroblastsfor defects in mitochondrial function. However, the patient

Figure 4Hypersensitivity of TDP2-mutated patient fibroblasts andA549 cells to DNAdamage induced by etoposide but notγ-rays

(A) Clonogenic survival of 1-BR normaland 850-BR patient primary fibro-blasts in a medium containing the in-dicated concentrations of etoposide.(B) Clonogenic survival of wild-typeA549 and TDP2−/− A549 cells in a me-dium containing the indicated con-centrations of etoposide. (C)Clonogenic survival of 1-BR normaland 850-BR patient primary fibro-blasts as above, following γ-irradia-tion. (D), DSBs were measured byγH2AX immunostaining in normal 1-BR and patient 850-BR primary fibro-blasts before and after the indicatedperiods after γ-irradiation (2 Gy). Dataare mean ± SEM of 3 independentexperiments, and statistically signifi-cant differences were determined bytwo-way ANOVA (ns = not significant,***p < 0.001). TDP2 = tyrosyl DNAphosphodiesterase 2.



fibroblasts failed to exhibit significant defects in respiratorychain complexes in biochemical assays compared witha range of different normal (control) human fibroblasts(figure 5A.a). Similar results were observed when wecompared wild-type A549 cells with A549 cells in which

TDP2 was mutated by gene editing (figure 5A.b). We alsofailed to identify any impact of the TDP2 mutation or de-letion in 850-BR fibroblasts and A549 cells, respectively, inthe level of respiratory chain proteins as measured by im-munoblotting (figure 5B).

Figure 5 Mitochondrial functionality in normal and TDP2-mutated cells

(A) Activity ofmitochondrial respiratorycomplexes in normal (red bars) andpatient 850-BR (blue bars) primaryhuman fibroblasts (A.a) and in wild-type human A549 (red bars) andTDP2−/− A549 cells (blue bars) (A.b) wasdetermined as previously described.15

The mean enzyme activities in controlcells (n = 8) are set to 100%, and errorbars represent SD. (B) Levels of mito-chondrial respiratory complex pro-teins in age-matched controlfibroblasts (C1 andC2) andpatient 850-BR fibroblasts (B.a) and in wild-typeA549 cells and TDP2−/− A549 cells (B.b),as measured by immunoblotting forsubunits of CI (NDUFB8), CII (SDHA), CIII(UQCRC2), CIV (COXI), and CV (ATP5A)and using VDAC1 as a mitochondrialloading control. (C) Subcellular locali-zation of TDP2 in HeLa cells (C.a) and innormal (1-BR) and patient (850-BR)primary human fibroblasts (C.b). Left,cell-equivalent amounts of HeLa totalcell lysate (30 g total protein; lane 1),cell lysate depleted of mitochondria(“post-mito spin”; lane 2), mitochondriatreated with proteinase K (to removeproteins associated with the outermembrane; lane 3), mitoplasts (mito-chondrial matrix plus inner membraneproteins; lane 4), mitoplasts treatedwith proteinase K (lane 5), and innermembrane mitochondrial proteins(extractedwith sodium carbonate; lane6) were immunoblotted for TDP2 andfor protein markers of the cytosol(eIF4E) and each of the mitochondrialcompartments, intermembrane space(AIF), mitochondrial matrix (EF-Tu), andinner mitochondrial membrane(NDUFB8). Right, cell-equivalentamounts of 1-BR or 850-BR fibroblasttotal cell lysates (30μg protein) de-pleted of mitochondria (“post-mitospin”) and of mitoplasts were immu-noblotted for TDP2 as above. The po-sition of full-length TDP2 (arrow) anda nonspecific band detected by theantibody (asterisk; *) are indicated.TDP2 = tyrosyl DNA phosphodiester-ase 2.



We next fractionated HeLa cells and both 1-BR and 850-BRfibroblasts into subcellular components, including intact mi-tochondria and mitoplasts, to examine whether TDP2 is lo-cated within the mitochondria and thus in the correct locationto repair mitochondrial DNA (figure 5C). Most of the ma-terial detected by the anti-TDP2 antibody and whichmigratedto the position expected for TDP2 was present, as expected, inthe cellular fractions containing cytosolic and nuclear proteins(figure 5C.a, lanes 1 and 2). However, a very small amount ofthis material was detected in intact mitochondria and mito-plasts (figure 5C.a, lanes 3–5). In 1-BR normal humanfibroblasts, most of the material detected by anti-TDP2 an-tibody was again present in the cytosolic/nuclear fraction, andthis material was TDP2 because it was absent from parallelpreparations from patient 850-BR fibroblasts (figure 5C.b).However, if TDP2 was present in mitoplasts prepared from1-BR cells, it was below the level of detection in our experi-ments (figure 5C.b, arrow). We did note the presence ofa slower migrating band that was detected by the TDP2 an-tibody in both HeLa and 1-BR mitoplasts, but this band wasalso present in parallel preparations from 850-BR patientfibroblasts, indicating that it was not TDP2 (figure 5C.b,asterisk). Collectively, although we cannot rule out the pres-ence of a small amount of endogenous TDP2 in mitochon-dria, we cannot detect defects in mitochondrial function inSCAR23 patient fibroblasts. Consequently, we suggest thatthe neurologic disease pathology in SCAR23 is most likely theresult of a defect in nuclear DSB repair.

DiscussionWe recently identified TDP2 mutations in a recessive hered-itary genetic disorder associated with intellectual disability,seizures, and ataxia, now denoted as spinocerebellar ataxiaautosomal recessive 23 (SCAR23)6. In the three Irish siblingsoriginally described, we also identified the mutation c.914AG(p.His305Arg) in ZNF193, a zinc finger protein of unknownfunction, which we concluded was not disease causing butwhich we could not exclude as a contributor to the disease.6

However, in contrast to the Irish patients, the new patientidentified in the United States and reported here harbors thesame TDP2-associated haplotype but lacks the mutation inZNF193, ruling out a contribution of the latter and confirmingthe TDP2 mutation as the cause of SCAR23.

The TDP2-mutated patient fibroblasts established here are thefirst such reported and have provided valuable informationconcerning the molecular and cellular defect in SCAR23.Consistent with TDP2 defects in other cell types, we observedreduced rates of DSB repair and elevated cellular sensitivityafter treatment with the TOP2 poison etoposide. These phe-notypes were the result of the TDP2 mutation in the patientfibroblasts because they were phenocopied in TDP2−/− A549cells created by gene editing and were complemented by thereintroduction of the wild-type human TDP2 transgene. Ofinterest, the magnitude of these phenotypes is greater in patient

fibroblasts and gene-edited A549 cells than what we have seenpreviously with our previous gene-edited cell lines.9 We believethat this is because the cell lines reported here are effectivelyTDP2 null, whereas our previously generated TDP2 gene-edited human cells retained a minor isoform of TDP2 that isexpressed at low levels and is primarily cytoplasmic.

It is notable that TOP2 poisons are used widely in the clinic totreat a variety of cancers, and we point out that the use of theseagents in patients with TDP2 mutations should be avoided orat the very least treated with extreme caution. AlthoughTDP2-defective cancer cells will be hypersensitive to this typeof chemotherapy, normal cells will also be hypersensitive, asillustrated in the current work by the etoposide sensitivityobserved in both TDP2−/− A549 lung cancer cells andSCAR23 primary fibroblasts.

The presence of mitochondrial dysfunction in the SCAR23patient described here is suggested by ETC studies and the highlactate and lactate/pyruvate ratio in muscle (see the PatientCase Study in the Methods section). In addition, the currentpatient exhibits phenotypes consistent with mitochondrialdysfunction such as hypotonia, low energy, and fatigability. It isunclear whether phenotypes consistent with mitochondrialdefects are also present in the Irish patients, although nonewere originally noted.6 Unfortunately, neither fibroblasts normuscle biopsy are available from these patients to address thispossibility. The lack of detectable mitochondrial defect in pri-mary skin fibroblasts from the current patient and in gene-edited A549 cells does not support amajor role for TDP2 in therepair of mitochondrial DNA. However, the absence of mito-chondrial defects in skin fibroblasts from patients with mito-chondrial defects in muscle is not uncommon.15 Nevertheless,given the dramatic impact of the TDP2 mutation on nuclearDNA repair, we suggest that any association of SCAR23 withmitochondrial disease is most likely an indirect or secondarydysfunction resulting from a nuclear disorder.16

Author contributionsG. Zagnoli-Vieira, F. Bruni, K. Thompson, and L. He designedand conducted mitochondrial experiments. S. Walkerdesigned TDP assay. R. Taylor supervised mitochondrialexperiments. A.P.M. de Brouwer conducted haplotype anal-ysis. D. Niyazov identified and assessed the patient andidentified the TDP2 mutation. K.W. Caldecott designed, di-rected, and coordinated the project and wrote the manuscript.

AcknowledgmentThe authors thank the patient and the patient’s family, andLimei Ju/Elena Korneeva for generation of primary 850-BRpatient fibroblasts and hTERT-immortalized derivatives.

Study fundingThis study was funded by Programme grants to KWC fromthe MRC (MR/P010121/1), CRUK (C6563/A16771), ERC(SIDSCA Advanced Grant, 694996), and by an MRC PhDStudentship for GZV (MRN/N50189X/1).



