An Official Journal of the American Academy of Neurology Neurology.org/ng • Online ISSN: 2376-7839 Volume 3, Number 5, October 2017

Moderate blast exposure Design and rationale for Genome-wide scan alters expression examining neuroimaging in Hispanics highlights and levels of amyloid genetics in ischemic stroke: candidate loci for precursor the MRI-GENIE study brain white matter hyperintensities Table of Contents Neurology.org/ng  Online ISSN: 2376-7839 Volume 3, Number 5, October 2017

EDITORIAL e195 No rare deleterious variants from STK32B, e196 Genomic links between blast exposure, brain injury, PPARGC1A,andCTNNA3 are associated with and Alzheimer disease essential tremor Y.P. Conley and R. Diaz-Arrastia G. Houle, A. Ambalavanan, J.-F. Schmouth, Companion article, e186 C.S. Leblond, D. Spiegelman, S.B. Laurent, C.V. Bourassa, C. Grayson, M. Panisset, S. Chouinard, ARTICLES N. Dupré, C. Vilariño-Güell, A. Rajput, S.L. Girard, e186 Moderate blast exposure alters and P.A. Dion, and G.A. Rouleau levels of amyloid precursor protein J. Gill, A. Cashion, N. Osier, L. Arcurio, V. Motamedi, K.C. Dell, W. Carr, H.-S. Kim, S. Yun, P. Walker, e183 Ataxia-pancytopenia syndrome with SAMD9L S. Ahlers, M. LoPresti, and A. Yarnell Editorial, e196 S. Gorcenco, J. Komulainen-Ebrahim, K. Nordborg, M. Suo-Palosaari, S. Andréasson, J. Krüger, e177 Whole-exome sequencing associates novel C. Nilsson, U. Kjellström, E. Rahikkala, CSMD1 gene mutations with familial Parkinson D. Turkiewicz, M. Karlberg, L. Nilsson, disease J. Cammenga, U. Tedgård, J. Davidsson, J. Ruiz-Martínez, L.J. Azcona, A. Bergareche, J. Uusimaa, and A. Puschmann J.F. Martí-Massó, and C. Paisán-Ruiz

e184 IBA57 mutations abrogate iron-sulfur cluster e178 Ultra-rare mutations in SRCAP segregate in assembly leading to cavitating Caribbean Hispanic families with Alzheimer leukoencephalopathy disease A. Ishiyama, C. Sakai, Y. Matsushima, S. Noguchi, B.N. Vardarajan, G. Tosto, R. Lefort, L. Yu, S. Mitsuhashi, Y. Endo, Y.K. Hayashi, Y. Saito, D.A. Bennett, P.L. De Jager, S. Barral, E. Nakagawa, H. Komaki, K. Sugai, M. Sasaki, N. Sato, D. Reyes-Dumeyer, P.L. Nagy, J.H. Lee, R. Cheng, I. Nonaka, Y.-i. Goto, and I. Nishino M. Medrano, R. Lantigua, E. Rogaeva, P. St George-Hyslop, and R. Mayeux e185 Genome-wide scan in Hispanics highlights candidate loci for brain white matter e180 Design and rationale for examining neuroimaging hyperintensities genetics in ischemic stroke: The MRI-GENIE study A. Beecham, C. Dong, C.B. Wright, N. Dueker, A.-K. Giese, M.D. Schirmer, K.L. Donahue, L. Cloonan, A.M. Brickman, L. Wang, C. DeCarli, R. Irie, S. Winzeck, M.J.R.J. Bouts, E.C. McIntosh, S.H. Blanton, T. Rundek, R. Mayeux, and S.J. Mocking, A.V. Dalca, R. Sridharan, H. Xu, P. Frid, R.L. Sacco E. Giralt-Steinhauer, L. Holmegaard, J. Roquer, J. Wasselius, J.W. Cole, P.F. McArdle, J.P. Broderick, J. Jimenez-Conde, C. Jern, B.M. Kissela, e187 De novo CTBP1 variant is associated with D.O. Kleindorfer, R. Lemmens, A. Lindgren, decreased mitochondrial respiratory chain J.F. Meschia, T. Rundek, R.L. Sacco, R. Schmidt, activities P. Sharma, A. Slowik, V. Thijs, D. Woo, B.B. Worrall, E.W. Sommerville, C.L. Alston, A. Pyle, L. He, S.J. Kittner, B.D. Mitchell, J. Rosand, P. Golland, G. Falkous, K. Naismith, P.F. Chinnery, O. Wu, and N.S. Rost, On behalf of MRI-GENIE R. McFarland, and R.W. Taylor Table of Contents continued e192 Children with 59-end NF1 gene mutations are more e189 Hyperventilation-athetosis in ASXL3 deficiency likely to have glioma (Bainbridge-Ropers) syndrome C. Anastasaki, S.M. Morris, F. Gao, and D.H. Gutmann R. Dad, S. Walker, S.W. Scherer, M.J. Hassan, S.Y. Kang, and B.A. Minassian

CLINICAL/SCIENTIFIC NOTES e179 Novel AMPD2 in pontocerebellar hypoplasia, e190 ARHGEF28 p.Lys280Metfs40Ter in an amyotrophic dysmorphisms, and teeth abnormalities lateral sclerosis family with a C9orf72 expansion A. Accogli, M. Iacomino, F. Pinto, A. Orsini, M.S. Vari, S.M.K. Farhan, T.F. Gendron, L. Petrucelli, R. Selmi, A. Torella, V. Nigro, C. Minetti, M. Severino, R.A. Hegele, and M.J. Strong P. Striano, V. Capra, and F. Zara

e191 Cervical artery dissection and iliac artery aneurysm e181 Novel POLG variants associated with late-onset in an SMAD-4 mutation carrier de novo status epilepticus and progressive ataxia E. Wiener, P. Martin, S. Mehta, and H.S. Markus Y.S. Ng, H. Powell, N. Hoggard, D.M. Turnbull, R.W. Taylor, and M. Hadjivassiliou e193 Autosomal recessive inheritance of ADCY5-related generalized dystonia and myoclonus e182 Novel intronic mutation in MTM1 detected by RNA M.J. Barrett, E.S. Williams, C. Chambers, and analysis in a case of X-linked myotubular R. Dhamija myopathy A. Al-Hashim, H.D. Gonorazky, K. Amburgey, ’ S. Das, and J.J. Dowling e194 The Alzheimer s Disease Sequencing Project: Study design and sample selection G.W. Beecham, J.C. Bis, E.R. Martin, S.-H. Choi, e188 Nonsyndromic Parkinson disease in a family with A.L. DeStefano, C.M. van Duijn, M. Fornage, Autosomal Dominant Optic Atrophy due to OPA1 S.B. Gabriel, D.C. Koboldt, D.E. Larson, A.C. Naj, mutations B.M. Psaty, W. Salerno, W.S. Bush, T.M. Foroud, D.S. Lynch, S.H.Y. Loh, J. Harley, A.J. Noyce, E. Wijsman, L.A. Farrer, A. Goate, J.L. Haines, L.M. Martins, N.W. Wood, H. Houlden, and M.A. Pericak-Vance, E. Boerwinkle, R. Mayeux, H. Plun-Favreau S. Seshadri, and G. Schellenberg

Podcast Video LOE classification LOE recommendation

tinyurl.com/NeurologyNG twitter.com/GreenJournal

Cover image: Confocal microscopy image of the mitochondrial network in fibroblasts cultured from an unaffected family member. In affected patients, the mitochondrial network was significantly fragmented. See “Nonsyndromic Parkinson disease in a family with autosomal dominant optic atrophy due to OPA1 mutations.” Neurology.org/ng  Online ISSN: 2376-7839 Volume 3, Number 5, October 2017

fi VISION: Neurology® Genetics will be the premier peer-reviewed Executive Of ce, American Academy of Neurology fi Catherine M. Rydell, CAE, Executive Director/CEO journal in the eld of neurogenetics. 201 Chicago Ave Minneapolis, MN 55415 MISSION: Neurology: Genetics will provide neurologists with Tel: 612-928-6100 outstanding original contributions that elucidate the role of genetic and epigenetic variations in diseases and biological traits Editorial Office of the central and peripheral nervous systems. Patricia K. Baskin, MS, Executive Editor Kathleen M. Pieper, Senior Managing Editor, Neurology Editor Lee Ann Kleffman, Managing Editor, Neurology: Genetics Neurology: Genetics Sharon L. Quimby, Managing Editor, Neurology® Clinical Practice Stefan M. Pulst, MD, Dr med, FAAN Morgan S. Sorenson, Managing Editor, Neurology® Neuroimmunology & fl Professor and Chair, Department of Neurology Neuroin ammation University of Utah Cynthia S. Abair, MA, Senior Graphics Editor Salt Lake City, UT Andrea R. Rahkola, Production Editor, Neurology [email protected] Robert J. Witherow, Senior Editorial Associate Specialties: Genetics, movement disorders Karen Skaja, Senior Editorial Associate Kaitlyn Aman Ramm, Editorial Assistant Editor-in-Chief Kristen Swendsrud, Editorial Assistant Neurology® Andrea Willgohs, Editorial Assistant Robert A. Gross, MD, PhD, FAAN Professor of Neurology and of Pharmacology and Physiology Strong Epilepsy Center Publisher University of Rochester Medical Center Wolters Kluwer Rochester, NY Baltimore, MD [email protected]

Editorial Inquiries Publishing Staff Tel: 612-928-6400 Kim Jansen, Executive Publisher Toll-free: 800-957-3182 (US) Alexandra Lazerow, Production Team Leader, AAN Journals Fax: 612-454-2748 Steve Rose, Editorial Assistant [email protected] Stacy Drossner, Production Associate

Academy Officers Ralph L. Sacco, MD, MS, FAAN, President James C. Stevens, MD, FAAN, President Elect Ann H. Tilton, MD, FAAN, Vice President Carlayne E. Jackson, MD, FAAN, Secretary Janis M. Miyasaki, MD, MEd, FRCPC, FAAN, Treasurer Terrence L. Cascino, MD, FAAN, Past President Neurology.org/ng  Online ISSN: 2376-7839 Volume 3, Number 5, October 2017

Editor Biostatistics Neurology® Genetics Richard J. Kryscio, PhD Stefan M. Pulst, MD, Dr med, FAAN University of Kentucky Professor and Chair, Department of Neurology Lexington, KY University of Utah Christopher A. Beck, PhD Salt Lake City, UT University of Rochester [email protected] Rochester, NY Specialties: Genetics, movement disorders Sue Leurgans, PhD Editor-in-Chief Rush University Medical Center Neurology® Chicago, IL Robert A. Gross, MD, PhD, FAAN Level of Evidence Evaluations Professor of Neurology and of Pharmacology and Physiology Gary S. Gronseth, MD, FAAN Strong Epilepsy Center University of Kansas University of Rochester Medical Center Kansas City, KS Rochester, NY [email protected] Podcasts Ted M. Burns, MD, Deputy Podcast Editor Deputy Editors University of Virginia Neurology Charlottesville, VA Bradford B. Worrall, MD, MSc, FAAN Andrew M. Southerland, MD, MSc Professor of Neurology and Public Health Sciences University of Virginia University of Virginia Charlottesville, VA Charlottesville, VA Ombudsman Neurology: Genetics David S. Knopman, MD, FAAN Nicholas Elwood Johnson, MD Mayo Clinic University of Utah Rochester, MN Salt Lake City, UT Specialties: Muscle diseases, patient-centered outcomes Scientific Integrity Advisor Robert B. Daroff, MD, FAAN Associate Editors Case Western Reserve University Neurology: Genetics Cleveland, OH Alexandra Durr, MD, PhD Editorial Board Hôpital de la Salpêtrière Hilary Coon, PhD, University of Utah, Salt Lake City, UT Paris, France Giovanni Coppola, MD, UCLA, Los Angeles, CA Specialties: Spastic paraplegia, Huntington disease, ataxia Chantal Depondt, MD, PhD, Université Libre de Bruxelles, Brussels, Belgium Massimo Pandolfo, MD, FAAN Brent L. Fogel, MD, PhD, FAAN, UCLA, Los Angeles, CA Hôpital Erasme Anthony J. Griswold, PhD, University of Miami, Miami, FL Université Libre de Bruxelles Orhun H. Kantarci, MD, Mayo Clinic, Rochester, MN Brussels, Belgium Julie R. Korenberg, PhD, MD, University of Utah, Salt Lake City, UT Specialties: Epilepsy, Friedreich ataxia, stem cells Margherita Milone, MD, PhD, Mayo Clinic, Rochester, MN Davide Pareyson, MD, C. Besta Neurological Institute IRCCS Foundation, Raymond P. Roos, MD, FAAN Milan, Italy University of Chicago Medical Center Shoji Tsuji, MD, PhD, University of Tokyo Graduate School of Medicine, Chicago, IL Tokyo, Japan Specialties: Motor neuron disease, prion disease, neurovirology/neuroimmunology Dineke S. Verbeek, PhD, University Medical Center Groningen, Groningen, The Netherlands Jeffery M. Vance, MD, PhD David Viskochil, MD, PhD, University of Utah, Salt Lake City, UT University of Miami Juliane Winkelmann, MD, Technische Universität München, Helmholtz Miami, FL Zentrum Münche, Munich, Germany Specialties: Parkinson, Alzheimer disease, inherited neuropathies Juan I. Young, PhD, University of Miami, Miami, FL Level of Evidence Review Team Melissa J. Armstrong, MD, Gainesville, FL Richard L. Barbano, MD, PhD, FAAN, Rochester, NY Richard M. Dubinsky, MD, MPH, FAAN, Kansas City, KS Jeffrey J. Fletcher, MD, MSc, Ann Arbor, MI Gary M. Franklin, MD, MPH, FAAN, Seattle, WA David S. Gloss II, MD, MPH&TM, Charleston, WV John J. Halperin, MD, FAAN, Summit, NJ Jason Lazarou, MSc, MD, Toronto, Ontario, Canada Steven R. Messé, MD, FAAN, Philadelphia, PA Pushpa Narayanaswami, MBBS, DM, FAAN, Boston, MA Alex Rae-Grant, MD, Cleveland, OH Neurology.org/ng  Online ISSN: 2376-7839 Volume 3, Number 5, October 2017

Publication Information 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 published online for the American Academy of Neurology, 201 Chicago Avenue, Minneapolis, MN 55415, by Wolters Kluwer Health, Inc. at 14700 Citicorp Drive, Bldg. 3, Hagerstown, MD 21742. Business offices are located at Two Commerce Square, 2001 Market Street, Philadelphia, PA 19103. Production offices are located at 351 West Camden Street, Baltimore, MD 21201-2436. © 2017 American Academy of Neurology. Neurology® Genetics is an official journal of the American Academy of Neurology. 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 “©Request Permissions” icon for the relevant article. Alternatively, send an email to [email protected]. General information about permissions can be found here: https://www.lww.com/journal-permission. Disclaimer: Opinions expressed by the authors and advertisers are not necessarily those of the American Academy of Neurology, its affiliates, or of the Publisher. The American Academy of Neurology, its affiliates, and the Publisher disclaim any liability to any party for the accuracy, completeness, efficacy, or availability of the material contained in this publication (including drug dosages) or for any damages arising out of the use or 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 Elizabeth S. Hall, tel: 267-804-8123, [email protected]. In Europe: Avia Potashnik, Wolters Kluwer, tel: 144 207 981 0722; 144 7919 397 933 or e-mail: [email protected]. Careers & Events: Monique McLaughlin, Wolters Kluwer, Two Commerce Square, 2001 Market Street, Philadelphia, PA 19103, tel: 215-521- 8468, fax: 215-521-8801; [email protected]. Reprints: Meredith Edelman, Commercial Reprint Sales, Wolters Kluwer, Two Commerce Square, 2001 Market Street, Philadelphia, PA 19103, tel: 215-555-1212 (office), 215-356-2721 (mobile); [email protected]; [email protected]. Special projects: US & Canada: Alan Moore, Wolters Kluwer, Two Commerce Square, 2001 Market Street, Philadelphia, PA 19103, tel: 215-521-8638, [email protected]. International: Andrew Wible, Senior Manager, Rights, Licensing, and Partnerships, Wolters Kluwer; [email protected]. EDITORIAL Genomic links between blast exposure, brain injury, and Alzheimer disease

Yvette P. Conley, PhD, Military personnel are at high risk of exposure to samples were collected on each day of training. FAAN blasts during military service, both during combat Gene expression data representing all were Ramon Diaz-Arrastia, and training. Evidence continues to build that sup- collected using RNA-Seq, and data before and after MD, PhD, FAAN ports exposure to blasts, particularly when repetitive, blast were compared. This approach has the advan- are associated with brain dysfunction which is usually tage that no a priori biological hypotheses are transient but can be prolonged or permanent. tested; instead, all genes are interrogated allowing Correspondence to Dr. Conley: Chronic traumatic encephalopathy (CTE), a conse- for the discovery of novel biological processes [email protected] quence to repetitive brain injury noted in professional affected by blast exposure. The biological network

Neurol Genet athletes, has received a lot of attention lately, and the that was most altered following blast injury was 2017;3:e196; doi: 10.1212/ suspicion that blast exposures in military personnel amyloid precursor protein (APP). The gene expres- NXG.0000000000000196 lead to CTE or a CTE-like syndrome1 has appropri- sion findings were followed up by the evaluation of ately received much attention in the scientific and lay APP protein levels, and the pattern of APP concen- press. Approximately 49% of injured military person- trations differed over time in the moderate blast nel are injured by explosions,2 and up to 15% of exposure group but not in the no or low blast expo- service members returning from a combat theater sure group. Of interest, APP levels significantly report having experienced traumatic brain injury decreased on days 8 and 9 after moderate blast (TBI) during a year-long deployment.3 Being able exposure returning to normal by day 10 and to identify the biological networks that are disrupted changes in APP correlated with blast force expo- by these injuries holds great promise for the develop- sure. While headache and concentration problems ment of therapeutics to improve neurologic recovery. occurredmoreofteninthegroupexposedtomod- In this issue of Neurology® Genetics,Gilletal.4 report erate blast, changes in APP levels were not corre- on the biological impact of moderate blast injury in lated with any symptom. military personnel. A unique feature of this study, which APP is a transmembrane glycoprotein that has allowed a high level of scientific rigor, is that the authors been extensively studied and is probably best known were able to measure the intensity of the blast, collect for its role in the pathophysiology of Alzheimer dis- biological data before blastinjury,andcomparethose ease (AD). The normal function of APP is still data with data after blast injury within an individual. unclear, but evidence so far supports its role in neuro- This was possible because participants were active-duty nal network formation including neuronal migration military service members enrolled in a blast training during brain development; progression in program, which involved repeated exposures to con- neural stem cells; maintenance of synapsis including trolled explosions. A total of 69 male participants (mean calcium homeostasis; and restoration of axonal and age of 30 years and mean duration of service of 10 years) neuronal functions after injury.5 Evidence presented were enrolled, and during training, 29 were exposed to by Gill et al. indicates that APP levels decrease after a moderate blast (mean peak pressure of 7.9 psi), while moderate blast exposure. This is seemingly in contra- the remaining 40 had no or low blast exposure. Blast diction to evidence in the TBI literature, which in- exposures were measured using bilateral sensors attached dicates that APP production in brain tissue increases to the participant’s helmet. Symptoms (i.e., headache, after injury and may play a neuroprotective role,6 dizziness, nausea/vomiting, sensitivity to sound, sleep indicating that more research is needed to fully under- disturbances, fatigue, irritability, depression/sadness, stand the peripheral patterns of APP after brain injury frustration, anxiety, memory, concentration, attention, and blast exposure. Additional directions for future visual disturbances, and balance issues) and blood research also include investigating other genes in the

From the University of Pittsburgh (Y.P.C.), PA; and University of Pennsylvania (R.D.-A.), Penn Presbyterian Medical Center, Philadelphia. Funding information and disclosures are provided at the end of the article. Go to Neurology.org/ng for full disclosure forms. The Article Processing Charge was funded by the authors. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Neurology.org/ng Copyright © 2017 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 APP biological network, particularly epistatic roles for DISCLOSURE genes involved in APP processing. Dr. Conley and Dr. Diaz-Arrastia report no disclosures. Go to The most important clinical implications for the Neurology.org/ng for full disclosure forms.

findings presented in the article by Gill et al. are REFERENCES that biological changes occur as a result of exposure 1. Goldstein LE, Fisher AM, Tagge CA, et al. Chronic to moderate blasts, that these changes involve bio- traumatic encephalopathy in blast-exposed military vet- markers implicated in neurologic disorders, and erans and a blast neurotrauma mouse model. Sci Transl that a discovery-based approach led to putative bio- Med 2012;4:134ra60. logical links between TBI and AD. The social con- 2. Ritenour AE, Blackbourne LH, Kelly JF, et al. Incidence of primary blast injury in US military overseas contingency cern regarding the effects of blast exposure during operations: a retrospective study. Ann Surg 2010;251: military service on long-term neurologic function 1140–1144. has led to large investments in research focused 3. Hoge CW, McGurk D, Thomas JL, et al. Mild traumatic on better understanding of the relationship brain injury in U.S. Soldiers returning from Iraq. N Engl J between TBI and AD and discovering therapeutics Med 2008;358:453–463. to prevent or mitigate TBI-related neurodegenera- 4. Gill J, Cashion A, Osier N, et al. Moderate blast exposure alters gene expression and levels of amyloid precursor pro- tion. With increased evidence for biomarkers in tein. Neurol Genet 2017;3:e186. doi: 10.1212/NXG. common between TBI and AD, the fruits of AD 0000000000000186. research may affect treatment for patients who have 5. Nalivaeva NN, Turner AJ. The amyloid precursor protein: sustained a TBI and vice versa. a biochemical enigma in brain development, function and disease. FEBS Lett 2013;587:2046–2054. STUDY FUNDING 6. Hefter D, Draguhn A. APP as a protective factor in acute No targeted funding reported. neuronal insults. Front Mol Neurosci 2017;10:22.

2 Neurology: Genetics Moderate blast exposure alters gene expression and levels of amyloid precursor protein

Jessica Gill, PhD ABSTRACT Ann Cashion, PhD Objective: To explore gene expression after moderate blast exposure (vs baseline) and proteomic Nicole Osier, PhD changes after moderate- (vs low-) blast exposure. Lindsay Arcurio, PhD Methods: Military personnel (N 5 69) donated blood for quantification of protein level, and peak Vida Motamedi, BA pressure exposures were detected by helmet sensors before and during a blast training program Kristine C. Dell, BA (10 days total). On day 7, some participants (n 5 29) sustained a moderate blast (mean peak Walter Carr, PhD pressure 5 7.9 psi) and were matched to participants with no/low-blast exposure during the Hyung-Suk Kim, PhD training (n 5 40). PAXgene tubes were collected from one training site at baseline and day 10; Sijung Yun, PhD RNA-sequencing day 10 expression was compared with each participant’s own baseline samples Peter Walker, PhD to identify genes and pathways differentially expressed in moderate blast-exposed participants. Stephen Ahlers, PhD Changes in amyloid precursor protein (APP) from baseline to the day of blast and following 2 days Matthew LoPresti, PhD were evaluated. Symptoms were assessed using a self-reported form. Angela Yarnell, PhD Results: We identified 1,803 differentially expressed genes after moderate blast exposure; the most altered network was APP. Significantly reduced levels of peripheral APP were detected the day after

Correspondence to the moderate blast exposure and the following day. Protein concentrations correlated with the mag- Dr. Jessica Gill, PhD: nitude of the moderate blast exposure on days 8 and 9. APP concentrations returned to baseline lev- [email protected] els 3 days following the blast, likely due to increases in the genetic expression of APP. Onset of concentration problems and headaches occurred after moderate blast. Conclusions: Moderate blast exposure results in a signature biological profile that includes acute APP reductions, followed by genetic expression increases and normalization of APP levels; these changes likely influence neuronal recovery. Neurol Genet 2017;3:e186; doi: 10.1212/ NXG.0000000000000186

GLOSSARY ANOVA 5 analysis of variance; APP 5 amyloid precursor protein; AUC 5 area under the curve; cDNA 5 complementary DNA; CV 5 coefficient of variation; qPCR 5 quantitative PCR; RNA-seq 5 RNA sequencing; TBI 5 traumatic brain injury.

Proximity to a blast explosion results in exposure to an overpressure wave and can result in injury to the brain and body. In the military, overpressures occur due to a variety of sources including artillery and improvised explosive devices. At least 30% of military personnel involved in the Operation Iraqi/ Enduring Freedom campaigns had a mild traumatic brain injury (TBI) as a result of blast exposure.1 Thus, there remains an impetus to better understand the molecular mechanisms underlying blast exposure, and the extent to which these biological changes account for behavioral, psychological, and cognitive symptoms to guide clinician decision-making and the development of novel therapeutics. Examination of gene activity through genetic expression analyses allows for the unbiased identification of pathways related to brain injury, including proteomic candidates. Previous studies in civilians with acute nonblast injuries,2 as well as in military personnel with a mix Supplemental data at Neurology.org/ng From the Intramural Research Program, CNRM Co Director Biomarkers Core, Uniformed Services University of the Health Sciences (J.G.) and National Institute of Nursing Research (A.C., N.O., L.A., V.M., H.-S.K., S.Y.), National Institutes of Health, Bethesda; Walter Reed Army Institute of Research (K.C.D., M.L., A.Y.), Silver Spring; Army Medical Research and Materiel Command (W.C.), Fort Detrick; and Naval Medical Research Center (P.W., S.A.), Silver Spring, MD. Funding information and disclosures are provided at the end of the article. Go to Neurology.org/ng for full disclosure forms. The Article Processing Charge was funded by NIH, Army, Navy. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Neurology.org/ng Copyright © 2017 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 of blunt force TBIs and blast-induced TBIs characteristics, duration of military service, and prior blast expo- 4 (biTBIs),3 suggest that gene activities change sure, using a previously described survey. Symptoms related to blast exposure were assessed using a symptom questionnaire4 to following head injuries; however, previous determine the onset of symptoms including headache, dizziness, studies have not acquired baseline samples, nausea/vomiting, sensitivity to sound, sleep disturbance, fatigue, resulting in an inability to control for individ- irritability, depression/sadness, frustration, anxiety, poor mem- ual variability. The objective of this study was ory, poor concentration, longer thinking, blurred vision, light sensitivity, double vision, and loss of balance. The onset of symp- to identify biological pathways implicated in toms was determined if the participant reported symptoms on day biTBIs in a cohort of military personnel who 7, 8, or 9, which were not reported at baseline. Participants also participated in a blast-related training. We provided blood samples on day 1 and on each day of training. Bilateral sensors (Micro Data Acquisition System, mDAS; hypothesized that moderate blast exposure Applied Research Associates, Inc., Albuquerque, NM) affixed would result in differential gene activity, deter- to each participant’s helmet were used to record blast exposures mined through RNA sequencing (RNA-seq), on all training days; the average of the right and left sensors was when comparing postmoderate blast blood used as raw data for peak overpressure (measured in psi). The helmet monitoring system has a threshold of 0.4 psi and records samples collected on day 10 with each partic- changes in ambient pressure, based on both the technological ipant’s own baseline. This study is an initial specifications of the sensors and considerations for signal-to- examination of the association of changes in noise ratio. gene activity following moderate blast expo- Blood sampling. Blood for protein quantification was drawn sure in . Following this, we confirmed from all participants on each training day between 1:600 and 1:800 hours, processed for serum, and stored at 280°C until gene expression by performing quantitative processing. At 1 of 3 training sites, PAXgene tubes were collected PCR (qPCR) and examined impli- at baseline and repeated on day 10; because all moderate blast cated in these gene-activity changes. cases (n 5 29) occurred on day 7 at this site, this enabled us to evaluate gene expression changes from baseline to 3 days follow- – METHODS Standard protocol approvals, registrations, ing moderate blast. Tubes were inverted 8 10 times and were and patient consents. Across 3 separate training sites, 108 placed upright for 2 hours at room temperature, at which point 2 active duty military personnel provided written informed consent they were then moved to a 20°C freezer for 24 hours, and then 2 to participate in this study while engaged in training as either to a 80°C freezer until processing. students or instructors. All protocols were approved by the Insti- Laboratory methods. Matching of cases and controls was com- tutional Review Boards at the Naval Medical Research Center pleted prior to running any laboratory assay. A complementary and Walter Reed Army Institute of Research DNA (cDNA) library was prepared by random fragmentation 4 (NMCR#2011.0002; WRAIR#1796). There is no clinical trial of the cDNA followed by 59 and 39 adapter ligation. A total of identification number associated with the present study. 58 samples from 29 participants in the moderate blast-exposed Participants. Inclusion criteria for the present study included group obtained at baseline and day 10 were assessed with RNA- – the following: (1) be active-duty military service member seq using an Illumina HiSeq 2500 ultra high-throughput participants, (2) be enrolled in an advanced blast training sequencing system with paired-end 101 base paired-end reads. A program at 1 of 3 sites, and (3) be willing to provide blood subset of 4 genes that were found to be dysregulated in RNA-seq samples. The 3 training sites used similar breaching blast ex- were validated using real-time qPCR with TaqMan assay-on- posures across a 2-week period, although the frequency and demand. Briefly, a high-capacity reverse kit type of explosive charges used differed. All sites had stan- (Applied Biosystems, Foster City, CA) was used to synthesize m dardization of data and biological sample collection. The cDNA from 10 L of total RNA. One-tenth of diluted cDNA participants in this study were representative of the larger was mixed with TaqMan assay-on-demand and TaqMan gene military population from which they were drawn. On day 7, expression master mix solution (Applied Biosystems). Triplicate m some participants (site 3) experienced a moderate blast, with reactions each totaling 6 L were analyzed using qPCR a peak pressure greater than 5 psi (n 5 29), and all were (QuantStudio 6 Flex; Thermo Fisher Scientific, Waltham, MA); included in the present study. Participants from the other 2 the data were processed with QuantStudio Real-Time PCR sitesservedasno/low-blastcontrols(n5 79) but were Software version 1.1 (Applied Biosystems). excluded if they experienced a blast exposure exceeding 2 psi For protein analysis, amyloid precursor protein (APP) was chosen during the entire training program (n 5 29), or did not have because (1) the statistical significance reflects the upregulation of APP blood sampling on the days examined (n 5 10). Prior to lab- after moderate blast (vs baseline); (2) the position of APP serves oratory analysis, the moderate blast and no/low-blast groups as a central hub in the ingenuity pathway analysis (IPA) diagram 5 were matched on demographic factors (age and duration of (figure 1) ; (3) the clear biological plausibility for a role of APP in 6– military service) and previous blast exposures; data for race/ biTBI pathology is well established in the preclinical blast literature. 10 ancestry were not available and thus were not used during Specifically, APP concentrations were measured for all participants matching. We undertook a group matching method, with at with an ELISA kit (Thermo Fisher Scientific); samples were random- least a 1:1 ratio of cases to controls. At the end of every training ized to the plates and run in duplicate. The coefficient of variation day, regardless of blast exposure, participants had blood drawn (CV) was assessed as a quality control measure; in the present study, and completed behavioral measures of symptomology. the average CV was 5.7%, and all samples had CVs less than 10%.

Demographic and blast characteristics. Baseline data were Statistical analysis. Overview of analysis. The Statistical Pack- collected on day 1 of training, including demographic age for the Social Sciences (SPSS; version 22; IBM Corporation,

2 Neurology: Genetics Figure 1 Ingenuity pathway analysis results depicting APP as the network with the highest score

The network was generated using IPA (QIAGEN INC., qiagenbioinformatics.com/products/ingenuity-pathway-analysis). APP 5 amyloid precursor protein.

Armonk, NY) was used for all analysis. All figures were developed was used to determine the performance of changes in APP levels using GraphPad Prism. Analysis of variance (ANOVA) was used (from baseline to days 8 and 9) in discriminating between the 2 to compare the 2 groups (no/low vs moderate blast exposed); p groups. values were 2 tailed and adjusted for multiple comparisons. Non- Symptom onset analysis. The onset of symptoms was deter- continuous variables were evaluated using the x2 test. Correla- mined if the participant reported having symptoms on days 2–10 tions between APP protein levels and blast parameters (psi) were that were mild, moderate, or severe and that were not reported at calculated using Pearson correlation. baseline. The onset of symptoms was determined as a dichoto- RNA-seq analysis. An accuracy of Q30 or higher was mous variable, and x2 was used to compare the rates of each of the achieved in 94.95% of base calls. A total of 52.2–75.5 million symptoms. We also used correlations to examine the relationships read counts were performed for each sample. Bioinformatics between APP changes on the onset of total and individual symp- quality control was performed using FastQC, version 0.11.5; toms, using Pearson correlations. both before and after moderate blast, all samples met the quality control criteria for sufficient integrity based on RNA Integrity RESULTS This sample included active-duty service Number and were sequenced. Spliced Transcripts Alignment to members in the Army who were men. The mean age of a Reference version 2.5 was used to align to the hg19 reference the sample was 30.42 years (SD 5 4.84) (table 1). The genome. The number of reads mapped to each gene was counted mean duration of service was 9.94 years (SD 5 4.88), using HTSeq version 0.6.1p1. Differentially expressed genes were and 46.3% had more than 40 previous blast exposures. identified using DESeq2 version 1.12.3; the cutoff for false dis- covery rate by Benjamini-Hochberg method was set to 0.05. The groups did not differ in number of exposures re- Pathway analysis was performed using IPA version 27821452. ported at baseline and duration of military service. Protein analysis. To compare changes within the 2 groups The onset of symptoms in the moderate blast cases over time, a repeated measures ANOVA was performed, and and no/low-blast controls was compared on days 7, 8, Bonferroni correction was applied to determine specific time and 9. On day 7, the onset of concentration problems points where the groups differed. Mean changes from baseline (x2 5 5.43, p 5 0.03) and headaches (x2 5 5.62, p 5 to days 7, 8, 9, and 10 were calculated and compared using an ANOVA. To evaluate the utility of APP as a peripheral biomarker 0.03) was greater in the moderate blast group, with for predicting moderate blast exposure, receiver operating charac- rates of 13.79% and 25.57%. On day 8, headache teristic plots were generated, and the area under the curve (AUC) onset was higher in the moderate blast groups (x2 5

Neurology: Genetics 3 Table 1 Demographics and characteristics for the moderate blast- and no/low blast-exposed groups

Moderate blast Low/no blast exposed (N 5 29) controls (N 5 40) Significance

Mean age in years (SD) 31.2 (4.4) 29.8 (5.1) F1,68 5 1.49, p 5 0.23

Mean years of service (SD) 11.2 (4.7) 9.1 (4.7) F1,61 5 3.22, p 5 0.08

No. of prior explosive breaches and artillery fires, % (no.) x2 5 3.01, p 5 0.32

0–9 20.7 (6) 12.5 (5)

10–39 34.8 (10) 40.0 (16)

40–99 17.2 (5) 17.5 (7)

100–199 20.6 (6) 20.0 (8)

2001 6.9 (2) 10.0 (4)

7.30, p 5 0.01), with a rate of 25.57%. There were exposure group had mean APP level reductions of no differences in any of the symptoms on days com- 3.28 pg/mL (SD 5 3.71), compared with reductions paring groups (moderate blast group and controls) on of 0.98 pg/mL (SD 5 3.01) in the no/low-blast days 1–6, or 9–10, p’s . 0.30, or in the onset of these group. symptoms between the groups on days 2–6or9–10 The mean changes in APP from baseline tended to (p’s . 0.30). Changes in APP did not correlate with correlate with the peak pressure measurements on day the total number of symptom onset (p . 0.30) or 8(r5 0.62, p , 0.01) and day 9 (r 5 0.59, p , with any individual symptom onset (p . 0.30). 0.01). The mean changes in APP from baseline did Statistical analysis of RNA-seq data identified 1,803 not correlate with day 7 peak pressure measurements genes that were differentially expressed after moderate (p . 0.10, figure 3, A–C). Last, the AUC analysis blast exposure (vs baseline), as evidenced by a log2 fold revealed that the change in APP from baseline to day change with an adjusted p value , 0.05. The 1,803 8 (AUC 5 0.802, 95% confidence interval: 0.683– dysregulated genes included 746 upregulated genes and 0.918, p , 0.001) and to day 9 (AUC 5 0.800, 95% 1,058 downregulated genes following a moderate blast confidence interval: 0.687–0.914, p , 0.001) distin- exposure (tables e-1 and e-2 at Neurology.org/ng); the guished moderate blast-exposed participants from top20up-anddown-regulatedgenesaresummarizedin those who experienced no/low-blast exposure. table 2. The IPA results showed that the APP gene network had the highest network score (value 5 43) DISCUSSION The primary finding of the present following moderate blast (figure 1). Real-time qPCR study was that APP protein concentrations signifi- was used to validate the expression changes for 4 genes cantly reduced acutely following a moderate blast related to APP, including genes involved in APP pro- exposure and returned to baseline levels within 3 days cessing and genes whose products interact with APP, of the moderate blast, addressing the objective of the namely amyloid precursor protein (APP), amyloid study, and proving that our hypothesis that gene activ- precursor–like protein 2 (APLP2), nicastrin (NCSTN), ity would change following moderate blast exposure. and NEDD8-Activating Enzyme E1 Subunit 1 This rapid recovery pattern of protein levels may relate (NAE1). Upon qPCR validation, dysregulation of these to the observed increases in APP gene activity on day genes was confirmed (p 5 0.00083476, p 5 0.000105, 10 in the moderate blast group. Taken together, these p 5 0.000673, and p 5 0.043145, respectively). findings suggest that blast exposure alters the activity of Differences were observed in the pattern of APP APP with temporal changes in the biological activity of 5 , concentrations over time (F1,68 12.39, p 0.01) the APP gene after moderate blast exposure. A second- with the moderate blast group having reductions on ary finding was the onset of headaches, concentration days 8 and 9 (all p’s , 0.05), which then returned to problems, and taking longer to think on the 2 days baseline levels on day 10 (figure 2). The mean change following the moderate blast exposure. These findings 5 in APP from baseline to day 8 differed (F1,60 provide unique insights into how peripheral gene activ- , 5 , 18.74, p 0.01) and day 9 (F1,59 20.49, p ity and proteomic activities change following blast and 0.01) in the moderate blast exposure group compared provide insights into how these changes may relate to with the no/low-blast group. The mean APP level neuronal and symptomatic risks. reductions on day 8 in the moderate blast group were The finding of reduced APP levels within the 2 days 4.17 pg/mL (SD 5 3.60), compared with mean following blast was surprising because most preclinical increases of 0.41 pg/mL (SD 5 2.90) in the no/ studies of blast exposures report acute elevations in low-blast group. On day 9, the moderate blast APP in brain tissue. Blast exposure of 22 psi resulted in

4 Neurology: Genetics Table 2 List of top 20 upregulated (A) and downregulated (B) genes in moderate blast-exposed participants

A. Top 20 upregulated genes in blast-exposed personnel B. Top 20 downregulated genes in blast-exposed personnel

Gene name Abbreviation Log2 fold change Adjusted p* Gene name Abbreviation Log2-fold change Adjusted p*

SH3 and PX domains 2B SH3PXD2B 0.9249 1.35E-05 COMM domain-containing 6 COMMD6 20.7681 4.02E-04

Transient receptor potential cation channel subfamily V member 4 TRPV4 0.7874 1.08E-04 Ribosomal protein L31 RPS3A 20.7529 9.77E-04

Disheveled associated activator of morphogenesis 2 DAAM2 0.7777 6.50E-04 Ribosomal protein 3A RPL31 20.7517 7.31E-04

MER proto-oncogene, tyrosine kinase MERTK 0.6767 9.75E-04 Keratinocyte growth factor–like protein 2 KGFLP2 20.7366 5.62E-04

Solute carrier family 9 member 7 pseudogene 1 SLC9A7P1 0.6504 1.35E-04 Ribosomal protein 7 RPS7 20.7309 1.38E-03

ELOVL fatty acid elongase 3 ELOVL3 0.6408 3.18E-03 Ribosomal protein L9 RPL9 20.7144 1.77E-03

Growth factor receptor–bound protein 10 GRB10 0.6395 2.28E-04 MicroRNA 5195 MIR5195 20.6867 1.44E-03

Syndecan 3 SDC3 0.6183 1.28E-03 Ribosomal protein S27L RPS27L 20.6828 1.02E-03

V-Set and immunoglobulin, domain containing 4 VSIG4 0.6086 2.12E-03 Ribosomal protein L39 RPL39 20.6782 2.22E-03

Semaphorin 6B SEMA6B 0.6026 1.05E-02 Zinc finger protein 571 ZNF571 20.6691 1.61E-03

Hypermethylated in cancer 1 HIC1 0.5902 1.32E-03 Zinc finger protein 157 ZNF157 20.6580 4.61E-03

Complement component 1, Q subcomponent, B chain C1QB 0.5842 1.13E-02 Ribosomal protein S24 RPS24 20.6534 1.97E-03

Calcium/calmodulin-dependent protein kinase kinase 1 CAMKK1 0.5763 3.51E-05 Long intergenic nonprotein–coding RNA 1,431 LINC01431 20.6497 3.22E-03

TBC1 domain family member 8 TBC1D8 0.5704 1.08E-04 Ribosomal protein L23 RPL23 20.6493 3.22E-03

Aldehyde dehydrogenase 2 family ALDH2 0.5688 1.08E-04 Ribosomal protein L26 RPL26 20.6440 3.71E-03

Phospholipase B domain containing 1 PLBD1 0.5672 2.14E-04 Ribosomal L24 domain-containing 1 RSL24D1 20.6437 4.58E-03

2

erlg:Genetics Neurology: Versican VCAN 0.5616 2.14E-04 Ribosomal protein 15A RPS15A 0.6420 3.79E-03

CD177 CD177 0.5616 1.00E-02 Cytochrome C oxidase subunit 7B COX7B 20.6415 3.52E-03

Potassium voltage-gated channel subfamily E regulatory subunit 1 KCNE1 0.5602 1.65E-03 Tubby bipartite transcription factor TUB 20.6275 6.50E-04

Gamma-glutamyltransferase 5 GGT5 0.5565 1.52E-02 Uncharacterized LOC100996579 20.6264 5.78E-03 5 peripheral levels of other biomarkers after TBI.19 How- Figure 2 Average APP concentrations in the moderate (n 5 29) and no/low (n 5 40) blast-exposed groups during the 10-day period ever, it is also possible that the APP detected was from well-studied noncentral sources, including platelets, mononuclear blood cells, and other extraneuronal cell types20–23; notably, given the whole-body nature of the blast exposure the military personnel sustained, periph- eral expression changes cannot be excluded. APP changes after blast exposure also may be relevant to the identification of potential therapeutic targets. Measuring symptoms acutely, and the months and years following a blast injury, is critical for creating an accurate clinical profile for diagnosing and predicting the trajectory of recovery for biTBI in military per- sonnel. The onset of reported headaches on days 7 and 8 builds on a previous study that found blast- exposed military personnel more frequently reported headaches and tinnitus than those with nonblast inju- ries.24 While there are previous reports of minimal Differences in APP concentrations over time (F1,68 5 12.39, p , 0.01) were observed, with the moderate blast group exhibiting reductions on days 8* and 9* (p’s , 0.05). APP 5 differences between the cognitive symptoms after 25 amyloid precursor protein. biTBIs and nonblast TBIs, very little is known regarding the specific cognitive sequelae of biTBI, APP reactive axons within the white matter tracks of the in clinical samples. In a study measuring the acute cervical spinal cord in rats at 6- and 24-hour postexpo- symptoms following a biTBI and long-term out- sure.7 In a similar study, bAPP1 axonal profiles were comes, loss of consciousness and altered mental status reported on days 1 and 2 in the white matter following predicted posttraumatic stress disorder and postcon- a70psiblast.8 Another study showed that APP levels cussive syndrome.26 In the present study, the onset of increased in the brains of rats and mice after blast over- symptoms, including taking longer to think and con- pressure, though no evidence of APP staining in axons centration problems observed on days 7 and 8, sug- was found.10 A similarly designed study reported that an gests that biTBIs may initiate immediate cognitive open-field blast exposure (peak pressure 5 26.1 psi) in repercussions that quickly resolve by day 9. A limita- rats led to decreased cerebral blood flow on days 3 and 5 tion of this study is the lack of longer-term protein postinjury and an increase in bAPP in the neuronal levels and symptom profiles; the trajectory of symp- perikarya of the cerebral cortex.9 Notably, most preclin- toms may prove critical for guiding the clinical and ical models use higher blast exposure levels to induce rehabilitation recommendations postinjury. Despite TBIs and examine central changes by analyzing brain this limitation, however, these findings have impor- samples directly. This method, however, is not feasible tant clinical implications for further understanding in clinical cohorts, where peripheral blood is predomi- acute symptomatic outcomes following blast expo- nantly studied. This study reports decreases in peripheral sure. Moreover, though the sample was generally rep- APP in a military cohort exposed to primary blast. resentative of the military, it was composed It is possible that the acute decrease on days 8 and exclusively of men of a narrow age range and may 9 reported in this study may have clinical consequen- not be generalizable to women or older individuals. ces. APP has been well established as being elevated in Developing a better understanding of APP changes the brain acutely after biTBIs,6–10 as well as blunt after blast is important because it is probable that force TBIs in preclinical models and penetrating in- APP-related pathology may contribute to risk of chronic juries in military personnel.11,12 However, it has been neurodegenerative conditions. One study of World War posited that APP plays a neuroprotective role after II veterans found that moderate and severe TBIs were TBI.13 Traditionally, the ubiquitously expressed associated with increased risk of Alzheimer disease,27 APP has been considered a marker of axonal injury, a condition also well characterized by changes in APP. known to increase after TBI and to relate to the Moreover, a study using post mortem brain tissue of severity of neuronal injury.14,15 APP has also been military members found that among those with a history implicated in neuronal plasticity and signaling,16 for- of blast exposure, APP1 axonal varicosities were present mation of neurites,17 as well as prevention of iron in the white matter of the frontal cortex, corpus callos- accumulation and subsequent iron-induced oxidative um, cerebellum, and superior cerebellar peduncle.28 The damage after TBI.18 Thus, the observed reductions in findings of this study, in the context of existing pub- the peripheral APP activity may relate to central lished evidence, suggest that changes in APP occur after changes, as it has been found for central and blast exposure and remain to be further clarified.

6 Neurology: Genetics The observations of a restoration of APP levels 3 days Figure 3 Changes in APP from baseline to days 7, 8, and 9 in the moderate (n 5 29) and no/low (n 5 40) blast-exposed groups following the blast are likely due to compensatory upre- gulation of gene expression; in this study, increased APP gene activity was observed at day 10 (compared with baseline) using RNA-seq with qPCR validation. Upregulation of APP messenger RNA has been previ- ously reported after experimental TBI,6 suggesting that this central activity may result in the rebound of APP concentrations. Still, future studies should clarify the mechanism of the changes in APP levels by including additional time points to relate protein changes to gene activity. It could be that unexplored changes in gene expression early after blast lead to downregulation of APP. Another possible explanation is that sequestration of APP is occurring somewhere in the CNS leading to decreased peripheral levels; this speculation may be the case if APP is playing a protective role within the brain. Likewise, blast may alter the flow of proteins from the brain to the periphery through alterations in glymphatic flow, clearance mechanisms, homeostatic control mech- anisms, or currently undiscovered mechanisms. This study reports temporal changes in APP biological activity after a moderate blast, suggesting a role for APP in biTBI pathophysiology; the rebound in APP levels observed could be due to compensatory mechanisms after mod- erate blast, representing avenues for further research. These findings from a well-characterized cohort pro- vide unique insights surrounding how peripheral changes in gene expression and levels of associated protein are altered after blast exposure. A major limitation of the cur- rent study is that an examination of the impact of lower blast exposure if not included on gene activity or APP. Our goal for this pilot study was to optimize discrimina- tion between groups, and for this reason we excluded those participants with blast between 2 and 5 psi on any training day, resulting in an inability to determine the impact of lower level blast exposures. Future studies would be strengthened by the recruitment of larger, more diverse samples and the inclusion of a variety of blast ex- posures and additional time points to gain a comprehen- sive understanding of the mechanism(s) underlying the changes in protein biomarkers reported in the present study. This study makes key contributions to the current understandings of the implications of biTBIs, while raising important questions to guide future inquiry.

AUTHOR CONTRIBUTIONS J.G., A.C., N.O., L.A., V.M., H.-S.K., and S.Y.: contributions to the con- ception and design of the work, laboratory analysis, interpretation of the data, and drafting the work. K.C.D., W.C., P.W., S.A., M.L., and A.Y.: There was no correlation between mean changes in APP from baseline to peak pressure contributions to the collection, interpretation of the data, and drafting of measurements on day 7 (p . 0.10) (A). The mean changes in APP from baseline correlated the work. All authors provided final approval of the manuscript. with peak pressure measurements (in psi) on day 8 (r 5 0.62, p , 0.01) (B) and day 9 (r 5 0.59, p , 0.01) (C). APP 5 amyloid precursor protein. ACKNOWLEDGMENT Thank you to the leadership from the U.S. Army Special Operations Command and U.S. Army Engineer School for their support. We also thank the military personnel for their service to our nation and their participation in this study.

Neurology: Genetics 7 STUDY FUNDING changes in rat brainstem and spinal cord. J Neurosci Rural This study was funded by the National Institutes of Health, Intramural Pract 2015;6:481–487. Department of Research, and the U.S. Army Medical Research and 8. del Mar N, von Buttlar X, Yu AS, Guley NH, Reiner A, Materiel Command and the U.S. Navy Bureau of Medicine and Surgery Honig MG. A novel closed-body model of spinal cord injury (WUN 603115HP.2380.001.A1304). caused by high-pressure air blasts produces extensive axonal injury and motor impairments. Exp Neurol 2015;271:53–71. DISCLOSURE 9. Verma SK, Kan EM, Lu J, et al. Multi-echo susceptibility- Material has been reviewed by the Walter Reed Army Institute of Research. weighted imaging and histology of open-field blast-induced There is no objection to its publication. The opinions and assertions con- traumatic brain injury in a rat model. NMR Biomed 2015; tained herein are the private views of the authors and are not to be con- 28:1069–1077. strued as official, or as reflecting true views of the Department of the 10. De Gasperi R, Gama Sosa MA, Kim SH, et al. Acute blast Navy, Department of the Army, the Department of Defense, the Uni- injury reduces brain abeta in two rodent species. Front formed Services University of the Health Sciences, or any other agency of the U.S. Government. The study protocol was approved by the Naval Med- Neurol 2012;3:177. ical Research Center and Walter Reed Army Institute of Research Institu- 11. Thelin EP, Just D, Frostell A, et al. Protein profiling in tional Review Boards in compliance with all applicable Federal regulations serum after traumatic brain injury in rats reveals potential governing the protection of subjects. The investigators have adhered injury markers. Behav Brain Res Epub 2016 Aug 31. to the policies for protection of human subjects as prescribed in AR 70–25. 12. Cartagena CM, Mountney A, Hwang H, et al. Subacute MAJ Carr, LCDR Walker, MAJ LoPresti, and MAJ Yarnell are military changes in cleavage processing of amyloid precursor pro- service members. Dr. Ahlers is an employee of the U.S. Government. This tein and tau following penetrating traumatic brain injury. work was prepared as part of their official duties. Title 17 U.S.C. §105 PLoS One 2016;11:e0158576. provides that ‘Copyright protection under this title is not available for any 13. Plummer S, Van den Heuvel C, Thornton E, Corrigan F, work of the United States Government.’ Title 17 U.S.C. §101 defines Cappai R. The neuroprotective properties of the amyloid a U.S. Government work as a work prepared by a military service member or employee of the U.S. Government as part of that person’s official duties. precursor protein following traumatic brain injury. Aging – J. Gill and A. Cashion have received research support from NIH. N. Osier, Dis 2016;7:163 179. L. Arcurio, and V. Motamedi report no disclosures. K.C. Dell is an 14. Blumbergs PC, Scott G, Manavis J, Wainwright H, Simp- employee of Walter Reed Army Institute of Research (WRAIR). W. Carr son DA, McLean AJ. Staining of amyloid precursor pro- has received research support from the United States Government (he is tein to study axonal damage in mild head injury. Lancet a military service member) from the U.S. Army Medical Research and 1994;344:1055–1056. Materiel Command and from the U.S. Navy Bureau of Medicine. 15. Gentleman SM, Nash MJ, Sweeting CJ, Graham DI, Rob- H.-S. Kim reports no disclosures. S. Yun is the CEO of Yotta Biomed, erts GW. Beta-amyloid precursor protein (beta APP) as LLC and has received research support from NINR, NIDDK, and NHLBI. a marker for axonal injury after head injury. Neurosci Lett P. Walker has received research support from Joint Program Committee 1993;160:139–144. FIVE, Occupational Standards for Repeated Blast Exposure, BUMED WII, and the Office of Naval Research. S. Ahlers has received research support 16. Kogel D, Deller T, Behl C. Roles of amyloid precursor from the U.S. Army (JPC5 intramural award, Blast Exposure Standards). protein family members in neuroprotection, stress signal- – M. LoPresti has received research support from the United States Govern- ing and aging. Exp Brain Res 2012;217:471 479. ment (he is a military service member). A. Yarnell has received research 17. Small DH, Nurcombe V, Reed G, et al. A heparin-binding support from the U.S. Army (JPC5 intramural award, Blast Exposure domain in the amyloid protein precursor of Alzheimer’s Standards). Go to Neurology.org/ng for full disclosure forms. disease is involved in the regulation of neurite outgrowth. J Neurosci 1994;14:2117–2127. Received April 11, 2017. Accepted in final form June 30, 2017. 18. Ayton S, Zhang M, Roberts BR, et al. Ceruloplasmin and beta-amyloid precursor protein confer neuroprotection in REFERENCES traumatic brain injury and lower neuronal iron. Free Radic 1. Tanielian T, Jaycox L. Invisible Wounds of War. Santa Biol Med 2014;69:331–337. Monica: RAND Corporation; 2008. 19. Pérez-Bárcena J, Ibañez J, Llompart-Pou JA, et al. Corre- 2. Livingston WS, Gill J, Cota M, et al. Differential gene lation of cytokine levels in peripheral blood, brain inter- expression associated with meningeal injury in acute mild stitial space and cerebrospinal fluid in traumatic head TBI. J Neurotrauma 2016;34:853–860. injury: P 012. Eur J Anaesthesiology 2008;25:4–5. 3. Heinzelmann M, Reddy SY, French LM, et al. Military 20. Shi C, Zhu X, Wang J, Long D. Estrogen receptor alpha personnel with chronic symptoms following blast trau- promotes non-amyloidogenic processing of platelet amy- matic brain injury have differential expression of neuronal loid precursor protein via the MAPK/ERK pathway. recovery and epidermal growth factor receptor genes. J Steroid Biochem Mol Biol 2014;144(pt B):280–285. Front Neurol 2014;5:198. 21. Padovani A, Pastorino L, Borroni B, et al. Amyloid precur- 4. Carr W, Yarnell AM, Ong R, et al. Ubiquitin carboxy-terminal sor protein in platelets: a peripheral marker for the diagnosis hydrolase-l1 as a serum neurotrauma biomarker for exposure to of sporadic AD. Neurology 2001;57:2243–2248. occupational low-level blast. Front Neurol 2015;6:49. 22. Jiang S, Zhang M, Ren D, et al. Enhanced production of 5. Kramer A, Green J, Pollard J Jr, Tugendreich S. Causal amyloid precursor protein mRNA by peripheral mononu- analysis approaches in ingenuity pathway analysis. Bioin- clear blood cell in Alzheimer’s disease. Am J Med Genet B formatics 2014;30:523–530. Neuropsychiatr Genet 2003;118B:99–102. 6. Van Den Heuvel C, Blumbergs P, Finnie J, et al. Upre- 23. Niederwolfsgruber E, Schmitt TL, Blasko I, et al. The pro- gulation of amyloid precursor protein and its mRNA in an duction of the Alzheimer amyloid precursor protein (APP) in experimental model of paediatric head injury. J Clin Neu- extraneuronal tissue does not increase in old age. J Gerontol A rosci 2000;7:140–145. Biol Sci Med Sci 1998;53:B186–B190. 7. Kallakuri S, Purkait HS, Dalavayi S, VandeVord P, 24. Wilk JE, Thomas JL, McGurk DM, Riviere LA, Castro CA, Cavanaugh JM. Blast overpressure induced axonal injury Hoge CW. Mild traumatic brain injury (concussion) during

8 Neurology: Genetics combat: lack of association of blast mechanism with persis- symptoms on mental health and service discharge out- tent postconcussive symptoms. J Head Trauma Rehabil comes. J Neurotrauma 2013;30:1391–1397. 2010;25:9–14. 27. Plassman BL, Havlik RJ, Steffens DC, et al. Docu- 25. Luethcke CA, Bryan CJ, Morrow CE, Isler WC. Compar- mented head injury in early adulthood and risk of ison of concussive symptoms, cognitive performance, and Alzheimer’s disease and other dementias. Neurology psychological symptoms between acute blast-versus non- 2000;55:1158–1166. blast-induced mild traumatic brain injury. J Int Neuro- 28. Ryu J, Horkayne-Szakaly I, Xu L, et al. The problem of psychol Soc 2011;17:36–45. axonal injury in the brains of veterans with histories 26. Eskridge SL, Macera CA, Galarneau MR, et al. Influence of blast exposure. Acta Neuropathol Commun 2014; of combat blast-related mild traumatic brain injury acute 2:153.

