Clinicogenetic and Functional Studies in Rare Hereditary

Clinicogenetic and functional studies in

rare hereditary neurodegenerative

movement disorders

Dr. Sarah Wiethoff

A thesis submitted to University College London for the degree of Doctor of Philosophy

Department of Molecular Neuroscience Institute of Neurology University College London (UCL) May 2016

I, Sarah Wiethoff, confirm that the work presented in this thesis is my own. Where information has been derived from other sources, I confirm that this has been indicated in the thesis. Collaborative work is also indicated in this thesis.

Signature:

Date:

Abstract

Neurodegenerative diseases are equally fascinating as they are devastating. They illustrate both function and pathology of neurons, the most complex cells in the human body.

In the past, technological progress has allowed the identification of genetic variation that can lead to neurodegenerative processes. However, for many patients with different neurodegenerative diseases to date, no genetic diagnosis is obtained despite thorough investigation. For another significant proportion of neurodegenerative diseases the genetic defect and the resulting clinical phenotype/spectrum is known, but exact pathomechanisms remain elusive. This delays successful translational research and eventual clinical treatment.

The objective of this thesis is to combine both aspects and employ two main techniques to further advance the search for better pathophysiologic understanding of neurodegenerative diseases: whole exome sequencing (WES) and induced pluripotent stem cell (iPSC) technology. Firstly, the thesis aims to improve clinical characterisation and genetic analysis of neurodegenerative patients to identify genetic causes and genetic modifiers of disease. Secondly, it aims to establish functional models in search of pathogenic and potentially druggable mechanisms using iPSCs in clinically and genetically characterised groups of patients.

This thesis thereby investigated the clinicogenetic and pathophysiological bases of rare hereditary movement disorders with focus on hereditary ataxias, trinucleotide repeat and complex parkinsonism disorders, including neurodegeneration with brain iron accumulation (NBIA). Clinical characterisation, WES, targeted sequencing, homozygosity mapping and traditional Sanger sequencing were used to yield clinical description, identify underlying genetic causes and to investigate disease modifiers in different trinucleotide repeat disorders.

To finally establish functional readouts of pathogenic mutations in a well characterised patient cohort, iPSC technology and directed differentiation was used to create a human neuronal model of spinocerebellar ataxia type 15 and to derive non-forebrain neuronal precursors to facilitate the study of cerebellar diseases using iPSCs in the future.

Table of contents

Abstract ...... 3

Table of contents ...... 4

Acknowledgements ...... 6

Abbreviations ...... 8

List of tables ...... 13

List of figures ...... 15

Publications/Book chapters ...... 18

1 Chapter 1: General introduction ...... 20 1.1 General introduction and aims of this thesis ...... 20 1.2 Genetics in neurodegenerative disease ...... 22 1.3 Clinicogenetics of neurodegeneration with brain iron accumulation ...... 35 1.4 Polyglutamine disorders and inherited cerebellar ataxias ...... 68 1.5 Induced pluripotent stem cell technology and its usefulness for cerebellar disease modelling ...... 74

2 Chapter 2: Methods ...... 91 2.1 Methods Chapter 3 ...... 91 2.2 Methods Chapter 4 ...... 105 2.3 Methods Chapter 5 ...... 109 2.4 Methods Chapter 6 ...... 115

3 Chapter 3: Dissecting neurodegeneration with brain iron accumulation clinically, genetically and cellularly ...... 127 3.1 Introduction ...... 127 3.2 Results ...... 129 3.3 Discussion ...... 152

4 Chapter 4: Genetics in families with rare neurodegenerative movement disorders: Examples from McLeod syndrome and hereditary ataxias ...... 168 4.1 McLeod syndrome: Identification of a novel single base-pair deletion ...... 168 4.2 SYNE1: Novel mutations and extension of the clinicogenetic spectrum ...... 175 4.3 PNPLA6-associated disorders: Novel homozygous variants in a large consanguineous Parsi kindred with pure cerebellar ataxia ...... 184

5 Chapter 5: Trinucleotide disorders: The conundrum of modifiers of age at onset, somatic and repeat instability and repeat interruptions ...... 193 5.1 DNA repair pathways underlie a common genetic mechanism modulating onset in polyglutamine diseases ...... 193 5.2 The role of repeat interruptions in SCA6 ...... 203

6 Chapter 6: Modelling cerebellar disease with human iPSCs ...... 211 6.1 Generating a human iPSC model to study spinocerebellar ataxia type 15 . 211 6.2 Generating non-cortical neuronal derivatives to improve the accuracy in modelling cerebellar disease ...... 241

7 Conclusions and future directions ...... 250

References ...... 258

8 APPENDIX ...... 282

Acknowledgements

I would like to thank my supervisors Professor Henry Houlden and Professor John Hardy for their excellent supervision and support, laboratory environment and mentorship on exciting projects. I would also like to express my thankfulness towards my subsidiary supervisors Dr. Conceicao Bettencourt and Dr. Rickie Patani for advising, guiding and supporting me throughout my different projects. Thanks to our patients and their families in providing their patience, time and material as well as to our numerous clinical collaborators who contribute with excellent clinical cases. A warm thanks to the members of the iPSC-dream-team rota who made my life in the second part of the PhD so much more wonderful. Special thanks to Prof. Andrew Singleton, Monica Federoff, Dr. Monia Ben-Hamad Hammer, Dr. Jinhui Ding, Steve Clipman, Dr. Raphael Gibbs and Dr. Dena Hernandez for their collaboration, guidance and support during my stay at the NIH and beyond as well as to Hallgeir Jonvick, Dr. Boniface Mok and Dr. Alan Pittman for helping with data transfer and analysis strategies at the IoN. Thanks to all members of the diagnostic lab, to Chris Steele, Dr. Lee Stanyer, June Smalley, Mark Gaskin, Deborah Hughes, Stephanie Efthymiou, Alexandra Zirra, Chris Lovejoy, Dr. Qiang Gang, Dr. Lucia Schottlaender, Dr. Niccolo Mencacci, Dr. Claudia Manzoni, Dr. Joshua Hersheson, Dr. Lasse Pihlstrom, Dr. Reema Paudel, Dr. Arianna Tucci, Dr. Alice Gardiner, Dr. Ellen Cottenie, Dr. David Lynch, Dr. Steven Lubbe, Claire Hall, Dr. Martina Hallegger, Dr. Chris Sibley, Dr. Zhi Yao, Dr. Sonia Gandhi, Dr. Rosella Abeti, Dr. Sarah Crisp, Dr. Monika Madej, Dr. Raffaele Ferrari, Dr. Davina Hensman, Dr. Enrico Bugiardini, Dr. Roman Praschberger, Dr. Mike Flower, Dr. Chelban, Andreea Manole and Dr. James Polke for their help and guidance at IoN as well as special thanks to Dr. Abi Li, Dr. Giulia Tyzack, Dr. Selina Wray, Professor Tom Warner, Dr. Rohan de Silva, Dr. Nuria Seto-Salvia, Dr. Charles Arber, Elisavet Preza, Luke Hill, Daniel Cotfas, Teresa Sposito, Dr. Rubika Balendra, Lee Darwent, Faiza Javad, Dr. Rina Bandopadhyay and Bimali Hapuarachchi for their support and advice at the Reta-Lila-Weston-Institute. Thanks to Prof. Bhatia and Prof. Wood for clinical teaching and cases. I would like to thank the following funding bodies for providing the financial support behind my projects and facilitating research exchange and travel to conferences: Brain Research Trust (BRT), Wellcome Trust, Medical Research Council (MRC), NIH, NBIA disorders association, NBIA Alliance, Hoffnungsbaum e.V. and Ministerium für Wissenschaft, Forschung und Kunst and Europäischen Sozialfonds, Baden-Württemberg within the Margarete von Wrangell programme. I want to thank my parents who taught me to wander this world with open eyes, and my 3 siblings Tina, David and Ben for being my most influential teachers during my early years. Special thanks to Dr. Stjepana Kovac and Dr. Rebecca Schuele for numerous stimulating discussions, to my wonderful friends outside the lab (especially Dr. Dr. Hassan Elbahtimy and Dr. Richard Rosch for delicious falafels, uncountable moral support and a resilient friendship) and to Dr. Thomas Ethofer and Professor Ludger Schoels for sparkling my interest in neurology.

This thesis is dedicated to all patients suffering from neurodegenerative conditions and, in loving memory, to Lucas. 6

Specific project acknowledgements and indication of collaborative work

Chapter 3: I am grateful to my supervisor Prof. Henry Houlden and our clinical collaborators who helped with case recruitment for this study. I was responsible for the design and overall organisation of the study, the clinical characterisation, post-hoc study of notes, DNA sample organisation and shipment, analysis of the data and organisation and analysis of Sanger sequencing and am grateful to Stephanie Efthymiou for any help with pipetting. I spent three months at the National Institutes of Health (NIH), Bethesda, MD, US to learn exome sequencing, including analysis. Thanks to Monica Federoff at NIH who, after my return to IoN, sequenced the majority of the exomes and brought over the raw data. I am equally grateful to Dr. Raphael Gibbs and Dr. Jinhui Ding in the US and Dr. Alan Pittman at IoN who helped and supervised me carrying out QC-curation of the data and gave important support in my bioinformatics data analysis. Additionally, I want to thank Dr. Charlie Arber, who collaboratively investigated the neuronal pathology in an iPSC-based model of PANK2 using provided fibroblasts of our patients, as well Dr. Francesca Mazzacuva (Institute of Child Health, UCL) and Dr. Yugo Tsuchya (Structural and Molecular Biology, UCL)) for lipid-, CoA- and VitB5-metabolism analyses in provided cell pellets, urine and serum of our clinically phenotyped atypical PKAN cases of Queen Square (QS). Chapter 4: 4.1: I am grateful to our colleagues in Greece who helped with clinical characterisation and sample collection of the studied family. 4.2: Thanks to my colleagues Dr. Joshua Hersheson and Debbie Hughes who did the initial exome and Miseq sequencing for the described study (1). 4.3: Regarding this subchapter, I want to specifically thank our colleagues in India who did the clinical characterisation and sample collection of the large pedigree. Thanks to Dr. Reema Paudel at Queen Square Brain Bank (QSBB) for help with sample handling and Sanger sequencing and to Dr. Joshua Hersheson for joint analysis of the genotyping and exome data. Chapter 5: I am grateful for this collaborative effort where several international groups and consortia, including ours, collected samples. Profs. P. Holmans, S. Tabrizi, L. Jones and H. Houlden designed the experiments, Drs. D. Hensman Moss, M. Flower and I collected clinical data. Drs. D. Hensman Moss, C. Bettencourt, M. Flower and me carried out the experiments and drafted figures, tables and manuscript jointly (2). Genotyping was done at LGC Genomics, Hertfordshire, UK. Thanks to Peter Holmans who guided and performed the statistical analyses. Chapter 6: Thanks to Chris Lovejoy and Elisavet Preza for help with initial biopsy handling and fibroblast culture, to Kerra Pearce for running the genotyping experiment, to Cell Guidance Systems Cambridge for karyotyping, to Dr. Sarah Crisp fo electrophysiological characterisation and patch-clamp, to Claire Hall, Debbie Hughes and Dr. Chris Sibley for help with RNA-sequencing and finally to Dr. Zhi Yao for support and guidance in live-cell imaging. Additional thanks to Dr. Rickie Patani for discussion of study design and data, to Dr. Giulia Tyzack for occasional help with cell culture and to Dr. Alexandra Zirra for RT-PCR guidance.

Abbreviations

AAO: Age at onset. aCP: Aceruloplasminemia. AD: Alzheimer’s disease or autosomal dominant. ADCA: Autosomal dominant spinocerebellar ataxia. ALS: Amyotrophic lateral sclerosis. AR: Autosomal recessive. ARCA: Autosomal recessive cerebellar ataxia. AT: Ataxia-telangiectasia. BPAN: Beta-propeller protein-associated neurodegeneration. BSA: Bovine serum albumin. CADD: Combined annotation dependent depletion. CBD: Corticobasal degeneration. CCDS: Consensus coding sequence. cDNA: Complementary DNA. CGH: Comparative genomic hybridization. cg69: Complete genomics 69 genomes data. Chr: Chromosome. CNS: Central nervous system. CNV: Copy number variation. CoA: Coenzyme A. CoPAN: COASY protein-associated neurodegeneration. CPNE: Cerebellar plate neuroepithelium. CSF: Cerebrospinal fluid. dbSNP: Database of short genetic variation. DCN: Deep cerebellar nuclei. ddH20: Double distilled water. DIV: Days in vitro. DM: Dorsomorphin. DMEM: Dulbecco’s Modified Eagle’s Medium. DMSO: Dimethyl sulfoxide. DNA: Deoxyribonucleic acid. DRD: Dopa responsive dystonia. DRPLA: Dentatorubral-pallidoluysian atrophy. DV: Dorsoventral. EDTA: Ethylenediaminetetraacetic acid. EEG: Electroencephalogram. EEX: Expanded exome oligos. 8

ENCODE: Encyclopedia of DNA elements. ER: Endoplasmic reticulum. ES: Embryonic stem or exome sequencing. ESC(s): Embryonic stem cell(s). ESP: Exome sequencing project (NHLBI). EVS: Exome variant server (NHLBI). ExAC: Exome Aggregation Consortium. FA: Friedreich’s ataxia. FAHN: Fatty acid hydroxylase-associated neurodegeneration. FBS: Fetal bovine serum. FGF: Fibroblast growth factor. FH: Family history. FTD: Frontotemporal dementia. g: Acceleration caused by gravity. gDNA: Genomic DNA. GERP: Genomic evolutionary rate profiling. GO: Gene ontology. GOSH: Great Ormond Street Hospital. GP: Globus pallidus. GWAS: Genome wide association analysis. HSS: Hallervorden-Spatz syndrome. HapMap: Haplotype map. HD: Huntington’s disease. HDL: Huntinton’s disease like. hES: Human embryonic stem. hESC(s): Human embryonic stem cell(s). het: Heterozygous. HGP: Human genome project. hiPSC(s): Human induced pluripotent stem cell(s). hom: Homozygous. HSP: Hereditary spastic paraplegia. ICC: Immunocytochemistry. IoN: Institute of Neurology, Queen Square London. iPS: Induced pluripotent stem. iPSC(s): Induced pluripotent stem cell(s). IP3R1: type 1 inositol 1,4,5-trisphosphate receptor. IsO: Isthmic organizer. ITPR1: Inositol-1, 4, 5,-trisphosphate-receptor-I gene.

9 kb: Kilobase. KRS: Kufor-Rakeb syndrome. LB: Lewy body. LD: Linkage disequilibrium. LoF: Loss-of-function. LTD: Long-term depression. LTP: Long-term potentiation. MAF: Minor allele frequency. max: Maximum. MCLDS: McLeod syndrome. MHB: Midbrain-hindbrain boundary/border. MIM #: Phenotype Mendelian Inheritance in Man number. min/mins: Minute(s) or minimum. MLPA: Multiplex ligation-dependent probe amplification. MPAN: Mitochondrial membrane protein-associated neurodegeneration. MR: Magnetic resonance. MRI: Magnetic resonance imaging. MSA: Multiple system atrophy. Mt/mut: Mutant. NA/n.a.: not available, not applicable. NB: nota bene. NBIA: Neurodegeneration with brain iron accumulation. NF: Neuroferritinopathy. NGS: Next-generation sequencing. NHNN: National Hospital for Neurology and Neurosurgery. NIH: National Institutes of Health.

NPSC(s): Neural precursor stem cell(s). NSC(s): Neural stem cell(s). nt: nucleotides. OMIM: Online Mendelian Inheritance in Man. OR: Odds ratio. PBS/DPBS: Phosphate-buffered saline/Dulbecco’s phosphate-buffered saline. PBST: PBS containing 0.3% triton-100. PC(s): Purkinje cell(s). PCR: Polymerase chain reaction. PD: Parkinson(´s) disease. PKAN: Pantothenate kinase-associated neurodegeneration. PLAN: PLA2G6- or Phospholipase A2-associated neurodegeneration.

PNP: Polyneuropathy. PSP: Progressive supranuclear palsy. QC: Quality control. qPCR: Quantitative PCR. QS: Queen Square (London). QSBB: Queen Square Brain Bank. RNA: ribonucleic acid. ROS: reactive oxygen species. rpm: revolutions per minute. rRNA: ribosomal RNA. RT: Room temperature. RT-PCR: Reverse transcriptase PCR. s: seconds. SB431542: TGF-beta/Activin/Nodal pathway-inhibitor. SBMA: Spinal and bulbar muscular atrophy. SBS: Sequencing by synthesis. SCA: Spinocerebellar ataxia. SCAR: Autosomal recessive spinocerebellar ataxia. SCA15: Spinocerebellar ataxia type 15. SD: Standard deviation. SEM: Standard error of the mean. SENDA: Static encephalopathy of childhood with neurodegeneration in adulthood. SIFT: Functional prediction tool provided by Craig Venter Institute. SN: Substantia nigra. SNP: Single-nucleotide polymorphism. SOCE: Store operated calcium entry.

SPB: Sample purification bead. SPG: Spastic gait gene/spastic paraplegia gene. TE: tris/EDTA. TCC: Thin corpus callosum. UK: United Kingdom. uk: unknown. USA/US: United States (of America). UTR: untranslated region. WB: Western Blot. (W)ES: (Whole) exome sequencing. WGS: Whole genome sequencing. WM: White matter.

WML: White matter lesions. WSS: Woodhouse-Sakati syndrome. WT: wildtype. XL(D): X-linked (dominant). 1000g: 1000 Genomes project.

List of tables

Table 1-1: Overview of NBIA conditions ...... 38 Table 1-2: Summary of trinucleotide repeat disorders ...... 69 Table 1-3: Neuronal differentiation protocols from iPSCs ...... 79 Table 2-1: Candidate genes investigated ...... 102 Table 2-2: Cohort characteristics ...... 110 Table 2-3: Characteristics of genotyped SNPs ...... 111 Table 2-4: Effects of repeat length of expanded allele on AAO ...... 114 Table 3-1: Clinical summary of NBIA cohort ...... 131 Table 3-2: Highly likely pathogenic variants NBIA cohort ...... 136 Table 3-3: Variants of unclear pathogenicity NBIA cohort ...... 137 Table 3-4: Interesting variants not confirmed by Sanger ...... 139 Table 4-1: Recently reported SYNE1 mutations after discovery paper ...... 182 Table 4-2: Novel homozygous PNPLA6 variants identified ...... 190 Table 5-1: Results of combined analysis of SNPs ...... 196 Table 5-2: Individual SNPs significantly associated with AAO ...... 197 Table 5-3: Characteristics of SCA6 cohorts ...... 205 Table 6-1: Clinical findings of SCA15 patients ...... 214 Table 6-2: Electrophysiological properties iPSC-derived neurons ...... 234 Table 8-0: Non-exhaustive list of important genes in neurodegeneration…………… 282 Table 8-1: Primer sequences candidate genes ...... 295 Table 8-2: Primer sequences XK ...... 297 Table 8-3: Primer sequences SYNE1 ...... 297 Table 8-4: Primer sequences PNPLA6 ...... 298 Table 8-5: Primer sequences ATCAY ...... 300 Table 8-6: Primer sequences CACNA1A ...... 300

Table 8-7: PCR master mix, default PCR reaction ...... 301 Table 8-8: PCR master mix default + DMSO ...... 301 Table 8-9: PCR master mix, default + DMSO + Solution Q ...... 302 Table 8-10: SCA6 PCR master mix ...... 302 Table 8-11: McLeod-TD70_65 programme ...... 303 Table 8-12: TD65_55long PCR programme ...... 303 Table 8-13: TD65_55 PCR programme ...... 304 Table 8-14: TD65_60 PCR programme ...... 304 Table 8-15: TD60_55 PCR programme ...... 305 Table 8-16: SCA6 PCR programme ...... 305 Table 8-17: Exosap PCR product cleanup ...... 306 Table 8-18: Pipetting scheme sequencing PCR ...... 306 13

Table 8-19: Sequencing PCR ...... 306 Table 8-20: Seed sense sequences for KASP assay design ...... 308 Table 8-21: Single SNP associations ...... 309 Table 8-22: Control iPSC and ESC lines ...... 312 Table 8-23: Stem cell culture and media reagents ...... 312 Table 8-24: N2B27 recipe and catalogue numbers ...... 313 Table 8-25: Extrinsic cues for differentiation of iPSCs ...... 314 Table 8-26: qPCR primers SCA15 iPSC project ...... 316 Table 8-27: qPCR primers developmental iPSC project ...... 317 Table 8-28: Reverse transcription using SuperScript III step 1 ...... 320 Table 8-29: Reverse transcription using SuperScript III step 2 ...... 320 Table 8-30: RT-PCR cycling conditions SuperScript III ...... 320 Table 8-31: Reverse transcription using RevertAid step 1 ...... 321 Table 8-32: Reverse transcription using RevertAid step 2 ...... 321 Table 8-33: RT-PCR cycling conditions RevertAid ...... 321 Table 8-34: qPCR master mix, 96-well format, Mx3000P qPCR System ...... 322 Table 8-35: qPCR cycling conditions on Mx3000P ...... 322 Table 8-36: qPCR master mix, 384-well format, QuantStudio 6 Flex Real-Time PCR System ...... 322 Table 8-37: qPCR cycling conditions on QuantStudio ...... 323 Table 8-38: Primary antibodies Chapter 6.1 ...... 323 Table 8-39: Primary antibodies Chapter 6.2 ...... 323 Table 8-40: Secondary antibodies ICC ...... 324

List of figures

Figure 1-1: Important milestones towards the sequencing of the human genome ...... 23 Figure 1-2: Sequencing costs per genome (2001-2015) ...... 25 Figure 1-3: Conventional versus second-generation sequencing ...... 28 Figure 1-4: Timeline of discoveries in genetics and neurodegeneration ...... 30 Figure 1-5: Missing heritability ...... 31 Figure 1-6: Imaging characteristics NBIA ...... 36 Figure 1-7: Neuropathology of PKAN, the most frequent NBIA disorder ...... 37 Figure 1-8: Neuropathology of PLAN ...... 47 Figure 1-9: Biochemical pathways and cellular processes in PKAN and PLAN ...... 49 Figure 1-10: WDR45: genetic architecture and reported mutations ...... 54 Figure 1-11: A clinical and radiographic diagnostic algorithm for NBIA ...... 62 Figure 1-12: NBIA genes and pathways ...... 66 Figure 1-13: Proposed pathomechanisms in trinucleotide repeat disorders ...... 71 Figure 1-14: Diagnostic algorithm progressive ataxias ...... 73 Figure 1-15: Scheme of induced pluripotent stem cell technology ...... 76 Figure 1-16: iPSC technology in cerebellar diseases ...... 82 Figure 1-17: Cell migration during cerebellar development ...... 85 Figure 1-18: Muguruma et al. protocol for hESC-derived cerebellar cells ...... 87 Figure 1-19: Timeline modelling cerebellar disease using hiPSC ...... 89 Figure 2-1: Exome sequencing workflow ...... 93 Figure 2-2: Next-generation sequencing steps ...... 94 Figure 2-3: Nextera Rapid Capture protocol overview ...... 95 Figure 2-4: Tagmentation process ...... 96 Figure 2-5: Timeline of experiments ...... 115 Figure 2-6: Six different neural induction conditions, experiment 1 ...... 124

Figure 3-1: NBIA subtypes and frequency distribution ...... 128 Figure 3-2: Clinical symptoms NBIA cohort ...... 130 Figure 3-3: Clinical symptoms NBIA cohort, absolute symptom counts ...... 130 Figure 3-4: Coverage metrics NBIA cohort ...... 133 Figure 3-5: Variant filtering strategy candidate genes ...... 134 Figure 3-6: Family tree, MRI and segregation status PLA2G6-case ...... 140 Figure 3-7: Family tree and segregation status FA2H-positive case ...... 142 Figure 3-8: PANK2 mutations ...... 143 Figure 3-9: Family tree, chromatogram and MRI of neuroferritinopathy case ...... 145 Figure 3-10: BAM-file and chromatograms FTL mutation ...... 146 Figure 3-11: WDR45 frameshift insertion and MRI findings ...... 147 Figure 3-12: Pathogenic stopgain mutation FUCA1 ...... 148 15

Figure 3-13: L1CAM variant and MRI ...... 150 Figure 4-1: Acanthocytes in peripheral blood ...... 170 Figure 4-2: T1- and T2-weighted MRI McLeod case ...... 171 Figure 4-3: Family tree, segregation and conservation status of novel XK mutation .. 172 Figure 4-4: Clinical and genetic findings in family I ...... 177 Figure 4-5: Genetic and MRI findings in patients II:1 and III:1 ...... 178 Figure 4-6: Recently reported and novel SYNE1 mutations, their location and phenotype ...... 181 Figure 4-7: PNPLA6-associated phenotypes ...... 185 Figure 4-8: Family tree of large Parsi kindred ...... 186 Figure 4-9: MRI of index case IV-7 ...... 187 Figure 4-10: Homozygous stretch shared by affected cousins ...... 188 Figure 4-11: Filtering strategy, segregation and conservation status PNPLA6 variants ...... 189 Figure 5-1: Residual AAO boxplot ...... 198 Figure 5-2: Linking DNA repair with somatic instability ...... 201 Figure 5-3: Inverse correlation of expanded allele size and AAO in SCA6 subcohort 206 Figure 5-4: Inverse correlation of sum allele size and AAO in SCA6 subcohort ...... 206 Figure 5-5: CACNA1A repeat sequencing schematic and exemplary results unexpanded allele ...... 207 Figure 5-6: CACNA1A repeat sequencing schematic and exemplary results with expanded allele ...... 208 Figure 6-1: Pedigrees SCA15 families and ITPR1 deletion sizes ...... 214 Figure 6-2: Sagittal MRI of individual ST at age 37 ...... 215 Figure 6-3: Developmental stages: Punch biopsy, fibroblasts, iPSCs, neuronal precursors ...... 216 Figure 6-4: Brightfield morphology iPSC clones ...... 217 Figure 6-5: Loss of heterozygosity over SUMF1/ITPR1 locus ...... 219 Figure 6-6: Normal G-band karyotyping results ...... 220 Figure 6-7: Pluripotency marker expression iPSC lines ...... 221 Figure 6-8: Pluripotency immunocytochemistry iPSC lines ...... 222 Figure 6-9: Pluripotency immunocytochemistry hESC line ...... 222 Figure 6-10: S100A4 expression fibroblasts, iPSC lines ...... 223 Figure 6-11: Endogenous and exogenous pluripotency marker expression ...... 224 Figure 6-12: Immunocytochemistry neural precursors ...... 225 Figure 6-13: Immunocytochemistry cortical control neurons DIV80 ...... 226 Figure 6-14: Immunocytochemistry cortical patient neurons DIV80 ...... 226 Figure 6-15: Immunocytochemistry control neurons, single channel data ...... 227