DisclosureG. Zagnoli-Vieira has no financial disclosures to report.F. Bruni serves on the editorial boards of Biomarkersjournal and Brain and Nervous System Current Research andreceives research support from ANVUR. K. Thompson,L. He, S. Walker, R. Taylor, and A.P.M de Brouwer reportno disclosures. D. Niyazov has served on the scientificadvisory boards of Alexion Pharmaceuticals and Mal-linckrodt ARD Inc.; has received funding for travel/speaker honoraria from Sanofi Genzyme Corporation;serves on the editorial board of Molecular Syndromology;has consulted with the Phase II clinical trial of Elamipre-tide in Mitochondrial Myopathy; is a current member ofthe Sanofi Genzyme Corporation Speakers’ Bureau; andhas served as an expert witness for a legal case regardingvaccine injury in a child with mitochondrial disease.K.W. Caldecott reports no disclosures. Full disclosureform information provided by the authors is available withthe full text of this article at Neurology.org/NG.


References1. Lindahl T. Instability and decay of the primary structure of DNA. Nature 1993;362:

709–715.2. Goodarzi AA, Jeggo PA. The repair and signaling responses to DNA double-strand

breaks. Adv Genet. Elsevier 2013;82:1–45.

3. Cortes Ledesma F, El-Khamisy SF, ZumaMC,Osborn K, Caldecott KW. A human 59-tyrosyl DNA phosphodiesterase that repairs topoisomerase-mediated DNA damage.Nature 2009;461:674–678.

4. Gomez-Herreros F, Romero-Granados R, Zeng Z, et al. TDP2-Dependent non-homologous end-joining protects against topoisomerase II-induced DNA breaks andgenome instability in cells and. Vivo Plos Genet 2013;9:e1003226.

5. Zeng Z, Cortes Ledesma F, El-Khamisy SF, Caldecott KW. TDP2/TTRAP is themajor 59-tyrosyl DNA phosphodiesterase activity in vertebrate cells and is critical for cellular re-sistance to topoisomerase II-induced DNA damage. J Biol Chem 2011;286:403–409.

6. Gomez-Herreros F, Schuurs-Hoeijmakers JHM, McCormack M, et al. TDP2 protectstranscription from abortive topoisomerase activity and is required for normal neuralfunction. Nat Genet 2014;46:516–521.

7. Bruni F, Gramegna P, Oliveira JMA, Lightowlers RN, Chrzanowska-LightowlersZMA. REXO2 is an oligoribonuclease active in human mitochondria. PLoS One2013;8:e64670.

8. Hoch NC, Hanzlikova H, Rulten SL, et al. XRCC1mutation is associated with PARP1hyperactivation and cerebellar ataxia. Nature 2016;541:87–91.

9. Gomez-Herreros F, Zagnoli-Vieira G, Ntai I, et al. TDP2 suppresses chromosomaltranslocations induced by DNA topoisomerase II during gene transcription. NatCommun 2017;8:233.

10. Thomson G, Watson A, Caldecott K, et al. Generation of assays and antibodies tofacilitate the study of human 59-tyrosyl DNA phosphodiesterase. Anal Biochem 2013;436:145–150.

11. Thompson K, Majd H, Dallabona C, et al. Recurrent de novo dominant mutations inSLC25A4 cause severe early-onset mitochondrial disease and loss of mitochondrialDNA copy number. Am J Hum Genet 2016;99:860–876.

12. Rogakou EP, Pilch DR, Orr AH, Ivanova VS, Bonner WM. DNA double-stranded breaksinduce histone H2AX phosphorylation on serine 139. J Biol Chem 1998;273:5858–5868.

13. Hoa NN, Shimizu T, Zhou Z-W, et al. Mre11 Is Essential for the Removal of LethalTopoisomerase 2 Covalent Cleavage Complexes. Mol Cell 2016;64:580–592.

14. Zagnoli-Vieira G, Caldecott KW. TDP2, TOP2, and SUMO: what is ZATT about?Cell Res 2017;17:182–1406.

15. Kirby DM, Thorburn DR, Turnbull DM, Taylor RW. Biochemical assays of re-spiratory chain complex activity. Methods Cell Biol 2007;80:93–119.

16. Niyazov DM, Kahler SG, Frye RE. Primary mitochondrial disease and secondarymitochondrial dysfunction: importance of distinction for diagnosis and treatment.Mol Syndromol 2016;7:122–137.




CLINICAL/SCIENTIFIC NOTES OPEN ACCESS

Atypical Alexander disease with dystonia,retinopathy, and a brain mass mimickingastrocytomaKeren Machol, MD, PhD, Joseph Jankovic, MD, Dhanya Vijayakumar, MD, Lindsay C. Burrage, MD, PhD,

Mahim Jain, MD, PhD, Richard A. Lewis, MD, Gregory N. Fuller, MD, PhD, Mingchu Xu, PhD,

Marta Penas-Prado, MD, Maria K. Gule-Monroe, MD, Jill A. Rosenfeld, MS, CGC, Rui Chen, PhD,

Christine M. Eng, MD, Yaping Yang, PhD, Brendan H. Lee, MD, PhD, Paolo M. Moretti, MD,

Undiagnosed Diseases Network, and Shweta U. Dhar, MD, MS


Correspondence

Dr. Dhar

[email protected]

Alexander disease (AD) is an autosomal dominant progressive astrogliopathy caused bypathogenic variants in glial fibrillary acidic protein (GFAP).1 Clinical presentation of ADincludes infantile AD, characterized by psychomotor retardation, seizures, pyramidal signs, andmegalencephaly; juvenile AD, characterized by bulbar/pseudobulbar signs, hyperreflexia, lowerlimb spasticity, ataxia, loss of intellectual function, and macrocephaly; and adult-onset AD,characterized by progressive bulbar symptoms, ataxia, palatal myoclonus, bladder dysfunction,and spastic paraparesis.1

Clinical reportThe patient is a 35-year-old woman with progressive left hemidystonia, retinitis pigmentosa(RP), and history of brain tumor. At age 8 years, she developed vision loss, limp, afferentpupillary defect, and optic disc edema. A pilocytic astrocytoma, World Health Organizationgrade I, was diagnosed and resected. At age 14 years, she developed gait abnormality andprogressive left hemidystonia that responded well to botulinum toxin injections. She thendeveloped bilateral optic disc fibrosis, bitemporal hemianopia, and bilateral posterior sub-capsular cataracts and was diagnosed with RP at 22 years (figure, A–D). She also reportedchronic constipation, functional megacolon, uterine leiomyomas, and dysphagia.

Family history includes nonconsanguineous unaffected parents. Physical examination showedrelative macrocephaly, absent ocular horizontal and vertical pursuit, dysarthria, circumductiongait and left-sided arm dystonia, muscle atrophy, hyperreflexia, and ankle clonus. Brain MRIdemonstrated a frontal resection cavity, cystic encephalomalacia, and fluctuating enhancementof the hypothalamus. At age 29 years, a new left middle cerebellar peduncle lesion was identified(figure, E and F).

The patient was enrolled in the NIH Undiagnosed Diseases Network (UDN). Whole-exomesequencing (WES) for her and her mother was completed. Her father was deceased at the timeof evaluation.

From the Department of Molecular and Human Genetics (K.M., L.C.B., M.J., R.A.L., M.X., J.A.R., R.C., C.M.E., Y.Y., B.H.L., P.M.M., S.U.D.), Department of Neurology (J.J., D.V., P.M.M), andDepartment of Ophthalmology (R.A.L.), Baylor College of Medicine; Department of Pathology (G.N.F.), Department of Neuro-Oncology (M.P.-P.), and Department of DiagnosticImaging (M.K.G.-M.), The University of Texas MD Anderson Cancer Center; Michael E. DeBakey VA Medical Center (P.M.M.); Baylor Genetics (C.M.E., Y.Y.); and Department ofMedicine (S.U.D.), Baylor College of Medicine, Houston, TX.


The Article Processing charge was funded by NIH.

Coinvestigators are listed at links.lww.com/NXG/A59.








ResultsWES revealed a novel heterozygous variant of uncertainsignificance (VUS) in GFAP (MIM_137780) c.989G>C(p.R330P, NM_002055) and a heterozygous VUS c.6196G>A(p.D2066N, NM_006269) in RP1 (MIM_603937), both ab-sent in the mother. The parental origin of the GFAP mutatedallele could not be concluded based on WES data and avail-able DNA samples. The variant in RP1 was reported in 127heterozygotes among 138,396 unrelated, unaffected individualsin gnomAD database (gnomad.broadinstitute.org). Both var-iants are predicted as damaging in SIFT (sorting intolerantfrom tolerant; sift.jcvi.org/) and Polyphen-2 (genetics.bwh.harvard.edu/pph2). No other contributory variants weredetected in 256 known retinal disease genes.

Pathology slides from the initial brain mass were re-examined.Histologic examination showed features of AD, includingprofuse Rosenthal fiber formation, Rosenthal fiber-like eosin-ophilic cytoplasmic inclusions in astrocyte cell bodies, andscattered cells with markedly atypical nuclei (figure, G and H).

DiscussionRadiologic findings in AD vary with age at onset, and typicalcharacteristics have been described for the various forms ofAD. Serial radiologic imaging of our patient showed sequelaeof right frontal lobe resection, development of periventricularwhite matter changes, and fluctuating brainstem and hypo-thalamic lesions. Waxing and waning imaging findings, in theabsence of treatment, were inconsistent with a brain tumor.