Neurology: Genetics 9 Whole-exome sequencing associates novel CSMD1 gene mutations with familial Parkinson disease

Javier Ruiz-Martínez, ABSTRACT MD, PhD Objective: Despite the enormous advancements made in deciphering the genetic architecture of Luis J. Azcona, BBA Parkinson disease (PD), the majority of PD is idiopathic, with single gene mutations explaining only Alberto Bergareche, MD a small proportion of the cases. Jose F. Martí-Massó, MD, Methods: In this study, we clinically evaluated 2 unrelated Spanish families diagnosed with PD, in PhD which known PD genes were previously excluded, and performed whole-exome sequencing anal- Coro Paisán-Ruiz, PhD yses in affected individuals for disease gene identification. Results: Patients were diagnosed with typical PD without relevant distinctive symptoms. Two dif- Correspondence to ferent novel mutations were identified in the CSMD1 gene. The CSMD1 gene, which encodes Dr. Paisán-Ruiz: a complement control protein that is known to participate in the complement activation and [email protected] inflammation in the developing CNS, was previously shown to be associated with the risk of PD in a genome-wide association study. Conclusions: We conclude that the CSMD1 mutations identified in this study might be responsible for the PD phenotype observed in our examined patients. This, along with previous reported studies, may suggest the complement pathway as an important therapeutic target for PD and other neurodegenerative diseases. Neurol Genet 2017;3:e177; doi: 10.1212/NXG.0000000000000177

GLOSSARY AD 5 Alzheimer disease; CCP 5 complement control protein; fPD 5 familial Parkinson disease; H&Y 5 Hoehn and Yahr; INDEL 5 insertions/deletions; LOPD 5 late-onset PD; PD 5 Parkinson disease; RBD 5 REM sleep behavior disorder; RLS 5 restless legs syndrome; SNV 5 single nucleotide variant; WES 5 whole-exome sequencing.

Parkinson disease (PD; MIM# 168600) is the second most common neurodegenerative disease affecting more than 4 million people worldwide.1,2 PD is thought to be caused by a combination of genetic and environmental risk factors, with only 5%–10% of patients reporting to have a mendelian pattern of inheritance.3–5 To date, 3 different genes, SNCA (MIM# 163890), LRRK2 (MIM# 609007), and VPS35 (MIM# 601501), have been linked to autosomal dom- inant PD,6–8 with 2 of these (SNCA and LRRK2) being well established as susceptibility loci for sporadic PD.9–11 In addition, common single nucleotide variations within GBA, MAPT, MCCC1, STK39, GAK, BST1, SYT11, ACMSD, and the HLA region have been reported to be associated with the risk of PD through genome-wide association studies performed by numerous different scientific groups.9,12–16 However, in the majority of these associated genes, no pathogenic (or rare) mutation has been identified in familial or sporadic PD, except for the ACMSD gene, in which pathogenic mutations have been reported in complex parkinsonism and idiopathic PD.17

From the Department of Neurology (J.R.-M., A.B., J.F.M.-M.), University Hospital Donostia, San Sebastián, Spain; Neurosciences Area, Bio- donostia Research Institute (J.R.-M., A.B., J.F.M.-M.), San Sebastián, Spain; Center for Biomedical Research in Neurodegenerative Diseases Network (CIBERNED) (J.R.-M., A.B., J.F.M.-M.), Madrid, Spain; Department of Neuroscience (L.J.A.), Departments of Neurology, Psychiatry, and Genetics and Genomic Sciences (C.P.-R.), Friedman Brain and Mindich Child Health and Development Institutes (C.P.-R.), Icahn School of Medicine at Mount Sinai, New York, NY; and Department of Neuroscience (J.F.M.-M.), University of the Basque Country—UPV-EHU, San Sebastián, Spain. Funding information and disclosures are provided at the end of the article. Go to Neurology.org/ng for full disclosure forms. The Article Processing Charge was funded by the NINDS (NIH). This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Neurology.org/ng Copyright © 2017 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology 1 In this study, we clinically evaluated 2 dif- for annotation of the resulting calls,24 and PICARD was used ’ ferent families suffering from late-onset PD to conduct exomes statistics (picard.sourceforge.net/). Genomic variations observed as common mutations (fre- (LOPD) without mutations in the known quency . 5%) in the latest dbSNP149 build, 1000 Genomes genes18 and performed whole-exome sequenc- Project Phase 3, the Exome Variant Server of the National Heart, ing (WES) analyses in 3 affected family mem- Lung, and Blood Institute (NHLBI) Exome Sequencing Project 25 bers to identify the genetic causes of disease to (evs.gs.washington.edu/EVS/), and the Exome Aggregation Consortium (exac.broadinstitute.org/) were excluded from sub- enhance our knowledge of the genetic archi- sequent analyses, as were genomic variations mapped to intra- tecture of PD. We identified 2 different muta- genic, intronic, and noncoding exonic regions. tions in a novel gene, previously reported to be Given the low frequency of pathogenic autosomal dominant PD mutations in public databases, which are either not present associated with the risk of PD, as possible (SNCA, VPS35) or identified with very low frequency (LRRK2, disease-causing mutations in both unrelated 1E-06), with the exception of LRRK2 p.Gly2019Ser mutation PD families. that is identified with a higher frequency (1E-04) because of its prevalence in European population, novel and genomic variations METHODS Standard protocol, approvals, registrations, with very low frequency were prioritized in follow-up analyses. and patient consents. Two individual families with LOPD and The pathogenicity of the identified mutations was predicted of Basque origin and 1 isolated familial PD (fPD) case were clin- using the following computational methods: MutPred ically examined. Patients of 3 families were diagnosed and treated (mutpred.mutdb.org/), SNPs&Go (snps-and-go.biocomp.uni- by a group of movement disorder specialists (J.R.-M., A.B., and bo.it/snps-and-go/), MutationTaster (mutationtaster.org/), and 26 J.F.M.-M.) at the University Hospital Donostia and were CADD tool. We also used the HomoloGene database (ncbi. included as having fPD without mutations in the known genes. Mu- nlm.nih.gov/homologene/) to examine the conservation of the tations in known genes were previously excluded through custom mutated CSMD1 protein across different species, as well as the targeted sequencing.18 Patients were diagnosed according to the professional Human Gene Mutation Database (HGMD) (portal. UK PD Brain Bank Society and Gelb criteria.19,20 Written informed biobase-international.com/hgmd/pro/start.php) and the NCBI consent, fully approved by the local ethics committee of the Hospital ClinVar database (ncbi.nlm.nih.gov/clinvar/) to determine Universitario Donostia, was obtained from all participants. whether CSMD1 mutations were known to be associated with DNA samples from 115 Spanish patients with PD and 94 a disease phenotype. DNA samples belonging to ethnicity-matched neurologically Candidate gene screening. To amplify all coding CSMD1 normal individuals (45 men and 49 women) without a family his- exons, genomic primers were designed using the ExonPrimer script tory of any movement disorders were also available for genetic (ihg.gsf.de/ihg/ExonPrimer.html) (primer sequences available on screening. The age at sample collection of the control individuals request), and PCR products were purified using the ExoSap-IT 17 ranged from 60 to 93 years with an average of 69.1 years. The reagent (Applied Biosytems, Foster City, CA). These were NDPT102 Parkinson panel from the Coriell Institute for Med- sequenced in both forward and reverse directions using Applied ical Research (coriell.org/) that contains DNA from 92 unique Biosystems BigDye terminator v3.1 sequencing chemistry as per and unrelated Caucasian individuals with idiopathic PD (59 men the manufacturer’s instructions, resolved on an ABI3730 genetic and 33 women) was also used for mutational screening. analyzer (Applied Biosytems), and analyzed using Sequencher All available DNA samples were isolated from whole blood 5.4.1 software (Gene Codes Corporation, Ann Arbor, MI). using standard procedures.

Whole-exome sequencing. Four DNA samples belonging to RESULTS Clinical history. Family A: Patient A1. This 2 different families with LOPD (cases A1, B1, and B2) and patient began feeling tired, clumsy, and very slow in 1 isolated familial case were subject to WES analyses (figure, movements at the age of 72 and was diagnosed with 21 A), which were conducted as previously described. The PD 1 year later. Initially, he did not take any treat- SureSelect Human All Exon 50Mb exon-capture kit was used for library enrichment (Agilent Technologies Inc., Santa ment, but then came back to the clinic with slowness, Clara, CA), and captured libraries were sequenced on the clumsiness, and a rest tremor in his right hand. On HiSeq2000 according to the manufacturer’s instructions for examination, he showed an amimic face, hypophonia, paired-end 100-bp reads (Illumina Inc., San Diego, CA), and dysarthria. He had a resting tremor (1/4) in the using a single flow cell lane per sample. Sequencing data were right hand as well as rigidity in the neck (3/4 neck put through a computational pipeline for WES data process- stiffness), right arm (2/4), and left arm (1/4). He ing and analysis following the general workflow adopted by had a moderate global akinesia (3/4) with axial distur- the 1000 Genomes Project,22 where raw sequence reads were aligned to the human reference genome sequence (NCBI bance. Reflexes were normal. He was diagnosed with GRCh37.p13) using the fast lightweight Burrows-Wheeler PD—score 2 according to the Hoehn and Yahr Alignment Tool (BWA),23 followed by a base-quality recali- (H&Y) scale. This condition was treated with carbi- bration and local realignment through the Genome Analysis dopa/levodopa (75/300 mg/d) with improvement, Toolkit (GATK v1.5-16-g58245bf). Single nucleotide sub- and a year later cabergoline (2 mg/d) and rasagiline stitutions (single nucleotide polymorphism/single nucleotide (1 mg/d) were administered. Later, cabergoline was variant) and short insertions/deletions (INDELs) were called using the GATK Unified Genotyper tool, where calls were replaced with rotigotine (8 mg/d). At the age of 78, he filtered based on mapping quality (q30 or higher) and depth was autonomous in daily life, but a year later, his of coverage (d10 or higher). Last, the AnnTools kit was used postural instability increased with more clumsiness

2 Neurology: Genetics Figure CSMD1 mutations in Parkinson disease

(A) Pedigree structures of 2 families with LOPD and CSMD1 mutations; family A is shown on the left side, whereas family B is shown on the right side. Wt/p.R1962H: heterozygous carriers for the CSMD1 p.R1962H mutation; Wt/p.G2987R: heterozygous carriers for the CSMD1 p.G2987R mutation; Wt/Wt: noncarriers. (B) Sanger chromatogram sequences for both CSMD1 mutations identified in this study are shown at the top, while conservation of both mutations across different species is shown at the bottom. The exact position of the mutations is highlighted with a red arrow. (C) CSMD1 encoding protein structure and the localization of both mutations identified in this study. LOPD 5 late-onset Parkinson disease. when turning in bed and hypophonia. At the age of disorder (RBD), associated with anxiety precipitated 80, he had an important gait disturbance with freez- by a family conflict and restless legs syndrome (RLS) ing but not falls, oscillations, and moderate dyskine- in the last 2 years. On examination, she showed mild sias without cognitive impairment. He died at the age amimia as well as neck and right arm rigidity. She had of 86 of respiratory infection. a slight decrease in arm movements in her walk, He had a family history of PD with a paternal aunt micrography, and bilateral babinski signs with mild with a diagnosis of PD, a father with a possible resting hyperreflexia. Brain MRI study was normal, and tremor in the elderly, a brother who died without PD DATSCAN showed bilateral striatal hypocaptation, at the age of 77, a healthy sister (87 years old), mainly on the left side. She was diagnosed with PD, a brother with cognitive impairment (86 years old), and initially began treatment with rasagiline (1 mg/d) a daughter with PD (58 years old; figure, A; patient and later with carbidopa-levodopa (75/300 mg/d). In A2), and a daughter without PD (53 years old). the following years, she has been presenting with Family A: Patient A2. This patient was first seen at 49 increasing oscillations and generalized dyskinesias years of age with a 6-month evolution of bradykinesia that have progressed to severe oscillations, difficulty in her right hand affecting her daily activities, in walking with some degree of ataxia, and severe including writing, moving, dressing, and so on. She instability with falls over the last year. Her cognitive also noticed in writing notes that her letters were state remains normal. smaller. She reported trouble sleeping because of Family B: Patient B1. This patient was examined for symptoms compatible with REM sleep behavior the first time at the age of 70, when he presented with

Neurology: Genetics 3 Table 1 CSMD1 mutations in familial PD

Computational prediction results Control population

Caucasian Position Nucleotide Protein Gene Protein MutPred SNPs&Go MutationTaster CADD Spanish (NDPT102) Brain FAM Chr (bp) change change (exon) domain expression

F.B 8 3,015,448 c.5885G.A p.R1962H CSMD1 (339) CCP 0.631 0.197 (Neutral) Disease 32.0 0/94 0/92 Yes

F.A 8 2,824,233 c.8959G.A p.G2987R CSMD1 (358) CCP 0.940 0.922 (Disease) Disease 29.1 0/94 0/92 Yes

Abbreviations: CCP 5 complement control protein; PD 5 Parkinson disease.

tremor in his right hand of 6 months of duration. At and was diagnosed with enhanced physiologic tremor examination, he showed rest tremor and akinesia with that aggravated by anxiety. At the age of 61, she came mild rigidity in his right hand. He was diagnosed with back to the clinic because of her tremor. She had shak- PD—score 1 according to the H&Y scale—and was ing hands without cephalic postural tremor, but nei- treated with carbidopa-levodopa (75/300 mg daily) ther rigidity nor slowness was observed. Her tremor and selegiline with good response. In the following 3 increased with nervousness. She reported to have years, his symptoms progressed to bilateral and he symptoms suggestive of RLS. No constipation, hypo- developed dyskinesias 6 years after levodopa initiation. smia, dizziness, or symptoms suggestive of RBD were He was autonomous in daily life until the age of 83, observed. The levels of thyroid hormones and other and he could go outside without assistance. He subse- analytic studies were normal. quently had postural instability, hypophonia, dysar- Genetic results. WES approaches were performed in 3 thria, and dysexecutive mild cognitive impairment, different family members (cases A1, A2, and B1) but never became demented. At the age of 93, he suf- belonging to 2 PD families (families A and B) and 1 iso- fered from lateral bulbar infarction related to embolism lated familial case. Between 96.38% (case A2) and because of atrial fibrillation, with minor sequels. Two 93.56% (case A1) of the target exome at 20-fold cov- months later, he had a fall with a chest contusion and erage or higher was captured for all sequenced samples. fracture of the 10th rib. During his hospital admission, This led to the identification of 979 coding genetic he had acute delirium, dyspnea, and finally died. variations for case A1, 1,560 for case A2, 1,132 for case He had a family history of PD. His mother, who B1, and 907 for the isolated familial case. After filtering died at the age of 75, suffered from PD and rest and by including only novel or with very low frequency tremor since the age of 70. Her brother died of pan- genomic variation (1E-04–1E-06), 26 SNVs (12 novel creatic cancer at the age of 79. This brother had and14withlowfrequency)werefoundtobesharedby 4children:asonwhowasdiagnosedwithPDat the 2 affected members of family A. We then searched if theageof51(heisnow59yearsold)aswellas2 any of the genes, where the shared SNVs were located, other sons and 1 daughter who died without neuro- were also mutated in family B and the isolated case, and degenerative disease. The patient (patient B1) had identified 2 novel heterozygous mutations, not pre- 3 children: a son who died of sepsis at the age of viously reported in public databases, in the CSMD1 53, a 68-year-old healthy daughter, and a 63- gene (MIM# 608397) in both affected families A and year-old daughter with postural tremor (figure, A; B. No novel genetic variation was found to be shared patient B2). between the 2 families (A and B) and the isolated fPD Family B: Patient B2. This patient was first examined case. A G-to-A transition (c.5885G.A) resulting in p. at the age of 57 when she presented with tremor in the Arg1962His amino acid substitution was identified in hands of 2 years of evolution. She feared having fPD both affected members from family A, while a G-to-A related to her father’s PD. She had postural tremor transition (c.8959G.A) resulting in p.Gly2987Arg of small amplitude and frequency 8 Hz in the hands amino acid substitution was identified in the only family member of family B subject to WES (table 1). Analysis of this second mutation (p.Gly2987Arg) in Table 2 Common CSMD1 variation: NGRC data set28 additional family members revealed that this mutation segregated with the disease status (figure, A and B). We Familial 1 sporadic Sporadic PD PD also found that both mutated CSMD1 amino acids are Allele highly conserved across different species (figure, B) as p p Chr SNP Position (bp) (minor/major) OR SE OR SE wellasCSMD2andCSMD3proteins(datanot 8 rs12681349 4,277,990 T/C 0.75 0.04 5E-08 0.78 0.04 7E-07 shown) and that both mutations are located in different

Abbreviations: OR 5 odds ratio; PD 5 Parkinson disease; SNP 5 single nucleotide poly- complement control protein (CCP) domains of the morphism. translated CSMD1 protein (figure, C). The CCP

4 Neurology: Genetics domains, containing approximately 60 amino acid res- CSMD1, to be mutated in both families. Both CSMD1 idues, have been identified in several proteins of the mutations segregated with disease status, were not pre- complement system that is part of the innate immune viously reported in public databases, were absent in a large system. The CSMD1 gene is also found to be weakly number of neurologically normal individuals, were expressed in most tissues, except in the brain, where it is highly conserved among other orthologous, and were expressed at an intermediate level in the cerebellum, predictedtobepathogenicby various computational substantia nigra, hippocampus, and fetal brain.27 Both methods. This, along with previously reported CSMD1 novel CSMD1 mutations were predicted highly patho- association with the risk of PD,28 led us to believe in genic by various computational methods. We then apossibleroleofCSMD1 genetic variability in the path- screened both CSMD1 mutations in 372 control ogenesis of PD. The CSMD1 gene contains 70 coding , including 188 ethnicity-matched control exons and encodes for a large protein (3,564 aa) that chromosomes, and did not identify any additional contains multiple CUB and Sushi, also known as mutation carrier. No mutation carrier was identified in CCP, domains. It is primarily synthesized in the devel- 115 DNA samples of Spanish patients with PD tested oping CNS and epithelial tissues, and its encoding pro- through Sanger sequencing. However, a genome-wide tein is known to act as a regulator of complement significant association between a CSMD1 nucleotide activation and inflammation in the developing CNS.29 variation (rs12681349; 2) and PD has been Complement activation is essential for synaptic prun- recently reported using the NeuroGenetics Research ing and plasticity and has recently been implicated in Consortium (NGRC) data set, which includes 435 fPD, several brain-related disorders and functions, includ- 1,565 sporadic PD, and 1,986 control cases (table 2).28 ing , Alzheimer disease (AD), immediate This association remained highly significant when episodic memory, and information processing.30–32 In including all PD (familial and sporadic) and sporadic particular, based on mice models of AD, it can be said PD cases. And although it was also observed when that both the microglia and the component pathway considering only fPD, this was not statistically signifi- might act as early mediators of hippocampal synapse cant, probably because of power limitations. loss and dysfunction before plaque formation and neu- roinflammation.32 In addition, copy number varia- DISCUSSION We here described the identification tions within the CSMD1 locus have been reported of a novel gene (CSMD1)tobemutatedinunrelated in patients with AD vs controls,33 further supporting Basque families with LOPD in which known PD genes a role of the CSMD1 genetic variability and function were previously excluded.18 Although patients from in the pathogenesis of AD. both families mainly presented with classical symptoms Therefore, taking into consideration the previous of PD, the disease course was observed with variable age associations of CSMD1 and the HLA region with at onset, ranging from 49 to 72 years, and variable the risk of PD and the identification of disease- phenotypic heterogeneity, with 2 patients presenting segregating CSMD1 mutations in familial LOPD, with resting tremor, akinesia, and rigidity, and later we hypothesize that CSMD1 genetic variability on postural instability (patients B1 and A1), 1 patient might also contribute to PD pathogenesis through (patient B2) presenting withtremorofthehandsthat mechanisms implicated in immune-related synaptic aggravated with anxiety and RLS, and the youngest dysfunction. Although it will be important to address patient with severe bradykinesia affecting her daily activ- these pathologic functions, this study may highlight ities, RBD, and RLS, as well as atypical symptoms such the complement pathway as an important therapeutic as bilateral Babinski signs and ataxic gait (patient A2). target in PD as it has been suggested in AD. We do not know whether the patient B2, currently diagnosed with postural tremor, will manifest PD in NOTE ADDED IN PROOF the near future, but we believe that she might develop The authors have added a citation for an article that was published online after acceptance of this manuscript.34 PD at a more advanced age as her father did. Despite the advance age of some of the patients, only 1 patient AUTHOR CONTRIBUTIONS presented with dysexecutive mild cognitive impairment Dr. Martí-Massó and Dr. Paisán-Ruiz: study concept and design. (patient B1). However, we cannot discard that the other Dr. Ruiz-Martínez, Dr. Azcona, Dr. Bergareche, and Dr. Paisán-Ruiz: acquisition of data. Dr. Ruiz-Martínez, Dr. Azcona, Dr. Bergareche, and 2 reported mutation carriers, who are still relatively Dr. Paisán-Ruiz: analysis and interpretation of data. Dr. Ruiz-Martínez, young, might develop some kind of cognitive dysfunc- Dr. Bergareche, Dr. Martí-Massó, and Dr. Paisán-Ruiz: critical revision tion at an advanced age. of the manuscript for important intellectual content. Dr. Martí-Massó Because of the same ethnicity and geographical region and Dr. Paisán-Ruiz: study supervision. of the examined patients, we first searched for novel and ACKNOWLEDGMENT rare genetic variations shared between all affected individ- The authors thank the families, their relatives, and other participants for uals, and although we did not find any mutation to be their contribution to this study. They also thank Vladimir Makarov for shared by all examined patients, we found the same gene, his contribution to data processing.

Neurology: Genetics 5 STUDY FUNDING 17. Marti-Masso JF, Bergareche A, Makarov V, et al. The This work was supported by the National Institute of Neurological Dis- ACMSD gene, involved in tryptophan metabolism, is orders and Stroke of the NIH (NINDS; R01NS079388; C.P.-R.). mutated in a family with cortical myoclonus, epilepsy, and parkinsonism. J Mol Med (Berl) 2013;91:1399–1406. DISCLOSURE 18. Gorostidi A, Marti-Masso JF, Bergareche A, Rodriguez- J. Ruiz-Martínez, L. Azcona, J. Bergareche, and J. Martí-Massó report no Oroz MC, Lopez de Munain A, Ruiz-Martinez J. Genetic disclosures. C. Paisán-Ruiz serves on the editorial board of BMC Neurology mutation analysis of Parkinson’s disease patients using and receives funding from the NINDS (R01NS079388) and the Amer- multigene next-generation sequencing panels. Mol Diagn ican Parkinson’s Disease Association (APDA). Go to Neurology.org/ng – for full disclosure forms. Ther 2016;20:481 491. 19. Gelb DJ, Oliver E, Gilman S. Diagnostic criteria for Par- Received March 2, 2017. Accepted in final form June 16, 2017. kinson disease. Arch Neurol 1999;56:33–39. 20. Hughes AJ, Daniel SE, Kilford L, Lees AJ. Accuracy of REFERENCES clinical diagnosis of idiopathic Parkinson’s disease: a clini- 1. Halliday G, Lees A, Stern M. Milestones in Parkinson’s co-pathological study of 100 cases. J Neurol Neurosurg disease–clinical and pathologic features. Mov Disord Psychiatry 1992;55:181–184. 2011;26:1015–1021. 21. Ruiz-Martinez J, Krebs CE, Makarov V, Gorostidi A, 2. Lees AJ, Hardy J, Revesz T. Parkinson’s disease. Lancet Marti-Masso JF, Paisan-Ruiz C. GIGYF2 mutation in 2009;373:2055–2066. late-onset Parkinson’s disease with cognitive impairment. 3. Hardy J, Lewis P, Revesz T, Lees A, Paisan-Ruiz C. The J Hum Genet 2015;60:637–640. genetics of Parkinson’s syndromes: a critical review. Curr 22. DePristo MA, Banks E, Poplin R, et al. A framework for Opin Genet Dev 2009;19:254–265. variation discovery and genotyping using next-generation 4. Kalinderi K, Bostantjopoulou S, Fidani L. The genetic DNA sequencing data. Nat Genet 2011;43:491–498. background of Parkinson’s disease: current progress and 23. Li H, Durbin R. Fast and accurate short read alignment future prospects. Acta Neurol Scand 2016;134:314–326. with Burrows-Wheeler transform. Bioinformatics 2009; 5. Hamza TH, Chen H, Hill-Burns EM, et al. Genome-wide 25:1754–1760. gene-environment study identifies glutamate receptor gene 24. Makarov V, O’Grady T, Cai G, Lihm J, Buxbaum JD, Yoon GRIN2A as a Parkinson’s disease modifier gene via inter- S. AnnTools: a comprehensive and versatile annotation tool- action with coffee. PLoS Genet 2011;7:e1002237. kit for genomic variants. Bioinformatics 2011;28:724–725. 6. Paisan-Ruiz C, Jain S, Evans EW, et al. Cloning of the 25. Exome Variant Server. NHLBI Exome Sequencing Project gene containing mutations that cause PARK8-linked Par- (ESP). Seattle. Available at: http://evsgswashingtonedu/ kinson’s disease. Neuron 2004;44:595–600. EVS/. Accessed July 20, 2017. 7. Polymeropoulos MH, Lavedan C, Leroy E, et al. Mutation 26. Kircher M, Witten DM, Jain P, O’Roak BJ, Cooper GM, in the alpha-synuclein gene identified in families with Shendure J. A general framework for estimating the rela- Parkinson’s disease. Science 1997;276:2045–2047. tive pathogenicity of human genetic variants. Nat Genet 8. Zimprich A, Benet-Pages A, Struhal W, et al. A mutation 2014;46:310–315. in VPS35, encoding a subunit of the complex, 27. Nagase T, Kikuno R, Ohara O. Prediction of the coding causes late-onset Parkinson disease. Am J Hum Genet sequences of unidentified human genes. XXI. The com- 2011;89:168–175. plete sequences of 60 new cDNA clones from brain which 9. Nalls MA, Pankratz N, Lill CM, et al. Large-scale meta- code for large proteins. DNA Res 2001;8:179–187. analysis of genome-wide association data identifies six new 28. Hill-Burns EM, Wissemann WT, Hamza TH, Factor SA, risk loci for Parkinson’s disease. Nat Genet 2014;46:989–993. Zabetian CP, Payami H. Identification of a novel Parkin- 10. Paisan-Ruiz C. LRRK2 gene variation and its contribution son’s disease locus via stratified genome-wide association to Parkinson disease. Hum Mutat 2009;30:1153–1160. study. BMC Genomics 2014;15:118. 11. Paisan-Ruiz C, Lewis PA, Singleton AB. LRRK2: cause, 29. Kraus DM, Elliott GS, Chute H, et al. CSMD1 is a novel risk, and mechanism. J Parkinsons Dis 2013;3:85–103. multiple domain complement-regulatory protein highly 12. Hamza TH, Zabetian CP, Tenesa A, et al. Common genetic expressed in the central nervous system and epithelial tis- variation in the HLA region is associated with late-onset sues. J Immunol 2006;176:4419–4430. sporadic Parkinson’s disease. Nat Genet 2010;42:781–785. 30. Athanasiu L, Giddaluru S, Fernandes C, et al. A genetic 13. International Parkinson Disease Genomics C, Nalls MA, association study of CSMD1 and CSMD2 with cognitive Plagnol V, Hernandez DG, et al. Imputation of sequence function. Brain Behav Immun 2016;61:209–216. variants for identification of genetic risks for Parkinson’s 31. Havik B, Le Hellard S, Rietschel M, et al. The complement disease: a meta-analysis of genome-wide association stud- control-related genes CSMD1 and CSMD2 associate to ies. Lancet 2011;377:641–649. schizophrenia. Biol Psychiatry 2011;70:35–42. 14. Pihlstrom L, Axelsson G, Bjornara KA, et al. Supportive 32. Hong S, Beja-Glasser VF, Nfonoyim BM, et al. Comple- evidence for 11 loci from genome-wide association studies ment and microglia mediate early synapse loss in Alz- in Parkinson’s disease. Neurobiol Aging 2013;34:1708. heimer mouse models. Science 2016;352:712–716. e7–1708.e13. 33. Swaminathan S, Kim S, Shen L, et al. Genomic copy 15. Sidransky E, Nalls MA, Aasly JO, et al. Multicenter anal- number analysis in Alzheimer’s disease and mild cognitive ysis of glucocerebrosidase mutations in Parkinson’s disease. impairment: an ADNI Study. Int J Alzheimers Dis 2011; N Engl J Med 2009;361:1651–1661. 2011:729478. 16. Foo JN, Tan LC, Irwan ID, et al. Genome-wide associa- 34. Vilas D, Fernández-Santiago R, Sanchez-Rodriguez E, et al. A tion study of Parkinson’s disease in East Asians. Hum Mol novel p.Glu298Lys mutation in the ACMSD gene in sporadic Genet 2017;26:226–232. Parkinson’s disease. J Parkinsons Dis Epub 2017 Jun 30.

6 Neurology: Genetics Ultra-rare mutations in SRCAP segregate in Caribbean Hispanic families with Alzheimer disease

Badri N. Vardarajan, PhD ABSTRACT Giuseppe Tosto, MD, Objective: To identify rare coding variants segregating with late-onset Alzheimer disease (LOAD) PhD in Caribbean Hispanic families. Roger Lefort, PhD Methods: Whole-exome sequencing (WES) was completed in 110 individuals from 31 Caribbean Lei Yu, PhD Hispanic families without APOE e4 homozygous carriers. Rare coding mutations segregating in David A. Bennett, MD families were subsequently genotyped in additional families and in an independent cohort of Carib- Philip L. De Jager, MD, bean Hispanic patients and controls. SRCAP messenger RNA (mRNA) expression was assessed in PhD whole blood from mutation carriers with LOAD, noncarriers with LOAD, and healthy elderly controls, Sandra Barral, PhD and also from autopsied brains in 2 clinical neuropathologic cohort studies of aging and dementia. Dolly Reyes-Dumeyer, BS Results: Ten ultra-rare missense mutations in the Snf2-related CREBBP, activator protein Peter L. Nagy, MD, PhD (SRCAP), were found in 12 unrelated families. Compared with the frequency in Caribbean His- Joseph H. Lee, PhD panic controls and the Latino population in the Exome Aggregation Consortium, the frequency of Rong Cheng, PhD SRCAP mutations among Caribbean Hispanic patients with LOAD was significantly enriched (p 5 Martin Medrano, MD 1.19e-16). mRNA expression of SRCAP in whole blood was significantly lower in mutation car- Rafael Lantigua, MD riers with LOAD, while the expression in whole blood and in the brain was significantly higher in Ekaterina Rogaeva, PhD nonmutation carriers with LOAD. Brain expression also correlated with clinical and neuropatho- Peter St George-Hyslop, logic endophenotypes. MD Richard Mayeux, MD Conclusions: WES in Caribbean Hispanic families with LOAD revealed ultra-rare missense muta- tions in SRCAP, a gene expressed in the brain and mutated in Floating-Harbor syndrome. SRCAP is a potent of the CREB-binding protein and a regulator of DNA damage response Correspondence to involving ATP-dependent chromatin remodeling. We hypothesize that increased expression in Dr. Mayeux: LOAD suggests a compensatory mechanism altered in mutation carriers. Neurol Genet 2017;3: [email protected] e178; doi: 10.1212/NXG.0000000000000178

GLOSSARY CBP 5 CREB-binding protein; CI 5 confidence interval; DLPFC 5 dorsolateral prefrontal cortex; ExAC 5 Exome Aggregation Consortium; FPKM 5 fragments per kilobase per million fragments mapped; GEE 5 generalized estimating equation; IRB 5 institu- tional review board; LOAD 5 late-onset Alzheimer disease; MCI 5 mild cognitive impairment; mRNA 5 messenger RNA; OR 5 odds ratio; QC 5 quality control; SNP 5 single nucleotide polymorphism; SNV 5 single nucleotide variant; WES 5 whole-exome sequencing.

Progress has been made in understanding the genetics of late-onset Alzheimer disease (LOAD),1,2 but gaps in its genetic influence still need investigation. Common variants play a role in disease risk, but functionally important rare or ultra-rare variants may help to explain the remaining heritability2,3 undetected by genome-wide association studies. Sequencing of large families multiply affected by LOAD increases the ability to detect novel variants conferring risk. Supplemental data at Neurology.org/ng From the Taub Institute for Research on Alzheimer’s Disease and the Aging Brain (B.N.V., G.T., R. Lefort, P.L.D.J., S.B., D.R.-D., J.H.L., R.C., R. Lantigua, R.M.); Gertrude H. Sergievsky Center (B.N.V., G.T., S.B., D.R.-D., J.H.L., R.C., R.M.); Department of Neurology (P.L.D.J., S.B., R. M.), Department of Psychiatry (R.M.), Department of Systems Biology (B.N.V.), Department of Medicine (R. Lantigua), and Department of Pathology and Cell Biology (R. Lefort, P.L.N.), College of Physicians and Surgeons, Columbia University, New York Presbyterian Hospital; Department of Epidemiology (J.H.L., R.M.), School of Public Health, Columbia University, New York; Tanz Centre for Research in Neurodegenerative Diseases (E.R., P.S.G.-H.) and Department of Medicine (E.R., P.S.G.-H.), University of Toronto, Krembil Discovery Tower, ON, Canada; Department of Clinical Neurosciences (P.S.G.-H.), Cambridge Institute for Medical Research, University of Cambridge, UK; Rush Alzheimer’s Disease Center (L.Y., D.A.B.), Rush University Medical Center, Chicago, IL; Program in Medical and Population Genetics (P.L.D.J.), Broad Institute, Cambridge, MA; and School of Medicine (M.M.), Mother and Teacher Pontifical Catholic University, Santiago, Dominican Republic. Funding information and disclosures are provided at the end of the article. Go to Neurology.org/ng for full disclosure forms. The Article Processing Charge was funded by the NIH. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Neurology.org/ng Copyright © 2017 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 The frequency of LOAD among Caribbean multiplexed in batches of 12 samples with index “bar- Hispanic multiplex families from the Domin- code” primers. These were sequenced using the Illu- ican Republic was 5 times higher than ex- mina Genome Analyzer IIx, HiSeq 2000, and MiSeq pected for similarly aged individuals in platforms (illumina.com) as paired-end reads over 82–307 cycles. Demultiplexing by barcode retrieved a non-Hispanic white population from the individual samples from sequencing pools. We ob- United States,4 and inbreeding was a signifi- tained a high coverage across the samples at an average cant predictor of LOAD in this population depth of .603 per sample. after adjusting for APOE-e4 genotype, an es- tablished genetic risk factor.5 Follow-up genotyping. Putative variants were confirmed and population frequencies estimated by genotyping To identify novel variants associated with the discovery samples, additional family members, the risk of LOAD, we conducted whole- and unrelated controls of the same ancestry (table exome sequencing in 31 Caribbean Hispanic e-1). Allele frequencies of novel variants were estimated families (table e-1 at Neurology.org/ng) with from 1,949 unrelated patients and 318 healthy elderly 4 or more affected individuals, no mutations controls similar in age and ancestry.6 Genotypes were in known AD genes, specifically PSEN1, generated using the KASP genotyping technology, PSEN2,orAPP, and no APOE e4 homozy- which uses allele-specific PCR for accurate calling of gotes. For each family, we sequenced at least 2 single nucleotide polymorphisms (SNPs) and Indels.18 affected and 1 unaffected member aged 65 Analytical methods. Burrows-Wheeler Aligner19 was years or older. used to align sequence reads to the reference genome build 37. Sequencing data quality control (QC) was METHODS Sample selection. Families were recruited as performed using the Genome Analysis Toolkit a part of a 15-year family-based study with institutional review (GATK),20 followed by variant calling using the 6 board (IRB) approval based in the Dominican Republic. Thirty- UnifiedGenotyper and VariantRecalibrator mod- one families (98 affected and 12 unaffected individuals) were selected for sequencing (mean age at onset was 74.8 wa8.3 years, ules. Variants that passed QC were annotated by 21 and 63.1% were women) (table e-1, a–c). All family members had ANNOVAR that included functional prediction standard neuropsychological tests and neurologic examinations to by SIFT22 and PolyPhen.23 verify their clinical status and for diagnoses based on NINCDS- ADRDA criteria.7,8 STATISTICAL METHODS Association tests. Variants that were validated by follow-up genotyping were Postmortem human brain samples. Data were obtained from tested for association with LOAD using generalized 2 clinical neuropathologic cohort studies: the ROS9 and the MAP.10 The IRB of Rush University Medical Center previously estimating equations (GEEs), which accounts for the approved both studies. Clinical evaluations were used to deter- familial correlation. We adjusted for age and sex using mine NINCDS-ADRDA7,8 criteria for dementia annually.11–13 At data from the families and unrelated healthy controls. death, a clinical diagnosis opinion was provided by a neurologist.14 We used GEE to conduct single variant and burden Neuropathologic evaluations included neuritic plaques, diffuse tests. plaques, and neurofibrillary tangles in 5 cortical regions, scaled For joint analyses of multiple variants, we summed and averaged to obtain a composite score.15 Participants who met intermediate or high likelihood were rendered pathologic diagno- the number of rare variant alleles found in all individ- sis of LOAD.16,17 uals and tested association using GEE, adjusting for familial correlation, age and sex, and APOE e4. The WHOLE-EXOME SEQUENCING Sample preparation. p value threshold required to define statistical signif- Qiagen’s Gentra Puregene and FlexiGene kits were icance using a rare variant burden analysis for 20,000 used to extract high-molecular-weight DNA from genes would be 2.5 3 1026. However, this would be fresh or frozen (,280°C) samples. DNA from saliva difficult to achieve by assessing rare or ultra-rare var- was isolated using prepIT.L2P (DNA Genoteck Inc., iants in a data set of this size. Thus, to determine Ottawa, ON, Canada). Cell lines from lymphocytes whether or not the variants discovered were enriched (in 13 probands) were used when high-quality blood in families with AD in subsequent analyses, we DNA was not available. The concentration of DNA used data from the Exome Aggregation Consortium was determined using a NanoDrop spectrophotometer. (ExAC) (exac.broadinstitute.org) combined with the Caribbean Hispanic controls. Sequencing. The Illumina TruSeq DNA preparation kit was used to prepare and index genomic DNA Exome Aggregation Consortium. The ExAC database24 libraries. Custom oligonucleotide baits in the TruSeq contains whole-exome data from 60,706 unrelated Exome Enrichment kit were used to capture coding adults sequenced as part of various disease-specific regions and splice sites and amplified according to (excluding AD) and population genetic studies from the Illumina protocol. The DNA samples were 6 ethnic groups. For disease association analyses, we

2 Neurology: Genetics with both unaffected unrelated controls and ExAC Figure 1 Workflow of the experiment and yield at each step combined using a Fisher exact test.

Gene expression in the postmortem human brain. RNA- Seq data came from the 541 ROS-MAP postmortem human brain samples (average age at death was 88.4 NA6.7 years; 63.0% were women and 97.4 self- identified Caucasian ancestry). RNA was extracted from the gray matter of frozen dorsolateral prefrontal cortex tissue using the Qiagen miRNeasy mini kit (Cat. no. 217004) and the RNase-free DNase Set (Cat. no. 79254) and quantified using NanoDrop. RNA-Seq library was prepared on the Broad In- stitute’s Genomics Platform using the strand-specific dUTP method25 with poly-A selection.26 Sequencing was performed on the Illumina HiSeq with 101bp paired-end reads and minimal coverage of 50M reads. Reads were aligned to the reference genome (GRCh37/hg19) using Bowtie and gene expression, measured as fragments per kilobase per million frag- ments mapped (FPKM), and estimated using RSEM software. Quantile normalization was applied to the FPKM calls, with batch effect removed using Com- bat.27 Logarithm base 2 transformation was applied to the gene expression level prior to the analyses. The Student t test compared brain expression levels between LOAD and controls. Regression models were used to examine the association of gene expression with LOAD and with level of cognition proximate to death. Regression models were used to examine the relation between gene expression and a postmortem diagnosis of LOAD or with LOAD pathology. All models were adjusted for age at death, sex, education, postmortem interval RNA integrity number, and APOE e4 status. Analyses were performed using SAS software, version 9.3, of the SAS(R) system for Linux.

RESULTS Whole-exome sequencing. Variant calling, recalibration, and application of QC filters resulted in 290,623 single nucleotide variants (SNVs) and in- dels called across the 110 individuals from the 31 CH 5 Caribbean Hispanic; ExAC 5 Exome Aggregation Consortium; GEE 5 generalized estimating equation; OR 5 odds ratio; QC 5 quality control; SNV 5 single nucleotide variant; families (figure 1). The mean depth of sequence cov- WES 5 whole-exome sequencing. erage was over 60-fold across individuals. We priori- tized damaging SNVs by SIFT22 or by PolyPhen28 compared the allele frequencies of suspected variants and indels in 78 genes segregating with LOAD in in the healthy controls from the Caribbean Hispanic families (figure e-1, A and B), which were confirmed cohort to the Latino, Caucasian, and African sub- by genotyping in the discovery families, additional populations of the ExAC database using a Fisher exact family members (265 affected and 61 unaffected in- test to avoid differences based on ancestry. Sub- dividuals), and a set of 318 unrelated controls of sequent disease-associated analyses were only con- similar age and ancestry. Using GEE adjusted for age, ducted when there were no statistically significant sex, and APOE genotype, we found 22 (28%) of the differences in allele frequency between the Caribbean 78 selected variants (in 22 genes) significantly asso- Hispanic controls and the ExAC subpopulations. To ciated with LOAD (p 5 6.4e-03 corrected for mul- confirm associations of genetic variants with LOAD, tiple testing) (table e-2). Only 4 of these 22 genes we compared the variant allele frequencies by select- were known to be expressed in brain-ADCY6, CIT, ing 1 patient from each Caribbean Hispanic family SRCAP, and SVOPL. The variants in SRCAP,

Neurology: Genetics 3 Table 1 Missense variants in SRCAP found in WES of Caribbean Hispanic families

Genotyping in discovery Discovery WESb familiesc Validation genotyping cohort ExAC frequencies

Frequency Frequency AFF AFF in familial in elderly Global Chr Start Gene Ref Obs AA changea samplesd FAMe samplesd FAMe load casesf unaffected (>65 y)f Latino frequency

16 30731484 SRCAP G A R940Q 1 1 1 1 4.30e-04 2.92e-03 8.64e-05 8.24e-06

16 30732143 SRCAP T C S1033P 5 3 12 4 8.89e-03 5.85e-03 7.45e-03 7.41e-04

16 30732558 SRCAP C A T1101K 3 1 9 1 7.63e-03 2.92e-03 2.59e-03 6.00e-03

16 30732600 SRCAP G A R1115H 4 1 6 1 1.28e-03 0 0 8.24e-06

16 30732644 SRCAP C T P1130S 1 1 2 2 4.71e-03 7.35e-03 2.94e-03 2.57e-03

16 30734934 SRCAP A G I1397V 3 1 4 1 1.93e-03 0 0.00e100 3.30e-04

16 30748932 SRCAP C G S2524C 2 1 1 1 1.27e-02 2.96e-03 1.30e-03 2.04e-03

16 30749583 SRCAP C G P2741R 3 1 6 2 1.28e-03 0 ——

16 30749652 SRCAP T G V2764G 1 1 2 1 6.66e-04 0 ——

16 30750116 SRCAP C T L2919F 1 1 5 1 2.58e-03 0 6.05e-04 1.35e-03

Abbreviations: ExAC 5 Exome Aggregation Consortium; WES 5 whole-exome sequencing. a AA change: amino acid change caused by the mutation. b Discovery WES: number of affected individuals and families identified as carrying an SRCAP variant in the WES experiment in 110 individuals (31 families). c Genotyping in discovery families: number of affected individuals and families identified as carrying an SRCAP variant in the validation genotyping in all the members of the 31 discovery families. d AFF samples: number of familial-affected samples carrying the SRCAP variant. e FAM: number of families carrying the SRCAP variant. f Frequency in the affected and unaffected individuals included familial and unrelated cases and controls.

ADCY6, and SVOPL were absent in unaffected of European, Latino, and African ancestry. p.R1115H Caribbean Hispanic individuals, but only SRCAP met was also found within 0.35 MB of a linkage peak (log- criteria for further analyses by having multiple, arithm (base 10) of odds 5 1.8). A single patient was putatively damaging variants segregating in more than homozygous for an SRCAP p.S2524C mutation and 1 of the 31 families. had an earlier age at onset compared with the average We observed 10 ultra-rare missense mutations, age in the cohort (63 vs 73.8 years). In family 1755 defined as having an allele frequency #0.5%, in (figure 2), there were 2 compound heterozygous carriers SRCAP in 12 families (table 1). Seven of these muta- (p.L2919F and p.S1033P) with earlier ages at onset tions were predicted to be damaging by either Poly- (55 and 59 years) compared with LOAD patients with- Phen or SIFT and had a Combined Annotation out SRCAP mutations (p 5 0.023) or patients who had Dependent Depletion29 .10. Twenty-four (24.4%) asingleSRCAP mutation (p 5 0.018). Of interest, no of the 98 affected individuals sequenced within the 31 unaffected family member or unrelated control was families were carriers of SRCAP mutations. This was a homozygous or compound heterozygous carrier. significantly higher than the expected frequency A joint burden analysis within the 31 discovery based on data from ExAC database24 (1.79 expected families using GEE adjusted for familial correlation, mutations; p 5 1.38e-05 using a binomial test). age, sex, and APOE genotype found a 6-fold increase All 10 SRCAP mutations were genotyped in addi- in the risk of LOAD (odds ratio [OR] 5 5.94; 95% tional members of the 31 families and segregated with confidence interval [CI] 1.59–22.27, p 5 8.2e-03). affection status but with incomplete penetrance (fig- Adding the additional 47 families resulted in a 2.5- ures e-3, A–D and e-4). Of the 265 patients with fold risk of LOAD in families (OR 5 2.54; 95% CI LOAD in the 31 families, 47 patients (17.67%) car- 1.03–6.21, p 5 0.04). We then tested the joint bur- ried at least 1 of the 10 missense mutations in SRCAP den of the 10 SRCAP variants in all the genotyped compared with 2 of the 61 (3.2%) unaffected indi- individuals in both sets of families, 1,949 unrelated viduals from these families. Variant p.R1115H LOAD cases and 318 unrelated elderly controls (table segregated with LOAD in the largest sibship in an e-1b). The risk of LOAD associated with SRCAP age-dependent manner (supplementary methods, remained increased (OR 5 1.92; 95% CI 1.13– figure e-2, tables e-3, a and b) and was absent in 3.28, p 5 0.016) after including APOE e4 in the Caribbean Hispanic controls and ExAC individuals model (OR 5 1.82; 95% CI 1.04–3.19, p 5 0.036).

4 Neurology: Genetics Figure 2 Segregation pattern of missense SRCAP mutations in pedigree 1755

Patients 99 and 9 are compound heterozygotes with ages at onset of 59 and 55 years, respectively. Four other affected siblings who were heterozygous (3 p.L2919F and 1 p.S1033P) for SRCAP mutation had ages at onset of 58, 58, 72, and 76 years, respectively. Import ID: Subject ID; AAO_Affected_Else_A- geLast_Seen: Age at onset of LOAD or age at last examination for healthy individuals. LOAD 5 late-onset Alzheimer disease.