Figure 6-16: Immunocytochemistry control neurons and glia, single channel data ..... 228 Figure 6-17: Immunocytochemistry control glia cells DIV100 ...... 229 Figure 6-18: Immunocytochemistry patient glia cells DIV100 ...... 229 Figure 6-19: Astrocyte and radial glia DIV100 ...... 230 Figure 6-20: Exemplary dynamic traces, thapsigargin induced calcium stores ...... 231 Figure 6-21: ER calcium and store operated calcium entry in iPSC-derived neurons 232 Figure 6-22: Electrophysiological activity of iPSC-derived cortical neurons ...... 233 Figure 6-23: Rationale of developmental experiments ...... 242 Figure 6-24: Pluripotency marker expression timecourse for six different conditions . 243 Figure 6-25: GATA4 and S100A4 expression timecourse for six different conditions . 243 Figure 6-26: Timecourse neural markers six different conditions ...... 244 Figure 6-27: Immunocytochemistry OTX2-positive cells DIV4 and 7, six different conditions ...... 245 Figure 6-28: qPCR expression profile of different regional markers at DIV 14 ...... 246 Figure 7-1: Future work iPSCs (SCA15 and cerebellar differentiation) ...... 252

Publications/Book chapters

Part of the work described in this thesis has been published in the following articles:

1) Wiethoff S#, Arber C, Li A, Wray S, Houlden H*l, Patani R*l. Using human induced pluripotent stem cells (hiPSC) to model cerebellar disease: Hope and hype. J Neurogenet. 2015;29: 95-102. 2) Wiethoff S, Bhatia KP, Houlden H. Genetics of NBIA Disorders. Book chapter in: Movement Disorder Genetics (pp. 263-291), 2015 Springer International Publishing. 3) Wiethoff S, Houlden H. Neurodegeneration with brain iron accumulation. Book chapter in: Kovacs and Alafuzoff: Neuropathology (Handbook of Clinical Neurology, submitted). 4) Pihlstrom L, Wiethoff S, Houlden H. Genetics of neurodegenerative diseases: An overview. Book chapter in: Kovacs and Alafuzoff: Neuropathology (Handbook of Clinical Neurology, submitted). 5) Wiethoff S*, Xiromerisiou G*, Bettencourt C, Kioumi A, Tsiptsios I, Tychalas A, Evaggelia M, George K, Makris V, Hardy J, Houlden H. Novel single base-pair deletion in exon 1 of XK gene leading to McLeod syndrome with chorea, muscle wasting, peripheral neuropathy, acanthocytosis and haemolysis. J Neurol Sci. 2014; 339: 220-222. 6) Wiethoff S#, Bettencourt C, Paudel R, Madon P, Liu YT, Hersheson J, Wadia N, Desai J, Houlden H. Pure cerebellar ataxia with homozygous mutations in the PNPLA6 gene. Cerebellum. 2016 Mar 19. [Epub ahead of print]. 7) Wiethoff S*, Hersheson J*, Bettencourt C, Wood NW, Houlden H. Heterogeneity and spectrum of severity in ataxia associated SYNE1 mutations. Journal of Neurology. 2016 May 13. [Epub ahead of print]. 8) Bettencourt C*, Hensman Moss D*, Flower M*, Wiethoff S*, Brice A, Goizet C, Stevanin G, Koutsis G, Karadima G, Panas M, Yescas-Gómez P, García-Velázquez LE, Alonso-Vilatela ME, Lima M, Raposo M, Traynor B, Sweeney M, Wood N, Giunti P, Durr A for the French SPATAX network, Holmans P*l, Houlden H*l, Tabrizi SJ*l, Jones L*l. DNA repair pathways underlie a common genetic mechanism modulating onset in polyglutamine diseases. Ann Neurol. 2016 Apr 4. doi: 10.1002/ana.24656. [Epub ahead of print]. 9) Li A, Wiethoff S, Arber C, Houlden H, Revesz T, Holton JL Pathology and Genetics of Neuroaxonal Dystrophy/ Neurodegeneration with Brain Iron Accumulation. Review Article in Developmental Neuropathology, 2nd Edition, ISN Book Series. (in press).2016.

Other publications (co)authored during this thesis: -Schüle R*, Wiethoff S*, Martus P, Karle KN, Otto S, Klebe S, Klimpe S, Gallenmüller C, Kurzwelly D, Henkel D, Rimmele F, Stolze H, Kohl Z, Kassubek J, Klockgether T, Vielhaber S, Kamm C, Klopstock T, Bauer P, Züchner S, Liepelt-Scarfone I, Schöls L. Hereditary Spastic Paraplegia -clinico-genetic lessons from 608 patients. Ann Neurol. 2016 Feb 9. doi: 10.1002/ana.24611. Epub 2016 Mar 11. -Zirra A, Wiethoff S, Patani R. Neural conversion and early patterning of human pluripotent stem cells: a developmental perspective. Stem Cells Int. 2016;2016:8291260. doi: 10.1155/2016/8291260. Epub 2016 Mar 16.

-Wiethoff S*, Hamada M*, Rothwell JC. Variability in response to transcranial direct current stimulation of the motor cortex. Brain Stimul. 2014;7: 468-475. -Bettencourt C, Forabosco P, Wiethoff S, Heidari M, Johnstone DM, Botía JA, Collingwood JF, Hardy J; UK Brain Expression Consortium (UKBEC), Milward EA, Ryten M*l, Houlden H*l. Gene co-expression networks shed light into diseases of brain iron accumulation. Neurobiol Dis. 2016;87:59-68. -Paudel R*, Li A*, Wiethoff S, Bandopadhyay R, Bhatia KP, de Silva R, Houlden H, Holton J. Neuropathology of Beta-propeller protein associated neurodegeneration (BPAN): a new tauopathy. Acta Neuropathol Commun. 2015: 30;3:39. -Mencacci N*, Erro R*, Wiethoff S, Hersheson J, Ryten M, Balint B, Ganos C, Stamelou M, Quinn N, Houlden H, Wood N, Bhatia KP. ADCY5 mutations, another cause of benign hereditary chorea. Neurology 2015;85:80-88. -Mencacci NE*, Kamsteeg EJ*, Nakashima K*, R’Bibo L, Lynch DS, Balint B, Willemsen M, Adams ME, Wiethoff S, Suzuki K, Davies CH, Ng J, Meyer E, Veneziano L, Giunti P, Hughes D, Raymond L, Carecchio M, Zorzi G, Nardocci N, Barzaghi C, Garavaglia B, Salpietro V, Hardy J, Pittman AM, Houlden H, Kurian M, Kimura H*l, Vissers L*l, Wood NW*l, Bhatia KP*l. De novo mutations in PDE10A cause childhood-onset chorea with bilateral striatal lesions. Am J Hum Genet. 2016 Apr 7;98(4):763-71. doi: 10.1016/j.ajhg.2016.02.015. -Madeo M*, Stewart M*, Sun Y, Sahir N, Wiethoff S, Chandrasekar I, Yarrow A, Cordeiro D, McCormick EM, Murarescu CC, Mokry JA, Yang Y, McBeth L, Jepperson TN, Seidahmed M, El Khashab H, Hamad M, Azzedine H, Clark K, Burn S, Myers A, Landsverk M, Crotwell PL, Houlden H, Padilla-Lopez S, Nolan PM, Singh B, Lifton R, Mane M, Bilguvar K*l, Falk M*l, Mercimek-Mahmutoglu S*l, Salih MA*l, Acevedo- Arozena A*l, Kruer MC*l #. Loss-of-function mutations in FRRS1L attenuate AMPA currents and lead to a hyperkinetic movement disorder in mice and humans. Am J Hum Genet. 2016 Jun 2;98(6):1249-55. -Brugger F, Kägi G, Pandolfo M, Mencacci N, Batla A, Wiethoff S, Bhatia K. Neurodegeneration with brain iron accumulation (NBIA) syndromes presenting with late-onset craniocervical dystonia – an illustrative case series . Movement Disorders Clinical Practice (accepted, ahead of publication). -Kara E*, Tucci A*, Manzoni C*, Lynch DS, Elpidorou M, Bettencourt C, Chelban V, Manole A, Hamed S, Haridy N, Federoff M, Preza E, Hughes D, Pittman A, Jaunmuktane Z, Brandner S, Xiromerisiou G, Wiethoff S, Schottlaender L, Proukakis C, Morris H, Warner T, Bhatia KP, Korlipara P, Singleton A, Hardy J, Wood NW, Lewis PA, Houlden H. Genetic, functional and phenotypic characterisation of complex hereditary spastic paraplegia. Brain. 2016 Jul;139(Pt 7):1904-18. -Wang Y, Hersheson J, Lopez D, Ben Hamad M, Liu Y, Lee K, Pinto V, Seinfeld J, Wiethoff S, Sun J, Amouri R, Hentati F, Baudry N, Tran J, Singleton AB, Coutelier M, Stevanin G, Bi X, Houlden H*l, Baudry M*l. Defects in the CAPN1 gene result in cerebellar ataxia and granule cell loss in mice and humans. Cell Rep. 2016 Jun 28;16(1):79-91. -Arber C, Angelova PR, Wiethoff S, Tsuchiya Y, Mazzacuva F, Houlden H, Clayton P, Mills K, Abramov A, Gout I, Duce J and Wray S. iPSC-derived neurons from patients with PANK2 mutations to investigate the molecular mechanisms of neurodegeneration (in preparation).

* = joint shared first with equal contributions (order not changed). # = corresponding author. *l = joint shared last with equal contributions (order not changed).

1 Chapter 1: General introduction

1.1 General introduction and aims of this thesis

Within the last decade, molecular and genetic techniques have improved significantly, allowing the generation of high-throughput genetic data within significantly shortened timeframes. Personalised genetic data of patients inform clinicians in research and clinical diagnostic settings about genetic variation that may lead to neurodegenerative disease. However, large parts of the obtained data remain difficult to interpret and genotypic heterogeneity, phenotypic variability within and beyond families, and phenomena such as reduced penetrance, missing heritability and variable expressivity only point towards the tip of the iceberg of unresolved factors towards successful and informed interpretation of the detailed genetic data available. Despite the advances in technology, data interpretation, data sharing and world-wide collaboration, no genetic diagnosis can be obtained for a number of patients with different inherited neurodegenerative diseases to date, even with thorough genetic investigation. For another significant proportion of neurodegenerative diseases the genetic defect and the resulting clinical phenotype is known. However, the underlying molecular and cellular pathogenic events and disease initiating cascades remain understudied. In many cases, it is challenging to create an appropriate disease model and the generated models might have inherent limitations (e.g. artificial overexpression systems in non-complex organisms or cell lines, tissue decay when dealing with post- mortem samples, biological differences between animal and human organism when studying an animal model, etc.). This hinders identification of key pathogenic events and the effective search for potential revertive treatments that can target the molecular and cellular pathways identified. Thereby, it significantly delays successful translational research and eventual clinical treatment.

In this thesis, both issues are addressed: Two main techniques that already have and will further advance the field in its search to understand, and potentially alter the pathogenesis of relentless neurodegenerative diseases are employed. Firstly, a number of genetic techniques, mainly dominated by whole exome and Sanger sequencing and targeted genotyping, is used on a cohort of clinically characterised idiopathic NBIA-like cases as well as on families/patients from the outpatients’ clinics with hereditary ataxias, and on a cohort of trinucleotide repeat disorders to find and study the underlying genetic mechanisms contributing to disease.

Secondly, induced pluripotent stem cell technology is used to better understand the functional pathomechanisms in rare Mendelian disorders with a focus on the orphan

20 disease spinocerebellar ataxia type 15 (SCA15, MIM #606658), a disorder where the underlying genetic defect was identified in 2007, but the exact pathomechanism leading to (cerebellar) neurodegeneration in humans remains inconclusive. Generation of the iPSC model as well as directed differentiation towards cortical neurons, establishment of functional readouts, and the attempt to generate cerebellar neuronal derivatives has proven challenging, however the work shown here might facilitate studying cerebellar diseases using iPSCs in the future.

1.2 Genetics in neurodegenerative disease

Introductory remarks This introductory subchapter prepares the reader for results chapters 3, 4 and 5 and lends a general overview to the role of genetics in neurodegeneration. Parts of the contents of this introductory subchapter are modified from a submitted overview article about the role of genetics in neurodegeneration in ‘Kovacs and Alafuzoff: Neuropathology (Handbook of Clinical Neurology)’. This introductory subchapter intends to refer to some of the major developments in the field of neurogenetics to date in order to highlight overarching historical trends and general insights with a focus on monogenic neurodegenerative diseases from the perspectives of both rare Mendelian forms of more common disorders, such as Alzheimer's (AD) and Parkinson's disease (PD), and heterogeneous less frequent inherited conditions, including ataxias and hereditary spastic paraplegias (HSPs). It does not intend to provide a detailed account of the vast number of genetic factors contributing to different neurodegenerative phenotypes but emphasizes important general insights through illustrative selected examples.

Genetic factors play a principal role in the aetiology of neurodegeneration. Genetic variation can act as a monogenic cause of heritable disease, as modifiers of susceptibility to complex, sporadic disorders and in a complex interplay of both. When looking at only the coding 1.5% of an individual’s genome, more than 20000 different variants can be observed in a healthy individual. For clinicians and clinically motivated scientists, the challenge for the present and next decades is and will be to disentangle disease causing from tolerated variants, and thereby not to restrict oneself to only the coding part of the human genome.

The publication of the initial drafts of the human genome in 2001 by two large initiatives originating from the public and the academic sector have been crucially important milestones in the early understanding and stimulation of further downstream genetic research (3, 4) – for an overview of important selected landmark efforts leading up to the decipherment of the human genome between 1990 and 2001, see Figure 1-1.

Figure 1-1: Important milestones towards the sequencing of the human genome

Timeline of large scale genomic projects with selected components of work on several non-vertebrate model organisms (red), the mouse (blue) and the human (green) leading to the sequencing of the human genome in 2001 from 1990 on. Abbreviations: SNPs, single-nucleotide polymorphisms; ESTs, expressed sequence tags. Figure reproduced from (3).

The 99,9%-complete publication of the human genome in 2004 (5, 6), and the further Human Genome Project (HGP) and succeeding ENCODE Project landmark papers on structural variation within the human genome (2008), and the analysis of functional deoxyribonucleic acid (DNA) elements (2007) further confirmed that the decipherment of the human genome in 2001 was not the end point, but rather the beginning to the understanding of human genetic makeup contributing to health and disease (7, 8). However, several surprising features from these initial studies are worth to be retained separately:

Firstly, even though the human genome was revealed to consist of around three billion bases, the HGP identified only about 20500 human genes, a number only twice as much as species such as worm or fly, and much lower than previously estimated with numbers ranging from 50000 genes to 140000. Secondly, segmental duplications consisting of nearly identical repeated sequences were much more frequent than previously thought, and finally, very few protein families encoded by the human DNA seemed to be vertebrate specific.

The initial findings were quickly stimulating the formation of several large international consortia such as the International HapMap Project (~2002-2009) aiming to develop a human haplotype map, the 1000 Genomes Project (1000g; ~2008-2015), UK10K (~2010-ongoing) and others. These efforts involve large scale deep-sequencing of more than 1000 genomes to uncover the human genetic reference variation throughout various different ethnicities, and in healthy, and affected individuals. The findings of these consortia have been published throughout the years (with some still being awaited), and have enriched current knowledge about human genome architecture, function and genetic variation (9-13) significantly. These landmark studies have also yielded novel and crucial insights regarding non-protein-coding ribonucleic acid (RNA) genes, gene number and density, high copy number repeat sequences, pervasive transcription, evolutionary conservation and disease causing and/or influencing variation.

The parallel advent of the first next-generation sequencers in 2004 (e.g. Roche GS 20 System; Roche 454; Branford, CT) and the ever since rapid expansion of high- throughput sequencing techniques to nowadays third-generation single molecules sequencers (e.g. Pacific Biosciences Menlo Park, CA; Helicos Biosciences, Cambridge, MA) renders such ambitions possible: However, the challenge slowly shifts from the technical possibilities in sequencing which are getting cheaper, better, quicker and more accurate rapidly, towards data storage, data analysis, data sharing and data interpretation which already does and further will represent the real challenge of the future.

Beyond promising and diverse applications in functional genomic studies, novel sequencing technologies and the constant advancement have led to “an inflation of sequencing cost and time” more rapidly than Moore’s law, enabling one now to sequence a whole human genome within days for a continuously dropping price (14) – see Figure 1-2.

Figure 1-2: Sequencing costs per genome (2001-2015)

Typical cost of sequencing a human-sized genome, on a logarithmic scale. Note the drastic trend faster than Moore's law beginning in 2008 as post-Sanger sequencing came online at sequencing centres. Figure and legend reproduced from (14).

All these precedents are necessary to foster the ongoing hunt for genetic factors contributing to human disease, and more specifically to human neurodegeneration. Over the last two decades, the identification of disease genes and risk loci has led to some of the greatest advances in medicine and invaluable insights into pathogenic mechanisms and disease pathways. Large scale research efforts, novel study designs and advances in methodology are rapidly expanding the understanding of the genome, thereby facilitating uncovering of the genetic architecture of neurodegenerative disease.

Genetics plays an essential role in translational research in neurology, ultimately aiming to develop novel disease-modifying therapies for neurodegenerative disorders. It has provided novel insights into the pathogenesis of neurodegeneration, and at the same time also uncovered novel layers of complexity. The field evolves at a dramatic pace and is expected to increasingly influence diagnostics and treatment. With advances in techniques and interpretation of findings, it is to be anticipated that individual genetic profiling will become increasingly relevant in a clinical context, with implications for patient care in line with the proposed ideal of personalised medicine (15).

Disease genetics – a brief historical overview with a focus on neurology The history of progress in disease genetics to some extent mirrors the technological advances in molecular biology, and the findings in neurodegeneration illustrate this. Ever since the pioneer mapping of the underlying genetic cause of Huntington's disease (HD, MIM #143100) in 1983 (16), neurodegenerative conditions have been central in this development. For this disease causing gene discovered in 1993, researchers were initially only able to demonstrate linkage of the disease phenotype to a marker on the short arm of chromosome 4 by using the analysis of restriction fragment-length polymorphisms. However, this achievement sparked motivation and early optimism that the identification of this and other disease genes and mutations would be feasible (17). For the case of HD, it would still take 10 years of intensive efforts before the causative mutation - an expansion of a CAG trinucleotide repeat in the gene nowadays known as HTT (18) - was discovered. This exemplarily illustrates how laborious and challenging gene identification was in this early phase, before comprehensive chromosomal maps as discussed above were available (19).

In the 1990s, linkage analysis and positional cloning evolved to become the paradigmatic approach for identification of disease causing genes and mutations. The aim of linkage analysis is to identify a set of chromosomal markers that co-segregates with a disease phenotype. Therefore, its statistical power and subsequent likelihood of success is dependent on large, multigenerational pedigrees with multiple affected individuals. If successful, the genetic locus may be narrowed down to a region of several million base pairs (Mb), where all coding genes can ideally be sequenced and searched for potential disease causing mutations. This strategy produced a wave of significant discoveries, including the genes for prion disease (20), familial AD (21) and spinocerebellar ataxia type 1 (SCA1, MIM #164400) (22) as prominent early examples from the neurodegenerative field – for a non-exhaustive table of genes implicated in neurodegeneration, their predominant phenotype and the respective year of discovery, see Table 8-0 in the Appendix.

Following the identification of novel disease genes, screening in larger cohorts of patients with similar phenotypes would follow, providing a gradually evolving picture of the clinicogenetic heterogeneity, mutation frequency and molecular interrelations of neurodegenerative disorders. Notably, sequencing in these years was done by elaborate Sanger sequencing, a technique that was invented 1977 by Fred Sanger (23), earning him his second Nobel Prize in 1980. Sanger sequencing is a technique revolutionary and ingenious in its own, however compared to nowadays’ approaches of high-throughput sequencing elaborate and time-consuming at the same time (see

2 6

Figure 1-3 for a generic comparison scheme between Sanger and next-generation sequencing (NGS)). Depending on the disease frequency, it can take years to identify patient groups big enough to do clinical studies and learn about the clinical course of the now genetically confirmed subgroups, including natural history, disease progression and longitudinal studies (24-26). With the discovery of the molecular genetic underpinnings, this tendency has even progressed further. This is illustrated nicely by the previously thought monogenic rare disorder HSP: For a long time HSP was associated only with a few causal genes, and a relatively easy clinical classification into complex and pure subforms ruled clinical practice for many years (27). However, over the years it has developed into a disease with more than 80 different known genes or loci and respective clinicogenetic subgroups rendering subgroup phenotype-genotype correlations challenging (26, 28). Thus, with the advent of detailed genetic knowledge, disease classification has become clearer on the one and more complex on the other hand.

However, powerful next-generation sequencing technologies such as sequencing panels are also being increasingly employed in diagnostic settings yielding a definite molecular diagnosis for more and more patients in shorter timespans. With regards to rare diseases, this is of immense value to individual patients who have had years and years without (genetic) diagnosis, ending their odyssey from hospital to hospital with a diagnosis influencing their prognosis and family planning. On very rare occasions, the identification of novel disease genes might immediately draw attention towards a specific pathway/molecule that is druggable/replacable such as riboflavin replacement treatment in the example of Brown-Vialetto-Van Laere syndrome 1 (BVVLS1, MIM #211530) (29). In most cases however, initial gene discoveries can only hope to improve treatment indirectly, by increasing the basic understanding of molecular pathogenic processes and generating hypotheses and stimuli for further mechanistically rationalised and translational research.

Figure 1-3: Conventional versus second-generation sequencing

(a) With high-throughput shotgun Sanger sequencing, genomic DNA is fragmented, then cloned to a plasmid vector and used to transform E. coli. For each sequencing reaction, a single bacterial colony is picked and plasmid DNA isolated. Each cycle sequencing reaction takes place within a microliter-scale volume, generating a ladder of ddNTP- terminated, dye-labeled products, which are subjected to high-resolution electrophoretic separation within one of 96 or 384 capillaries in one run of a sequencing instrument. As fluorescently labeled fragments of discrete sizes pass a detector, the four-channel emission spectrum is used to generate a sequencing trace. (b) In shotgun sequencing with cyclic-array methods, common adaptors are ligated to fragmented genomic DNA, which is then subjected to one of several protocols that results in an array of millions of spatially immobilized PCR colonies or 'polonies'. Each polony consists of many copies of a single shotgun library fragment. As all polonies are tethered to a planar array, a single microliter-scale reagent volume (e.g., for primer hybridisation and then for enzymatic extension reactions) can be applied to manipulate all array features in parallel. Similarly, imaging-based detection of fluorescent labels incorporated with each extension can be used to acquire sequencing data on all features in parallel. Successive iterations of enzymatic interrogation and imaging are used to build up a contiguous sequencing read for each array feature. Figure and legend reproduced from (30).

Table 8-0 illustrates how the field has expanded since the initial gene discoveries in the early 90s and how the number of discovered genes per disease category has exponentially grown since usage of next-generation sequencing has become widely available. Over the years, linkage and cloning approaches have been replaced by (linkage analysis followed by) exome sequencing, but both strategies have proven extremely successful in large families for disorders with classical Mendelian inheritance. However, the identification of genetic factors modulating disease risk in complex sporadic disorders is an important additional aim. One influential early example of striking effect size was genetic variability in apolipoprotein E, where one copy of the common ε4 variant increases risk to develop AD 2-3-fold (31, 32). However, it soon became clear, that common variants of similar effect size are generally rare and difficult to identify in complex disorders. One of the difficulties herein is caused by the fact that identification of loci with smaller impact on risk requires large scale case-control association studies for adequate statistical power.

Briefly summarised, more than a decade of genome-wide association studies (GWAS) has revealed that most common susceptibility variants carry only a very small effect on risk (odds ratio (OR) range of 1.0 to 1.4), and other than the APOE alleles that are protein altering single-nucleotide polymorphisms (SNPs), most GWAS signals originate from non-coding regions. As the reported SNPs are only markers tagging variation across different regions along the genome, the disease risk affecting variants are usually unknown, and it is thought they alter different functional elements of regulatory DNA. The disease-relevant gene can be challenging to determine if a GWAS locus spans several genes and no obvious candidate with convincing expression patterns in disease-relevant cell types or previous implication in Mendelian disease is present. A minority of signals revealed by GWAS may immediately serve to highlight a mechanism or disease pathway. As an example, significant associations with SNPs in the HLA region on chromosome 6 have been found for both AD, PD and for frontotemporal dementia (FTD), implicating immune mechanisms in the pathogenesis of these diseases, however, the implications are usually not straightforward, and significant post-GWAS efforts are needed to link GWAS findings to molecular pathways and disease mechanisms, as recently exemplified in schizophrenia (33).

In summary, from the turn of the millennium, developments in technology, computational resources and data analysis tools have been crucial in facilitating novel breakthroughs in both monogenic and complex disease genetics. The sequencing of the human genome provided an essential scientific fundament (3, 4) to be followed by international large scale efforts such as the International HapMap and the 1000 29

Genomes projects as introduced further above, to characterize the genetic variability between individuals. Additionally and simultaneously, the range of methodological advances, including high coverage genotyping arrays and high-throughput parallel sequencing facilitated genetic data generation on unprecedented scales within significantly reduced timeframes. This was an advantage in the fields of Mendelian and complex genetics equally. For complex genetics, the genome-wide association study became the dominant research design from 2005 onwards. In this approach, several million single-nucleotide polymorphism markers across the genome are investigated for their association with specific features such as disease or trait in an unbiased, hypothesis-free approach (34). For the monogenic diseases, high-throughput exome sequencing facilitates novel disease identification strategies that are independent of precedent time- and cost-intensive single-gene sequencing within previously identified linkage regions from large pedigrees. High-throughput sequencing facilitated a second wave of Mendelian gene discoveries from approximately 2010 onwards (see Figure 1-4 for a summary of selected important advances shaping technical issues around sequencing, reference variation of the human genome and disease implicating discoveries over time).

Figure 1-4: Timeline of discoveries in genetics and neurodegeneration

A non-exhaustive selection of important events shaping the field of (neuro-)genetics since 1983.