Figure Clinical, histologic, and imaging findings

A 35-year-old woman with juvenile-onset Alexander disease (AD) (A and B).Note relative macrocephaly (FOC 56cm, 94th centile) and esotropia. Thephotomontage of the right (C) and left(D) eyes shows the atrophic gliosisfrom prior optic disc edema and themarked retinal atrophy and secondarypigmentary degeneration. The retinalvasculature appears to show glialsheathing and peripheral vaso-obliter-ation in each eye. Brain MRI with con-trast at 29 years (E and F). (E) Axial T2image demonstrates a T2 hyperintenselesion centered in the left middle cer-ebellar peduncle (yellow arrow). (F)Axial T1 image with contrast shows anarea of central enhancement withinthe lesion (yellow arrow). Partially im-aged is cystic encephalomalacia withinthe frontal lobes bilaterally (yellow as-terisk) with surgical changes in the rightfrontal lobe not included. Slides frompatient’s brain lesion (G and H) dem-onstrating AD histologic features. Theprototypical histologic feature of AD isprominent Rosenthal fiber formation(G, solid arrows). An additional featureunique to AD is the presence ofRosenthal fiber-like eosinophilic cyto-plasmic inclusions in astrocyte cellbodies (G, open arrows). The cell bodyinclusions are generally not seen inother Rosenthal fiber–rich conditionssuch as low-grade primary braintumors. Astrocytes with markedlyatypical nuclei (H, solid arrow) are alsoa characteristic morphologic feature ofAD and can mimic the neoplastic cellsof astrocytoma or ganglioglioma. FOC= fronto-occipital circumference.


http://gnomad.broadinstitute.org

http://sift.jcvi.org/

http://genetics.bwh.harvard.edu/pph2

http://genetics.bwh.harvard.edu/pph2


Therefore, despite her previous diagnosis of a brain tumor,these new findings were interpreted as non-neoplastic, andthe working diagnosis was non-MS demyelinating lesions.Similar atypical MRI findings had been described before inpatients with molecularly proven AD.2 Dystonia and retinalabnormalities have never been previously reported with AD.Left hemidystonia in our patient may be secondary to AD-related encephalomalacia or a possible delayed consequenceof right frontal brain resection.3 TheRP1VUS in our patient isless likely to be pathogenic, given the presence of multipleheterozygous carriers in population databases. It is unclearwhether RP is a phenotypic expansion of AD or an unrelatedfinding (i.e., dual genetic diagnosis).4

Several histologic features of AD can mimic brain tumors.Rosenthal fiber formation in astrocytes of the brain and spinalcord, a pathologic hallmark of AD, is also seen in severallow-grade primary brain tumors (e.g., pilocytic astrocytoma,ganglioglioma, and pleomorphic xanthoastrocytoma). Cytolog-ically atypical astrocytes, a recognized feature of AD,5 can alsolead tomisdiagnosis of pilocytic astrocytoma/ganglioglioma andunnecessary invasive interventions. The presence ofRosenthal fiber–like eosinophilic cytoplasmic inclusions inastrocyte cell bodies serves as a “red flag,” raising the index ofsuspicion of AD.

Over 120 pathogenic variants inGFAP have been reported witha possible dominant gain-of-function mechanism.6 GFAPp.R330P substitutes arginine, located in the evolutionarily well-conserved structure of GFAP, with proline. Another change inthe same amino acid, p.R330G, was associated previously withadult AD.7 The potential functional significance of the p.R330Psubstitution, our patient’s phenotype, and the histologic find-ings all strongly supportGFAP p.R330P being a disease-causingvariant.

Although relative macrocephaly, spasticity, ocular movementabnormalities, and autonomic disturbance (functional mega-colon), as seen in our patient, have been described in juvenileAD,1 the diagnosis of AD was delayed because of an atypicalpresentation, including RP, left hemidystonia, atypical MRIfindings, and histologic features confounding the diagnosis. Thiscase emphasizes the clinical, histologic, and radiologic challengesin diagnosing AD and demonstrates the importance of consid-ering AD in the differential diagnosis of specific Rosenthalfiber–rich brain lesions such as pilocytic astrocytoma.

Author contributionsK. Machol: drafting the manuscript and analysis of whole-exome sequencing. J. Jankovic: neurologic management of thepatient. D. Vijayakumar: drafting the manuscript. L.C. Bur-rage and M. Jain: analysis of whole-exome sequencing. R.A.Lewis: ophthalmologic evaluation of the patient and reviewand editing of the manuscript. G.N. Fuller: revision of histo-pathology and drafting the manuscript. M. Xu: review ofRP-related genes in whole-exome sequence. M. Penas-Prado:oncologic management of the patient, drafting the manuscript,

and critical revision of the manuscript. M. K. Gule-Monroe:radiology interpretation and drafting the manuscript.J. A. Rosenfeld: coordination of patient evaluation and criticalrevision of the manuscript. R. Chen: review of RP-relatedgenes in whole-exome sequence. Y. Yang and C.M. Eng:analysis of whole-exome sequencing. B.H. Lee: critical re-vision of the manuscript. P.M. Moretti: neurologic evaluationof the patient and drafting the manuscript. S.U. Dhar: draftingthe manuscript, critical revision of the manuscript, and patientmanagement.

AcknowledgmentThe authors thank the patient and her family for participatingin this research. They thank Alyssa Tran, Deanna Erwin,and Alex Espana in the Molecular and Human GeneticsDepartment, Baylor College of Medicine for administrativeassistance.

Study fundingThis work was supported by the BCM Intellectual and De-velopmental Disabilities Research Center (HD024064) fromthe NICHD and NIH/NHGRI U01 HG007709 Clinical Sitefor a UDN. The content is solely the responsibility of theauthors and does not necessarily represent the official views ofthe NIH. The Department of Molecular and Human Geneticsat Baylor College of Medicine receives revenue from clinicalgenetic testing offered by Baylor Genetics.

DisclosureK.Machol reports no disclosures. J. Jankovic has served on thescientific advisory boards of, received gifts from, and receivedfunding for travel or speaker honoraria from Adamas Phar-maceuticals Inc., Allergan Inc., and Teva Pharmaceutical In-dustries; has served on the editorial boards ofActa NeurologicaScandinavica, the Journal of Neurological Sciences, NeurologyMedlink, Neurotherapeutics, Journal of Parkinson’s Disease,and Toxins; receives publishing royalties from Cambridge,Elsevier, Future Science Group, Hodder Arnold, LippincottWilliams and Wilkins, and Wiley-Blackwell; and has receivedresearch support from Allergan Inc., CHDI Foundation,Civitas/Acorda Therapeutics, Dystonia Medical ResearchFoundation, Huntington Study Group, Ipsen Limited, KyowaHaako Kirin Pharma Inc., Lundbeck Inc., Medtronic, MerzPharmaceuticals, Michael J. Fox Foundation for ParkinsonResearch, NIH, National Parkinson Foundation, ParkinsonStudy Group, Pfizer, Prothena Biosicneces Inc., PsyadonPharmaceuticals Inc., St. Jude Medical, and Teva Pharma-ceutical Industries Ltd.; D. Vijayakumar reports no dis-closures. L.C. Burrage has received funding for travel orspeaker honoraria from the Burroughs Welcome, the NIH,and the Society for Inherited Metabolic Disorders and hasreceived research support from the NIH/NIDDK, the NIHRDCRN, the Burroughs Welcome Fund, and the LigumsFamily. M. Jain, R.A. Lewis, G.N. Fuller, and M. Xu report nodisclosures. M. Penas-Prado has received funding for travelfrom AGIOS and LILLY and serves on the editorial boardof Neuro-Oncology Practice. M.K. Gule-Monroe reports no



disclosures. J.A. Rosenfeld serves on the editorial board ofPrenatal Diagnosis and Molecular Syndromology and receivesresearch support from the NIH/National Human GenomeResearch Institute. R. Chen has received research supportfrom the NIH. C.M. Eng is employed by Baylor College ofMedicine, which has a joint venture withMiraca Genetics Lab.Y. Yang reports no disclosures. B.H. Lee serves on the edi-torial advisory boards of the Journal of Clinical Investigation,Human Molecular Genetics, and the Journal of Bone MineralResearch and holds a patent license for helper dependentadenoviral gene therapy for osteoarthritis and phenylbutyrateuse in Maple Syrup Urine Disease. P.M. Moretti receivedresearch support from the NHGRI/NIH. S.U. Dhar isemployed by the Baylor College of Medicine, which hasa joint venture with Miraca Life Services. Full disclosure form

information provided by the authors is available with the fulltext of this article at Neurology.org/NG.


References1. Prust M, Wang J, Morizono H, et al. GFAP mutations, age at onset, and clinical

subtypes in Alexander disease. Neurology 2011;77:1287–1294.2. Barreau P, Prust MJ, Crane J, et al. Focal central white matter lesions in Alexander

disease. J Child Neurol 2011;26:1422–1424.3. Wijemanne S, Jankovic J. Hemidystonia-hemiatrophy syndrome. Mov Disord 2009;

24:583–589.4. Posey J, Harel T, Liu P, et al. Resolution of disease phenotypes resulting from

multilocus genomic variation. N Engl J Med 2017;376:21–31.5. Borrett D, Becker L. Alexander’s disease: a disease of astrocytes. Brain 1985;108:

367–385.6. Quinlan R, Brenner M, Goldman J, Messing A. GFAP and its role in Alexander

disease. Exp Cell Res 2007;313:2077–2087.7. Balbi P, Seri M, Ceccherini I, et al. Adult-onset Alexander disease. J Neurol 2007;255:

24–30.