To reach genome-wide significance (p 5 2.5e-06) and in the top 0.15 and 0.3 percentiles of genes in a gene-based test, assuming an OR of 2.0 and 5% ranked by the Residual Variation Intolerance Score causal SNPs and b 5 0.8, a gene of average size, such (RVIS) using the ExAC and Exome Sequencing Pro- as SRCAP, requires an extremely large sample size ject data sets, respectively. (N 5 42,000). Therefore, we compared the combined Replication in the AD Neuroimaging Initiative data set. genotyped data mentioned above with data from the Using data from the Alzheimer Disease Neuroimag- ExAC database.24 Using a Fisher exact test, we found ing Initiative (ADNI) data set,31 we repeated the anal- no statistically significant differences in the allele fre- ysis of the exonic regions of SRCAP. The analyses quencies of the SRCAP variants in healthy Caribbean included 213 patients with LOAD, 304 individuals Hispanic controls compared with the Latino, African with mild cognitive impairment (MCI), and 214 American, and white, non-Hispanics in the ExAC healthy controls of white, non-Hispanic ancestry. database. We then compared the frequencies of We found a higher frequency of nonsynonymous SRCAP variants of LOAD patients with Caribbean mutations in LOAD and MCI cases (10.4%) vs Hispanic controls alone or with ExAC Latino controls, controls (7.4%) representing a 40% enrichment. ExAC African American controls, or ExAC white, Because of the sample size, we once again compared non-Hispanic controls. Using one mutation carrier the frequency of nonsynonymous mutations in per family, we compared the number of variant al- LOAD with the ExAC database. Of interest, there leles in the LOAD patients with the number of al- was a higher frequency in LOAD vs ExAC controls leles in the ExAC database. Two variants, p.P2741R (OR 5 1.78, p 5 2.3e-04). and p.V2764G, were not found in the ExAC data- base (table 1). We used the total number of alleles SRCAP gene expression. Among the ROS-MAP co- genotyped or sequenced for the 10 mutations as the horts, 218 of 541 (40.3%) individuals were diagnosed denominator for the Fisher exact test. The enrich- with LOAD at death, and compared with non- ment of SRCAP variants in the Caribbean Hispanic demented persons, SRCAP expression was signifi- patients with LOAD was highly significant at p 5 cantly higher (OR 5 4.9 [2.1–11.6], p 5 0.0002) 1.19e-16. Similarly, the enrichment was statistically (figure 3A). Individuals with higher SRCAP expres- significant for white, non-Hispanics, and the African sion also had lower cognitive scores prior to death American ethnic groups in the ExAC database (table (b 520.96, p , 0.0001) (figure 3B). Similar as- e-4; p 5 1.51e-21 for white, non-Hispanics; p 5 3e- sociations were observed in episodic, semantic and 04 for African Americans). Five of the 10 SRCAP working memory, and perceptual speed (table e-5). mutations were not found in the Caribbean His- Using the clinical diagnosis, higher SRCAP expression panic controls, and 4 of these were absent in the was associated with greater likelihood of meeting the ExAC Latino data (table 1). Restricting analyses to NIA-Reagan criteria for AD (OR 5 4.0 [1.7–9.2], these 5 SNPs only, the enrichment remained signif- p 5 0.0015) (figure 3C). Individuals with higher icant p 5 3.55e-07. SRCAP expression had a greater burden of pathology SRCAP is highly intolerant toward loss of function (b 5 0.3, p , 0.0001) (figure 3D). We also accessed and missense mutations.24,30 SRCAP’s intolerance whole-genome sequencing data from 63 participants score (2.23) in the ExAC ranks in the 18th percentile with brain expression. Two of 69 individuals

Neurology: Genetics 5 Figure 3 Association of expression level of SRCAP in autopsy brain tissue with Figure 4 Comparison of SRCAP mRNA (A) AD clinical diagnoses, (B) cognitive performance, (C) AD pathologic expression in affected mutation diagnoses, and (D) global AD pathology carriers, affected noncarriers, and controls

AD 5 Alzheimer disease. mRNA 5 messenger RNA. sequenced carried 1 of the 10 SRCAP missense mu- tations (S1033P and L2919F), and both were diag- Reduced mRNA expression of SRCAP was found in nosed with mild cognitive impairment. whole blood of mutation carriers compared with Mutation-specific gene expression. Relative messenger noncarriers and controls. Higher SRCAP expression RNA (mRNA) expression of SRCAP (supplementary was expressed in noncarriers in both blood of methods) in whole blood was measured in 5 mutation Caribbean Hispanic patients with LOAD and brains carriers with LOAD, 12 individuals without muta- of Caucasian patients with LOAD, irrespective of tions with LOAD, and 12 elderly controls without differences in ancestry. Pathologic evidence indicates dementia, all unrelated. After normalization against that SRCAP expression is altered in LOAD with or GAPDH, SRCAP mRNA expression was significantly without mutations, correlating with characteristic lower in the mutation carriers with LOAD than in the clinical manifestations. Taken together, the com- other 2 groups (figure 4) (LOAD carriers 0.74, bined data implicate a putative role for SRCAP in LOAD noncarriers 1.128, and controls 0.99, p 5 LOAD. – 0.004). We compared the levels of exogenous mature Snf2-related cyclic AMP-responsive element wild-type SRCAP mRNA in HEK293 cells to the one binding protein (SRCAP) binds to the CREB-binding carrying an ultra-rare mutation (p.P2741R; minor protein (CBP), which influences the transcription of allele frequency 5 0.00128 in Caribbean Hispanic CREB. CREB and the related transcription factors cases and absent in ExAC). mRNA expression of are involved in memory retention32 and consolidation mutant SRCAP (p.P2741R) was downregulated by by hippocampal neurogenesis.33 In AD-transgenic 40% relative to wild-type SRCAP. mice and in humans with LOAD, CREB and CBP are decreased or disrupted by the accumulation DISCUSSION A statistically significant enrichment of amyloid-b.34–36 Reduced levels have also been of SRCAP ultra-rare mutations was found among observed for pCREB, CBP, and related CREB co- patients within Caribbean Hispanic families multiply activators p300 and cAMP-dependent protein kinase affected by LOAD. These mutations were observed to in the human AD brain and blood compared with segregate imperfectly with disease consistent with the elderly controls.37 We also observed reduced mRNA presence of phenocopies in large LOAD families. expression of SRCAP in mutation carriers compared

6 Neurology: Genetics with healthy elderly controls and unrelated, noncar- (ADRC), and the Department of Defense. R. Lefort has received research ’ riers with LOAD, and a reduction in SRCAP expres- support from Columbia University Alzheimer s Disease Research Center (ADRC) and the Alzheimer’s Association. L. Yu has received research sion in cell lines transfected with SRCAP mutations. support from NIH/NIA. D.A. Bennett has served on the scientific advi- We reported an association between episodic memory sory boards of Vigorous Minds, Takeda Pharmaceuticals, and AbbVie; and SNP rs2526690 in CBP as well as nominal sig- has served on the editorial boards of Neurology, Current Alzheimer Research, and Neuroepidemiology; and has received research support from nificance for common variants in CREB1 and NIH. P.L. De Jager has served on the scientific advisory boards of TEVA 38 RBAP48 in the CREB pathway. Neuroscience, Sanofi/Genzyme, and Celgene; has received speaker hon- SRCAP regulates the CREB pathway by catalyzing oraria from Biogen Idec, Source Healthcare Analytics, Pfizer Inc., and H2A.Z into chromatin, which is required for gene TEVA; has served on the editorial boards of the Journal of Neuroimmu- nology, Neuroepigenetics, and Multiple Sclerosis; and has received research expression and interactions with co-activators of CREB support from Biogen, Eisai, UCB, Pfizer, Sanofi/Genzyme, NIH, and the 39,40 such as CBP. In the absence of mutations, the National MS Society. S. Barral and D. Reyes-Dumeyer report no disclo- increased mRNA expression of SRCAP in postmortem sures. P.L. Nagy has been an employee of MNG Laboratories. J.H. Lee LOAD brain suggests a compensatory response to has received research support from NIH/NIA, NIH/NCATS, and b BrightFocus Foundation. R. Cheng, M. Medrano, and R. Lantigue promote CREB activation in the presence of A accu- report no disclosures. E. Rogaeva has received research support from mulation. Paradoxically, in persons with SRCAP mu- the Ontario Research Fund and the W. Garfield Weston Foundation. tations, disruption of the binding of SRCAP to CBP P. St George-Hyslop has received research support from the Canadian Institute of Health Research, Medical Research Council, and Wellcome would decrease CREB-mediated transcription in brain Trust. R. Mayeux has received research support from NIH. Go to over the lifetime. CREB may have neuroprotective Neurology.org/ng for full disclosure forms. qualities against Ab toxicity, and persistent downregu- lation of CREB resulting from decreased SRCAP could Received February 23, 2017. Accepted in final form June 29, 2017. promote neurodegeneration.41 This suggests that REFERENCES SRCAP has a regulatory role in the CREB pathway 1. Lambert JC, Ibrahim-Verbaas CA, Harold D, et al. Meta- in LOAD, regardless of the presence or absence of analysis of 74,046 individuals identifies 11 new suscepti- SRCAP mutations. Ample evidence indicates disrup- bility loci for Alzheimer’s disease. Nat Genet 2013;45: tion of CREB transcription by Ab, but the effects of 1452–1458. these ultra-rare mutations on protein function and the 2. Jonsson T, Stefansson H, Steinberg S, et al. Variant of ’ LOAD phenotype will need to be fully elucidated. TREM2 associated with the risk of Alzheimer s disease. N Engl J Med 2013;368:107–116. Truncating mutations in SRCAP cause Floating- 3. Guerreiro R, Wojtas A, Bras J, et al. TREM2 variants in 42 Harbor syndrome, a childhood disorder characterized Alzheimer’s disease. N Engl J Med 2013;368:117–127. by short stature, delayed speech, and facial abnor- 4. Vardarajan BN, Faber KM, Bird TD, et al. Age-specific malities. The majority of SRCAP mutations reduced incidence rates for dementia and Alzheimer disease in expression and create a nonfunctioning protein. NIA-LOAD/NCRAD and EFIGA families: National Mutations in CBP, the substrate of SRCAP activation, Institute on Aging Genetics Initiative for Late-Onset Alz- heimer Disease/National Cell Repository for Alzheimer cause a developmental disorder similar to Floating- Disease (NIA-LOAD/NCRAD) and Estudio Familiar de Harbor syndrome and Rubinstein-Taybi syndrome. Influencia Genetica en Alzheimer (EFIGA). JAMA Neurol This study does have limitations. We found a mod- 2014;71:315–323. est increase in the risk of LOAD among carriers of these 5. Vardarajan BN, Schaid DJ, Reitz C, et al. Inbreeding ultra-rare SRCAP mutations in Caribbean Hispanic among Caribbean Hispanics from the Dominican Repub- families, which makes replication difficult. Validation lic and its effects on risk of Alzheimer disease. Genet Med – and replication efforts in other ethnic backgrounds 2015;17:639 643. 6. Lee JH, Cheng R, Barral S, et al. Identification of novel loci would be essential to generalize the findings. for Alzheimer disease and replication of CLU, PICALM, and BIN1 in Caribbean Hispanic individuals. Arch Neurol AUTHOR CONTRIBUTIONS 2011;68:320–328. Conception and design of the study: B.N.V., G.T., S.B., R. Lefort, 7. McKhann G, Drachman D, Folstein M, Katzman R, Price D.A.B., P.L.D.J., E.R., P.S.G.-H., and R.M. Acquisition and analysis D, Stadlan EM. Clinical diagnosis of Alzheimer’s disease: of data: B.N.V., R. Lefort, P.L.N., D.R.-D., J.H.L., R.C., L.Y., D.A.B., report of the NINCDS-ADRDA Work group under the P.L.D.J., M.M., R. Lantigua, and R.M. Drafting the manuscript or figures: B.N.V., G.T., E.R., P.S.G.-H., and R.M. auspices of Department of Health and Human Services Task Force on Alzheimer’s disease. Neurology 1984;34:939–944. 8. Bennett DA, Schneider JA, Aggarwal NT, et al. Decision STUDY FUNDING rules guiding the clinical diagnosis of Alzheimer’s disease Supported by grants from the National Institute on Aging and the NIH, in two community-based cohort studies compared to stan- RF1AG015473, R01AG037212, P50AG008702 (R.M.), P30AG10161, RF1AG1518, R01AG17917, R01AG36936, and U01AG46152 (D.A.B. dard practice in a clinic-based cohort study. Neuroepi- – and P.L.D.J.). demiology 2006;27:169 176. 9. Bennett DA, Schneider JA, Arvanitakis Z, Wilson RS. DISCLOSURE Overview and findings from the Religious orders study. – B.N. Vardarajan has served on the scientific advisory board of the Im- Curr Alzheimer Res 2012;9:628 645. muneering Corporation. G. Tosto has received research support from 10. Bennett DA, Schneider JA, Buchman AS, Barnes LL, NIH/NIA, Columbia University Alzheimer’s Disease Research Center Boyle PA, Wilson RS. Overview and findings from the

Neurology: Genetics 7 rush memory and aging project. Curr Alzheimer Res 2012; 28. Adzhubei I, Jordan DM, Sunyaev SR. Predicting func- 9:646–663. tional effect of human missense mutations using Poly- 11. Wilson RS, Beckett LA, Barnes LL, et al. Individual differ- Phen-2. Curr Protoc Hum Genet 2013;Chap 7. Unit7.20. ences in rates of change in cognitive abilities of older 29. Kircher M, Witten DM, Jain P, O’Roak BJ, Cooper GM, persons. Psychol Aging 2002;17:179–193. Shendure J. A general framework for estimating the rela- 12. Wilson RS, Boyle PA, Yu L, Segawa E, Sytsma J, Bennett tive pathogenicity of human genetic variants. Nat Genet DA. Conscientiousness, dementia related pathology, and 2014;46:310–315. trajectories of cognitive aging. Psychol Aging 2015;30: 30. Petrovski S, Wang Q, Heinzen EL, Allen AS, Goldstein 74–82. DB. Genic intolerance to functional variation and the 13. Wilson R, Barnes L, Bennett D. Assessment of lifetime interpretation of personal genomes. PLoS Genet 2013;9: participation in cognitively stimulating activities. J Clin e1003709. Exp Neuropsychol 2003;25:634–642. 31. Saykin AJ, Shen L, Foroud TM, et al. Alzheimer’s Disease 14. Schneider JA, Arvanitakis Z, Bang W, Bennett DA. Mixed Neuroimaging Initiative biomarkers as quantitative pheno- brain pathologies account for most dementia cases in types: genetics core aims, progress, and plans. Alzheimers community-dwelling older persons. Neurology 2007;69: Dement 2010;6:265–273. 2197–2204. 32. Lakhina V, Arey RN, Kaletsky R, et al. Genome-wide 15. Bennett DA, Wilson RS, Schneider JA, et al. Apolipopro- functional analysis of CREB/long-term memory-depen- tein E epsilon4 allele, AD pathology, and the clinical dent transcription reveals distinct basal and memory gene expression of Alzheimer’s disease. Neurology 2003;60: expression programs. Neuron 2015;85:330–345. 246–252. 33. Ortega-Martinez S. A new perspective on the role of the 16. Schneider JA, Arvanitakis Z, Leurgans SE, Bennett DA. CREB family of transcription factors in memory consoli- The neuropathology of probable Alzheimer disease and dation via adult hippocampal neurogenesis. Front Mol mild cognitive impairment. Ann Neurol 2009;66: Neurosci 2015;8:46. 200–208. 34. Duclot F, Meffre J, Jacquet C, Gongora C, Maurice T. 17. Hyman BT, Trojanowski JQ. Consensus recommenda- Mice knock out for the histone acetyltransferase tions for the postmortem diagnosis of Alzheimer disease p300/CREB binding protein-associated factor develop from the National Institute on Aging and the Reagan a resistance to amyloid toxicity. Neuroscience 2010;167: Institute Working Group on diagnostic criteria for the 850–863. neuropathological assessment of Alzheimer disease. 35. Pugazhenthi S, Wang M, Pham S, Sze CI, Eckman CB. J Neuropathol Exp Neurol 1997;56:1095–1097. Downregulation of CREB expression in Alzheimer’s brain 18. Abrams KR, Gillies CL, Lambert PC. Meta-analysis of and in Abeta-treated rat hippocampal neurons. Mol Neu- heterogeneously reported trials assessing change from base- rodegener 2011;6:60. line. Stat Med 2005;24:3823–3844. 36. Satoh J, Tabunoki H, Arima K. Molecular network anal- 19. Li H, Durbin R. Fast and accurate short read alignment ysis suggests aberrant CREB-mediated gene regulation in with Burrows-Wheeler transform. Bioinformatics 2009; the Alzheimer disease hippocampus. Dis Markers 2009; 25:1754–1760. 27:239–252. 20. McKenna A, Hanna M, Banks E, et al. The Genome 37. Bartolotti N, Bennett DA, Lazarov O. Reduced pCREB in Analysis Toolkit: a MapReduce framework for analyzing Alzheimer’s disease prefrontal cortex is reflected in periph- next-generation DNA sequencing data. Genome Res eral blood mononuclear cells. Mol Psychiatry 2016;21: 2010;20:1297–1303. 1158–1166. 21. Wang K, Li M, Hakonarson H. ANNOVAR: functional 38. Barral S, Reitz C, Small SA, Mayeux R. Genetic variants in annotation of genetic variants from high-throughput a “cAMP element binding protein” (CREB)-dependent sequencing data. Nucleic Acids Res 2010;38:e164. histone acetylation pathway influence memory perfor- 22. Kumar P, Henikoff S, Ng PC. Predicting the effects of mance in cognitively healthy elderly individuals. Neuro- coding non-synonymous variants on protein function biol Aging 2014;35:2881.e7–2881.e10. using the SIFT algorithm. Nat Protoc 2009;4:1073–1081. 39. Wong MM, Cox LK, Chrivia JC. The chromatin remod- 23. Adzhubei IA, Schmidt S, Peshkin L, et al. A method and eling protein, SRCAP, is critical for deposition of the his- server for predicting damaging missense mutations. Nat tone variant H2A.Z at promoters. J Biol Chem 2007;282: Methods 2010;7:248–249. 26132–26139. 24. Lek M, Karczewski KJ, Minikel EV, et al. Analysis of 40. Zovkic IB, Paulukaitis BS, Day JJ, Etikala DM, Sweatt protein-coding genetic variation in 60,706 humans. JD. Histone H2A.Z subunit exchange controls consolida- Nature 2016;536:285–291. tion of recent and remote memory. Nature 2014;515: 25. Levin JZ, Yassour M, Adiconis X, et al. Comprehensive 582–586. comparative analysis of strand-specific RNA sequencing 41. Monroy MA, Ruhl DD, Xu X, Granner DK, Yaciuk P, methods. Nat Methods 2010;7:709–715. Chrivia JC. Regulation of cAMP-responsive element-binding 26. Adiconis X, Borges-Rivera D, Satija R, et al. Comparative protein-mediated transcription by the SNF2/SWI-related analysis of RNA sequencing methods for degraded or low- protein, SRCAP. J Biol Chem 2001;276:40721–40726. input samples. Nat Methods 2013;10:623–629. 42. Nikkel SM, Dauber A, de Munnik S, et al. The phenotype 27. Johnson WE, Li C, Rabinovic A. Adjusting batch effects of Floating-Harbor syndrome: clinical characterization of in microarray expression data using empirical Bayes meth- 52 individuals with mutations in exon 34 of SRCAP. ods. Biostatistics 2007;8:118–127. Orphanet J Rare Dis 2013;8:63.

8 Neurology: Genetics Design and rationale for examining neuroimaging genetics in ischemic stroke The MRI-GENIE study

Anne-Katrin Giese, MD ABSTRACT Markus D. Schirmer, PhD Objective: To describe the design and rationale for the genetic analysis of acute and chronic cere- Kathleen L. Donahue, BS brovascular neuroimaging phenotypes detected on clinical MRI in patients with acute ischemic Lisa Cloonan, BA stroke (AIS) within the scope of the MRI–GENetics Interface Exploration (MRI-GENIE) study. Robert Irie, PhD Methods: MRI-GENIE capitalizes on the existing infrastructure of the Stroke Genetics Network Stefan Winzeck, MS (SiGN). In total, 12 international SiGN sites contributed MRIs of 3,301 patients with AIS. Detailed Mark J.R.J. Bouts, PhD clinical phenotyping with the web-based Causative Classification of Stroke (CCS) system and Elissa C. McIntosh, MA genome-wide genotyping data were available for all participants. Neuroimaging analyses include Steven J. Mocking, MSc the manual and automated assessments of established MRI markers. A high-throughput MRI anal- Adrian V. Dalca, PhD ysis pipeline for the automated assessment of cerebrovascular lesions on clinical scans will be Ramesh Sridharan, PhD developed in a subset of scans for both acute and chronic lesions, validated against gold stan- Huichun Xu, MD, PhD dard, and applied to all available scans. The extracted neuroimaging phenotypes will improve Petrea Frid, MD characterization of acute and chronic cerebrovascular lesions in ischemic stroke, including Eva Giralt-Steinhauer, CCS subtypes, and their effect on functional outcomes after stroke. Moreover, genetic testing MD, PhD will uncover variants associated with acute and chronic MRI manifestations of cerebrovascular Lukas Holmegaard, MD disease. Jaume Roquer, MD, PhD Johan Wasselius, MD, PhD Conclusions: The MRI-GENIE study aims to develop, validate, and distribute the MRI analysis plat- John W. Cole, MD form for scans acquired as part of clinical care for patients with AIS, which will lead to (1) novel Patrick F. McArdle, PhD genetic discoveries in ischemic stroke, (2) strategies for personalized stroke risk assessment, and Neurol Genet Joseph P. Broderick, MD (3) personalized stroke outcome assessment. 2017;3:e180; doi: 10.1212/ Jordi Jimenez-Conde, NXG.0000000000000180 MD, PhD Christina Jern, MD, PhD GLOSSARY 5 5 5 5 Brett M. Kissela, MD, MS ADC apparent diffusion coefficient; AIS acute ischemic stroke; CE cardioembolic; CCS Causative Classification of Stroke; CCSc 5 causative CCS; DICOM 5 Digital Imaging and Communications in Medicine; DWI 5 diffusion-weighted Dawn O. Kleindorfer, MD imaging; DWIv 5 DWI volume; FLAIR 5 fluid-attenuated inversion recovery; GISCOME 5 Genetics of Ischemic Stroke Robin Lemmens, MD, PhD Functional Outcome; GWAS 5 genome-wide association studies; ICC 5 intraclass correlation coefficient; LAA 5 large artery atherosclerosis; MGH 5 Massachusetts General Hospital; MRI-GENIE 5 MRI–GENetics Interface Exploration; mRS 5 Arne Lindgren, MD, PhD modified Rankin Scale; PHI 5 protected health information; QC 5 quality control; SAO 5 small artery occlusion; SiGN 5 James F. Meschia, MD Stroke Genetics Network; SNP 5 single nucleotide polymorphism; SWI 5 susceptibility-weighted imaging; TOAST 5 Trial of Org 10172 Acute Stroke Treatment; VLSM 5 voxel-based lesion–symptom mapping; WMHv 5 white matter hyperintensity Tatjana Rundek, MD, PhD volume; XNAT 5 eXtensible Neuroimaging Archive Toolkit. Ralph L. Sacco, MD, MS Reinhold Schmidt, MD Genome-wide association studies (GWAS) have been instrumental in elucidating the genetics of Pankaj Sharma, MD, PhD complex vascular traits (ischemic stroke1,2 and coronary artery disease3,4) and their risk factors Agnieszka Slowik, MD, (blood pressure,5 atrial fibrillation,6 hyperlipidemia,7 and diabetes mellitus8). Despite recent PhD advances in prevention and treatment, stroke remains a leading cause of adult neurologic Vincent Thijs, MD, PhD 9 Daniel Woo, MD, MS disability and death in the United States and worldwide. Recent GWAS have uncovered several 10,11 Bradford B. Worrall, MD, risk loci for ischemic stroke and its subtypes, specifically PITX2 and ZFHX3 for cardioem- 11,12 11,12 11 11 MSc bolic (CE) stroke, HDAC9 and TSPAN2 for large artery stroke, and ALDH2 for Steven J. Kittner, MD, small artery stroke. These results highlight the necessity for large-scale collaborations such as MPH

Braxton D. Mitchell, *These authors contributed equally to the manuscript. PhD, MPH Author affiliations are provided at the end of the article. Funding information and disclosures are provided at the end of the article. Go to Neurology.org/ng for full disclosure forms. The Article Processing Charge was funded by the NIH. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC Author list continued on next page BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Neurology.org/ng Copyright © 2017 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology 1 Jonathan Rosand, MD, METASTROKE and Stroke Genetics Net- study participants to data sharing was mandatory for all sites. MSc work (SiGN) to identify risk loci for these Shared data include basic demographics, vascular risk factors and Polina Golland, PhD detailed Causative Classification of Stroke (CCS) phenotyping, complex diseases. genotypic data, and clinical MRIs. Ona Wu, PhD* Improving our understanding of the mech- Study oversight. The primary aims and progress of the MRI- Natalia S. Rost, MD, anisms underlying stroke is crucial, as treat- GENIE study are overseen by a Scientific Steering Committee. In MPH* ment options remain limited for this conjunction with the SiGN Publication Committee, the MRI- On behalf of MRI-GENIE GENIE Steering Committee critically reviews project proposals debilitating disease. Unlike clinical diagnosis by collaborators to avoid potential overlap with existing SiGN of stroke, cerebrovascular phenotypes detected projects and to assess feasibility of the proposed projects. This Correspondence to on brain MRI can be characterized with high effort is supported by the Phenotyping Committee, which is in Dr. Rost: validity and quantified with good precision charge of the data access to phenotypes previously obtained [email protected] through SiGN, current data acquisition for neuroimaging using advanced neuroimaging analysis techni- markers, and quality control (QC) of new neuroimaging phe- 13 Supplemental data ques. Targeting these specific endopheno- notypes. In addition, it is responsible for the oversight of statis- at Neurology.org/ng types will lead to substantial pathophysiologic tical analysis of MRI-derived phenotypes and functional outcomes related to specific stroke subtypes. The Neuroimaging insight into stroke mechanisms and foster Analysis Committee is in charge of designing, validating, and future advances in individualized therapy and implementing the MRI pipeline to automatically assess acute prevention. and chronic neuroimaging markers. Moreover, it facilitates and The MRI–Genetics Interface Exploration monitors the assessment of manually obtained neuroimaging markers. The Genetic Analysis Committee conducts the primary (MRI-GENIE) study aims to bridge current genetic analyses for the MRI-GENIE study to identify genetic knowledge gaps by facilitating genetic discov- variants associated with acute and chronic MRI-based manifes- ery and developing novel therapeutic and tations of cerebrovascular disease. It is also essential in conducting preventive strategies in stroke through auto- secondary analyses and additional projects as proposed by col- laborators (detailed listings of committee memberships are mated, multimodal MRI analysis. Leveraging available in coinvestigator appendix e-1 at Neurology.org/ng). the existing infrastructure of SiGN and har- Imaging platform. An integral part of the MRI-GENIE study nessing its expertise, MRI-GENIE focuses on is the centrally maintained imaging platform hosting deidentified the subset of richly phenotyped and genotyped acute or subacute brain MRIs obtained within 48 hours of participants for whom clinical MRIs have been symptom onset from all contributing SiGN sites. The MRI- obtained. This article outlines the premises, GENIE Imaging Platform is maintained centrally at MGH and has been described previously.14 Initially developed in the scope methodology, and aims of MRI-GENIE. of an NIH-funded project to create a centralized system to share canonical human stroke data (R01 NS063925-01A1, O. Wu/ METHODS MRI-GENIE capitalizes on SiGN, an ongoing Sorensen—PI), the underlying technology for the imaging plat- multicenter, NIH-funded collaboration within the community form integrates the eXtensible Neuroimaging Archive Toolkit of stroke neurologists, geneticists, and neuroimaging analysts, (XNAT)15 as the back-end data repository with a flexible, open- which enabled the initial development of the SiGN Imaging source content management system with user-friendly features 14 Platform. We have amassed the largest-to-date collection of (conglomeration of Plone,16 Deliverance,17 and NGINX18) as the ischemic stroke cases with comprehensively ascertained cerebro- front end for the users. An example of such features is the ability vascular phenotypes and genome-wide data. The project is fun- to search the imaging repository by stroke-specific clinical phe- — ded by the NIH-NINDS (R01NS086905, N.R. Rost PI) to notypic variables (e.g., age, sex, or infarct location). Images can be undertake the first major study to jointly model MRI-derived viewed for semiquantitative analysis via a web-based XNAT traits obtained during acute ischemic stroke (AIS) evaluation, plugin.15 causative and phenotypic stroke subtypes, and traditional vascular Upon receipt of MRIs from the individual sites in the “Digital risk factors to accelerate the pace of genetic discoveries and Imaging and Communications in Medicine” (DICOM) format, advance clinical applications in risk and outcome prediction in all images were deidentified to remove protected health informa- ischemic stroke. tion (PHI) potentially embedded into the DICOM headers. In Structure of MRI-GENIE. Participating study sites. addition, DICOM files with image type indicative of screen shots MRI-GENIE is founded on the existing collaborations between were removed to ensure elimination of files with potential PHI members of the multidisciplinary clinical stroke research team that may be “burned” into the screen shots. Each site provided at Massachusetts General Hospital (MGH), Massachusetts Institute phenotypic data (e.g., SiGN ID, sex, race, ethnicity, age, and of Technology (MIT), and the NINDS SiGN investigators. infarct location). Age and sex were cross-referenced with pheno- To date, 12 sites from the initial SiGN study have contributed types documented in the SiGN Phenotype Database, to flag phenotypes, images, and genotypic data of 3,301 participants to potential discrepancies between databases. MRI-GENIE (For a summary, see table and figure; detailed study The imaging platform is open to collaborators for the explo- descriptions have been published previously11). Of those sites, 7 are ration of phenotypic and genetic underpinnings of AIS. Individ- European centers (BASICMAR—Spain, BRAINS—United ual investigators receive access in a project-based manner, after Kingdom, GRAZ—Austria, KRAKOW—Poland, LEUVEN— the MRI-GENIE Scientific Steering Committee and the SiGN Belgium, LUND STROKE REGISTER—Sweden, and SAHLSIS Publication Committee have reviewed the project proposal. —Sweden) and 5 are based in the United States (GASROS, Incorporating the computational workflow for automated seg- GCNKSS, GEOS, ISGS, and MIAMISR). Informed consent of mentation of acute and chronic cerebrovascular phenotypes with

2 Neurology: Genetics Table Basic demographic data for the MRI-GENIE study sites (n 5 12)

Sex Race Study name Center Total scans Mean age (SD) (% female) (% Caucasian)

BASICMAR IMIM-Hospital del Mar, Spain 124 69.8 6 11.0 37.1 94.4

BRAINS Imperial College—London, UK 70 63.2 6 16.4 47.1 94.3

GASROS Mass General Hospital 457 64.9 6 14.5 35.4 93.2

GCNKSS U Cincinnati 245 64.3 6 14.3 49.0 72.7

GEOS U Maryland 76 41.8 6 6.5 26.3 52.6

SAHLSIS U of Gothenburg—Sweden 401 52.4 6 11.7 38.7 100

GRAZ Medical University—Graz, Austria 373 63.3 6 13.7 30.0 100

ISGS Mayo Clinic—Florida 425 65.1 6 14.7 40.9 84.2

KRAKOW Jagiellonian University—Poland 224 60.5 6 13.9 46.4 100

LEUVEN U Hospitals—Leuven, Belgium 448 66.9 6 14.7 42.0 99.6

LUND STROKE Lund University Hospital, 196 63.4 6 12.8 39.3 100 REGISTER Sweden

MIAMISR U Miami 262 62.1 6 13.8 37.0 59.5

Abbreviation: MRI-GENIE 5 MRI–GENetics Interface Exploration. the imaging platform will complete the process for centralized will be drawn. The automatically determined acute DWI volume data abstraction, collection, and sharing for future genetic studies. (DWIv) will be used to evaluate the effect of stroke lesion volume and topography on poststroke outcomes. Voxel-based lesion– MRI-derived phenotypes. The analysis of MRI-derived phe- symptom mapping (VLSM) has been used previously to investi- notypes will include both automated, volumetric analyses as well gate the relationship between ischemic stroke lesion location and as manual, semiquantitative analyses of acute and chronic cere- stroke severity and outcome, but so far has been limited to the brovascular phenotypes. middle cerebral artery territory. The sample size of MRI-GENIE Image preprocessing. Prior to phenotypic analyses, imaging will allow for investigating all vascular territories with VLSM.21 data from some sites required a manual review. As part of the de- The pre-existing burden of cerebrovascular white matter dis- identification process, some sites removed all metadata informa- ease, measured as white matter hyperintensity volume (WMHv), tion from the DICOM headers, including series description, will be computed by a separate automatic analysis algorithm prior to transmission of images. This necessitated a visual review based on T2-FLAIR images.22 Key components of the WMH of cases by a trained operator for sites for which no series descrip- analysis will be the registration of all T2-FLAIR images to a com- tions were available. For these sites, sequence labels for T2 fluid- mon atlas space and segmentation of WMH based on the voxel attenuated inversion recovery (T2 FLAIR), diffusion-weighted intensity information and spatial priors that pattern WMH in the imaging (DWI), susceptibility-weighted imaging (SWI), and brain.23 The algorithm will be trained on manually outlined magnetic resonance angiography were ascertained. Although the WMH segmentations. Validation across sites will be performed quality of specific sequences may preclude future analysis of for each site on a random sample of scans representing the entire specific phenotypes, no participants were excluded for image severity spectrum of WMH. quality reasons (e.g., motion artifact and low-quality images) or QC of automated, quantitative phenotypes. The results of incomplete imaging data (e.g., missing FLAIR sequence) at this automatic acute DWI and WMH lesion analysis will be cross- stage to maximize the overall sample size in analysis of individual validated with manually drawn lesion volumes obtained on phenotypes. For manual semiquantitative analyses, useful infor- a random subset of scans from each site representative of the mation can often be extracted (e.g., location of acute infarct) by disease spectrum. Intraclass correlation coefficient (ICC) will be expert image analysts despite artifacts (e.g., motion). used to assess the agreement between the manual and the auto- Image QC. Image sequences will be excluded from auto- mated volumes. Further evaluation of the automated algorithms mated and semiquantitative analysis based on visual inspection will involve voxel-based comparison of the automated masks with due to excessive motion artifact, incomplete sequence acquisition, manual outlines and classifying true positives, true negatives, false or severe bilateral brain pathology that precludes accurate assess- positives, and false negatives. Performance of the automated al- ment. Details of the MRI data acquisition that are embedded in gorithms will be assessed by analysis of sensitivity, specificity, and the DICOM headers (e.g., MRI manufacturer, model name, field Dice similarity coefficient.24 Subset analysis of algorithm accuracy strength, dimensions, echo time, and repetition time) will be re- as a function of MRI data acquisition parameters will be per- tained to allow for subset analysis of phenotypic information as formed to assess for bias as a function of scanner data quality. a function of data acquisition variability. Semiquantitative analysis of acute and chronic neuroimaging Quantitative analyses of acute and chronic neuroimaging markers. MRIs will be systematically reviewed for acute and chronic markers. The acute infarct volume measured on DWI will be markers of cerebrovascular disease to facilitate topography-based and analyzed through an automated segmentation algorithm.19,20 stroke subtype–specific analyses. Specific data sets may be excluded The algorithm will be trained on an independent cohort of acute on a case-by-case basis if artifacts preclude an accurate reading. DWI stroke data and then applied to MRI-GENIE participants. The will be used to ascertain acute infarct location (vascular territories: results will be validated across sites against a random sample set of middle cerebral artery, anterior cerebral artery, posterior cerebral images selected from each site for which manual lesion volumes artery, vertebrobasilar artery, and multiple vascular territories),

Neurology: Genetics 3 Figure Organizational structure of MRI-GENIE

MRI-GENIE 5 MRI–GENetics Interface Exploration; SiGN 5 Stroke Genetics Network.

number of acute infarct lesions, and DWI-based stroke subtypes major categories are large artery atherosclerosis (LAA), small (cortical, subcortical, and watershed infarct). When available, artery occlusion (SAO), CE stroke, other, and undetermined maps of the apparent diffusion coefficient (ADC) will be cross- causes of stroke. Known rare causes of ischemic stroke were referenced to minimize inclusion of lesions that are not pertinent excluded from MRI-GENIE. The exact workup was performed as to the index stroke. This is necessary because in the subacute to clinically directed by individual study centers. This includes pa- chronic stage, within hyperintense DWI lesions, the corre- tients with rare monogenic causes of stroke (cerebral autosomal sponding ADC values may be pseudonormal or elevated as dominant arteriopathy with subcortical infarcts and leukoence- a result of vasogenic edema.25 T2-FLAIR sequences will be used phalopathy, mitochondrial encephalomyopathy, lactic acidosis, to screen for subacute and chronic infarcts, as well as WMH and stroke-like episodes, and sickle cell syndrome), infectious severity using the Fazekas scale.26 Cerebral microbleed (CMB, # causes (infective endocarditis, meningitis, and primary infection 10 mm) count and location, macrohemorrhages (.10 mm),27 of the arterial wall), hypercoagulability (acute disseminated and hemorrhagic infarct transformation will be rated on 2D T2* intravascular coagulation and heparin-induced thrombocytopenia gradient echo or 3D SWI sequences. In addition, the location of type II), distinct vascular and cardiac disorders (acute arterial arterial occlusion, collateral circulation grade, and evidence of dissection, dilated cardiomyopathy, papillary fibroelastoma, left significant extracranial or intracranial large artery stenosis will be atrial myxoma, and cerebral venous thrombosis), as well as evaluated on MR angiography. migraine-related, drug-induced, or iatrogenic causes of ischemic QC of semiquantitative phenotypes. Readers will be system- stroke. atically trained on a standardized training set of AIS MRIs to In addition to CCS, Trial of Org 10172 Acute Stroke Treat- attain the independent rater status (e.g., Fazekas score or CMBs). ment (TOAST)28 subtyping was conducted by the sites if Furthermore, agreement between raters will be evaluated with the required by the individual study protocols. Ischemic stroke cases ICC for ordinal and continuous data, and a Cohen kappa will be are assigned to the most likely cause of stroke: LAA, SAO, CE, used to assess interrater agreement for categorical data. other, or undetermined. In SiGN, CCS and TOAST agree mod- erately (agreement rate 70%), varying by site and stroke subtype AIS subtyping. All patients with AIS in SiGN underwent exten- with LAA having the highest agreement across CCS and TOAST sive phenotyping through the web-based, standardized algorithm and SAO having the lowest agreement rate.29 TOAST subtyping “Causative Classification of Ischemic Stroke” (details on CCS will be valuable in instances if future replication cohorts will not phenotyping in SiGN have been previously published).14 In brief, have appropriate CCS subtyping. CCS incorporates multiple clinical symptoms, clinical examina- tions, and testing results obtained throughout the clinical stroke Functional outcomes and clinical characteristics. Baseline evaluation and assigns both a phenotypic CCS subtype based on stroke severity was assessed using the NIH Stroke Scale,30,31 abnormal testing results at the time of stroke and causative CCS and the modified Rankin Scale (mRS)32 was used to assess func- (CCSc) subtyping based on prior medical history. If challenged tional outcomes at 3–6 months after stroke in a subset of the with multiple potential causes of ischemic stroke, CCSc assigns SiGN sites. These data will be available through collaboration the most likely cause of stroke based on clinical data. CCSc al- with the Genetics of Ischemic Stroke Functional Outcome locates one of 5 different causative and phenotypic subtypes based (GISCOME) study. The mRS measures the degree of depen- on symptoms, vascular risk factors, and diagnostic tests. The 5 dence and disability after neurologic injury, ranging from

4 Neurology: Genetics 0 (asymptomatic) to 6 (death). Other clinical baseline character- CONCLUSIONS Quantitative neuroimaging has istics were collected per protocol of the individual studies. recently been used to gain further insight into physi- ology and anatomy in both healthy participants (e.g., Genotyping and analysis strategy. Stroke cases were either intracranial volume35) and clinical cohorts (e.g., hip- genotyped previously and genotypes submitted to SiGN, or they pocampal volumes36 and structural neuroimaging bi- were genotyped as part of SiGN, at the Center for Inherited Dis- omarkers37 in Alzheimer disease). In this article, we ease Research on the Illumina HumanOmni5Exome-4v1 array.14 aim to analyze acute (e.g., cerebral infarct volume) This platform includes 4.5 million single nucleotide poly- morphisms (SNPs) genotyped across the genome, resulting in and chronic (e.g., WMH volume and cerebral micro- excellent coverage of both common and infrequent variants bleeds) neuroimaging phenotypes in patients with (.1%). In addition, 240,000 rare polymorphic variants (#1%) AIS. While genetic loci associated with WMH have were genotyped. The majority of cases are of European ancestry previously been reported in healthy populations,38,39 5 (table), but a small number of African Americans (n 249, self- no loci associated with WMH have yet been identi- reported) and Hispanics (n 5 153, self-reported) are available for fied in patients with AIS, despite efforts with large analysis. The SiGN Data Management Core has undertaken the 40 data cleaning and QC procedures for the primary analysis in sample sizes. Similarly, no studies have linked spe- SiGN. QC procedures included data cleaning by subjects with cific genetic loci with cerebral microbleeds or stroke removal of samples with (1) a poor genotyping rate (,98%), (2) lesion volume. To facilitate these genetic studies, identity problems (sex mismatch, unexpected duplicates, and MRI-GENIE will develop, validate, and disseminate cryptic relatedness), (3) chromosomal anomalies, (4) batch ef- an automated analysis pipeline for large-scale phe- fects, and (5) ethnic outliers. Poor-quality SNPs were identified notypic analysis of clinical brain MRI, as part of the on the basis of high levels of missingness. The cleaned data are maintained at the SiGN Data Management Core. The analysis future advances in personalized prediction modeling plan for the MRI-GENIE portion of SiGN includes a principal of stroke risk and outcomes. component analysis to identify and account for population stratification in subsequent genome-wide association testing. AUTHOR AFFILIATIONS Genotype imputation has been performed on the University of From the Department of Neurology (A.-K.G., M.D.S, K.L.D., L.C., Michigan Imputation Server (imputationserver.sph.umich.edu/ J.R., N.S.R.), Massachusetts General Hospital, Harvard Medical School, index.html)33 using the Haplotype Reference Consortium Boston; Program in Medical and Population Genetics (A.-K.G., J.R.), panel.34 The primary genetic analysis will be to test for the Broad Institute of MIT and Harvard; Computer Science and Artificial association of SNPs with the DWI and WMH volumes. The Intelligence Laboratory (M.D.S., A.V.D., R.S., P.G.), MIT, Cambridge, association analyses will be performed under a linear regression MA; Department of Radiology (R.I., S.W., M.J.R.J.B., E.C.M., S.J.M., A.V.D., O.W.), Athinoula A. Martinos Center for Biomedical Imaging, model with allelic dose (0, 1, or 2 copies of the reference allele) as Massachusetts General Hospital (MGH), Harvard Medical School, Charles- the independent variable adjusted for age, sex, and population town; Division of Endocrinology, Diabetes and Nutrition (H.X., P.F.M., stratification as calculated by principal component analysis. DWI B.D.M.), Department of Medicine, University of Maryland School of Med- and WMH volumes will be adjusted for average head size and icine, Baltimore; Department of Clinical Sciences Lund, Neurology (P.F., natural log transformed to facilitate modeling with linear A.L.) and Department of Clinical Sciences Lund, Radiology (J.W.), Lund regression because of nonlinear volume distributions. The current University, Sweden; Department of Neurology and Rehabilitation Medi- sample size of 3,301 participants will provide 80% power to cine, Neurology (P.F.) and Department of Radiology, Neuroradiology detect variants accounting for as little as 1.2% of the variation in (J.W.), Skåne University Hospital, Malmö, Sweden; Department of Neu- DWI or WMH at genome-wide levels of statistical significance (i. rology (E.G.-S., J.R., J.J.-C.), Neurovascular Research Group (NEUVAS), ’ e., p , 5 3 1028). Secondary analyses will include the modeling IMIM-Hospital del Mar (Institut Hospital del Mar d Investigacions Mèd- iques), Universitat Autonoma de Barcelona, Spain; Institute of Neurosci- of CCS subtypes and ethnic group–specific analyses. SNPs found ence and Physiology (L.H.) and Institute of Biomedicine (C.J.), Sahlgrenska to be associated with WMHv and/or DWIv will be tested for Academy, University of Gothenburg, Sweden; Department of Neurology association with 90-day mRS data available in the GISCOME (J.W.C., S.J.K.), University of Maryland School of Medicine and Veterans study. Additional analyses will be performed to examine stroke Affairs Maryland Health Care System, Baltimore; Department of Neurol- – subtype specific genetic effects on DWIv and WMHv as well as ogy and Rehabilitation Medicine (J.P.B., B.M.K., D.O.K., D.W.), Univer- the genetic underpinnings of additional neuroimaging markers. sity of Cincinnati College of Medicine, OH; Department of Neurosciences (R.L.), Experimental Neurology, KU Leuven—University of Leuven; VIB —Center for Brain & Disease Research (R.L.); Department of Neurology Incoming sites. In the spirit of open collaboration, the MRI- (R.L.), University Hospitals Leuven, Belgium; Department of Neurology GENIE Imaging Platform is available to collaborators. We are and Rehabilitation Medicine, Neurology (A.L.), Skåne University Hospital, currently incorporating 8 new sites (estimated additional total Lund, Sweden; Department of Neurology (J.F.M.), Mayo Clinic, Jackson- n 5 3,890) including (1) Secondary Prevention of Small Sub- ville; Department of Neurology (T.R.), Miller School of Medicine, Univer- cortical Strokes, n z 1,000, (2) Siblings With Ischemic Stroke sity of Miami, FL; Department of Neurology (R.S.), Clinical Division of Study, n z 300, and (3) Washington State University–St. Louis Neurogeriatrics, Medical University Graz, Austria; Institute of Cardiovas- stroke patient collection, n z 640, (4) Helsinki-2000 study, n z cular Research (P.S.), Royal Holloway University of London (ICR2UL), ’ 300, (5) Australian Stroke Genetics Cohort, n z 100, (6) Stroke Egham; Ashford and St Peter s Hospital (P.S.), UK; Department of Neu- in Young Fabry Patients, n z 800, (7) University of Campinas rology (A.S.), Jagiellonian University Medical College, Krakow, Poland; Stroke Division (V.T.), Florey Institute of Neuroscience and Mental stroke patient collection, n z 150, and (8) Follow-up of Tran- Health; Department of Neurology (V.T.), Austin Health, Heidelberg, Aus- sient Ischemic Attack and Unelucidated Risk Factor Evaluation tralia; Departments of Neurology and Public Health Sciences (B.B.W.), Study/Observational Dutch Young Symptomatic StrokE studY, University of Virginia, Charlottesville; Geriatric Research and Education z n 600. These collaborations will lead to the development of Clinical Center (B.D.M.), Veterans Administration Medical Center, Balti- one of the largest (n z 7,000) databases of patients with ischemic more, MD; and Center for Human Genetic Research (J.R.), Massachusetts stroke with MRI and genome-wide genotyping available to date. General Hospital, Boston.

Neurology: Genetics 5 AUTHOR CONTRIBUTIONS and Elsevier; has received research support from Massachusetts Institute Anne-Katrin Giese: review of imaging data, data analysis, drafting and of Technology; and has been an employee of Massachusetts Institute of revising the manuscript. Markus D. Schirmer: design of white matter hy- Technology and Massachusetts General Hospital (part of Partners perintensity lesion outlining algorithm and critical revision for intellectual HealthCare, a nonprofit); in addition, his wife is employed by Partners content. Kathleen L. Donahue and Lisa Cloonan: maintenance and HealthCare. R. Sridharan reports no disclosures. H. Xu holds a patent for review of imaging data and critical revision for intellectual content. Methods for diagnosing etiologies of ischemic stroke by detecting expres- Robert Irie: design, development, and maintenance of imaging database sion levels of genes and has received research support from the NIH, and critical revision for intellectual content. Stefan Winzeck: design of University of Maryland Claude D. Pepper Older Americans Indepen- acute lesion outlining algorithm and critical revision for intellectual con- dence Center, and AHA. P. Frid, E. Giralt-Steinhauer, and L. Holme- tent. Mark J.R.J. Bouts: maintenance of imaging database and critical gaard report no disclosures. J. Roquer has received research support from ’ revision for intellectual content. Elissa C. McIntosh: maintenance and Spain s Ministry of Health. J. Wasselius reports no disclosures. J.W. Cole review of imaging data and critical revision for intellectual content. has received publishing royalties from Springer. P.F. McArdle has Steven J. Mocking: maintenance of imaging database, design of acute received research support from the NIH. J.P. Broderick has served on lesion outlining algorithm, and critical revision for intellectual content. a scientific advisory board for Astra Zeneca; has received travel funding Adrian V. Dalca and Ramesh Sridharan: design of white matter hyperin- from universities as invited visiting professor over the past 2 years – tensity lesion outlining algorithm and critical revision for intellectual con- (approximately 4 5 times per year); has received honoraria for speaking tent. Huichun Xu: analysis of genetic data and critical revision for placed in faculty academic fund within department; holds a patent for intellectual content. Petrea Frid and Eva Giralt-Steinhauer: acquisition Method for Controlling the Lysis of Coagulated Blood with Apolipopro- of imaging data and critical revision for intellectual content. Lukas tein e4 Phenotype; has received publishing royalties from Elsevier; has Holmegaard: acquisition of data and critical revision for intellectual con- been a consultant for Pfizer; and has received research support from tent. Jaume Roquer and Johan Wasselius: acquisition of imaging data Genentech and the NIH/NINDS. J. Jimenez-Conde has been an and critical revision for intellectual content. John W. Cole, Patrick employee of IMIM-Hospital del Mar and has received research support F. McArdle, Joseph P. Broderick, Jordi Jimenez-Conde, Christina Jern, from the NIH/NINDS, ISCIII/Spanish Government Grants, and La Brett M. Kissela, James F. Meschia, Tatjana Rundek, Ralph L. Sacco, Marató TV3. C. Jern has received research support from the Sahlgrenska Reinhold Schmidt, Pankaj Sharma, Agnieszka Slowik, Vincent Thijs, Academy and the Sahlgrenska University, the Swedish Medical Research Daniel Woo, and Bradford B. Worrall: acquisition of data and critical Council, the Swedish Heart Lung Foundation, and the Swedish Stroke revision for intellectual content. Steven J. Kittner: study concept and Foundation. B.M. Kissela has served on the scientific advisory board of design and critical revision for intellectual content. Braxton D. Mitchell: Ipsen Pharmaceuticals; has received compensation to perform event study concept and design, genetic analyses, and critical revision for intel- adjudication for clinical trials sponsored by Janssen and AbbVie; and lectual content. Jonathan Rosand: study concept and design and critical has received research support from the NIH. D.O. Kleindorfer has revision for intellectual content. Polina Golland: study concept and received travel funding from Genentech; has served on the speakers design, design of white matter hyperintensity lesion outlining algorithm, bureau of Genentech; and has received research support from the and critical revision for intellectual content. Ona Wu: study concept and NIH/NINDS. R. Lemmens has received research support from FWO design, design and development of imaging database, design of acute Flanders. A. Lindgren has served on scientific advisory boards for Astra lesion outlining algorithm, and critical revision for intellectual content. Zeneca, Bristol-Myers Squibb, Boehringer Ingelheim, Sanofi, Pfizer, Natalia S. Rost: study concept and design, study supervision, drafting Bayer, and ReNeuron; has received travel funding/speaker honoraria from the manuscript, and critical revision for intellectual content. Astra Zeneca, Bristol-Myers Squibb, Boehringer Ingelheim, Sanofi, Pfizer, and Bayer; and has served on the editorial board of Neurologi i ACKNOWLEDGMENT Sverige. J.F. Meschia has served on a scientific advisory board for the NINDS NeuroNEXT clinical trial network; has served on the editorial On behalf of the BASICMAR project, the authors thank Carolina Soriano- boards of Mayo Clinic Proceedings and the European Journal of Neurology; Tarraga and Alejandra Gomez for their critical support of the study. and has received research support from the NIH/NINDS. T. Rundek has R. Lemmens is a senior clinical investigator of FWO Flanders. served on the editorial boards of Stroke, Neurology, Cerebrovascular Dis- ease, and the Journal of Ultrasound in Medicine; and has received research STUDY FUNDING support from the NIH/NINDS. R.L. Sacco has served on scientific — NIH-NINDS (MRI-GENIE: R01NS086905 PI N.S. Rost; advisory boards for SOCRATES Trial through UCSF (indirect sponsor — — K23NS064052, R01NS082285 N.S. Rost; SiGN: U01 NS069208 Astra Zeneca) and EUCLID trial through Duke Clinical Research Insti- J. Rosand and S.J. Kittner; and R01NS059775, R01NS063925, tute (indirect sponsor Astra Zeneca); has served on the editorial boards of R01NS082285, P50NS051343, R01NS086905, and U01 NS069208 Stroke and Neuroepidemiology; has been a consultant for Boehringer In- — — O. Wu), NIH NIBIB (P41EB015902 P. Golland; U01NS030678 gelheim; and has received research support from Boehringer Ingelheim, — — B.M. Kissela and D.O. Kleindorfer; and EB015325 O. Wu), Swed- the NIH/NINDS, the Evelyn McKnight Brain Institute, and the Amer- — ish Heart and Lung Foundation, the Swedish Stroke Association A. ican Heart Association Bugher Foundation. R. Schmidt has served on Lindgren, Swedish Research Council and the Swedish Heart and Lung scientific advisory boards for Rivastigmine Patch; Novartis JSW Life — Foundation C. Jern, Spanish Ministry of Science and Innovation, In- Sciences, and Dementia Trial Axon Neuroscience; has received travel stituto de Salud Carlos III (Funding for Research in Health [PI051737], expenses and/or honoraria for lectures from Pfizer, Novartis, Merz Aus- — [PI10/02064], [PI12/01238], and [PI15/00451 J. Jimenez-Conde]), tria, and Lundbeck, and for educational activity from Takeda and Axon Fondos FEDER/EDRF Red de Enfermedades Vasculares Cerebrales Neuroscience; has served on the editorial board of Clinical Neurology and — (RD16/0019/0002 J. Roquer), Fundació la Marató TV3 (76/C/2011 Neurosurgery; has been an employee of Pfizer, Novartis, Merz Austria, — ’ — J. Jimenez-Conde) and Recercaixa 13 (JJ086116 J. Jimenez-Conde), Lundbeck, and Takeda; has been a consultant for QPS and Austria Axon Wistron Corporation (P. Golland). Neuroscience; and has received research support from Austrian Science Fund. P. Sharma has served on scientific advisory boards for Bohringer DISCLOSURE and Bayer and has served on the editorial board of the Journal of Royal A.-K. Giese has received research support from the NIH/NINDS. Society of Medicine Cardiovascular Disease. A. Slowik reports no disclosures. M.D. Schirmer reports no disclosures. K.L. Donahue has received V. Thijs has served on scientific advisory boards for Bayer, Boehringer In- research support from the NIH/NINDS. L. Cloonan has been an gelheim, Amgen, Pfizer/BMS, Daiichi Sankyo, and Medtronic; has served on employee of Decision Resources Group. R. Irie has received travel fund- the editorial boards of Acta Neurologica Belgica, Stroke, International Journal of ing from the Department of Defense. S. Winzeck reports no disclosures. Stroke, European Stroke Journal,andEuropean Journal of Emergency Medicine; M.J.R.J. Bouts has received research support from Leiden University. has been a consultant for Bayer, Boehringer Ingelheim, Amgen, Pfizer/BMS, E.C. McIntosh reports no disclosures. S.J. Mocking has been an Daiichi Sankyo, and Medtronic; and has received research support from the employee of Google. A.V. Dalca has served on the editorial board of NIH/NINDS and VIB Flanders. D. Woo has served on the editorial board Imaging Genetics; has received publishing royalties from Imaging Genetics of Stroke Research and Treatment and has received research support from the

6 Neurology: Genetics NIH. B.B. Worrall has served on the editorial board of Neurology; receives 11. NINDS Stroke Genetics Network (SiGN), International publishing royalties from Merritt’s Neurology; and has received research Stroke Genetics Consortium. Loci associated with ischae- support from the NIH/NINDS and the University of Virginia. S.J. Kittner mic stroke and its subtypes (SiGN): a genome-wide asso- has served on the editorial board of Neuroepidemiology; has received research ciation study. Lancet Neurol 2016;15:174–184. support from the NIH/NINDS and has participated in legal proceedings 12. Traylor M, Farrall M, Holliday EG, et al. Genetic risk regarding Atlantic States Insurance (review of medical case). B.D. Mitchell has served on the editorial board of Diabetes and has received research factors for ischaemic stroke and its subtypes (the META- support from the NIH/NINDS. J. Rosand has served on scientific advisory STROKE collaboration): a meta-analysis of genome-wide boards for Pfizer and DSMB; has served on the editorial boards of Lancet association studies. Lancet Neurol 2012;11:951–962. Neurology and Stroke; and has received research support from the NIH/ 13. Masdeu JC, Irimia P, Asenbaum S, et al. EFNS guideline NINDS. P. Golland has received travel funding/speaker honoraria from on neuroimaging in acute stroke. Report of an EFNS task the University of Utah School of Computing Visiting Committee; has served force. Eur J Neurol 2006;13:1271–1283. on the editorial board of Journal of Medical Image Analysis; and has received 14. Meschia JF, Arnett DK, Ay H, et al. Stroke genetics net- research support from Wistron Corporation and the NIH. O. Wu has work (SiGN) study design and rationale for a genome- received travel funding/speaker honoraria from RSNA and the NIH; holds wide association study of ischemic stroke subtypes. Stroke patents for Delay-compensated calculation of tissue blood flow and Method – for evaluating novel stroke treatments using a tissue risk map; receives pub- 2013;44:2694 2702. lishing royalties from Nobel Prize Internet Archive; has been a consultant for 15. Marcus DS, Olsen TR, Ramaratnam M, Buckner RL. The Penumbra; has received research support from Genentech and the NIH/ extensible neuroimaging archive toolkit: an informatics NINDS; and receives royalty payments from General Electric, Olea Medical, platform for managing, exploring, and sharing neuroimag- and Imaging Biometrics. N.S. Rost has served on scientific advisory boards ing data. Neuroinformatics 2007;5:11–33. for Genzyme Co., Merck, Boston Biomedical Associates, Omniox, and Bio- 16. The Plone Foundation. Plone CMS: open source content Telemetry; has served on the editorial boards of Stroke and Current Treatment management. Available at: plone.org/homepage. Accessed Options in Cardiovascular Medicine; receives publishing royalties from UpTo- March 1, 2017. Date; has been a consultant for Genzyme Co., Merck, Boston Biomedical 17. The open planning project deliverance. Available at: py- Associates, and Omniox; and has received research support from the NIH/ NINDS. Go to Neurology.org/ng for full disclosure forms. thonhosted.org/Deliverance/. Accessed March 1, 2017. 18. NGINX Inc. NGINX: high performance load balancer, web server, & reverse proxy. Available at: nginx.com/. Received March 9, 2017. Accepted in final form June 30, 2017. Accessed March 1, 2017. 19. Winzeck S, Bouts M, McIntosh E, et al. Early identifica- REFERENCES tion of tissue at risk of infarction after Acute Ischemic 1. Markus HS, Bevan S. Mechanisms and treatment of is- Stroke. 2015 NIPS Workshop on Machine Learning for chaemic stroke—insights from genetic associations. Nat Healthcare; December 11, 2015; Montreal, Canada. – Rev Neurol 2014;10:723 730. 20. Wu O, Winzeck S, Mocking SJT, et al. 19th Workshop of 2. Lindgren A. Stroke genetics: a review and update. J Stroke the International Stroke Genetics Consortium, April – 2014;16:114 123. 28–29, 2016, Boston, Massachusetts, USA: 2016.002 3. Pjanic M, Miller CL, Wirka R, Kim JB, DiRenzo DM, Modeling acute lesion volume and topography for Quertermous T. Genetics and genomics of coronary artery genetic analysis (MRI-GENIE study). Neurol Genet 2017;3 disease. Curr Cardiol Rep 2016;18:102. (1 suppl 1):S2–S11. doi: 10.1212/NXG.0000000000000100. 4. McPherson R, Tybjaerg-Hansen A. Genetics of coronary 21. Wu O, Cloonan L, Mocking SJ, et al. Role of acute – artery disease. Circ Res 2016;118:564 578. lesion topography in initial ischemic stroke severity and 5. Geller DS. New developments in the genetics of hyper- long-term functional outcomes. Stroke 2015;46:2438– tension: what should clinicians know? Curr Cardiol Rep 2444. 2015;17:122. 22. Schirmer MD, Dalca A, Sridharan R, et al. 19th Workshop 6. Tucker NR, Clauss S, Ellinor PT. Common variation in of the International Stroke Genetics Consortium, April atrial fibrillation: navigating the path from genetic associ- 28–29, 2016, Boston, Massachusetts, USA: 2016.008 The – ation to mechanism. Cardiovasc Res 2016;109:493 501. automated pipeline for white matter hyperintensity detec- 7. Hegele RA, Ginsberg HN, Chapman MJ, et al. The poly- tion: the MRI-GENIE study. Neurol Genet 2017;3(1 suppl genic nature of hypertriglyceridaemia: implications for def- 1):S2–S11. doi: 10.1212/NXG.0000000000000100. inition, diagnosis, and management. Lancet Diabetes 23. Dalca AV, Sridharan R, Cloonan L, et al. Segmentation of – Endocrinol 2014;2:655 666. cerebrovascular pathologies in stroke patients with spatial 8. Karaderi T, Drong AW, Lindgren CM. Insights into the and shape priors. Med Image Comput Comput Assist genetic susceptibility to type 2 diabetes from genome-wide Interv 2014;17:773–780. association studies of obesity-related traits. Curr Diab Rep 24. Mocking S, Rost NS, Fitzpatrick KM, et al. Automatic 2015;15:83. Segmentation of Diffusion MRI from the Genes Associ- 9. Centers for Disease Control and Prevention, National ated with Stroke Risk and Outcomes Study. ISMRM 21st Center for Health Statistics. Underlying cause of death Annual Meeting & Exhibition; April 20–26, 2013; Salt 1999-2014 on CDC WONDER online database, released Lake City, UT. 2015. Data are from the multiple cause of death files, 25. Copen WA, Schwamm LH, González RG, et al. Ischemic 1999-2014, as compiled from data provided by the 57 stroke: effects of etiology and patient age on the time vital statistics jurisdictions through the vital statistics course of the core apparent diffusion coefficient. Radiology Cooperative Program. Available at: wonder.cdc.gov/ucd- 2001;221:27–34. icd10.html. Accessed November 21, 2016. 26. Fazekas F, Chawluk JB, Alavi A, Hurtig HI, Zimmerman 10. Malik R, Traylor M, Pulit SL, et al. Low-frequency and RA. MR signal abnormalities at 1.5 T in Alzheimer’s common genetic variation in ischemic stroke the META- dementia and normal aging. AJR Am J Roentgenol STROKE collaboration. Neurology 2016;86:1217–1226. 1987;149:351–356.