With more complex pictures arising, sequencing approaches are also showing increasing support for the hypothesis that multiple rare variants may carry an intermediate effect on disease risk. This indicates a continuous spectrum of causality from the common susceptibility variants of GWAS to the highly penetrant disease causing mutations.

While the rare variants are normally classified as strong disease influencing factors, there is clearly no sharp line between these and variants that can be called dominant

mutations with reduced penetrance. Supportive of this, several recent studies have found enrichment of rare coding variants in monogenic genes in patients with late- onset or sporadic forms of PD and AD (35, 36). It is evident that current sequencing- based genetic discoveries overthrow the traditional dichotomy between Mendelian and complex genetics in favour of a continuous spectrum model.

Unresolved issues and future prospects: Missing heritability, genetic heterogeneity and personalised medicine In theory, the sum effect of all genetic risk factors for one disease should mirror its heritability as estimated from twin studies or observations of familial clustering in epidemiological studies. It became evident early from GWAS that even though a number of susceptibility variants were found, these only explained a small proportion of the overall heritable risk. This phenomenon has been described as the 'missing heritability' problem (37). Novel strategies are required to capture missing heritability and various hypotheses have been brought forward to explain where it is hiding (see Figure 1-5 for a graphic conceptualisation).

Figure 1-5: Missing heritability

Relationship of effect sizes and minor allele frequencies of different disease associated variants. Most emphasis and interest lies in identifying associations with characteristics shown within diagonal dotted lines. However, missing heritability may hide outside those, being more difficult to identify. Figure reproduced from (37).

As mentioned earlier, rare high-risk variants undetectable by GWAS may be an 31

important contributor of the missing heritability. Several international consortia are currently expanding the genetic investigations towards large scale cohorts employing and combining a range of different techniques from standard genotyping arrays to whole exome sequencing, whole genome sequencing or custom arrays enriched with rare variants of potential relevance to neurodegenerative disorders. These datasets call for novel approaches towards data sharing, collaboration, data storage and importantly towards statistical analysis. Traditional genetic case-control studies typically test each variant individually for association with the trait of disease under study. Even in very large datasets, this strategy will not capture the effect of very rare variants, or even singletons. As an alternative, different ways of collapsing many variants across one or a set of genes into a single compound variable are being developed. As the direction and magnitude of effect for each variant is difficult to predict, such methods involve many non-trivial choices and assumptions, and the statistical framework for such wider burden tests is constantly evolving (38, 39).

Other proposed sources of missing heritability include gene-environment and gene- gene interaction that remain inherently difficult to investigate. It seems highly plausible however, that subsets of genetic variants may affect disease risk only in a specific combination or under given biological and environmental circumstances. The simplest model of polygenic disease assumes a log-additive relationship between risk alleles (multiplicative), whereas gene-gene interaction, also termed epistasis, would implicate a synergistic effect resulting in a larger combined effect than expected under the multiplicative model. The large number of potential hypotheses additionally complicates statistical testing for epistasis on a genome-wide scale (40). To date, most gene-gene interaction studies for neurodegenerative diseases have looked at candidate pairs of SNPs (41, 42), but significant signals from hypothesis-free genome-wide analyses emerge (43). Undetected interaction phenomena such as digenic inheritance may additionally underlie unresolved cases of familial disease, as recently shown for OTUD4 and RNF216 in ataxia with dementia and hypogonadotropism (44). Examples mentioned above represent phenomena of heritability that can be missed by current standard genetic approaches.

Additionally, hereditary neurodegenerative disorders are clinically and genetically heterogeneous. A rare inherited disease with a distinct phenotype may occasionally appear to be caused by mutations in one gene only (there is a small and ever shrinking group of extremely rare neurodegenerative phenotypes where currently only one gene is believed to explain a specific condition). However, as there is always the possibility of detection of novel families where a different gene causes a clinically 32

indistinguishable disease such genetic homogeneity can never be ultimately proven and non-existence cannot be claimed until everyone’s genome is sequenced and centrally analysed and even then the occurrence of newly de novo mutations developing over time cannot be ruled out. The general picture emerging from expanding clinical and genetic studies of hereditary neurologic disorders is characterised by profound clinicogenetic heterogeneity. However, the multitude of relevant genes and mutations varies across different phenotypes and ethnicities, with implications for diagnostics, genetic counselling and research. Furthermore, establishing causality in patients from different ethnicities, especially in non-Caucasians can be challenging. It is vital to refer to the ethnicity matched control databases, where available, and with growing databases and genetic studies of diverse ethnicities, this becomes achievable.

Not only do genes act on a continuum, also different phenotypes can begin to resemble each other or merge into each other as diagnostic tools sharpen and there is not always a clear line between an affected and an unaffected individual. However, traditional genetic studies have largely been focussed on disease status as a binary variable, as in linkage analysis or case-control association studies. These study designs therefore do not account for the remarkable clinical heterogeneity, reduced penetrance, variable expressivity and phenotypic variability observed in many neurodegenerative disorders. Individual patients may vary widely with respect to parameters such as age at onset, clinical subtype, progression rate and response to therapy, representing variations in disease course that are likely to be partially determined by unidentified genetic factors. Disentangling the genetic underpinnings of clinical heterogeneity is challenging, as it will require careful collection of high quality, standardised clinical data at multiple time-points in large cohorts.

As a future prospect of a more refined understanding of genotype-phenotype correlations in neurodegenerative disorders, individual genetic profiling may also prove relevant for prognosis and treatment. As a first step, data on genetic vulnerability linked to distinct symptoms or molecular pathways will help optimize patient selection in clinical trials and stratification of disease groups according to their underlying genetic make-up. As optimistic researcher, one anticipates a future scenario where a range of disease-modifying therapies targeting specific pathogenic mechanisms may become available. To establish these aims, the interpretation of individual genetic data is becoming increasingly important in health care for patients with non-neurological or neurological disorders equally, in line with the proposed ideal of personalised medicine.

Conclusions Genetics is vital to modern understanding of neurodegenerative disorders. The characterisation of Mendelian disease genes has revealed important starting points lending a molecular angle to identification of underlying disease mechanisms by stimulating hypotheses and powerful tools for further research. Additionally, large scale GWAS have identified common susceptibility variants for complex (neurodegenerative) diseases motivating further studies to refine their interpretation and unravel underlying mechanisms of conferred risk. Despite encouraging progress in the understanding of their pathogenesis, causal therapies for most neurodegenerative disorders are still lacking. The ultimate aim of disease genetics is to positively impact health care and improve the lives of patients and their families beyond the yield of a diagnosis and genetic counselling as practised widely. Rapidly evolving molecular genetic techniques, including induced pluripotent stem cell technology and genome editing, hold promise that further significant advances in neurodegenerative genetics can be achieved in the near future. In this PhD the aim was to learn some of the current approaches in next-generation sequencing and iPSC culture in order to clinically characterise neurodegenerative patients, investigate their underlying genetic architecture and develop ways to functionally assess pathogenic variance identified. The approaches employed here will continue to fertilise clinical, pathomechanistic and translational perspectives collectively in the years to come. In the following, the two principal groups of rare inherited movement disorders mainly worked on during the generation of this thesis will be introduced: Neurodegeneration with brain iron accumulation disorders, and ataxias, including genetically overlapping polyglutamine disorders. Finally, induced pluripotent stem cell technology and its usefulness for studying rare inherited neurodegenerative diseases, especially cerebellar ataxias, will be critically introduced to prepare the reader for the subsequent results chapters.

1.3 Clinicogenetics of neurodegeneration with brain iron accumulation

Introductory remarks This introductory subchapter prepares the reader for results chapter 3 and parts of results chapter 4. Modified parts of this introduction subchapter were published in the Movement Disorders Genetics book (45) and have been submitted as another short book chapter about NBIA to ‘Kovacs and Alafuzoff: Neuropathology (Handbook of Clinical Neurology)’.

Neurodegeneration with brain iron accumulation, also referred to as pallidopyramidal disorders, includes several clinical and genetic entities. Increased regional brain iron stores, especially in the basal ganglia, are often seen as the unifying hallmark (e.g. for review (46, 47)) even though the occurrence or iron deposits in the course of disease can be highly variable or even undetectable or absent in certain genetically-confirmed NBIA disorders (48-51). Clinically, a combination of severe and progressive dystonia, spasticity, parkinsonism and dysarthria are common presenting features. Phenotypic variation in terms of manifestation and progression can be vast and challenging in clinical practice, but prominent iron accumulation detectable on magnetic resonance imaging (MRI) remains – despite of its caveats – an important red flag in the diagnostic workup (52, 53).

At least ten different genes with causal implication in NBIA and responsible for subforms of the disease have been proposed and will be discussed in more detail in separate sections below. Certain clinical presentations are enriched in specific genetic subforms of NBIA (54, 55) and genetic tests usually make for the final confirmation of the disease in most of the cases where neuropathology is unavailable during lifetime. Figure 1-6 gives an overview of important imaging characteristics of the four most prevalent types of NBIA disorders. Two recent reviews propose detailed strategies on an updated clinicoradiological approach to diagnosis in NBIA (56, 57) for the interested reader.

Figure 1-6: Imaging characteristics NBIA

Figure and figure legend adapted and reproduced from (58): Imaging characteristics of the four major subtypes of NBIA. A: T2-weighted MRI in PKAN shows GP hypointensity indicating iron accumulation (white arrow) with an area of central hyperintensity (red arrowhead) known as the “eye of the tiger” sign. B: Substantia nigra (SN) in the same patient showing hypointensity in the medial aspect of the nucleus. C: T2-weighted sequence in a patient with MPAN showing pallidal hypointensity with hyperintense streaking in the region of the medial medullary lamina. D: Iron accumulation in the SN. E: T2 sequence of GP and (F) SN in a 9 year-old child with PLAN showing iron accumulation. Imaging performed earlier in the disease course had shown no signal changes. Inset (F) shows cerebellar atrophy in the same child. G: T2 imaging showing the GP in a young adult with BPAN, after the onset of parkinsonian symptoms. In (H) note the marked hypointensity in the SN, and, in the inset in (H), the same region on T1-weighted sequence showing the characteristic hyperintense “halo” thought to represent neuromelanin release from degenerating neurons. All images performed on 3.0T magnet except (G) and (H) which were performed on 1.5T. Abbreviations: PKAN: pantothenate kinase-associated neurodegeneration; GP: globus pallidus; SN: substantia nigra; MPAN: mitochondrial membrane protein-associated neurodegeneration; PLAN: phospholipase A2-associated neurodegeneration; BPAN: beta-propeller protein-associated neurodegeneration; T: tesla.

Neuropathological findings in NBIA most commonly include axonal spheroids in predominantly central nerves in addition to variable quantities of iron accumulation in the basal ganglia (59, 60). For detailed neuropathology of the most frequent NBIA subtype, Pantothenate kinase-associated neurodegeneration (PKAN), see Figure 1-7.

Figure 1-7: Neuropathology of PKAN, the most frequent NBIA disorder

In the globus pallidus axonal swellings are readily apparent (A – E) consisting of large eosinophilic swellings with a granular appearance (A, arrows, haematoxylin and eosin) in addition to smaller swellings with homogeneous appearance (A, arrowhead). Frequent widespread iron deposition within large axonal swellings (B, arrow, Perls stain). Ubiquitin immunohistochemistry identifies axonal swellings (C) where small axonal swellings contain phosphorylated neurofilaments (D, arrows, SMI31 immunohistochemistry) and express amyloid precursor protein (E, arrow) absent in large swellings (E, arrowhead). Tau-positive neurofibrillary tangle pathology varies in severity between cases (F, frontal cortex, AT8 immunohistochemistry). Bar in A represents 50 µm in A, C and F; 25 µm in B, D and E. Image courtesy of Prof. J. Holton, T. Revesz, and Dr. Abi Li, Queen Square Brain Bank, UCL; also published in (61).

In compliance with its clinical heterogeneity, post-mortem alpha-synuclein and Lewy body (LB) pathology (59, 62, 63) as well as TDP-43 and tau pathology (64-67) have additionally been observed for some subtypes of NBIA (summary and overview, see Table 1-1). 37

Table 1-1: Overview of NBIA conditions

Condition Synonym Neuropathologic Gene MI Age at onset (Acronym) (s) inclusion type* Pantothenate kinase-associated Mostly NBIA1 PANK2 AR Tau or none neurodegeneration childhood (PKAN) Phospholipase A2- Alpha- associated synucleinopathy neurodegeneration NBIA2, Childhood to with cortically PLA2G6 AR (PLAN) PARK14 adulthood widespread Lewy bodies, additional tau pathology Mitochondrial Alpha- membrane protein- synucleinopathy associated NBIA4, Mostly with cortically C19orf12 AR neurodegeneration SPG43 childhood widespread Lewy (MPAN) bodies, additional tau pathology Beta-propeller NBIA5, protein-associated Adolescence to formerly: WDR45 XLD Tau neurodegeneration adulthood SENDA (BPAN) Fatty acid hydroxylase- Mostly associated SPG35 FA2H AR NA childhood neurodegeneration (FAHN) Kufor-Rakeb Not known yet - disease (KRS) clinically and genetically PARK9 ATP13A2 AR Adolescence significant overlap with lipofuscinoses Neuroferritinopathy NBIA3, Adolescence to FTL AD Tau (NF) NF adulthood Aceruloplasminemia No alpha- (aCP) synucleinopathy, Adolescence to aCP CP AR primarily adulthood astrocytic pathology Coenzyme A synthase- associated NBIA6 COASY AR Childhood NA neurodegeneration (CoPAN) Woodhouse-Sakati Adolescence to WSS DCAF17 AR NA syndrome (WSS) adulthood

Overview of NBIA conditions, genes, MRI findings and neuropathology. Abbreviations: MI=mode of inheritance; NA=not available; SENDA= static encephalopathy of childhood with neurodegeneration in adulthood; AD=autosomal dominant; AR=autosomal recessive; XLD=X-linked dominant. *BG-iron not specifically stated. A slightly modified table has been submitted to ‘Kovacs and Alafuzoff: Neuropathology (Handbook of Clinical Neurology)’.

Despite ongoing excitement and progress in elucidating underlying genes and molecular pathophysiology, treatment of NBIA disorders up to now remains largely symptomatic. As inherited conditions, NBIA disorders can be inherited in autosomal dominant (AD), autosomal recessive (AR) and X-linked (XL/XLD) fashion. The genetic underpinnings of the disease, which are in the focus of this section, are as heterogeneous as its clinical phenotypes (68).

Broadly speaking, one group of NBIA disorders can be caused by mutations in genes which have a direct link to iron metabolism. The two examples here are mutations in the ferroxidase encoding gene ceruloplasmin (CP) which lead to the multisystem disease aceruloplasminemia (69, 70), or pathogenic mutations in the ferritin light polypeptide (FTL) gene that lead to adult onset hereditary ferritinopathies with the gene encoding the light subunit of the ferritin protein responsible for iron release and storage (71-74). However, most of the disease associated mutations in NBIA impair genes in phospho- and sphingolipid metabolic pathways (75), mitochondrial or lysosomal activity (76, 77) and pathomechanisms linking those to iron metabolism are less evident (also see Figure 1-12 later). Finally, the ultimate hope in studying NBIA disorders is to use this presence of overlapping clinical features and biological pathways in order to identify new therapeutic alleys, not only for these rare disorders, but also for more common disorders as PD, AD and forms of genetic dystonia that share important elements of pathophysiology.

As an introduction into the field this subchapter hereby gives an overview about the distinct genetic defects that lead to the different, yet overlapping phenotypes of NBIA disorders. This introduction of the NBIA disorders highlights important phenotypic presentations that direct genetic testing where possible, provides information about treatment options when present and discusses remaining controversies of this highly dynamic field of neurodegeneration.

Pantothenate kinase-associated neurodegeneration PKAN (MIM #234200), also recognized as NBIA1, accounts for about 35-50% of NBIA cases (58, 78). As always when studying rare diseases, there is a possibility of overestimation depending on the cases/populations screened and possibly confounding selection bias that is hard to avoid. PKAN has an estimated prevalence of approximately 1:1000000 world-wide with regionally higher prevalences due to founder effects (e.g. Dominican Republic). Together with other genetically unconfirmed NBIA disorders it was formerly classified under the name Hallervorden-Spatz syndrome (HSS) – a terminology widely abandoned due to the unethical implications of these

39 researchers under the Nazi-regime, but still to be found in the literature sporadically for historic reasons (79). Clinically, PKAN can be classified into two subforms, typical and atypical PKAN. Typical PKAN patients show an early-onset and usually become symptomatic with gait and postural abnormalities around the age of 3-4 (78, 80). Genetically, their mutational spectrum includes variations in PANK2 leading to protein truncation more often than missense mutations that seem to be more frequent in atypical, late-onset PKAN cases (78). Typical PKAN patients present with a more rapid and non-linear progression to the full clinical spectrum of severe (in its early manifestation predominantly oromandibular and non-axial) dystonia (78, 81), parkinsonism, spasticity and hyperreflexia, chorea, impaired saccadic pursuits as well as decelerated and hypometric vertical saccades, mental decline and behavioural abnormalities (78, 82). Pigmentary retinopathy can be present in up to 40% of typical PKAN cases whereas optic atrophy appears to be infrequent (78, 83). Given the disease progression with rapid deterioration intertwined with phases of stability, the majority of patients are non- ambulatory within approximately 15 years into the disease (78).

Atypical PKAN cases tend to be significantly older when the first obvious symptoms appear, usually with an age at onset at 13.7±5.9 years [range: 1 to 28] in atypical vs. 3.4±3.0 years [range: 0.5 to 12] in typical cases as reported in the original paper by Hayflick et al. (78). Pyramidal and extrapyramidal manifestations tend to be generally milder and progress less rapidly, however most atypical PKAN patients eventually lose their ambulatory independence in the progress of the disease (78). Furthermore, in atypical cases speech defects like dysarthria or palilalia as well as psychiatric features seem to be reported more frequently (78).

In terms of radiographic findings, the eye of the tiger sign, a T2-weighted hypointense signal in the globus pallidus with a central region of hyperintensity (see Figure 1-6, A) remains a red flag for PKAN, but has lost its previously acclaimed pathognomonic status (78, 84). MRIs from patients with other genetic NBIA subforms or with a different neurodegenerative disease have been found to show this sign (53, 85) as well as PANK2-positive patients radiographically lacking it (86). The eye of the tiger sign can even be present in individuals without clinical symptoms (87). Despite these reports it remains an important flag in clinical diagnostics pointing towards the group of NBIA disorders.

Neuropathologically, PKAN forms part of the neuroaxonal dystrophy spectrum. Two main classes of spheroids are commonly observed: large, granular bodies representing

40 degenerating neurons and smaller eosinophilic spheroid-like structures indicative of dystrophic neurites (52), also see Figure 1-7 for more details. Prior to gene identification numerous cases of heterogeneous and mostly unknown aetiology under the former umbrella term “Hallervorden-Spatz syndrome” have been neuropathologically investigated. Lewy body pathology was a predominant finding (62, 88-90). However, given the undetermined underlying genetic cause the interpretation of this histopathological association is uncertain and less trustworthy. These initial reports have been genetically unconfirmed cases that are likely to be PANK2-negative with possibly heterogeneous, but finally undetermined underlying genetic defects. However, a recent neuropathological investigation strictly including only PANK2- positive patients observed no alpha-synuclein positivity, but some tau-positive inclusions and concluded that PKAN is not a synucleinopathy (65) which neuropathologically distinguishes it from other NBIA disorders where alpha-synuclein pathology can occur (e.g. Phospholipase A2- or PLA2G6-associated neurodegeneration (PLAN) and mitochondrial membrane protein-associated neurodegeneration (MPAN)).

As mentioned above, the human gene associated with PKAN is PANK2. PANK2 is located on the short arm (p13) of chromosome 20 (91). It encodes the only human protein from the pantothenate-family that is thought to be expressed in mitochondria and seems responsible for the intramitochondrial de novo synthesis of Coenzyme A (CoA) (91). Specifically, pantothenate kinase is the key regulatory enzyme in Coenzyme A synthesis and catalyses the first step from pantothenic acid/vitamin B5 to 4’phosphopantothenic acid in this universally vital pathway. Given the role of CoA in fatty acid metabolism, amino acid synthesis and the Krebs cycle, CoA is substantial for the basic key metabolism of the cell (92).

Several animal models of PKAN have been generated and have shown to recapitulate different aspects of the human disease with varying degree. One possible reason for the absence of clear neurodegenerative phenotypes in these models might be that other PANK isoforms (1, 3, and 4) exist and are able to compensate for PANK2 loss in these species. Selected examples include PANK2 null mice with retinal dysfunction and infertility which did not show gait abnormalities, brain iron accumulation or obvious basal ganglia pathology despite detected altered mitochondrial membrane potential in neurons from these mice (93, 94) and with sudden precipitous death when deprived of dietary pantothenic acid (95). Interestingly, pantothenic acid deprivation in control mice resulted in azoospermia (as had been shown to occur in the PANK2 null mice without pantothenic acid deprivation) and an additional movement disorder, confirming the

41 unspecific importance of pantothenic acid pathways in healthy as well as PANK2 null mice, illustrating the usefulness equally as the limitations of this mouse model (95). Further supporting these findings, Brunetti et al. showed induction of a PKAN-like phenotype with severe motor dysfunction, abnormal mitochondrial morphology and neurodegeneration by introducing a ketogenic diet to these mice with a rescue of the phenotype upon pantethine treatment (96). Several PKAN drosophila and zebrafish models have been generated confirming and extending findings of abnormal mitochondrial function, reduced life span, increased oxidative damage, and an important emerging role of PANK2 during neurogenesis and central nervous system (CNS) development (97-100). This last aspect is particularly interesting, since neurodegenerative diseases classically often are presented to be distinct of so called developmental disorders. I believe these categorical dogmata will be proven more and more limited and overlaps of contributing developmental and degenerative defects due to disease causing mutations will become an accepted paradigm as more data emerges. Given all known limitations introduced by usage of animal models, the first exciting human based (non post-mortem) data is presented on human fibroblast and human induced neurons carrying PANK2 mutations confirming previous findings such as altered oxidative status, reduced antioxidant defence, impaired cytosolic and mitochondrial aconitase activities, increased mitochondrial labile iron pool, increased mitochondrial reactive oxygen species (ROS) production, decreased mitochondrial membrane potential and ATP production and altered mitochondrial shape in human fibroblasts with defects partially mirrored in the induced neurons (101). Looking back, the limitations of animal models and the lack of human neuronal cell models that faithfully recapitulate disease phenotypes in vitro has been one of the major hurdles in the efforts to study the mechanisms of PKAN and other (NBIA) disorders effectively. However, the recent discovery and rapid spread of iPSC technology may yield new important insights into NBIA pathophysiology within the upcoming years.

PKAN is inherited in autosomal recessive manner. The causal mutational spectrum on PANK2 includes mostly missense and nonsense mutations, which - depending on the location of the mutation, the corresponding evolutionary conservation and functional consequence - result in reduced or abolished activity of pantothenate kinase in mitochondria. Homozygous and compound heterozygous mutations have been reported. Additionally, deletions up to 6 kilobases, whole exon deletions, splice site mutations and duplications have been observed and all 7 exons of the gene can be affected (102). A PANK2 screening study in 2006 enrolled 72 patients with a clinical diagnosis of NBIA and identified both mutated PANK2-alleles in 48 patients. In total, 33 different mutations – mostly missense and nonsense - were identified, and deletions

42 were only found in rare cases (around 4% of the mutated PANK2-alleles) (102). The most frequent mutations in this study were c.1583C>T/p.T528M (n = 11), c.573delC/p.S191RfsX13 (n = 10) and c.1561G>A/p.G521R (n = 10). Over the coming years, authors used different transcripts to refer to the same (common) mutations in PANK2, and in effect c.1561G>A is found to be equal to c.1231G>A and c.1583C>T to c.1253C>T and all four annotations exist variably in the literature. In order to reduce confusion, the amino acid change only will be mentioned in the following although even a certain amount of different annotations for the protein consequences can be found throughout the literature. With increasing evidence, p.G521R (allele-frequency ~ 25%), p.T528M (allele-frequency ~ 8%) and p.R451X (allele-frequency ~3%) seem to be the most common pathogenic variants (103, 104), whilst the residual spectrum of PANK2 mutations includes mainly private mutations (78) – meaning that they are only found in one family (or sometimes a very small population). Two thirds of PANK2-positive patients are compound heterozygotes (104). As per example, one of the most common point mutations, p.G521R, has been shown to result in reduced production of the mature isoform of the PANK2 enzyme (92) and loss-of-function seems to be the assumed predominant pathomechanism. Regarding the molecular pathology of PKAN it has therefore been suggested that mutations in PANK2 lead to reduced PANK2- protein levels, impaired catalytic activity of PANK2 and consecutively altered neuronal mitochondrial lipid formation and metabolism (92). However, it has to be noted that not all disease associated point mutations result in significantly reduced catalytic activity of the enzyme (92). The final transformation of molecular dysfunction into clinical symptoms still remains equally elusive and only few attempts have been made to combine the mutational location and corresponding clinical features: As one example, Hartig et al. introduced an activity score to take into account the severity of the identified mutations and the predicted rest activity of the enzyme. They found the activity score of the mutations to be correlated with the age at onset of the disease, but not the loss of ambulation in their PANK2-positive cases (102). In a different attempt, a correlation between the location of missense mutations and their biochemical properties with the disease features found mutations located in the dimer interface, the ATP-binding-site and the interior of the protein to negatively impact the enzyme activity and to be mostly associated with classic, young-onset and rapidly progressing disease. Mutations located on the surface that lead to slight protein instability, but seemed to preserve most of the enzyme’s catalytic activity, were associated with milder symptom- presentation of PKAN (105). Similar to other rare inherited neurodegenerative diseases genotype-phenotype correlations in PKAN remain controversial and difficult. However, it is these arduous and small steps to try and correlate the specific location of a mutation with functional and phenotypic properties that will shed further light on the gap

43 between the genetic underlying causes and their transformation into clinical manifestations and will ultimately facilitate genetic counselling. One important restriction in the field of extremely rare conditions will always be the inherently small amount of cases available and phenotypic heterogeneity even within families sharing the same underlying genetic defect and much of the genetic background information. These observations point towards unidentified environmental and genetic modifiers of disease and are an important subject of study to be addressed more thoroughly throughout all genetic diseases in the upcoming and present eras of research.