De novo DNM1L mutation associated withmitochondrial epilepsy syndrome with feversensitivityEmma Ladds, MBBCh, Andrea Whitney, MB, BS, Eszter Dombi, BSc, Monika Hofer, BMSc, MBChB,

Geetha Anand, MRCPCH, MD, Victoria Harrison, MBChB, Carl Fratter, MPhil, MChem, Janet Carver, MSc,

Ines A. Barbosa, DPhil,Michael Simpson, PhD, Sandeep Jayawant,MD, FRCPCH, and JoannaPoulton, DM, FRCP


Correspondence

Dr. Poulton

[email protected]

Catastrophic epileptic encephalopathy of unclear etiology following a mild metabolic insultgenerally has a poor outcome. Here, we present 2 such unrelated individuals in whom whole-exome sequencing identified the same de novo recurrent mutation (c.1207C>T p.Arg403Cys)in the gene encoding the guanosine triphosphatase (GTPase) Dynamin-1 like Protein(DNM1L) (reference sequence NM_012062.4).

The dynamic fission and fusion of the intracellular mitochondrial network are essential tofacilitatemitophagy and thusmitochondrial quality and function.1 Duringmitochondrial division,the GTPaseDNM1L forms multimeric collars at specific fission sites, constricting portions of themitochondrial reticulum and generating fragments for engulfment and degradation.2

DNM1L has been implicated in several presentations of refractory epilepsy.3 Both of ourpatients exhibited signs of preexisting developmental delay and presented with epilepsy during,or recently following, a febrile illness or exercise. Elevated lactate levels, epilepsia partialiscontinua, nonspecific imaging, and evidence of lipid storage myopathy all support mitochon-drial dysfunction (See table for presentation summary and e-case report for details, links.lww.com/NXG/A63). This evidence supports an etiological role for DNM1L in mitochondrialepilepsy syndrome with fever sensitivity (MEFS).

MethodsWe used IN Cell Analyzer 1000 (IN Cell 1000), a previously validated high-throughput imagingmethod for quantifying mitophagy and mitochondrial DNA (mtDNA) in cultured fibroblastsfrom patients compared with cultures derived from karyotypically normal controls. Cells wereimmunostained for the autophagy marker Light Chain 3 (LC3) and the mitochondrial importreceptor, translocase of outer membrane 20 (TOM20) and analyzed with IN Cell1000. Thereadout for mitophagy was colocalization of LC3 puncta with TOM20-positive mitochondria.

ResultsWe showed that the mitochondria in fibroblasts from both patients are lengthened andhyperpolarized (figure e-1, A and B, links.lww.com/NXG/A62). Total mitophagic flux wasincreased, showing that mitophagy is activated (figure e-1C, i). This is consistent with the

From the Harvard Chan School of Public Health (E.L.), Harvard University, Boston, MA; Department of Paediatrics (A.W.), University Hospital Southampton NHS Foundation Trust;Nuffield Department Women’s + Reproductive Health (E.D., J.C., J.P.), University of Oxford, The Women’s Centre; Department of Neuropathology (M.H.), Oxford University HospitalsNHS Foundation Trust; Oxford Children’s Hospital (G.A., S.J.), Oxford University Hospitals NHS Foundation Trust; Wessex Clinical Genetics Service (V.H.), University Hospital South-ampton NHS Foundation Trust; and Department of Medical and Molecular Genetics (C.F., I.A.B., M.S.), King’s College London School of Basic and Medical Biosciences, London, UnitedKingdom.












mtDNA depletion documented in fibroblasts (figure e-1C, iii)and the borderline low mtDNA content in skeletal musclefrom patient 1 (table).

DiscussionIn 2 patients, we identified a de novo dominant mutation inDMN1L, the same mutation having now been identified in 4unrelated patients with refractory epilepsy.3 Presentationfeatures and investigation findings supported an underlyingmitochondrial pathology. Altered mitochondrial dynamics arenow a well-established cause of disease (see supplementaryinformation for reference, links.lww.com/NXG/A63), anddefective mitochondrial fission may both cause synaptic dys-function1 and impair responses to infection.4

In mice, constitutive homozygous knockouts of Drp1 (themurine homolog of DNM1L) do not survive embryogenesis,while conditional ablation leads to developmental defects,both associated with abnormal mitochondrial fission.5 Indi-viduals heterozygous for DNM1L p.Arg403Cys display

a milder phenotype, both in terms of mitochondrial structuraland functional abnormalities and symptom severity.

Previously described cases similarly display several years ofrelatively normal development followed by severe, refractoryepilepsy following a mild metabolic insult, vaccinations, or low-grade fever, resulting in profound global developmental delay.3

As in case 1, nonspecific thalamic hyperintensities were seen onMRI scans on the 2 previously reported probands with thep.Arg403Cys mutation.3 Case 2 demonstrated diffuse cerebralvolume atrophy. These changes are in keeping with previouslyreported nonspecific T2 MRI hyperintensities and cerebral orcerebellar atrophy seen in other mitochondrial disorders. Itremains to be seen whether MEFS is part of the same clinicalspectrum of other conditions associated with status epilepticusrelated to febrile illnesses, i.e., new-onset refractory status epi-lepticus or febrile illness–related epilepsy syndrome.6

DNM1L is required for division of mitochondria and perox-isomes, interacting with receptor Mff and endoplasmic re-ticulum (ER) components. The missense mutation shared in

Table Comparison of clinical presentations and investigations

Case 1 Case 2

Prepresentation development Moderate language delay Mild general developmental delay

Seizure features Age 3: focal status epilepticus; continuous partialseizures and a right hemiparesis

Age 5: asymmetric generalized tonic-clonicseizures provoked by febrile illness.

Age 6: focal and convulsive status epilepticus

Persistent right-sided clonic and occasional non-or generalized convulsive status epilepticus

Etiological insult 5 days after influenza vaccination; concurrentmild febrile illness

Shortly after febrile illness or exercise

Biochemical findings Increased lactate level (2.7 mmol/L) Increased lactate level (2.5 mmol/L)

EEG findings Left frontotemporal eliptogenic focus Focal right-sided clonic seizures associated withleft frontal ictal onset; nonconvulsive statusepilepticus with bilateral slow spike waves

MRI findings Transient left thalamic lesion with progression ofT2 white matter hyperintensity

Normal morphology and myelination; globalbilateral cerebral volume loss

Muscle biopsy Variable muscle fiber size and type II atrophy Type II atrophy and cytochrome oxidase negative(COX-negative) succinate dehydrogenase-positive (SDH-positive) fibers

Muscle lipid Increased Prominent in type I fibers

Muscle mtDNA 33% of expected (borderline low) 71% of expected (normal)

Mitochondrial appearance Prominent—consistent with long mitochondriain fibroblasts

Prominent—consistent with long mitochondriain fibroblasts

Mitochondrial activity Borderline low complex IV Normal

Whole mitochondrial genome sequencing No pathologic variants No pathologic variants

Genes implicated in autosomal mitochondrialdisorders

No abnormalities in 17 genes Normal NPC1, NPC2, and POLG

Whole-exome sequencing c.1207C>T p.Arg403Cys in DNML1 c.1207C>T p.Arg403Cys in DNML1

Treatment Moderate response to carbamazepine,clobazam, and coenzyme Q

Moderate response to levetiracetam,carbamazepine, and clobazam




our cases lies in the middle domain of DNM1L,3 impairingoligomerization and recruitment to mitochondria3 consistentwith our findings of elongated mitochondria in fibroblastsfrom both patients (figure e-1A, links.lww.com/NXG/A62).Furthermore, knockdown of the DNM1L ortholog Drp1 inmouse cells and its ligand Mff can each cause mtDNA de-pletion and mitochondrial dysfunction,7 consistent with theborderline low mtDNA content and COX activity in skeletalmuscle we demonstrated in case 1. This is likely due to in-creased mitophagic flux (figure e-1C, i).

Our findings suggest that DNM1L is implicated as a geneticcontributor to MEFS. DNM1L p.Arg403Cys mutationscreening could therefore be useful in patients with similarpresentations and in identifying impaired mitochondrial fis-sion causing synaptic dysfunction1 and defective response toinfection.

Author contributionsE. Ladds wrote the initial case reports and manuscript draft,was part of the editing and submissions process, and preparedthe cases for presentation at the British Paediatric NeurologyAssociation Conference 2018. A. Whitney was the pediatri-cian responsible for identifying SP and summarizing his case.E. Dombi and M. Hofer performed the laboratory tests andanalysis of the data. G. Anand was the pediatrician responsiblefor initially summarizing MN’s clinical presentation with thekey learning points, bringing together the coauthors, andinitiating and supervising case presentation at the NationalBritish Paediatric Neurology Association Conference 2018.He also contributed to the final editing process. V. Harrisoninitiated the laboratory tests and analysis of the data. C.Fratter, J. Carver, and I.A. Barbosa performed the laboratorytests and analysis of the data. M. Simpson performed thewhole-exome sequencing. S. Jayawant was the pediatric neu-rology consultant responsible for MN’s clinical care andidentifying the case for publication. J. Poulton performed

conceptualization of the study and analysis of the data andrevised the manuscript.