Neurology: Genetics 7 27. Wardlaw JM, Smith EE, Biessels GJ, et al. Neuroimaging 34. Loh P-R, Danecek P, Palamara PF, et al. Reference-based standards for research into small vessel disease and its con- phasing using the haplotype reference consortium panel. tribution to ageing and neurodegeneration. Lancet Neurol Nat Genet 2016;48:1443–1448. 2013;12:822–838. 35. Adams HHH, Hibar DP, Chouraki V, et al. Novel genetic 28. Adams HP, Bendixen BH, Kappelle LJ, et al. Classifica- loci underlying human intracranial volume identified tion of subtype of acute ischemic stroke. Definitions for through genome-wide association. Nat Neurosci 2016; use in a multicenter clinical trial. TOAST. Trial of Org 19:1569–1582. 10172 in Acute Stroke Treatment. Stroke 1993;24:35–41. 36. Song A, Yan J, Kim S, et al. Network-based analysis of genetic 29. McArdle PF, Kittner SJ, Ay H, et al. Agreement between variants associated with hippocampal volume in Alzheimer’s TOAST and CCS ischemic stroke classification the disease: a study of ADNI cohorts. BioData Min 2016;9:3. NINDS SiGN Study. Neurology 2014;83:1653–1660. 37. Moon SW, Dinov ID, Kim J, et al. Structural neuroimag- 30. Brott T, Adams HP, Olinger CP, et al. Measurements of ing genetics interactions in Alzheimer’s disease. acute cerebral infarction: a clinical examination scale. J Alzheimers Dis 2015;48:1051–1063. Stroke 1989;20:864–870. 38. Fornage M, Debette S, Bis JC, et al. Genome-wide asso- 31. Lyden P, Brott T, Tilley B, et al. Improved reliability of ciation studies of cerebral white matter lesion burden: the the NIH Stroke Scale using video training. NINDS TPA CHARGE consortium. Ann Neurol 2011;69:928–939. Stroke Study Group. Stroke 1994;25:2220–2226. 39. Verhaaren BFJ, Debette S, Bis JC, et al. Multiethnic genome- 32. Bonita R, Beaglehole R. Recovery of motor function after wide association study of cerebral white matter hyperinten- stroke. Stroke 1988;19:1497–1500. sities on MRI. Circ Cardiovasc Genet 2015;8:398–409. 33. Das S, Forer L, Schönherr S, et al. Next-generation geno- 40. Traylor M, Zhang CR, Adib-Samii P, et al. Genome-wide type imputation service and methods. Nat Genet 2016;48: meta-analysis of cerebral white matter hyperintensities in 1284–1287. patients with stroke. Neurology 2016;86:146–153.

8 Neurology: Genetics No rare deleterious variants from STK32B, PPARGC1A, and CTNNA3 are associated with essential tremor

Gabrielle Houle, BSc ABSTRACT Amirthagowri Objective: To assess the contribution of variants in STK32B, PPARGC1A,andCTNNA3 as Ambalavanan, PhD essential tremor (ET) predisposing factors following their association in a 2-stage genome-wide Jean-François Schmouth, association study (GWAS). PhD Methods: The coding regions of these genes was examined for the presence of rare variants Claire S. Leblond, PhD using two approaches: (1) Looking at whole-exome and whole-genome sequencing data of 14 Dan Spiegelman, MSc autosomal dominant multiplex ET families. (2) Conducting a targeted massive parallel sequencing Sandra B. Laurent, BTS to examine the three genes in cohorts of 269 ET cases and 287 control individuals. The cumula- Cynthia V. Bourassa, MSc tive impact of rare variants was assessed using SKAT-O analyses using (1) all variants, (2) only Celene Grayson, PhD rare variants, and (3) only the rare variants altering the mRNA. Michel Panisset, MD, FRCPC Results: Thirty-four variants were identified. No difference emerged regarding the distributions of Sylvain Chouinard, MD, individual variants (or gene) between cases and controls. FRCPC Conclusion: No rare exonic variants further validated one of these genes as a risk factor for ET. Nicolas Dupré, MD, The recent GWAS offers promising avenues, but the genetic heterogeneity of ET is nonetheless FRCPC challenging for the validation of risk factors, and ultimately larger cohorts of cases should help to Carles Vilariño-Güell, overcome this task. Neurol Genet 2017;3:e195; doi: 10.1212/NXG.0000000000000195 PhD Alex Rajput, MD, GLOSSARY FRCPC ET 5 essential tremor; ExAC 5 Exome Aggregation Consortium; GWAS 5 genome-wide association study; QC 5 quality 5 5 5 Simon L. Girard, PhD control; SKAT-O sequence kernel association test; WES whole-exome sequencing; WGS whole-genome sequencing. Patrick A. Dion, PhD With a worldwide prevalence of 0.9% across age groups (#4.6% in individuals older than 65 Guy A. Rouleau, PhD, 1 MD, FRCPC years), essential tremor (ET) is one of the most common human movement disorders. ET is characterized by involuntary oscillations of a body part, primarily in upper limbs, during postural control and voluntary motion.2,3 Despite strong evidence supporting ET to be an Correspondence to Dr. Rouleau: inherited predisposition, very few predisposing genes have been identified.4 Recently, a 2-stage [email protected] genome-wide association study (GWAS) using 2,807 cases and 6,441 controls of European descent5 was reported. This study revealed disease associations for intronic variants within 3 genes: a serine/threonine kinase (STK32BI), a transcriptional coactivator (PPARGC1A), and a cell-adhesion molecule (CTNNA3). The authors performed linkage disequilibrium analysis and showed no additional associated markers in neighboring genes. This study aims to establish whether coding variants from these genes might be associated with ET. Because low-frequency and rare variants are not tagged by conventional genome-wide genotyping arrays, they may represent an important and understudied component of complex trait genetics. Using higher Supplemental data at Neurology.org/ng From the Montreal Neurological Institute and Hospital (G.H., A.A., J.-F.S., C.S.L., D.S., S.B.L., C.V.B., P.A.D., G.A.R.), Quebec, Canada; Department of Human Genetics (G.H., A.A., C.S.L., P.A.D., G.A.R.) and Department of Neurology and Neurosurgery (J.-F.S., P.A.D., G.A.R.), McGill University, Montreal, Quebec, Canada; Xenon Pharmaceuticals Inc (C.G.), Burnaby, British Columbia, Canada; André Barbeau Movement Disorders Unit (M.P., S.C.), Centre Hospitalier Universitaire de Montréal (CHUM)-Notre-Dame, Quebec, Canada; Department of Medicine (N.D.), Faculty of Medicine, Laval University, Quebec, Canada; Département des Sciences Neurologiques (N.D.), CHU de Québec (Enfant-Jésus), Quebec, Canada; Department of Medical Genetics (C.V.-G.), University of British Columbia, Vancouver, British Columbia, Canada; Division of Neurology (A.R.), Saskatchewan Movement Disorders Program, University of Saskatchewan, Saskatoon Health Region, Saskatoon, Canada; and Département des Sciences Fondamentales (S.L.G.), Université du Québec à Chicoutimi, Saguenay, Canada. Funding information and disclosures are provided at the end of the article. Go to Neurology.org/ng for full disclosure forms. The Article Processing Charge was funded by the authors. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Neurology.org/ng Copyright © 2017 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 resolution methods to interrogate variants distribution between cases and controls (Fisher exact across the entire frequency spectrum, this test after Bonferroni correction). To assess the cumu- study has the potential to find additional evi- lative impact of rare variants, we performed SKAT-O dence to support the role of the genes in the analyses using (1) all variants, (2) only rare variants, disease etiology. and (3) only the rare variants altering the mRNA. Using individual genes as bin delimiters, none of the SKAT-O tests led to a rejection of the null hypothesis METHODS Patients were recruited in different centers across (p . 0.05 after Bonferroni correction); thus, no dif- Canada, and a senior neurologist trained to evaluate movement ference in variant distribution for any of the genes was disorders reviewed their ET diagnoses. Exclusion criteria included (1) an identified cause of exaggerated physiologic tremor, (2) pres- observed between the ET cases and the controls. ence of other neurologic deficits (parkinsonisms, polyneuritis, and others), and (3) an orthostatic tremor or (4) a psychogenic- DISCUSSION In this study, we performed a combina- like tremor. Signed written informed consent forms were ob- tion of exonic and targeted DNA sequencing of tained from each individual studied. 3 genes. ET-affected cases and matching controls In an effort to identify potentially deleterious variants in the from European descents were recruited to identify rare genes recently associated with ET by GWAS, we first examined variants associated with ET. Genes were chosen for the whole-exome and whole-genome sequencing (WES/WGS) data from 54 cases across 14 multiplex families with autosomal analysis on the basis of a recently published study that dominant ET. Secondly, we selected 269 unrelated patients with showed association between variants located in their ET and 287 ethnically matched unrelated individuals with no intronic regions and ET. It is important that this neurologic disorder is known for a case-control study. European previous study relied on GWAS approaches, which decent participants were recruited from Canadian movement dis- generally do not interrogate rare genetic variants. order clinics. Targeted massive parallel sequencing was performed Both an examination of WES/WGS data obtained across the coding regions of STK32B (NM_018401), PPARGC1A (NM_013261), and CTNNA3 (NM_013266). Read mapping, from a cohort of familial ET cases and a case-control variant calling, and quality controls (QCs) are described in the study (Canadians of European decent) analysis failed supplementary material at Neurology.org/ng. A total of 34 var- to identify additional STK32B, PPARGC1A,and iants passed the QC validation. CTNNA3 variants that are associated with ET. Single-variant case-control associations were analyzed using Although a few rare coding variants were identified 6 a Fisher exact test (PLINK v1.90). In addition, a gene-based, across the genes, SKAT-O did not reveal those to have variance-component test was performed using an optimal a cumulative effect toward ET. Rare variants are inher- sequence kernel association test (SKAT-O).7,8 Results were con- sidered statistically significant when p values were #0.05 after ently uncommon, and the size of the cohorts examined Bonferroni correction for multiple testing. here was modest. Making an allowance for the genetic heterogeneity of ET and that up to 50% of individuals RESULTS All familial ET samples for which WES or diagnosed with ET have been suggested to be misdiag- WGS data were available had $97% of the targeted nosed,9 it is likely that the increased power of detection sequences covered at $15X. After genotype and variant of a larger cohort might be warranted to further vali- QC, 12 variants were identified and 7 of these altered date these genes. Nonetheless, this study identified the amino acid sequence. Only 1 variant (CTNNA3 coding variants in 3 genes recently associated with ET.5 c.1453A.T) was rare in the general population (fre- quency ,0.01 in Exome Aggregation Consortium AUTHOR CONTRIBUTIONS [ExAC]) databases,8 but it was observed only in 3 of Ms. Houle: design or conceptualization of the study; analysis or interpretation of the data; and drafting or revising the manuscript for intellectual content. the 4 affected individuals of a single family. Despite the Drs. Ambalavanan and Schmouth: analysis or interpretation of the data. absence of segregating rare deleterious variants across Dr. Leblond: design or conceptualization of the study. Mr. Spiegelman, familial cases, we proceeded with the analysis of the Ms. Laurent, and Ms. Bourassa: analysis or interpretation of the data. targeted sequencing in cases and controls. After remov- Drs. Grayson, Panisset, Chouinard, Dupré, Vilariño-Güell, Rajput, and 5 Girard: drafting or revising the manuscript for intellectual content. Drs. Dion ing poorly captured samples (n 8), the remaining and Rouleau: design or conceptualization of the study; analysis or interpretation ones had $80% of the targeted sequences covered at of the data; and drafting or revising the manuscript for intellectual content. .15X, and 34 coding variants were identified. Briefly, 20 nonsynonymous variants were found, among which STUDY FUNDING 3werecommonones(frequency.0.01 in ExAC) and This study was fully supported by a Foundation Scheme application (#332971) from the Canadian Institutes of Health and in part Xenon® 17 were rare. In addition, we found 1 rare nonframe- Pharmaceutical which of reference sequencing data but not of the tar- shifting in STK32B of a control individual; geted sequencing presented in this manuscript. unfortunately, no DNA from family members was available to test for co-segregation with the phenotype DISCLOSURE (see table e-1 for a detailed list of variants). G. Houle signed an acknowledgment of Research Agreement for her involve- ment in the study conducted within the research agreement concluded Of the 34 single nucleotide polymorphisms between Xenon Pharmaceutical and Dr. Rouleau and has received research identified, none had a significantly different allelic support from the Canadian Institute of Health Research (CIHR). A.

2 Neurology: Genetics Ambalavanan and J. Schmouth report no disclosures. C. Leblond has received REFERENCES research support from ALS Canada. D. Spiegelman, S. Laurent, and C. Bour- 1. Louis ED, Ferreira JJ. How common is the most common assa report no disclosures. C. Grayson serves as a Research Scientist in Clinical adult movement disorder? Update on the worldwide prev- Genetics for Xenon Pharmaceuticals Inc and acted as a project manager in alence of essential tremor. Mov Disord 2010;25:534–541. a separate essential tremor study conducted within a research agreement con- 2. Louis ED. Clinical practice. Essential tremor. N Engl J Med cluded between Xenon Pharmaceutical and Dr. Rouleau (September 25, 2001;345:887–891. 2014, to February 1, 2016). A part of the high-throughput sequencing data – used in this study were provided by Xenon Pharmaceuticals to be freely used 3. Louis ED. Essential tremor. Lancet Neurol 2005;4:100 by the team of Dr. Rouleau outside their Research Agreement. M. Panisset 110. has served on the scientific advisory boards of Allergan and Merz and has 4. Schmouth JF, Dion PA, Rouleau GA. Genetics of essential received research support from Allergan, Merz, Medtronic, and the Weston tremor: from phenotype to genes, insights from both Foundation. S. Chouinard reports no disclosures. N. Dupré has received travel human and mouse studies. Prog Neurobiol 2014;119– funding/speaker honoraria from Actelion Pharmaceuticals and has served on 120:1–19. the editorial board of Cerebellum & Ataxias.C.Vilariño-Güellreportsno 5. Muller SH, Girard SL, Hopfner F, et al. Genome-wide disclosures. A. Rajput has served on a scientific advisory board sponsored by association study in essential tremor identifies three new Ipsen Biopharmaceuticals Canada Inc.; has received speaker honoraria from loci. Brain 2016;139:3163–3169. Teva and Parkinson Society Canada; has served on the editorial boards of 6. Purcell S, Neale B, Todd-Brown K, et al. PLINK: a tool set Canadian Journal of Neurological Sciences and Parkinsonism and Related Dis- orders; and receives research support from the Dr. Ali Rajput Endowment for for whole-genome association and population-based linkage – Parkinson’s Disease and Movement Disorders. S. Girard reports no disclo- analyses. Am J Hum Genet 2007;81:559 575. sures. P. Dion signed an acknowledgment of Research Agreement for his 7. Lee S, Emond MJ, Bamshad MJ, et al. Optimal unified involvement in the study conducted within the research agreement con- approach for rare-variant association testing with application cluded between Xenon Pharmaceutical and Dr. Rouleau. G. Rouleau to small-sample case-control whole-exome sequencing stud- has received research support from the CIHR Foundation Scheme, ALS ies. Am J Hum Genet 2012;91:224–237. Society of Canada, and ALS Association and signed a research agree- 8. Wu MC, Lee S, Cai T, Li Y, Boehnke M, Lin X. Rare- ment (via The Royal Institution for Advancement of Learning/McGill variant association testing for sequencing data with the University) regarding a separate essential tremor study conducted between his laboratory and Xenon Pharmaceutical (September 25, sequence kernel association test. Am J Hum Genet 2011; – 2014, to February 01, 2016). Go to Neurology.org/ng for full disclo- 89:82 93. sure forms. 9. Schrag A, Munchau A, Bhatia KP, Quinn NP, Marsden CD. Essential tremor: an overdiagnosed condition? Received April 14, 2017. Accepted in final form July 24, 2017. J Neurol 2000;247:955–959.

Neurology: Genetics 3 Ataxia-pancytopenia syndrome with SAMD9L mutations

Sorina Gorcenco, MD ABSTRACT Jonna Komulainen- Objective: We describe the neurologic, neuroradiologic, and ophthalmologic phenotype of 1 Ebrahim, MD Swedish and 1 Finnish family with autosomal dominant ataxia-pancytopenia (ATXPC) syndrome Karin Nordborg, MD and SAMD9L mutations. Maria Suo-Palosaari, MD, Methods: Members of these families with germline SAMD9L c.2956C.T, p.Arg986Cys, or PhD c.2672T.C, p.Ile891Thr mutations underwent structured interviews and neurologic and oph- Sten Andréasson, MD, thalmologic examinations. Neuroimaging was performed, and medical records were reviewed. PhD Previous publications on SAMD9L-ATXPC were reviewed. Johanna Krüger, MD, PhD Results: Twelve individuals in both families were affected clinically. All mutation carriers examined Christer Nilsson, MD, had balance impairment, although severity was very variable. All but 1 had nystagmus, and all but PhD 1 had pyramidal tract signs. Neurologic features were generally present from childhood on and Ulrika Kjellström, MD, progressed slowly. Two adult patients, who experienced increasing clumsiness, glare, and diffi- PhD culties with gaze fixation, had paracentral retinal dysfunction verified by multifocal electroreti- Elisa Rahikkala, MD, nography. Brain MRI showed early, marked cerebellar atrophy in most carriers and variable PhD cerebral periventricular white matter T2 hyperintensities. Two children were treated with hema- Dominik Turkiewicz, topoietic stem cell transplantation for hematologic malignancies, and the neurologic symptoms of MD, PhD one of these worsened after treatment. Three affected individuals had attention deficit hyperac- Mikael Karlberg, MD, tivity disorder or cognitive problems. Retinal dysfunction was not previously reported in individu- PhD als with ATXPC. Lars Nilsson, MD, PhD Conclusions: The neurologic phenotype of this syndrome is defined by balance or gait impairment, Jörg Cammenga, MD, nystagmus, hyperreflexia in the lower limbs and, frequently, marked cerebellar atrophy. Paracen- PhD tral retinal dysfunction may contribute to glare, reading problems, and clumsiness. Timely diagno- Ulf Tedgård, MD, PhD sis of ATXPC is important to address the risk for severe hemorrhage, infection, and hematologic Josef Davidsson, MD, malignancies inherent in this syndrome; regular hematologic follow-up might be beneficial. Neurol PhD Genet 2017;3:e183; doi: 10.1212/NXG.0000000000000183 Johanna Uusimaa, MD, PhD GLOSSARY Andreas Puschmann, ADHD 5 attention deficit hyperactivity disorder; ATXPC 5 ataxia-pancytopenia; mfERG 5 multifocal electroretinography; 5 MD, PhD HSCT hematopoietic stem cell transplantation. Ataxia-pancytopenia (ATXPC; MIM 159550) syndrome is an autosomal dominant disease with

Correspondence to early-onset gait and balance impairment, nystagmus, mild pyramidal signs, and marked cerebel- Dr. Puschmann: lar atrophy. Hematological abnormalities in ATXPC include pancytopenia, which may remain [email protected] subclinical but can cause severe infections or hemorrhages. Through different genetic mecha- nisms in blood cells, the hematopoietic phenotype may be reverted. Elimination of the germ- 1 Supplemental data line SAMD9L mutation by loss of 7(q) can result in myeloid malignancies. at Neurology.org/ng From the Sections of Neurology (S.G., C.N., A.P.), Pediatric Neurology (K.N.), Ophthalmology (S.A., U.K.), Pediatric Oncolocgy and Hematology (D.T., U.T., J.D.), Otorhinolaryngology (M.K.), and Hematology (L.N.), Department of Clinical Sciences, Skåne University Hospital, Lund University, Sweden; Department of Children and Adolescents (J.K.-E., J.U.), Department of Diagnostic Radiology (M.S.-P.), Department of Neurology (J.K.), and Department of Clinical Genetics, (E.R.), Oulu University Hospital; PEDEGO Research Unit (J.K.-E., E.R., J.U.), Medical Research Center Oulu (J.K.-E., M.S.-P., E.R., J.K., J.U.), Biocenter Oulu (J.K.-E., J.U.), and Research Unit of Clinical Neuro- science (J.K.), University of Oulu, Finland; Department of Hematology (J.C.), Linköping University Hospital and IKE Linköping University (J.C.), Sweden; and Division of Molecular Hematology (J.D.), Institution for Laboratory Medicine, Lund University, Sweden. Funding information and disclosures are provided at the end of the article. Go to Neurology.org/ng for full disclosure forms. The Article Processing Charge was funded by the authors. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Neurology.org/ng Copyright © 2017 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Figure 1 Pedigrees of the Swedish and Finnish families with Ataxia-pancytopenia syndrome and SAMD9L mutations

Standard symbols are used. Several members’ sex is disguised (diamond-shaped symbols) and some family members are not included for reasons of confidentiality or lack of consent for this study. Solid symbols indicate neurologic signs or symptoms. (A) Family 1 (Swedish): mut: heterozygote SAMD9L c.2956C.T, p.Arg986Cys mutations in DNA from buccal swabs. Wt: wild-type. Asterisk (*): 3 individuals also carry the rare variant c.689C.A p.Thr233Asn, located in trans in IV- 3 and IV-4. The c.2956C.T mutation leads to a gain of function of SAMD9L’s inhibitory functions on cell-cycle regula- tion, whereas c.689C.A was hypothesized to ameliorate this effect through a loss of function. (B) Family 2 (Finnish): mut: heterozygote SAMD9L c.2672T.C, p.Ile891Thr mutations. Wt: wild-type in DNA extracted from the peripheral blood.

ATXPC has so far been described in 5 fami- including optical coherence tomography, full-field electro- lies,1–5 including the Swedish family (F1) and retinography, and multifocal electroretinography (mfERG; appendix e-1 at Neurology.org/ng). Finnish family (F2) reported here. Recently, missense mutations in sterile alpha motif Standard protocol approvals, registrations, and patient – consents. This study was approved by the responsible regional domain containing protein 9-like (SAMD9L) ethical standards committees. Informed consent was obtained have been identified as the cause of ATXPC in from all participants. 4 of these families.1,5 Following the genetic discovery, we studied the neurologic pheno- RESULTS Data were compiled from 21 members of F1 and 6 members of F2. The pedigrees, individual type in F1 and F2. neurologic phenotype, and SAMD9L genotype are METHODS Affected and unaffected family members of both shown in figure 1 and table, and patient descriptions families were examined by a neurologist. A structured inter- in appendix e-1 and videos 1–3. DNA was obtained view and examination were performed including all clinical from 18 members of F1 and 6 members from F2. All – features previously described in ATXPC families,2 5 and the mutation carriers who were examined within this 6 Scale for the assessment and rating of ataxia. Neuroradiologic study showed neurologic signs except the 4-year-old images were compiled. Eye movement examination was mutation carrier, F1: V-2, who had no obvious filmed and reviewed by a neuro-otologist. Medical records were obtained whenever possible. Results of previous blood neurologic signs or symptoms according to her pa- cell counts were reviewed, and new blood samples were taken. rents. Individual F1: IV-5 declined participation. DNA was extracted using standard methods from the Thirteen family members were examined who were peripheral blood or buccal swabs.1 Some of the genetic anal- subsequently shown not to carry SAMD9L muta- yses and clinical (hematologic) information have previously tions; these did not have neurologic signs or been published.1 For this study, F1 was expanded; newly symptoms. included family members were tested with a custom droplet digital PCR TaqMan genotyping assay for the SAMD9L F1: III-4 presented at our neurology clinic at age c.2956C.T mutation (primers available on request). Two 53 years because of mild balance problems and a pos- patients underwent detailed ophthalmologic examination, itive family history (figure 1). No SAMD9L mutation

2 Neurology: Genetics Table Summary of the neurologic phenotype of ATXPC with SAMD9L mutations

ID: ref SAMD9L mutation(s) AE/AD Balance Dysm Dysart Nystag Strab Retina Pyram Sens Cog/beh Cytopenia MD

F1: I-2 ND 80a 11,55 NA NA NA NA 2 NA NA 2 NA 2

F1: II-3 ND 87a 111,53 11 GEN NA 1 NA NA 222

F1: III-41 c.2956C.T, p.Arg986Cys 64 1115 11 GEN 2 PRD AC, Bab, LEHR Vibr 2 P; Neu 2 (bnonhematopoietic cells/buccal swabs only)

F1: III-51 c.2956C.T, p.Arg986Cys 58a 1 NA 2 NA NA 1 NA NA 2 R, P, L; Inf MDS (56)

F1: IV-31 c.2956C.T, p.Arg986Cys; 38 122DBN 12 LEHR 22 2 2 c.689C.A, p.Thr233Asn

F1: IV-41 c.2956C.T, p.Arg986Cys; 33 122GEN, DBN 2 PRD LEHR 22 2 2 c.689C.A, p.Thr233Asn

F1: IV-51 c.2956C.T, p.Arg986Cys 28 NA NA NA NA NA NA NA 2 tR, tP, tL 2

F1: V-11 c.2956C.T, p.Arg986Cys 6 11(4c) 12 V 22 LEHR, S, Bab, AC 22 R, P, L MDS (4)

F1: V-21 c.2956C.T, p.Arg986Cys 4 2 NA NA NA NA NA NA NA 2 tP 2

F2: I-21 c.2672T.C, p.Ile891Thr 34 1(32) 12 GEN, DBN 2 NA LEHR, LLW 2 LI, MI, ADHD susp tP; Inf 2

F2: II-11 c.2672T.C, p.Ile891Thr 14 1(5) 22 2 2NA 22ADHD tR; Inf

F2: II-41 c.2672T.C, p.Ile891Thr 9 11(7) 12 H 2 NA LEHR, LLW, ppv 2 ADHD (4) R, P, L; Inf MDS (1,5)

I-12,3,5 c.3587G.C, p.Cys1196Ser 54 111adol 11 H 1 NR AC, Bab Vibr 2 NR 2

II-12,3,5 c.3587G.C, p.Cys1196Ser 10a 11child NR NR NR 1 NR NR NR 2 Neu AML (10)

II-22,3,5 c.3587G.C, p.Cys1196Ser 5a 11child NR NR NR 1 NR NR NR 2 R, P, L; Hem, Inf 2

II-32,3,5 c.3587G.C, p.Cys1196Ser 9a 11 NR NR V 1 H 1 NR NR NR 2 R, Neu; Inf 2

II-42,3,5 c.3587G.C, p.Cys1196Ser 54 11adol 11 V 1 H NR NR Bab Vibr 2 R 2

II-52,3,5 c.3587G.C, p.Cys1196Ser 7a 111child 1 NR V 1 NR NR NR 2 R, P, L AMMoL (7)

UW-AP II-35 c.2640C.A, p.His880Gln 79a 1145 to 50 NR NR NR NR NR NR NR dem 69 Hem 2

UW-AP II-45 c.2640C.A, p.His880Gln 85a 162 11 V 1 H NR NR Bab, LEHR NR 2 P 2

UW-AP III-35 c.2640C.A, p.His880Gln 60 130 11 H . V NR NR AC, LEHR NR 2 R, P, L 2

UW-AP III-55 c.2640C.A, p.His880Gln 16a 216 NR NR NR NR NR NR NR 2 R, P, L; Hem 2 erlg:Genetics Neurology: UW-AP III-65 c.2640C.A, p.His880Gln 55 1125 1 111 V 1 H NR NR AC, LEHR NR 2 NR 2

UW-AP III-85 c.2640C.A, p.His880Gln 50 125 1 NR V 1 H NR NR AC, Bab, LEHR NR 2 NR 2

UW-AP IV-15 c.2640C.A, p.His880Gln 32 218 12 V 1 H NR NR AC, LEHR NR 222

UW-AP IV-25 c.2640C.A, p.His880Gln 30 216 12 V 1 HNRNR2 NR 2 R, P, L 2

UW-AP IV-35 c.2640C.A, p.His880Gln 25 211 12 H NR NR AC, LEHR NR 2 R, P, L 2

Continued 3 5 was identified in DNA from the peripheral blood, but nonhematopoietic cells from buccal swabs carried heterozygous SAMD9L c.2956C.T (p.Arg986Cys) normocytic 2

gaze-evoked 1 not reported; 5 mutations. His mother and maternal grandmother 1 5 5 strabismus; t lower extremity had similar symptoms that had progressed slowly. 5 5

mutations have been None of the other members of this family sought

cognitive dysfunction or medical attention because of neurologic or ophthal- 5 mologic symptoms, but on direct inquiry, all patients R, P, LP d SAMD9L age at most recent examination neutropenia; NR 5 examined reported mild balance impairment, for – 5 5 2

leucopenia; LEHR example during sports, and problems with gaze

5 fixation. F2: II-4 was referred to a pediatric neurologist at the age of 7 years after a single seizure during sleep. 2 2 222 He had balance impairment, nystagmus, dysmetria, not assessed; Neu during childhood; Cog/beh 5 5

abs, because of mosaicism in blood cells. brisk Achilles tendon reflexes and muscle stiffness in in adolescence; AE/AD 5

ow aspiration revealed hypocellularity without evidence for the lower limbs. Symptoms had developed during the last months. His mother, F2: I-2, was investigated at the same time at the age of 32 years by a neurologist signs of sensory polyneuropathy on EMG; Strab dysmetria on nose-finger or heel-knee testing; GEN

5 for memory problems, nystagmus, lower limb weak- 5 recurrent or severe infection(s); L ness, and balance problems. His brother, F2: II-1, Babinski sign; child 5

5 had mild neurologic symptoms. The neurologic phe- memory impairment; NA pyramidal tract signs and lower limb weakness or foot deformity; R 5 notype within the family ranged from nearly normal 5 to marked disability. As described previously, a het- .

and from previous publications on families with ATXPC. erozygous SAMD9L c.2672T C (p.Ile891Thr) 1

sensory deficits; sPNP mutation was identified in DNA from the peripheral

5 blood from all three affected individuals.1 individual identifier; Inf

cerebellar dysarthria; Dysm The hematologic disorders of the family members 5 attention deficit hyperactivity disorder; adol 5 ataxia-pancytopenia; Bab HH NR NR NR NR Bab, LLW Bab, LLW sPNP sPNP H NR NR S, LEHR sPNP 1

5 have been described. Transient or permanent cyto- 5 1 1 1 V V penia was detected in 8 of 12 patients. Two children myelodysplastic syndrome; MI

spasticity; Sens underwent hematopoietic stem cell transplantation 5 5 (HSCT) because of myelodysplastic syndrome at 2 NR V planovalgus foot deformity; Pyram and 4.5 years. One of these patients (F1:V-1) had 5 dementia; Dysart

5 no neurologic symptoms prior to HSCT, but 3 ankle clonus; ADHD 1 11 1 111

5 months later, ataxic gait and nystagmus were noted. Similar treatment did not lead to neurologic worsen-

5 ing in F2: II-4, in whom follow-up MRI showed myelodysplasia; MDS 3 7 retinal dysfunction; S 5 severe; AC decreasing white matter abnormalities (figure 2B, 5 11 11 111, appendix e-1). hematopoietic stem cell transplantation; ID acute myelomonocytic leukemia; ATXPC

111 5 F1: III-4 had noticed difficulties with gaze fixation 5 5 during several years, and sometimes experienced diffi- culty reading or tendency to manual clumsiness, for down-beating nystagmus; dem who, however, had a very similar clinical picture. moderate; paracentral retinal dysfunction; ppv

4 example, when dining. Also, he sometimes found it 5 5 cumbersome to see in conditions with low ambient 11 5 lower limb weakness; MD light, such as during dawn, dusk, or in foggy weather 5 reference number; Retina

5 conditions. Examination by mfERG, which reflects cone function in the posterior pole, demonstrated severe hemorrhage; HSCT cytopenia; DBN a reduction of paramacular cone function (figure 5 5 reduced sensitivity to vibration.

mild (or, present); 2C). In F1: IV-4, there was similar but milder para- 5 recurrent; ref mutation(s) AE/AD Balance Dysm Dysart Nystag Strab Retina Pyram Sens Cog/beh Cytopenia MD thrombocytopenia; PCD central retinal dysfunction. Other patients from both 15 5 5 acute myelogenous leukemia; AMMoL families reported very similar symptoms. 5 but 1 previously reported family none; learning impairment; LLW SAMD9L NENENE 5 12 34 horizontal; Hem 3,5 5 – 25 vertical; Vibr 1 DISCUSSION The neurologic phenotype of ATXPC 5 Continued 5

4 4 4 with SAMD9L mutations is characterized by nystag- nystagmus; P

5 mus and slowly progressive balance impairment. These signs were found in almost all SAMD9L muta- Patient 1 ID: ref Patient 2 Patient 3 Table Testing of blood-DNA from this individual did not revealTreated a with mutation, daunomycin, but cytosine the arabinoside, mutation 6-thioguanine, was for identified suspected in malignancy, DNA died from of nonhematopoietic pancytopenia 6 cells weeks from later; buccal postmortem sw bone marr Deceased. After HSCT. behavioral abnormalities; Cytop c leukemia. Figures indicate age in years. or age at death; AML nystagmus; H hyperreflexia; LI Abbreviations: a b d identified in all transient; V This table includes information on all individuals with ATXPC from this study (Swedish family F1 and Finnish family F2) normochromic anemia; r Nystag tion carriers examined in this study (table). The

4 Neurology: Genetics Figure 2 Neuroimaging and multifocal electroretinography (mfERG)

(A and B) Neuroradiologic findings in affected members of family 1 (A) and family 2 (B). Age at examination is provided in parentheses. pHSCT, exam- ination performed posthematopoietic stem cell transplantation for hematologic malignancy. *Also carry the rare variant SAMD9L c.689C.Ain trans. Sagittal MRIs reveal cerebellar atrophy in all individuals examined in adult age. Patient F1: V-1 had cerebellar atrophy at 5 years of age, (A) but patient F2: II-1 did not have clear cerebellar atrophy at 5 years of age (not shown). Later, at the age of 12 years and 9 months F2: II-1 and F2: II-4 at the age of 7 years and 7 months, both had mild cerebellar atrophy (B). Bilateral hyperintense signal changes were visible in the frontoparietal periventricular white matter on T2-FLAIR images to a variable degree in all patients, except the 2 adults with the rare variant SAMD9L c.689C.A(AandB). Frontoparietal white matter changes of F2: II-4 and F2: II-1 decreased during childhood (B). F1: V-1 showed the most prominent white matter changes after hematologic malignancy and 3 months post-HSCT (A), but the white matter changes in F2: II-4, who also had hematologic malignancy and HSCT, decreased (B). For F1: V-1, the follow-up examinations 9 and 21 months post-HSCT showed unchanged widespread white matter abnormalities (not shown). FLAIR images revealed white matter abnormalities in the peritrigonal area in the individuals examined at 5 years and younger (B) that might be a sign of incomplete myelination normal to this age (see appendix e-1). Since there were also peritrigonal white matter signal abnormalities in the adult patient F2: I-2 (B), they might indicate vulnerability of these areas in SAMD9L-related ATXPC. CT of patient F1: II-3 showed cerebellar atrophy and diffuse periventricular hypodensities of the cerebral white matter (A). (C) mfERG of 2 patients, compared to normal findings in an adult, shows reduced paracentral function in both patients, to a different degree. N, normal; Y, significantly reduced function. Fundus photography and optical coherence tomography showed normal results (appendix e-1). ATXPC 5 ataxia-pancytopenia; FLAIR 5 fluid-attenuated inversion recovery; ms and MS, milliseconds.

Neurology: Genetics 5 neuro-otological findings indicate midline cerebellar and cellular homeostasis pathways and interacts with dysfunction. Balance impairment was clearly visible another tumor suppressor, p53, in the cell-cycle con- on examination,6 but it was mild or very mild in some trol system.10 Cerebellar Purkinje cells and retinal patients with only difficulty with tandem walking or cells might be particularly vulnerable to the effect of standing in a tandem position. Additional signs SAMD9L mutations. include mild pyramidal signs, strabismus, and decreased vibration sense. Neuroimaging revealed AUTHOR CONTRIBUTIONS marked cerebellar atrophy in all 8 patients and peri- Sorina Gorcenco: drafting/revising the manuscript for content, including medical writing for content, study concept or design, analysis or interpre- ventricular white matter changes in 6 of 8 (figure 2). tation of data, acquisition of data, and other: examination of 13 adult pa- Attention deficit hyperactivity disorder or cognitive tients, chart review. Jonna Komulainen-Ebrahim: drafting/revising the problems were noticed in 4 patients, but it remains manuscript for content, including medical writing for content, analysis uncertain if these are related to the disorder. or interpretation of data, and other: clinical data on 2 Finnish pediatric patients. Karin Nordborg: drafting/revising the manuscript for content, Ophthalmological symptoms or signs have not including medical writing for content, analysis or interpretation of data, previously been described for ATXPC with SAMD9L and other: clinical data on 1 Swedish pediatric patient and sibling. Maria mutations. Adult patients from both families reported Suo-Palosaari: drafting/revising the manuscript for content, including reading and focusing difficulties. mfERG showed medical writing for content, analysis or interpretation of data, and other: neuroradiologic data on 3 Finnish patients and interpretation of neurora- intact function in the most central area but marked diologic data. Sten Andréasson: drafting/revising the manuscript for con- paracentral (cone) dysfunction, which correlates with tent, including medical writing for content, analysis or interpretation of these visual symptoms. Similar ophthalmologic find- data, acquisition of data, obtaining funding, and other: ophthalmologic examination of 2 adult patients and evaluation of results. Johanna ings have been reported, for example, in patients Krüger: drafting/revising the manuscript for content, including medical 7 with SCA7. We suggest that some of the “atactic” writing for content, analysis or interpretation of data, and other: clinical signs in the upper extremities noted in patients data on 1 adult patient. Christer Nilsson: revising the manuscript for with SAMD9L mutations may in fact be caused content and other: clinical data on Swedish index patient (1 patient). Ulrika Kjellström: drafting/revising the manuscript for content, including by a decrease in visual control of hand movements, medical writing for content, analysis or interpretation of data, and other: rather than cerebellar ataxia. ophthalmologic examination of 2 adult patients. Elisa Rahikkala: revising In our as well as all previously reported families, all the manuscript for content, including medical writing for content, anal- carriers of pathogenic SAMD9L mutations developed ysis or interpretation of data, and other: genetic and genealogical studies of Finnish family, consenting. Dominik Turkiewicz: drafting/revising the neurologic signs. Expressivity was variable from indi- manuscript for content, including medical writing for content, and clin- viduals experiencing marked balance impairment to ical data on HSCT treatment of 1 Swedish pediatric patient. Mikael those who only noticed mild problems with balance Karlberg: drafting/revising the manuscript for content, including medical writing for content, analysis or interpretation of data, and other: review of or gaze fixation. Careful neurologic examination, videos of eye movement examination. Lars Nilsson: revising the manu- however, revealed clearly abnormal signs in all carriers script for content and other: contact with adult hematology patients. Jörg with complete penetrance. The hematologic disorders Cammenga: revising the manuscript for content and other: providing in members of both families have previously been data on hematologic phenotype. Ulf Tedgård: drafting/revising the man- uscript for content, including medical writing for content, analysis or 1 described in detail. In these and the other families, interpretation of data, study supervision or coordination, and other: clin- there was pancytopenia with a risk of severe hemor- ical data and contact with Swedish pediatric patient and sibling. Josef rhages or infections on the one hand, and myelodys- Davidsson: revising the manuscript for content, analysis or interpretation of data, acquisition of data, and other: genetic analyses of additional plastic syndrome on the other hand, but these Swedish family members. Johanna Uusimaa: revising the manuscript hematologic abnormalities were not noticed in all for content, study concept or design, analysis or interpretation of data, mutation carriers (table). study supervision or coordination, obtaining funding, and other: study One of SAMD9L’s roles is that of a tumor sup- supervision and coordination in Finland, interpretation of results, and ethical approvals (Finland). Andreas Puschmann: drafting/revising the 8 pressor or inhibitor of uncontrolled cell division. manuscript for content, including medical writing for content, study There is strong evidence that the disease-causing mu- concept or design, acquisition of data, study supervision or coordination, tations exert a gain-of-function mechanism, resulting obtaining funding, and other: examination of 13 patients and ethical in more pronounced inhibition of cell division, which approvals (Sweden). in blood cell lineages causes pancytopenia and genetic ACKNOWLEDGMENT pressure to eliminate the mutant copy of the The authors thank the members of the 2 families who participated in the 1 SAMD9L gene. Also in ataxia-telangiectasia, cerebel- study. Research nurse, Emma Pettersson, coordinated patient visits in lar degeneration co-occurs with immunologic defi- Lund, Sweden. ciencies, lymphopenia, and increased risk of lymphoid malignancies.9 While dysfunction of STUDY FUNDING DNA repair caused by mutations in ATM underlies This study was supported by ALF, Skåne University Hospital research grants, and Cancerfonden, all Sweden. some of the features of ataxia-telangiectasia, the cause

of cerebellar degeneration in ataxia-telangiectasia re- DISCLOSURE mains difficult to explain. Perhaps similar to S. Gorcenco has received research support from ALF, Sweden and Skåne SAMD9L, ATM also influences cell-cycle signaling University Hospital research grants (Sweden). J. Komulainen-Ebrahim

6 Neurology: Genetics has received research grants from the Alma and K.A. Snellman Founda- immunodeficiency, MDS and neurological symptoms. tion (Oulu, Finland) and Finnish Cultural Foundation, North Blood 2017;129:2266–2279. Ostrobothnia Regional Fund. K. Nordborg, M. Suo-Palosaari, 2. Li FP, Potter NU, Buchanan GR, Vawter G, Whang-Peng S. Andréasson, and J. Krüger report no disclosures. C. Nilsson has J, Rosen RB. A family with acute leukemia, hypoplastic received travel funding/speaker honoraria from the Movement Disorder anemia and cerebellar ataxia: association with bone mar- Society; has served on the editorial board of the Journal of Neurodegen- row C-monosomy. Am J Med 1978;65:933–940. erative Diseases; and has received publishing royalties from Norstedts. U. Kjellström and E. Rahikkala report no disclosures. D. Turkiewicz 3. Li FP, Hecht F, Kaiser-McCaw B, Baranko PV, Potter has received a research grant from Barncancerfonden. M. Karlberg has NU. Ataxia-pancytopenia: syndrome of cerebellar served as an unpaid scientific advisor to Synphora AB (Stockholm, ataxia, hypoplastic anemia, monosomy 7, and acute Sweden); has received payments from AstraZeneca AB (Sweden), Medical myelogenous leukemia. Cancer Genet Cytogenet Need AB (Sweden), and Interacoustics A/S (Denmark, Sweden, Norway, 1981;4:189–196. and Finland) for lectures and instructional courses; and has been a con- 4. Daghistani D, Curless R, Toledano SR, Ayyar DR. Ataxia- sultant for Sensorion Pharmaceuticals SA (France). L. Nilsson reports pancytopenia and monosomy 7 syndrome. J Pediatr 1989; no disclosures. J. Cammenga has received research support from 115:108–110. Barncancerfonden, ALF, Skåne University Hospital (Sweden), and 5. Chen DH, Below JE, Shimamura A, et al. Ataxia-pancy- Cancerfonden. U. Tedgård reports no disclosures. J. Davidsson has topenia syndrome is caused by missense mutations in received research support from ALF, The Swedish Society for Medical – Research, and The Swedish Childhood Cancer Foundation. J. Uusimaa SAMD9L. Am J Hum Genet 2016;98:1146 1158. reports no disclosures. A. Puschmann has received travel funding/speaker 6. Schmitz-Hubsch T, du Montcel ST, Baliko L, et al. honoraria from International Parkinson and Movement Disorder Scale for the assessment and rating of ataxia: develop- Society, Scandinavian Movement Disorder Society, International Associ- ment of a new clinical scale. Neurology 2006;66: ation of Parkinsonism and Related Disorders, and Swedish Movement 1717–1720. Disorder Society (SWEMODIS); serves as an editorial board member of 7. Hugosson T, Gränse L, Ponjavic V, Andréasson S. Mac- Parkinsonism & Related Disorders and BMC Neurology; has been a consul- ular dysfunction and morphology in spinocerebellar ataxia tant for Swedish National Board of Health and Welfare; has received type 7 (SCA 7). Ophthalmic Genet 2009;30:1–6. research support from ALF, Region Skåne, Trolle Wachtmeister Stiftelse, 8. Wang Q, Zhai YY, Dai JH, Li KY, Deng Q, Han ZG. and Parkinsonfonden; and has received institutional support from SAMD9L inactivation promotes cell proliferation via facil- MultiPark—a Strategic Research Environment at Lund University, and Skåne University Hospital research grants (Sweden) for neurogenetics itating G1-S transition in hepatitis B virus-associated hepa- – research. Go to Neurology.org/ng for full disclosure forms. tocellular carcinoma. Int J Biol Sci 2014;10:807 816. 9. Rothblum-Oviatt C, Wright J, Lefton-Greif MA, McGrath-Morrow SA, Crawford TO, Lederman HM. Received May 2, 2017. Accepted in final form July 21, 2017. Ataxia telangiectasia: a review. Orphanet J Rare Dis 2016;11:159. REFERENCES 10. Shiloh Y, Ziv Y. The ATM protein kinase: regulating the 1. Tesi B, Davidsson J, Voss M, et al. Gain-of-function cellular response to genotoxic stress, and more. Nat Rev SAMD9L mutations cause a syndrome of cytopenia, Mol Cell Biol 2013;14:197–210.

Neurology: Genetics 7 IBA57 mutations abrogate iron-sulfur cluster assembly leading to cavitating leukoencephalopathy

Akihiko Ishiyama, MD* ABSTRACT * Chika Sakai, PhD Objective: To determine the molecular factors contributing to progressive cavitating leukoence- Yuichi Matsushima, PhD phalopathy (PCL) to help resolve the underlying genotype-phenotype associations in the mito- Satoru Noguchi, PhD chondrial iron-sulfur cluster (ISC) assembly system. Satomi Mitsuhashi, MD, Methods: The subjects were 3 patients from 2 families who showed no inconsistencies in either PhD clinical or brain MRI findings as PCL. We used exome sequencing, immunoblotting, and enzyme Yukari Endo, MD, PhD activity assays to establish a molecular diagnosis and determine the roles of ISC-associated fac- Yukiko K. Hayashi, MD, tors in PCL. PhD Yoshiaki Saito, MD, PhD Results: We performed genetic analyses on these 3 patients and identified compound heterozy- Eiji Nakagawa, MD, PhD gosity for the IBA57 gene, which encodes the mitochondrial iron-sulfur protein assembly factor. Hirofumi Komaki, MD, Protein expression analysis revealed substantial decreases in IBA57 protein expression in myo- PhD blasts and fibroblasts. Immunoblotting revealed substantially reduced expression of SDHB, Kenji Sugai, MD, PhD a subunit of complex II, and lipoic acid synthetase (LIAS). Levels of pyruvate dehydrogenase a Masayuki Sasaki, MD, complex-E2 and -ketoglutarate dehydrogenase-E2, which use lipoic acid as a cofactor, were PhD also reduced. In activity staining, SDH activity was clearly reduced, but it was ameliorated in Noriko Sato, MD, PhD mitochondrial fractions from rescued myoblasts. In addition, NFU1 protein expression was also Ikuya Nonaka, MD, PhD decreased, which is required for the assembly of a subset of iron-sulfur proteins to SDH and LIAS Yu-ichi Goto, MD, PhD in the mitochondrial ISC assembly system. Ichizo Nishino, MD, Conclusions: Defects in IBA57 essentially regulate NFU1 expression, and aberrant NFU1 ulti- PhD mately affects SDH activity and LIAS expression in the ISC biogenesis pathway. This study pro- vides new insights into the role of the iron-sulfur protein assembly system in disorders related to mitochondrial energy metabolism associated with leukoencephalopathy with cavities. Neurol Correspondence to Genet 2017;3:e184; doi: 10.1212/NXG.0000000000000184 Dr. Ishiyama: [email protected] GLOSSARY BN 5 blue native; cDNA 5 complementary DNA; CS 5 citrate synthase; ISC 5 iron-sulfur cluster; KDH 5 ketoglutarate dehydrogenase; LA 5 lipoic acid; LIAS 5 lipoic acid synthetase; PAGE 5 polyacrylamide gel electrophoresis; PCL 5 pro- gressive cavitating leukoencephalopathy; PDH 5 pyruvate dehydrogenase.