In summary, with the current state of knowledge and despite the remaining gaps, some conclusions regarding PKAN can be retained. As shown in the original work by Hayflick and colleagues, patients homozygous for the common p.G521R amino acid change have classic disease (78). Furthermore, biallelic null mutations seem to result in rapidly progressive, early-onset PKAN, whilst missense mutations with diminished, but preserved enzyme function are likely to lead to atypical PKAN with later-onset and slower progression. Several compound heterozygous cases with atypical PKAN have been seen at the Institute of Neurology (IoN) and the National Hospital for Neurology and Neurosurgery (NHNN), Queen Square, over the years: Late disease onset in these cases goes along with rather preserved cognition, slow progression, longer life span and longer preserved ambulation (personal observations).

Furthermore, PKAN has been shown to be allelic to HARP syndrome, including hypoprebetalipoproteinemia, acanthocytosis, retinitis pigmentosa, and pallidal degeneration (106) which is now considered to be part of the PKAN continuum. PKAN is the first NBIA disorder where iron-mediating medical treatment has systematically been tested (e.g. see Clinicaltrials.gov identifier NTC00907283) and whilst treatment with the iron chelator deferiprone led to a median reduction of 30% of pallidal iron content, clinical benefits unfortunately remained non-significant in a pilot study (107). The four years follow-up data including 5 patients with genetically confirmed PKAN however suggests clinical stabilisation upon treatment in 5 out of 6 patients (108). Furthermore, PKAN is the only NBIA subgroup where preliminary data about the use of deep brain stimulation (DBS) of the bilateral globi pallidi internii in the treatment of dystonia is available (109, 110). DBS treatment seems to show some degree of clinical benefit in some cases for the postoperational time observed, but does not stop the ongoing neuronal degeneration.

PLA2G6-associated neurodegeneration PLA2G6-associated neurodegeneration accounts for roughly 20% of NBIA disorders and thereby for the largest proportion after PKAN (58, 111). Three main distinct phenotypes have been described within the spectrum of PLA2G6-associated disease which represents a striking example of allelic heterogeneity: classic infantile neuroaxonal dystrophy (INAD) (112), childhood-onset atypical neuroaxonal dystrophy (NAD) (63, 113) and adult-onset dystonia-parkinsonism (50). The main unifying characteristics are progressive cognitive decline and rapid deterioration of motor skills, but the three syndromes can be distinguished via the following: -INAD (MIM #256600) shows a disease onset between 6-36 months and usually manifests with initial hypotonia, progressive psychomotor delay and developmental regression, cerebellar features, bulbar dysfunction and progressive spastic tetraparesis (113-116). MRI findings include cerebellar atrophy (see Figure 1-6, F and inset F) and electromyography typically shows denervation atrophy (117, 118). Additionally, fast rhythms on electroencephalogram (EEG), seizures and optic atrophy can be present. Prior to gene identification, the definitive diagnosis was confirmed on sural biopsy, depicting typical dystrophic axons in the peripheral nerves (115). Most INAD patients die within the first decade of life, isolated evidence reports some affected patients surviving up to their early twenties (103). -Atypical NAD (MIM #610217) is less frequent than INAD and usually becomes apparent between year one and eight of life, whilst a delayed onset in the late twenties has sporadically been reported (116). The disease course is less aggressive with a slower progression rate compared to INAD. Prominent symptoms are dystonia, spastic tetraparesis, cerebellar ataxia, speech delay and behavioural abnormalities ranging from impulsivity, poor attention span, and emotional lability to autistic behaviour (63, 113). Seizures, optic atrophy

and nystagmus seem to occur in NAD as frequently as in INAD whilst fast rhythms, truncal hypotonia and strabism are predominantly part of INAD (63). -PLA2G6-related adult-onset dystonia-parkinsonism (MIM #612953) represents the third phenotype and is known for a later (second to third decade), subacute onset of dystonia-parkinsonism with rapid decline and initial dramatic response to levodopa. Dystonia, dysarthria, psychiatric features, cognitive decline, oculomotor abnormalities, pyramidal involvement and gait abnormalities are further common clinical characteristics. Brain iron deposits could not be detected on MR-imaging of these patients, whereas cerebral and cerebellar atrophy were prominent features (50).

PLAN is inherited in autosomal recessive manner and caused by mutations in the widely expressed gene PLA2G6 (119). PLA2G6 is located on the long arm of chromosome 22 and entails 17 exons which undergo alternative splicing resulting in the occurrence of multiple different isoforms (119). The disease causing gene was discovered in 2006 following a linkage approach in families with INAD and NBIA initially (112) and was subsequently found to be mutated in the related Karak-syndrome (120) which is now recognized to be part of atypical NAD and the PLAN spectrum. Prior to gene discovery the diagnosis was solely achieved clinically or histopathologically and confirming the diagnosis via molecular genetic testing has since proven useful to reduce invasive tissue biopsies and alleviate genetic counselling.

Histopathologically, the hallmark finding in PLAN is the presence of neuroaxonal spheroids thought to represent the degeneration of the inner mitochondrial membrane (121, 122). This neuroaxonal degeneration with distended axons (spheroid bodies) and axonal swellings is mainly present in the CNS, but has been reported in peripheral nerves as well (115). Furthermore, Lewy body pathology alongside tau-positive inclusions has been observed (123). A second common finding is progressive excess accumulation of brain iron that can be detected neuropathologically as well as neuroradiologically as the disease progresses. For illustration of PLAN neuropathology, see Figure 1-8.

Figure 1-8: Neuropathology of PLAN

Histological appearances in PLAN. In the globus pallidus (A – C) axonal swellings are numerous and are visible in haematoxylin and eosin stained preparations (A). These may be highlighted by ubiquitin immunohistochemistry (B) and staining for phosphorylated neurofilaments (C, SMI32). Iron deposition is evident on routine staining (A, arrow). Alpha-synuclein immunoreactive Lewy bodies and neurites may be widespread (D, nucleus basalis of Meynert) and there may also be tau-positive neurofibrillary tangles and neuropil threads (E, nucleus basalis of Meynert). Bar in A represents 50 µm in A, B and E; 25 µm in C and D. Image courtesy of Prof. J. Holton, T. Revesz, and Dr. Abi Li, Queen Square Brain Bank, UCL, also published in (61).

Initially, typical iron deposits in the globus pallidus can be missing on brain imaging of INAD cases, but usually these develop with progress of the disease course (112, 117). Further common neuroradiographic features in PLAN are cerebellar hypoplasia or atrophy and T2-weighted high signal in the cerebellum, hypoplastic optic tracts/chiasm and an elongated, vertically orientated splenium (117). 47

In the initial discovery study, the majority of 44 identified unique mutations were homozygous missense mutations (33), followed by frameshift-inducing deletions (5), nonsense mutations (3), deletions without subsequent frameshift (1), a splice site mutation and a large deletion. As in PKAN, most mutations are private mutations. This frequency spectrum has since been reproduced in a slightly larger cohort (63), the occurrence of copy number variants reported (124) and various single reports have been added summing up to more than seventy disease causing mutations reported so far. Obvious genotype-phenotype correlations are missing, however homozygous null mutations were mostly found in patients with INAD, and not in individuals of the broader clinical category with iron dyshomeostasis without axonal spheroids (112) so far. This suggests that mutations resulting in complete absence of protein seem to be associated with more severe phenotypes. Comparable to PANK2, compound heterozygous missense mutations in PLA2G6, potentially suggestive of remaining protein function, correlate with the atypical, milder NAD phenotype (63). Additionally, in contrast to mutations reported in INAD cases, the observed mutations in PLA2G6- related dystonia-parkinsonism do not impair the catalytic activity of the PLA2G6 enzyme (125).

The 85 kDa calcium-independent group 6 phospholipase A2 (iPLA2-VI) protein encoded by the PLA2G6 gene is one of several calcium independent phospholipases and is an active enzyme in its tetrameric state. It plays an important role in the maintenance of cell membrane homeostasis, regulation of apoptosis, leukotriene and prostaglandine synthesis, catalysis of glycerophospholipid hydrolysis and in phospholipid remodelling (50, 63, 103, 111, 119, 126, 127). Multiple isoforms are encoded by the PLA2G6 transcript variants, whereof two are enzymatically active (119). A subset of the reported mutations in the literature was found to alter the shorter enzymatically inactive isoforms and act as dominant-negative inhibitors when incorporated into the tetramer iPLA2-VI (119, 126). Therefore, defects in iPLA2-VI result in impaired phospholipid-remodelling representing an essential component of physiological membrane repair and homeostasis (128). Additionally, iPLA2-VI can initiate apoptosis when exposed to oxidative stress (129), and dysregulation of these vital cell protection and repair mechanisms result in neuronal damage and degeneration. Several PLAN mouse models (knock-out and knock-in) have been generated and findings seem to reflect human post-mortem findings with the degeneration of the inner mitochondrial membrane increasing along the course of the axon as an important event in pathology; for a recent review see (130). The results of potential iPSC models

48 of this orphan disease still await publication at this point in time. The two most common NBIA subforms PKAN and PLAN share important pathophysiological pathways and these disorders are molecularly connected via Acetyl-CoA and lipid metabolism. For an overview, see Figure 1-9 and the recent review of Kurian and Hayflick about this topic (131).

Figure 1-9: Biochemical pathways and cellular processes in PKAN and PLAN

PLA2G6 and PANK2 enzymes: biochemical pathway and implicated cellular processes in PLAN and PKAN. Figure reproduced from (131).

Mitochondrial membrane protein-associated neurodegeneration Mitochondrial membrane protein-associated neurodegeneration (MIM #614298), also termed “neurodegeneration associated with C19orf12 mutations”, is the third most frequent group after PKAN and PLAN and accounts for roughly 6-10% of NBIA disorders (58, 111, 132, 133). Disease onset usually spans from childhood to early adulthood and common features include speech and gait difficulties due to spasticity and dysarthria, cognitive decline resulting in dementia, optic atrophy, dystonia- parkinsonism, psychiatric features and motor neuronopathy with early signs of upper motor neuron dysfunction, followed by affection of the second motor neuron later. The disease is slowly progressive, and survival into adulthood is observed frequently. However, MPAN is more than other NBIA disorders associated with rapid progressive cognitive decline (111, 132, 133). 49

Post-mortem brain histopathology in two published cases showed extensive neuronal loss, iron deposits within neurons, astrocytes and macrophages of the globus pallidus (GP) and substantia nigra (SN), Lewy neurites in the globus pallidus, Lewy bodies, axonal spheroids, neurofibrillatory tangles and tau-positive inclusions. The pyramidal tracts of the spinal cord and optic nerve displayed severe loss of myelin in one of them (132, 133).

Neuroimaging shows excess iron accumulation in the globus pallidus and substantia nigra as well as cortical and cerebellar atrophy. Hyperintense streaking of the medial medullary lamina between the globus pallidus interna and externa on T2-weighted images can be additionally present and is sometimes mistaken for an eye of the tiger sign (133), see Figure 1-6, C and D.

MPAN is caused by mutations in the orphan gene C19orf12 and is inherited in autosomal recessive fashion. A combined approach of homozygosity mapping followed by candidate gene sequencing identified the initial common founder mutation c.204_214del11, p.G69RfsX10 in C19orf12 (NM_001031726.3/NP_001026896.2) in a Polish family in 2009 (132). Subsequent screening in a Polish cohort of additional 23 index patients matching the clinical phenotype described above and negative for mutations in PANK2, PLA2G6, FTL and CP revealed biallelic mutations in C19orf12 in 18 of those. Twelve patients carried the homozygous founder mutation, thought to have arisen in a common founder at least 50 to 100 generations ago. Three remaining cases carried the deletion in combination with a different missense mutation each (p.T11M, p.G53R or p.G65Q) in the compound heterozygous state. One further case carried two missense mutations, p.G69R and p.K142Q, and two cases with homozygous missense mutations (p.T11M and p.G69R) were identified. One additional patient carrying compound heterozygous missense mutations (p.G69R and p.K142Q) in C19orf12 was identified who displayed a late-onset (49 years) predominantly parkinsonian phenotype. The same combination of mutations was identified in one young Polish patient with a mild impairment of fine motor skills and therefore led the authors to speculate that this mutation combination could be resulting in milder phenotypes. Most evidence in C19orf12-pathogenesis has been collected regarding the comon founder mutation c.204_214del11, p.G69RfsX10. This homozygous 11-bp deletion results in a frameshift that introduces a premature stop codon, and is predicted to cause early truncation of the protein. Even though the deletion did not modify C19orf12-mRNA levels in blood, the protein was undetectable via immunoblot analysis in patients’ fibroblasts (132). The missense mutation p.K142Q changes a highly

50 conserved positive lysine to a negatively charged glutamate. Three different missense mutations p.G53R, p.G65Q, and p.G69R replace highly conserved glycines in the transmembrane region of C19orf12 with a charged amino acid and therefore are likely to be diminishing its function. A follow-up screening study on a larger and ethnically more diverse cohort of 161 idiopathic NBIA cases was published in 2013. It identified two mutated alleles in 23 cases and established MPAN as the third most common subform of NBIA to date. In addition to the common founder mutation, a variety of different, mostly private mutations (nonsense, frameshift and missense) were identified throughout the C19orf12 gene without evidence for large deletions and duplications. However, the occurrence of a single heterozygous mutant allele in three cases suggests the presence of occult mutations in intronic and/or regulatory sequences (133) or undetected deletions and duplications. Furthermore, in their study, one single mutation case had a family history compatible with autosomal dominant inheritance backed up by histological findings in the deceased father of the family. The identified single unique frameshift mutation induces a series of 32 amino acid substitutions and ultimately results in a premature stop codon 2 amino acids away from the physiological termination. Therefore, the authors discuss the possibility of this aberrant protein product being unamenable to physiological nonsense-mediated mRNA decay, and thereby presumably exerting dominant negative effects on the normal protein. However, further studies are awaited to elucidate the pathogenesis of this single mutation and finally clarify the possibility of additional autosomal dominant inheritance in C19orf12 (133).

The exact gene function of C19orf12 is still unknown, but expression analysis suggests it has mitochondrial location (132). However, analysis of respiratory chain complexes and mitochondrial morphology in fibroblasts of C19orf12-positive patients could not detect any significant abnormalities. Analysis of C19orf12-coregulated genes in whole blood interestingly revealed strongest coregulation with genes involved in valine, leucine and isoleucine degradation as well as fatty acid biogenesis (132). These processes are related to Coenzyme A metabolism in mitochondria and therefore suggest a link between MPAN and other NBIA disorders with impairment in CoA- pathways and lipid homeostasis (PKAN, PLAN and COASY protein-associated neurodegeneration (CoPAN) (see below)). With roughly 70 patients reported in the literature so far, more is known about clinicogenetic characteristics than its molecular underpinnings with addition of a few successful animal models only. However, a recent drosophila model based on double heterozygous deletion of the two drosophila orthologues of C19orf12, CG3740 and CG11671, seems to recapitulate parts of a

51 neurodegenerative phenotype with vacuole formation in the brain, bang sensitivity, reduced life span and uncoordinated climbing behaviour. It could be used in future functional studies and subsequent screening of therapeutic targets once the molecular cascade initiating the observed dysfunction becomes clearer (134). Results from potential iPSC model studies for this orphan disease still await publication at the time of writing.

Beta-propeller protein-associated neurodegeneration Beta-propeller protein-associated neurodegeneration (BPAN, MIM #300894) was previously known as SENDA syndrome (static encephalopathy of childhood with neurodegeneration in adulthood) before gene discovery in 2012 and 2013 (76, 135). It accounts for about 7% of NBIA disorders based on its prevalence in the International Registry for NBIA and Related Disorders and the North American database from the Hayflick laboratory (58, 111).

Affected cases show a distinct phenotype of psychomotor developmental delay (mental retardation and early-onset spasticity of the lower limbs) in early childhood remaining fairly stable until a sudden onset and rapidly progressive complex of symptoms including dystonia, parkinsonism and spasticity plus dementia occurs in adulthood. Eye movement disorder, sleep disturbances, dysautonomia, seizures and frontal release signs can additionally be present. Parkinsonian features usually show good response to L-dopa even though early motor fluctuations and dyskinesias often develop (46, 136). Preliminary brain pathology showed widespread tangles and threads in addition to excess iron deposition, gliosis, axonal swellings and severe neuronal loss in the substantia nigra, identified in a single case (137), and clearly more BPAN patients will need to be autopsied to complement these observations.

Brain MRI shows a specific substantia nigra hyperintensity with central hypointensity on T1-weighted MRI, iron accumulation in GP and SN, and mild cerebellar atrophy along with white matter (WM) changes (57). An interesting recent report found persistently elevated neuron specific enolase in serum (NSE) and cerebrospinal fluid (CSF) of a BPAN case and suggested these to be added as additional diagnostic cues for BPAN (138).

Only three to four years ago, exome sequencing identified the underlying previously unknown genetic cause of this disease: De novo mutations in the beta-propeller- scaffold-protein WDR45 on chromosome Xp11.23 were reported in affected SENDA cases establishing BPAN as the first X-linked-dominant NBIA disorder (76, 135). All

52 reported variants had unique occurrence in the index patient of the sequenced families suggesting de novo events.

Loss-of-function mutations account for the majority of so far reported variants, followed by missense mutations in highly conserved residues (76, 136). The majority of reported WDR45-positive cases are females (135, 136), whilst phenotypes of affected males do not substantially differ from females (76) which might be due to somatic mosaicism in the few surviving males, somatic and or germline mutations and skewing X- chromosome inactivation in females.

In the last years, mutations in WDR45 have additionally been found to be implicated in Rett-like-syndrome, familial brain calcification, epileptic spasms, mild cognitive delay, and autistic regression, and mutations seem to be enriched in patients with young- onset parkinsonism and intellectual disability extending to the Japanese population (139-143).

Hoffjan et al. give a recent and fairly exhaustive collection of WDR45 mutations reported in the literature with their associated phenotypes, see Figure 1-10 (139). These findings are another classical example of high clinicogenetic and phenotypic variability in WDR45-associated disorders that is constantly evolving, as has been observed for many other inherited neurodegenerative conditions.

Figure 1-10: WDR45: genetic architecture and reported mutations

Localisation and structure of the WDR45 gene with the mutations described for different phenotypes in the literature as well as the study by Hoffjan et al. Colour/symbol code: bold: Rett (-like) syndrome; °: NBIA (SENDA subtype); §: moderate to severe intellectual disability; #: mild cognitive delay; +: epileptic spasms; ǂ: originally described as c.342- 2 A > C; box: novel mutation from the study by Hoffjan et al. Figure reproduced from (139).

WDR45 encodes WD-repeat-domain 45, one of the mammalian homologues of the autophagy related gene Atg18 in yeast. It is thought to be a human core autophagy gene important in the early autophagy pathw ay (135, 144, 145). In an experiment involving human cells, WDR45-positive patient-derived lymphoblastoid cell lines showed accumulation of aberrant early autophagic bodies and decreased autophagic activity (135). Furthermore, first data from a CNS-specific WDR45 knock-out mouse model has been reported with the mice showing greatly impaired learning and memory, poor motor coordination and extensive axon swellings with wide-ranging axonal spheroids and defective autophagy pathways which partly recapitulate the human phenotype and could become a useful tool for further study (146).

Together with the original findings (76) this recent data provides further evidence for a strong connection between neurodegeneration and genetically defective autophagy. In

54 addition to this, iPSC-based models of BPAN are awaited to contribute a human neuronal model to the dynamic field and to hopefully recapitulate important features of BPAN.

Fatty acid hydroxylase-associated neurodegeneration Fatty acid hydroxylase-associated neurodegeneration (FAHN, MIM #612319) is another rare subform of NBIA which accounts for about 1% of the NBIA spectrum (58). It comprises dysmyelinating leukodystrophy and spastic paraparesis with or without dystonia/Spastic Paraplegia 35 (SPG35) (147, 148), which had been recognized as distinct clinical entities previously but have now been included into the spectrum of FAHN (149, 150). Clinically, FAHN is characterised by a pyramidal-extrapyramidal movement disorder with predominant gait difficulties due to severe spasticity, ataxic and dystonic features together with progressive intellectual impairment and a likely occurrence of optic atrophy, oculomotor abnormalities and seizures. For most affected, disease onset occurs in childhood, but at the latest during the first or second decade of life. Typical neuroimaging features are bilateral globus pallidus hypointensity on T2- weighted images, white matter lesions (WML), mild cortical and more pronounced pontocerebellar atrophy and a thin corpus callosum (TCC) (151). TCC is usually not commonly observed in the remaining NBIA disorders and should therefore direct genetic testing towards mutations in the FA2H gene, especially when it co-occurs with brain iron accumulation. However, it has to be kept in mind that TCC can be observed more frequently in other complex hereditary spastic paraplegia subtypes, mostly SPG11 and SPG15 (152, 153), which represent important differential diagnoses for this disease. Neuropathological data from post-mortem brain analysis has not yet been reported in this disorder. Sural nerve biopsy of one FA2H-positive case revealed a decrease of myelinated nerve fibres with intact myelin compaction and a normal amount of non- myelinated fibres and normal ratio of large to small fibres (148). A muscle biopsy from another case from Dick et al. showed denervation and reinnervation without evidence of mitochondrial disease (147). However, in another reported case, biopsies of skin, muscle, and sural nerve were found to be normal (154). Although not mandatory for diagnosis, bone marrow biopsy in FA2H-positive cases may show accumulation of granular histiocytes.

FAHN is inherited in autosomal recessive manner and is caused by mutations in the gene FA2H on chromosome 16q23.1. Human FA2H has seven exons and encodes the fatty acid 2 hydroxylase, an enzyme which alpha-hydroxylates incipient fatty acids. This hydroxylation step is required for the fatty acids in order to enable them structurally to

55 become incorporated into subsequent lipids that make up cell membranes and form myelin structures which are essential for impulse propagation along axons (155). The crystal structure of the human protein has not yet been reported, however recently, the 2.6 Å crystal structure of sphingolipid α-hydroxylase Scs7p, a yeast homolog of FA2H, was revealed, adding further information regarding active domains, substrate binding and substrate recognition of this enzyme (156).

Up to now, more than 30 patients/families from different ethnic backgrounds are reported (147-149, 151, 154, 157, 158) and the number is constantly growing (159, 160), especially in the current era of widespread usage of NGS-techniques in research and diagnostics. Nonetheless, reliable world-wide prevalence data is still missing and only available for few populations, e.g. the North American database registry as cited above. Additionally, characteristics of the phenotype as well as the natural history of FAHN can therefore only be reported with limited certainty and needs to await further confirmation with more cases arising (157-160). Reported sequence variants in FA2H to date are mainly private mutations and it is thought that they exert their effect through a loss-of-function mechanism. Identified mutations can lead to premature termination of the protein, or abolished/reduced catalytic activity; and confirmatory of the loss-of- function mechanism, reduced protein levels of FA2H have been detected in RT4- D6P2T cells transfected with pcDNA-hFA2H p.R154C (151). Further functional studies include transient transfection of COS7 cells to overexpress the D35Y mutation detected in the original study. Results from this model suggested functionally inactivating effects of the D35Y mutation towards FA2H activity, thereby contributing to pathogenesis (148). However, due to the rare occurrence of FAHN, the functional implications of further mutations in FA2H driving the pathogenesis so far remain understudied and therefore poorly elucidated.

In murine animal models of FA2H, demyelination and profound axonal loss in the CNS as well as cerebellar abnormalities have been observed after 12 months, whilst at that time-point peripheral nerves were principally normal in function and structure and only exhibited later-onset peripheral demyelinating and axonal neuropathy (161). Another previously published model has reported scattered axonal and myelin sheath degeneration in the spinal cord and more pronounced in the sciatic nerves of FA2H- knock-out mice from age 18 months onwards only, suggesting association of pathogenesis in mice with long term maintenance of axonal function (162). Results from potential iPSC model studies for this orphan disease still await publication at this point of time.

Kufor-Rakeb syndrome (ATP13A2 mutations/PARK9) Kufor-Rakeb syndrome (KRS, MIM #606693) was originally described in a family with a consanguineous background from Kufor-Rakeb, a small village in Jordan in 1994 (163). The initial paper described juvenile-onset severe parkinsonian features in combination with pyramidal symptoms, dementia and supranuclear upgaze paresis in this family. Subsequent studies showed levodopa therapy to improve the extrapyramidal dysfunction observed in this condition, whilst levodopa-induced dyskinesias are reported to develop early (164, 165). Subsequent papers after the original descriptions in 1994 report oculogyric dystonic spasms, facial-faucial-finger myoclonus, visual hallucinations, autonomic dysfunction, slowing of saccadic pursuit and horizontal and vertical saccades in addition to juvenile-onset parkinsonism, pyramidal signs, dementia and supranuclear gaze palsy (164-167). Initially KRS was categorized as Parkinson disease 9/PARK9. It was only twelve years after the original report from 1994 when its underlying genetic defect, a compound heterozygous loss- of-function mutation (c.3057delC/c.1306+5G>A) in a previously uncharacterised neuronal P-type ATPase-gene, ATP13A2, was identified in a large non- consanguineous Chilean sibship in 2006. Subsequent mutation screening in the original Jordanian family revealed a disease segregating homozygous 22bp duplication (1632_1653dup22) in ATP13A2 in all affected individuals (77). Reports on the incidence of brain iron accumulation in the putamen and caudate (166, 168) in a proportion of ATP13A2-positive patients led to the inclusion of this disease entity into the NBIA spectrum (168). However, it has been hypothesized that occurrence of iron might be associated exclusively with more severe mutations or may only develop later in the disease course (169). Further MRI findings comprise severe atrophy of the globus pallidus, the pyramids and more generalized brain atrophy in later course of the disease.

KRS is inherited in autosomal recessive manner and as described above has been attributed to mutations in the ATP13A2 gene on chromosome 1p36. Affected patients predominantly carry homozygous mutations (77, 164, 168, 170, 171), but compound heterozygous cases have also been reported and mutations seem to occur in diverse ethnical populations (77, 166, 172). The human ATP13A2-protein is a member of the lysosomal P5-subfamily of transport ATP-ases and contains 10 transmembrane domains (173). The 1180 amino acid protein is encoded by 29 exons and expressed in dopaminergic neurons of the substantia nigra as well as neurons from the pyramidal cortex (174).