Study fundingThe authors acknowledge Kate Sergeant and Charu Desh-pande for their involvement in whole-exome sequencing andparental testing of case 1, which was supported by the LilyFoundation.

DisclosureE. Ladds reports no disclosures. A Whitney has served on thescientific advisory board of Zogenix and has received fundingfor speaker honoraria from Zogenix. E. Dombi, M. Hofer, G.Anand, V. Harrison, C. Fratter, and J. Carver report no dis-closures. I.A. Barbosa has received research funding from theLily Foundation. M. Simpson is employed by Genomics Plc.S. Jayawant reports no disclosures. J. Poulton receives re-search support from the Lily Foundation, the WellcomeTrust, and the UK Medical Research Council. Full disclosureform information provided by the authors is available with thefull text of this article at Neurology.org/NG.


References1. Itoh K, Nakamura K, Iijima M, Sesaki H. Mitochondrial dynamics in neuro-

degeneration. Trends Cell Biol 2013;23:64–71.2. Smirnova E, Griparic L, Shurland DL, van der Bliek AM. Dynamin-related protein

Drp1 is required for mitochondrial division in mammalian cells. Mol Biol Cell 2001;12:2245–2256.

3. Fahrner JA, Liu R, Perry MS, Klein J, Chan DC. A novel de novo dominant negativemutation in DNM1L impairs mitochondrial fission and presents as childhood epi-leptic encephalopathy. Am J Med Genet A 2016;170:2002–2011.

4. Shahni R, Cale CM, Anderson G, et al. Signal transducer and activator of transcription2 deficiency is a novel disorder of mitochondrial fission. Brain 2015;138:2834–2846.

5. Cahill TJ, Leo V, Kelly M, et al. Resistance of dynamin-related protein 1 oligomers todisassembly impairs mitophagy, resulting in myocardial inflammation and heart fail-ure. J Biol Chem 2015;290:25907–25919.

6. Gaspard N, Hirsch LJ, Sculier C, et al. New-onset refractory status epilepticus(NORSE) and febrile infection-related epilepsy syndrome (FIRES): state of the artand perspectives. Epilepsia 2018;59:745–752.

7. Parone PA, Da Cruz S, Tondera D, et al. Preventing mitochondrial fission impairs mito-chondrial function and leads to loss of mitochondrial DNA. PLoS One 2008;3:e3257.






GLRA1 mutation and long-term follow-up of thefirst hyperekplexia familyMartin Paucar, PhD, MD, Josefine Waldthaler, MD, and Per Svenningsson, PhD, MD


Correspondence

Dr. Svenningsson

[email protected]

Hyperekplexia (HPX) is a rare familial disorder characterized by an exaggerated startle reflexand stiffness at birth. In 1958, Boris P. Silfverskiold published a report on a Swedish familyaffected by “emotionally precipitated drop seizures.”1 This first description of HPX becameseminal, but it would take 35 years before mutations in the glycine receptor subunit alpha-1(GLRA1) gene were discovered as the cause of this disease.2 Subsequently, SLC6A5 andGLRBmutations were discovered as causes of HPX.3 Here, we present a 60-year follow-up of theSilfverskiold family found to harbor the R271Q mutation in the GLRA1 gene. Some affectedpatients in this family display unreported features for HPX.

MethodsThis report was made within the frame of a study approved by the local ethics committee(Etikprovningsnamnden 2016/1661-31). The family consisted of 4 affected patients (figure).Phenotype details are provided in the original article and summarized in table e-1 (links.lww.com/NXG/A64). Briefly, 3 patients had early-onset violent and injurious falls triggered by unexpectedstimuli (II-1, II-2, and III-1) causing a skull fracture in 2 (II-1 and II-2). Three patients hadhypnagogic myoclonus and a good response to phenobarbital. At times, symptoms recededspontaneously in patient II-1. Reported onset in I-1 (J.E.) was at age 40 years; he had startle withfalls once or twice per year. Siblings II-1 (A.W.) and II-2 (B.E.) lost consciousness sometimes afterstartle-related falls.1 Patient II-1 was diagnosed with late-onset dementia and died at age 87 years;patient II-2 was found drowned in a bathtub at age 76 years. Patient III-1 was aged 10 years at thetime of the publication; at present, she is a 66-year-old retired school teacher. She was diagnosedwith stiff baby syndrome. Patient III-1 had symptoms until age 13 years. At that point, symptomsreceded spontaneously until age 27 years. When her symptoms reappeared, beneficial treatmentwith clonazepamwas started and continued since then. Patient IV-1, born in 1976, presented withinsidious clumsiness starting in childhood and later startle reactions. She underwent surgery forumbilical hernia at age 1 year. Stiffness became gradually persistent between startle reactionsduring adolescence; treatment with clonazepam was started at age 22 years. She has also been oncontinuous antidepressant treatment; Gabapentin was added later because of diffuse pain. Sheworks part time as a preschool teacher. At age 32 years, she developed anxiety, weight loss, andgait difficulties. On examination, an exaggerated head-retraction reflex, hesitant gait, tremulousjerks in all extremities, and tremor of variable frequency were evident; the latter indicatesa functional overlay. EMG displayed 80 ms polymorph bursts with a constant frequency of 7Hz; there was no evidence of neuropathy; MRI of her brain was normal. The heterozygousmutation c.896G>A (R271Q) in GLRA1 was found in patients III-1 and IV-1.

DiscussionThe R271Q mutation in GLRA1 found in the Silfverskiold family is consistent with a phenotypepreviously described in association with GLAR1 mutations. HPX phenotype was variable in this

From the Section of Neurology, Department of Clinical Neuroscience, Karolinska Institutet, Stockholm, Sweden.











family, but there are some novel features. For instance, thewaxing and waning course (patient II-1) and adult onset (I-1)have not been described in other GLRA1 mutations. In addi-tion, another patient (III-1) had a spontaneous remission for 14years. Anxiety is unreported in HPX; in the index case, thisfeature could be related to fears of fall. Also unusual is theabsence of perinatal stiffness, considered as one of the di-agnostic criteria for HPX, and the functional overlay in theindex case. Incongruence and variable frequency of tremor inthe index case are compatible with a psychogenic movementdisorder. Reports on the occurrence of psychogenic movementdisorders in the context of a positive family history of hyper-kinesias are scarce.4 HPX is responsive to clonazepam, and theVigevano maneuver is effective during the neonatal period.3

HPX associated with GLRA1 mutations is in most cases anautosomal recessive disease.2,3 Hypnagogic myoclonus andumbilical hernia have been described in association withGLRA1 mutations. We did not find any of the other featuresassociated with HPX such as apnea, seizures, developmental

delay, learning disabilities, or sudden death.3 These features aremore likely to occur in patients with biallelicGLRA1mutationsor in HPX associated with mutations in SLC6A5 and GLRB.3

Less often are heterozygous GLAR1mutations associated withmild developmental delay.3 R271Q is the most commondominant GLRA1 mutation identified in families of differentethnicities; a founder effect among Caucasian patients has beenproposed.3,5 R271Q is located in the second membrane-spanning domain of the receptor, which is expressed on the cellsurface but displays reduced current and channel opening.6

HPX occurs also as part of severe neurodevelopmental syn-dromes associated with mutations in ARHGEF9 and GPHN,the latter is a lethal condition, which illustrates the importanceof etiologic diagnosis. A variety of animal models and in vitrostudies have provided valuable knowledge on HPX associatedwith GLRA1 mutations,3,6,7 but questions about pathophysi-ology and variable expressivity still await answers.

Author contributionsM. Paucar, J. Waldthaler, and P. Svenningsson: study concept,data collection, and writing of the manuscript. P. Svennings-son: editing of the manuscript.

AcknowledgmentThe authors are grateful to the patients for consenting to thisreport and to Dr. Ruth H. Walker for review of the draft.

Study fundingThis study was supported by the Stockholm County Council.

DisclosureM. Paucar and J. Waldthaler have received research supportfrom the Stockholm County Council. P. Svenningsson servesor has served on the scientific advisory board of CBD sol-utions AB; serves or has served on the editorial boards of PLoSOne and Neuropharmacology; and has received research sup-port from the Swedish Research Council, ALF Stockholm,andWallenberg. Full disclosure form information provided bythe authors is available with the full text of this article atNeurology.org/NG.

Received March 23, 2018. Accepted in final form June 6, 2018.

References1. Kirstein L, Silfverskiold BP. A family with emotionally precipitated drop seizures. Acta

Psychiatr Neurol Scand 1958;33:471–476.2. Shiang R, Ryan SG, Zhu YZ, et al. Mutations in the alpha 1 subunit of the inhibitory

glycine receptor cause the dominant neurologic disorder, hyperekplexia. Nat Genet1993;5:351–358.

3. Thomas RH, Chung SK, Wood SE, et al. Genotype-phenotype correlations inhyperekplexia: apnoeas, learning difficulties and speech delay. Brain 2013;136:3085–3095.

4. Bentivoglio AR, Loi M, Valente EM, Ialongo T, Tonali P, Albanese A. Phenotypicvariability of DYT1-PTD: does the clinical spectrum include psychogenic dystonia?Mov Disord 2002;17:1058–1063.

5. Thomas RH, Drew CJG, Wood SE, et al. Ethnicity can predict GLRA1 genotypes inhyperekplexia. J Neurol Neurosurg Psychiatry 2015;86:341–343.