Progressive cavitating leukoencephalopathy (PCL) is an autosomal recessive neurodegenerative disorder, characterized by leukoencephalopathy with cavities showing episodic clinical deterio- ration, and vascular permeability in the corpus callosum, centrum semiovale, or cerebellum on brain MRI.1 The pathogenesis is currently not well understood. White matter abnormalities occur in patients harboring mutations in genes encoding complex I–III assembly factors acting in mitochondrial energy metabolism.2,3 Iron-sulfur clusters (ISCs) participate in electron trans- fer in the mitochondrial respiratory chain and act as sulfur donors in the synthesis of lipoic acid Supplemental data at Neurology.org/ng *These authors contributed equally to this article. From the Department of Child Neurology (A.I., Y.S., E.N., H.K, K.S., M.S.), National Center Hospital; Department of Neuromuscular Research (A.I., S.N., S.M., Y.E., Y.K.H., I. Nonaka, I. Nishino.), National Institute of Neuroscience; Department of Mental Retardation and Birth Defect Research (C.S., Y.M., Y.-i.G.), National Institute of Neuroscience; Department of Radiology (N.S.), National Center Hospital, National Center of Neurology and Psychiatry, Tokyo; Department of Pharmacology (A.I.), Interdisciplinary Graduate School of Medicine and Engineering, University of Yamanashi; and Department of Pathophysiology (Y.K.H), Tokyo Medical University, Japan. Funding information and disclosures are provided at the end of the article. Go to Neurology.org/ng for full disclosure forms. The Article Processing Charge was funded by the authors. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Neurology.org/ng Copyright © 2017 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 (LA) by mitochondrial lipoic acid synthetase whole-exome sequencing on 2 affected (II-2a and II-3a) and 3 (LIAS).4–8 The [2Fe-2S] cluster, synthesized nonaffected members (I-1a, I-2a, and II-1a) of family A, and 1 affected (II-1b) and 2 nonaffected members (I-1b and I-2b) of on the ISC unit, is converted by a protein family B (figure 2A). Genetic and bioinformatics analysis was complex consisting of ISCA1, ISCA2, and performed according to the procedure as described.22–24 Candi- IBA57 into the [4Fe-4S] form, which is in- date mutations were validated by Sanger sequencing on an ABI serted into the apoprotein via complexes I Prism 3130 DNA Analyzer (Applied Biosystems). Within-family IBA57 variant segregation was analyzed using primers designed to and II, LIAS, and aconitase encoded by amplify exons corresponding to the sequence under accession – NFU1, BOLA3, and IND1.9 14 number NG_042231—PCR conditions, and primer sequences Neuroradiologically diverse phenotypes of are available upon request. white matter abnormalities, associated with Cell culture. Patient-derived cells were cultured in Dulbecco ISC systems, have been reported. Patients modified Eagle medium/F-12 (GIBCO BRL, USA) supple- mented with heat-inactivated fetal bovine serum (20% [vol/vol] with BOLA3 or NFU1 mutations present with for myoblasts and 10% for fibroblasts; GIBCO BRL). encephalopathy and neurologic regression, Establishment of immortalized patient-derived myoblasts. 15–18 leading to death in infancy. Patients with Patient-derived and control myoblasts were transfected IBA57 mutations show severe encephalopathy with pEF321-T plasmid (a kind gift from Dr. Sumio Sugano, to hereditary spastic paraplegia.19–21 However, University of Tokyo) and serially cultured for more than 10 the genotype-phenotype correlation remains population doublings, until morphological alteration was observed.25 elusive with respect to ISC biogenesis- Preparation of mitochondrial fractions. Mitochondrial frac- associated factors. We evaluated the factors tions from skeletal muscle and cultured patient-derived cells were that might play a role in ISC biogenesis path- prepared as reported,26 with slight modification. way regulation, resulting in cavitating Immunoblotting. Mitochondrial fractions were subjected to leukoencephalopathy. sodium dodecyl sulfate–polyacrylamide gel electrophoresis (PAGE). Thirty micrograms of mitochondrial protein was sepa- METHODS Patients. Patients with PCL were selected based rated on 4%–12% Bis-Tris gradient gels (Life Technologies, on clinical and radiologic examinations. They showed episodic, Carlsbad, CA) and transferred to polyvinylidene fluoride mem- acute onset of irritability or neurologic deficits between branes as described.27 The membranes were incubated with pri- 11 months and 2 years of age, followed by an intermittent mary antibodies against IBA57 (Sigma, St. Louis, MO), VDAC1 increase in lactate and clinical deterioration. Brain MRI showed (Abcam, Cambridge, UK), LA (Calbiochem, San Diego, CA), patchy leukoencephalopathy with cavities and diffusion restric- NDUFS1, NDUFS8, NDUFV1, SDHB, and UQCRFS1 (all tion in the affected region (figure 1). Patients 1 (II-2a) and 2 from GeneTex, Irvine, CA). Protein levels of IBA57 and ISC (II-3a) are siblings. All clinical materials used in this study were were quantified with ImageJ software (National Institutes of obtained with written informed consent. All experiments were Health, Bethesda, MD). approved by the Ethical Committee of the National Center of Measurement of respiratory enzymatic activity. Enzymatic Neurology and Psychiatry. activities of mitochondrial respiratory complexes I–V and citrate Exome sequencing and bioinformatics analysis. To identify synthase (CS) were measured in mitochondrial fractions prepared the genetic cause of recessive inherited PCL, we performed from the patients’ skeletal muscle and patient-derived fibroblasts.

Figure 1 Illustrations of the characteristic MRI pattern in patients with progressive cavitating leukoencephalopathy

MRIs of patient 1 (A–D) obtained at the disease onset of 11 months. Axial fluid-attenuated inversion recovery hyperintensity images (A), diffusion-weight images (B), apparent diffusion coefficient map images (C), and sagittal T1-weighted images (D) are shown. Abnormal high intensities are noted in the cerebral periventricular and subcortical white matter, which contained multiple cavities (A and B; arrow), as well as in the corpus callosum (D; arrowhead). Diffusion restriction is present in the solid part of the abnormal white matter (C).

2 Neurology: Genetics Figure 2 Genetic analysis of recessively inherited leukoencephalopathy and IBA57 protein expression

(A) Pedigree structures for recessive leukoencephalopathy family A and family B, with filled symbols indicating affected in- dividuals. *Individuals who were analyzed by exome sequencing. IBA57 genotypes are shown below the individuals in the pedigree. (B) IBA57 expression was analyzed by immunoblotting. C1, C2: controls; P1–3: patients. Mitochondrial fractions prepared from patient-derived myoblasts and fibroblasts were analyzed via immunoblotting. IBA57 levels were significantly lower in all patients than in the controls. Porin was used as a loading control. The experiments were performed in triplicate and normalized to porin levels. (C) IBA57 protein expression in mitochondrial fractions prepared from myoblasts. Error bars indicate SDs (**p , 0.005, Student t test).

The assays for complexes I–IV and CS were conducted as II, UQCRC2 as complex III, MTCO1 as complex IV, and described.27 The assay for complex V was performed according to ATP5B as complex V. Morava et al.28 Expression vector construction and transfection. For mam- High-resolution clear native PAGE and enzyme activity malian expression vector construction, full-length IBA57 was staining. High-resolution clear native PAGE and activity amplified from complementary DNA (cDNA) prepared from the staining were performed as described.29 In brief, isolated mito- control subject by PCR. The PCR product was cloned into chondria were solubilized with Native PAGE Sample pEBMulti-Pur (Wako, Osaka, Japan) and verified by Sanger Buffer (Invitrogen) containing 0.3% n-dodecyl-b-D-maltoside sequencing. Mammalian expression vector, pEBMulti-Pur con- (Dojindo, Kumamoto, Japan). Thirty micrograms of protein struct containing IBA57 cDNA, was transfected into immortal- was applied to NativePAGE Novex 3%–12% Bis-Tris Gel ized myoblasts using Lipofectamine LTX Reagent (Thermo (Invitrogen). Native PAGE Buffer (Invitrogen) was used as the Fisher Scientific, Waltham, MA). anode buffer, and Native PAGE Buffer containing 0.02% n-dodecyl-b-D-maltoside and 0.05% deoxycholate was used as RESULTS Clinical findings. The detailed clinical and the cathode buffer. laboratory findings are summarized in table e-1 at For in-gel catalytic activity assays, the gels were incubated as Neurology.org/ng. Characteristic symptoms were described.29 intermittent episodic clinical deterioration and an Blue native PAGE and immunoblotting. Blue native infection-associated increase in lactate. Psychomotor (BN)-PAGE and immunoblotting were performed as development was normal until clinical onset between described.28 In brief, isolated mitochondria were solubilized 11 months and 2 years of age. The patients showed with 0.5% (w/v) n-dodecyl-b-D-maltoside. Protein (30 mg) was applied to NativePAGE Novex 3%–12% Bis-Tris Gel. marked irritability and lost acquired skills due to Immunoblotting was conducted as described above. Primary repeated infections. These episodes led to psycho- antibodies targeted NDUFA9 as complex I, SDHA as complex motor regression with axial hypotonia accompanied

Neurology: Genetics 3 Figure 3 Blue native polyacrylamide gel electrophoresis of respiratory chain complexes in the 3 patients (P1, P2, and P3) and controls (C1 and C2)

(A) Blue native polyacrylamide gel electrophoresis (PAGE) was performed using mitochondrial fractions prepared from myoblasts. No patient showed severely reduced assembly of respiratory chain complexes as compared to the control. (B) In-gel activity staining and Western blotting of complexes I and II. In-gel activity staining of complex II after separation by high-resolution clear native (hrCN)-PAGE using mitochondrial fractions prepared from the myoblasts. In parallel, complexes I, II, and V were detected by Western blotting after their separation by hrCN-PAGE. All patients showed clearly reduced complex II activity when compared with the control, whereas no such decrease was detected by Western blotting. The activity and amount of assembled complex I were not significantly different from those of controls. (C, D) The experiments were performed in triplicate and normalized to complex V (n.s. 5 not significant, **p , 0.005, Student t test).

by spasticity of the legs. Treatment with a mito- low apparent diffusion coefficient values was chondrial cocktail recovered some of the lost skills observed in the noncavitated abnormal white during periods of remission. Plasma and CSF lactate matter. levels were slightly increased intermittently, and Genetic analyses. In total, 13 candidate genes from serum amino acid and urine organic acid levels were family A and 20 candidate genes from family B were normal based on gas chromatography/mass spec- assessed based on the assumption of autosomal reces- trometry. Muscle biopsy showed normal histology sive inheritance and with reference to the dbSNP135, (data not shown). The last neurologic examination at 1000 Genomes Project database, NHLBI Exome the age of 6–7 years varied between normal and mild Variant Server, and human genetic variation database cognitive and acquisitive deficiency, assessed as for Japanese Genetic Variants databases. Among all a score of 78–82 and 73–85, respectively, according extracted variants, we identified only 1 associated to the Kaufmann Assessment Battery for Children, gene, IBA57, common to both families, with com- Second Edition. pound heterozygous mutations of c.323A.C and MRI abnormalities. Brain MRI in all 3 patients re- c.940C.T in family A and c.323A.C and vealed progressive leukoencephalopathy with exten- c.150C.A in family B (RefSeq accession number sive signal abnormalities in the periventricular NM_001010867). These variants in IBA57, encod- cerebral white matter and corpus callosum (figure 1, ing the mitochondrial iron-sulfur protein assembly figure e-1). The abnormal white matter in the factor, are predicted to lead to a missense and a non- genu and splenium of the corpus callosum showed sense mutation (figure 2A). The mutations were partial cavitation. The basal ganglia, cerebellum, confirmed by Sanger sequencing and the compound and brain stem were speared. Magnetic resonance heterozygosity by transcript analysis. In silico analyses spectroscopy showed an increased lactate concentra- of mutation function demonstrated that p.Tyr108Ser tion in the white matter. Diffusion restriction with is predicted as damaging based on scale-invariant

4 Neurology: Genetics feature transform and Polymorphism Phenotyping To examine the effects of the IBA57 mutations v2. The amino acids involved are moderately to further, we assessed the expression level of each mito- highly conserved. chondrial [4Fe-4S] protein, including 5 subunits of mitochondrial respiratory chain complexes Functional studies. IBA57 protein expression was eval- (NDUFS1, NDUFS7, NDUFS8, and NDUFV1 in uated by immunoblotting using mitochondrial frac- respiratory complex I, and SDHB in respiratory com- tions prepared from patient-derived myoblasts and plex II), ACO2, and LIAS. Protein expression of the 4 fibroblasts. IBA57 protein was substantially decreased subunits in complex I and ACO2 were not substan- in the patients; a small amount of full-length IBA57, tially reduced in patient vs control myoblasts (figure but no truncated IBA57, was detected in the patients 4A). The levels of LIAS and the SDHB, but not the (figure 2, B and C). This result indicated that the SDHA subunit of complex II, were lower in patient mutations severely affect the level of IBA57 protein. myoblasts (figure 4, B–D). To examine the effect of Given that IBA57 is a mitochondrial iron-sulfur the defect in IBA57 on LIAS function, an anti-LA protein assembly factor, we first tested the effects of antibody was used to detect LA bound to the key the mutations on the respiratory complexes by mea- mitochondrial enzymes pyruvate dehydrogenase suring enzyme activity and analyzing complex (PDH) complex and a-ketoglutarate dehydrogenase assembly by BN-PAGE (figure 3A) in myoblast (KDH), which use LA as a cofactor to form LA- mitochondrial fractions. Although the activities of PDH-E2 and LA-KDH-E2, respectively. LA-PDH- these complexes were slightly lowered, the differences E2 and LA-KDH-E2 proteins were decreased in all were not significant, and there were no differences in patient-derived myoblasts (figure 4B), suggesting that their amounts (figure 3A). As we did not detect clear a defect in LIAS significantly decreases LA synthesis decreases in activity or assembly, we next used activity in patient myoblasts. To evaluate the potential effects staining to determine any change in the respiratory of IBA57 deficiency on other ISC components, we complexes. The staining intensity for SDH was assessed NFU1, which inserts [4Fe-4S] into the apo- weaker in all patients than in controls, indicating protein after [2Fe-2S] is converted to [4Fe-4S] by a defect in SDH activity (figure 3, B and C). By IBA57. NFU1 protein was significantly decreased in contrast, there was no clear decrease in complex I all patients, although no mutation in NFU1 was de- activity staining (figure 3, C and D). tected in any of the cases (figure 4, B and E).

Figure 4 Expression of mitochondrial [4Fe-4S] proteins

(A, B) Mitochondrial [4Fe-4S] protein expression levels were analyzed by immunoblotting. Mitochondrial fractions prepared from myoblasts were analyzed. Levels of mitochondrial [4Fe-4S] proteins in all patients were comparable with those in con- trols (A), except lipoic acid synthetase (LIAS) and SDHB, which were clearly lower than in controls. Immunoblotting for lip- oylated residues in mitochondrial fractions prepared from the myoblasts. All patients showed a severely reduced signal for the bands corresponding to the lipoylated residues pyruvate dehydrogenase (PDH)-E2 and a-ketoglutarate dehydrogenase (KDH)-E2 (B). NFU1 protein expression in mitochondrial fractions prepared from myoblasts as analyzed by immunoblotting (B). NFU1 protein level was clearly lower in all patients than in controls. The experiments were performed in triplicate and normalized to porin (C–E). Error bars indicate SDs (**p , 0.005, *p , 0.05, Student t test).

Neurology: Genetics 5 A phenotype rescue experiment was used to test proper functioning is of vital importance to all eu- whether forced expression of wild-type IBA57 in karyotic organisms.4–8 All of our patients harbored immortalized patient-derived myoblasts would com- compound heterozygous mutations in IBA57, which plement the SDH activity defect. We confirmed that encodes a protein involved in [4Fe-4S] cluster IBA57 protein expression was lower in patient- assembly in the ISC assembly machinery. To date, 3 derived than in control myoblasts than in those of phenotypes linked to IBA57 mutations have been controls, and then transduced an empty expression described.19–21 Homozygous mutations in IBA57 vector, pEBMulti-Pur (Wako), or a pEBMulti-Pur were first described in association with lethal construct containing full-length, wild-type IBA57 encephalopathy, lactic acidosis, and profound respi- cDNA into the immortalized patient-derived myo- ratory complex I and II deficiencies in skeletal muscle, blasts. IBA57 protein expression was confirmed in and brain MRI showed frontoparietal polymicrogyria mitochondrial fractions prepared from the rescued with white matter abnormalities and hypoplasia of the myoblasts (figure 5A), and SDH activity staining medulla oblongata.19 A mild phenotype associated showed nearly the same intensity as observed in con- with homozygous mutations in IBA57 was reported trols (figure 5B). in a highly consanguineous family with hereditary spastic paraplegia accompanied by optic atrophy and DISCUSSION Mitochondrial ISC assembly systems peripheral neuropathy, with complex I and II defi- are highly conserved from yeast to humans, and their ciencies in lymphoblastoid cells; MRI indicated mild

Figure 5 IBA57 protein expression and activity staining of complex II in genetically rescued cells

(A) Immortalized patient-derived myoblasts were transduced with empty vector or a wild-type IBA57-expressing construct, and rescued cells were established. iC1, iC2 5 immortalized control myoblasts; iP1 5 immortalized patient 1–derived myoblasts; iP1-vector 5 immortalized patient 1–derived myoblasts transfected with empty vector; Rescued 5 immortal- ized patient 1–derived myoblasts expressing wild-type IBA57; AS 5 activity staining; WB 5 Western blotting. IBA57 protein expression was confirmed by immunoblotting in mitochondrial fractions prepared from rescued cells. Porin was used as a loading control. Activity staining of complex II was performed to examine its restoration in rescued cells after high-res- olution clear native (hrCN) polyacrylamide gel electrophoresis (PAGE). (B) In parallel, complexes II and V were detected by Western blotting after hrCN-PAGE. The band intensity of complex II was normalized to that of complex V. (C) Third step of the mitochondrial iron-sulfur cluster (ISC) biosynthesis pathway. Dedicated maturation of mitochondrial [4Fe-4S] proteins by specific ISC-targeting factors. Yellow boxes highlight ISC proteins whose genes are mutated in human ISC-related disease. LIAS 5 lipoic acid synthetase; PCL 5 progressive cavitating leukoencephalopathy.

6 Neurology: Genetics corpus callosum and cerebellar atrophy, but no enzyme ACO2, another reported target of IBA57 that involvement of the white matter except for cystic contains 1 [4Fe-4S] cluster, was not reduced.8,9,19–21 cavitation.20 A third case of homozygous mutations in This biochemical profile, i.e., an observed defect in IBA57 was described as severe leukodystrophy, LA synthesis without a corresponding defect in characterized by neurologic deterioration with com- ACO2, appears to be inconsistent with the changes plex I and II deficiencies, leading to death at 17 to mitochondrial [4Fe-4S] proteins identified in pre- months.21 These previous reports suggest that IBA57 vious reports on IBA57 deficiencies.19–21 The present deficiency affects the mitochondrial respiratory chain profile is more reminiscent of the phenotype related complexes I and II. to alterations in NFU1, rather than that related Defects in the ISC biosynthesis pathway lead to IBA57 mutations. abnormal function of oxidative phosphorylation as NFU1 is characterized as a late-targeting factor of well as deficiencies in many proteins involved in the the ISC assembly machinery and is specifically enzyme-bound cofactor lipoate and intermediary required for the proper assembly of a subset of metabolic pathways.11–14 The ISC biosynthetic pro- [4Fe-4S] proteins to respiratory complexes I, II, cess is initiated by the synthesis of [2Fe-2S] clusters and LIAS.16–18 We determined the effect of IBA57 on a scaffold protein, followed by the generation of mutation on NFU1 to evaluate potential links of [4Fe-4S] clusters; their into appropriate IBA57 deficiency to other aspects of the ISC assem- apoproteins is facilitated by the coordinated action bly machinery. The NFU1 protein level was also of additional ISC maturation factors, including tar- decreased in patients, corroborating that IBA57 reg- geting factors such as NFU1 and BOLA3. Along with ulates the expression of NFU1, which subsequently IBA57, the ISC proteins, ISC1 and ISC2, help to affects the expression of lipoate-containing protein convert [2Fe-2S] into [4Fe-4S] clusters. Defects in as well as respiratory complex II (figure 5C). There IBA57 have been reported to lead to a general mito- have been 5 reports describing a total of 22 patients chondrial [4Fe-4S] protein deficiency, such as in with NFU1 mutations (table e-2). Seven of 9 cases respiratory complexes I and II.13,14 In our cases, there with brain MRI data presented with leukoencephal- were no differences in the respiratory complexes in opathy in the periventricular white matter and cor- terms of overall activity and complex assembly, pus callosum, with partial cystic degeneration or although complex activities were slightly lower in pa- cavitation, characterized by neurologic regression, tients. However, SDH activity was substantially lactic acidosis, and hyperglycinemia, leading to weaker in all patients than controls. These results death within 2.5 years.16,18,35–38 Navarro-Sastre suggested that IBA57 defect affected SDH function, et al.,18 Nizon et al.,38 and Invernizzi et al.36 reported whereas there was no effect on protein expression that the activity of complex II was significantly involved in ISC biosynthesis. Therefore, IBA57 decreased, while that of complex I was just below defect might reduce the efficiency of the ISC biosyn- normal. Biochemical profiles of patients with muta- thesis process. Furthermore, detailed assessment of tions in NFU1 also suggest that NFU1 mainly af- complex II function of mitochondrial [4Fe-4S] pro- fects complex II activity. The neuroimaging features teins by Western blotting revealed that the SDHB and biochemical profiles in our cases are reminiscent subunit of complex II was the only factor decreased of those reported in patients harboring mutations in in patient myoblasts, with no severe reductions in the NFU1. This suggests that the IBA57 mutation af- subunits of complexes I and III. These results suggest fects NFU1 protein expression and function. To that defects in IBA57 mainly affect the expression of date, the exact role of IBA57 and its functional rela- SDHB. tionship to NFU1 have remained unclear. Our re- Defects in the ISC biosynthesis pathway have sults provide new insights into the pathogenic a negative effect not only on the respiratory com- mechanisms in that defects in IBA57 essentially plexes but also on lipoate-containing protein, likely cause decreased respiratory complex II and LIAS as a result of deficient maturation of LIAS. LIAS con- activities through the aberrant function of NFU1, tains 2 [4Fe-4S] clusters that are essential for the syn- leading to white matter abnormalities with disturbed thesis of LA, a cofactor of PDH and a-KDH. In mitochondrial energy metabolism. PDH and a-KDH, which comprise certain subunits, This study demonstrated that IBA57 plays essen- measurement of the lipoylation status of the E2 sub- tial roles in the biosynthesis of ISC proteins to regu- units serves as an indirect but reliable estimate of the late mitochondrial energy metabolism. IBA57 gene iron-sulfur maturation efficiency of LIAS.30–34 In our mutations resulted in decreased expression of SDHB cases, detection of PDH and a-KDH revealed a severe as well as deficiency of LIAS through aberrant NFU1 decrease in LA in patient myoblasts, providing further function. This new understanding of the molecular evidence of disturbance in the ISC biosynthesis path- basis of PCL pathophysiology should contribute to way. However, the mitochondrial citric acid cycle a better definition of the phenotypic spectrum and

Neurology: Genetics 7 provide critical information to develop potential ther- Resonance of Myelination and Myelin Disorders. Berlin, apeutic strategies for PCL. Heidelberg: Springer-Verlag; 1989:20–24. 4. Lill R, Hoffmann B, Molik S, et al. The role of mitochon- AUTHOR CONTRIBUTIONS dria in cellular iron-sulfur protein biogenesis and iron – Akihiko Ishiyama and Chika Sakai: data acquisition, analysis and inter- metabolism. Biochim Biophys Acta 2012;1823:1491 1508. pretation, and manuscript drafting and editing. Yuichi Matsushima, 5. Booker SJ, Cicchillo RM, Grove TL. Self-sacrifice in rad- Satoru Noguchi, and Satomi Mitsuhashi: bioinformatics analysis. ical S-adenosylmethionine proteins. Curr Opin Chem Biol Yukari Endo and Yukiko K. Hayashi: exome data interpretation. 2007;11:543–552. Yoshiaki Saito, Eiji Nakagawa, Hirofumi Komaki, and Kenji Sugai: 6. Johnson DC, Dean DR, Smith AD, Johnson MK. Struc- patient clinical information collection. Masayuki Sasaki and Noriko ture, function, and formation of biological iron-sulfur clus- Sato: MRI review and clinical and imaging data analysis. Ikuya ters. Annu Rev Biochem 2005;74:247–281. Nonaka, Yu-ichi Goto, and Ichizo Nishino: study supervision and 7. Beinert H. Iron-sulfur proteins: ancient structures, still full manuscript preparation and editing. All authors were responsible for of surprises. J Biol Inorg Chem 2000;5:2–15. data acquisition and analysis. 8. Lill R. Function and biogenesis of iron-sulphur proteins. Nature 2009;460:831–838. STUDY FUNDING 9. Sheftel AD, Wilbrecht C, Stehling O, et al. The human This study was supported partly by the Intramural Research Grant (29-4 mitochondrial ISCA1, ISCA2, and IBA57 proteins are to Ichizo Nishino) for Neurological and Psychiatric Disorders of NCNP and Research and Development Grants for Practical Research Project for required for [4Fe-4S] protein maturation. Mol Biol Cell – Rare/Intractable Diseases (17ek0109285h0001 to Ichizo Nishino) from 2012;23:1157 1166. the Japan Agency for Medical Research and Development, AMED. 10. Mühlenhoff U, Richter N, Pines O, et al. Specialized function of yeast Isa1 and Isa2 proteins in the maturation DISCLOSURE of mitochondrial [4Fe-4S] proteins. J Biol Chem 2011; A. Ishiyama and C. Sakai report no disclosures. Y. Matsushima has 286:41205–42116. received research support from the Japan Society for the Promotion of 11. Rouault TA, Tong WH. Iron-sulfur cluster biogenesis and Science (JSPS). S. Noguchi holds patent WO 2010/131712A1; has human disease. Trends Genet 2008;24:398–407. received research support from Novartis Pharmaceuticals, Japan; has 12. Rouault TA. Biogenesis of iron-sulfur clusters in mamma- received governmental support from the Japan Agency for Medical lian cells: new insights and relevance to human disease. Dis Research and Development (AMED); and has received institutional sup- Model Mech 2012;5:155–164. port from the National Center of Neurology and Psychiatry (NCNP). 13. Stehling O, Wilbrecht C, Lill R. Mitochondrial iron-sulfur S. Mitsuhashi and Y. Endo report no disclosures. Y.K. Hayashi serves protein biogenesis and human disease. Biochimie 2014; as an editorial board member of Neuromuscular Disorders and Journal of – Physiological Sciences and received institutional support from the NCNP, 100:61 77. the Japan Society for the Promotion of Science (JSPS), The Promotion 14. Ye H, Rouault TA. Human iron-sulfur cluster assembly, and Mutual Aid Corporation for Private Schools of Japan, and Tokyo cellular iron homeostasis, and disease. Biochemistry 2010; Medical University. Y. Saito reports no disclosures. E. Nakagawa received 49:4945–4956. institutional support from the NCNP. H. Komaki has received research 15. Liu Y, Qi W, Cowan JA. Iron-sulfur cluster biosynthesis: support from Sanofi, Daiichi Sankyo, Taiho, Pfizer, and Nippon functional characterization of the N- and C-terminal do- Shinyaku and has served as a consultant for PTC Therapeutics. K. Sugai mains of human NFU. Biochemistry 2009;48:973–980. has served on the editorial board of the Journal of the Japanese Epilepsy 16. Cameron JM, Janer A, Levandovskiy V, et al. Mutations Society. M. Sasaki, N. Sato, and I. Nonaka report no disclosures. Y.-i. in iron-sulfur cluster scaffold genes NFU1 and BOLA3 Goto serves on the editorial board of Mitochondrion; has received cause a fatal deficiency of multiple respiratory chain and governmental support from the Japan Agency for Medical Research and Development (AMED) and the Ministry of Health, Labour and 2-oxoacid dehydrogenase enzymes. Am J Hum Genet – Welfare of Japan; and has received institutional support from the NCNP. 2011;89:486 495. I. Nishino has received travel funding/speaker honoraria from Sanofi, 17. Haack TB, Rolinski B, Haberberger B, et al. Homozygous Daiichi Sankyo, and Japan Blood Products Organization; serves on the missense mutation in BOLA3 causes multiple mitochon- editorial boards of Neuromuscular Disorders, Therapeutic Advances in drial dysfunctions syndrome in two siblings. J Inherit Neurological Disorders, Journal of the Neurological Sciences, Journal of Metab Dis 2013;36:55–62. Neuromuscular Diseases, Skeletal Muscle, and Neurology and Clinical 18. Navarro-Sastre A, Tort F, Stehling O, et al. A fatal mito- Neuroscience; holds a patent for Efficacious agents to pathologic condition chondrial disease is associated with defective NFU1 func- due to GNE protein dysfunction; has served on the scientific advisory tion in the maturation of a subset of mitochondrial Fe-S boards of Ultragenyx and Novartis Pharmaceuticals; and has received proteins. Am J Hum Genet 2011;89:656–667. research support from Sanofi, Daiichi Sankyo, and Astellas, in addition to governmental support from the AMED and institutional support from 19. Ajit Bolar N, Vanlander AV, Wilbrecht C, et al. Mutation the NCNP. Go to Neurology.org/ng for full disclosure forms. of the iron-sulfur cluster assembly gene IBA57 causes severe myopathy and encephalopathy. Hum Mol Genet Received May 9, 2017. Accepted in final form July 27, 2017. 2013;22:2590–2602. 20. Lossos A, Stümpfig C, Stevanin G, et al. Fe/S protein REFERENCES assembly gene IBA57 mutation causes hereditary spastic 1. Naidu S, Bibat G, Lin D, et al. Progressive cavitating paraplegia. Neurology 2015;84:659–667. leukoencephalopathy: a novel childhood disease. Ann 21. Debray FG, Stümpfig C, Vanlander AV, et al. Mutation of Neurol 2005;58:929–938. the iron-sulfur cluster assembly gene IBA57 causes fatal 2. Morató L, Bertini E, Verrigni D, et al. Mitochondrial infantile leukodystrophy. J Inherit Metab Dis 2015;38: dysfunction in central nervous system white matter disor- 1147–1153. ders. Glia 2014;62:1878–1894. 22. YukariEndoY,NoguchiS,HaraY,etal.Dominant 3. Van der Knaap MS, Valk J. Classification of myelin dis- mutations in ORAI1 cause tubular aggregate myopathy orders. In: Van der Knaap MS, Valk J, editors. Magnetic with hypocalcemia via constitutive activation of

8 Neurology: Genetics store-operated Ca21 channels. Hum Mol Genet 2015; 31. Tort F, Ferrer-Cortes X, Ribes A. Differential diagnosis of 24:637–648. lipoic acid synthesis defects. J Inherit Metab Dis 2016;39: 23. DePristo MA, Banks E, Poplin R, et al. A framework 781–793. for variation discovery and genotyping using next- 32. Mochel F, Knight MA, Tong WH, et al. Splice mutation generation DNA sequencing data. Nat Genet 2011; in the iron-sulfur cluster scaffold protein ISCU causes 43:491–498. myopathy with exercise intolerance. Am J Hum Genet 24. Wang K, Li M, Hakonarson H. ANNOVAR: functional 2008;82:652–660. annotation of genetic variants from high-throughput 33. Olsson A, Lind L, Thornell LE, Holmberg M. Myopathy sequencing data. Nucleic Acids Res 2010;38:e164. with lactic acidosis is linked to chromosome 12q23.3-24.11 25. Kim DW, Uetsuki T, Kaziro Y, et al. Use of the human andcausedbyanintronmutationintheISCU gene resulting elongation factor 1 alpha promoter as a versatile and effi- in a splicing defect. Hum Mol Genet 2008;17:1666–1672. cient expression system. Gene 1990;91:217–223. 34. Gelling C, Dawes IW, Richhardt N, et al. Mitochondrial 26. Frezza C, Cipolat S, Scorrano L. Organelle isolation: func- Iba57p is required for Fe/S cluster formation on aconitase tional mitochondria from mouse liver, muscle and cul- and activation of radical SAM enzymes. Mol Cell Biol tured fibroblasts. Nat Protoc 2007;2:287–295. 2008;28:1851–1861. 27. Shimazaki H, Takiyama Y, Ishiura H, et al. A homozygous 35. Seyda A, Newbold RF, Hudson TJ, et al. A novel syn- mutation of C12orf65 causes spastic paraplegia with optic drome affecting multiple mitochondrial functions, located atrophy and neuropathy (SPG55). J Med Genet 2012;49: by microcell-mediated transfer to chromosome 2p14- 777–784. 2p13. Am J Hum Genet 2001;68:386–396. 28. Morava E, Rodenburg RJ, Hol F, et al. Clinical and bio- 36. Invernizzi F, Ardissone A, Lamantea E, et al. Cavitating chemical characteristics in patients with a high mutant leukoencephalopathy with multiple mitochondrial dys- load of the mitochondrial T8993G/C mutations. Am J function syndrome and NFU1 mutations. Front Genet Med Genet A 2006;140:863–838. 2014;5:412. 29. Mitsuhashi S, Hatakeyama H, Karahashi M, et al. Muscle 37. Ahting U, Mayr JA, Vanlander AV, et al. Clinical, bio- choline kinase beta defect causes mitochondrial dysfunc- chemical, and genetic spectrum of seven patients with tion and increased mitophagy. Hum Mol Genet 2011;20: NFU1 deficiency. Front Genet 2015;6:123. 3841–3851. 38. Nizon M, Boutron A, Boddaert N, et al. Leukoencephal- 30. Mayr JA, Feichtinger RG, Tort F, et al. Lipoic acid bio- opathy with cysts and hyperglycinemia may result from synthesis defects. J Inherit Metab Dis 2014;37:553–563. NFU1 deficiency. Mitochondrion 2014;15:59–64.

Neurology: Genetics 9 Genome-wide scan in Hispanics highlights candidate loci for brain white matter hyperintensities

Ashley Beecham, MS ABSTRACT Chuanhui Dong, PhD Objective: To investigate genetic variants influencing white matter hyperintensities (WMHs) in the Clinton B. Wright, MD understudied Hispanic population. Nicole Dueker, PhD Methods: Using 6.8 million single nucleotide polymorphisms (SNPs), we conducted a genome- Adam M. Brickman, PhD wide association study (GWAS) to identify SNPs associated with WMH volume (WMHV) in 922 Liyong Wang, PhD Hispanics who underwent brain MRI as a cross-section of 2 community-based cohorts in the Charles DeCarli, MD Northern Manhattan Study and the Washington Heights–Inwood Columbia Aging Project. Multi- Susan H. Blanton, PhD ple linear modeling with PLINK was performed to examine the additive genetic effects on ln Tatjana Rundek, MD, (WMHV) after controlling for age, sex, total intracranial volume, and principal components of PhD ancestry. Gene-based tests of association were performed using VEGAS. Replication was per- Richard Mayeux, MD, formed in independent samples of Europeans, African Americans, and Asians. MS Ralph L. Sacco, MD Results: From the SNP analysis, a total of 17 independent SNPs in 7 genes had suggestive evi- dence of association with WMHV in Hispanics (p , 1 3 1025) and 5 genes from the gene- based analysis with p , 1 3 1023. One SNP (rs9957475 in GATA6)and1gene(UBE2C) , Correspondence to demonstrated evidence of association (p 0.05) in the African American sample. Four SNPs Dr. Sacco: with p , 1 3 1025 were shown to affect binding of SPI1 using RegulomeDB. [email protected] Conclusions: This GWAS of 2 community-based Hispanic cohorts revealed several novel WMH- associated genetic variants. Further replication is needed in independent Hispanic samples to val- idate these suggestive associations, and fine mapping is needed to pinpoint causal variants. Neurol Genet 2017;3:e185; doi: 10.1212/NXG.0000000000000185

GLOSSARY GWAS 5 genome-wide association study; HDAC 5 histone deacetylase; LD 5 linkage disequilibrium; mRNA 5 messenger RNA; NOMAS 5 Northern Manhattan Study; SNP 5 single nucleotide polymorphism; WHICAP 5 Washington Heights– Inwood Columbia Aging Project; WMH 5 white matter hyperintensity; WMHV 5 WMH volume.

White matter hyperintensities (WMHs) are frequently detected by MRI in the aging brain1 and are associated with a range of negative health outcomes.2–6 Prevalence ranges from 40% to 70% in the fifth decade7 and increases with age. It is important that Hispanics and African Americans have shown more severe WMHs than ancestral Europeans.8 Although heterogeneous in etiol- ogy, there is a consistent link between WMH burden and cerebrovascular risk factors, as well as with retinal microvascular abnormalities and vascular pathology.9 Thus, WMHs could be considered a quantitative marker of small vessel injury.10 Moreover, WMHs have a significant genetic component, with heritability estimates from 0.45 to 0.80.10–13 Identifying genetic determinants of WMHs has been a challenge. Linkage studies have yielded conflicting findings regarding the genetic loci influencing WMHs.12–14 The most Supplemental data at Neurology.org/ng From the John T. McDonald Department of Human Genetics (A.B., L.W., S.H.B., R.L.S.), John P Hussman Institute for Human Genomics (A.B., N.D., L.W., S.H.B.), Evelyn F. McKnight Brain Institute (C.D., C.B.W., T.R., R.L.S.), Department of Neurology (C.D., C.B.W., T.R., R.L.S.), and Department of Epidemiology and Public Health (C.B.W., T.R., R.L.S.), Miller School of Medicine, University of Miami, FL; Gertrude H. Sergievsky Center (A.M.B., R.M.), Taub Institute for Research on Alzheimer’s Disease and the Aging Brain (A.M.B., R.M.), and Department of Neurology (A.M.B., R.M.), College of Physicians and Surgeons, Columbia University, New York; and Department of Neurology and Center for Neuroscience (C.D.), University of California at Davis, Sacramento. Funding information and disclosures are provided at the end of the article. Go to Neurology.org/ng for full disclosure forms. The Article Processing Charge was funded by the NIH. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Neurology.org/ng Copyright © 2017 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 consistently replicated locus identified for nonuniformities were removed from the image, the corrected WMHs is on chromosome 17q25 and was image was modeled as a mixture of 2 Gaussian probability ; functions with the segmentation threshold determined at first identified in 9,500 individuals of Euro- the minimum probability between these 2 distributions. Once pean descent.15 This locus was later replicated brain matter segmentation was achieved, a single Gaussian dis- and additional loci on chromosomes 10q24 tribution was fitted to image data, and a segmentation threshold and 2p21 identified.16 A multiethnic meta- for WMHs was determined a priori as 3.5 SDs in pixel intensity ; above the mean of the fitted distribution of the brain pa- analysis, including 2,000 African Ameri- renchyma.20 For the NOMAS, high interrater reliabilities cans, ;800 Hispanics, and ;400 Asians, were documented for both the intracranial volume (0.97) and identified additional loci on chromosomes WMH (0.99).2 16 1q22 and 2p16. Genotyping, imputation, and quality control. DNA sam- Although Hispanics and African Americans ples were obtained through whole blood extraction. For the NO- show more severe WMHs than Europeans,8 MAS, genotyping of DNA samples was performed using Affymetrix Genome-Wide Human SNP Array 6.0 chips ac- these populations have not been the focus cording to Affymetrix procedures at the Genotyping Core of the of genetic discovery. To address this, we John P. Hussman Institute for Human Genomics at the Uni- performed a genome-wide association scan to versity of Miami. Genotype calling was performed using Affy- identify genetic variants influencing WMHs in metrix Power Tools v.1.15.0. For the WHICAP, genotyping was performed in 2 batches on the Illumina HumanHap 650Y Hispanics in 2 community-based cohorts from (WHICAP1) and Omni 1M (WHICAP2) chips, according to the Northern Manhattan Study (NOMAS) and Illumina procedures at Columbia University. Genotype calling the Washington Heights–Inwood Columbia was performed separately for WHICAP1 and WHICAP2 using Aging Project (WHICAP). GenomeStudio v.1.0. Genome-wide genotype data were available on a total of 1,067 Hispanics meeting all inclusion criteria from either NOMAS (811) or WHICAP (191 WHICAP1 and 65 METHODS Participants. Hispanic participants from NOMAS WHICAP2). Because both NOMAS and WHICAP were sam- and WHICAP with WMHs and genotype data available were pled from Northern Manhattan, we expected some participants included in this study. All participants provided written informed to be enrolled in both studies. Therefore, we checked for unex- consent. The study was approved by the Institutional Review Boards pected relationships with identical-by-descent proportion .0.25 of the University of Miami and Columbia University Medical using PLINK21 and found 34 duplicate and 15 related pairs Center. The research design and detailed ascertainment scheme for between NOMAS and WHICAP1 or WHICAP2. After drop- NOMAS and WHICAP were described in detail previously.17,18 ping at least 1 from each pair, typically the one genotyped on the To be eligible for recruitment in the NOMAS, participants had to smaller array, a total of 1,019 unrelated individuals remained for be at least 40 years of age, never been diagnosed with a stroke, and the final joint analysis. resided for at least 3 months in a household with a telephone in Quality control was applied to both DNA samples and sin- Northern Manhattan. A total of 3,298 participants were enrolled gle nucleotide polymorphisms (SNPs). Specifically, samples from 1993 to 2001, and 199 unrelated household members were were removed from further analysis if they had call rates below recruited from 2003 to 2008. Data were collected at enrollment 95%, sex discrepancies, relatedness, autosomal heterozygosity using standardized data collection instruments, medical records, beyond 4 SDs from the mean. In addition, a subset of and fasting blood samples. ;51,000 independent SNPs (R2 # 0.2) across the genome were For the WHICAP, participants were recruited in 2 waves used to infer principal components using Eigenstrat software.22 (1992–1994 and 1999–2002) through the use of a probability Samples were eliminated if they were outliers beyond 6 SDs sample of Medicare beneficiaries who were older than 65 years from the mean based on the first 10 principal components. and residing in a defined area of Northern Manhattan. Each An additional 50 participants from the WHICAP with a diag- WHICAP participant underwent an in-person interview of gen- nosis of stroke were excluded from further analysis. This led to eral health and functional ability at the time of study entry, 760, 97, and 65 samples available for analysis from NOMAS, followed by a standardized assessment, including medical history, WHICAP1, and WHICAP2, respectively. SNPs were excluded physical and neurologic examination, and a neuropsychological if they were not in the Hardy-Weinberg equilibrium (p , 1 3 battery. 1026) or had a genotyping call rate less than 95%. This led to Quantitative measurement of WMH volume. Selected a total of 813,769; 624,026; and 970,293 SNPs available in NOMAS and WHICAP participants were scanned on the same NOMAS, WHICAP1, and WHICAP2, respectively. A total of 1.5T Philips Intera scanner at Columbia University Medical Cen- 6%, 1%, and 11% of SNPs were dropped because of low call ter using the same sequences. Participants were selected for MRI, rate for NOMAS, WHICAP1, and WHICAP2, respectively. on average 6 years after recruitment, if they were 55 years or older, Only ;100,000 SNPs overlapped between all 3 data sets, so were classified as nondemented at a previous study visit, had no imputation was performed on each data set using IMPUTE223 contraindications for MRI, and signed informed consent.17,18 with the 1000 Genomes Phase I (interim) NCBI Build b37 Data sets were transferred electronically to UC Davis for mor- (June 2011) reference panel. This reference panel included phometric analysis as previously described.19 Briefly, nonbrain 1,094 individuals from across Europe, Africa, Asia, and the elements were manually removed from the image by operator- Americas. SNPs were subsequently removed if the imputation guided tracing of the dura mater within the cranial vault, quality score was less than 0.8 or the minor allele frequency was including the middle cranial fossa but excluding the posterior less than 0.01 in any of the NOMAS, WHICAP1, or WHI- fossa and cerebellum. The resulting measure of the cranial vault CAP2 samples, resulting in a total of 6,783,792 SNPs for was defined as the total intracranial volume. After image intensity analysis.

2 Neurology: Genetics Statistical analyses. To compare the sample characteristics for our Hispanic cohorts), VEGAS2 does not allow this option. between the NOMAS and WHICAP Hispanic cohorts, a x2 test VEGAS2 does, however, provide more extensive reference panel was used to test for frequency differences for categorical variables data than VEGAS for the analysis of the European, African and a student t test was performed to test for mean or median American, and Asian samples used in the CHARGE analyses. differences for continuous variables using R.24 Population substruc- Nonetheless, the VEGAS and VEGAS2 algorithms are based ture was assessed using ADMIXTURE,25 including reference pop- on the same simulation strategy and produce similar results, given ulations of Europeans (Utah residents with northern and western the same input data set. Sensitivity analyses were conducted using European ancestry) and Africans (Yoruba in Ibadan, Nigeria) from MAGMA (v.1.06),30 which produced similar results (data not 1000 Genomes as well as Native Americans from the Human shown). Genome Diversity Project.26 Because of the skewed distribution, WMHs were log transformed. To evaluate the association of RESULTS The sample characteristics of the NOMAS WMHs with each SNP, using PLINK,21 an additive genetic model and WHICAP Hispanic MRI cohorts are presented in was fit by regressing log-transformed WMHs on genotype dosage table 1. Among the 922 Hispanic participants overall, (0–2 copies of the variant allele) after adjusting for age at MRI, sex, the mean age was 71.1 6 9.0 years, and 63.7% were total intracranial volume, and the top 2 principal components of ancestry to account for population substructure. No adjustment for women. Compared with NOMAS Hispanic partici- treatment was needed, as all patients were classified as non- pants, WHICAP Hispanic participants had a greater demented. These analyses were performed separately in each of the WMH volume (WMHV) and smaller intracranial NOMAS, WHICAP1, and WHICAP2 cohorts. An inverse- volumes because of the older age distribution. The variance meta-analysis under a fixed-effects model was used in population substructure of each of the NOMAS and METAL (released March 27, 2011)27 to summarize results across WHICAP Hispanic MRI cohorts is presented in figure 1. the 3 strata. SNPs with I2 heterogeneity estimates greater than 50% were removed post hoc. VEGAS (v.01)28 was used to assign SNPs The overall cohort was on average 57% European, to genes (within 50KB using NCBI Build b37) and produce gene- 31% African, and 12% Native American, expected, based test statistics and empirical p values by simulation. A random given the largely Dominican Hispanic population of sample of 200 Hispanics from the NOMAS was used to estimate Northern Manhattan. linkage disequilibrium (LD) patterns for each gene. In the meta-analysis of Hispanic samples, the To determine the relevance of our findings to other race and quantile-quantile plot (figure e-1 at Neurology.org/ng) ethnic groups, summary statistics were obtained from CHARGE analyses of ;18,000 Europeans, ;2,000 African Americans, and did not reveal an obvious departure from the distri- ;400 Asians16 for each of our top SNPs from the single SNP bution of expected p values by chance (genomic analysis (using an a priori threshold of p , 1 3 1025 and inde- inflation factor [l] of 0.997), suggesting no signifi- pendence with R2 # 0.2 based on an LD-clumping procedure) cant inflation of probability values due to population and each SNP within our top genes (using an a priori threshold of substructure. Genome-wide association results are , 3 23 16 p 1 10 ). Although the CHARGE analyses included shown in figure 2. While no SNP reached genome- ;800 Hispanics, these samples overlapped largely with NOMAS wide significance, 17 independent SNPs (R2 # 0.2) and WHICAP and therefore did not provide a suitable and inde- , 3 25 pendent replication of our study. A multiethnic meta-analysis was were associated with WMHV at p 1 10 . performed using a z-score–based fixed-effects model in METAL Frequencies and effect estimates for these 17 SNPs (released March 27, 2011). An inverse-variance meta-analysis was are shown in table 2. The mean changes in ln not possible, given the data format produced by the previous (WMHV) for per copy increase in these risk alleles – CHARGE z-score based meta-analyses. ranged from 0.14 to 0.55, corresponding to 14.4%– VEGAS2 (v.02)29 was used to produce gene-based empirical 21.7% of the mean WMHV. Seven of the 17 SNPs p values from the CHARGE summary statistics using 1000 Genomes reference panels of Europeans, Africans, and Asians were within a known gene. While none of the 17 to estimate LD patterns for each gene. While VEGAS allows SNPs demonstrated statistical significance at a multi- 2 input of a customized reference data set to estimate LD (necessary ple testing threshold of 2.9 3 10 3 (0.05/17), one of

Table 1 WMH measures and associated risk factors

Risk factors and Overall NOMAS WHICAP1 WHICAP2 NOMAS vs WHICAP1 NOMAS vs WHICAP1 vs phenotypes (N 5 922) (N 5 760) (N 5 97) (N 5 65) WHICAP2 WHICAP2 p Value (Pearson x2) N (%)

Female 587 (63.7) 474 (62.4) 68 (70.8) 45 (69.2) 1.69 3 10201 3.34 3 10201 1.00

Mean 6 SD p Value (Student t test)

Age at MRI 71.1 6 9.0 69.1 6 8.4 80.2 6 5.3 81.0 6 5.1 1.42 3 10240 9.92 3 10231 3.52 3 10201

WMHV, mL 8.0 6 9.7 6.9 6 8.7 11.3 6 11.4 15.5 6 13.2 3.88 3 10204 2.72 3 10206 4.13 3 10202

Intracranial volume, mL 1,133.7 6 112.5 1,139.0 6 112.9 1,114.9 6 108.5 1,099.7 6 105.3 4.23 3 10202 5.20 3 10203 3.74 3 10201

Abbreviations: NOMAS 5 Northern Manhattan Study; WHICAP 5 Washington Heights–Inwood Columbia Aging Project; WMH 5 white matter hyper- intensity; WMHV 5 WMH volume.

Neurology: Genetics 3 Figure 1 Global genetic ancestry

The plots depict the ancestral proportions observed within each cohort, where individual is represented on the x-axis and proportion ancestry on the y-axis. Global ancestry was calculated using 50,327 independent single nucleotide polymor- phisms (R2 # 0.2), with reference populations including 46 Europeans (Utah residents with northern and western European ancestry from 1000 Genomes), 46 Africans (Yoruba in Ibadan, Nigeria from 1000 Genomes), and 46 Native Americans (Surui and Karitiana of Brazil and Maya and Pima of Mexico from the Diversity Project). NOMAS 5 Northern Manhattan Study; WHICAP 5 Washington Heights–Inwood Columbia Aging Project.

these 17 SNPs, rs9957475 (risk allele 5 C, beta 5 Using RegulomeDB,31 we sought to determine 0.188, p 5 6.19 3 1026) located in GATA-binding whether any SNP associated with WMHV had reg- protein 6 (GATA6), demonstrated evidence of asso- ulatory function. Of the 95 SNPs with p , 1 3 ciation with p , 0.05 in an independent sample of 1025 (17 independent SNPs with R2 # 0.2), 5 were African Americans (rs9957475, risk allele 5 C, p 5 likely to affect protein binding: one of these variants, 0.020). In the NOMAS Hispanic sample, the average rs2670314, was located downstream of GDNF fam- WMHV for 0, 1, and 2 copies of the risk allele were ily receptor a 4(GFRA4) on chromosome 20p13, 6.08, 8.58, and 8.95 mL, respectively. No SNPs while the remaining 4 variants (rs73306471, reached genome-wide significance in the z-score– rs73306449, rs73306445, and rs73306424) were based or sample size–weighted, multiethnic meta- located on chromosome 20q13 within an LD block analysis, not surprising given the substantial weight spanning 3 genes, including WAP four-disulfide (per the sizable sample) of the largely insignificant core domain 3 (WFDC3), deoxynucleotidyltransfer- European cohort. ase terminal-interacting protein 1 (DNTTIP1), and

Figure 2 Single nucleotide polymorphism association results for meta-analysis of white matter hyperintensity volume

The Manhattan plot depicts the –log 10 p values by position from the Hispanic meta-analysis of Northern Manhattan Study and Washington Heights–Inwood Columbia Aging Project single single nucleotide polymorphism association results.