To date, no definitive and detailed brain pathology is available in ATP13A2-positive KRS patients. However, ATP13A2 mutations were identified retrospectively in a family with juvenile neuronal ceroid-lipofuscinosis (NCL) displaying progressive spinocerebellar ataxia, intellectual decline, bulbar syndrome and pyramidal and extrapyramidal syndrome. In this family, a post-mortem brain examination had shown widespread neuronal and glial lipofuscinosis along with whorled lamellar inclusions in electron microscopy, sparing the white matter (175). Furthermore, sural nerve biopsy in ATP13A2-positive cases revealed acute axonal degeneration with an accompanying mild chronic inflammation response with endoneurial and perineurial T-cells. Abundant cytoplasmic inclusion bodies were seen within Schwann cells, epi- and perineurial cells (not within axons) and electron microscopy suggested their resemblance to irregular primary lysosomes (176).

Pathomechanistically, mutations in ATP13A2 have been associated with mislocalisation of the mutant ATP13A2, subsequent ER stress, cathepsin D dysfunction, modifications in the proteasomal and endolysosomal pathways and premature and insufficient degradation of mutant ATP13A2 (172, 177). Additional functional studies revealed the importance of ATP13A2 for mitochondrial maintenance and renewal when ATP13A2-deficient cells showed impairments in autophagy, increased mitochondrial mass, affected mitochondrial renewal and increased ROS production (178). ATP13A2-positive patient fibroblasts showed deficient mitochondrial clearance, increased oxygen consumption rates, increased fragmentation of the mitochondrial network and increased occurrence of mitochondrial DNA lesions (179). Murine models showed early neuropathological changes of widespread gliosis, increased lipofuscin granules and large ubiquitin positive aggregates with phenotypically evident age-dependent motor impairment in the ATP13A2 deficient mice (180, 181) mirroring important aspects of the human disease. Additionally, a zebrafish model was created, where knock-down of the zebrafish homologue of human ATP13A2 resulted in embryonic lethality (182). Another in vivo model using ATP13A2 deficient medaka fish revealed lysosomal dysfunction and consecutive selective dopaminergic neuronal degeneration, however, mutant fish did not exhibit locomotor impairment or any other evident and potentially treatable phenotype (177). Due to known limitations of animal models, obtained results are important for further understanding of pathogenic mechanisms, but might be difficult to translate to patients suffering from KRS. At the time of writing I am not aware of any published pathomechanistic findings from the study of human iPSC-derived ATP13A2 deficient neurons.

Hereditary ferritinopathy Hereditary ferritinopathy or neuroferritinopathy (NF, NBIA3, MIM #606159) typically presents with adult-onset extrapyramidal features including dystonia, oromandibular dyskinesias, blepharospasm and chorea. It is often thought to be phenotypically similar to neuroacanthocytosis or Huntington’s disease (74, 183). Less frequent symptoms comprise parkinsonism, cerebellar symptoms including dysarthria and tremor, and psychiatric features with depression, psychosis and cognitive decline (74, 183). Reported cases seem to cluster in Cumbria, Northern England, with the first and most common identified mutation c.460InsA in exon 4 of ferritin light chain gene FTL (also found as FTL1 in the literature), but affected patients with at least 9 different mutations from India, France, Japan and elsewhere in the world have been reported in the meantime (71, 183-185). Neuropathology revealed tau- and ubiquitin-positive neuroaxonal spheroids and neurofilaments as well as ferritin-positive inclusions in the cerebellum, the posterior putamen, and other iron-rich areas (186, 187).

MRI findings include iron accumulation in the caudate, GP, putamen, SN and red nuclei alongside to cystic necrosis in the basal ganglia and bilateral pallidum. Interestingly, hypointense signal changes predated the onset of symptoms and the burden of T2*-abnormalities increased with age (53, 188). Radiological findings alongside with decreased ferritin levels can therefore give unmissable diagnostic hints and should lead to genetic testing of FTL in order to identify this progressive disease ideally before onset of neurological symptoms and facilitate genetic counselling of families.

Hereditary ferritinopathy is caused by heterozygous mutations in ferritin light chain gene FTL and is the only NBIA disorder inherited in autosomal dominant manner. Being an essential component of the ferritin molecule, an important iron storage protein in the human body, FTL is implicated in iron pathways directly and therefore displays a unique link between iron metabolism molecularly and NBIA clinically, that necessitates mechanistic study. The most common mutation, c.460InsA, leads to an extension of the peptide-chain which alters its important ferritin dodecahedron structure and ultimately causes accumulation of ferritin and iron predominantly in central neurons that further triggers oxidative stress, neuroinflammation and neurodegeneration (71, 74, 187, 189). Most other reported mutations to date are frameshift mutations due to nucleotide(s) insertion in exon 4 of the L-ferritin gene which result in structural conformation changes of the L-ferritin subunit’s c-terminus. The heterozygous mutations hereby exert a dominant negative effect and lead to dysfunction in the iron storage efficiency of the ferritin molecule, for a recent review, see (185).

Various mouse models of the disease commonly revealed progressive brain (and systemic) iron dysregulation, with or without motor impairment and variable degree of neurodegeneration (190-192). Human iPSC-based models of this disease are awaited.

Aceruloplasminemia Aceruloplasminemia (aCP, MIM #604290) typically shows an adult-onset (mean: 51, range: 16 – 71 years) movement disorder consisting of craniofacial dyskinesias, cerebellar ataxia and predominant cognitive impairment, dysarthria and retinal degeneration (70). However, its multisystem features microcytic anemia and diabetes mellitus often predate neurologic symptoms by up to two decades and can give an important hint in diagnostics (193). The prevalence of aceruloplasminemia is estimated to be 1:2000000 in Japan where most of up to now reported cases originate from (194). The disease is inherited in autosomal recessive manner and caused by loss-of- function mutations in the ceruloplasmin (CP) gene located on the long arm of chromosome 3. Thereby, it represents the second NBIA gene with a direct link to iron homeostasis. More than 40 different homozygous nonsense, missense and frameshift mutations, as well as small deletions and insertions are reported. Few heterozygous cases appeared in the literature with generally milder and more diverse phenotypes ranging from choreoathetosis to milder cerebellar signs and tremor (70).

Neuropathologically, abundant iron deposits within mainly astrocytes, and to a lesser extent within neurons of the basal ganglia, thalamus, cerebellum and to a negligible amount in the frontal cortex, characterise the condition (195, 196). Astrocytic deformation with included globular structures proportional to the degree of iron deposition and positive for anti-ubiquitin-antibody and lipid peroxidation marker anti-4- hydroxynonenal-antibody staining are characteristic features in aCP brains. No positivity to anti-α-synuclein antibodies has been reported.

Radiographically, the homozygous mutation carriers display characteristic features on T2-weighted MRI with marked hypointensity of the thalami, the nuclei of the basal ganglia and the dentate and of cerebral and cerebellar cortices, and T2-weighted hyperintensity of the white matter henceforth showing simultaneous and confluent involvement of basal ganglia and cortex (70). For more information on the typical radiographic appearance of this and other NBIA-related disorders, see Figure 1-11 for a clinical and neuroimaging diagnostic algorithm (reproduced from (57)).

Diagnostically, due to homozygous mutations in CP, ceruloplasmin is undetectable in serum, ferritin levels are elevated and copper and iron are low (70). The encoded protein ceruloplasmin represents the most important carrier of copper in the human blood system. It is an alpha-2-glycoprotein with an incorporated ferroxidase- activity and is crucial to iron mobilisation (197). Loss of its ferroxidase function leads to excessive iron accumulation within the pancreas, liver and cerebral tissue and lipid peroxidation in astrocytes as main presumed pathomechanism in the CNS. Iron deposition is a trigger of free radical generation consecutively causing oxidative damage to the vulnerable cells and leading to the described pathology in liver, pancreas, haematological system and neuronal tissue (198-200).

Model investigations of CP mutations and their role in iron homeostasis include a knock-out mutant mouse model that revealed increased iron deposits in the brain stem and cerebellum, loss of dopaminergic neurons and associated deficits in motor coordination (201). At this point, I am not aware of published iPSC-based human neuronal models, or further animal models including drosophila or zebrafish. Despite limitations, reported models confirm CP as an important player in iron homeostasis and further study might offer therapeutic possibilities.

Figure 1-11: A clinical and radiographic diagnostic algorithm for NBIA

A proposed algorithm based on clinical and radiographic findings for evaluation of patients with suspected NBIA. Abbreviations: BG: ganglia; WM: white matter; FFF: facial- faucial-finger. Figure reproduced from (57).

COASY protein-associated neurodegeneration More recently, Dusi et al. identified homozygous and compound heterozygous missense mutations in the human gene encoding Coenzyme A synthase (COASY) in two unrelated Italian individuals with previously idiopathic NBIA that led to the inclusion of COASY protein-associated neurodegeneration (CoPAN, NBIA6, MIM #615643) into the spectrum of NBIA disorders (202). CoPAN is thought to be inherited in autosomal recessive fashion. Clinically, patients presented with early-onset spastic paraplegia, subsequently developing oromandibular dystonia, dysarthria, bradykinesia and distal amyotrophy/arreflexia accompanied by severe progressive cognitive impairment, and psychiatric and behavioural abnormalities (obsessive compulsive behaviour, depression, complex motor tics). Axonal neuropathy was reported in one of the two patients, and both reported patients had lost independent ambulation within the second decade of their lives (at age 15 or 20, respectively). No post-mortem brain analysis is available for this disease. However, brain MRI showed progressive iron accumulation in the caudate, putamina, thalami, and globi pallidi and bilateral hypointensity in the globus pallidus associated with a central region of hyperintensity in the anteromedial portion, mimicking the eye of the tiger sign. So far, more COASY-positive cases are

62 awaited to further assess its prevalence, typical clinical presentation and natural history.

The human gene COASY comprises 9 exons and is located on the long arm of chromosome 17. CoPAN is strongly linked to PKAN, as both disorders are implicated in Coenzyme A synthesis – a vital process for the cell´s energy maintenance and regulation. In a healthy organism, the encoded enzyme CoA synthase facilitates the coupling of ATP and phosphopantetheine to generate dephospho-CoA which is subsequently enzymatically phosphorylated again by the bifunctional CoA synthase to finally generate CoA. The two reported mutations in the Coenzyme A synthase have been shown to lead to alterations in RNA- and protein expression and Coenzyme A- levels in fibroblasts from the affected patients as well as in a yeast model (202). The pathogenic mechanism therefore is strongly connected to CoA deficiency; however, animal models and iPSC-based neuronal models for further pathomechanistic enquiry have not yet been reported.

More loosely associated genes: DCAF17, FBXO7, RAB39B Woodhouse-Sakati syndrome Woodhouse-Sakati syndrome (WSS, MIM #241080) was first introduced in 1983 and describes a rare autosomal recessive multisystem condition characterised by learning disabilities, extrapydamidal manifestations, hypogonadism, hearing impairment, alopecia and diabetes mellitus (203). Homozygous sequence variants in the gene C2orf37/DCAF17, encoding a nucleolar substrate receptor, were discovered as the underlying genetic cause of WSS in several Saudi families in 2008 (204). Ever since its first description, approximately 30 patients from families from the Middle East and few Caucasian, Pakistani and Indian cases have been described (205-207) with homozygous or compound heterozygous variants that lead to protein truncation with presumed loss-of-function pathomechanism. MRI characteristics indicating iron accumulation, along with adolescence-onset endocrine (gonadal dysfunction, hair loss, diabetes) and further multisystem features like hearing loss and intellectual disability can be indicative of this syndrome that is often included in the wider circle of NBIA- related disorders.

FBXO7-associated early-onset complex parkinsonism FBXO7-associated Parkinson disease 15 (MIM #260300) describes an early-onset parkinsonian pyramidal disorder with equinovarus deformities, lower limb spasticity and dopa-responsive extrapyramidal features developing later in the disease course. Biallelic mutations in the FBXO7 gene were discovered in 2008 and ever since only

63 few further families have been described (208, 209). The phenotypic presentation, natural course of the disease and its prevalence are largely unknown, however, this rare gene associated disease is sometimes mentioned in the wider circle of NBIA- associated diseases. The FBXO7 gene, located on the long arm of chromosome 22, encodes an F-box protein with a presumed role as one component of the modular E3 ubiquitin protein ligase SCF (SKP1, cullin, F-box protein), hereby playing an important role in phosphorylation-dependent ubiquitination (210).

RAB39B-associated X-linked mental retardation RAB39B-associated X-linked mental retardation (MIM #300271) sometimes appears in wider contexts with a mention as another NBIA-associated gene. Its phenotype reported in the initial discovery study includes autism spectrum disorder, epileptic seizures, and macrocephaly (211) until loss-of-function mutations in RAB39B were found in 2014 in patients with early-onset parkinsonism and intellectual disability with alpha-synuclein and iron accumulation in post-mortem brain analyses (212). RAB39B localizes to the Golgi compartment in neurons and its downregulation was shown to lead to changes in the number and morphology of neurite growth cones and presynaptic buttons in the discovery study that suggests its important role for healthy synapse formation and maintenance.

GTPBP2-associated neurodegeneration with brain iron accumulation: Another novel gene? Very recently, Jaberi et al. proposed homozygous mutations in GTPBP2 as the underlying genetic cause of neurodegenerative symptoms in combination with brain iron accumulation in one consanguineous family with three affected siblings. A splice site mutation causing deletion of exon 9 of the GTP-binding protein 2 encoding GTPBP2 gene was identified via linkage analysis and exome sequencing followed by

Sanger sequencing. Affected siblings presented with ataxic and dystonic features and mental retardation (213). Even though interesting, this single finding will need careful interpretation and reproduction in independent cohorts in the future.

Brain iron accumulation in other neurodegenerative diseases Additionally, brain iron accumulation can also occur in different neurological disorders which brings back the unresolved question of iron accumulation as an accompanying epiphenomenon or result of neuroinflammation/-degeneration or of iron as the causal and incipient trigger of the neuronal loss observed in these diseases. With different loci of highest iron content in the brain, brain iron accumulation and dysregulation has been observed inter alia in multiple sclerosis (214), Woodhouse-Sakati syndrome (as

64 discussed above) (204, 215), GM1-gangliosidosis (MIM #230500) (216, 217), alpha- mannosidosis (MIM #248500) (218), Huntington’s disease (219), fronto-temporal dementia amyotrophic lateral sclerosis complex (FTD-ALS, MIM #105550 for C9orf72- associated disease) (220), restless leg syndrome associated with haemochromatosis (221), fucosidosis (MIM #230000) (222), mucolipidosis type 4 (MIM #252650) (223), superficial siderosis (224, 225), Friedreich´s ataxia (FA, MIM #229300) (226), atypical Parkinson syndromes like multiple system atrophy, corticobasal degeneration and progressive supranuclear palsy (MIM #601104), and classical PD - even though still under dynamic controversy (227-231). Of those disorders, only Woodhouse-Sakati syndrome, caused by homozygous mutations in DCAF17, is included into the wider NBIA disorder spectrum in some literature and therefore has been briefly discussed above. However, it is beyond the scope of this introduction to review all diseases with observed brain iron accumulation mentioned above in detail, and interested readers must be relegated to seminal review literature at this point (e.g. (232)).

Idiopathic NBIA Even though declining due to the spread of next-generation sequencing and the advent of whole genome sequencing technologies, there is still a proportion of NBIA cases classified as idiopathic, meaning without detected genetic origin (58, 133). Careful dissection of their family history with collection of large pedigrees and powerful combinations of genetic techniques (linkage analysis, homozygosity mapping and whole exome or whole genome sequencing) will further nurture the ongoing excitement and dynamics in the gene hunt and additionally inform phenotype-genotype correlations in NBIA disorders. Collection of the idiopathic cases in international consortia and subsequent screening of those once a new gene has been identified will further reduce idiopathic NBIA in number and sharpen the understanding of the prevalence of the genetic subforms and their associated clinical spectrum.

In the coming years, the advent of whole genome sequencing may introduce novel intronic variants associated with the disease that are linked to alternative splicing, nonsense-mediated mRNA-decay and altered transcription of disease-relevant genes. These may feed into the known or point to yet unknown novel therapeutically targetable pathways and may account for additional parts of the idiopathic cases within the spectrum. NBIA or pallidopyramidal syndromes are - despite their scarcity – an exciting and expanding field of study and may, due to their significant overlap in pathophysiology and phenotype with common diseases like PD and AD, pave the avenue in understanding common pathways in clinical neurodegeneration.

Can genetics help to clarify which pathways to target and how? To date, therapeutics in NBIA remain largely symptomatic. They can bring relief and are therefore important, but they do not target the origin of the disease and are incapable of stopping the neurodegenerative process. The development of specific, causal and successful therapeutic agents in the future will therefore heavily depend upon the full understanding of the implicated molecular pathways in this disease. Therefore, genetics have been, are and will be crucial. In the past, the identification of the genes mutated in NBIA disorders as discussed above have elucidated common pathways in NBIA that include iron aggregation, mitochondrial deficiency and lysosome and ceramide metabolism (for an overview of the different genes associated with NBIA to date, and their role in the different pathways, see

Figure 1-12). Primarily the lysosomal function seems to be an emerging candidate as the main metaboliser of ceramides which are crucial for many cellular processes (233) and seem to be involved in tauopathies equally as in Lewy body formation (151, 234) bonding NBIA closer to more common neurodegenerative diseases like atypical and typical Parkinson syndromes and potentially Alzheimer’s dementia.

Figure 1-12: NBIA genes and pathways

NBIA genes and their suspected settlement within the different and interacting disease associated pathways directly and indirectly linked to iron accumulation (reproduced from (45)).

Closing the gaps: How iPSC technology could help to foster the understanding of NBIA So far, the lack of human neuronal cell models that faithfully recapitulate disease phenotypes in vitro has been one of the major hurdles in the efforts to study the mechanisms of NBIA disorders effectively. However, the recent discovery and rapid spread of iPSC technology may yield new important insights into NBIA pathophysiology within the upcoming years. This fascinating breakthrough-technique enables one to model PKAN and other NBIA disorders with a defined genetic mutation in vitro via the pathophysiologic study of induced pluripotent stem cell derived neurons that carry the exact genetic information as the affected patients. A joint collaborative project with Dr. Selina Wray and Dr. Charlie Arber has started to tackle these models for PANK2 and COASY mutations (see Discussion of results chapter 3 below).

Briefly, iPSC technology involves collecting fibroblasts from genetically confirmed NBIA patients via skin biopsy and consecutive laboratory reprogramming into pluripotent stem cell state via episomal gene delivery (235). Subsequently, neural induction is initiated via treatment with dual SMAD inhibitors (236) followed by exposition to default neural inductive conditions in the presence of retinoids for 80-100 days to produce cortical neurons (237) derived from the patient’s original skin cells and ready for experimental study. Thereby, extensive characterisation of the molecular events in these iPSC-derived neurons inclusive screening for possible therapeutic rescues becomes possible. Significant time during this PhD was spent to learn these exciting and necessary techniques around iPSC disease modelling that will be introduced with obtained results and limitations in one of the following chapters (Chapter 6) of this thesis in more detail.

1.4 Polyglutamine disorders and inherited cerebellar ataxias

Introductory remarks This subchapter prepares the reader for parts of the results chapter 4, the full results chapter 5 and parts of the results chapter 6.

Polyglutamine diseases are a subgroup of unstable repeat expansion disorders characterised by CAG-triplet repeat expansions (coding for polyglutamine chains) of various sizes in a gene, specific for each disease. They might be individually rare, however as a group they represent a significant fraction of neurodegenerative diseases that share important mechanisms of pathophysiology. Functionally and genetically, the disorders of unstable repeat expansion include tri-, tetra- and pentanucleotide repeats and can be classified into a) diseases caused by expansions of non-coding repeats resulting in putative loss of protein function, b) disorders due to underlying expansions of coding repeats resulting in altered protein function and c) diseases that are caused by expansions of non-coding repeats but result in altered RNA function – please note that these classes and respective classifications are dynamic and under active, ongoing study as mechanisms emerge and novel techniques of study become available (238). Whilst the overall group of diseases of unstable repeat expansion continuously grows and includes more than 30 distinct conditions (239), CAG-triplet repeat disorders make up approximately one third of these with currently nine entities described: Huntington’s disease, spinocerebellar ataxias 1, 2 (SCA2, MIM #183090), 3 (SCA3, MIM #109150; also known as Machado-Joseph disease (MJD)), 6 (SCA6, MIM #183086), 7 (SCA7, MIM #164500) and 17 (SCA17, MIM #607136), dentatorubral- pallidoluysian atrophy (DRPLA, MIM #125370) and spinal and bulbar muscular atrophy (SBMA, MIM #313200). The CAG repeats encode polyglutamine chains of different size that once above a certain threshold exert pathological function within the respective genes. Even though they are phen otypically distinct disorders where the underlying pathologic triplet repeat expansion lies in a different gene for each disease, they share common pathogenic principles such as pathological intracellular insoluble inclusion aggregates, prominent neurological involvement, autosomal dominant inheritance (apart from SBMA which is inherited in X-linked fashion), an inverse correlation of the expanded allele size with age at onset and/or clinical severity, anticipation in successive generations, observed somatic instability and the presence of interruptions of the expanded repeats for most but not all trinucleotide repeat disorders. Table 1-2 gives an overview of these nine trinucleotide repeat disorders with their respective genes harbouring the repeat, their prevalence, phenotype, normal and pathogenic repeat sizes, the variance in age at onset explained by repeat length and the observations regarding somatic instability and interruption of expanded repeats.

Table 1-2: Summary of trinucleotide repeat disorders

Repeat Variance in AAO Normal Pathogenic SI disorder Prev. Phenotype explained by repeat range range /Int. and gene length Involuntary movements, HD: Yes 3-10 cognitive 50-60% (40-60%) 6-35 40-121 HTT /Yes impairment

Ataxia, ophthalmoplegia, SCA1: 64-76% (no detected Yes 0.16 pyramidal and 6-38 45-83 ATXN1 heritable component) /Yes extrapyramidal features Ataxia, neuropathy, SCA2: ophthalmoplegia, Yes 0.2 50-80% (17-59%) 15-31 33-500 ATXN2 extrapyramidal /Yes features Ataxia, pyramidal signs, neuropathy, SCA3: Yes 0.4 extrapyramidal 45-80% (46%) 12-44 52-87 ATXN3 /Yes features, ophthalmoplegia SCA6: 26-52% (no detected Unk./ 0.04 Ataxia 4-18 20-33 CACNA1A heritable component) Unk. Ataxia, macular degeneration, SCA7: ophthalmoplegia, 71-88% (no detected Yes 0.12 3-19 37-460 ATXN7 pyramidal and heritable component) /No extrapyramidal features Ataxia, pyramidal signs, dementia, SCA17: Unk. <0.02 seizures, Unknown 25-40 49-66 TBP /Yes extrapyramidal features Myoclonus, DRPLA: 0.005 Yes epilepsy, ataxia, 50-68% 6-35 48-93 ATN1 -0.04 /No dementia Limb and bulbar SBMA: 0.65- weakness, Yes 29% 9-34 38-72 AR 2 neuropathy, /No endocrine features

NB for table above: Epidemiology and CAG repeat ranges of trinucleotide repeat disorders. Prevalence is given/100000 European population. Abbreviations: Prev. – prevalence; AAO – age at onset; SI – somatic instability; Int=interruptions within repeat sequence; Unk. – unknown; HD – Huntington’s disease; SCA – spinocerebellar ataxia; DRPLA – dentatorubral-pallidoluysian atrophy; SBMA – spinal and bulbar muscular atrophy. This table is modified from (2).

The expanded repeat introduces a sequence change in the innate wildtype sequence that – once above a certain threshold and actively transcribed – loses compatibility with normal conformation and function of the protein. Resulting common molecular themes proposed to contribute to pathogenesis in trinucleotide repeat disorders include transcriptional dysregulation, mitochondrial dysfunction, altered RNA-metabolism, dysfunctional calcium homeostasis, aberrant proteolytic cleavage, defective axonal transport, cytoskeletal abnormalities, protein misfolding and protein accumulation 69 escaping intracellular clearance mechanisms and potentially contributing to cell death (see Figure 1-13b for a graphic overview and review articles (238, 240) for details of implicated pathways). Observed differences in clinical phenotype and underlying cell vulnerability/atrophy of distinct brain regions/neuronal subpopulations might be determined by the intrinsic function of the respective disease causing protein, its expression pattern and interaction partners.

However, several conundrums as to why some diseases show stronger anticipation and more pronounced somatic and repeat instability than others do remain, and the identification of additional disease modifiers might contribute towards a more clear picture of pathogenesis in the trinucleotide repeat disorders. More specific details about the trinucleotide repeat disorders helpful to understand the motivation behind the studies described in results chapter 5 can be found in the respective introductions to the two subchapters of Chapter 5 in this thesis.

Figure 1-13: Proposed pathomechanisms in trinucleotide repeat disorders

a) A common theme among the nine polyglutamine diseases is altered protein conformation, leading to protein accumulation and aberrant interactions. The figure shows hypothetical aberrant interactions; some of these might be inappropriately enhanced, whereas others might be lost or unchanged. b) Expanded polyglutamine proteins might mediate pathogenesis through a range of mechanisms. Accumulation of mutant polyglutamine protein to form insoluble inclusions recruits components of the ubiquitin–proteasome system and other protein quality-control pathways, although the implications of this for pathogenesis are unclear. The full-length polyglutamine protein can also be cleaved by proteases to form fragments, which might also mediate pathogenic effects. The mutant protein affects many cellular processes, including transcription and RNA metabolism. The effects on transcription can occur through interaction of the mutant protein with transcriptional activators and repressors (including CREB-binding protein (CBP), Sp1 transcription factor (SP1) and nuclear co-repressor proteins), or through effects on chromatin. Alternatively, the mutant protein itself might serve as a transcription factor. Other cellular processes that can be affected include mitochondrial function, calcium homeostasis, and axonal transport, ultimately leading to neuronal dysfunction and death. Abbreviations: HDACs, histone deacetylases; Q, glutamine; Ub, ubiquitin. Figure and figure legend reproduced from (238).

Seven of nine trinucleotide repeat disorders present with ataxic phenotypes and are included in the group of autosomal dominant cerebellar ataxias, a subgroup of inherited cerebellar diseases. In the following, I will give a brief overview introduction to the inherited cerebellar ataxias to lend motivation to the studies described in Chapters 4, 5 and 6.