6. Chung SK, Vanbellinghen JF, Mullins JG, et al. Pathophysiological mechanisms ofdominant and recessive GLRA1 mutations in hyperekplexia. J Neurosci 2010;30:9612–9620.

7. Harvey RJ, Topf M, Harvey K, Rees MI. The genetics of hyperekplexia: more thanstartle! Trends Genet 2008;24:439–447.

FigureUpdated pedigree of the Silfverskiold hyperekplexiafamily described originally in 1958

Phenotype details on patients III-1 and IV-1 are provided in the text, bothpatients harbor the recurrent R271Q mutation in GLRA1.





Case of late-onset Sandhoff disease due to a novelmutation in the HEXB geneAngela R. Sung, MD, Paolo Moretti, MD, and Aziz Shaibani, MD


Correspondence

Dr. Shaibani

[email protected]

Sandhoff disease is one of a group of autosomal recessive conditions known as the GM2gangliosidoses. Normal breakdown of GM2 gangliosides is performed by the enzyme β-hex-osaminidase A. This enzyme consists of 2 subunits (α and β), which are encoded by theHEXAandHEXB genes, respectively. Mutations in either of these genes result in buildup of the GM2gangliosides, with HEXA mutations producing a phenotype of Tay-Sachs disease and HEXBmutations causing Sandhoff disease.

The classic form of Sandhoff disease presents in infancy with symptom onset between ages 2and 9 months. Symptoms include progressive weakness, intellectual disability, vision andhearing impairment, exaggerated startle response, seizures, and death usually before age 3. Late-onset forms of Sandhoff disease have been described but are much rarer. Adult-onset cases canpresent with a wide spectrum of symptoms, including spinocerebellar ataxia, motor neurondisease, sensorimotor neuropathy, tremor, dystonia, and psychosis.1 Specifically, in reviewingcases with the motor neuron disease phenotype, most reports describe predominant lowermotor neuron features, with few cases showing both upper and lower motor neuron findingssimilar to amyotrophic lateral sclerosis.2,3 More recent reports have identified an increasingnumber of novel sequence variants (often compound heterozygous point mutations) that areassociated with the motor neuron disease phenotype,4,5 although the mechanism by whichthese variants produce this specific phenotype is not well understood.

Case presentationA 40-year-old woman was referred for evaluation of slowly progressive weakness of the lowerextremities over the course of 3 years. She reported weakness affecting both legs, difficultystanding from a chair, poor balance, and twitching in her thigh muscles. Additional symptomsincluded anxiety, mood fluctuations, decreased concentration, generalized fatigue, and poorsleep. She denied numbness, pain, arm weakness, difficulty with breathing, speaking, orswallowing.

On physical examination, she had mild weakness in the hip flexors, knee flexors, and kneeextensors, with full strength in all other muscle groups. There was mild atrophy of the kneeextensors with rare fasciculations. Deep tendon reflexes were reduced at the knees, but intactelsewhere. Tandem gait was slightly impaired. The rest of the neurologic examination waswithin normal limits.

Brain MRI revealed mild cerebral and cerebellar atrophy with preferential involvement of thesuperior vermis. Lumbar spine MRI was significant for mild degenerative changes at the L3-L4levels and bilateral neural foraminal stenosis at the L5-S1 levels. EMG and nerve conduction

From the Department of Neurology (A.R.S.), Baylor College of Medicine; Neurology Care Line (P.M.), Michael E. DeBakey Veterans Affairs Medical Center; and Nerve and MuscleCenter of Texas (A.S.), Houston, TX.









studies was normal except for neurogenic units in the kneeextensors and hip flexors. The creatine kinase level was normal.

The pure motor nature of the findings and the EMG resultssuggested a chronic form of lower motor neuron disease, andthe MRI findings did not adequately explain her symptoms.The picture was not typical of any of the common lowermotor neuron diseases. Because there is no inclusive panel forgenetic motor neuron diseases and the affected gene could notbe pinpointed clinically, whole-exome sequencing (WES) wasperformed for further evaluation. WES revealed a compoundheterozygous mutation in the HEXB gene on chromosome 5(c.298delC pathogenic variant and G473S likely pathogenicvariant). Subsequent hexosaminidase enzymatic testingrevealed reduced total hexosaminidase activity in leukocytes(14% that of normal controls), consistent with a GM2 gan-gliosidosis. In addition, there was a high ratio of hexosamin-idase A to B activity (79%), supporting a diagnosis of Sandhoffdisease.

DiscussionThis case describes 2 new sequence variants found in a patientwith clinical symptoms of Sandhoff disease. The populationfrequency of these 2 sequence variants was investigated usingthe Genome Aggregation Database (gnomAD), whichaggregates data on 123,136 exomes and 15,495 genomes fromunrelated individuals.6 The c.298delC variant in HEXB ispredicted to result in loss of protein function and is absentfrom the gnomAD, indicating a very rare, likely pathogenicvariant. The G473S variant in HEXB has been observed onlyonce in the heterozygous state in the gnomAD. The aminoacid position 473 is highly conserved across evolution, with aninvariant glycine residue present from Caenorhabditis elegans

and Drosophila melanogaster to various species of vertebrates(figure).7 The presence of these 2 sequence variants in thecompound heterozygous state, combined with the enzymaticassay showing decreased hexosaminidase activity, implies thatthese sequence variants are pathogenic and are responsible forthe loss of enzymatic activity.

Late-onset Sandhoff disease is rare but should remain a di-agnostic consideration in adults presenting with slowly pro-gressive lower motor neuron disease, and discovery of newpathogenic sequence variants such as the ones discussed inthis case will help further understanding of this disease andfacilitate diagnosis in future patients.

Author contributionsA.R. Sung reviewed the literature, drafted the initial manu-script, and made multiple subsequent revisions. P. Morettiinvestigated the 2 sequence variants discussed in this manu-script in genetic databases and interpreted the data. A. Shai-bani collected clinical data during clinical consultation withthe patient and performed testing that led to the discovery ofthe discussed sequence variants; he also made multiple revi-sions to the manuscript.


DisclosureA.R. Sung reports no disclosures. P. Moretti has received re-search support from the New Jersey Commission on BrainInjury Research and the Department of Veterans Affairs. A.Shaibani receives publishing royalties from Oxford UniversityPress. Full disclosure form information provided by the authorsis available with the full text of this article at Neurology.org/NG.


References1. JohnsonWG. The clinical spectrum of hexosaminidase deficiency diseases. Neurology

1981;31:1453–1456.2. Delnooz CCS, Lefeber DJ, Langemeijer SMC, et al. New cases of adult-onset Sandhoff

disease with a cerebellar or lower motor neuron phenotype. J Neurol NeurosurgPsychiatry 2010;81:968–972.

3. Scarpelli M, Tomelleri G, Bertolasi L, Salviati A. Natural history of motor neurondisease in adult onset GM2-gangiosidosis: a case report with 25 years of follow up.Mol Genet Metab Rep 2014;1:269–272.

4. Yoshizawa T, Kohno Y, Nissato S, Shoji S. Compound heterozygosity with two novelmutations in the HEXB gene produces adult Sandhoff disease presenting as a motorneuron disease phenotype. J Neurol Sci 2002;195:129–138.

5. Banerjee P, Boyers MJ, Berry-Kravis E, Dawson G. Preferential beta-hexosaminidase(Hex) A (alpha beta) formation in the absence of beta-Hex B (beta beta) due toheterozygous point mutations present in beta-Hex beta-chain alleles of a motorneuron disease patient. J Biol Chem 1994;269:4819–4826.

6. Lek M, Karczewski KJ, Minikel EV, et al. Analysis of protein-coding genetic variationin 60,706 humans. Nature 2016;536:285–291.

7. MARRVEL [Database online]. Houston: Baylor College of Medicine; 2017. Availableat: marrvel.org/. Accessed May 9, 2007.

Figure G473S conservation across species

Multiple protein alignment highlighting the evolutionary conservation ofglycine 473 in the human hexosaminidase subunit beta (HEXB) amino acidsequence. Labels: Hs = Homo sapiens; Mm =Mus musculus; Dr = Danio rerio;Dm =Drosophila melanogaster; Ce = Caenorhabditis elegans. The numbers onthe right side of each sequence correspond to the last amino acid depicted.



http://marrvel.org/



Independent NF1mutations underlie cafe-au-laitmacule development in a woman with segmentalNF1Morgan E. Freret, PhD, Corina Anastasaki, PhD, and David H. Gutmann, MD, PhD


Correspondence

Dr. Gutmann

[email protected]

Segmental neurofibromatosis type 1 (NF1) is an under-recognized form of NF1 caused bypostzygotic somatic loss-of-functionNF1 gene mutations that affect a subset of cells in the body.1

This is in contrast to classic or generalizedNF1, in which a germlineNF1 genemutation affects alldiploid cells in the body. In the segmental NF1 variant, individuals typically exhibit clinicalfeatures characteristic of generalized NF1, such as cafe-au-lait macules (CALMs), skinfoldfreckling, and neurofibromas, restricted to one segment of the body. For this reason, establishingthe diagnosis can be challenging because the underlyingNF1 gene mutation is often not detectedin the blood. Underscoring the challenges of caring for individuals with this variant of NF1, wedescribe a woman with segmental NF1 referred to us at 22 years of age for evaluation.