4 Neurology: Genetics Table 2 Single SNP association results (independent with R2 < 0.2) with p £ 1.0 3 1025 among Hispanics

Hispanic cohorts CHARGE

Chr:MB rsID Gene:Func RA N FRQ W1 FRQ W2 FRQ b p I2 AFR FRQ AFR p ASN FRQ ASN p EUR FRQ EUR p Meta-p

1:92.0 rs13447479 CDC7:intron G 0.16 0.17 0.11 0.19 8.12 3 10206 42.6 0.10 9.40 3 10201 0.26 13.53 3 10201 0.18 18.22 3 10202 8.83 3 10203

2:107.4 rs76479473 ST6GAL2:UTR3 A 0.97 0.97 0.96 0.44 3.16 3 10206 0.0 0.98 9.70 3 10201 0.95 6.33 3 10201 0.95 13.04 3 10201 6.61 3 10202

4:18.4 rs6832924 — T 0.93 0.95 0.90 0.28 5.62 3 10206 0.0 0.79 15.39 3 10201 NA NA NA NA 1.97 3 10203

4:58.4 rs1372094 — T 0.29 0.22 0.21 0.15 5.81 3 10206 0.0 0.29 1.12 3 10201 0.02 12.80 3 10201 0.28 16.57 3 10201 3.35 3 10201

5:87.4 rs75027182 — T 0.02 0.01 0.02 0.55 8.56 3 10206 0.0 0.06 16.31 3 10201 NA NA NA NA 4.44 3 10203

5:119.0 rs162603 — G 0.33 0.31 0.34 0.18 1.26 3 10207 31.3 0.33 18.59 3 10201 0.33 12.36 3 10201 0.31 5.72 3 10201 4.77 3 10201

6:72.4 rs272210 — A 0.03 0.02 0.08 0.42 5.23 3 10206 11.9 0.09 13.11 3 10201 NA NA NA NA 7.67 3 10204

8:136.1 rs79780482 — A 0.02 0.03 0.01 0.46 9.98 3 10206 0.0 0.07 8.46 3 10201 0.01 7.33 3 10201 NA NA 1.16 3 10202

15:98.7 rs62024995 — C 0.75 0.75 0.79 0.18 4.78 3 10207 25.8 0.89 17.41 3 10201 0.75 13.67 3 10201 0.63 15.98 3 10202 2.59 3 10203

18:19.8 rs9957475 GATA6:intron C 0.19 0.18 0.18 0.19 6.19 3 10206 0.0 0.45 11.99 3 10202 0.09 14.98 3 10201 0.02 6.65 3 10201 1.12 3 10201

18:27.0 rs9947408 — G 0.10 0.07 0.10 0.26 8.51 3 10206 0.0 0.31 4.74 3 10201 NA NA NA NA 7.27 3 10202

18:29.8 rs10502586 — G 0.11 0.13 0.08 0.23 9.13 3 10206 0.0 0.19 19.40 3 10201 0.03 6.87 3 10203 0.07 9.24 3 10201 6.53 3 10201

20:3.6 rs2670314 GFRA4:ds G 0.61 0.58 0.53 0.14 7.74 3 10206 0.0 0.39 6.75 3 10201 0.54 13.66 3 10201 0.71 18.34 3 10201 2.70 3 10201

20:5.9 rs73596147 TRMT6:intron T 0.92 0.91 0.96 0.24 7.95 3 10206 0.0 0.89 7.02 3 10201 0.91 8.75 3 10201 0.98 18.40 3 10201 3.27 3 10201

20:44.4 rs6032525 WFDC3:intron G 0.03 0.02 0.03 0.44 1.17 3 10206 18.7 0.06 11.50 3 10201 NA NA NA NA 1.02 3 10204

22:35.4 rs5999656 — C 0.85 0.87 0.87 0.20 3.34 3 10206 0.0 0.68 14.39 3 10201 0.94 19.22 3 10201 0.93 3.00 3 10201 8.18 3 10201

22:45.7 rs6007590 FAM118A:intron G 0.08 0.13 0.11 0.24 8.77 3 10206 0.0 0.09 17.83 3 10201 0.31 16.44 3 10201 0.11 19.25 3 10201 2.59 3 10201 erlg:Genetics Neurology: Abbreviations: NOMAS 5 Northern Manhattan Study; SNP 5 single nucleotide polymorphism; WHICAP 5 Washington Heights–Inwood Columbia Aging Project. N FRQ, W1 FRQ, W2 FRQ, AFR FRQ, ASN FRQ, and EUR FRQ - NOMAS, WHICAP1, WHICAP2, CHARGE African American, CHARGE Asian, and CHARGE European allele frequencies, respectively. NA - not applicable, indicating the variant was monomorphic or not well imputed (INFO , 0.4) in the replication sample. CHARGE results with p , 0.05 are indicated in bold text. All BP are in HG19 coordinates in the form chromosome:megabase. I2 5 heterogeneity statistic describing the percentage of variation across NOMAS, WHICAP1, and WHICAP2 due to heterogeneity. CHARGE results are based on 17,936 Europeans, 1,943 Africans, and 405 Asians. SNPs which have not been analyzed in select populations are monomorphic. A “1” sign indicates that the direction of effect is the same as that seen in Hispanics. The Meta-p was calculated using a z-score–based fixed-effect meta-analysis of Hispanic and CHARGE cohorts. No I2 is available due to the z-score–based method. 5 ubiquitin-conjugating enzyme E2 C (UBE2C). In 1025, all on chromosomes 20q13 and 20p13, influ- total, these 5 SNPs were predicted to influence the enced the binding of SPI1. A gene-based analysis binding of 70 proteins. A list of the proteins affected implicated UBE2C on chromosome 20q13 in our by each SNP and associated RegulomeDB scores are Hispanic sample, a finding which was supported in an shown in table e-1. Notably, Spi-1 proto-oncogene independent sample of African Americans. This study (SPI1) was the only protein influenced by focuses on a minority population as a means of dis- rs2670314 on 20q13 and rs73306449 on 20p13. covery for novel genetic variants that influence white Two additional SNPs within the WFDC3- matter lesion burden. DNTTIP1-UBE2C LD block influenced the GATA6 encodes a member of the GATA family of binding of SPI1. No other proteins were influenced transcription factors and has been shown to regulate by4ormoreSNPs.Ofthe6,395SNPswithp , 1 angiogenesis and promote survival of endothelial 3 1023, 110 were likely to affect the binding of cells.32 Expression in early embryogenesis is crucial 130proteins(tablee-1).Themostcommonly to heart development, and mutations in this gene affected proteins were RNA polymerase II subunit have been shown to be associated with a number of A(POLR2A), which was influenced by 24 variants congenital abnormalities including dilated cardiomy- across 12 chromosomes, and CCCTC-binding fac- opathy33 and atrial fibrillation.34 In addition to tor (CTCF), which was influenced by 23 variants GATA6 displaying a single SNP association with across 13 chromosomes (table e-2). WMHV in Hispanics and African Americans, 18 Table 3 indicates the results from our gene-based variants with p , 1 3 1023 in the Hispanic analysis association tests (p , 1.00 3 1023). The most strongly affect the binding of at least 1 of GATA6, GATA1, associated gene in our Hispanic sample, UBE2C (p 5 GATA2, and GATA3. Similar to GATA6, the other 3.61 3 1024), demonstrated evidence of association in members of the GATA family of transcription factors the African American sample (p 5 0.034). The single are important in the vascular and lymphatic systems SNP results of all variants within this gene region in through regulation in the development of erythrocytes the Hispanic and all replication samples are shown in (GATA1) and lymphocytes (GATA2 and GATA3).35 table e-3. The majority of SNPs within 50KB of Other notable transcription factors affected by our UBE2C with p , 1.00 3 1025 in Hispanics were top SNP associations include CTCF and POLR2A. monomorphic in the European and Asian samples. CTCF is a regulator protein, which can bind histone acetyltransferase–containing complex or histone de- DISCUSSION In this genome-wide association study acetylase (HDAC)-containing complex, functioning as (GWAS) of 2 community-based Hispanic cohorts, 17 a transcriptional activator or repressor. Protein acety- independent SNPs located within 7 genes showed lation has been implicated in a host of neurologic dis- suggestive associations with WMHV with p , 1 3 eases in recent years due primarily to HDAC-associated 1025. One SNP in GATA6 demonstrated evidence of decrease in the level of a protein or widespread genomic association in an independent sample of African transcriptional losses.36 Thus, CTCF also plays a critical Americans. A total of 110 unique variants with p , 1 role in this pathway for neurodegeneration. POLR2A 3 1023 were shown to influence the binding of plays a role in messenger RNA (mRNA) processing 130 unique proteins, most commonly CTCF and through the transcription of pre-mRNA from the geno- POLR2A. Four of 5 of the 110 variants with p , 1 3 micDNA;thus,inhibitionofPOLR2A binding could

Table 3 Gene-based association results with p £ 1.0 3 1023 among Hispanics

CHARGE

AFR ASN EUR

Chr Start (MB) End (MB) Gene # SNP p # SNP p # SNP p # SNP p

20 44.39 44.50 UBE2C 285 3.61 3 10204 479 3.35 3 10202 244 7.14 3 10201 327 8.34 3 10201

10 17.58 17.71 PTPLA 251 2.15 3 10204 775 3.75 3 10201 433 9.68 3 10201 535 8.20 3 10201

2 108.4 108.6 RGPD4 193 5.37 3 10204 316 8.49 3 10201 189 5.95 3 10201 284 7.52 3 10201

2 73.56 73.89 ALMS1 641 6.01 3 10204 1,251 9.79 3 10201 412 8.01 3 10201 673 4.85 3 10201

15 56.87 57.08 ZNF280D 265 3.70 3 10204 312 9.09 3 10202 258 7.15 3 10201 428 8.78 3 10201

Abbreviation: AFR, ASN, EUR - CHARGE African American, Asian, and European cohorts. SNP 5 single nucleotide polymorphism. All BP are in HG19 coordinates in the form chromosome:megabase. CHARGE results are based on 17,936 Europeans, 1,943 Africans, and 405 Asians. Gene start and end positions are 650KB from the gene. # SNPs 5 # nonmonomorphic SNPs analyzed.

6 Neurology: Genetics also lead to widespread RNA dysregulation, an integral community-based cohorts which used the same element in a number of neurologic disease.37 MRI protocol and quantitative measurement of Perhaps, one of the most interesting findings in WMHV at the same medical center. However, our this study is the implication of UBE2C in the gene- study also has limitations. First, with 922 Hispanic based test for both Hispanics and African Americans. participants, we had limited power, at a genome- This ubiquitin-conjugating enzyme is an important wide significance level, to uncover an association with regulator in cell cycle progression, and overexpression small-to-moderate effect sizes. Second, as comprehen- has been associated with progression of malignant sive genotype data are not available for Hispanic pop- gliomas.38 There is a substantial amount of LD in ulations in the 1000 Genome database, imputations Hispanics between SNPs within 50KB of UBE2C, were based on pooled samples from populations of including SNPs in WFDC3, DNTTIP1, TNNC2, the Americas, Europe, Africa, and Asia. Bias in allele SNX21, ACOT8, and ZSWIM3. WFDC3 functions frequency estimates cannot be excluded for those as a protease inhibitor, and notably, both protein SNPs with large allele differences across subpopula- degradation though ubiquitination and protease inhi- tions, although high imputation quality was observed. bition are represented among neurodegenerative Third, since this GWAS was performed in Hispanics, diseases.39 Furthermore, DNTTPI1 acts as a chroma- findings may not necessarily extend to other race- tin-binding module in complex with HDAC1 and ethnic groups. For instance, we see varying allele fre- HDAC2. As previously mentioned, 3 SNPs with quency differences across notable SNPs such as the p , 1 3 1025 within WFDC3 and DNTTIP1 on one found in GATA6. The majority of SNPs we see chromosome 20q13 affect protein binding of SPI1, as associated in Hispanics within 50KB of UBE2C are with both genes also showing nominal significance in in fact monomorphic in the European and Asian rep- Hispanics with p 5 9.55 3 1023 and p 5 2.40 3 1023, lication samples, findings which are underscored by respectively. This is a transcription factor which is the population substructure observed in our study involved in the activation of gene expression during sample. This is a weakness for generalizability but myeloid and B-lymphocyte development and is crit- emphasizes the importance of genetic study in diverse ical for viability and function of human brain micro- samples. glia, with a loss reducing the ability of microglia to This GWAS of 2 community-based Hispanic co- clear debris by phagocytosis.40 Outside chromosome horts revealed several novel genetic loci for WMHV. 20q13, 1 additional SNP on chromosome 20p13 Further replication is needed in an independent His- with p , 1 3 1025 and 7 additional SNPs with panic sample as well as samples from additional race/ p , 1 3 1023 across 5 additional chromosomes affect ethnic groups. In addition, local ancestry computa- binding of SPI1. tion may provide additional insight into the ancestral The CHARGE consortium identified loci for risk haplotypes. While fine mapping is needed to pin- WMHV on chromosomes 17q25, 10q24 2p21, point causal variation for drug targets in relevant pop- 1q22, and 2p16 in primarily individuals of European ulations, this study represents progress toward descent but also through multiethnic meta-analysis elucidating the genetic underpinnings of increased and replication, inclusive of ;2,000 African Americans, WMHV potentially underlying cognitive impairment ;400 Asians, in addition to the ;18,000 Europeans and vascular dementia. (which we have now used as replication of our findings) and ;800 Hispanics (a majority subset of AUTHOR CONTRIBUTIONS our NOMAS and WHICAP samples).16 Only 2 of Ashley Beecham: study design, analysis and interpretation of data, and drafting and revising the manuscript. Chuanhui Dong: analysis and inter- these loci (PDCD11 on 10q24 and PMF1 on 1q22) pretation of data and drafting and revising the manuscript. Clinton B. demonstrated a nominal level of significance (p , Wright: study design, analysis and interpretation of data, and drafting 0.05) in the Hispanic sample, with the same risk allele and revising the manuscript. Nicole Dueker: drafting and revising the manuscript. Adam M. Brickman and Liyong Wang: study design and as seen in Europeans. This may be due to differences revising the manuscript. Charles DeCarli: study design, analysis and in allele frequency across populations and supports interpretation of data, and revising the manuscript. Susan H. Blanton, the importance of genetic discovery in minority pop- Tatjana Rundek, Richard Mayeux, and Ralph L. Sacco: study design ulations. However, SNPs implicated by CHARGE and revising the manuscript. Dr. DeCarli is a consultant of Novartis Pharmaceuticals. within 17q25, 10q24, and 2p16 have been shown to influence binding of 30 proteins,30 29 of which ACKNOWLEDGMENT , 3 23 are also influenced by SNPs with p 1 10 in NOMAS: The authors acknowledge the Center for Genome Technology the Hispanic analysis, most notably POLR2A. This within the University of Miami John P. Hussman Institute for Human may indicate common pathways across populations. Genomics for generating the genotype data for this project. This investiga- A strength of the current study is that it focuses on tion was funded by the Evelyn F. McKnight Institute, the NIH (R01 NS 29993), and the National Institute of Aging (P30 AG010120—through discovery of novel variation influencing WMHV in the contributions of Dr. DeCarli). They also thank the study participants a Hispanic sample. It includes analysis of 2 and their families for their willingness to participate in our research studies.

Neurology: Genetics 7 WHICAP: Data collection and sharing for this project was supported by 6. Arsava EM, Rahman R, Rosand J, et al. Severity of leu- the Washington Heights–Inwood Columbia Aging Project (WHICAP, koaraiosis correlates with clinical outcome after ischemic PO1AG07232, R01AG037212, and RF1AG054023) funded by the stroke. Neurology 2009;16:1403–1410. National Institute on Aging (NIA) and by the National Center for 7. Launer LJ. Epidemiology of white matter lesions. Top Advancing Translational Sciences, NIH, through Grant Number Magn Reson Imaging 2004;6:365–367. UL1TR001873. This manuscript has been reviewed by WHICAP inves- tigators for scientific content and consistency of data interpretation 8. Brickman AM, Schupf N, Manly JJ, et al. Brain morphology with previous WHICAP Study publications. They acknowledge the in older African Americans, Caribbean Hispanics and Whites WHICAP study participants and the WHICAP research and support staff from Northern Manhattan. Arch Neurol 2008;8:1053–1061. for their contributions to this study. 9. Jeerakathil T, Wolf PA, Beiser A, et al. Stroke risk profile predicts white matter hyperintensity volume: the Framing- STUDY FUNDING ham Study. Stroke 2004;8:1857–1861. This study was supported by the Evelyn F. McKnight Brain Institute, the 10. Atwood LD, Wolf PA, Heard-Costa NL, et al. Genetic National Institute on Aging (NIA), and by the National Center for variation in white matter hyperintensity volume in the Advancing Translational Sciences, NIH through grants R01NS29993, Framingham Study. Stroke 2004;7:1609–1613. P30 AG010120, PO1AG07232, R01AG037212, RF1AG054023, and 11. Carmelli D, DeCarli C, Swan GE, et al. Evidence for UL1TR001873. genetic variance in white matter hyperintensity volume – DISCLOSURE in normal elderly male twins. Stroke 1998;6:1177 1181. A. Beecham has received research support from the NIH, the National 12. Kochunov P, Glahn D, Winkler A, et al. Analysis of Multiple Sclerosis Society, and the Evelyn F. McKnight Institute. genetic variability and whole genome linkage of whole- C. Dong has performed data analysis as needed for CSR Incorporated brain, subcortical, and ependymal hyperintense white mat- and has received research support from the NIH, the Intersocietal Accred- ter volume. Stroke 2009;12:3685–3690. itation Commission, and the American Heart Association/ASA-Bugher 13. Turner ST, Fornage M, Jack CR, et al. Genomic suscep- Foundation. C.B. Wright has served on a scientific advisory board for tibility Loci for brain atrophy, ventricular volume, and the NIH; receives publishing royalties from UpToDate; and has received leukoaraiosis in hypertensive sibships. Arch Neurol 2009; research support from the NIH and the American Stroke Association. 7:847–857. N. Dueker has received research support from the NIH. A.M. Brickman serves on the scientific advisory boards of Keystone Heart and ProPhase; 14. DeStefano AL, Atwood LD, Massaro JM, et al. Genome- serves on the editorial boards of the Journal of the International Neuro- wide scan for white matter hyperintensity: the Framing- psychological Society and Neuropsychology Review; has been a consultant for ham Heart Study. Stroke 2006;1:77–81. ProPhase and Keystone Heart; and has received research support from the 15. Fornage M, Debette S, Bis JC, et al. Genome-wide asso- NIH, Columbia University, Alzheimer’s Association, and the Mary ciation studies of cerebral white matter lesion burden: the E. Groff Surgical Medical Research and Education Charitable Trust. CHARGE consortium. Ann Neurol 2011;6:928–939. L. Wang has received research support from the NINDS. C. DeCarli 16. Verhaaren BF, Debette S, Bis JC, et al. Multiethnic genome- has served on the scientific advisory board of and has been a consultant wide association study of cerebral white matter hyperinten- for Novartis Pharmaceuticals and has received research support from sities on MRI. Circ Cardiovasc Genet 2015;2:398–409. the NIH. S.H. Blanton has received research support from the NIH/NINDS. T. Rundek has served on the editorial boards of Stroke, 17. Warsch JRL, Rundek T, Paik MC, et al. Association of the Neurology, Cerebrovascular Disease,andtheJournal of Ultrasound in Medicine Northern Manhattan study Global vascular risk score and and has received research support from the NINDS. R. Mayeux has Successful aging. J Am Geriatr Soc 2013;61:519–524. received research support from the NIH. R.L. Sacco has served on the 18. Scarmeas N, Luchsinger JA, Stern Y, et al. Mediterranean scientific advisory boards of the SOCRATES Trial through UCSF (indi- diet and magnetic resonance imaging-assessed cerebrovas- rect sponsor Astra Zeneca) and the EUCLID trial through the Duke cular disease. Ann Neurol 2011;69:257–268. Clinical Research Institute (indirect sponsor Astra Zeneca); has served 19. DeCarli C, Maisog J, Murphy DG, Teichberg D, Rapo- on the editorial boards of Stroke and Neuroepidemiology; and has received port SI, Horwitz B. Method for quantification of brain, research support from Boehringer Ingelheim, the NINDS, the Evelyn ventricular, and subarachnoid CSF volumes from MR im- McKnight Brain Institute, and the American Heart Association-Bugher – Foundation. Go to Neurology.org/ng for full disclosure forms. ages. J Comput Assist Tomogr 1992;2:274 284. 20. DeCarli C, Murphy DG, Teichberg D, Campbell G, Received April 11, 2017. Accepted in final form August 1, 2017. Sobering GS. Local histogram correction of MRI spatially dependent image pixel intensity nonuniformity. J Magn – REFERENCES Reson Imaging 1996;3:519 528. 1. de Leeuw FE, de Groot JC, Achten E, et al. Prevalence of 21. Purcell S, Neale B, Todd-Brown K, et al. PLINK: a tool cerebral white matter lesions in elderly people: a population set for whole-genome association and population-based – based magnetic resonance imaging study. The Rotterdam linkage analyses. Am J Hum Genet 2007;3:559 575. Scan Study. J Neurol Neurosurg Psychiatry 2001;1:9–14. 22. Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shad- 2. Wright CB, Festa JR, Paik MC, et al. White matter hyper- ick NA, Reich D. Principal components analysis corrects intensities and subclinical infarction: associations with psy- for stratification in genome-wide association studies. Nat chomotor speed and cognitive flexibility. Stroke 2008;3: Genet 2006;8:904–909. 800–805. 23. HowieB,MarchiniJ,StephensM.Genotypeimputation 3. Prins ND, van Dijk EJ, den Heijer T, et al. Cerebral white with thousands of genomes. G3 (Bethesda) 2011;6:457–470. matter lesions and the risk of dementia. Arch Neurol 24. R Core Team. R: a Language and Environment for Statis- 2004;10:1531–1534. tical Computing. Vienna: R Foundation for Statistical Com- 4. Guttmann CR, Benson R, Warfield SK, et al. White mat- puting; 2013. Available at: http://www.R-project.org/. ter abnormalities in mobility-impaired older persons. Neu- Accessed January, 2017. rology 2000;6:1277–1283. 25. Alexander DH, Novembre J, Lange K. Fast model-based 5. Smith EE. Leukoaraiosis and stroke. Stroke 2010;41(10 estimation of ancestry in unrelated individuals. Genome suppl):S139–S143. Res 2009;19:1655–1664.

8 Neurology: Genetics 26. Stanford University. Human Genome Diversity Project. 34. Li J, Liu WD, Yang ZL, Yang YQ. Novel GATA6 loss-of- 2007. Available at: http://www.hagsc.org/hgdp/. Accessed function mutation responsible for familial atrial fibrilla- January 5, 2016. tion. Int J Mol Med 2012;4:783–790. 27. Willer CJ, Li Y, Abecasis GR. METAL: fast and efficient 35. Gao J, Chen YH, and Peterson LC. GATA family tran- meta-analysis of genomewide association scans. Bioinfor- scriptional factors: emerging suspects in hematologic dis- matics 2010;17:2190–2191. orders. Exp Hematol Oncol 2015;4:28. 28. Liu JZ, McRae AF, Nyholt DR, et al. A versatile gene- 36. Didonna A, Opal P. The promise and perils of HDAC based test for genome-wide association studies. Am J Hum inhibitors in neurodegeneration. Ann Clin Transl Neurol Genet 2010;1:139–145. 2015;1:79–101. 29. Mishra A, Macgregor S. VEGAS2: Software for more flexible 37. Walsh MJ, Cooper-Knock J, Dodd JE, et al. Invited gene-based testing. Twin Res Hum Genet 2015;1:86–91. review: decoding the pathophysiological mechanisms that 30. de Leeuw CA, Mooij JM, Heskes T, et al. MAGMA: underlie RNA dysregulation in neurodegenerative disor- generalized gene-set analysis of GWAS data. PLoS Com- ders: a review of the current state of the art. Neuropathol put Biol 2015;11:e1004219. Appl Neurobiol 2015;2:109–134. 31. Boyle AP, Hong EL, Hariharan M, et al. Annotation of 38. Ma R, Kang X, Zhang G, et al. High expression of functional variation in personal genomes using Regulo- UBE2C is associated with the aggressive progression and meDB. Genome Res 2012;9:1790–1797. poor outcome of malignant glioma. Oncol Lett 2016;3: 32. Froese N, Kattih B, Breitbart A, et al. GATA6 promotes 2300–2304. angiogenic function and survival in endothelial cells by 39. Simonsen AH, Hagnelius NO, Waldemar G, Nilsson TK, suppression of autocrine transforming growth factor be- and McGuire J. Protein markers for the differential diag- ta/activin receptor-like kinase 5 signaling. J Biol Chem nosis of vascular dementia and Alzheimer’s disease. Int J 2011;7:5680–5690. Proteomics 2012;2012:824024. 33. Xu L, Zhao L, Yuan F, et al. GATA6 loss-of-function 40. Smith AM, Gibbons HM, Oldfield RL, et al. The tran- mutations contribute to familial dilated cardiomyopathy. scription factor PU.1 is critical for viability and function of Int J Mol Med 2014;5:1315–1322. human brain microglia. Glia 2013;6:929–942.

Neurology: Genetics 9 De novo CTBP1 variant is associated with decreased mitochondrial respiratory chain activities

Ewen W. Sommerville, ABSTRACT PhD Objective: To determine the genetic etiology of a young woman presenting an early-onset, pro- Charlotte L. Alston, PhD gressive neurodegenerative disorder with evidence of decreased mitochondrial complex I and Angela Pyle, PhD IV activities in skeletal muscle suggestive of a mitochondrial disorder. Langping He, PhD Methods: A case report including diagnostic workup, whole-exome sequencing of the affected Gavin Falkous, MPhil patient, filtering, and prioritization of candidate variants assuming a suspected autosomal reces- Karen Naismith, sive mitochondrial disorder and segregation studies. MBChB, DCH, FRCPCH Results: After excluding candidate variants for an autosomal recessive mitochondrial disorder, re- Patrick F. Chinnery, evaluation of rare and novel heterozygous variants identified a recently reported, recurrent path- . FRCP, FMedSci ogenic heterozygous CTBP1 missense change (c.991C T, p.Arg331Trp), which was confirmed Robert McFarland, to have arisen de novo. MRCPCH, MRCP, Conclusions: We report the fifth known patient harboring a recurrent pathogenic de novo PhD c.991C.T p.(Arg331Trp) CTBP1 variant, who was initially suspected to have an autosomal Robert W. Taylor, PhD, recessive mitochondrial disorder. Inheritance of suspected early-onset mitochondrial disease FRCPath could wrongly be assumed to be autosomal recessive. Hence, this warrants continued re- evaluation of rare and novel heterozygous variants in patients with apparently unsolved sus- pected mitochondrial disease investigated using next-generation sequencing. Neurol Genet Correspondence to 2017;3:e187; doi: 10.1212/NXG.0000000000000187 Dr. Taylor: [email protected] GLOSSARY ATP 5 adenosine triphosphate; COX 5 cytochrome c oxidase; GVGD 5 Grantham Variation with Grantham Deviation; MEF 5 mouse embryonic fibroblast; mtDNA 5 mitochondrial DNA; SDH 5 succinate dehydrogenase; WES 5 whole-exome sequencing.

Mitochondrial respiratory chain disorders are among the most common early-onset metabolic dis- orders with an estimated minimum point prevalence of 1 in 5,000 live births.1 Early-onset respi- ratory chain disorders are characterized by extreme clinical, molecular, and genetic heterogeneity due to pathogenic mitochondrial DNA (mtDNA) or nuclear gene variants affecting proteins essential for mitochondrial functions. The diagnosis of suspected mendelian mitochondrial disorders is especially challenging given that an estimated 1,200 mitochondrial proteins are encoded by the nuclear genome. Whole-exome sequencing (WES) has been widely used as a cost-effective tool for the diagnosis of suspected mendelian, early-onset mitochondrial respiratory chain disorders.2–5 How- ever, secondary respiratory chain defects in skeletal muscle have also been described in other neuromuscular and neurologic disorders due to pathogenic variants of genes encoding nonmito- chondrial proteins,6,7 providing additional challenges to candidate variant prioritization. We describe a young woman presenting an early-onset, progressive neurodegenerative disor- der with decreased mitochondrial complex I and IV activities in skeletal muscle. Using WES, we

From the Wellcome Centre for Mitochondrial Research (E.W.S., C.L.A., L.H., G.F., R.M., R.W.T.), Institute of Neuroscience, Newcastle University, United Kingdom; Department of Molecular and Human Genetics (E.W.S.), Baylor College of Medicine, Houston, TX; Wellcome Centre for Mitochondrial Research (A.P.), Institute of Genetic Medicine, Newcastle University; Armistead Child Development Centre (K.N.), Kings Cross Hospital, Dundee, Scotland; Department of Clinical Neurosciences (P.F.C.), School of Clinical Medicine, University of Cambridge; and MRC Mitochondrial Biology Unit (P.F.C.), University of Cambridge, United Kingdom. Funding information and disclosures are provided at the end of the article. Go to Neurology.org/ng for full disclosure forms. The Article Processing Charge was funded by the Wellcome Trust. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (CC BY), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Neurology.org/ng Copyright © 2017 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 identified a recurrent pathogenic de novo het- successful resuscitation, she was notably weaker and erozygous CTBP1 variant previously associ- dysphagic, becoming dependent on a percutaneous ated with an early-onset neurologic disorder.8 endoscopic gastrostomy tube for feeding. Brain MRI We highlight the challenges of variant priori- about 2 years prior to this event had revealed mild cerebellar and brainstem atrophy, but no neuro- tization in a patient who was initially sus- imaging was performed afterward. Clinical examina- pected to have an autosomal recessive tion, at the age of 16 years, revealed a young woman mitochondrial disorder. with sunken eyes and thin tapering fingers. Move- ments of her limbs and face were infrequent and slow. METHODS Muscle histopathological, biochemical, and genetic studies. Diagnostic muscle biopsy was performed accord- Lower limbs were cold and edematous with livedo ing to standard procedures. Muscle biopsy was subjected to cyto- reticularis of the overlying skin. Although her muscle chrome c oxidase (COX) and succinate dehydrogenase (SDH) tone was generally low, she had increased tone at the histochemical reactions.9 Mitochondrial respiratory chain complex elbows, wrists, knees, and ankles with contractures activities (complexes I-IV) relative to citrate synthase were measured of the wrists and elbows. There was a marked spinal 10 in the skeletal muscle homogenate as previously described. Whole scoliosis, and muscle mass was generally decreased. mitochondrial genome sequencing was performed to exclude patho- At 16 years of age, she died of acute respiratory genic point mutations and small insertions or deletions, while diag- nostic quantitative real-time PCR and long-range PCR assays were failure. performed for mtDNA quantification and detection of rearrange- Her family history was unremarkable for neuro- ments, according to standard protocols. logic disorders; she had 1 unaffected younger sibling. Whole-exome sequencing, analysis, and interpretation. Diagnostic Sanger sequencing of PLA2G6 Exome capture was attained using the Illumina TruSeq 62 Mb (NM_003560) and APTX (NM_001195251) per- capture kit, sequenced using the Illumina HiSeq 2000 system formed at another center was negative for pathogenic in 100 (bp) reads, and aligned to the human reference or likely pathogenic variants. genome (UCSC hg19). Variants with a minor allele frequency A skeletal muscle biopsy showed histopathologic $ 0.01 (1%) from 378 in-house controls and external variant abnormalities including variation in muscle fiber size, databases including the Exome Aggregation Consortium, National Heart, Lung, and Blood Institute Exome Variant Server occasional internal nuclei, and evidence of denerva- (Exome Sequencing Project), and 1000 Genomes Project were tion atrophy. These were accompanied by increasing excluded. Autosomal recessive (compound heterozygous or accumulation of fibroadipose connective tissue and homozygous) variants in nuclear genes encoding mitochondrial- patchy lipid accumulation within the muscle, with localized proteins were prioritized. Autosomal recessive variants of excess lipid also surrounding the muscle. Oxidative all nuclear genes were next examined, followed by examination of enzyme histochemistry showed patchy loss of COX rare and novel heterozygous variants. PolyPhen-2, SIFT, and activity and clumped SDH reactivity; unfortunately, Align Grantham Variation with Grantham Deviation (GVGD) were used to assess the pathogenicity of candidate variants. sequential COX-SDH histochemistry was not per- formed. Biochemical analysis of mitochondrial respi- Sanger sequencing validation. Candidate variants were vali- ratory chain activities showed decreased complex I dated by Sanger sequencing using standard protocols.2 Custom forward and reverse primers were designed for all exons and in- (57% of control) and IV (35% of control) activities, tronic regions of EARS2 (NM_001083614) and exon 9 of with sparing of the complex II activity (figure, A). CTBP1 (NM_001012614). Based on these findings, whole mitochondrial genome sequencing was performed but failed to RESULTS Case report. This female patient was born detect known pathogenic or likely pathogenic var- to nonconsanguineous parents following a normal iants. Quantitative real-time PCR and long-range pregnancy. Very early development was considered PCR assays also excluded mtDNA depletion and normal though, while the patient was able to sit inde- mtDNA rearrangements. pendently at 8 months, she disliked being placed prone and never acquired the ability to crawl. She Whole-exome sequencing analyses. Informed parental did manage to walk with hand-holding support at consent was obtained, and WES was performed. the age of 16 months but generalized hypotonia and Anticipating a possible autosomal recessive mito- weakness meant that independent ambulation was chondrial disorder, nuclear genes encoding never achieved and she started to use a wheelchair. mitochondrial-localized proteins were prioritized. Fine motor skills were variable in that she could turn However, no likely candidate variants were iden- the pages of a book and hold a crayon, but had dif- tified, and inspection of all rare autosomal recessive ficulty drawing. Language acquisition was limited to variants did not reveal persuasive candidates. A a few words in early life and these were lost by the age rare heterozygous missense variant (c.452A.G, of 4 years. Scoliosis, requiring corrective surgery, had p.Lys151Arg) in EARS2 (NM_001083614) encoding developed by 10 years, and at 14 years, she had mitochondrial glutamyl-transfer RNA synthetase was pneumonia with cardiorespiratory arrest. Following identified; recessive pathogenic EARS2 variants cause

2 Neurology: Genetics Figure Biochemical and molecular genetic characterization of the de novo c.991C>T p.(Arg331Trp) CTBP1 variant

(A) Mitochondrial respiratory chain activities normalized to citrate synthase for complexes I-IV in patient’s skeletal muscle (green) compared with controls (red). Asterisk denotes decreased complex I and IV activities. (B) Family pedigree and Sanger sequencing confirming that the c.991C.T p.(Arg331Trp) variant arose de novo. (C) Multiple sequence alignment demon- strating high conservation of the Arg331 residue. the mitochondrial disorder leukoencephalopathy with c.991C.T p.(Arg331Trp) variant had arisen de no- thalamus and brainstem involvement and high vo or 1 parent was low level mosaic (figure, B). The lactate.11 However, the c.452A.Gp.(Lys151Arg) c.991C.T p.(Arg331Trp) variant was absent from variant was not previously reported as pathogenic. our in-house exomes and external databases, affected In addition, Sanger sequencing of all exons and a highly conserved residue, and was predicted to be intronic regions of EARS2 did not identify a second deleterious by PolyPhen-2, SIFT, and Align GVGD candidate variant. Examination of WES read cov- (figure, C). erage also did not disclose copy number variants or rearrangements involving EARS2. DISCUSSION We performed WES of a woman pre- Inspection of rare and novel heterozygous variants senting an early-onset, progressive neurodegenerative that were predicted to be damaging to protein func- disorder and suggested mitochondrial etiology on the tion revealed a c.991C.T p.(Arg331Trp) missense basis of decreased respiratory chain activities of change in CTBP1 (NM_001012614.1). Following complexes I and IV, with sparing of the complex II the report of 4 patients sharing the identical hetero- activity, in a muscle biopsy. Suspecting an autosomal zygous CTBP1 variant that arose de novo in each recessive disorder of mitochondrial translation, fil- case,8 Sanger sequencing with extracted blood DNA tering and interpretation of variants in nuclear genes from our patient and her parents confirmed that the encoding mitochondrially targeted proteins failed to

Neurology: Genetics 3 identify likely candidates. Only after the evaluation of with a recurrent de novo CTBP1 variant did not rare and novel heterozygous variants was a previously report respiratory chain defects in skeletal muscle, reported pathogenic c.991C.T p.(Arg331Trp) although one patient was investigated for “pathogenic CTBP1 variant identified and confirmed to have mitochondrial mutations”.8 As yet, it is unclear arisen de novo. All 4 previously reported patients whether mitochondrial dysfunction contributes to presented with comparable phenotypes of hypotonia, the pathogenesis of this disorder or is a secondary ataxia, severe developmental delay, intellectual dis- consequence. Therefore, future studies of mitochon- ability, and problems with weight gain.8 Similarly, drial function and morphology in tissue from affected our patient had generally low muscle tone, develop- patients could provide further insights into the path- mental regression, severe intellectual disability, and ologic mechanisms involved. was nonambulatory. Cerebellar atrophy was also Pathogenic de novo heterozygous variants have been noted in our patient and 2 reported patients. Of widely associated with early-onset phenotypes that interest, all previously described patients also had include autism and epileptic encephalopathies.18,19 striking tooth enamel defects manifesting as hypo- By contrast, reported de novo variants associated with plasia or discoloration. By contrast, our patient had early-onset mitochondrial disorders are rare. Recently, 5 normal teeth on clinical examination, although there patients sharing a de novo c.1582C.T p.(Arg528Trp) was no comment on the enamel. ATAD3A variant with global developmental delay, CTBP1 encodes C-terminal binding protein 1 hypotonia, optic atrophy, axonal neuropathy, and (CtBP1), a transcriptional first identified hypertrophic cardiomyopathy were identified.20 In addi- as an E1A oncogene binding protein,12 but has also tion, de novo c.239G.A p.(Arg80His) and c.703C.G been shown to associate with DNA-specific repressors p.(Arg235Gly) SLC25A4 variants were reported in 7 during human development and tumorigenesis,13,14 patients with neonatal-onset respiratory insufficiency thereby regulating gene expression and DNA repair. and severe mtDNA depletion in skeletal muscle.21 Can- Although CtBP1 does not localize to mitochondria, it didate variant filtering strategies of early-onset mito- has been identified as a key regulator of mitochondrial chondrial disorders investigated using next-generation function and morphology.15 Knockout of Ctbp1 in sequencing have typically prioritized autosomal recessive mouse embryonic fibroblasts (MEFs) resulted in or X-linked variants in nuclear genes encoding elongated mitochondria, swollen cristae, and decreased mitochondrial-localized proteins.2–5 Although the cellular adenosine triphosphate (ATP), oxygen con- diagnostic yield of early-onset mitochondrial disor- sumption, and mitochondrial membrane potential.15 ders using next-generation sequencing can be as high Altered mitochondrial function and morphology were as 60%, there remains a considerable proportion of attributed to transcriptional repression of Bcl-2- cases without a diagnosis. Similarly, we initially pri- associated X protein (Bax)byCtBP1.BAX has been oritized autosomal recessive candidate variants based demonstrated to be essential for Drp1-dependent fis- on the decreased complex I and IV activities in pa- sion,16 while also regulating mitochondrial-dependent tient’s skeletal muscle and an unremarkable family apoptosis in response to glucose deprivation, for which history. Although the respiratory chain defects due its expression is modulated by association of CtBP1 to the pathogenic CTBP1 variant are a likely second- with the BAX promotor.15 When Bax and Ctbp1 were ary consequence, our experience suggests that this by knocked out together in MEFs, this led to increased no means precludes the identification of pathogenic ATP production, oxygen consumption rate, and mem- variants in nonmitochondrial disease genes or an brane potential, confirming a role for CtBP1 in main- autosomal dominant etiology. taining mitochondrial functions. The identification of a recurrent pathogenic de The de novo c.991C.T p.(Arg331Trp) CTBP1 novo c.991C.T p.(Arg331Trp) CTBP1 variant variant occurs in the PLDLS C-terminal binding using WES in a patient with decreased complex I cleft, which is required for the recruitment of and IV activities serves to caution that inheritance DNA-binding factors and components of the CtBP1 of suspected early-onset mitochondrial disorders corepressor complex.17 Decreased complex I and IV could wrongly be assumed autosomal recessive. Fur- activities in skeletal muscle of our patient suggests thermore, possible identification of patients with that BAX expression is altered by mutated CtBP1, respiratory chain defects due to variants involving with the c.991C.T p.(Arg331Trp) variant acting “non-classical mitochondrial disease” genes should in a dominant-negative mechanism to disrupt mito- not be overlooked.6,7 Our observations suggest that chondrial function. CtBP1 may also regulate the pathogenic de novo heterozygous variants could be expression of additional nuclear genes involved in under-recognized in suspected early-onset mitochon- mitochondrial functions, including nuclear genes en- drial disease, emphasizing the need to evaluate het- coding essential mitochondrial protein synthesis erozygous variants with segregation studies in machinery. The original description of 4 patients apparently unsolved and prospective cases.

4 Neurology: Genetics AUTHOR CONTRIBUTIONS mitochondrial patients in clinical practice. J Inherit Metab E.W. Sommerville: analysis and interpretation of whole-exome sequenc- Dis 2015;38:437–443. ing, clinical and molecular genetic data, drafting and revision of the man- 6. Pyle A, Nightingale HJ, Griffin H, et al. Respiratory chain uscript, and figure preparation. C.L. Alston: analysis and interpretation of deficiency in nonmitochondrial disease. Neurol Genet molecular genetic data. A. Pyle: analysis and interpretation of whole- 2015;1:e6. doi: 10.1212/NXG.0000000000000006. exome sequencing data. L. He: analysis and interpretation of biochemical 7. Keogh MJ, Daud D, Pyle A, et al. A novel de novo data. G. Falkous: analysis and interpretation of histochemical and histo- STXBP1 mutation is associated with mitochondrial com- logic data. K. Naismith: clinical care of the patient and revision of the plex I deficiency and late-onset juvenile-onset parkinson- manuscript. P.F. Chinnery: study concept and design, and analysis and – interpretation of whole-exome sequencing data. R. McFarland and ism. Neurogenetics 2015;16:65 67. R.W. Taylor: study concept and design, analysis and interpretation of 8. Beck DB, Cho MT, Millan F, et al. A recurrent de novo clinical, molecular genetic and biochemical data, and drafting and revi- CTBP1 mutation is associated with developmental delay, sion of the manuscript. hypotonia, ataxia, and tooth enamel defects. Neuroge- netics 2016;17:173–178. STUDY FUNDING 9. Taylor RW, Schaefer AM, Barron MJ, McFarland R, The clinical and diagnostic mitochondrial service in Newcastle upon Tyne is Turnbull DM. The diagnosis of mitochondrial muscle funded by the UK NHS Highly Specialised Commissioners to provide the disease. Neuromuscul Disord 2004;14:237–245. “Rare Mitochondrial Disorders of Adults and Children” Service. 10. Kirby DM, Thorburn DR, Turnbull DM, Taylor RW. Biochemical assays of respiratory chain complex activity. DISCLOSURE Methods Cell Biol 2007;80:93–119. E.W. Sommerville reports no disclosures. C.L. Alston is supported by the 11. Steenweg ME, Ghezzi D, Haack T, et al. Leukoencephal- National Institute for Health Research (NIHR) doctoral fellowship opathy with thalamus and brainstem involvement and (NIHR-HCS-D12-03-04). A. Pyle, L. He, G. Falkous, and K. Naismith high lactate “LTBL” caused by EARS2 mutations. Brain report no disclosures. P.F. Chinnery has served on the editorial board of 2012;135:1387–1394. BRAIN; in addition, he is a Wellcome Trust Senior Fellow in Clinical 12. Schaeper U, Boyd JM, Verma S, Uhlmann E, Subrama- Science (101876/Z/13/Z) and a UK NIHR Senior Investigator, who receives support from the Medical Research Council Mitochondrial Biology Unit nian T, Chinnadurai G. Molecular cloning and character- (MC_UP_1501/2). R. McFarland has served on the scientific advisory boards ization of a cellular phosphoprotein that interacts with of the United Mitochondrial Disease Foundation and the Lily Foundation a conserved C-terminal domain of adenovirus E1A and receives research support from the Wellcome Centre for Mitochondrial involved in negative modulation of oncogenic transforma- Research (203105/Z/16/Z), the Medical Research Council Centre for Trans- tion. Proc Natl Acad Sci USA 1995;92:10467–10471. lational Research in Neuromuscular Disease Mitochondrial Disease Patient 13. Bergman LM, Blaydes JP. C-terminal binding proteins: Cohort (UK) (G0800674), the Lily Foundation, the Ryan Stanford Appeal, emerging roles in cell survival and tumorigenesis. Apopto- “ and the UK NHS Highly Specialised Rare Mitochondrial Disorders of Adults sis 2006;11:879–888. and Children” Service. R.W. Taylor receives research support from the Well- 14. Chinnadurai G. Transcriptional regulation by C-terminal come Centre for Mitochondrial Research (203105/Z/16/Z), the Medical binding proteins. Int J Biochem Cell Biol 2007;39:1593– Research Council Centre for Translational Research in Neuromuscular Dis- ease Mitochondrial Disease Patient Cohort (UK) (G0800674), the Lily Foun- 1607. dation, the UK NHS Highly Specialised “Rare Mitochondrial Disorders of 15. Kim JH, Youn HD. C-terminal binding protein main- Adults and Children” Service, and the UK NIHR Biomedical Research Cen- tains mitochondrial activities. Cell Death Differ 2009; tre for Ageing and Age-related disease award to the Newcastle upon Tyne 16:584–592. Foundation Hospitals NHS Trust. Go to Neurology.org/ng for full disclosure 16. Wu S, Zhou F, Zhang Z, Xing D. Bax is essential for forms. Drp1-mediated mitochondrial fission but not for mito- chondrial outer membrane permeabilization caused by Received May 19, 2017. Accepted in final form July 21, 2017. photodynamic therapy. J Cell Physiol 2011;226:530–541. 17. Kuppuswamy M, Vijayalingam S, Zhao L-J, et al. Role of REFERENCES the PLDLS-binding cleft region of CtBP1 in recruitment 1. Skladal D, Halliday J, Thorburn DR. Minimum birth of core and auxiliary components of the corepressor com- prevalence of mitochondrial respiratory chain disorders plex. Mol Cell Biol 2008;28:269–281. in children. Brain 2003;126:1905–1912. 18. Epi KC; Epilepsy Phenome/Genome P. De novo muta- 2. Taylor RW, Pyle A, Griffin H, et al. Use of whole-exome tions in epileptic encephalopathies. Nature 2013;501: sequencing to determine the genetic basis of multiple 217–221. mitochondrial respiratory chain complex deficiencies. JA- 19. Iossifov I, O’Roak BJ, Sanders SJ, et al. The contribution MA 2014;312:68–77. of de novo coding mutations to autism spectrum disorder. 3. Pronicka E, Piekutowska-Abramczuk D, Ciara E, et al. Nature 2014;515:216–221. New perspective in diagnostics of mitochondrial disorders: 20. Harel T, Yoon Wan H, Garone C, et al. Recurrent de novo two years’ experience with whole-exome sequencing at and biallelic variation of ATAD3A, encoding a mitochon- a national paediatric centre. J Transl Med 2016;14:174. drial membrane protein, results in distinct neurological 4. Kohda M, Tokuzawa Y, Kishita Y, et al. A comprehensive syndromes. Am J Hum Genet 2016;99:831–845. genomic analysis reveals the genetic landscape of mito- 21. Thompson K, Majd H, Dallabona C, et al. Recurrent chondrial respiratory chain complex deficiencies. PLoS de novo dominant mutations in SLC25A4 cause severe Genet 2016;12:e1005679. early-onset mitochondrial disease and loss of mitochon- 5. Wortmann S, Koolen D, Smeitink J, van den Heuvel L, drial DNA copy number. Am J Hum Genet 2016;99: Rodenburg R. Whole exome sequencing of suspected 860–876.

Neurology: Genetics 5 Children with 59-end NF1 gene mutations are more likely to have glioma

Corina Anastasaki, PhD ABSTRACT Stephanie M. Morris, MD Objective: To ascertain the relationship between the germline NF1 gene mutation and glioma Feng Gao, MD, PhD development in patients with neurofibromatosis type 1 (NF1). David H. Gutmann, MD, Methods: The relationship between the type and location of the germline NF1 mutation and the PhD presence of a glioma was analyzed in 37 participants with NF1 from one institution (Washington University School of Medicine [WUSM]) with a clinical diagnosis of NF1. Odds ratios (ORs) were calculated using both unadjusted and weighted analyses of this data set in combination with 4 Correspondence to Dr. Gutmann: previously published data sets. [email protected] Results: While no statistical significance was observed between the location and type of the NF1 mutation and glioma in the WUSM cohort, power calculations revealed that a sample size of 307 participants would be required to determine the predictive value of the position or type of the NF1 gene mutation. Combining our data set with 4 previously published data sets (n 5 310), children with glioma were found to be more likely to harbor 59-end gene mutations (OR 5 2; p 5 0.006). Moreover, while not clinically predictive due to insufficient sensitivity and specificity, this asso- ciation with glioma was stronger for participants with 59-end truncating (OR 5 2.32; p 5 0.005) or 59-end nonsense (OR 5 3.93; p 5 0.005) mutations relative to those without glioma. Conclusions: Individuals with NF1 and glioma are more likely to harbor nonsense mutations in the 59 end of the NF1 gene, suggesting that the NF1 mutation may be one predictive factor for glioma in this at-risk population. Neurol Genet 2017;3:e192; doi: 10.1212/NXG.0000000000000192

GLOSSARY CI 5 confidence interval; NF1 5 neurofibromatosis type 1; OR 5 odds ratio; WUSM 5 Washington University School of Medicine.

Neurofibromatosis type 1 (NF1; OMIM162200) is characterized by substantial clinical variabil- ity, with individuals prone to the development of numerous medical complications, ranging from neurofibromas and bone defects to autism and nervous system tumors. The wide pheno- typic variation seen in this disorder creates a particular challenge for clinicians when counseling families. For this reason, there is a pressing need to identify potential predictive risk factors. NF1 is caused by germline pathogenic variants in the NF1 locus, containing a large gene with 57 exons.1 While thousands of different NF1 gene mutations have been identified, recent studies have suggested that genotype-phenotype correlations may exist. In particular, 2 types of NF1 mutations have been described, in which patients do not develop neurofibromas (c.2970_72de- lAAT2; p.Arg18093). However, several groups have evaluated the predictive value of the location of the NF1 mutation as a potential risk factor for optic glioma with conflicting conclusions.4–7 An important limitation to these analyses is the small sample size of each cohort. This is particularly relevant to brain tumors, which arise in only ;20% of children with NF1.8,9 For this reason, we combined participants from our own institution (Washington University School of Medicine [WUSM] cohort) with the 4 previously published studies that used MRI scans to Supplemental data at Neurology.org/ng From the Department of Neurology (C.A., S.M.M., D.H.G.) and Department of Surgery (F.G.), Washington University in St. Louis, MO. Funding information and disclosures are provided at the end of the article. Go to Neurology.org/ng for full disclosure forms. The Article Processing Charge was funded by the authors. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Neurology.org/ng Copyright © 2017 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 determine the presence or absence of glio- type or location and glioma diagnosis was found mas.4–7 As such, we performed adequately when each of the 3 previously published cohorts that powered unadjusted and weighted analyses contained participants with and without glioma were 4,6,7 to determine the relationship between the type analyzed individually (table e-2). A priori power analysis using the G*power and location of the NF1 mutation and glioma 3.0.10 program (Universität Kiel, Germany) indi- in over 300 individuals with NF1. cated that a total sample of 307 participants would be needed to detect an effect size of 0.16 with 80% METHODS Study population. This retrospective analysis power using a x2 test to detect differences in pro- was performed using pre-existing clinical data in the electronic medical records under an approved Human Studies protocol at portions with an alpha level of 0.05. Since each of the WUSM. Only those participants with a known NF1 gene the studies are limited by small participant numbers, mutation (performed at the University of Alabama, Birmingham we combined our data with the 4 previously pub- Medical Genomics Laboratory) and brain MRI (performed at lished data sets, and after excluding participants with ’ Barnes-Jewish Hospital/St. Louis Children s Hospital) were unknown mutations or genomic microdeletions, included in this study. A total of 37 participants met these cri- 101 participants with “glioma” and 209 controls teria: 14 participants were identified with glioma (optic, cere- “ ” bellar, brainstem, and temporal lobe glioma) using previously with a no glioma diagnosis were available (table published radiographic criteria,8 while 23 did not (negative scan e-3). Using this combined data set, more 59-end after age 10 years). Deidentified data analyzed included patient (exons 1–26) mutations were found in participants sex (male/female), family history of NF1 (yes/no), NF1 gene with glioma than in controls (67.3% vs 50.7%; OR 5 mutation (DNA and predicted protein change), and glioma 2.00; 95% CI: 1.22–3.29; p 5 0.006) (figure 1A). (presence/absence) (table e-1 at Neurology.org/ng). In addition, this significance persisted when the Standard protocol approvals, registrations, and patient OR was adjusted for sample size and heterogeneity consents. This is a retrospective analysis performed using pre- using a meta-analysis with control groups (n 5 existing clinical data in the electronic medical records under 296) (OR 5 2.21; 95% CI: 1.22–4.00; p 5 a WUSM-approved Human Studies protocol (IRB#201703143). 0.009; figure 1B and table e-4A). As such, no patient consent forms were required. When mutations were stratified by type, the Statistical analysis. All categorical variables were analyzed using unadjusted OR for glioma was 3.93 (95% CI: x2 or Fisher exact tests (SPSS, v23). Odds ratios (ORs) were 1.52–10.18; p 5 0.005; figure 2A), and following reported with 95% confidence intervals (CIs), estimated using 2 3 2 contingency tables to compare outcomes (presence of meta-analysis, the weighted OR was 3.46 (95% CI: – 5 glioma) with the variant type and location within the NF1 gene. 1.19 10.10; p 0.023; figure 2B and table e-4B) Sensitivity analyses were performed by meta-analysis,4,6,7 exclud- forparticipantswith59-end nonsense mutations. ing 1 data set that lacked a control group.5 Adjusted ORs and Similarly, the unadjusted OR for glioma was 2.32 associated 95% CIs were calculated using weights, based on the (95% CI: 1.23–4.18; p 5 0.005; figure 2C), while inverse of the variance, while individual studies included in the the adjusted OR was 2.65 (95% CI: 1.12–6.24; p 5 meta-analysis were assumed to be random samples of the target population. Statistical significance was defined as a p value of 0.026; figure 2D and table e-4C) in participants 9 ,0.05 (2 sided). with 5 -end truncation mutations relative to those with 39-end mutations. When the 3 WUSM pa- RESULTS A total of 37 individuals with identified tients with cerebellar brainstem and temporal lobe NF1 gene mutations were included (4 splice, 3 mis- glioma were excluded from the analyses, the signif- sense, 18 nonsense, 9 frameshift deletion, and 3 icance of the findings was unchanged (table e-5). Of frameshift insertion mutations; WUSM cohort; table note, when we expanded the 59 end to include mu- e-1). Based on radiographic criteria,8 14 participants tations within the RAS-GAP domain (exons 1–34), had a glioma, while 23 participants did not. There the significance observed in unadjusted ORs for gli- was no association between patient sex and family oma (table e-6) was lost when weighted analyses history of NF1 and glioma, and 59-end (exons 1– were performed (table e-6). 261) clustering of mutations in patients with glioma The utility of these findings in the clinic setting was not observed. In addition, no differences were depends on their specificity and sensitivity in predict- observed when mutations were stratified by variant ing the presence or absence of a glioma. Unfortu- type (table e-1). While there was a larger proportion nately, as stand-alone factors, the position and type of participants with glioma who harbored mutations of the NF1 gene mutation lack sufficient specificity predicted to cause premature protein truncation (0.49, 0.71, and 0.54) or sensitivity (0.67, 0.62, and (nonsense, frameshift insertions/deletions) compared 0.67 for all 59-end, 59-end nonsense, and 59-end with participants without glioma (64.3% vs 30.4%), truncation mutations, respectively) to enable accurate no statistically significant difference between groups risk assessment. This point is underscored by the was observed (p 5 0.09). Similarly, with one excep- finding that the same mutation can be found in par- tion,6 no significant association between the mutation ticipants with and without glioma (table e-7).

2 Neurology: Genetics Figure 1 Distribution of NF1 gene mutations in neurofibromatosis type 1 participants with and without glioma

(A) Histogram indicating the location of NF1 gene mutations identified in participants with glioma (n 5 101; black bars) or without glioma (n 5 209; gray bars). The different regions of the NF1 gene are highlighted below the exons (59 end: exons 1– 26 and 39 end: exons 27–57, including the RAS-GAP domain). The specificity, sensitivity, and unadjusted odds ratio (OR) for glioma in participants with any NF1 gene mutation are shown in the table above the graph. (B) Summary of the combined weighted OR after sensitivity meta-analysis (n 5 296 patients). CI 5 confidence interval; WUSM 5 Washington University School of Medicine.