The hereditary ataxias are a heterogeneous group of genetic disorders united by occurrence of slowly progressive incoordination of gait, fine motor skill tasks, speech, and eye movements. Atrophy of the cerebellum is observed on a frequent basis, and the dysfunction of the cerebellum and its associated systems is at the core of the clinical symptoms. There is variable involvement of additional systems leading to changing frequencies of accompanying features such as optic atrophy, neuropathy, retinopathy, extrapyramidal and pyramidal symptoms, seizures, intellectual disabilities, dementia, sensorineural deafness, endocrine manifestations and more (241). Inherited cerebellar ataxias are often classified according to their mode of inheritance in the first and using the mutated gene in the second instance. This approach creates numerous subclassifications within the four broad inheritance categories autosomal dominant, autosomal recessive, maternal and X-linked inheritance. Figure 1-14 gives some guidance for the multifaceted diagnosis of slowly progressive cerebellar disease in clinical practice, a field constantly evolving. Past efforts have yielded more than 20 genes causally implicated in autosomal dominant ataxias and over 70 genes that can cause recessive ataxia (242, 243) and the numbers are constantly increasing. However, individually and subsumed, hereditary ataxias are rare: The overall estimated prevalence of the autosomal dominant cerebellar ataxias (ADCAs/SCAs) lies at approximately 1-5/100000 and is estimated at 3/100000 for the autosomal recessive ataxias (ARCAs/SCARs) with evident differences between populations (244-246). Amongst the autosomal dominant spinocerebellar ataxias, SCA3 is the most frequent disorder, whereas Friedreich’s Ataxia, ataxia-telangiectasia (AT, MIM #208900) and ataxia oculomotor apraxia syndromes represent the most common recessive forms (247). With further spreading of next-generation sequencing techniques, the absolute count of causal genes, as well as their frequency spectrum, associated phenotypic manifestations and observed mutational spectrum undergo constant change and expansion with implications for research, diagnostics and clinical practice equally (243, 248, 249). Table 8-0 (genes under “Ataxia”) gives an overview of the important genes implicated in autosomal dominant, autosomal recessive, spastic and X-linked recessive ataxias with their respective mode of inheritance, year of gene discovery and associated phenotypes. In this introduction, I will not be able to give a detailed account of each of

72 the genes and their respective pheno- and genotypic spectrum and the pathogenetic mechanisms implicated in hereditary cerebellar ataxias. However, respective introductions to subchapters of Chapter 4, 5 and 6 will contain more specific information on some autosomal recessive (e.g. SYNE1- and PNPLA6-associated ataxias) and autosomal dominant spinocerebellar ataxias (trinucleotide repeat ataxias and SCA15) as they are more closely investigated in this thesis.

Figure 1-14: Diagnostic algorithm progressive ataxias

Red represents clinical considerations, blue represents MRI considerations, purple represents anamnestic and clinical genetic considerations, and green represents molecular genetic considerations. The pathway for the diagnosis of patients in whom the condition seems to be sporadic is depicted with red arrows. (1) If clinically justified and without a family history of Friedreich’s ataxia. (2) If clinically justified. Abbreviations: AOA, ataxia with oculomotor apraxia; ARSACS, autosomal recessive spastic ataxia of Charlevoix–Saguenay; AVED, ataxia with vitamin E deficiency; FRDA, Friedreich’s ataxia; ILOCA, idiopathic late-onset ataxia; MSA, multiple system atrophy; POLG, mitochondrial DNA polymerase gamma. Figure and legend reproduced from (250).

1.5 Induced pluripotent stem cell technology and its usefulness for cerebellar disease modelling

Introductory remarks This introductory subchapter prepares the reader for the results chapter 6. Modified contents of this introduction subchapter have been published in parts and figures in (251).

As evident from introductory subchapters above, molecular genetic studies have formed the basic understanding of human disease, especially of neurodegeneration. However, this knowledge is limited and generation of functional models of neurodegeneration to elucidate the underlying pathology and to indicate potential areas of treatment has been complicated by the intrinsic biological complexity and respective limitations of model systems. The same holds true for the cerebellum, and its inherited disorders described in the preceding subchapters have motivated the functional part of this research. This last introductory subchapter shall give an introduction into the development of the cerebellum and into the field of induced pluripotent stem cell models for neurodegenerative diseases, more specifically cerebellar disorders. It shall provide an overview of the potential promise and important shortcomings of human induced pluripotent stem cells for regenerative neurology, with a particular emphasis on in vitro modelling of cerebellar degeneration, laying the foundations to understand the iPSC work conducted and presented in this thesis later on.

Consisting of several distinct cellular subtypes, the cerebellum constitutes a highly organized and irreplaceable coordinate of motor function within the adult neuraxis. Cerebellar disease, attributable to a variety of different genetic and environmental factors, results in the loss of function of defined subclasses of neurons, and represents a pivotal and untreatable healthcare burden. The unavailability of therapies for most cerebellar disorders can partially be attributed to unresolved disease mechanisms due to inaccessibility of human cerebellar neurons in relevant experimental contexts where initiating cascades triggering disease mechanisms and neurodegeneration could be functionally elucidated, and where drug screens could be performed in the search of required therapeutic agents. It is enthusiastically believed that with the advent of human induced pluripotent stem cell (hiPSC) technology, these shortcomings mentioned above could be reversed advancing the field in its search of potential treatments and towards a better understanding of pathomechanisms leading to cerebellar disease.

Initial remarks: The cerebellum and iPSC technology The cerebellum, also referred to as the “little brain”, is an extensively researched component of the neuraxis. Nevertheless, the molecular pathogenesis underlying cerebellar disease remains unsatisfactorily understood, resulting in the vast majority of such disorders still being immedicable to date (252). Representing one of the dominant neuraxial structures displaying high cellular and structural complexity, the cerebellum integrates and orchestrates ~80% of the total number of CNS neurons in only 10% of the total CNS volume (253). Anatomically positioned in the posterior cranial fossa the cerebellum functionally coordinates the major input from pontine and vestibular nucleii, the inferior olive and from the spinal cord towards its major output pathways via the ventrolateral thalamus and pontine, medullary reticular, red and vestibular nucleii (254- 257). Using non-iPSC technology strategies, the cerebellum has been substantially investigated and is found to have well-established functions in balance, posture, motor control, motor learning and (fine) coordination (255, 258, 259). Furthermore, there exists an ever-growing literature about the cerebellum and its implications in emotion and cognitive functions (260-263). The unique organisation, precise wiring and distinct pathophysiology render this particular structure equally fascinating as vulnerable to injury and very difficult to repair.

Lately, approaches in regenerative medicine have been transformed drastically by the potential of disease modelling, disease manipulation and potential disease cure via human iPSC technology, a ground-breaking technological discovery made in 2007 and later recognized by the Nobel Prize in Physiology or Medicine (264, 265). Briefly, via transfection with e.g. lenti-, Sendai-, retroviruses, minicircles, piggyBac, miRNA, mRNA or episomal plasmids that carry different pluripotency-associated transcription factors (the most commonly used combination of reprogramming factors consists of c-myc, SOX2, OCT4 and KLF4, also named ‘Yamanaka factors’) this reprogramming technique initiates a cascade of erasure and remodelling of epigenetic marks that enables one to generate so called induced pluripotent stem cells. These pluripotent cells display features similar to true human embryonic stem cells (hESCs) and can be generated readily from adult human skin cells, or subsequently most other human fully differentiated tissues (see Figure 1-15 for a schematic). Depending on the technique utilised during reprogramming (miRNA, episomal plasmids, viral transfection, etc.) and the adult cell type starting the process with (e.g. fibroblasts, blood cells, renal epithelial cells from urine, etc.), the reprogramming process takes around 2-8 weeks and is variably efficient. Upon introduction of reprogramming factors, the cells reprogrammed to 100% will slowly start forming colonies of pluripotent stem cells. These gain significant growth advantage over the

75 non- or only partially reprogrammed cells in the dish given their proliferative capacity. They can then be isolated based on expression of surface markers or reporter genes, their morphology or based on media and surface conditions that additionally select for their growth. Extensive validation and rigorous quality control of iPSC clones is necessary, including expression of endogenous pluripotency markers, differentiation capacity in vitro and in vivo, silencing of exogenous transcription factors and exclusion of chromosomal abnormalities or transgene integration.

Figure 1-15: Scheme of induced pluripotent stem cell technology

Somatic cells from adults can be cultured and reprogrammed to induced pluripotent stem cells by transient transfection with exogenous pluripotency factors. These factors are thought to change the epigenetic landscape of transfected cells towards silencing of differentiation-associated genes and enhanced transcription of endogenous pluripotency genes. Figure reproduced from (266).

Purified and fully validated iPSCs can then be differentiated into almost any cell type of the human body where efficient differentiation protocols that suppress pluripotency and guide the cells to their desired fate exist. Depending on the cell type and the individual hypotheses of the respective study design, the generated cells can then be analysed using classical cellular phenotyping assays investigating cell cycle, cell migration, cell metabolism, cell growth, apoptosis, nuclear or cytoplasmic foci, cell shape, membrane texture, cytoskeletal reorganisation, neurite outgrowth, mitochondrial mass, mitochondrial membrane potential, ROS production, protein localisation, expression and quantification, etc. For the field of neurology, the differentiation of human iPSCs from healthy and diseased individuals into disease relevant neuronal cells permits for the first time in vitro study and modelling of human neurodevelopmental and neurodegenerative processes. It thereby holds the potential to significantly foster the understanding of pathological processes governing different developmental and degenerative diseases and thus to inform strategies aimed at stopping disease progression, with the ultimate but more long-term aim to restore structure and function (236, 237, 267). Specific advantages of iPSCs include directed differentiation to any human cell type to model disease in a highly reductionist fashion. Furthermore, these cells will express mutations at the normal pathophysiological level and do contain the genetic background

76 information of the diseased individual, removing the need for artificial over expression, knock-down or knock-out. Additionally, given their theoretically limitless potential to proliferate before directed differentiation into disease relevant cell types, this renewable resource for experimental material both holds the potential to recapitulate disease phenotypes and to further enable high-throughput drug-screening to identify a revertive therapeutic agent (268-270). In the fields of disease modelling this technique already has and might further reduce the amount of animal research in specific areas. Another long term promise of hiPSCs will be patient-targeted individual therapy through transplantation, even though accurate and safe restoration of circuit integrity and function after established damage, as well as safe, surgical procedures and informative readouts to safely deliver and monitor these approaches will be formidable challenges.

The advent of hiPSC technology is of unprecedented relevance to a structure such as the cerebellum both from developmental and clinical neurological perspective: In this field, most of the mysteries remain to be unravelled (252, 271, 272) and no curing treatment is readily available for the vast majority of cerebellar disorders. Against this background however, understanding the developmental processes involved in cerebellar differentiation is crucial to enable robust and reproducible directed differentiation of hiPSCs in vitro, in order to be able to generate accurate disease recapitulating models for the variety of cerebellar disorders.

Since cerebellar disorders have so far rather reluctantly benefited from the iPSC technology, I want to first briefly introduce the potential and remaining challenges of iPSC technology in the general fields of (regenerative) medicine and neurology, before focusing on current approaches, achievements and limiting factors for directed differentiation to cerebellar neurons.

iPSC technology and its implications in general medicine As briefly discussed above, the discovery of efficient reprogramming of adult human fibroblasts into iPSCs in 2007 (264) has caused an unprecedented fundamental paradigm shift in regenerative medicine. The unique potential and importance of human iPSC-based modelling has only been confirmed by subsequent technique improvement and optimisation as well as refinement of reproducibility, safety and scalability (235, 273-277). By introducing a continuously reduced number of transgenes it is becoming a standardized experimental procedure to revert primary somatic cells of diseased patients and healthy control individuals into pluripotent stem cells for many laboratories world-wide. Naturally, the technology has attracted significant (and continually growing) commercial interests (e.g. see Bioinformant L.L.C.:

77 http://www.bioinformant.com/bioinformant-announces-release-of-complete-2015-16- induced-pluripotent-stem-cell-industry-report/ accessed: May 2016).

These reprogramming methods circumnavigate ethical concerns and limitations accompanied by usage of primary human embryonic stem (hES) cells for medical research, even though novel and updated ethical guidelines for iPSC-related research are and should be continuously conceptualized as the field develops (e.g. (278), and http://www.eurostemcell.org/factsheet/ethics-and-reprogramming-ethical-questions- after-discovery-ips-cells accessed: May 2016). iPSC technology revolutionizes predating approaches including somatic nuclear transfer into oocytes or fusion with embryonic stem cells, leading to tetraploidy (266) and stimulates numerous further discoveries. The theoretically endless self-renewing capacities and the preserved genetic background (where mutated genes are expressed at a representative pathophysiological level in the context of the patient’s genetic background contributing to disease and susceptibility) render human iPSCs an unparalleled and unprecedented resource for basic science and clinically motivated research. Combined with ongoing progress in the field of directed differentiation towards disease-specific cell types (237, 265, 279-282), hiPSC in vitro models have been shown to faithfully recapitulate key aspects of human development, as well as of sporadic and genetic disease that would otherwise be inaccessible for thorough study.

Thus, and for the first time in regenerative medicine, iPSC technology permits detailed temporal and spatial interrogation of occurring pathomechanisms in a patient-specific way by usage of a fully humanized model, circumventing previous needs for artificial overexpression, knock-down or knock-out. For a schematic on iPSC technology and its promises and limitations in cerebellar diseases, see Figure 1-16.

iPSC technology and its implications in neurology Neuronal cells have been amidst the earliest cell types to be differentiated with the help of efficient and robust differentiation approaches from hESCs (283, 284) and hiPSCs (236, 237). Employed protocols include co-culture with neural inducing feeder cells, direct pharmacological inhibition of transforming growth factor beta 1 (TGF-beta)- and bone morphogenetic protein (BMP)-signalling performed on a monolayer of confluent stem cells (dual SMAD inhibition) or in vitro generation of suspension based three- dimensional embryoid bodies exposed to retinoids in a stage specific manner. All outlined approaches can reliably initiate ‘neural induction’ and efficient ‘neural conversion’ that represent the first steps in any protocol that aims to generate neurons from iPSCs. The resulting neural precursor cells can be “patterned“ (second step) by

78 application or absence of developmentally rationalised morphogenetic extrinsic cues, and finally be terminally differentiated (third step) towards region-specific neuronal (neural crest, spinal motor, cerebellar, hypothalamic, dopaminergic, cortical) and glial subtypes. Please see Table 1-3 for a non-exhaustive overview of neuronal subtype specific differentiation approaches available.

Table 1-3: Neuronal differentiation protocols from iPSCs

Neuronal cell First author, Developmental cues (neural type year of Culture method Duration (days) conversion and patterning) publication Cortical Watanabe et Serum free BMP ant. (BMPRIA-Fc) 35 precursors al. 2007 embryoid body Activin/Nodal ant. (LeftyA) (SFEB)-like Wnt ant. (Dkk1) Cortical Eiraku et al. SFEB derivative BMP ant. (BMPRIA-Fc) 45-60 neurons 2008 Activin/Nodal ant. (LeftyA) Wnt ant. (Dkk1) Cortical Chambers et Monolayer BMP ant. (NOGGIN) 19 neurons al. 2009 Activin/Nodal ant. (SB431542) Cortical Li et al. 2009 Suspension None for cortical (endogenous Wnt) 30-35 neurons For MGE and LGE derivatives: and MGE / Wnt ant. (Dkk1) LGE neurons Sonic hedgehog (SHH) Cortical Shi et al. 2012 Monolayer BMP antagonist (NOGGIN) 80-100 neurons Activin/Nodal antagonist (SB431542) Midbrain Kricks et al. Monolayer BMP ant. (NOGGIN or LDN) 80 dopaminergic 2011 Activin/Nodal ant. (SB431542) neurons SHH and purmorphamine Fibroblast growth factor 8b (FGF8b) Wnt agonist (CHIR99021) Midbrain Kirkeby et al. Embryoid body BMP antagonist (NOGGIN) 35 dopaminergic 2012 Activin/Nodal antagonist (SB431542) neurons Wnt agonist (CT99021) Sonic hedgehog (SHH-C24II) Midbrain Jaeger et al. Monolayer BMP ant. (NOGGIN) 30-35 dopaminergic 2011 Activin/Nodal ant. (SB431542) neurons FGF/ERK ant. (PD0325901) Fibroblast growth factor 8b (FGF8b) SHH Cerebellar Erceg et al. Embryoid body Fibroblast growth factors (FGF8, 35 neurons 2012 FGF4, FGF2) Retinoic acid (RA) Wnt agonists (Wnt1, Wnt3a) BMPs (BMP4, BMP6, BMP7, GDF7) Sonic hedgehog (SHH) Cerebellar Muguruma et SFEBq Activin/Nodal ant. (SB431542) 35-135 neurons al. 2015 FGFs (FGF2, FGF19) Insulin Stromal-cell derived factor 1 (SDF-1) (co-culture with mouse granule cells to generate Purkinje cells) Cerebellar Wang et al. Embryoid body FGF2 20-65 neurons 2015 Insulin Sonic hedgehog ant. (cyclopamine) (co-culture with rat organotypic cerebellar slice to generate Purkinje cells) Spinal cord Li et al. 2005 Monolayer Retinoic acid (RA) 21-35 motor SHH neurons FGF2 Spinal cord Patani et al. Suspension Activin/Nodal ant. (SB431542) 21-35 motor 2011 Sonic hedgehog (purmorphamine) neurons Fibroblast growth factor (FGF2) Spinal cord Calder et al. Monolayer Activin/Nodal ant. (SB431542) 35-40 motor 2015 BMP ant. (LDN193189) neurons Retinoic acid (RA)

This table is modified from (285). Abbreviations:ant.=antagonist; all other abbreviations explained in the table.

Unless differentiated iPSC-derived neurons are “forced“ to age (e.g. via progerin expression (286)), this technology models human developmental processes and diseases (287, 288) in an essentially fetal and premature system (289). Interestingly, an ever-growing number of studies nonetheless confirm mutant hiPSC-derived neurons obtained from patients with inherited disease - including adult-onset disorders - do successfully recapitulate central cellular pathomechanisms despite their fetal characteristics (281, 290-295). Furthermore, in addition to confirming known and uncovering novel cellular and molecular pathomechanisms for a range of sporadic and inherited disorders (290, 293), human iPSCs have also been utilised to screen for novel potential therapeutics successfully (270, 296, 297). They have been helpful in delineating genetic and non-genetic factors driving neuronal degeneration as lately demonstrated for the case of monogenic twins discordant in their clinical phenotype of PD (295). With the constantly growing literature around iPSC technology in neuronal disease, an exhaustive and detailed summary of the most important findings from iPSC studies in the field of neurology is not at the goal of this introduction subchapter. For recent reviews on this topic the interested reader is referred to the following cited manuscripts (267, 298, 299).

Despite its undisputed contribution towards better understanding of development and disease in the human nervous system, there are remaining general caveats to be aware of: Low-efficiency and non-directed differentiation that might cause proliferation and teratoma formation in vivo, undiscovered transgene integration, uncontrolled epigenetic and genetic modifications of produced stem cells and precursors from in vitro culture conditions, time and cost-intensive, laborious culture and experimental settings with high intrinsic biological variability and possible oncogene reactivation via reprogramming procedures are only some that merit mentioning and warrant caution. These caveats will need further thorough optimisation until iPSC technology can harvest its full potential for disease modelling, drug discovery and ultimately autologous transplantation. The latter so far remains a long-term aim with several unresolved issues to date (265, 300-305).

What distinguishes cerebellar disease modelling using iPSCs from the more successful modelling of other subregions of the human brain?

Compared to other neurological diseases, successful modelling of cerebellar disorders using human iPSC technology is reported with much lower frequency and has not yet become a broadly used routine approach. One reason for this is that the achieved progress in modelling cortical (237, 290), striatal (306-309), midbrain (293, 310-312), spinal (313-318), peripheral sensory (319- 322) and autonomic nervous system (322) pathologies with human pluripotent stem 80 cells has one crucial and distinct advantage over cerebellar disease modelling: standardized, reliable and efficient protocols for directed differentiation of region- and subtype-specific neurons (e.g. see Table 1-3) and glia are available and novel refined protocols get developed with high speed. This naturally fosters the applicability of hiPSCs in clinical research and further paves its way towards translation. This disadvantage is clearly reflected in the number of publications generated in these respective fields compared to the low finding turnout from iPSC-based cerebellar modelling. Comparing the number and content of iPSC publications studying neurological non-cerebellar disorders to the amount and tenor of published work harnessing iPSC technology to study cerebellar pathology one currently faces a pronounced disequilibrium: 1) Interestingly, the few published reports studying cerebellar diseases with iPSC technology (291, 323-325) have derived non-cerebellar cells from patients with primarily cerebellar symptoms, 2) no reported study thus far has reproduced or employed the published protocol to specify cerebellar-like cells from hiPSCs (326) which clearly was available by the time of the majority of the publications.

Why is the cerebellum hanging back? It seems plausible that the inherent and notorious developmental complexity of cerebellar granule- and especially Purkinje cells (PCs) has caused subsequent difficulties generating these neuronal subtypes from human pluripotent stem cells. This might have hampered significant progress in this field, despite both early (327-329) and indeed more recent developmental breakthroughs (330); see “bottleneck“, Figure 1-16. However, another important contribution to the lack of reliable and efficient cerebellar differentiation protocols might additionally come from the rareness of cerebellar diseases. One could argue that simply the socio-economic pressure and medical necessity to generate human iPSC models of cerebellar disorders is lower. Therefore orchestrated efforts of developmental biologists might have focussed on deriving more ‘deserving’ cell types for more common disorder such as PD and AD. In these fields, significant progress for efficient directed differentiation of disease-relevant neurons has indeed been made (237, 310-312). Even though (hereditary) degenerative disorders of the cerebellum might be less frequent than degenerative conditions of the cortex or basal ganglia, the cerebellum is affected in a range of common and rare diseases, amongst them cerebellar degeneration due to trauma, alcohol abuse, nutritional deficiency and paraneoplastic syndromes, or degeneration due to rare inherited conditions (dominant and recessive spinocerebellar ataxias). Furthermore, and most importantly, multiple sclerosis - the commonest cause for neurological disability in young people from developed countries

- frequently affects the cerebellum, and tumours of the granule cell (medulloblastomas) represent up to 20-40% of brain tumours in children and young adults (327, 331). Additionally, cerebellar malformations (e.g. Dandy-Walker- or Arnold-Chiari- malformations, Joubert-Syndrome, pontocerebellar hypoplasias) are highly disabling both to the affected patients and their dependants, significantly reducing life expectancy and quality of life. Currently, one cannot offere treatment to influence healthy cerebellar development or to prevent ongoing cerebellar degeneration. Unless one understands the molecular and cellular mechanisms underlying human cerebellar degeneration more precisely, one will not be able to do so in the near future. The generation of reliable cerebellar disease models will be crucial to progress in this field. It therefore needs highlighting that a deeper mechanistic understanding of cerebellar developmental processes will permit establishment of accurate disease models using iPSC technology. Insights into the molecular pathogenesis of cerebellar conditions will in turn allow more mechanistically rationalized approaches to therapies in this field.

Figure 1-16: iPSC technology in cerebellar diseases

iPSC technology has already started to foster the study of the human cerebellum and its pathologies in a patient-specific way. Overcoming the current bottleneck of directed differentiation will further facilitate the beneficial effects of this technology on the disorders of the cerebellum. Cerebellar Patient taken from Netter’s Concise Neurology, p.85, Elsevier, Inc. (copyright). This figure is published elsewhere, see (251).

What has iPSC technology contributed to the study of cerebellar diseases so far? One publication reports successful generation of human iPSC-derived long-term self- renewing neuroepithelial-like stem (lt-NES) cells from patients with spinocerebellar ataxia 3 to study the cellular pathology of the most common autosomal dominant degenerative cerebellar disorder (291). Here, the inherent difficulty and the lack of efficient in vitro cerebellar differentiation protocols are circumvented by studying lt- NES-cells, a surrogate neuronal subpopulation. These cells do express a hindbrain-like transcriptional signature with markers usually observed in cells of a ventral anterior hindbrain fate (332). This transcriptionally (and possibly functionally) approximates these cells to the cerebellum, however lt-NES-cells share few characteristics of bona fide cerebellar granule cells and even less of physiological Purkinje cells. Using this proxy model, the authors successfully detected neuron-specific early aberrant protein processing that was aggravated by excitatory activation in SCA3-mutant iPSC-derived neurons. This study in its early appearance was breaking a milestone for the field in modelling late-onset cerebellar “disease in a dish” (291). Since then, Friedreich’s ataxia, the most common autosomal recessive ataxia with additional multisystem affection where cerebellar pathology is a late, secondary process in the disease, has been studied via various attempts using iPSC technology. Again, investigators in these studies utilised non-cerebellar cells to study the cellular phenotype (323-325, 333). I am not aware of many other cerebellar diseases being successfully modelled using iPSC technology at this point.

In summary, none of the few studies above reported successful utilisation or modification of the – by that time - only published protocol describing cerebellar-like granule cell differentiation from human iPSCs (326) and all studies used surrogate cells to establish their findings.

In order to explain the lack of studies using patient-derived iPSCs to generate cerebellar cells and to study their pathology, it makes sense to address the following, developmental question:

What is known about the development of cerebellar cells that might distinguish them from other neuronal cells? In humans as well as most other higher organisms, cerebellar development occurs over a prolonged time span ranging from the early embryonic period to the first postnatal years. It follows its own distinct stereotyped pattern whilst it happens in

83 parallel to the development of the forebrain, midbrain, spinal cord and their arising substructures. In brief, four basic processes can be delineated in human cerebellar development: Firstly, the cerebellar primordium forms at the midbrain-hindbrain boundary (MHB) under close transcriptional influence of the isthmic organizer (IsO; for more information on this structure, see (334)). Subsequently, two different proliferative compartments are induced. These give rise to the two principal and distinct cerebellar cell type precursors, which will later migrate and/or differentiate to form the two dominating classes of neurons: inhibitory Purkinje cells and the excitatory granule cells; as well as all remaining glial, neuronal and interneuron subtypes residing the cerebellum (interneurons, unipolar brush cells, stellate cells, basket cells, Bergmann glia, etc.). In a third step, the aforementioned granule precursors migrate tangentially to form the external granule layer (EGL) where significant maturation takes place before they eventually migrate radially inwards to their ultimate residence in the internal granule layer (IGL). Finally, cell maturation and establishment of cellular connections characterise the last process giving rise to the three-layered cerebellar cortex architecture with its distinct circuitry and conserved foliation pattern that organises the cerebellum into 10 lobules (for more detailed information on cerebellar development, please see (335, 336) and Figure 1-17).