The patient first came tomedical attention at four months of age when her pediatrician noted thatshe had more than six CALMs localized to her right abdomen, back, and hip. Over the next fiveyears, she developed additional hyperpigmented macules in this region, followed by freckling inthe right inguinal region. Collectively, these findings met the criteria for a diagnosis of segmentalNF1. On examination, the CALMs and skinfold freckling in the affected region did not crosseither the anterior or posteriormidlines, and there were no neurofibromas, other hyperpigmentedlesions, or Lisch nodules detected. Because the affected body segment could potentially includeher right ovary, at 28 years of age, she underwent genetic testing (direct Sanger sequencing, mixedligation-dependent probe amplification, and interphase fluorescence in situ hybridization[FISH]), which failed to identify an NF1 gene mutation in her blood. She subsequently un-derwent biopsy of three of herCALMs (right lower back, hip, and abdomen), followed by primaryculture of these skin melanocytes for genetic testing as above.2 As shown in the figure, sequencingand FISH revealed that all three CALM-derived melanocyte cultures shared a common 1.4megabase deletion of the NF1 gene (type 1 total gene deletion [TGD]). Furthermore, in eachof the three CALMs, a different second-hit mutation was identified that affected the remaining(non-deleted) NF1 gene copy.

This case report is instructive for several reasons. First, it underscores the utility of analyzing theaffected tissues (e.g., melanocytes frommultiple CALMs), rather than the blood, in establishinga diagnosis of segmental NF1. Although this patient met the diagnostic criteria for segmentalNF1 on clinical grounds (regional distribution of CALMs and inguinal freckling), the geneticfindings provide a reference for future prenatal genetic counseling and diagnostic testing.3

Second, genetic testing of affected tissue suggests that this patient sustained a total NF1 genedeletion during post-zygotic development and thus affected only a subset of her cells. Indi-viduals with the generalized form of NF1 who harbor this type of mutation (type 1 TGD)frequently have a more severe phenotype, with greater numbers of neurofibromas and anincreased risk of cancer, compared with individuals with generalized NF1 caused by other NF1genemutations.4 By contrast, this patient with segmental NF1 has no visible neurofibromas andis otherwise healthy, which is different from a previous report of several patients with segmental

From the Department of Neurology (M.E.F., C.A., D.H.G.), Washington University School of Medicine, Saint Louis, MO; and Harvard Medical School (M.E.F.), Boston, MA.


The Article Processing Charge was funded by NIH.







NF1 and type 1 TGDs.5 In that series, four of the five caseshad either numerous dermal neurofibromas or a plexiformneurofibroma. The less severe phenotype in the currentsubject suggests that, compared with previously reportedcases, (1) her total NF1 gene deletion may have arisen ata later developmental stage, (2) it is restricted to a differentpopulation of cells, or (3) modifier genes additionally in-fluence clinical expression. Third, each of the CALMs in thispatient contained melanocytes harboring a unique secondNF1 gene mutation that presumably resulted in loss of orimpaired function of the NF1-encoded protein (neuro-fibromin) in those cells. This finding is consistent withprevious reports demonstrating biallelic NF1 gene in-activation in CALMs,6,7 similar to that observed in benigntumors from patients with NF1.2 The finding that thesecond-hit mutations were distinct in each CALM arguesthat each macule arose independently, similar to benign

neoplasms (i.e., neurofibromas) that arise in this populationof tumor-prone individuals.

Author contributionsM.E. Freret contributed to manuscript preparation and madethe figure. C. Anastasaki contributed to manuscript prepara-tion. D.H. Gutmann wrote the first draft of the manuscript.

Study fundingThis work was supported by the NIH (1-R35-NS07211-01,D.H. Gutmann).

DisclosureM.E. Freret and C. Anastasaki report no disclosures. D.H.Gutmann holds patents for the identification of the ND1 geneand mTOR regulator; has received research support from theUS Army Department of Defense, the Giorgio Foundation,

Figure Unique second neurofibromatosis type 1 (NF1) gene mutations in multiple cafe-au-lait macules (CALMs) from anindividual with segmental NF1

(A) Segmental ormosaic NF1 is caused by de novoNF1 genemutations that occur during postzygotic development and affect only a subset of cells in the body.Genetic testing of melanocytes derived from three different CALMs in the patient’s affected body segment (denoted by the diagonal stripes) revealeda common 1.4megabase (MB) deletion, resulting in total loss of one copy of theNF1 gene, and three unique second-hitmutations affecting the remainingNF1allele. The insets depict both NF1 gene copies in the indicated cell types. (B) Locations (colored pinheads) and predicted messenger RNA (mRNA) and proteinchanges resulting from the secondNF1 genemutations inmelanocytes derived from three different CALMs. Two intronicmutationswere locatedwithin splicesites, and one exonic mutation resulted in a premature stop codon (nonsense mutation). Short black boxes denote introns; tall gray boxes denote exons.



the Children’s Tumor Foundation, and the Neurofibroma-tosis Acceleration Therapeutics Program; receives license feepayments for the TSC1 knockout mouse; and receives royaltypayments for the NF1 gene patent. Full disclosure form in-formation provided by the authors is available with the fulltext of this article at Neurology.org/NG.

Received March 20, 2018. Accepted in final form April 24, 2018.

References1. Listernick R, Mancini AJ, Charrow J. Segmental neurofibromatosis in childhood. Am J

Med Genet A 2003;121A:132–135.

2. Maertens O, De Schepper S, Vandesompele J, et al. Molecular dissection of isolateddisease features in mosaic neurofibromatosis type 1. Am J Hum Genet 2007;81:243–251.

3. Ko Y, Lee C, Lee H, Lee M, Lee JS. Clinical application of next-generation sequencingfor the diagnosis of segmental neurofibromatosis. J Dermatol Sci 2017;88:370–372.

4. Kehrer-Sawatzki H, Mautner VF, Cooper DN. Emerging genotype-phenotype rela-tionships in patients with large NF1 deletions. Hum Genet 2017;136:349–376.

5. Messiaen L, Vogt J, Bengesser K, et al. Mosaic type-1 NF1microdeletions as a cause ofboth generalized and segmental neurofibromatosis type-1 (NF1). Hum Mutat 2011;32:213–219.

6. Eisenbarth I, Assum G, Kaufmann D, Krone W. Evidence for the presence of thesecond allele of the neurofibromatosis type 1 gene in melanocytes derived from cafeau laitmacules of NF1 patients. Biochem Biophys Res Commun 1997;237:138–141.

7. De Schepper S, Maertens O, Callens T, Naeyaert JM, Lambert J, Messiaen L. Somaticmutation analysis in NF1 cafe au lait spots reveals two NF1 hits in the melanocytes.J Invest Dermatol 2008;128:1050–1053.





Novel ELOVL4 mutation associated witherythrokeratodermia and spinocerebellar ataxia(SCA 34)Pierre R. Bourque, MD, Jodi Warman-Chardon, MD, Daniel A. Lelli, MD, Lauren LaBerge, MD,

Carly Kirshen, MD, Scott H. Bradshaw, MD, Taila Hartley, MSc, and Kym M. Boycott, PhD, MD


Correspondence

Dr. Bourque

[email protected]

Erythrokeratodermia (EK) is a rare skin disorder, likely genetic and usually present frominfancy.1 There is patchy symmetrical involvement over the body surface, manifested inprogressive figurate plaques of hyperkeratosis and more transient areas of erythema.There is significant overlap in the clinical and genetic features of the “variabilis” and“progressiva” forms of EK. Restricted cutaneous syndromes of EK have been describedassociated with mutations in the connexin (GJB3, GJB4, and GJA1) and loricrin (LOR)genes. The majority of patients with EK, however, have no pathogenic mutations in theGJB genes or LOR.

Three different mutations in the ELOVL4 gene have so far been described in patients pre-senting with a combination of EK and spinocerebellar ataxia (SCA34).2–4 We describe herea novel variant in ELOVL4 associated with this syndrome. This case draws attention to theimportance of assessing subtle chronic neurologic dysfunction in patients presenting with EKand, conversely, being aware of the occurrence of characteristic cutaneous lesions in this newlydescribed syndrome of spinocerebellar ataxia.

Case description and genetic resultsA 60-year-old womanwas referred for assessment of visual blurring and diplopia. Her cutaneousdisorder manifested around age 4 years, with widespread variable itchy areas of erythema andprogressive localized skin thickening. She had no success with topical Tazorac and com-pounded creams. Oral acitretin (25 mg) was helpful but discontinued because of alopecia.Although her visual symptoms had only gradually appeared at age 50 years, she reportednoticing ataxia of gait as far back as teenage years. Her parents who were deceased at age 73years and age 84 years had reported no symptoms of rash or ataxia. Her 5 siblings, similarly,were asymptomatic from the dermatologic and neurologic standpoint. Two children werebiologically unrelated (adopted).

On examination, there were widespread demarcated brown erythematous keratotic plaquestypical of EK over her wrists, hands, inner thighs, knees, and ankles (figure, A and B). Neuro-ophthalmic deficits included square wave jerks, saccadic pursuit, and alternating skew deviationwith superimposed periodic alternating skew deviation in primary position (period 2.5minutes). Myotatic reflexes were diffusely reduced (1+) and absent at the ankles. There wasmoderate rombergism and marked tandem gait ataxia, but no appendicular dysmetria. A skinbiopsy showed irregular acanthosis, papillomatosis, and hyperkeratosis, in keeping with EK.