DISCUSSION In the era of precision medicine, it be- we perform the largest genotype-phenotype analysis to comes critical to identify risk factors that influence dis- determinewhetherthepositionortypeoftheNF1 gene ease pathogenesis and clinical variability. This problem mutation is a risk factor for glioma. is illustrated by NF1, where no robust predictive factors First, NF1 participants who harbor NF1 muta- have been identified. Based on insights derived from the tions proximal to the RAS-GAP domain are more use of NF1-patient–induced pluripotent stem cells10 likely to develop gliomas, similar to two previously coupled with emerging genotype-phenotype correla- published reports.5,6 Of interest, NF1 mutations pre- tions,2,3 one potential factor important for conferring dicted to produce premature stop codons were also disease variability is the germline NF1 mutation. Here, associated with higher glioma risk.4 This mutational

Neurology: Genetics 3 Figure 2 Specificity, sensitivity, and odds ratios for glioma in participants with neurofibromatosis type 1 and nonsense or truncating NF1 gene mutations

(A) Summary of the specificity, sensitivity, and unadjusted odds ratio (OR) for glioma in participants with nonsense mutations in the 59 end of the NF1 gene (exons 1–26). (B) Summary of the combined weighted OR for glioma after sensitivity meta-analysis (n 5 73 patients). (C) Summary of the specificity, sensitivity, and unadjusted OR for glioma in participants with truncating mutations in the 59 end of the NF1 gene (exons 1–26). (D) Summary of the combined weighted OR for glioma after sensitivity meta-analysis (n 5 193). CI 5 confidence interval; WUSM 5 Washington University School of Medicine.

specificity could reflect the impact of the mutation on information to families. Future studies that incorpo- protein expression10; however, this hypothesis will rate multiple independently significant risk factors require further mechanistic exploration. may yield useful predictive assessment profiles for Second, each of the previous analyses used differ- children with NF1. ent statistical methods to look for genotype- phenotype correlations and included only modest AUTHOR CONTRIBUTIONS numbers of participants. It is important that future C.A. collated the data, performed the analyses, and wrote the manu- script. S.M.M. performed the statistical analyses and participated in studies incorporate power calculations to determine the generation of the final manuscript. F.G., as the study biostatisti- whether the analyses planned are sufficiently powered cian, performed the meta-analysis and reviewed all biostatistical anal- to detect differences. One limitation inherent in the yses. D.H.G. collated the data from Washington University, combined analysis of multiple data sets is that the pre- provided the funding, and was responsible for generation of the final manuscript. cise radiographic criteria used by each institution to diagnose gliomas could vary slightly. Moreover, it is STUDY FUNDING critical to use clinically applicable analyses, such as D.H.G. is supported by a Research Program Award from the NINDS OR calculations, as measures of risk factor (1-R35-NS097211-01), while S.M.M. is supported by the Neurological significance. Sciences Academic Development Award (K12 NS001690). Finally, while 59-end mutations, and 59 nonsense DISCLOSURE or truncation mutations in particular, are associated C. Anastasaki reports no disclosures. S.M. Morris has received research with glioma, neither the location nor the mutation support from the NIH. F. Gao reports no disclosures. D. Gutmann type is a highly specific marker of glioma (figure 1 and has served on the American Association for Cancer Research grant review tables e-2 and e-4–e-7). We interpret this result to panel and the National Institute for Neurological Disorders and Stroke Advisory Council; has served on the editorial boards of Glia and Familial indicate that the germline mutation alone is not a suf- Cancer; holds the following patents: Identification of the NF1 gene and ficiently robust risk factor to provide prognostic protein (NF1 Gene, 1999) Identification of NF1 protein as mTOR

4 Neurology: Genetics regulator (mTOR, 2012); has received research support from US Army 4. Ars E, Kruyer H, Morell M, et al. Recurrent mutations in Department of Defense, NCI Tumor Microenvironment Network the NF1 gene are common among neurofibromatosis type (TMEN) U01, the NIH, The Giorgio Foundation, the NF1-Dermal 1 patients. J Med Genet 2003;40:e82. Neurofibroma Consortium, the Children’s Tumor Foundation Synodos, 5. Sharif S, Upadhyaya M, Ferner R, et al. A molecular anal- and the Low-Grade Glioma Consortium; receives license fee payments ysis of individuals with neurofibromatosis type 1 (NF1) for TSC1 knockout mouse; and receives royalty payments from the University of Michigan. Go to Neurology.org/ng for full disclosure and optic pathway gliomas (OPGs), and an assessment of forms. genotype-phenotype correlations. J Med Genet 2011;48: 256–260. Received June 16, 2017. Accepted in final form August 15, 2017. 6. Bolcekova A, Nemethova M, Zatkova A, et al. Clustering of mutations in the 5’ tertile of the NF1 gene in Slovakia REFERENCES patients with optic pathway glioma. Neoplasma 2013;60: – 1. Anastasaki C, Le LQ, Kesterson RA, Gutmann DH. Up- 655 665. dated nomenclature for human and mouse neurofibroma- 7. Hutter S, Piro RM, Waszak SM, et al. No correlation tosis type 1 genes. Neurol Genet 2017;3:e169. doi: 10. between NF1 mutation position and risk of optic pathway 1212/NXG.0000000000000169. glioma in 77 unrelated NF1 patients. Hum Genet 2016; – 2. Upadhyaya M, Huson SM, Davies M, et al. An absence of 135:469 475. cutaneous neurofibromas associated with a 3-bp inframe 8. Mahdi J, Shah AC, Sato A, et al. A multi-institutional deletion in exon 17 of the NF1 gene (c.2970-2972 study of brainstem gliomas in children with neurofibro- – delAAT): evidence of a clinically significant NF1 genotype- matosis type 1. Neurology 2017;88:1584 1589. phenotype correlation. Am J Hum Genet 2007;80:140– 9. Listernick R, Ferner RE, Liu GT, Gutmann DH. Optic 151. pathway gliomas in neurofibromatosis-1: controversies and – 3. Rojnueangnit K, Xie J, Gomes A, et al. High incidence of recommendations. Ann Neurol 2007;61:189 198. noonan syndrome features including short stature and pul- 10. Anastasaki C, Woo AS, Messiaen LM, Gutmann DH. monic stenosis in patients carrying NF1 missense mutations Elucidating the impact of neurofibromatosis-1 germline affecting p.Arg1809: genotype-phenotype correlation. Hum mutations on neurofibromin function and dopamine- Mutat 2015;36:1052–1063. based learning. Hum Mol Genet 2015;24:3518–3528.

Neurology: Genetics 5 Clinical/Scientific Notes

Andrea Accogli, MD NOVEL AMPD2 MUTATION IN common dysmorphic features, including sloping Michele Iacomino, MS PONTOCEREBELLAR HYPOPLASIA, forehead, large and posterior rotated ears, upper Francesca Pinto, MD DYSMORPHISMS, AND TEETH ABNORMALITIES lateral incisor agenesis, mottled and fragile teeth Alessandro Orsini, MD with multiple cavities, mandibular hypoplasia, and – Maria Stella Vari, MD Pontocerebellar hypoplasias (PCHs) are a genetically clinodactyly of the V fingers (figure 1, A C). In all Raed Selmi, MD and clinically heterogeneous group of autosomal individuals, brain MRI showed PCH, severe callosal Annalaura Torella, PhD recessive inherited neurodevelopmental disorders. In hypoplasia, leukoencephalopathy, and basal ganglia – Vincenzo Nigro, MD the past decade, whole-exome sequencing (WES) involvement (figure 1, D F). Diffusion tensor Carlo Minetti, MD has led to the identification of new genes, allowing imaging revealed marked thinning of the cortico- Mariasavina Severino, the recognition of at least 10 different PCH types spinal tracts (figure e-2), and MR spectroscopy MD with broad and overlapping phenotypes.1 showed low N-acetyl aspartate in the thalamic Pasquale Striano, MD, PCH type 9 (PCH9) (MIM 615809) is caused by region (data not shown) in patient II-3. Electro- PhD homozygous mutations in the adenosine monophos- physiologic studies performed in individual IV-2 Valeria Capra, MD phate deaminase 2 (AMPD2) gene. did not reveal signs of axonal neuropathy. At the Federico Zara, PhD AMPD2 plays an evolutionary conserved role in final follow-up, all patients showed cortical blind- purine nucleotide metabolism by regulating the gua- ness and were using wheelchair or were bedridden. Neurol Genet Their clinical and imaging details are summarized in 2017;3:e179; doi: 10.1212/ nine nucleotide biosynthesis and protein transla- NXG.0000000000000179 tion.2 Patients typically present with severe table e-1. developmental delay, microcephaly, axonal neurop- Genetic study. We first excluded copy number and athy, and epilepsy. Neuroimaging in patients with structural DNA variations by karyotyping and array– “ ” PCH9 shows a typical midbrain figure of 8 comparative genomic hybridization. We then per- appearance, callosal hypoplasia, and periventricular formed WES on DNA samples of all family members, 3 white matter involvement. using a standard Illumina pipeline (e-Methods, tables Previous reports have demonstrated the pivotal e-2 and e-3) and identified a novel homozygous role of AMPD2 during neurogenesis and showed c.495delG (p.R165fs*21) in the a potentially rescue therapy in vitro by administration gene encoding for the adenosine monophosphate 2 of purine precursors. Since the original report of 5 deaminase 2 enzyme (AMPD2, NM_001257360.1). 2 PCH9 families, only 6 additional patients from 2 Sanger sequencing showed that the mutation segre- 3,4 families have been described. Furthermore, a homo- gates according to a recessive model of inheritance zygous frameshift mutation in the AMPD2 gene has (figure e-2). The institutional review board approved been identified in 2 members of a consanguineous the use of human samples for this study. family affected by spastic paraplegia without PCH.5 We report a novel homozygous AMPD2 mutation Discussion. AMPD2 encodes 1 of 3 adenosine in 3 siblings with severe PCH9 phenotype. monophosphate (AMP) deaminase enzyme homo- logs, which convert AMP to IMP. AMPD2 defi- ciency results in accumulation of adenosine Clinical description. The probands are 2 girls nucleotides and depletion of guanine nucleotide, im- (IV:1, aged 9 years and IV:3 aged 8 years) and 1 pairing the guanosine-5’-triphosphate–dependent boy (IV:2, aged 7 years) born to first-cousin initiation of protein translation, similarly to what parents from the Middle East (figure e-1 at was observed in transfer RNA splicing endonuclease Supplemental data Neurology.org/ng). Family history was unremark- complex genes, linked to other PCH types.2,6 at Neurology.org/ng able. All individuals were born at term by normal Recently, AMPD2 mutant mice and Ampd22/2 delivery, following an uneventful pregnancy. Since mice have been associated with nephrotic syndrome the first months of life, developmental delay with- and proteinuria in the absence of any brain abnor- out any motor skills acquisition, drug-resistant focal mality.7 A neurodegenerative phenotype has been and tonic-clonic seizures, and progressive spasticity observed when both Ampd2 and Ampd3 are knocked were noticed. All patients shared microcephaly and out, suggesting a functional redundancy among AMP

Neurology.org/ng Copyright © 2017 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology 1 Figure 1 Clinical and neuroradiologic features of patients carrying homozygous AMP2 mutations

(A–C) Facial images demonstrate shared features of microcephaly, sloping forehead, large and posterior rotated ears, and mandibular hypoplasia. In insets, mottled teeth with multiple cavities. (D–F) Brain MRIs from each patient showing characteristic “figure of 8” midbrain appearance (dotted ovals) and small hyperintense basal ganglia and thalami (arrowheads) on axial T2-weighted images. Sagittal T1-weighted images show small pons (red arrows) and extremely severe callosal hypoplasia (empty arrows). Coronal FLAIR images reveal hypoplasia/atrophy of the cerebellar hemispheres (arrowheads) with relative sparing of the vermis. Leukoencephalopathy similar to periventricular leukomalacia present in all patients, with loss of white matter bulk, periven- tricular hyperintensity, and enlarged lateral ventricles (empty arrows). FLAIR 5 fluid-attenuated inversion recovery.

deaminase homologs. However, it remains to be abnormalities that might be an additional feature elucidated whether humans carrying AMPD2 muta- associated with PCH9. Genetic studies did not show tions may have renal involvement. any additional homozygous variant that could readily To date, 14 patients with PCH9 have been re- explain these findings (table e-3). Further studies will ported. All but 1 AMPD2 mutation reside within clarify the phenotypic spectrum associated with the conserved catalytic AMP deaminase domain2–4 AMPD2 mutations. (table e-1). The novel identified homozygous From the Istituto G. Gaslini (A.A., M.I., F.P., A.O., M.S.V., C.M., AMPD2 mutation (c.495delG p.R165RfsX21) lies M.S., P.S., V.C., F.Z.), Genova; Università degli Studi di Genova outside the catalytic domain and is associated with (A.A., M.I., C.M., P.S.); Ospedale San Paolo (R.S.), Milano, Italy; a very severe phenotype as described for c.751C.T Dipartimento di Biochimica Biofisica e Patologia Generale (A.T., V.N.), Seconda Università di Napoli; and Telethon Institute of (p.R251W).4 Neuroimaging revealed the involve- Genetics and Medicine (A.T., V.N.). ment of the basal ganglia and thalami and marked Author contributions: A.A., P.S., V.C., and F.Z. contributed to the hypoplasia/atrophy of the corticospinal tracts, unrav- conception and design of the study and the acquisition, analysis, and eling the anatomical basis of the characteristic mid- interpretation of data. M.I., A.T., and V.N. performed NGS data brain “figure of 8” appearance and pontine flattening analysis. F.P., A.O., M.S.V., R.S., C.M., and M.S. contributed to the acquisition, analysis, and interpretation of data. All authors observed in PCH9 individuals. However, our patients contributed to the critical revision of the final version of the manu- showed peculiar facial dysmorphisms and teeth script for important intellectual content.

2 Neurology: Genetics Study funding: F.Z. has been supported by the Italian Ministry of 1. Rudnik-Schoneborn S, Barth PG, Zerres K. Pontocerebellar Health (RF-2010-2314356). P.S. has been supported by the Italian hypoplasia. Am J Med Genet C Semin Med Genet 2014; Ministry of Health (GR-2011-02346749). A.A., V.C., M.I., A.T., 166C:173–183. V.N., F.P., A.O., M.S.V., R.S., C.M., and M.S. report no disclo- 2. Akizu N, Cantagrel V, Schroth J, et al. AMPD2 regulates sures. The Article Processing Charge was funded by the authors. GTP synthesis and is mutated in a potentially treatable neu- Disclosure: A. Accogli, M. Iacomino, F. Pinto, A. Orsini, M. S. Vari, rodegenerative brainstem disorder. Cell 2013;154:505–517. R. Selmi, and A. Torella report no disclosures. V. Nigro has served on 3. Marsh AP, Lukic V, Pope K, et al. Complete callosal agen- the editorial board of Neuromuscular Disorders. C. Minetti and M. Severino report no disclosures. P. Striano has served on the sci- esis, pontocerebellar hypoplasia, and axonal neuropathy due entific advisory board of the Italian Agency of the Drug (AIFA) and to AMPD2 loss. Neurol Genet 2015;1:e16. has served on the editorial board of Epilepsia. V. Capra reports no 4. Marsh AP, Yap P, Tan T, et al. A novel AMPD2 mutation disclosures. F. Zara has received research support from the Italian outside the AMP deaminase domain causes pontocerebellar Ministry of Health. Go to Neurology.org/ng for full disclosure forms. hypoplasia type 9. Am J Med Genet A 2017;173:820–823. This is an open access article distributed under the terms of the 5. Novarino G, Fenstermaker AG, Zaki MS, et al. Exome Creative Commons Attribution-NonCommercial-NoDerivatives Li- sequencing links corticospinal motor neuron disease to com- cense 4.0 (CC BY-NC-ND), which permits downloading and shar- mon neurodegenerative disorders. Science 2014;343:506–511. ing the work provided it is properly cited. The work cannot be 6. Budde BS, Namavar Y, Barth PG, et al. tRNA splicing changed in any way or used commercially without permission from endonuclease mutations cause pontocerebellar hypoplasia. the journal. Nat Genet 2008;40:1113–1118. Received March 30, 2017. Accepted in final form June 30, 2017. 7. Helmering J, Juan T, Li CM, et al. A mutation in Ampd2 is associated with nephrotic syndrome and hypercholesterol- Correspondence to Dr. Zara: [email protected] emia in mice. Lipids Health Dis 2014;13:167.

Neurology: Genetics 3 Clinical/Scientific Notes

Yi Shiau Ng, PhD, NOVEL POLG VARIANTS ASSOCIATED WITH and CSF analysis were unremarkable, except MRCP LATE-ONSET DE NOVO STATUS EPILEPTICUS a slightly raised serum lactate level at 3.3 mol/L Helen Powell, BSc AND PROGRESSIVE ATAXIA (normal: ,2.2 mmol/L). EEG showed encephalo- Nigel Hoggard, MD pathic changes. MRI head T2 and fluid attenuation Doug M. Turnbull, PhD, Mitochondrial disease is phenotypically and geneti- and inversion recovery sequences revealed stroke- FRCP cally heterogeneous with an estimated prevalence of like lesions (figure, A), in addition to the previously Robert W. Taylor, PhD, 1 in 4,300.1 Mutations in the POLG gene, encoding documented changes in the clinic (figure, B). Epi- FRCPath the catalytic subunit of DNA polymerase gamma, are lepsia partialis continua, affecting the left face, arm, Marios Hadjivassiliou, an important cause of mitochondrial disease. The and leg, emerged on day 12 of admission. Her seiz- MD spectrum of clinical manifestations in POLG-related ures were suprarefractory to treatment, despite mitochondrial disease is variable,2 with disease onset receiving a combination of phenytoin, levetirace- Neurol Genet tam, clonazepam, propofol, midazolam, and pulse 2017;3:e181; doi: 10.1212/ ranging from adulthood-onset dominant or recessive NXG.0000000000000181 progressive external ophthalmoplegia (chronic progres- methylprednisolone. She died of worsening epilep- sive external ophthalmoplegia), ataxia-neuropathy tic encephalopathy and multiorgan failure after 2 spectrum, myoclonic epilepsy, myopathy, and sensory weeks of hospitalization. ataxia to childhood-onset Alpers syndrome, which is This patient was tested negative for common mito- characterized by intractable seizures, psychomotor chondrial DNA (mtDNA) point mutations, including . . . regression, and hepatic impairment. Epilepsy is a poor m.3243A G, m.8344A G, and m.8993T C/G. prognostic factor in POLG mutations,3 and the onset She was also tested negative for 3 common POLG of epilepsy often clusters in childhood (,5years)and mutations (p.Ala467Thr, p.Trp748Ser, and p. teenage.4 However, late-onset epileptic encephalopathy Gly848Ser). Her muscle biopsy revealed histochemical is uncommon.4,5 Herein, we describe a patient who and molecular genetic evidence of mitochondrial dys- – died of de novo, late-onset refractory status epilepticus function, including cytochrome c oxidase deficient with the identification of 2 novel variants in the POLG fibers (figure, C) and variable mtDNA deletions (fig- gene. ure, D). No pathogenic variant was identified in TWNK and RRM2B. Direct sequencing of the POLG Case report. A 69-year-old woman presented with gene (GenBank accession number NM_002693.2) an 8-year history of slowly progressive gait ataxia identified 2 rare variants, c.1232T.C, p.(Leu411Pro) associated with dysarthria to the regional ataxia cen- and c.1721G.A, p.(Arg574Gln), both affecting ter. She also noted to have generalized myoclonic conserved amino acids and predicted to be damaging jerks for 9 months. There was no other medical his- (figure, E). Familial segregation studies were not fea- tory or relevant family history of any neurologic dis- sible, as she was the only child and both her parents order. On examination, she had evidence of were deceased. ophthalmoplegia in all directions of gaze. She was found to have prominent gait and lower limb ataxia. Discussion. Our patient’s initial presentation of Myoclonus was demonstrable with outstretched a progressive cerebellar ataxia plus other neurologic arms. Reflexes were present and symmetrical. She was features including external ophthalmoplegia and just able to walk with a stick and required 1 person’s myoclonus is highly suggestive of a mitochondrial assistance. Mitochondrial disease was suspected, and etiology. Moreover, her neuroimaging findings of she underwent a muscle biopsy. bilateral signal abnormalities in thalami, cerebellar She was admitted acutely to the hospital following dentate nuclei, and cerebellar atrophy have previously 2 episodes of generalized tonic-clonic seizures at age been reported in POLG-related mitochondrial dis- 71. She was treated with IV phenytoin and levetirace- ease.6 However, the development of fatal epileptic tam. Her management was rapidly escalated to the encephalopathy is rather surprising, given the insidi- administration of general anesthesia due to convulsive ous onset of her illness. Our case highlights the pro- status epilepticus. Laboratory investigations including gressive nature of POLG-related mitochondrial routine biochemistry, autoantibodies, septic screens, disease, the overlap of clinical syndromes and

Neurology.org/ng Copyright © 2017 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Figure Neuroimaging, muscle biopsy, and molecular genetic findings

(A) Head MRI performed at age 71 following admission in status epilepticus. Diffusion-weighted imaging sequence showed restricted diffusion in occipital, parietal, and frontal lobes, thalami, and with low ADC map in the right frontal lobe (red arrow). (B) Head MRI performed at onset of ataxia aged 69. Sagittal T1 view showed cerebellar atrophy and axial T2 view showed symmetrical hyperintensities in the cerebellar dentate nuclei and thalami. (C) Sequential cytochrome c oxidase (COX)–succinate dehydrogenase histochemistry demonstrates a mosaic distribution of COX-deficient muscle fibers (blue) among fibers exhibiting normal COX activity (brown). (D) Long range PCR amplification of muscle DNA across the major arc confirms multiple mitochondrial DNA (mtDNA) deletions in patient muscle (lane 1) compared with age-matched control muscle (lane 2); MW 5 molecular weight marker. (E) Alignments of mutation-containing POLG regions across multiple species show the evolutionary conservation of the heterozygous c.1232T.C, p.(Leu411Pro) and c.1721G.A, p. (Arg574Gln) missense POLG variants. The c.1232T.C, p.(Leu411Pro) variant is absent from both the ExAC browser (exac. broadinstitute.org)and the NHLBI ESP (evs.gs.washington.edu/EVS/) database (both accessed on October 6, 2016), thus representing a novel missense change, while the c.1721 G.A, p.(Arg574Gln) variant has only identified in 3/120480 alleles on the ExAc browser. A different POLG variant affecting the same amino acid c.1720C.T, p.(Arg574Trp) has been pre- viously reported in trans with other known pathogenic variants in 4 unrelated patients, according to the Human DNA Polymerase Gamma Mutation Database (tools.niehs.nih.gov/polg/index.cfm/main/search) (accessed on May 17, 2017). Affected amino acids are highlighted by an asterisk; sequence identity is shown by bold, red typeface.

difficulty of predicting the trajectory of disease pro- speculate that our patient had a late-onset recessive gression, and the management challenge of refractory POLG disease, given that recessive POLG disease is mitochondrial epilepsy.4 The presence of focal onset more common than dominant presentations accord- motor status, together with the acute stroke-like le- ing to our experience and reported cases in the liter- sions, is likely related to the neuronal energy failure6 ature. Both variants are located in the linker domain of which inhibitory interneurons have been shown of POLG, and we have recently showed that muta- to be particularly vulnerable to mitochondrial tions (homozygous or compound heterozygous) in dysfunction.7 this region are associated with later disease presenta- We were unable to unequivocally conclude tion and longer survival compared with other do- whether these 2 variants were in cis or in trans. We mains within the POLG protein.4

2 Neurology: Genetics We propose that POLG-related mitochondrial dis- to the Newcastle upon Tyne Foundation Hospitals NHS Trust, and “ ease should be a differential diagnosis in cases of de novo the UK NHS Highly Specialized Rare Mitochondrial Disorders of Adults and Children” Service. M. Hadjivassiliou has served on the status epilepticus, particularly with other clinical features editorial board of Cerebellum & Ataxias and has been a member of such as ataxia and external ophthalmoplegia, irrespective the medical advisory boards of Ataxia UK and Coeliac UK. Go to Neurology.org/ng for full disclosure forms. The Article Processing of age. Full sequencing of POLG should be performed Charge was funded by the Wellcome Centre for Mitochondrial because more than 20% of patients do not carry 1 of the Research. 3commonmutations,4 as exemplified by this case. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (CC BY), which permits From the Wellcome Centre for Mitochondrial Research (Y.S.N., unrestricted use, distribution, and reproduction in any medium, pro- H.P., D.M.T., R.W.T.), Institute of Neuroscience, Newcastle vided the original work is properly cited. University, Newcastle upon Tyne; and Sheffield Teaching Hospitals NHS Trust and University of Sheffield (N.H., M.H.), Royal Hal- Received March 15, 2017. Accepted in final form June 30, 2017. lamshire Hospital, Sheffield, United Kingdom. Correspondence to Dr. Hadjivassiliou: [email protected] Author contributions: Y. S. Ng: analysis and interpretation of data and drafting and revising the manuscript. H. Powell: acquisition of 1. Gorman GS, Schaefer AM, Ng Y, et al. Prevalence of data and genetic analysis and interpretation of data. N. Hoggard: nuclear and mitochondrial DNA mutations related to adult acquisition, analysis, and interpretation of imaging data. D. M. mitochondrial disease. Ann Neurol 2015;77:753–759. Turnbull: interpretation of data and revising the manuscript. 2. Cohen BH, Chinnery PF, Copeland WC. POLG-related R. W. Taylor: study concept and design, analysis and interpretation disorders. In: Pagon RA, Adam MP, Ardinger HH, et al, of data, and revising the manuscript. M. Hadjivassiliou: study con- editors. GeneReviews® [Internet]. Seattle: University of cept and design, analysis and interpretation of data, and revising the Washington, Seattle; 2010: 1993–2016. Available at: manuscript. http://www.ncbi.nlm.nih.gov/books/NBK26471/. Accessed Acknowledgement: The clinical and diagnostic mitochondrial services December 18, 2014. in Newcastle upon Tyne is funded by the UK NHS Highly Speci- 3. Neeve VC, Samuels DC, Bindoff LA, et al. What is influ- alised Service for Rare Mitochondrial Disorders of Adults and Children. encing the phenotype of the common homozygous poly- merase- mutation p.Ala467Thr? Brain 2012;135:3614– Study funding: No target funding reported. 3626. Disclosure: Y. S. Ng holds an NIHR Clinical Lectureship and 4. Anagnostou ME, Ng YS, Taylor RW, McFarland R. Epi- received funding from the MRC Centre for Neuromuscular Diseases for his doctoral study. H. Powell reports no disclosures. N. Hoggard lepsy due to mutations in the mitochondrial polymerase has served on the editorial board of British Journal of Radiology gamma (POLG) gene: a clinical and molecular genetic and has received research support from the Medical Research Council. review. Epilepsia 2016;57:1531–1545. D. M. Turnbull is supported by the Wellcome Centre for Mitochon- 5. Cheldi A, Ronchi D, Bordoni A, et al. POLG1 mutations drial Research (203105/Z/16/Z) (newcastle-mitochondria.com/), the and stroke like episodes: a distinct clinical entity rather MRC Centre for Translational Research in Neuromuscular Disease than an atypical MELAS syndrome. BMC Neurol 2013; Mitochondrial Disease Patient Cohort (UK) (G0800674), the Lily 13:8. Foundation and the UK NIHR Biomedical Research Centre for 6. Tzoulis C, Neckelmann G, Mørk SJ, et al. Localized cere- Ageing, and Age-related disease award to the Newcastle upon Tyne bral energy failure in DNA polymerase gamma-associated Foundation Hospitals NHS Trust. R. W. Taylor is supported by the encephalopathy syndromes. Brain 2010;133:1428–1437. Wellcome Centre for Mitochondrial Research (203105/Z/16/Z) (newcastle-mitochondria.com/), the MRC Centre for Translational 7. Lax NZ, Grady J, Laude A, et al. Extensive respiratory Research in Neuromuscular Disease Mitochondrial Disease Patient chain defects in inhibitory interneurones in patients with Cohort (UK) (G0800674), the Lily Foundation, the UK NIHR mitochondrial disease. Neuropathol Appl Neurobiol 2016; Biomedical Research Centre for Aging and Age-related disease award 42:180–193.

Neurology: Genetics 3 Clinical/Scientific Notes

Aqeela Al-Hashim, MD NOVEL INTRONIC MUTATION IN MTM1 percentile, diffuse extremity weakness, and long fin- Hernan D. Gonorazky, MD DETECTED BY RNA ANALYSIS IN A CASE OF X- gers, a combination of clinical features highly sugges- Kimberly Amburgey, LINKED MYOTUBULAR MYOPATHY tive of a mutation in MTM1.1 MSc, CGC Diagnostic muscle biopsy performed at age 11 Soma Das, PhD X-linked myotubular myopathy (XLMTM) is a rare months showed the classic features of an XLMTM James J. Dowling, MD, neuromuscular condition that presents with neonatal (figure, A); thus, he was presumptively diagnosed with PhD hypotonia and weakness and is associated with severe the condition. However, genetic testing of MTM1, – morbidities (including wheelchair, feeding tube, and accomplished on a next-generation sequencing based Neurol Genet 1 panel of 17 genes implicated in congenital myopathy 2017;3:e182; doi: 10.1212/ ventilator dependence) and early death. It is defined NXG.0000000000000182 by muscle biopsy features, including central nuclei, (including all exons plus 10 base pairs of surrounding abnormal oxidative stain distribution, and type I fiber intronic sequence as performed at the University of hypotrophy.2 Mutations in myotubularin (MTM1) Chicago), did not reveal a mutation, and multiplex account for all genetically solved cases of XLMTM, ligation-dependent probe amplification (MPLA) testing but have not been discovered in all individuals with for deletion/duplication of these genes was also normal. characteristic clinical and biopsy features.3 Of note, Given the high suspicion of XLMTM, we investi- there are some forms of autosomal centronuclear gated MTM1 transcript and protein from the pro- ’ myopathy that can resemble XLMTM, such as those band s skin fibroblasts. Reverse transcription PCR associated with mutations in BIN1, DNM2, RYR1, (RT-PCR) analysis using sets of overlapping exonic and SPEG, although rarely are such cases a complete primers identified abnormalities in the MTM1 tran- phenocopy of XLMTM.4 In this study, we present script including primarily deletion of exon 5 (figure, a case that illustrates the importance of considering B and C). Western blot analysis showed reduction of noncoding mutations as a cause of XLMTM and full-length MTM1 protein and the presence of illustrate the utility of RNA analysis in individuals a smaller molecular weight fragment (figure, D). with a phenotype suggestive of a particular genetic Genomic DNA Sanger sequencing of intron 4 re- diagnosis. vealed a previously unreported 4 base pair deletion The proband is a 27-year-old man with severe (c.232-26_232-23delGACT), 23 base pairs from the generalized weakness and prominent disease morbid- splice acceptor junction (figure, E). In silico analysis ities, including wheelchair, feeding tube, and ventila- predicted that this mutation interrupts the splice tor dependence. The family history was unremarkable branch point, the consequence of which is skipping for neuromuscular disorders. He was born at term of exon 5 (figure, F and G). In all, these data confirm from a healthy mother after a pregnancy complicated this as a case of XLMTM due to a novel intronic by diminished fetal movements. He was noted at mutation that was detected by RNA analysis in vitro. birth to be weak and hypotonic, although he did This case is instructive for several reasons. First, it not require invasive respiratory support as a neonate. adds to the growing list of noncoding mutations iden- He had multiple episodes of respiratory failure that tified as causes of rare mendelian disorders. Given resulted in repeated hospitalizations, and at age 6.5 the estimate that approximately 50% of all neuroge- years underwent tracheostomy and thereafter netic diseases are not solved by exome sequencing required continuous ventilatory support. He achieved or gene panels, cases such as ours support a hypothesis the ability to walk independently at age 30 months that many of these unsolved cases are the result of in- and lost the ability at 48 months after prolonged hos- tronic or regulatory sequence mutations.5 Second, the pitalization for a respiratory infection. Since that case illustrates the utility of RNA analysis as a means time, he has been supported by wheelchair. He is able of identifying such mutations. In settings like this one to eat independently, although he has difficulties with of a high index of suspicion for a specific genetic chewing and swallowing. Physical examination is cause, RT-PCR is an easy modality for interrogating notable for myopathic facies, ptosis, ophthalmopare- transcripts. However, when mutations in many genes sis, head circumference greater than the 90th may be responsible for a given phenotype, large-scale

Neurology.org/ng Copyright © 2017 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Figure Novel intronic mutation in MTM1 as a cause of X-linked myotubular myopathy

(A) Hematoxylin & eosin and succinate dehydrogenase (SDH) staining of the patient’s muscle biopsy showing the classic features of centronuclear myopathy, including myofiber hypotrophy, central nuclei, and aggregation of oxidative material; (B) RT-PCR of amplicon 1 (F1, spanning exons 2–8) on RNA from patient fibroblasts revealing an abnormal transcript of 567 bp (exon 5 skipping) and 472 bp (exons 4 and 5 skipping) compared with control 678 bp; (C) Western blot of myotubularin (MTM1) protein from patient (Pt) and control (Ctrl) fibroblasts (65 kDa normal band) revealing that MTM1 is almost not detectable in the patient (actin for loading control); (D) DNA chromatogram of the hemizygous c. 232-26_232-23 deletion in intron 4 of MTM1 gene; (E) Schematic diagram of the MTM1 mutation and its position within a branch point site. (F) Schematic depicting the consequences of the branch on RNA processing of MTM1 (i.e., skipping of exon 5 or of exons 4 and 5) as determined by RT-PCR (B). (G) Validation of exon 5 skipping as revealed in the chromatogram from Sanger sequencing of the middle band (*) identified by RT-PCR. Sequence analysis confirms the loss of exon 5 sequence. cDNA 5 complementary DNA; RT-PCR 5 reverse transcription PCR.

2 Neurology: Genetics transcriptome analysis using RNA sequencing is an boards of the RYR1 Foundation, the MTM/CNM patient registry, effective methodology.5,6 Last, the case identifies an and the World Muscle Society; has received donations to SickKids Foundation, Where There’s a Will Foundation, and Joshua Frase unexpected shortcoming of gene panels and exome Foundation; has received travel funding from Audentes; has served on sequencing; unlike Sanger sequencing, to enable cost- the editorial boards of Muscle & Nerve, PLoS Currents: Muscular effective screening of many exons, these technologies Dystrophy, Journal of Neuromuscular Diseases, Neuromuscular use short primers that are typically within 10 base Disorders, and Disease Models & Mechanisms; has been a consul- tant for Guidepoint Global Advisors and GLG Group; and has pairs of the exon, and thus miss out on noncoding received research funding from Valerion Therapeutics, CIHR, mutations outside the immediate splice region. This NIH, NSERC, Genome Canada, Hospital for Sick Children, Cure is particularly relevant for splice lariat mutations (as CMD, Joshua Frase Foundation, Team Joseph, Myotubular Trust, – and Muscular Dystrophy Association. Go to Neurology.org/ng for full in our situation), as the branch point is typically 20 disclosure forms. The Article Processing Charge was funded by 40 base pairs from the intron boundary. Of note, the Genome Canada. milder phenotype of our patient may be explained by This is an open access article distributed under the terms of the our observation of a faint amount of full-length Creative Commons Attribution-NonCommercial-NoDerivatives Li- MTM1 transcript and protein, and/or by the fact that cense 4.0 (CC BY-NC-ND), which permits downloading and shar- ing the work provided it is properly cited. The work cannot be his MTM1 mutation produces exon 5 skipping, the changed in any way or used commercially without permission from consequence of which is an in-frame deletion and the journal. predicted production of a truncated protein. Received May 11, 2017. Accepted in final form July 11, 2017. From the Division of Neurology (A.A.-H., H.G., K.A., J.J.D.) and Correspondence to Dr. Dowling: [email protected] Program for Genetics and Genome Biology (J.J.D.), Hospital for Sick Children; Department of Paediatrics (J.J.D.) and Deparmtent of 1. Das S, Dowling J, Pierson CR. X-linked centronuclear Molecular Genetics (J.J.D.), University of Toronto, ON, Canada; myopathy. In: Pagon RA, Adam MP, Ardinger HH, et al, Division of Neurology (A.A.-H.), King Fahad Medical City, Riyadh, editors. GeneReviews(R). Seattle: University of Washing- Saudi Arabia; and Department of Human Genetics (S.D.), Univer- ton; 1993. sity of Chicago, IL. 2. Lawlor MW, Beggs AH, Buj-Bello A, et al. Skeletal muscle Author contributions: Dr. Al-Hashim: study concept and design, data pathology in X-linked myotubular myopathy: review with acquisition, and writing of manuscript. Dr. Gonorazky and Ms. cross-species comparisons. J Neuropathol Exp Neurol 2016; Amburgey: data acquisition and interpretation and critical revision 75:102–110. of manuscript. Dr. Das: data interpretation and critical revision of 3. Jungbluth H, Wallgren-Pettersson C, Laporte J. Centronu- manuscript for intellectual content. Dr. Dowling: study concept and clear (myotubular) myopathy. Orphanet J Rare Dis 2008;3: design, study supervision, acquisition and interpretation of data, and writing and critical revision of manuscript for intellectual content. 26. 4. North KN, Wang CH, Clarke N, et al. Approach to the Acknowledgment: The authors thank Etsuko Tsuchiya for assistance for REB and Valerion Therapeutics for support of longitudinal nat- diagnosis of congenital myopathies. Neuromuscul Disord – ural history study. 2014;24:97 116. Study funding: Study funded by the Department of Paediatrics Start- 5. Cummings BB, Marshall JL, Tukiainen T, et al. Improving up Fund at the Hospital for Sick Children. Study participant seen as genetic diagnosis in Mendelian disease with transcriptome part of an ongoing longitudinal natural history study funded by sequencing. Sci Transl Med 2017:9. Valerion Therapeutics. 6. Gonorazky H, Liang M, Cummings B, et al. RNAseq anal- Disclosure: A. Al-Hashim, H. Gonorazky, K. Amburgey, and S. Das ysis for the diagnosis of muscular dystrophy. Ann Clin report no disclosures. J.J. Dowling serves on the scientific advisory Transl Neurol 2015;3:55–60.

Neurology: Genetics 3 Clinical/Scientific Notes

David S. Lynch, MRCPI NONSYNDROMIC PARKINSON DISEASE IN indicating anosmia. Her sister, P3, scored 33/40 Samantha H.Y. Loh, PhD A FAMILY WITH AUTOSOMAL DOMINANT OPTIC in the UPSIT, which is within the normal range. OPA1 Jasmine Harley, BSc ATROPHY DUE TO MUTATIONS A validated assessment of bradykinesia and incoor- Alastair J. Noyce, PhD dination, the BRAIN test, was performed by both L. Miguel Martins, PhD Many genes implicated in familial Parkinson disease P2 and P3. P2 scored in the low normal range for Nicholas W. Wood, PhD, (PD) code for proteins with mitochondrial function.1 speed with a prolonged dwell time. She was also FMedSci Several of these genes, including PINK1 and PARK2, inaccurate and had poor rhythm. P3 scored in the Henry Houlden, PhD are involved in mitophagy, a mitochondrial quality normal range for speed, accuracy, and dwell time. Helene Plun-Favreau, control pathway.2 We describe a family with 3 mem- These findings support the clinical observation that PhD bers affected by autosomal dominant optic atrophy in P3 does not fulfill motor criteria for PD, nor does which 2 affected members also developed PD. she have objective olfactory or motor dysfunction Neurol Genet that might suggest prodromal disease. She is, how- 2017;3:e188; doi: 10.1212/ While the role of mitophagy-related genes in PD is NXG.0000000000000188 well established, this report provides further evidence ever, 10 years younger than her affected sister. of PD risk conferred through abnormal mitochon- All 3 patients were found to carry the same hetero- drial fusion and cristae morphology. zygous deletion of exons 28 and 29 of OPA1.We excluded point mutations in autosomal dominant Clinical description. Patient 1 (P1) developed PD genes via focused exome sequencing in patient visual impairment in childhood. She was severely P2 (Illumina TruSight One). Specifically, no rare affected by optic atrophy but otherwise well until variants were detected in SNCA, LRRK2, GBA, age 70 years. She then developed depression, fol- PINK1, PARK2, or HTRA2. An unrelated patient lowed by asymmetric upper limb tremor. This (OC) carrying a heterozygous frameshift mutation was accompanied by postural instability and a bra- at the beginning of exon 28 (c.2708_2711 delTTAG) dykinetic gait. She was diagnosed with PD and has was used as a control. He was not affected with and a moderate response to levodopa. Patient 2 (P2) had no family history of PD. developed visual symptoms at age 9 years and was Methods. All 5 members of the family and patient diagnosed with optic atrophy. Her vision deterio- OC underwent skin biopsy and dermal fibroblast rated in her 20s, but she remained well until age culture. OPA1 protein levels were determined by 40 years when she developed gait difficulty and Western blotting from whole cell lysates using the upper limb rest tremor. She was diagnosed with anti-OPA1 antibody (BD Biosciences), which rec- PD and has a moderate response to both levodopa ognizes both long and short isoforms of OPA1. and dopamine agonists. During examination at age Anti-b-actin was used as a loading control (Abcam, 53 years, she had Hoehn and Yahr Stage 2 PD. Cambridge, United Kingdom). Transmission Additional features, such as dementia, neuropathy, electron microscopy (TEM) was used to investigate or deafness, were not present in P1 or P2. The mitochondrial cristae morphology. Cells were youngest sibling, patient 3 (P3), was diagnosed fixed and imaged, and mitochondrial morphology with optic atrophy at age 14 years, and while visu- was scored in a masked way in which a score of ally impaired, remained mildly affected and can 3 indicated normal cristae morphology, a score drive and read without magnification at age 40 of 2 indicated internal disorganization, and a score years. She does not fulfill clinical diagnostic criteria of 1 indicated mitochondria with severely damaged for PD at this time. See figure, A for pedigree. cristae.4,5 Mitochondrial networks were assessed in BothP2andP3wereclinicallyassessedusing live cells after staining the mitochondria with tet- a test battery used in the PREDICT-PD study, ramethylrhodamine and imaging at 403 under a population-based cohort in which participants an Opera Phenix Microscope (Perkin Elmer, are risk stratified for future PD using a variety of Waltham, MA). determinants.3 P2, who already carried a diagnosis of PD, scored 16/40 in the University of Pennsyl- Results. All affected patients showed an approxi- vania Smell Identification Test (UPSIT), mate 50% reduction in OPA1 levels by Western

Neurology.org/ng Copyright © 2017 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology 1 Figure Pedigree, immunoblot mitochondrial morphology score, and confocal microscopy images of mitochondrial networks showing representative cells

(A) Pedigree, in which filled boxes represent OPA1 mutation carriers. (B) Immunoblot demonstrating reduced OPA1 protein levels in all affected patients. (C) Mitochondrial morphology score demonstrating a significant increase in the number of abnormal mitochondrial cristae found in patients P1, P2, and P3. Examples of each score are shown using white arrows. (D) Confocal microscopy images of mitochondrial networks showing repre- sentative cells. OPA1 patient cells showed an increased number of fragmented mitochondrial networks compared with controls. PD 5 Parkinson disease.

blotting (figure, B), in keeping with haploinsuffi- implicates OPA1 in nonsyndromic, idiopathic PD ciency. Examination of mitochondrial morphology associated with abnormal cristae morphology and by TEM revealed a significant increase in abnormal mitochondrial networks. cristae in P1, P2, and P3, compared with their unaffected family members (figure, C). This phe- From the Department of Molecular Neuroscience (D.S.L., J.H., N.W.W., H.H., H.P.-F.), UCL Institute of Neurology, London, UK; notype was not shared by the unaffected OPA1 MRC Toxicology Unit (S.H.Y.L., L.M.M.), Leicester, UK; Reta Lila patient OC, despite all affected patients showing Weston Institute of Neurological Studies (A.J.N.), UCL Institute of similar disruption of the mitochondrial network Neurology, London, UK; and Neurogenetics Laboratory (H.H.), National Hospital for Neurology and Neurosurgery, London, UK. in live cell imaging (figure, D). Author contributions: David S. Lynch: study concept and design, Discussion. Mutations in a number of nuclear- acquisition of data, and writing the manuscript. Samantha H.Y. encoded mitochondrial genes are known to cause Loh: acquisition of data and critical revision of the manuscript. Jasmine Harley and Alastair J. Noyce: acquisition of data. L. Miguel PD, some of which lead to dysfunction of mito- Martins: interpretation of data and critical revision of the manu- chondrial quality control (mitophagy). In 1 pre- script. Nicholas W. Wood, Henry Houlden, and Helene Plun- vious report, a form of syndromic PD with Favreau: critical revision of the manuscript. dementia, neuropathy, and deafness was associated Acknowledgment: The authors thank the patients and their family for participating in this study. with OPA1 mutations.6 Dysregulated mitophagy Study funding: Funding for this work was provided by the Leonard and mitochondrial networks have also been de- Wolfson Experimental Neurology Centre, the Medical Research tected in OPA1 patient cells.7 This report further Council, and the Wellcome Trust.

2 Neurology: Genetics Disclosure: D.S. Lynch, S.H.Y. Loh, and J. Harley report no disclo- 1. Klein C, Westenberger A. Genetics of Parkinson’s sures. A.J. Noyce has received speaker honoraria from Britannia disease. Cold Spring Harb Perspect Med 2012;2: Pharmaceuticals, Office Octopus, and Global Kinetics Corporation a008888. and has received research support from Elan Pharmaceuticals, GE 2. Deas E, Wood NW, Plun-Favreau H. Mitophagy and Par- Healthcare, Parkinson’s UK, and LifeLab Online. L.M. Martins kinson’s disease: the PINK1-parkin link. Biochim Biophys reports no disclosures. N.W. Wood has served on the scientific advi- Acta 2011;1813:623–633. sory board of Luxembourg Health; has served on the editorial boards ’ of Neurogenetics, the Journal of Neurodegenerative Disease, and 3. Noyce AJ, R Bibo L, Peress L, et al. PREDICT-PD: Progress in Neurology and Psychiatry; and has received research an online approach to prospectively identify risk indica- ’ support from the Medical Research Council (MRC) UK and Well- tors of Parkinson s disease. Mov Disord 2017;32: come Trust Parkinson’s UK. H. Houlden has received research sup- 219–226. port from the Medical Research Council (MRC) UK, the BRT, the 4. Tufi R, Gandhi S, de Castro IP, et al. Enhancing nucleotide MDA USA, Muscular Dystrophy UK, Ataxia UK, Muscular Dys- metabolism protects against mitochondrial dysfunction and trophy UK, Rosetrees Trust, The Wellcome Trust, and the National neurodegeneration in a PINK1 model of Parkinson’s dis- Institute for Health (NIHR) UCL/UCLH BRC. H. Plun-Favreau ease. Nat Cell Biol 2014;16:1–12. reports no disclosures. Go to Neurology.org/ng for full disclosure forms. 5. Lehmann S, Loh SHY, Martins LM. Enhancing NAD 1 The Article Processing Charge was funded by the University College salvage metabolism is neuroprotective in a PINK1 model of London. Parkinson’s disease. Biol Open 2017;6:141–147. This is an open access article distributed under the terms of the 6. Carelli V, Musumeci O, Caporali L, et al. Syndromic par- Creative Commons Attribution License 4.0 (CC BY), which permits kinsonism and dementia associated with OPA1 missense unrestricted use, distribution, and reproduction in any medium, pro- – vided the original work is properly cited. mutations. Ann Neurol 2015;78:21 38. 7. Liao C, Ashley N, Diot A, et al. Dysregulated mitophagy Received May 2, 2017. Accepted in final form July 5, 2017. and mitochondrial organization in optic atrophy due to Correspondence to Dr. Lynch: [email protected] OPA1 mutations. Neurology 2017;88:131–142.

Neurology: Genetics 3 Clinical/Scientific Notes

Rubina Dad, MPhil HYPERVENTILATION-ATHETOSIS IN ASXL3 Since age 10, the patient has been exhibiting fre- Susan Walker, PhD DEFICIENCY (BAINBRIDGE-ROPERS) SYNDROME quent daily episodes of deep and constant hyperven- Stephen W. Scherer, PhD tilation. His neurodevelopmental presentation places Muhammad Jawad The protein product of the Drosophila additional sex him at the severe end of the autism spectrum. The Hassan, PhD combs-like (Asx) gene was shown to be a regulator, hyperventilation escalates with any nervousness, and Suk Yun Kang, MD, PhD both a suppressor and an activator, of Hox develop- as it builds, he develops athetotic movements of both Berge A. Minassian, MD mental genes. Mammals, including humans, possess 3 upper extremities, especially of the hands (videos 1 Asx orthologs: 2 expressed ubiquitously, while the and 2 at Neurology.org/ng). The young man also Neurol Genet hyperventilates in bed prior to falling asleep again 2017;3:e189; doi: 10.1212/ third, ASXL3, is predominantly expressed in the NXG.0000000000000189 brain. All 3 are involved in transcriptional regulation developing the athetoid movements (video 3). Both of many genes through direct actions or epigenetically hyperventilation and movements cease with sleep. via histone modifications. Specific genes regulated by Written informed consent was obtained from the ASXL3 have not been identified.1,2 family for participation in the study and showing the The clinical phenotype associated with heterozy- video recordings. Whole-exome sequencing was per- ’ gous loss of ASXL3 function was first described in formed on the patient s DNA using the Agilent Sure- 2013 (Bainbridge-Ropers syndrome; OMIM Select Human All Exon V5 target enrichment kit 615485) and subsequently expanded through a total followed by sequencing on Illumina HiSeq 2500. Bases of 27 patients to date. Key clinical features are as fol- were called using bcl2fastq v2.17 and reads mapped to lows: intellectual disability with profound speech the hg19 reference sequence using the Burrows-Wheeler impairment, severe early feeding difficulty, autistic be- Aligner backtrack algorithm (v0.5.9). A previously unre- haviors, failure to thrive, severe muscular hypotonia, ported heterozygous substitution-deletion mutation in and a characteristic long face with arched eyebrows, the ASXL3 gene was identified (NM_030632; exon11: downslanting palpebral fissures, and poor expressiv- c.1314_1316delinsA; p. S439Rfs*7), resulting in ity.3–6 We describe a new case with a striking pheno- frameshift and predicted premature truncation. Sanger type, namely hyperventilation-induced athetosis. sequencing confirmed the mutation in the patient (fig- The patient is the 16-year-old son of unrelated ure, B) and its absence in his parents. Korean parents. His birth weight was 3.2 kg. In the Specific loci for which transcription is regulated by neonatal period, he had episodes of apnea, for which the ASXL3 are not known, nor which of these target EEG did not support an epileptic cause. He subse- genes underlie the resultant neurodevelopmental out- quently exhibited severe delays in all aspects of devel- come. One study in patient fibroblasts identified 564 opment. He walked at 9 years. He is short of stature misregulated genes (approximately half up and the and has microcephaly (head circumference ,third rest downregulated), most of which had known func- percentile), hypertelorism, and a hypoplastic face tions in development and proliferation, or were them- (figure, A). He has severe intellectual disability and selves transcriptional regulators, suggesting that cannot understand or express any language. He has ASXL3 may function in upstream modulation of neu- no eye contact, has never focused on any object, and rodevelopmental regulator genes.2 ASXL3 is clearly is considered blind. Brainstem auditory evoked po- critical to brain development, and profound develop- tentials obtained recently were normal. He does not mental disturbance manifested in patients resulting have feeding or swallowing difficulties. Current exam- from the loss of 1 of the 2 copies of the gene suggests Supplemental data at ination reveals scoliosis and postural instability. He haploinsufficiency as the likely mechanism. Neurology.org/ng has an ataxic wide-based and staggering gait, which Athetosis with hyperventilation is seen in the pres- has been gradually deteriorating. He is not presently ent case; however, only hyperventilation was previously hypotonic, and motor strength and sensory functions reported in 2 cases.4,5 Hyperventilation in the context appear to be normal, as are his deep tendon and of severe neurodevelopmental disturbance is seen in plantar reflexes. Brain MRI and several EEGs have several neurogenetic disorders, including Rett, Joubert, been normal. Pitt-Hopkins, and Pallister-Killian syndromes. Recent

Neurology.org/ng Copyright © 2017 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology 1 Figure Facial appearance and the mutation of the patient with Bainbridge-Ropers syndrome

(A) Patient photograph: microcephaly (head circumference ,third percentile), hypertelorism, and facial hypoplasia. (B) Electropherograms from exon 11 of the ASXL3 gene showing de novo heterozygous deletion of 2 base pairs (shaded area) and a substitution (arrow) of a third (C.A) resulting in frameshift/ premature termination (ASXL3 NM_030632: c.1314_1316delinsA; p.S439Rfs*7).

work in Rett syndrome mice revealed that expression (HEC-IRSIP). This work was funded by the Ontario Brain Institute of Mecp2 broadly in the medulla, rather than in any and Genome Canada. Berge A. Minassian holds the University of Toronto Michael Bahen Chair in Epilepsy Research and the Univer- particular medullary nucleus or pathway, is required sity of Texas Southwestern Jimmy Elizabeth Westcott Distinguished for normal breathing.7 Similar future work in autopsy Chair in Pediatric Neurology. and animal models of the other above diseases, and in Study funding: No targeted funding reported. Bainbridge-Ropers disease, will gradually uncover the Disclosure: R. Dad has received research support from the Higher pathways, systems, and functions involved in the neu- Education Commission of Pakistan. S. Walker reports no disclosures. S.W. Scherer has served on the editorial boards of Genomic Med- rodevelopmental symptoms in affected patients and aid icine, Genes, Genomes, Genetics, the Journal of Personalized in the understanding of normal brain development and Medicine, the Open Genomics Journal, the Hugo Journal, function. Our particular case suggests a neural connec- Genome Medicine, the Journal of Neurodevelopmental Disor- tion, in the context of ASXL3 deficiency, between ders, Autism Research, PathoGenetics, Comparative and Func- tional Genomics, BioMed Central Medical Genomics, and pathways of respiration and of motor control. Cytogenetics and Genome Research; and has received research From the Atta-ur Rahman School of Applied Biosciences (R.D., M.J.H.), support from the Genome Canada/Ontario Genomics Institute, the National University of Sciences and Technology (NUST), Canadian Institutes of Health Research, the Canadian Institute for Islamabad, Pakistan; Program in Genetics and Genome Biology Advanced Research, the McLaughlin Centre, the Canada Founda- (R.D.) and The Centre for Applied Genomics, Genetics and Genome tion for Innovation, the government of Ontario, the NIH, Autism Biology (S.W., S.W.S.), The Hospital for Sick Children, Department Speaks, and the SickKids Foundation. M.J. Hassan reports no dis- of Molecular Genetics (S.W.S.), and McLaughlin Centre (S.W.S.), closures. S.Y. Kang has served on the editorial board of the Journal of University of Toronto, Ontario, Canada; Department of Neurolgy Clinical Neurology. B.A. Minassian holds patents for diagnostic (S.Y.K.), Dongtan Sacred Heart Hospital, Hallym University Col- testing of the following genes: EPM2A, EPM2B, MECP2, and lege of Medicine, Hwaseong, Gyeonggi-do, Republic of Korea; Pro- VMA21; has received research support from the NINDS of the gram in Genetics and Genome Biology (B.A.M.), Division of NIH; and receives license fee payments/royalty payments from Neurology, Department of Pediatrics, The Hospital for Sick Chil- patents for diagnostic testing of the following genes: EPM2A, dren, University of Toronto, Ontario, Canada; and Departments EPM2B, MECP2, and VMA21. Go to Neurology.org/ng for full of Pediatrics, Neurology, and Neurotherapeutics (B.A.M.), Univer- disclosure forms. The Article Processing Charge was funded by the sity of Texas Southwestern, Dallas. authors. Author contributions: study concept and design by Berge A. Minassian This is an open access article distributed under the terms of the and Suk Yun Kang. Acquisition of clinical data by Suk Yun Kang. Creative Commons Attribution-NonCommercial-NoDerivatives Li- Acquisition of genetic data by Rubina Dad, Susan Walker, and cense 4.0 (CC BY-NC-ND), which permits downloading and shar- Stephen W. Scherer. Analysis and interpretation of data by ing the work provided it is properly cited. The work cannot be Rubina Dad and Susan Walker. Study supervision by Berge A. changed in any way or used commercially without permission from Minassian and Muhammad Jawad Hassan. Critical revision of the the journal. manuscript for intellectual content by Berge A. Minassian and Suk Received April 13, 2017. Accepted in final form July 25, 2017. Yun Kang. Correspondence to Dr. Minassian: [email protected] Acknowledgement: The authors are thankful to the family members for participation in the study. Rubina Dad gratefully acknowledges 1. Micol JB, Abdel-Wahab O. The role of additional sex her funding by the Higher Education Commission of Pakistan combs-like proteins in Cancer. Cold Spring Harb Perspect under the International Research Support Initiative Program Med 2016;6:a026526.