Figure 1-17: Cell migration during cerebellar development

Development of the cerebellum and corticogenesis. (A–B) Dorsal and lateral aspect of the mouse embryo at embryonic day E10-E11, showing outline of cerebellar primordium (indicated by arrows) and mesencephalon (midbrain) (m). (C) Schematic illustration of the developing cerebellum at about E11 indicates ventricular germinal zone (green) and genes that are involved in neurogenesis of GABAergic neurons such as Purkinje cells. The rhombic lip is a germinal zone that under control of genes such as Wls, Bmp, math1, pax6, and Lmx1a generates almost all glutamatergic neurons in cerebellum. Rhombic lip derived precursors produce cerebellar nuclei neurons (red), and external germinal zone precursors (orange). An arrow from mesencephalon (m) indicates a germinal zone for a group of cells derived from the mesencephalon to the cerebellar primordium. (D) A section of the cerebellum at around P4 indicate external germinal zone (EGZ) cells that after proliferation migrate through the Purkinje cells to the granular layer that is the location of granule cells. Abbreviations: Iso: isthmic organizer, m: mesencephalon, EGZ: external germinal zone, ML: molecular layer, PCL: Purkinje cell layer, GL: granular layer, WM: white matter, 4thV: fourth ventricle, r: rostral, c: caudal, d: dorsal, v: ventral. Figure and figure legend reproduced from (337).

Finally, it is noteworthy that the cerebellar cortex is connected to the remaining neuraxis only via the long projecting axons arising from the PC-somas in the central PC-layer. These exert modulatory inhibitory input to deep cerebellar nuclei (DCN), the key communicators in the wider cerebellar circuitry.

What is known about the development of Purkinje cells that might distinguish them from other cerebellar cells? Purkinje cells are not only the cerebellum’s key effectors but they are also thought to be primarily affected in most spinocerebellar ataxias (338) among other cerebellar conditions. Their pronounced sensitivity, complex morphology, and unique functional circuitry connecting them with neighbouring cells permits a carefully coordinated output, which integrates various temporal and spatial molecular events at the synapses

85 to long-term depression (LTD) or long-term potentiation (LTP) with precision and fidelity (336, 339, 340).

As mentioned above, their embryonic origin stems from neural progenitor cells of the cerebellar ventricular zone (341, 342), one of the two germinal layers of the cerebellar primordium. However, it has been intrinsically difficult for developmental biologists to investigate or influence Purkinje cell generation either in vivo or in vitro. Publication bias towards positive findings confounds an accurate and comprehensive analysis of past and current attempts at cerebellar neurogenesis, and it is likely that many failed attempts have not been published. However, one notable recent study demonstrates successful transplantation of cerebellar neural stem cells (NSC) into the cerebellum of SCA3-mutant mice with functional reorganisation (343). The report found significant reduction of mutant ATXN3 inclusions, degeneration and atrophy within the cellular layers of the cerebellum and specific preservation of the number of Purkinje cells post- transplant compared to the non-transplanted mutant mice. Notably this work failed to demonstrate the specification of Purkinje cells from the cerebellar NSC-graft in vivo and in vitro, thus suggesting an indirect beneficial effect of the graft on endogeneous Purkinje cell survival and cerebellar integrity. At the expense of Purkinje cells, the differentiation of TUJ1-positive neurons, GFAP-positive astrocytes and NG2-positive oligodendrocytes and their subsequent integration into the cerebellum was observed in vivo and in vitro in their mouse model. These exemplary results seem to be mirrored in some extent by the advances in the field of cerebellar cell differentiation from human iPSCs: whilst for years there has been no successful report of mature fully functional Purkinje cell differentiation from human iPSCs or embryonic stem cells (ESCs) until early 2015 (330, 344), the generation of MATH1+-cerebellar-like granule cells from hiPSCs had been reported to be possible several years earlier (326).

Modelling cerebellar disorders, especially spinocerebellar ataxias, with iPSC technology: Doom to bloom, or doomed to fail? The recent publication of Muguruma et al. (330) is a landmark study which imitates key developmental steps in cerebellar neurogenesis in order to achieve efficient directed differentiation of hiPSCs. In brief, the authors report regionalising ESC-derived embryoid bodies to the MHB in vitro and subsequent generation of self-organising cerebellar plate neuroepithelium (CPNE) that gives rise to mature, fully functional Purkinje and granule cells, as well DCN-neurons and various interneurons in specific coculture and FAC-sorting settings. By the precise timing of sequentially administered extrinsic morphogenetic signals (fibroblast growth factor 2 (FGF2), FGF19 and stromal cell-derived factor 1 (SDF1)) the team promoted self-formation of continuous CPNE

86 with “dorsal-ventral” and “apical-basal organisation”, mimicking cerebellar development for the first time in a dish (see Figure 1-18 for summarised schematic of devised protocol, and Table 1-3).

Figure 1-18: Muguruma et al. protocol for hESC-derived cerebellar cells

Schematic depicting timeline of administered extrinsic cues and concentrations, as well as FAC-sorting (FACS) and coculture (CC) settings needed to generate mature Purkinje and granule cells as reported in (330). Abbreviations: PC=Purkinje cells; DCN=deep cerebellar nuclei; DV=dorsoventral; CPNE: cerebellar plate neural epithelium; FACS: fluorescence activated cell sorting; SDF1: stromal cell-derived factor 1; FGF: fibroblast growth factor; SB: SB431542; ROCK-i.: ROCK-inhibitor; gfCDM: growth-factor-free chemically defined medium; cell specific markers are printed in red behind their cell types (printed in black) at the right side of the figure.

As described in more detail in Figure 1-18 the investigators managed to harness the so far limited knowledge of developmental cues required for cerebellar development in vitro by reducing them to a reductionist 3-D-culture with external administration of three extrinsic cues in a timely defined manner only. This proof of principle of cerebellar like development in vitro was achieved using hESCs, although the authors confirm that the same protocol applied to two human iPSC lines yielded very similar preliminary findings (330). This study will certainly inspire researchers to reproduce, modify (344) and further optimise the protocol. One important modification would be the development of a monolayer culture system in fully chemically defined medium at every step in order to avoid the uncontrolled and intrinsic cell-signalling promoted by 3D-morphology and spatial structure formation of cells. Besides the uncontrolled signalling cascades evoked in 3D-cultured cells, another disadvantage of 3D-structures is its inherent difficulty to establish functional live-cell imaging readouts on 3D-cell clusters. There is hope that, once a 2D monolayer system to generate Purkinje cells becomes available, more researchers will use this platform to further investigate 87 human cerebellar development and degeneration (see “Bottleneck” and what can be done for cerebellar diseases once it is successfully surpassed, Figure 1-16).

The future of iPSC technology in modelling cerebellar diseases If goal(s) mentioned above can be met, the future of modelling cerebellar diseases with iPSC technology will be bright. I am cautiously optimistic that further efficient and reproducible protocols (especially monolayer protocols) for the generation of Purkinje and granule cells from human iPSCs will be developed within the upcoming years using insights from developmental biology and also guided by recent breakthroughs (291, 330, 343). For the reasons stated above, this will be crucially important to promote this currently underrepresented field within human iPSC modelling of neurodegeneration. Advances in differentiation efficiency will permit accurate study of developmental and degenerative processes. These in turn will guide mechanistically rationalised therapies that ameliorate, halt or revert cerebellar diseases. Having lagged behind the main field previously might actually become a strategic advantage for this since important platforms for disease-modelling (for recent reviews, please see e.g (345)), drug-discovery (e.g. (346)) and evaluation of cell-replacement therapies (347) have been established in the meantime successfully for other iPSC-derived cell types. These can now be adapted to iPSC-derived cerebellar cells more quickly by building on this important knowledge and learning from previous mistakes. Successfully established and translated to cerebellar cells, these platforms will exert their full potential towards a better understanding of cerebellar development, disease-pathways and therapeutic avenues in cerebellar disease. Figure 1-19 provides a time bar summarising past efforts and gives future-oriented guidance and personal opinion on which genetic cerebellar diseases could be of special interest to model with iPSC technology.

Figure 1-19: Timeline modelling cerebellar disease using hiPSC

Modelling cerebellar disease with human iPSC technology: past, present and future. SCA3 and Friedreich’s Ataxia (FA), as the most common autosomal dominant and recessive cerebellar degenerative conditions, should be remodelled using Purkinje cells differentiated from patient-derived iPSCs once possible. Furthermore, preferential choice should target cerebellar diseases due to genetic defects in pathophysiologically widely linked genes, e.g. ITPR1 in SCA15 (339), and employ suitable cellular readouts with a focus on electrophysiology and live cell imaging to widen the knowledge about cerebellar diseases as “impaired network and impaired plasticity”-disorders. This figure has been published elsewhere, see (251).

Finally, progress in establishing accurate cerebellar in vitro disease models will - as it happened in other fields already (348) - be complemented by the generation of isogenic hiPSC lines via modern genomic editing techniques (349, 350). Such approaches allow introducing and/or ‘correcting’ different mutations that solely affect cerebellar integrity against an identical genetic background to precisely dissect disease causing mechanisms, culture-inherent factors and background genomic variation between different lines or to possibly study environmental and epigenetic factors contributing to ‘sporadic’ diseases. Finally, the gathered knowledge and progress towards efficient cerebellar differentiation protocols will impact on direct lineage reprogramming efforts (for recent reviews, please see (351, 352)) where so far no direct generation of Purkinje or granule cells from other fully differentiated nonneuronal cells has been reported in the literature.

To summarise this introductory subchapter, the cerebellum currently fails to obtain comparable experimental recognition as other neuraxial regions using hiPSCs. This is 89 largely attributed to the poor translation of its developmental logic into efficient directed differentiation strategies. Nonetheless, human iPSCs should represent a powerful as yet unrealised method to better understand cerebellar development and degeneration, highlighted by recent advances in this field (330). These should be taken forward to correct the current disequilibrium observed (see Figure 1-19) even though this might prove challenging and significantly time-consuming as will be reported in the last results chapter (Chapter 6) of this thesis.

Introductory final section With the background provided in the introduction, the aims of this thesis are to clinically characterise unexplored neurodegenerative patient cohorts, to genetically examine these in order to establish diagnoses, influence genetic counselling and family planning, identify disease causing mutations in novel and known genes, and investigate potential disease modifiers. Ultimately, iPSC technology allows functional characterisation of potential cellular defects in patient-derived neurons from a clinically and genetically well characterised group of SCA15-patients and prompts the developmental necessity of generation of novel cerebellar differentiation protocols from iPSCs.

2 Chapter 2: Methods

2.1 Methods Chapter 3

Sample collection With the help of various clinical collaborators from the IoN and the NHNN at Queen Square, and other world-wide based clinicians a cohort of sporadic and familial patients with an NBIA-like phenotype was collected. Some of the patients were recruited in the neurogenetics outpatients’ clinic at NHNN. For the remainder, hospitals in the UK and abroad were contacted by sending out collaboration emails and letters. When patients and collaborating physicians had consented to participate in this study, DNA samples and clinical information of suspected NBIA cases were collected. Patient inclusion criteria consisted of iron deposition on MRI and/or a clinical phenotype strongly suggesting an underlying NBIA disorder. Therefore required symptoms included (early- onset) movement disorder with dystonia of predominant oromandibular involvement and/or complex parkinsonism and/or developmental delay and/or psychiatric features or (in most cases) a combination of those. Patients where mutations in most common NBIA genes (and other genes responsible for common neurodegenerative diseases according to the individual clinical presentation) had been excluded previously in diagnostic labs and various research settings were included preferentially. The ethics committee of Queen Square had approved the study. All patients had given their consent prior to the study. DNA was readily available and already extracted via the diagnostic neurogenetics laboratory at NHNN for the majority of cases.

Clinical characterisation Clinical characterisation was carried out during the neurogenetics outpatients’ clinics and included clinical history taking, clinical examination, and individual presentation- based additional diagnostics (MRI, electrophysiology, electroencephalogram, blood tests, photography/videotaping) for cases recruited via the outpatients’ clinics. For the remainder of cases clinical characterisation was done according to retrospective study of clinical notes and or clinical information provided by collaborators where possible. Gender, ethnicity, age at onset and family history, as well as other notifiable characteristics if present were noted.

Symptoms were recorded in two ways: 1) Uncategorized descriptive summary of all noted findings with detailed timeline. 2) Dichotomic (present: yes/no) record of presence of following symptoms:

Parkinsonism, Dystonia, Intellectual disability/developmental delay, Cognitive decline/dementia, Psychiatric/behavioural disturbances, Myoclonus, Epilepsy, Chorea, Spasticity, Ataxia, Neuropathy, Tics, Retinopathy.

This approach allowed subclassification of patients according to family history, age at onset, predominant clinical phenotype and MRI findings for generation of smaller and more homogeneous subcohorts for subsequent analysis or variant filtering.

Genetics Whole exome sequencing was performed for 91 selected, clinically characterised idiopathic NBIA cases in collaboration with the group of Andrew Singleton from the Laboratory of Neurogenetics at the National Institute on Aging (NIA), National Institutes of Health at Bethesda, MD, USA (main collaborators: Monica Federoff, Dr. Raphael Gibbs, Dr. Jinhui Ding, Steve Clipman). In the initial phase of this PhD (January 2014- April 2014) three months were spent at the NIH to launch the collaboration, plan the experiment, provide patient DNA samples and learn the techniques of exome sequencing and data analysis.

For exome sequencing data generation, expanded exome oligos (EEX) were used within the recommendations of the Nextera Rapid Capture Enrichment (Illumina) protocol for the generation of paired-end reads which were then sequenced on the Illumina HiSeq 2000. Figure 2-1 and Figure 2-2 give a generic overview of the necessary steps in exome sequencing which will be described in more detail for the Nextera Rapid Capture Enrichment kit (see overview Figure 2-3) employed.

Figure 2-1: Exome sequencing workflow

Input DNA gets sheared and adaptors get ligated. Enrichment for coding parts (exons) by hybridisation to biotinylated DNA or RNA baits, pulldown, amplification, sequencing and bioinformatic analyses follow. Figure reproduced from (353).

Figure 2-2: Next-generation sequencing steps

Four designated steps with several substeps can be distinguished to obtain next- generation sequencing data: Library preparation, cluster amplification, sequencing and data analysis. Figure reproduced from www.illumina.com.

Figure 2-3 gives a detailed account of the Nextera Rapid Capture protocol used to generate the exome data that will be described in single steps below.

Figure 2-3: Nextera Rapid Capture protocol overview

Nextera Rapid Capture Enrichment Kit User flow-chart, figure reproduced from www.illumina.com. (See kit recommendations for abbreviations of consumables and plates).

The Nextera Rapid Capture Enrichment Kit sequencing protocol comprises several steps:

1) DNA library preparation and enrichment (consisting of 14 substeps as below)

1. Quantification: DNA library preparation with the Illumina Nextera Rapid Capture Enrichment Kit requires a minimum of 50 ng of genomic DNA (concentration: 5 ng/µl), quantified with the QubitTM fluorometric quantitation system (Thermo Fisher Scientific) and diluted accordingly. 2. gDNA tagmentation: Each sample gets tagmentated to oligonucleotide adapters and subsequently cleaned with Sample Purification Beads (SPB/SPBs). To these ends, 10 µl gDNA, 25 µl Tagment DNA Buffer and 15 µl Tagment DNA Enzyme 1 get mixed in a 96-well MIDI plate and placed on a microplate shaker at 1800 rpm for 1 minute, followed by centrifugation at 280 xg for 1 minute (Note: these are frequently recurring standard shaking and centrifugation conditions and will be referred to as such from now on). Subsequently, the plate is incubated at 58°C for 10 minutes before 15 µl of Stop Tagment Buffer is added to each sample followed by shaking and centrifuging at standard conditions. Finally, the plate is incubated at room temperature (RT) for 4 minutes. See Figure 2-4 for schematics of tagmentation process.

Figure 2-4: Tagmentation process

Tagmentation-based DNA-seq library construction: Genomic DNA is attacked by transposase homodimers loaded with synthetic, discontinuous oligos (yellow, purple) that allow for fragmentation and adaptor incorporation in a single step. Subsequent PCR appends outer flowcell-compatible primers (pink, green). Figure and legend reproduced from http://www.biotechniques.com/news/More-from-Less/biotechniques- 328886.html?service=print.

3. Tagmentated DNA clean up: (Note: This cleaning process is frequently recurring throughout the protocol and will therefore be described in detail here and referenced in later sections.) 65 µl of magnetic cleaning beads (also called SPB as above) are added to each sample in a deep well MIDI plate which is shaken and centrifuged at standard conditions subsequently. In the following, the plate is placed on the magnetic stand for ~2 minutes (until the liquid clears). All supernatant is then discarded and 200 µl of 80% ethanol is added to each well carefully without disturbance of magnetic beads. After 30 seconds, the ethanol is discarded and the cleaning process repeated once more to allow thorough cleaning. After the second wash, all ethanol must be thoroughly and carefully removed to avoid carry over. Beads must remain intact at this point. The plate is left to dry on the magnetic stand for 10 minutes. Next, 22.5 µl of resuspension buffer is added to each sample well, followed by shaking at standard conditions for 1 minute. In the following, the plate is incubated at RT for 2 minutes and subsequently centrifuged for 1 minute at standard conditions as described above. The plate is then put back on the magnetic stand for 2 minutes (or until liquid appears clear). Finally, 20 µl of clear supernatant is transferred from each well to a new standard 96 well plate.

4. First PCR Amplification: After each sample is tagged with two distinct series of indices, the first polymerase chain reaction (PCR) amplification is carried out. For this, PCR mixture contains of 5 µl Index 1 primer, 5 µl Index 2 primer, and 20 µl Nextera Library Amplification Mix (NLM) and 20 µl of cleaned sample obtained from the step above. The solution will be shaken and centrifuged at standard conditions described earlier. Next, the following programme is run on the thermal cycler: Pre-heat lid option (100°C) - 72°C for 3 minutes - 98°C for 30 seconds followed by 10 cycles of: 98°C for 10 seconds, 60°C for 30 seconds, 72°C for 30 seconds, 72°C for 5 minutes, final hold at 10°C.

5. First PCR clean up: Clean up steps are the same as described in 3: Tagmentated DNA clean up above with following difference: sample starting volume here is 50 µl and therefore 90 µl SPB are added to each well.

6. Quality check on Bioanalyzer: One µl of each sample is bioanalyzed using the Agilent DNA 1000 chip. Successful sample tagmentation should yield DNA fragments ranging from 150-1000bp in size.

7. First hybridisation: In this step the binding of each DNA library to biotinylated oligos (bates) is facilitated to enrich for certain parts of the sheared DNA fragments in the process. In preparation for hybridisation, each sample is pooled into a library of 12 samples. In order to obtain balanced amounts of DNA for each sample in the pool to ensure non-biased, well- covered libraries throughout, samples get quantified using the Qubit fluorometric quantitation system and 500 ng per sample are inputted to the pool (Speedvac concentrator or resuspension buffer to reduce or enlarge volume may be necessary to obtain a volume of 40 µl per sample). This prerequisite allows for the pooling of samples into a single library, and the potential to run 12 times the number of samples per lane, rendering the process time and cost effective. Next, 40 µl of each library pool is mixed with 10 µl Coding Exome Oligos and 50 µl Enrichment Hybridisation Buffer (total volume per sample is 100 µl). The plate is shaken on the microshaker followed by centrifugation (standard conditions). The samples are hybridized using the hybridisation programme NRC HYB on the thermal cycler (conditions: Pre-heat lid to 100°C - 95°C for 10 minutes - 18 cycles of 1 minute incubation, starting at 94°C, then decreasing 2°C per cycle - 58°C for >90 minutes but <24hours).

8. First capture: Post-hybridisation capture works with the help of SPBs used to separate genomic DNA-bait hybrids by binding to the biotinylated probes. The starting volume from the hybridisation is 100 µl. It is transferred completely to a deep well MIDI plate to start the SPB cleaning process. 250 µl SPBs are added to each well and shaken on the microplate shaker at a speed of 1200 rpm for 5 minutes. Next follows plate incubation at RT for 25 minutes and centrifugation at standard conditions after which the plate is placed on the magnetic stand for 2 minutes or until the liquid appears clear. Without disturbing the SPBs, all supernatant is discarded. Subsequently, the plate needs to be removed from the magnetic stand and 200 µl of Enrichment Wash Solution (EWS) is added to each well and the plate placed on the microplate shaker at standard conditions for 4 minutes. After this, the plate is incubated at 50°C for 30 minutes using a thermal cycler. The plate is then placed back on the magnetic stand for 2 minutes before all supernatant is discarded. To allow for increased purification of the target regions 200 µl EWS is added for a second wash as described above.

9. First elution: 28.5 µl Enrichment Elution Buffer 1 (EEB1) and 1.5 µl 2N NaOH are mixed into a premix, making 30 µl in total for each pool. 23 µl of this premix is added to each well/pool followed by 2 minutes shaking on the microshaker at standard conditions.

Incubation at RT for 2 minutes follows before the plate is centrifuged at standard conditions. Following centrifugation, the plate is placed on the magnetic stand until the liquid turns clear. 21 µl of the supernatant is added to 4 µl of Elute Target Buffer 2 (ETB2). The plate is shaken and centrifuged at standard conditions.

10. Second hybridisation: To further amplify and ensure high specificity of the capture regions, a second hybridisation step is required, lasting a minimum of 14.5 hours and a maximum of 24 hours. The second hybridisation is analogous to the first hybridisation (step 7, but without library pooling). Additionally (since the starting volume of each pool is 25 µl after the first elution from the first hybridisation and capture), 15 µl resuspension buffer is added to the final solution to make a total of 100 µl.

11. Second capture: As described in 8 (First capture), samples are thoroughly captured analogous to the steps following the first hybridisation.

12. Capture sample clean up: In this step, samples are cleaned before the final enrichment. With a starting volume of 45 µl SPBs, the capture sample clean up follows the steps as described in 3 and 5 (Tagmentation DNA clean up and PCR clean up).

13. Second PCR amplification: Finally, an additional PCR amplification step is performed to maximally enrich the library prior to clustering. 20 µl Nextera Enrichment Amplification Mix and 5 µl PCR Primer Cocktail are added to the 25 µl of each pool. The plate is placed on the microshaker, then centrifuged under standard conditions and subsequently the NEM

AMP10 programme on the thermal cycler is started (NEM AMP10 programme: pre- heat lid to 100°C - 98°C for 30 seconds - 12 cycles of: 98°C for 10 seconds - 60°C for 30 seconds - 72°C for 30 seconds -- 72°C for 5 minutes – final hold at 10°C).

14. Second PCR clean up: Samples are cleaned up as described previously (steps 3, 5, and 12) with 90 µl SPBs at the start. Quantification on the bioanalyzer using an Agilent DNA High Sensitivity chip follows prior to the clustering phase.

2) DNA amplification and clustering on the C-bot Special glass microfabricated devices called flow cells whose technique and makeup allow for the cluster generation on the automatic cluster generator termed C-bot, are essential in this second step of the WES protocol. Each library pool consisting of 12 libraries at this point is run on a single lane within the eight channels located inside each flow cell. DNA polymerases initiate the cluster generation process by amplifying DNA fragments via bridge formation (see Figure 2-2, upper right panel). This ultimately produces millions of DNA clusters. Within each distinct cluster roughly 1 million copies of the original fragment get generated, a quantity required for accurate signal fluorescence and fluorescence detection during the high-throughput sequencing process on the Illumina HiSeq 2000.

3) Parallel sequencing by synthesis (SBS) on the Illumina HiSeq 2000 Massive parallel SBS takes place in the third step of the WES process - in this study libraries were sequenced on the Illumina HiSeq 2000. In brief, all four nucleotides (nt) get fluorescently labeled with a unique colour and are randomly incorporated into the oligo-primed cluster fragments on the flow cell during the sequencing process. First, DNA linearisation happens via the cleavage of a single adaptor followed by denaturation. This process yields single-stranded DNA to which sequencing primers are added in combination with four reversible terminators, unique to each nucleotide. Upon the addition of a new nucleotide via the DNA polymerase (sequencing by synthesis), each base has a chemically blocked 3-prime hydroxyl (-OH) group where at this moment no further polymerase activity and thereby no chain elongation can happen. At this stage, the optic lens captures an image after every addition of a fluorescently labeled nucleotide. Each flow cell lane gets imaged in 3 distinct 100-tile segments at an approximate cluster density of 30000 clusters per tile. Upon successful capture of the image, the 3-prime blocking OH-group on the chain gets chemically removed which naturally allows for incorporation of the subsequent complementary base. The process is repeated for approximately 200 cycles in total with final read lengths in the 50-100 base pair range. The sequencing process takes place on both single-strands of the DNA, creating a paired-end read to facilitate accuracy when mapping data downstream. The overall run time on the Illumina HiSeq 2000 consists of approximately 10 days.

4) Bioinformatic analyses Obtained sample fastq-files are retrieved from the sequencer, trimmed based upon quality over read cycles, aligned using hg19 as genome reference file, and written into binary format using sam- (http://samtools.sourceforge.net/) and bamtools

100

(https://sourceforge.net/projects/bamtools/) for each sample. Obtained bam-files were indexed and a local realignment around small insertions and deletions (so called ‘indels’) as well as recalibration of quality scores was performed as recommended by Broad Institute’s best practices paper (354).

The GATK toolbox with implemented HaplotypeCaller as developed by the Broad Institute was used for variant calling and produced genomic Variant Call Format files (gVCFs) for each sample. In brief, this programme assigns the likely genotypes to the sample for each potential variant site by employing Bayes' rule using the likelihoods of alleles given the read data, to calculate the likelihoods of each genotype per sample given the read data observed for that sample. For more details, see: https://www.broadinstitute.org/gatk/gatkdocs/org_broadinstitute_gatk_tools_walkers_h aplotypecaller_HaplotypeCaller.php.

Sample gVCFs were combined per cohort into a group gVCF using vcftools (http://vcftools.sourceforge.net/).Applied quality control (QC) steps on the group gVCF included determination and exclusion of call rate outliers, samples failing gender check, samples with outliers from heterozygosity rate test either for biological reasons, or other reasons such as contaminations or samples that were suspicious of duplication or undisclosed high relatedness due to too many variants shared.

Using the terminal and developed batch scripts employing awk language, quality checked group VCFs were queried for nonsynonymous, splice site, nonsense and frameshift variants in known genes causing NBIA-like disorders, dystonia syndromes, parkinsonism, spasticity and other neurodegenerative diseases with common symptom overlap in the described cohort (see Table 2-1 for list of genes examined; see

Appendix, Section 4: “Bioinformatics” for details on bioinformatic analyses).