From the Department of Medicine (Neurology) (P.R.B., J.W-C., D.A.L.), University of Ottawa; Ottawa Hospital Research Institute (P.R.B., J.W-C.); Department of Medicine (Dermatology)(L.L., C.K.), University of Ottawa; Department of Anatomical Pathology (S.H.B.), University of Ottawa; and Department of Genetics (J.W-C., T.H., K.M.B.), Children’s Hospital of EasternOntario, Ottawa, Canada.









MRI of the brain showed only subtle flattening of the ventralpons and mild global cerebellar atrophy. Nerve conductionstudies and needle EMG studies were normal.

The clinical presentation suggested the possibility of SCA34.ELOVL4 gene sequencing (Prevention Genetics, Marshfield,WI) identified a variant (c.698C>T; p.Thr233Met), which hasnot been reported in the literature or public databases. Theamino acid substitution programs CADD (27.9), PolyPhen-2

(1.0), and MutationTaster (1.0) predict the variant to be dam-aging, and the Sorting Intolerant From Tolerant (SIFT) pro-gram predicts it to be tolerated (table). It was not possible to testparents, but 2 clinically unaffected siblings were tested and didnot carry this variant. The patient was started on baclofen,starting at 5 mg tid and titrating up to 15 mg tid over 1 month.She reported significant improvement in the perception ofoscillopsia and the duration and frequency of bouts of diplopia.She also elected to try application of beeswax cream (reported tocontain 30-32 carbon very long chain fatty acids) on her legs for2 months, which was of no benefit.

DiscussionThis case broadens the genetic mutation profile of EK byreporting a patient with a fourth variant in the ELOVL4 geneassociated with SCA34 (table). The ELOVL4 gene catalyzes therate-limiting reaction in elongation of fatty acids with a chainlength greater than C26. It is ubiquitously expressed, but par-ticularly enriched in the retina. It is critical in the development ofthe outer layer of the epidermis, the stratum corneum.5 Rarediseases associated with dysregulation of ELOVL4 in addition toSCA34 include 2 autosomal recessive conditions: hereditarymacular degeneration (Stargardt disease) and a syndrome ofichthyosis, spastic quadriplegia, and mental retardation.6

The clinical presentation of our patient is quite similar to whatwas reported in a large French Canadian family with 32 af-fected members.2 In this large SCA34 family, cutaneous in-volvement began in infancy, with a preponderance ofhyperkeratosis over erythema. Gait ataxia was noted to start

Figure Erythrokeratodermia (lower limbs)

Photographs of symmetrical brownish-red hyperkeratosis, most prominentover the knees, lower anterior legs and dorsal foot region. There were similarsymmetrical lesions over the inner thigh and more subtle involvement of thedorsal forearm and wrist region.

Table Clinical and MRI features of patients with ELOVL4 mutations associated with clinical features of SCA34

Present case Ozaki et al.4 Bourassa et al.3 Cadieux-Dion et al.2

Mutation ELOVL4; c.698C>T,p.Thr233Met; heterozygous

ELOVL4; c.736T>G, p.W246G;heterozygous

ELOVL4; c.539A>C,p.Gln180Pro; heterozygous

ELOVL4; c.504 G>C, p.L168F;heterozygous

Ethnicity English Canadian Japanese South American French Canadian

Number affected 1 clinically affected mutationcarrier

9 clinically affected from 2separated families

1 clinically affected mutationcarrier

19 mutation carriers, 4clinically unaffected

Age or mean age,onset of gait ataxia, y

15 34 Mid 20s 51

Type of ataxia Gait (no dysarthria or limbdysmetria)

Gait, speech, and limb Speech, gait > limbs Gait (12/19), limb (9/19), andspeech (6/19)

Progression of ataxia Independent ambulationwithout aid at age 60 y

Slow, ambulatory aid afterage 60 y

Only 10-y follow-up Slow: use of cane after 10 y,use of walker by age 70 y

Oculomotor signs Saccadic pursuit, squarewave jerks, and periodicalternating skew deviation

Horizontal nystagmus (78%),supranuclear palsy (33%),and impaired smooth pursuit(56%)

Horizontal nystagmus andmild bilateralophthalmoparesis

Horizontal nystagmus in 7/19and slow ocular pursuit in 5/19

Myotatic reflexes Hyporeflexia Hyperreflexia and Babinskisigns in 89%

Normal Hyporeflexia in 7/19

Erythrokeratodermia Infantile onset andpersistent, though showingcyclical exacerbations.

None EK of forearms and legs EK of infantile onset butmostly resolving after age 25y, noted in 14/19

Continued



usually around age 50 years. Our patient may be unique inreporting prominent oculomotor symptoms. A periodic al-ternating skew deviation was significantly improved whentreated with baclofen.

The possibility of SCA34 should be considered in patientswith EK, and clinicians should be aware that cerebellarmanifestations are relatively late and variable, whereas cuta-neous involvement has an onset early in childhood. SCA34should thus be considered in the differential diagnosis of ge-netic neurocutaneous disorders. Oculomotor manifestationsshould be specifically assessed because they may be improvedwith treatment.

Author contributionsP.R. Bourque: analysis and interpretation and production andrevision of the manuscript. J. Warman Chardon: analysis andinterpretation and critical revision of the manuscript. D.A.Lelli: acquisition and interpretation of neuro-ophthalmologydata and critical revision of the manuscript. L. LaBerge: ac-quisition and revision of dermatology data and critical revisionof the manuscript. C. Kirshen: acquisition and revision ofclinical dermatology data and critical revision of the manu-script. S.H. Bradshaw: acquisition and interpretation ofdermatopathology data and critical revision of the manuscript.T. Hartley: preparation of the table and assessment of ge-netic classification. K.M. Boycott: study supervision, critical

revision of the manuscript, and overview of genetic data ac-quisition and interpretation.


DisclosureAll authors have no financial disclosures relevant to this re-search and publication. Full disclosure form informationprovided by the authors is available with the full text of thisarticle at Neurology.org/NG.


References1. Ishida-Yamamoto A. Erythrokeratodermia variabilis et progressiva. J Dermatol 2016;

43:280–285.2. Cadieux-Dion M, Turcotte-Gauthier M, Noreau A, et al. Expanding the clinical

phenotype associated with ELOVL4 mutation: study of a large French-Canadianfamily with autosomal dominant spinocerebellar ataxia and erythrokeratodermia.JAMA Neurol 2014;71:470–475.

3. Bourassa C, Raskin S, Serafini S, Teive HA, Dion PA, Rouleau GA. A new ELOVL4mutation in a case of spinocerebellar ataxia with erythrokeratodermia. JAMA Neurol2015;72:942–943.

4. Ozaki K, Doi H, Misui J, et al. A novel mutation in ELOVL4 leading to spinocerebellarataxia (SCA) with the hot cross bun sign but lacking erythrokeratodermia: a broad-ened spectrum of SCA34. JAMA Neurol 2015;72:797–805.

5. Cameron DJ, Tong Z, Yang Z, et al. Essential role of ELOVL4 in very long chain fattyacid synthesis, skin permeability barrier function, and neonatal survival. Int J Biol Sci2007;3:111–119.

6. Aldahmesh MA, Mohamed JY, Alkuraya HS, et al. Recessive mutations in ELOVL4cause ichthyosis, intellectual disability, and spastic quadriplegia. Am J Hum Genet2011;89:745–750.

Table Clinical and MRI features of patients with ELOVL4 mutations associated with clinical features of SCA34 (continued)

Present case Ozaki et al.4 Bourassa et al.3 Cadieux-Dion et al.2

Brain MRI Mild cerebellar and pontinebasal atrophy

Pontine and cerebellaratrophy in 100%, hot crossbun sing in 67%

Moderate cerebellar andpontine atrophy

Cerebellar and pontineatrophy in 6/9 carriers tested

Support forpathogenicity

Two asymptomatic siblingsdid not carry the mutation.Variant not previouslyreported in public databases(gnomAD, EVS, and1000Genomes). CADD (27.9),PolyPhen-2 (1.0), andMutationTaster (1.0) predictdamaging. SIFT predictstolerated.

Segregation with affectedmembers of 2 differentfamilies. A third family hasbeen reported in ClinVar(RCV000522609.1). Variantcontinues to be absent frompublic databases (gnomAD,EVS, and 1000Genomes).CADD (23.5), PolyPhen-2(0.99), andMutationTaster (1)predict damaging. SIFTpredicts tolerated.

Variant continues to beabsent from publicdatabases (gnomAD, EVS,and 1000Genomes). CADD(26.5), PolyPhen-2 (0.99), andMutationTaster (1) predictdamaging. SIFT predictstolerated. No segregationdata.

Segregation with diseasehaplotype in 19 individuals.Variant continues to beabsent from publicdatabases (gnomAD, EVS,and 1000Genomes). CADD(24), PolyPhen-2 (0.99), andMutationTaster (0.87) predictdamaging. SIFT predictstolerated.

ACMG variantclassification

VUS (PM2+PP3+PP4) Likely pathogenic(PS1+PM2+PP1+PP3+PP4)

VUS (PM2+PP3+PP4) VUS (PM2+PP1+PP3+PP4)

Abbreviations: ACMG = American College of Medical Genetics and Genomics; EK = erythrokeratodermia; EVS = Exome Variant Server database; SIFT = SortingIntolerant From Tolerant program; VUS = Variant of Unknown Significance.




lww nng nngcover 4 4 - genetics · academy officers ralphl.sacco,md,ms,faan,president...

Documents