2 Neurology: Genetics 2. Srivastava A, Ritesh KC, Tsan YC, et al. De novo dominant 5. Balasubramanian M, WilloughbyJ,FryAE,etal.Delineating ASXL3 mutations alter H2A deubiquitination and tran- the phenotypic spectrum of Bainbridge-Ropers syndrome: 12 scription in Bainbridge-Ropers syndrome. Hum Mol Genet new patients with de novo, heterozygous, loss-of-function mu- 2016;25:597–608. tations in ASXL3 and review of published literature. J Med 3. Bainbridge MN, Hu H, Muzny DM, et al. De novo trun- Genet 2017;54:537–543. cating mutations in ASXL3 are associated with a novel clin- 6. Dinwiddie DL, Soden SE, Saunders CJ, et al. De novo ical phenotype with similarities to Bohring-Opitz frameshift mutation in ASXL3 in a patient with global syndrome. Genome Med 2013;5:11. developmental delay, microcephaly, and craniofacial anom- 4. Kuechler A, Czeschik JC, Graf E, et al. Bainbridge-Ropers alies. BMC Med Genomics 2013;17:32. syndrome caused by loss-of-function variants in ASXL3: 7. Huang TW, Kochukov MY, Ward CS, et al. Progressive changes a recognizable condition. Eur J Hum Genet 2017;25: in a distributed neural circuit underlie breathing abnormalities in 183–191. mice lacking MeCP2. J Neurosci 2016;36:5572–5586.

Neurology: Genetics 3 Clinical/Scientific Notes

Sali M.K. Farhan, PhD ARHGEF28 P.LYS280METFS40TER IN AN do not have any features suggestive of ALS. There is Tania F. Gendron, PhD AMYOTROPHIC LATERAL SCLEROSIS FAMILY currently no individual affected with ALS in the sub- Leonard Petrucelli, PhD WITH A C9ORF72 EXPANSION sequent generation; however, all are younger than the Robert A. Hegele, MD typical age at onset (table e-2). Michael J. Strong, MD We previously discovered a frameshift variant in Genetic variants in patients with ALS. We genotyped ARHGEF28, encoding rho guanine nucleotide family members for ARHGEF28 p.Lys280Metf- Neurol Genet 2017;3:e190; doi: 10.1212/ exchange factor (RGNEF), in a patient with amyo- s40Ter and observed an additional 8 heterozygous NXG.0000000000000190 trophic lateral sclerosis (ALS) following our observa- carriers (figures e-1 and e-2). Of note, we also identi- tion of RGNEF cytoplasmic inclusions in the fied ARHGEF28 p.Lys280Metfs40Ter in 2 unre- patient’s motor neurons.1 This variant was subse- lated spouses (IV-4 and IV-7), despite the rarity of quently observed in a Chinese cohort of patients with the ARHGEF28 variant based on its absence from ALS and in ALS-discordant identical twins.2,3 Herein, public databases such as the Exome Aggregation we describe the expanded analyses performed on the Consortium, Genome Aggregation Database, family of our index patient, who carries both ARH- Human Gene Mutation Database, and ClinVar, and GEF28 p.Lys280Metfs40Ter and a C9orf72 repeat from the Amyotrophic Lateral Sclerosis Online expansion. We performed neuromuscular examina- Genetics Database. tion and screened for cognitive impairment using DNA of participants was also tested for the C9orf72 the Montreal Cognitive Assessment on 24 family repeat expansion as previously described (table e-1).5 members. We also screened for the C9orf72 repeat We identified 7 additional C9orf72 expansion carriers, expansion and ARHGEF28 p.Lys280Metfs40Ter. In and these results were confirmed by the Clinical Lab- total, we observed 8 C9orf72 expansion carriers and 9 oratory Improvement Amendments–certified labora- ARHGEF28 p.Lys280Metfs40Ter carriers, with 4 tories. Of the identified carriers, 2 died with a diagnosis double heterozygotes carrying both variants. We also of ALS. It is unknown whether the other 6 carriers, used ONDRISeq, a next-generation sequencing who showed no neurologic abnormalities at the time of panel of neurodegeneration-related genes, to identify evaluation, will eventually develop symptoms of dis- additional variations, with no variants detected. We ease. We also used ONDRISeq6 to identify additional postulate that ARHGEF28 p.Lys280Metfs40Ter variation in known ALS genes in III-2 and III-5, with may modify the incremental neurodegeneration risk no further variants identified. present in individuals with a C9orf72 expansion. C9orf72 dipeptide repeat protein immunostaining. Cere- bellar sections from III-2 and III-5 were obtained Results. Clinical characteristics. III-2 presented with from archived neuropathologic specimens and im- left upper extremity weakness at age 60 and later munostained for dipeptide repeat proteins [poly(GP), developed features of bulbar dysfunction, diffuse limb poly(GA), and poly(GR)] produced from C9orf72 wasting and weakness, and pathologically brisk re- G4C2 expansions through repeat associated non-ATG flexes (table e-1 at Neurology.org/ng). Given the translation. In each instance, dipeptide repeat pro- history of ALS in the family, the diagnosis was defi- teins were observed (figure 1). nite familial ALS using the El Escorial criteria.4 The postmortem neuropathology confirmed the diagnosis Discussion. We expand the analyses in the ALS fam- in the absence of frontotemporal lobar degeneration. ily described in our previous report.1 We previously Her brothers, III-3 and III-5, were also diagnosed identified individuals with ALS who have RGNEF Supplemental data at with ALS but presented at a much younger age (table pathology, which are cytoplasmic inclusions of Neurology.org/ng e-1). III-3 presented with bilateral upper extremity RGNEF (encoded by ARHGEF28)inmotorneurons, weakness as well as limb wasting. Similar to III-2, as well as a C9orf72 expansion. Of interest, we the first symptomology observed in III-5 was weak- observed individuals with ALS who have RGNEF ness in the left upper extremity. He later developed pathology but without a C9orf72 expansion, which bulbar dysfunction, limb wasting, and pyramidal suggests that RGNEF pathology may be sufficient to weakness. The 2 surviving siblings, III-7 and III-9, cause ALS or there may be other as yet uncharacterized

Neurology.org/ng Copyright © 2017 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology 1 Figure 1 Cerebellar dipeptide repeat protein pathology

Characteristic neuropathologic lesions immunopositive for poly(GP), poly(GA), or poly(GR) proteins in Purkinje cells (PC), the granule cell layer (GL), or the molecular layer (ML) of the cerebellum. Scale bar 10 mm.

pathogenic factors involved. Although ARHGEF28 Department of Clinical Neurological Sciences (M.J.S.), Robarts p.Lys280Metfs40Ter is present in currently unaffected Research Institute, Schulich School of Medicine and Dentistry, Western University, London, Ontario, Canada. individuals, which may suggest that the variant is Author contributions: Sali M.K. Farhan: study concept and design, benign, it is more likely that the variant modifies dis- acquisition of data, analysis and interpretation of data, and writing ease risk, as we have previously observed RGNEF neu- and critical revision of the manuscript for important intellectual ronal cytoplasmic inclusions in spinal cord motor content. Tania F. Gendron and Leonard Petrucelli: acquisition of 1 7 data, analysis and interpretation of data, and critical revision of the neurons of the index case (III-5) and in other cases. manuscript for important intellectual content. Robert A. Hegele: In addition, the variant may also reflect common study concept and design, analysis and interpretation of data, critical ancestral origins as both unrelated individuals reported revision of the manuscript for important intellectual content, and that they are from the same Northern Netherlands study supervision. Michael J. Strong: study concept and design, anal- ysis and interpretation of data, critical revision of the manuscript for region as other members of the family. Based on the important intellectual content, study supervision, and lead location of the mutation (amino acid 280 of 1731) investigator. leading to the eventual termination of the RGNEF Acknowledgment: The authors are grateful to the participating fam- polypeptide, it is likely the variant affects RGNEF ilies in the study and to Mrs. Ann Rowe for assistance in sample acquisition. They also thank Dr. Zhang for experimental guidance function. Whether the disease mechanism is haploin- and Ms. Castanedes-Casey and Dr. Dickson for assistance with the sufficiency or cellular toxicity is unclear. immunohistochemistry assays. While it is presently unknown whether carriers Study funding: This research was supported by the Canadian Insti- will eventually develop symptoms of disease, we pos- tutes of Health Research (CIHR), ALS Canada, and the McFeat tulate that double heterozygotes of both the C9orf72 Family Fund. S.M.K. Farhan was supported by the CIHR Fredrick Banting and Charles Best Canada Graduate Scholarship. expansion and ARHGEF28 p.Lys280Metfs40Ter Disclosure: S.M.K. Farhan reports no disclosures. T.F. Gendron has may be at a greater risk of developing ALS earlier than received speaker honoraria from Johns Hopkins and the NIH; holds individuals who only carry a C9orf72 expansion. a patent for Methods and materials for detecting poly(GP) proteins in Given the limited sample size, precludes us from tissues from C9ORF72 repeat expansion carriers; has received research support from the NIH, the ALS Association, and the Mus- definitively determining the effect of the ARHGEF28 cular Dystrophy Association; and receives license fee payments for variant; we will investigate these variants in vitro to C9ORF72 repeat expansion constructs and viruses and antibodies evaluate their dual effect on motor neurons. In addi- against C9ORF72 dipeptide repeat proteins. L.P. Petrucelli has tion, we plan to sequence a larger cohort of ALS cases served on the scientific advisory boards of Denali and Biogen; has served on the editorial boards of the Journal of Neuroscience, to determine the frequency of ARHGEF28 variation. Molecular Neurodegeneration, and PLoS ONE; holds a patent for “C9orf72 antibody as a Diagnostic Test for C9orf72 Mutation From the Department of Biochemistry (S.M.K.F., R.A.H.), Robarts Carriers. Development of an assay to detect c9orf72 RAN translation Research Institute, Schulich School of Medicine and Dentistry, products in CSF/blood”; has received research support from Western University, London, Ontario, Canada; Department of Lundbeck, Biogen, the NIH/NINDS, the Mayo Clinic Foundation, Neuroscience (T.F.G., L.P.), Mayo Clinic, Jacksonville, FL; and the Amyotrophic Lateral Sclerosis Association, Robert Packard Center

2 Neurology: Genetics for ALS Research at Johns Hopkins, Target ALS, and the Association 3. Zhang M, Xi Z, Ghani M, et al. Genetic and epigenetic for Frontotemporal Degeneration; receives license fee payments from study of ALS-discordant identical twins with double muta- Lundbeck, Biogen, and Denali; and receives royalty payments for tions in SOD1 and ARHGEF28. J Neurol Neurosurg Psy- Licensing of TDP-43 antibody. R.A. Hegele reports no disclosures. chiatry 2016;87:1268–1270. M.J. Strong serves on the editorial board of Amyotrophic Lateral 4. Brooks BR. El Escorial World Federation of Neurology Sclerosis and receives research support from the Canadian Institutes of Health Research. Go to Neurology.org/ng for full disclosure forms. criteria for the diagnosis of amyotrophic lateral sclerosis. The Article Processing Charge was funded by the authors. Subcommittee on Motor Neuron Diseases/Amyotrophic This is an open access article distributed under the terms of the Lateral Sclerosis of the World Federation of Neurology Creative Commons Attribution-NonCommercial-NoDerivatives Li- Research Group on Neuromuscular Diseases and the El “ ” cense 4.0 (CC BY-NC-ND), which permits downloading and shar- Escorial Clinical limits of amyotrophic lateral sclerosis ing the work provided it is properly cited. The work cannot be workshop contributors. J Neurol Sci 1994;124(suppl): changed in any way or used commercially without permission from 96–107. the journal. 5. DeJesus-Hernandez M, Mackenzie IR, Boeve BF, et al. Received April 27, 2017. Accepted in final form August 7, 2017. Expanded GGGGCC hexanucleotide repeat in noncoding Correspondence to Dr. Strong: [email protected] region of C9ORF72 causes chromosome 9p-linked FTD and ALS. Neuron 2011;72:245–256. 1. Droppelmann CA, Wang J, Campos-Melo D, et al. Detec- 6. Farhan SM, Dilliott AA, Ghani M, et al. The ONDRISeq tion of a novel frameshift mutation and regions with homo- panel: custom-designed next-generation sequencing of genes zygosis within ARHGEF28 gene in familial amyotrophic related to neurodegeneration. NPJ Genom Med 2016. doi: lateral sclerosis. Amyotroph Lateral Scler Frontotemporal 10.1038/npjgenmed.2016.32. Degener 2013;14:444–451. 7. Keller BA, Volkening K, Droppelmann CA, Ang LC, 2. Ma Y, Tang L, Chen L, et al. ARHGEF28 gene exon Rademakers R, Strong MJ. Co-aggregation of RNA binding 6/intron 6 junction mutations in Chinese amyotrophic lat- proteins in ALS spinal motor neurons: evidence of a com- eral sclerosis cohort. Amyotroph Lateral Scler Frontotempo- mon pathogenic mechanism. Acta Neuropathol 2012;124: ral Degener 2014;15:309–311. 733–747.

Neurology: Genetics 3 Clinical/Scientific Notes

Emmanuel Wiener, MD CERVICAL ARTERY DISSECTION AND ILIAC was not possible to date the dissection. No intracra- Peter Martin, MD ARTERY ANEURYSM IN AN SMAD-4 MUTATION nial abnormality, aneurysm, or arteriovenous malfor- Sarju Mehta, MD CARRIER mation was found. Hugh Stephen Markus, Owing to his multifocal arteriopathy, genetic screen- MD, FRCP ing was performed using an arteriopathy gene panel. A heterozygous, loss-of-function, nonsense mutation was Neurol Genet Case report. Cervical artery dissection is an impor- found in exon 10 of the SMAD-4 gene (c.1336C.T; 2017;3:e191; doi: 10.1212/ tant cause of stroke, especially in younger individuals. NXG.0000000000000191 protein change p.Gln446Ter). This Class 4, likely path- Most cases appear to be sporadic, but rare monogenic ogenic mutation, is predicted to create a premature ter- conditions can cause dissection, such as vascular mination codon, resulting in a truncated protein. This 1 Ehlers-Danlos syndrome. The wider use of sequenc- mutation has not been identified within in-house data- ing techniques in clinical diagnosis of genetic disorders bases or in individuals from the Exome Aggregation has made it more realistic to test for multiple genes in Consortium project, although a similar c.1333C.T potential monogenic stroke conditions. Here, we mutation (p.Arg445Ter) has been reported in a patient report the case of a cervical artery dissection in with JPS, aortic root dilation, and mitral valvulopathy.2 a SMAD-4 mutation carrier. Segregation studies have been advised to confirm path- This patient is a 48-year-old man with a history of ogenicity via the Clinical Genetics department. hypertension, aortic root dilation that led to surgical replacement at age 41, and a subsequent right com- Discussion. This case illustrates that cervical artery mon iliac artery aneurysm. He had experienced sev- abnormal tortuosity and dissection can occur as part eral episodes of mild epistaxis, but none had needed of a systemic arteriopathy due to SMAD-4 mutation. specific treatment. There was no personal history of SMAD-4 protein is involved in the transforming gastrointestinal symptoms or bleeding. growth factor beta signaling pathway. The chromo- His family history was suggestive of juvenile poly- somal location of the SMAD-4 gene is 18q21.1. Sev- posis syndrome (JPS). Several colon polyps were diag- eral SMAD-4 loss-of-function mutations have been nosed in his mother; his brother had experienced described among patients with JPS and hereditary several episodes of serious digestive bleeding related hemorrhagic telangiectasia (HHT), 2 genetically het- to multiple colon polyps. The patient has 2 sons and erogeneous autosomal dominant disorders. 2 daughters. At age 4, his eldest daughter has had an JPS, occurring in approximately 1/100,000 indi- episode of rectal bleeding that led to the diagnosis of viduals, is characterized by early-onset colorectal colon polyps; his youngest son has experienced an epi- and gastric polyp formation, with a subsequent risk sode of epistaxis. His youngest daughter and eldest son of digestive bleeding, anemia, and malignancy. are doing well. SMAD-4 and BMPR1A mutations are the most fre- Physical examination showed slightly translucent quently found. The largest JPS series to date has chest skin and a possible small telangiectasia in his found SMAD-4 mutations in 21% of cases.3 right cheek. There were no other skin abnormalities; HHT, affecting approximately 1/5,000–10,000 in- joint examination was unremarkable; heart and neck dividuals, is a multifocal vascular dysplasia, with skin auscultation was normal. and mucosal telangiectases and visceral arteriovenous Because of the aortic and iliac arteriopathy, cere- malformations (lung, brain, and liver). Recurrent bral CT angiogram was performed to screen for cere- epistaxis is the most frequent manifestation. The most bral artery abnormalities. The angiogram revealed 3 frequently involved genes are ACVRL1, ENG, and findings (figure): firstly, abnormal increased tortuos- SMAD-4. Patient series suggest that SMAD-4 muta- ity of both internal carotids; secondly, mild arterial tions account for 2% of cases.4 JPS and HHT can be diameter irregularities; and thirdly, a short but dis- associated in the same individual, usually in case of tinct dissection flap in the V2 segment of the right a SMAD-4 mutation. vertebral artery, at the level of C5. The patient had no Cardiovascular abnormalities have been described history of stroke, neck pain, or trauma; therefore, it in SMAD-4 mutation carriers. Thoracic aortopathy

Neurology.org/ng Copyright © 2017 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Figure Cervical artery abnormalities

(A–C) Abnormal tortuosity of carotid arteries. (D) Mild diameter irregularities. (E) Right vertebral artery dissection flap.

(aortic root dilation) has been reported in 9%–37% This is an open access article distributed under the terms of the of patients5–7; heart valve regurgitation in 6%,5 and Creative Commons Attribution-NonCommercial-NoDerivatives Li- 5 cense 4.0 (CC BY-NC-ND), which permits downloading and shar- intracranial aneurysm in 3%. However, involvement ing the work provided it is properly cited. The work cannot be of the cervical and iliac arteries does not seem to be changed in any way or used commercially without permission from a typical feature of this condition. the journal. This report extends the number of monogenic Received June 26, 2017. Accepted in final form July 25, 2017. causes of cervical artery dissection and highlights that SMAD-4 mutations can cause systemic multifocal ar- Correspondence to Dr. Wiener: [email protected] teriopathy, involving not only the aortic root but also the cervical and iliac arteries. Vascular imaging of the 1. Debette S, Markus HS. The genetics of cervical artery dissec- tion: a systematic review. Stroke 2009;40:459–466. arterial tree is advised in at-risk individuals. 2. Andrabi S, Bekheirnia MR, Robbins-Furman P, Lewis RA, From the Department of Stroke Medicine (E.W., P.M., H.S.M.) Prior TW, Potocki L. SMAD4 mutation segregating in a family and Department of Genetics (S.M.), Addenbrooke’s Hospital, with juvenile polyposis, aortopathy, and mitral valve dysfunc- Cambridge, United Kingdom. tion. Am J Med Genet A 2011;155A:1165–1169. Author contributions: Dr. Martin collected clinical data during clin- 3. Calva-Cerqueira D, Chinnathambi S, Pechman B, Bair J, ical consultations with the patient. Dr. Wiener reviewed the litera- Larsen-Haidle J, Howe JR. The rate of germline mutations ture and wrote the case. Pr. Markus reviewed the article. Dr. Mehta and large deletions of SMAD4 and BMPR1A in juvenile conducted the genetic analysis of the patient. polyposis. Clin Genet 2009;75:79–85. Study funding: No targeted funding reported. 4. Prigoda NL, Savas S, Abdalla SA, et al. Hereditary haemor- Disclosure: E. Wiener and P. Martin report no disclosures. S. Mehta rhagic telangiectasia: mutation detection, test sensitivity and receives publishing royalties from Hodder Arnold. H.S. Markus has novel mutations. J Med Genet 2006;43:722–728. served on the editorial boards of the International Journal of Stroke, 5. Wain KE, Ellingson MS, McDonald J, et al. Appreciating Clinical Neurology and Neurosurgery, BMC Medicine, and the broad clinical features of SMAD4 mutation carriers: Frontiers in Neurology; receives publishing royalties from Oxford a multicenter chart review. Genet Med 2014;16:588–593. University Press; has been a consultant for Astra Zeneca; has received 6. Heald B, Rigelsky C, Moran R, et al. Prevalence of thoracic reimbursement from Astra Zeneca for teaching session to employees; aortopathy in patients with juvenile Polyposis Syndrome- and has received research support from Medical Research Council Experimental Medicine Grant, National Institute for Health Hereditary Hemorrhagic Telangiectasia due to SMAD4. – Research, Stroke Association, EU, Wellcome Trust, The British Heart Am J Med Genet A 2015;167A:1758 1762. Foundation, and Alzheimer Research UK. Go to Neurology.org/ng 7. Jelsig AM, Tørring PM, Kjeldsen AD, et al. JP-HHT phe- for full disclosure forms. The Article Processing Charge was funded by notype in Danish patients with SMAD4 mutations. Clin the authors. Genet 2016;90:55–62.

2 Neurology: Genetics Clinical/Scientific Notes

Matthew J. Barrett, MD, AUTOSOMAL RECESSIVE INHERITANCE OF evaluations, trio exome sequencing was pursued to MSc ADCY5-RELATED GENERALIZED DYSTONIA AND establish a definitive diagnosis (Ambry Genetics Eli S. Williams, PhD MYOCLONUS Corp., Aliso Viejo, CA). Testing revealed the pres- Chelsea Chambers, MS, ence of a maternally inherited frameshift variant * CGC Monoallelic ADCY5 mutations have been associated (c.409_428del20; p.G137Cfs 184) and a paternally . Radhika Dhamija, MD with a benign hereditary chorea-like phenotype and inherited missense variant (c.3037C T; p.R1013C) a mixed hyperkinetic syndrome with dystonia, cho- in ADCY5 (NM_183357). These alleles were con- Neurol Genet 1–4 firmed in the proband’s brother using Sanger 2017;3:e193; doi: 10.1212/ rea, and myoclonus. We report 2 siblings with NXG.0000000000000193 generalized dystonia and myoclonus associated with sequencing. biallelic pathogenic variants in ADCY5 inherited in Discussion. ADCY5 mutations are believed to be an autosomal recessive pattern. inherited in an autosomal dominant fashion based on observations from the original pedigree: there were Case report. A 27-year-old woman presented with affected individuals in all 5 generations, complete a movement disorder. She was born full term. She penetrance in males and females, and evidence of walked at 18 months, and speech development was male-to-male transmission.5 Subsequent reported delayed. Involuntary movements were first noted at 3 families confirmed autosomal dominant inheri- years with walking and increased in amplitude over tance.3,4 In addition, mutations described in ADCY5- time. Dystonia was first diagnosed at age 3. She had related dyskinesia are recurrent, with the p.R418W frequent falls due to involuntary myoclonic move- and the p.A726T mutation described in multiple ments in the setting of lower extremity dystonia. unrelated individuals.1–3,6 Subsequent functional Movements lessened with sleep and did not improve analysis characterized the effect of these mutations as with alcohol or caffeine. Trihexyphenidyl, levodopa, gain of function: there was increased adenylyl cyclase clonazepam, and amantadine failed to provide bene- activity in cells containing mutant ADCY5 compared fit. On examination, she was found to have general- with wild type.6 However, in a previously reported ized dystonia with superimposed myoclonus. See family affected with autosomal dominant chorea and video at Neurology.org/ng. Chromosomal microarray dystonia, a c.208811G.A splice site mutation seg- was negative. Alpha-fetoprotein and vitamin E levels regated with disease.4 Functional studies showed no were normal. Her 24-year-old brother was born at expression of the variant allele and reduced expression term and walked prior to 1 year. His speech devel- of ADCY5 messenger RNA. It appears that both gain opment was delayed. After age 8, he developed of function and haploinsufficiency may result in cramping in his hands and neck. His gait worsened at ADCY5-related dyskinesia. age 15 secondary to worsening lower extremity dys- Here, we report 2 siblings with a phenotype of tonia following a growth spurt. Levodopa, clonaze- generalized dystonia and myoclonus associated with pam, gabapentin, propranolol, primidone, and biallelic pathogenic variants in ADCY5. A phenotype topiramate did not provide benefit. His neurologic of autosomal dominant myoclonus dystonia has been examination revealed generalized dystonia with su- reported with ADCY5 mutations, and dystonia and perimposed myoclonus. See video. Creatine kinase, myoclonus have been reported in other published lactic acid, comprehensive metabolic panel, cerulo- cases.7 Our cases differ from previously published plasmin, and 24-hour urine copper levels were nor- cases because they inherited 2 pathogenic variants Supplemental data at mal. No pathogenic variants or large deletions/ from unaffected parents, a paternally inherited mis- Neurology.org/ng duplications were found in SGCE. Brain MRI was sense variant and a maternally inherited frameshift unremarkable. There were no other siblings. Neither variant. There is evidence supporting a role for both parent nor other relatives had a history of a movement these variants in the clinical manifestation of disease. disorder. The paternal p.R1013C variant occurs at an evolu- Because of familial juvenile-onset generalized tionarily conserved residue, is predicted to be damag- dystonia with myoclonus and previous negative ing by PolyPhen and deleterious by scale-invariant

Neurology.org/ng Copyright © 2017 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 feature transform, and is not present in over 60,000 the Neurotoxin Exposure Treatment Parkinson’s Research Program. individuals in the Exome Aggregation Consortium E.S. Williams has served on the editorial board of Current Genetic Medicine Reports. C. Chambers and R. Dhamija report no disclo- browser. These data together with cosegregation with sures. Go to Neurology.org/ng for full disclosure forms. The Article the movement disorder phenotype suggest that Processing Charge was funded by the authors. p.R1013C is not a normal population variant and This is an open access article distributed under the terms of the has clinical significance. We interpret the maternally Creative Commons Attribution-NonCommercial-NoDerivatives Li- cense 4.0 (CC BY-NC-ND), which permits downloading and shar- inherited 20 base-pair deletion as a loss of function ing the work provided it is properly cited. The work cannot be allele because it results in an early stop codon after changed in any way or used commercially without permission from 184 amino acids. the journal. In this family, both variants appear to be Received February 22, 2017. Accepted in final form August 15, 2017. required for a movement disorder phenotype. The missense variant may result in a change in Correspondence to Dr. Barrett: [email protected] the function of adenylate cyclase that does not result in a phenotype in the presence of a wild- 1. Chen DH, Meneret A, Friedman JR, et al. ADCY5-related type allele. However, in the presence of a null dyskinesia: broader spectrum and genotype-phenotype cor- relations. Neurology 2015;85:2026–2035. allele, the mutant protein is insufficient to main- 2. Mencacci NE, Erro R, Wiethoff S, et al. ADCY5 mutations tain adequate cyclic adenosine monophosphate lev- are another cause of benign hereditary chorea. Neurology els resulting in dystonia and myoclonus. Autosomal 2015;85:80–88. recessive inheritance has also been reported for 3. Chang FC, Westenberger A, Dale RC, et al. Phenotypic other genes typically associated with autosomal insights into ADCY5-associated disease. Mov Disord – dominantly inherited dystonia, specifically THAP1 2016;31:1033 1040. 4. Carapito R, Paul N, Untrau M, et al. A de novo ADCY5 and GNAL.8,9 This family expands the understand- mutation causes early-onset autosomal dominant chorea ing of ADCY5-related dyskinesia by showing that and dystonia. Mov Disord 2015;30:423–427. certain pathogenic variants inherited in an autoso- 5. Chen YZ, Matsushita MM, Robertson P, et al. Autosomal mal recessive pattern may produce a phenotype dominant familial dyskinesia and facial myokymia: single identical to autosomal dominant mutations. exome sequencing identifies a mutation in adenylyl cyclase 5. Arch Neurol 2012;69:630–635. From the Department of Neurology (M.J.B., C.C.), and Department 6. Chen YZ, Friedman JR, Chen DH, et al. Gain-of-function of Pathology (E.S.W.), University of Virginia Health System, Char- ADCY5 mutations in familial dyskinesia with facial myo- lottesville; and Department of Clinical Genomics and Neurology kymia. Ann Neurol 2014;75:542–549. (R.D.), Mayo Clinic, Phoenix, AZ. 7. Douglas AG, Andreoletti G, Talbot K, et al. ADCY5- Author contributions: Dr. Barrett: drafting the original manuscript related dyskinesia presenting as familial myoclonus-dysto- and analysis and interpretation of data. Dr. Williams, Ms. Cham- – bers, and Dr. Dhamija: analysis and interpretation of data and nia. Neurogenetics 2017;18:111 117. revising the manuscript for intellectual content. 8. Schneider SA, Ramirez A, Shafiee K, et al. Homozygous Study funding: No targeted funding reported. THAP1 mutations as cause of early-onset generalized dys- tonia. Mov Disord 2011;26:858–861. Disclosure: M.J. Barrett has received research support from Axovant Sciences Inc., Azevan Pharmaceuticals, and Merck Sharp & Dohme 9. Masuho I, Fang M, Geng C, et al. Homozygous GNAL Corp. and receives support from the Commonwealth of Virginia’s mutation associated with familial childhood-onset general- Alzheimer’s and Related Diseases Research Award Fund and the ized dystonia. Neurol Genet 2016;2:e78. doi: 10.1212/ Office of the Assistant Secretary of Defense for Health Affairs through NXG.0000000000000078.

2 Neurology: Genetics Clinical/Scientific Notes

Gary W. Beecham, PhD THE ALZHEIMER’S DISEASE SEQUENCING Families meeting initial criteria were assigned a priority J.C. Bis, PhD PROJECT: STUDY DESIGN AND SAMPLE rank based on number and age at onset of affected in- E.R. Martin, PhD SELECTION dividuals, number of generations affected, and presence S.-H. Choi, PhD of APOE e4 alleles. Priority was given to families heavily $ A. L. DeStefano, PhD Late-onset Alzheimer disease (LOAD) is the leading loaded for AD ( 4 affected members with DNA avail- e C.M. van Duijn, PhD cause of dementia worldwide, with substantial economic able) with minimal APOE 4 alleles. Cases met National – ’ M. Fornage, PhD and public health implications.1 LOAD is a neurodegen- Institute of Neurological Diseases Alzheimer s S.B. Gabriel, PhD erative disease characterized by progressive dementia NINCDS-ADRDA (National Institute of Neurological D.C. Koboldt, PhD typically manifesting in the seventh to ninth decades. and Communicative Disorders and Stroke, and the ’ D.E. Larson, PhD Neuropathological changes precede clinical symptoms Alzheimer s Disease and related Disorders Association; ’ A.C. Naj, PhD by 10–20 years, resulting in clinically asymptomatic now, Alzheimer s Association) criteria for possible, B.M. Psaty, PhD individuals carrying neuropathologic features of probable, or definite AD. Controls were free of clinical W. Salerno, PhD LOAD.2 Much of the heritability of LOAD remains AD on cognitive assessment. A detailed description of W.S. Bush, PhD unexplained, despite LOAD having a high heritability the family design is in Appendix 1 at Neurology.org/ng. T.M. Foroud, PhD (60%–80%) and despite the identification of the APOE In total, we selected 582 individuals (498 affected E. Wijsman, PhD locus, a major genetic determinant for LOAD.3 Genetic and 84 unaffected) from 111 families for WGS to L.A. Farrer, PhD analyses have identified more than 25 other variants identify genomic regions associated with increased A. Goate, PhD associated with smaller individual effects on disease risk.4 risk of LOAD. Selected individuals include 229 Euro- J.L. Haines, PhD To identify novel genetic variation influencing AD pean ancestry and 353 Caribbean Hispanic (CH) in- Margaret A. Pericak- risk and protection, the Alzheimer’s Disease Sequenc- dividuals (table). The European ancestry families Vance, PhD ing Project (ADSP) was implemented as a collabora- included 2 large Dutch families from the Erasmus 5 E. Boerwinkle, PhD tive effort of the National Institutes on Aging, the Rucphen Family study. Most of these families were R. Mayeux, MD National Human Genome Research Institute, and recently analyzed for genetic linkage, an analysis that 6,7 S. Seshadri, MD the Alzheimer disease research community. Individ- will be used in the analysis of the sequence data. By e e G. Schellenberg, PhD ual contributors include the Alzheimer’s Disease design, no 4/ 4 individuals were selected for e e Genetics Consortium, the Neurology Phenotype sequencing, and we prioritized 3/ 4 individuals with Neurol Genet earlier disease onset. Twenty-seven percent of families 2017;3:e194; doi: 10.1212/ Working Group of the Cohorts for Heart and Aging NXG.0000000000000194 Research in Genomic Epidemiology consortium, had at least 1 case with autopsy confirmation. and the Large Scale Sequencing and Analysis Centers at Baylor University, the Broad Institute, ADSP case-control design. Over 30,000 samples and Washington University. were considered for inclusion in the case-control Study design and sample selection were conducted design. All cases met NINCDS-ADRDA criteria for to address issues of phenotypic heterogeneity and max- possible, probable, or definite AD, had documented imize statistical power. The study design includes 2 pri- age at onset or age at death (for pathologically verified mary phases: a whole-genome sequencing (WGS) cases), and APOE genotyping. All controls were at family-based study and a whole-exome sequencing least 60 years old and were free of dementia by direct, (WES) case-control study. The WGS study was de- documented cognitive assessment. Three primary signed to target rarer variation through allelic segrega- case-control selection strategies were evaluated, and tion and linkage analyses in multiplex AD families. ultimately, a design was chosen that targeted cases The WES case-control study was designed to target with minimal risk as predicted by known risk factors Supplemental data at (age, sex, and APOE) and targeted controls with the Neurology.org/ng low-frequency coding variation in genes that contrib- ute to AD risk or protection. least probability of conversion to AD by age 85 years. The details and rationale of the case-control selection ADSP family study design. Approximately 1,400 process and the evaluation of alternate study designs multiplex LOAD families were reviewed for inclusion. are described in detail in Appendix 2. Families were required to have multiple members with In total, we selected 5,096 cases and 4,965 con- LOAD, genomic DNA, and available APOE genotypes. trols under the chosen design (table). We selected

Neurology.org/ng Copyright © 2017 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Department of Biostatistics and Epidemiology (A.C.N.) and Perel- Table Sample demographics for family and case-control studies man School of Medicine (G.S.), University of Pennsylvania, Phila- delphia; Group Health Research Institute (B.M.P.), Group Health Family study Case-control study Cooperative, Seattle, WA; Human Genome Sequencing Center (W.S., E.B.), Baylor College of Medicine, Houston, TX; Department Affected Unaffected Case Control of Epidemiology and Biostatistics (W.S.B., J.L.H.), Case Western N 498 84 5,778 5,136 Reserve University, Cleveland, OH; Department of Medical and Molecular Genetics (T.M.F.), Indiana University School of Medi- Age at onset/examination (SD) 73.7 (9.4) 68.0 (11.0) 76.0 (9.2) 86.1 (5.2) cine, Indianapolis; Department of Medicine (Biomedical Genetics) Male (%) 38.8 38.1 43.1 40.7 (L.A.F.), Department of Ophthalmology (L.A.F.), and Department of Epidemiology (L.A.F.), Boston University School of Medicine and Hispanic/Latino (%) 60.6 59.5 3.7 3.4 Public Health, MA; Department of Neuroscience (A.G.), Icahn Autopsy confirmation (%) 15.6 0.0 32.8 7.0 School of Medicine at Mount Sinai, New York, NY; Human Genet- ics Center (E.B.), UT Health School of Public Health, Houston, TX; APOE genotype Taub Institute for Research on Alzheimer’s Disease and the Aging e3e3 268 (54%) 60 (71%) 2,915 (50%) 3,394 (66%) Brain (R.M.) and Gertrude H. Sergievsky Center (R.M.), Columbia University Medical Center, New York, NY; Department of Neurol- e3e4 196 (39%) 9 (11%) 2,198 (38%) 679 (13%) ogy (R.M.), Columbia University Medical Center and New York e4e4 0 (0%) 0 (0%) 161 (3%) 17 (,1%) Presbyterian Hospital, NY; and Department of Epidemiology (R.M.), Mailman School of Public Health, Columbia University, e2e2 1(,1%) 1 (1%) 23 (,1%) 48 (1%) New York, NY. e2e3 29 (6%) 11 (13%) 359 (6%) 925 (18%) Author contributions: All authors contributed to the work presented e2e4 4 (1%) 2 (2%) 122 (2%) 73 (1%) in this article. Drafting: the primary manuscript was prepared by G.W.B., with significant contributions from S.S., E.B., G.S., M.A.P.-V., and J.C.B. All authors participated in the revision and editing of the manuscript. Concept and design: primary study concept and design was by G.W.B., with significant contributions from E.R.M., J.C.B., 682 additional unrelated cases from additional mul- M.A.P.-V., J.L.H., R.M., S.S., E.B., G.S., L.A.F., A.G., C.M.v.D., A.C.N., and A.D. Analysis and interpretation: review of family data was tiplex families that had a strong family history for performed by M.A.P.-V., R.M., E.B., S.S., C.M.v.D., and T.M.F. LOAD. Because some of these 682 cases arose from Primary statistical analyses were performed by G.W.B., with additions CH multiplex families, we included 171 cognitively from J.C.B., A.C.N., E.R.M., S.-H.C., A.D., and S.S. All authors normal CH control samples in the WES. participated in the interpretation and discussion of results. Acquisition of data: sample data were contributed by C.M.v.D., A.D., T.M.F., L.A.F., The sequencing of the nearly 600 whole genomes A.G., J.L.H., M.A.P.-V., E.B., R.M., S.S., and G.S. Statistical and 11,000 whole exomes has been completed; the analyses: statistical analyses were primarily conducted by G.W.B.; data sets are currently available to the research com- additional analyses conducted by J.C.B., A.C.N., E.R.M., S.-H.C., munity through qualified access (dbGaP study A.D., and S.S. (affiliations noted above, all academic). Study supervision and coordination: primary study supervision and coordination was by phs000572.v7.p4). This data set will be used to iden- G.S., R.M., E.B., M.A.P.-V., J.L.H., S.S., A.G., L.A.F., and E.W. tify genetic factors influencing AD risk and protection Acknowledgment: The Alzheimer’s Disease Sequencing Project and will be a critical resource for the LOAD research (ADSP) comprises 2 Alzheimer’s Disease (AD) genetics consortia community. and 3 National Human Genome Research Institute (NHGRI)- funded Large Scale Sequencing and Analysis Centers (LSAC). Standard protocol approvals, registrations, and The 2 AD genetics consortia are the Alzheimer’sDiseaseGenetics patient consents. This study has the approval of the Consortium (ADGC) funded by the NIA (U01 AG032984), and institutional review boards of participating institu- the Cohorts for Heart and Aging Research in Genomic Epidemi- ology (CHARGE) funded by the NIA (R01 AG033193), the tions, and informed consent was obtained from all National Heart, Lung, and Blood Institute (NHLBI), other patients. NIH institutes, and other foreign governmental and nongovern- From the John P. Hussman Institute for Human Genomics (G.W.B., mental organizations. The Discovery Phase analysis of sequence E.R.M., M.A.P.-V.) and Dr. John T. Macdonald Foundation data is supported through UF1AG047133 (to G. Schellenberg, Department of Human Genetics (G.W.B., E.R.M., M.A.P.-V.), L.A. Farrer, M.A. Pericak-Vance, R. Mayeux, and J.L. Haines); Miller School of Medicine, University of Miami, FL; Cardiovascular U01AG049505 to S. Seshadri; U01AG049506 to E. Boerwin- Health Research Unit (J.C.B.), Department of Medicine, Cardiovas- kle; U01AG049507 to E. Wijsman; and U01AG049508 to cular Health Research Unit (B.M.P.), Departments of Medicine, A. Goate. Data generation and harmonization in the Follow- Epidemiology, Health Services, Department of Biostatistics (E.W.), up Phases is supported by U54AG052427 (to G. Schellenberg and Division of Medical Genetics (E.W.), Department of Medicine, and Wang). The ADGC cohorts include Adult Changes in ’ University of Washington, Seattle; Department of Biostatistics (S.-H.C., Thought (ACT), the Alzheimer sDiseaseCenters(ADC),the A.D., L.A.F.), Boston University School of Public Health, MA; Chicago Health and Aging Project (CHAP), the Memory and The Framingham Heart Study (A.D., S.S.), MA; Department of Aging Project (MAP), Mayo Clinic (MAYO), Mayo Parkinson’s Neurology (A.D., L.A.F., S.S.), Boston University School of Medi- Disease controls, the University of Miami, the Multi- cine, MA; Department of Epidemiology (C.M.v.D), Erasmus MC, Institutional Research in Alzheimer’s Genetic Epidemiology Rotterdam, Netherlands; Brown Foundation Institute of Molecular Study (MIRAGE), the National Cell Repository for Alzheimer’s Medicine (M.F.) and Human Genetics Center (M.F.), University Disease (NCRAD), the National Institute on Aging Late Onset of Texas Health Science Center, Houston; The Eli and Edythe L. Alzheimer’s Disease Family Study (NIA-LOAD), the Religious Broad Institute of Massachusetts Institute of Technology (S.B.G.), Orders Study (ROS), the Texas Alzheimer’s Research and Care Cambridge; Harvard University (S.B.G.), Cambridge, MA; The Consortium (TARC), Vanderbilt University/Case Western McDonnell Genome Institute (D.C.K., D.E.L.) and Department Reserve University (VAN/CWRU), the Washington Heights- of Genetics (D.E.L.), Washington University, St. Louis, MO; Inwood Columbia Aging Project (WHICAP) and the

2 Neurology: Genetics Washington University Sequencing Project (WUSP), the Colum- Acknowledgment section. W. Salerno has been a consultant for Laser- bia University Hispanic–Estudio Familiar de Influencia Ge- gen. W.S. Bush serves on the editorial boards of BMC BioData netica de Alzheimer (EFIGA), the University of Toronto (UT), Mining and PLoS One; and has received research support from and Genetic Differences (GD). The CHARGE cohorts with the NIA. T.M. Foroud has served on the scientific advisory boards funding provided by 5RC2HL102419 and HL105756, include of the National Advisory Council on Alcohol Abuse and Alcoholism, the following: the Atherosclerosis Risk in Communities (ARIC) the Washington University Alzheimer’s Disease Research Center, and Study which is conducted as a collaborative study supported by the NIA Genetics of Alzheimer’s Disease Data Storage Site; has NHLBI contracts (HHSN268201100005C, received travel funding from the Michael J. Fox Foundation for HHSN268201100006C, HHSN268201100007C, Parkinson’s Research, the NIH, the University of Pittsburgh, and HHSN268201100008C, HHSN268201100009C, the University of Chicago; has received travel funding and speaker HHSN268201100010C, HHSN268201100011C, and honoraria from the University of Texas at Austin; and has received HHSN268201100012C), the Austrian Stroke Prevention Study research support from the NIH, the US Department of Defense, (ASPS), the Cardiovascular Health Study (CHS), the Erasmus Columbia University, San Diego State University, the University Rucphen Family Study (ERF), the Framingham Heart Study (FHS), of California, San Diego, the University of Massachusetts, the Uni- and the Rotterdam Study (RS). The 3 LSACs are the Human versity of Pennsylvania, the State University of New York, and the ’ Genome Sequencing Center at the Baylor College of Medicine (U54 Michael J. Fox Foundation for Parkinson s Research. E. Wijsman has HG003273), the Broad Institute Genome Center served on the scientific advisory board of NIH NHLBI National (U54HG003067), and the Washington University Genome Insti- Registry of Genetically Triggered Thoracic Aortic Aneurysms and tute (U54HG003079). Biological samples and associated phenotypic Cardiovascular Conditions; has served on the editorial boards of data used in primary data analyses were stored at Study Investigators BMC Proceedings and Bioinformatics; and has received research support from the NIH and the Metropolitan Life Foundation Award institutions and at the National Cell Repository for Alzheimer’s for Medical Research. L.A. Farrer has served on the editorial boards Disease (NCRAD, U24AG021886) at Indiana University funded of the American Journal of Alzheimer’s Disease & Other Demen- by the NIA. Associated Phenotypic Data used in primary and sec- tias and Clinical Genetics; has 1 patent pending for the use of ondary data analyses were provided by Study Investigators, the NIA- PLXNA4 as a drug target and biomarker for Alzheimer disease; funded Alzheimer’s Disease Centers (ADCs), and the National Alz- has been a consultant for Novartis Pharmaceuticals, Gerson Lehr- heimer’s Coordinating Center (NACC, U01AG016976) and the man, Guidepoint Global, and Finnegan & Associates, LLP; and has National Institute on Aging Genetics of Alzheimer’s Disease Data received research support from the NIH, the Fidelity Foundation, Storage Site (NIAGADS, U24AG041689) at the University of and the Thome Memorial Foundation. A. Goate has served on the Pennsylvania, funded by the NIA and at the Database for Genotypes scientific advisory board of Denali Therapeutics; has received travel and Phenotypes (dbGaP) funded by the NIH. This research was funding from the Rainwater Foundation; has served on the editorial supported in part by the Intramural Research Program of the NIH board of eLife; holds patents for PSEN mutations in AD, Tau and the National Library of Medicine. Contributors to the Genetic mutations in FTD, and TDP43 mutations in ALS\FTD; has been Analysis Data included Study Investigators on projects that were a consultant for Cognition Therapeutics and AbbVie; has received individually funded by the NIA and other NIH institutes, and by research support from F-Prime, the NIA, the Rainwater Charitable private U.S. organizations, or foreign governmental or nongovern- Foundation, and the JPB Foundation; and receives royalty payments mental organizations. from Taconic Industries for tau mutation patent. J.L. Haines has Study funding: Supported by the NIH, primarily the NIA, NHLBI, served on the editorial boards of Neurogenetics, Current Protocols and NHGRI. Primary support includes the Alzheimer’s Disease in Human Genetics, and Human Molecular Genetics; receives Genetics Consortium (ADGC) funded by the NIA (U01 publishing royalties from John Wiley & Sons; and has received AG032984), and the Cohorts for Heart and Aging Research in research support from the NIH. M.A. Pericak-Vance serves on the Genomic Epidemiology (CHARGE) funded by the NIA (R01 editorial boards of Genetic Epidemiology, Molecular Autism, and AG033193), the Human Genome Sequencing Center at the Baylor Advances in Genomics and Gene Expression; her immediate fam- College of Medicine (U54 HG003273), the Broad Institute Genome ily member Dr. Jeffery Vance has served on the editorial boards of Center (U54HG003067), and the Washington University Genome Neurology Genetics and American Journal of Neurodegenerative Institute (U54HG003079). Additional funding of contributing sites Disease; and she has received research support from the NIH and the is noted in the Acknowledgment section. JJ Vance Foundation. E. Boerwinkle has received a speaker honorar- Disclosures: G.W. Beecham receives funding from the NIH and the ium from the American Society for Bone and Mineral Research; is Department of Defense. J.C. Bis reports no disclosures. E.R. Martin a Scientific Officer at Codified Genomics, LLC; and has received has served on the editorial board of Frontiers in Statistical Genetics research support from the NIH. R. Mayeux has received research and Methodology and holds US Patent No. 6697739 Test for support from the NIH. S. Seshadri serves on the editorial boards of Linkage and Association in General Pedigrees: The Pedigree Disequi- Journal of Alzheimer’s Disease, Stroke, and Neurology and has librium Test. S.-H. Choi reports no disclosures. A. DeStefano has received research support from the NIA. G. Schellenberg has served on received research support from the NIH. C.M. van Duijn and M. the scientific advisory boards of Alzheimer’s Association, the Society of Fornage report no disclosures. S.B. Gabriel is an employee of a non- Progressive Supranuclear Palsy, the Alzheimer Research Consortium, profit entity and has been a consultant for WilmerHale Guidepoint the Peebler PSP Research Foundation, the United Kingdom Parkin- Global. D.C. Koboldt receives a coinventor’s share of license revenue son Disease Center, University College London, the Alzheimer’s Dis- for VarScan (a software tool for next-generation sequencing analysis), ease Sequence Project, the Structural Variant Work Group, Mayo with licensing and disbursements handled by his former institution, Clinic, Rochester, Udall Center, the University of Miami, and the Washington University in St. Louis. In the past 2 years, paying Oxford Parkinson’s Disease Centre; has received travel funding/ licensees included Bina Technologies, Janssen, Fera Science, Philips speaker honoraria from the Alzheimer’s Disease Center, CurePSP, Electronics, and WuXi NextCODE. D.E. Larson has received the University of California, San Diego, Keystone Symposia, the research support from the NIH and St. Jude Children’s Research University of California, the Institute for Memory Impairment and Hospital. A.C. Naj has received speaker honoraria from Pfizer; has Neurological Disorders, Biomarkers in Neuropsychiatric Disorders served on the editorial board of PLoS One; and has received research (Toronto, Canada), the NIH, Novartis, the McKnight Brain Insti- support from the NIA, the BrightFocus Foundation, and Penn Insti- tute, the University of Florida, the NIA, the Keep Memory Alive tute on Aging. B.M. Psaty serves on the DSMB of a clinical trial for Center (Cleveland Clinic), the Lou Ruvo Center for Brain Health, a device funded by the manufacturer (Zoll Lifecor) and on the Steer- PSP/Lewy Body Disease Think-Tank, the American Association of ing Committee for the Yale Open Data Access project funded by Neuropathologists, the Fusion Conference, “What does the Johnson & Johnson; is a contributing writer for JAMA; and has future hold?” (Tucson, AZ), “Progressive supranuclear palsy received research support from an entity/entities listed in the genetics—update” (La Jolla, CA), the Center for Public Health

Neurology: Genetics 3 Genomics, Genome Sciences Seminar, the University of Virginia, 2. Bateman RJ, Xiong C, Benzinger TL, et al; Dominantly Neurology Grand Rounds, and Columbia University; has served Inherited Alzheimer Network. Clinical and biomarker on the editorial boards of the JournalofNeuralTransmission, changes in dominantly inherited Alzheimer’s disease. N ’ ’ Alzheimer s Research,theAmerican Journal of Alzheimer s Engl J Med 2012;367:795–804. Disease and other Dementias, Neurodegenerative Diseases, 3. Corder EH, Saunders AM, Strittmatter WJ, et al. Gene Current Alzheimer Research,andPathology and Laboratory dose of apolipoprotein E type 4 allele and the risk of Alz- Medicine International; is a professor at the University of Penn- ’ sylvania; and has received research support from the NIA/NIH, heimer s disease in late onset families. Science 1993;261: – CurePSP, and CBD Solutions. Go to Neurology.org/ng for full 921 923. disclosure forms. The Article Processing Charge was funded by the 4. Lambert JC, Ibrahim-Verbaas CA, Harold D, et al. Meta- authors. analysis of 74,046 individuals identifies 11 new suscepti- ’ This is an open access article distributed under the terms of the bility loci for Alzheimer s disease. Nat Genet 2013;45: Creative Commons Attribution-NonCommercial-NoDerivatives 1452–1458. License 4.0 (CC BY-NC-ND), which permits downloading and 5. Liu F, Arias-Vásquez A, Sleegers K, et al. A genomewide sharing the work provided it is properly cited. The work cannot be screen for late-onset Alzheimer disease in a genetically iso- changed in any way or used commercially without permission from lated Dutch population. Am J Hum Genet 2007;81:17–31. the journal. 6. Barral S, Cheng R, Reitz C, et al. Linkage analyses in Received May 1, 2017. Accepted in final form August 17, 2017. Caribbean Hispanic families identify novel loci associated ’ Correspondence to Dr. Beecham: [email protected] with familial late-onset Alzheimer s disease. Alzheimers De- ment 2015;11:1397–1406. 1. Prince M, Jackson J. Alzheimer’s Disease International, 7. Kunkle BW, Jaworski J, Barral S, et al. Genome-wide link- World Alzheimer Report 2009: The Global Prevalence of age analyses of non-Hispanic white families identify novel Dementia; 2009. Available at: alz.co.uk/research/files/Worl- loci for familial late-onset Alzheimer’s disease. Alzheimers dAlzheimerReport.pdf. Accessed October, 2016. Dement 2016;12:2–10.

4 Neurology: Genetics