Note that for some genes (e.g. PLA2G6, FA2H) disease categorisation is somewhat arbitrary due to overlapping phenotypes and causality for a broader range of disease (HSP, NBIA, dystonia-parkinsonism). Genes were selected exhaustively from the literature (and for relatively novel genes from review processes or word of mouth (conferences, abstracts, etc.)) describing implicated genes for the respective diseases. Obtained variants within these genes were annotated using ANNOVAR via the command line with a virtual machine on the allocated space of the neurogenetics share of the server ‘Kronos’ at IoN (see Appendix, Section 4: “Bioinformatics” for details on bioinformatic analyses).

101

Variants within genes previously described for each disease category (NBIA, dystonia- parkinsonism, HSP, broader neurodegeneration; see Table 2-1) were separated by recessive, dominant or X-linked implication and checked against the presumed mode of inheritance in the respective family to facilitate downstream analysis. Variants were manually looked through, and further filtered according to previous description in the literature and frequency in annotated control databases (Exome Aggregation Consortium (ExAC), database of short genetic variation (dbSNP), 1000 Genomes project and NHLBI Exome Variant Server (NHLBI-EVS)), hereby accounting for inheritance mode and gene in question. As a general rule, known variants in dominant genes with frequencies > 0.0001 unless previously described as pathogenic or disease-implicated were discarded; for recessive genes variants with a frequency > 0.01 were removed unless interesting or implicated previously.

Table 2-1: Candidate genes investigated

WDR45 TOR1A ATL1 MAPT RAB39B TUBB4A SPAST GRN PANK2 GCH1 NIPA1 CSF1R PLA2G6 THAP1 KIAA0196 Dementia/ SLCA3 c19orf12 PNKD KIF5A Neuropathy/ PSEN-2 Dystonia FA2H SLC2A1 Hereditary RTN2 Leuko- APP NBIA- - ATP13A2 PRRT2 Spastic HSPD1 dystrophy/ PSEN-1 like Parkin- Paraplegia Other neuro- CP sonism SGCE BSCL2 TREM2 DCAF17 ATP1A3 REEP1 degenerative SCN9A COASY CIZ1 ZFYVE27 diseases WDR73 ANO3 SLC33A1 SCP2 GNAL CYP7B1 FTL CHRNA4 SPG7 GTPBP2 ATP6AP2 SLC16A2 SLC20A2 SPG11

NKX2-1 ZFYVE26 DRD5 ERLIN2 CHMP2B SPG20 ATN1 SPG21 ACTB DDHD1 CACNA1A KIF1A CACNB4 FA2H COL4A1 PNPLA6 MAPT c19orf12 PDE10A GJC2 NDUFA10 GBA2 SDHA AP4B1 NDUFS4 KIAA0415 NDUFAF2 TECPR2 NDUFA2 AP4M1 102

DLD AP4E1 C8orf38 AP4S1 SURF1 VPS37A COX15 DDHD2 NDUFS3 C12orf65 NDUFS8 CYP2U1 FOXRED1 ALS2 NDUFA9 KIF1C NDUFA12 USP8 NDUFS7 WDR48 ATXN3 ARL6IP1 PRKCG ERLIN1 TBP AMPD2 JPH3 ENTPD1 VPS13A NT5C2 VPS13B ARSI TAF1 PGAP1 TIMM8A FLRT1 XK RAB3GAP2 ARX MARS MECP2 ZFR MT-ND1 BICD2 MT-ND4 REEP2 MT-ND3 MAG MT-ND6 ARG1 PRKRA AFG3L2 TREX1 DDHD3 ATM DDHD4 SLC19A3 L1CAM CLN3 PLP1 SPR ETHE1 FUCA1 GCDH GLB1 GAMT HPRT SCP2 ARSA C2orf25 SERAC1 SDHAF1 NPC1 NPC2 SLC6A3 PRKN DJ1 LRRK2 PLP1 SEPSECS

103

ALDH7A1 TH AP4B1 ALDH5A1 HEXA ATP7B DDC MUT SNCA LRRK2 VPS35 EIF4G1 DNAJC13 UCHL1 CHCHD2 PARK2 PARK7 DJ1 PINK1 COQ2 FBXO7 DNAJC6 SYNJ1 GIGYF2

Selected candidate genes implicated in diseases with highest clinical overlap to the observed clinical symptoms in the described cohort. Variants observed in these genes that were present in the cohort were further analysed.

Sanger sequencing (as described in Appendix, Section 2: “Sanger Sequencing Workflow” and in more detail below) was used to confirm variants found (see Appendix, Section 1 and Section 2 for primers, pipetting schemes and PCR cycling conditions). Where accessible, DNA from further family members was collected and segregation analysis performed.

Briefly, Sanger sequencing involves a first PCR where per sample PCR-mastermix, water, respective primers and DNA are mixed in a standard 96-well plate, briefly centrifuged and put on a thermal cycler (see Appendix, Section 2 “Sanger Sequencing Workflow” for respective pipetting volumes, reagents and cycling conditions). In a second step, the product is checked for its expected size on an agarose gel using the principle of gelelectrophoresis (3 µl of PCR product and 3 µl of loading dye get mixed and loaded into the agarose gel next to a standard DNA ladder, 100 V gets applied for 20 mins which makes the PCR fragments move at different speeds according to their size. The gel is then visualised using UV light). If correct size of the amplicon is

104 confirmed, the remaining PCR product is cleaned up using the ExoSap method as detailed in Appendix, Section 2. The cleaned PCR product undergoes sequencing PCR as detailed in Appendix, Section 2 followed by cleaning via the Sephadex method (again detailed in the Appendix, Section 2) and sequencing on a 3730 DNA analyser. Obtained chromatograms are analysed using the demo version of Sequencher 5.1, Gene Codes Corporation, Ann Arbor, MI.

2.2 Methods Chapter 4

2.2.1 Methods Chapter 4.1

Patient examination included anamnesis, clinical and neurological investigation, blood testing, echocardiogram, electrocardiogram, electroencephalogram, neuropsychiatric testing, brain MRI and nerve conduction studies (Greece). DNA was extracted from peripheral blood (Greece) of the affected proband, his son and his daughter and sent to IoN, London where I performed molecular genetic studies, including Sanger sequencing as described in Methods Chapter 3 above and in the Appendix, Section 2. Primer sequences, pipetting scheme and cycling conditions for PCR and Sanger sequencing of XK can be found in the Appendix, Section 1, Chapter 4, “XK gene“, and Appendix, Section 2, “Sanger Sequencing Workflow” and cycling programmes. Primers were designed using the online tool Primer 3 (http://primer3.ut.ee/). Due to lack of further biomaterial (e.g. skin fibroblasts, or fresh blood), no confirmatory functional analyses assessing the functional consequences of the deletion could be performed.

2.2.2 Methods Chapter 4.2

Patients All patients were recruited through the Neurogenetics outpatients’ clinic at the National Hospital for Neurology and Neurosurgery, Queen Square, London and gave written consent. Patients all had a diagnosis of progressive cerebellar ataxia with either known autosomal recessive or presumed sporadic inheritance with a relatively early onset of disease (here: < 35 years). In total 196 patients were screened for SYNE1 mutations through either exome sequencing (110 patients) or targeted next-generation sequencing (86 patients).

Genetic Analysis DNA was extracted from peripheral leucocytes of all patients in the diagnostic lab, using standard procedures. Additional samples were taken from affected or unaffected relatives to test mutation segregation where appropriate.

105

Exome sequencing libraries were prepared using Illumina Nextera Rapid Capture Exome Kits following the manufacturer’s recommendation and as described in more detail in Methods Chapter 3 above. Libraries were indexed and sequenced on an Illumina HiSeq 2500 machine.

A custom sequencing panel was designed to amplify the coding exons of SYNE1 using the Illumina Truseq Custom Amplicon v1.5 Kits. Libraries were prepared in keeping with the standard recommended protocol and then sequenced on an Illumina MiSeq machine. The custom sequencing panel was designed and ran by my colleague Dr. J. Hersheson, therefore the interested reader has to be referred to his thesis for a detailed account of the methods for the custom panel sequencing process employed.

Bioinformatic analysis was the same for both exome sequencing and targeted next- generation sequencing. Reads were aligned to the hg19 genome build using Novoalign with variant calling performed using SAMtools and Genome Analysis Toolkit Best Practices (GATK, Broad Institute). Variant annotation was performed with ANNOVAR and coverage metrics were investigated using a modified inhouse Bedtools coverageBed script. For more details on bioinformatic analysis, see Methods Chapter 3 subsection as well as Appendix, Section 4, Bioinformatics. All SYNE1 annotations and mutation locations given below are for the refseq NM_033071 transcript (ENST00000423061).

The final list of called variants in SYNE1 was filtered according to the following criteria: 1) nonsynonymous variants present in a homozygous or compound heterozygous state only, 2) quality > 30, 3) depth > 10,

4) minor allele frequency in Exome Variant Server, ExAC and 1000 Genomes databases <0.005.

Identified variants were confirmed using Sanger sequencing in affected cases and also in parents or unaffected siblings where available to confirm segregation or mutation phase in compound heterozygous mutations. Sanger sequencing was performed as described in Methods Chapter 3 above. Respective primer sequences, protocols, pipetting schemes and cycling conditions for PCR and Sanger sequencing can be found in Appendix under Section 1, Chapter 4, SYNE1 gene and Section 2.

106

2.2.3 Patients and Methods Chapter 4.3

Patients The patients were from a large, multigenerational consanguineous family of Zorastrians (Parsis) of Indian origin with two affected cousins, both born to consanguineous first- cousin marriages (see Figure 4-8). The Zoroastrians (Parsis) are recognised as a small endogamous and closed community, distinct from the rest of Indian cultural subgroups. They have maintained a separate identity by not accepting conversions and not permitting intermarriage. They are followers of the pre-Christian prophet Zoroaster/Zarathustra and fled from Pars (Fars) in Iran/former Persia to the west coast of India after the downfall of their Sassanian Empire, around the 8th century A.D.. The Iranis are co-religionists who migrated later around the 19th century leaving behind a small number in Iran who still practice their religion. Most Iranis have settled down in Mumbai (Bombay) and Parsis are a declining community with an estimated population of 71630 individuals in India in 1990. Recent data from UNESCO’s Parzor project suggests that there are only about 60000 Parsis left in India today (http://pulitzercenter.org/reporting/asia-india-government-jiyo-parsi-population accessed: March 2016).

The index patient was first seen 55 years ago by Dr. N. H. Wadia (now Director Emeritus at the Department of Neurology, Jaslok Hospital, Mumbai, India), and was diagnosed as having spinocerebellar ataxia. The patient underwent periodic clinical follow-up. At the age of 57 in 1999, he was seen at the Department of Assisted Reproduction and Genetics, Jaslok Hospital, Mumbai, together with his affected cousin. Here, thorough neurological and general medical examination on both affected cousins was performed by Dr. Desai. Pedigree analysis, karyotyping and chromosome breakage study were carried out, genetic counselling given and further testing was coordinated. DNA analysis was carried out for SCAs 1, 2, 3, 6, 7, DRPLA, Friedreich’s ataxia and Huntington’s disease in the Department of Molecular Medicine and Biology, Jaslok Hospital, Mumbai. Fragile-X testing by Southern blot was performed at the University of Delhi South Campus (UDSC) in 2005. For targeted DNA mutation analysis of a number of rare Mendelian diseases, saliva was sent to Counsyl, USA, while the IlluminaHumanCytoSNP-12 v2.1 assay was carried out at the Centre for Genomic Application (TCGA), Delhi, in 2010. Despite these extensive tests, the genetic diagnosis could not be reached for many years.

107

With informed consent, DNA and fresh blood samples of the index patient, the affected female cousin, and 20 family members were sent to the Institute of Genomics and Integrated Biology (IGBI, Delhi) and subsequently to the Institute of Neurology, London, where less common SCA11 and 15 were excluded prior to this study.

No other biomaterial (fibroblasts, CSF, autopsy) was available. All participants had given their written consent. The study was approved by the joint ethics committee of UCL Institute of Neurology and the National Hospital for Neurology and Neurosurgery, London, UK.

Homozygosity mapping, whole exome sequencing, Sanger sequencing and in silico predictions DNA and RNA were extracted from blood using standardized protocols similar to the ones described for DNA and RNA extraction detailed in Appendix, Section 5 and 6. In order to determine the genetic cause of the disease in this study, homozygosity mapping using SNP-array data and the online-tool “homozygosity mapper” (http://www.homozygositymapper.org/) was performed for the two affected cousins, and several unaffected family members. Genes in all stretches of homozygosity > 1Mb were cross-queried with a list of inherited neurodegenerative conditions. Subsequently, WES was carried out as a company service in the index case using NimbleGen SeqCap Target Enrichment EZ-system (Roche Sequencing) and Illumina paired End Genomic DNA-Prep Kit for library preparation, amplification and sequencing on a GAII platform. The exome sequencing in the index case and subsequent file generation with provision of the VCF file was outsourced and done as part of a commercial service by a company. The exome was aligned using Burrows-Wheeler Aligner (BWA) and variants called according to GATK best practices (https://www.broadinstitute.org/gatk/guide/best-practices.php). The provided files were analysed by prioritising homozygous non-synonymous variants in genes within the homozygous stretches, but other known genetic causes of cerebellar ataxia outside of the homozygous regions were additionally excluded. Sanger sequencing was performed as described in Methods Chapter 3 with primer sequences and cycling conditions listed in the Appendix, Section 1 and 2 to confirm variants, to test segregation within the family, to rule out additional variants within the target gene that were missed by exome sequencing, and to screen an additional cohort of 40 British pure cerebellar ataxia patients. To rule out possible effects of variants on splicing, RNA was extracted from blood for both affected cousins and their siblings, complementary DNA (cDNA) created and subsequently Sanger sequenced.

108

2.3 Methods Chapter 5

2.3.1 Methods Chapter 5.1

Patients Subject cohorts were gathered from the following different sites: the Neurogenetics Unit of the National Hospital for Neurology and Neurosurgery (London, UK), TRACK- HD (Europe) (355), SPATAX network (France), the University of Athens Medical School/Eginition Hospital (Athens, Greece), the National Institute of Neurology and Neurosurgery, Manuel Velasco Suarez (Mexico), and the University of Azores (Ponta Delgada, Portugal) (see Table 2-2).

All subjects with polyglutamine diseases seen at any of the collaborators’ sites that agreed to participate in research were enrolled without preselection of CAG repeat size or AAO. All studies were approved by local ethics committees and all subjects gave written informed consent. For this study samples and data for HD and SCAs 1, 2, 3, 6, 7, and 17 were gathered; however, very few DRPLA (n=4) and SBMA (n=8) samples were available without detailed enough clinical characterisation so these diseases were not included.

In total, AAO and CAG repeat size was available for 1462 patients (see Table 2-2). Given the varied phenotypes of polyglutamine diseases, motor onset (HD) or the onset of the first progressive symptom as reported by the patient (e.g. ataxia, visual disturbance, myoclonus) was used to determine AAO throughout all cohorts. Given the small number of patients with age at onset and CAG repeat size available (n=7), SCA17 was only considered in the combined SCA analysis.

109

Table 2-2: Cohort characteristics

Disease Cohort SCA HD SCA1 SCA2 SCA3 SCA6 SCA17 Total 7 Athens, Greece 351 0 0 0 0 0 0 351 Azores, Portugal 0 0 0 91 0 0 0 91 London, UK 0 30 66 45 69 7 1 218 Mexico 0 0 113 0 0 66 6 185 Paris, France 0 147 115 261 0 0 0 523 TRACK-HD, 94 0 0 0 0 0 0 94 Europe Total 445 177 294 397 69 73 7 1462

%M 49.4 54.2 48.6* 52.6 60.9 56.2 85.7 51.8*

45 ± 37 ± 33 ± 39 ± 57 ± 35 ± Mean AAO ± SD 12.1 30 ± 13.4 10.5 12.9 11.6 10.5 17.6 (range) (6- (8-44) (16-65) (8-73) (9-74) (18-76) (5-84) 82) 44 ± 48 ± Mean (CAG)n 48 ± 42 ± 71 ± 22 ± 5.0 11.1 51 ± 6.4 length ± SD 5.3 4.5 4.4 0.9 (37- (36- (42-58) (range) (39-66) (33-64) (50-82) (21-26) 92) 100)

Abbreviations: HD – Huntington’s disease; SCA – spinocerebellar ataxia; %M – percentage of males; AAO – age at onset; SD – standard deviation. *one subject had no gender information. This table is published elsewhere (2).

SNP selection criteria and genotyping SNPs were selected from the most significant genes (gene-wide p<0.1) in the “DNA repair pathway cluster” from the GeM-HD analysis (listed in Table S4 of GeM-HD paper) (356). Additionally, SNPs from RRM2B and UBR5 were added to this list since they are both members of GO:0006281 “DNA Repair” (which, although nominally significant in GeM-HD GWAS, did not reach q<0.05 and was therefore not used to create the pathway cluster), both lie within a genome-wide significant association peak in GeM-HD GWAS, and both have significant gene-wide p-values (see Table S5 of the GeM-HD paper) (356). For each gene, the most significant SNP was selected, along with a small number of proxy SNPs in close LD (r2>0.8) with the most significant SNP that also showed association in GeM-HD GWAS. Where possible, these proxy SNPs were chosen to have functional annotation available

110

(http://browser.1000genomes.org/index.html: accessed 12/6/14). If a gene contained two independent significant signals in GeM-HD GWAS (for example, FAN1), then the lead SNP for the second signal was included. Note that this chosen selection procedure is not intended to give comprehensive coverage of the genes in question, but instead to highlight SNPs likely to be disease relevant. To guard against the effects of population stratification, SNPs were removed from the analysis if they had a Hardy- Weinberg p-value <0.001 in the whole dataset. Procedures detailed above yielded 22 genotyped SNPs with success rates ranging from 94.2-98%. Table 2-3 gives an overview of the genotyped SNPs in this study.

SNP genotyping was performed using custom KASP assays at LGC Genomics (Hertfordshire, UK). Gene level sense sequences were used to design SNP assays (see Appendix, Section 3, Table 8-20). The assays for several SNPs were designed in reverse orientation to the chromosome (rs4150407, rs1805323, rs1037700, rs1037699, rs3512, and rs20579; assay sequences highlighted in red in the table). For this reason, for all SNPs in reverse orientation to the chromosome (rs4150407, rs1805323, rs1037700, rs1037699, rs3512, and rs20579) genotypes resulting from these KASP assays will be complementary to those using HGVS nomenclature. This is reflected in Appendix, Section 3, Table 8-21, where the minor allele for these SNPs differs from GeM-HD (356), but corresponds to the same allele.

Table 2-3: Characteristics of genotyped SNPs

Chr: Geno- position Functional P typing P SNPs (bp) Gene MAF* annotation (GeM-HD) success (HWE)* (GRCh37/hg rate* 19) rs1800937 2:48025764 MSH6 Stop_gained 4.30E-03 0.074 0.973 0.840 2:12804963 rs4150407 1 ERCC3 Intron_variant 4.60E-04 0.479 0.964 0.003 2:19064931 NMD_transcript rs5742933 6 PMS1 _variant 9.49E-04 0.205 0.972 1.000 Missense rs1799977 3:37053568 MLH1 _variant 7.16E-07 0.28 0.966 0.354 rs6151792 5:80056961 MSH3 Intron_variant 2.09E-04 0.117 0.978 0.706 rs115109737 5:80102444 MSH3 Intron_variant 4.50E-04 0.041 0.980 0.489 rs71636247 5:80118976 MSH3 Intron_variant 2.55E-04 0.034 0.976 1.000 Missense rs1805323 7:6026942 PMS2 _variant 3.04E-02 0.043 0.975 0.736

111

rs12531179 7:6028687 PMS2 Intron_variant 3.84E-05 0.169 0.971 0.925 8:10321769 rs3735721 5 RRM2B 3’_UTR_variant 5.68E-07 0.083 0.971 0.058 8:10325077 rs1037700 5 RRM2B Intron_variant 5.03E-08 0.094 0.973 0.002 8:10325083 Frameshift rs5893603 9 RRM2B _variant 4.28E-08 0.093 0.973 0.007 8:10325093 Missense rs1037699 0 RRM2B _variant 2.70E-08 0.094 0.976 0.002 8:10330603 Synonymous rs16869352 3 UBR5 _variant 4.01E-07 0.08 0.975 0.030 8:10331115 Synonymous rs61752302 3 UBR5 _variant 3.03E-03 0.026 0.977 0.621 14:7549505 rs72734283 9 MLH3 Intron_variant 4.32E-03 0.089 0.971 0.623 14:7551382 Missense rs175080 8 MLH3 _variant 7.72E-03 0.435 0.971 0.447 15:3112640 rs146353869 1 FAN1 Intron_variant 4.30E-20 0.017 0.973 1.000 15:3119797 Synonymous rs114136100 6 FAN1 _variant 8.49E-16 0.019 0.976 0.423 15:3120296 Missense rs150393409 1 FAN1 _variant 9.34E-18 0.013 0.975 1.000 15:3123500 rs3512 5 FAN1 3’_UTR_variant 5.28E-13 0.283 0.973 1.000 19:4866883 NMD_transcript rs20579 0 LIG1 _variant 6.65E-03 0.134 0.942 0.732

SNPs were selected from the most significant genes (gene-wide p<0.1) in the “DNA repair pathway cluster” from the GeM-HD GWAS (listed in Table S4 of GeM-HD paper). Genes annotated by the SNPs are indicated. *Refers to the current study. Abbreviations: Chr =chromosome; MAF=minor allele frequency; HWE=Hardy–Weinberg equilibrium; FAN1=FANCD2/FANCI-Associated Nuclease 1; ERCC3=Excision Repair Cross- Complementation Group 3; RRM2B=Ribonucleotide Reductase M2 B (TP53 inducible); UBR5=Ubiquitin Protein Ligase E3 Component N-Recognin 5; MLH3=MutL Homolog 3; MSH3=MutS Homolog 3; MSH6=MutS homolog 6; PMS1=Postmeiotic Segregation Increased 1 (S. Cerevisiae); PMS2=PMS2 Postmeiotic Segregation Increased 2 (S. Cerevisiae); MLH1=MutL Homolog 1; LIG1=Ligase I, DNA, ATP-dependent. This table is published elsewhere (2).

112

Statistical analyses Ages at onset for the various diseases were corrected for repeat length using a similar method to the GeM-HD GWAS (356). A linear regression was performed for each disease separately of ln(AAO) on expanded repeat length. See Table 2-4 for the regression parameters. These parameters were used to construct an expected value of AAO for each individual, based on their repeat length, which was subtracted from their actual AAO to give a residual. Association of each SNP with AAO was tested by performing a linear regression of these residuals on the number of minor alleles in the genotype in PLINK (357). The effect of gender on AAO (after accounting for CAG length) was also tested. Since this was non-significant for all disorders (results not shown), gender was not included in the calculation of residuals. The primary analysis in this report tested whether there was an overall association of AAO across all 22 SNPs. This was done by combining the association p-values for each SNP using Brown’s method (358). Essentially, this is Fisher’s method for combining p-values corrected for linkage disequilibrium between SNPs. The primary analysis used one-sided p-values for association in the same direction as that observed in the GeM-HD GWAS. In order to assess the overall directionality of the associations, the significance to that obtained from a similar analysis using two-sided p-values was compared. The analyses were performed on eight disease groups: all polyglutamine diseases (HD+SCAs), HD, all SCAs, SCA1, SCA2, SCA3, SCA6 and SCA7. P-values were Bonferroni corrected for eight tests – this is conservative since the disease groups are not independent. Individual SNPs significantly associated with AAO in each disease group were also noted. Due to small sample size (n=7), SCA17 was not analyzed independently, but was included in the analyses of disease groups ‘all SCAs’ and ‘HD+SCAs’.

113

Table 2-4: Effects of repeat length of expanded allele on AAO

Disease Sample N A B P HD 445 6.119939 -0.052966 <2e-16 SCA1 177 5.682974 -0.043694 <2e-16 SCA2 294 5.799343 -0.056682 <2e-16 SCA3 397 7.137211 -0.049477 <2e-16 SCA6 69 5.96740 -0.08686 0.00268 SCA7 73 4.643231 -0.026023 2.94e-5 SCA17 7 2.38659 0.01716 0.70

Results of fitting a linear regression ln(AAO) = A + B*(CAG)n. P-value refers to the significance of the regression parameter (B) indexing the effect of repeat length. Abbreviations: HD – Huntington’s disease; SCA – spinocerebellar ataxia. This table is published elsewhere (2).

2.3.2 Methods Chapter 5.2

In total, DNA from 173 individuals with SCA6 due to expanded repeats in CACNA1A was available. The SCA6 repeats in these 173 samples were Sanger sequenced with primers and conditions as indicated in Appendix, Section 1 and 2. The process of Sanger sequencing was identical as described in Methods Chapter 3.

Repeat sizes for the normal and expanded allele were determined by thorough analysis and manual counting of visualised repeats on forward and reverse sequencing in BioEdit v7.2.5 (Tom Hall, Ibis Biosciences, Carlsbad, CA 92008) and Sequencher (demo version of Sequencher 5.1, Gene Codes Corporation, Ann Arbor, MI). Normal and expanded allele were read separately and obtained sequence was compared to the CACNA1A wildtype sequence in order to investigate the possibility of presence of repeat interruptions. When a discrepancy for the repeat count between forward and reverse existed, the longer repeat was taken consistently.

Additional AAO and gender was available for 65 individuals and correlation with AAO and expanded allele as well as sum of normal and expanded allele was calculated using linear regression analysis in excel.

114

2.4 Methods Chapter 6

2.4.1 Methods Chapter 6.1

Three patients from two different families with ITPR1 deletions were clinically characterised. Skin biopsies were taken and two iPSC clones per patient (6 patient- derived clones in total – these will appear as CT1, CT2, ST1, ST2, MD3 and MD5 in subsequent text and figures) were generated and validated. The patient-derived iPSCs were validated alongside to control iPSCs and ESCs and differentiated into cortical neurons to investigate them using a) immunocytochemistry (ICC) (day in vitro (DIV)15, DIV34, DIV80, DIV100 – in order to validate the model and investigate a pathomechanistically presumed haploinsufficiency of ITPR1), b) live cell imaging (DIV60-90 - in order to assess potential calcium handling defects in ITPR1 mutant cells), c) electrophysiology (DIV80-100 – in order to assess their maturity and functionality as neurons) and d) RNA-sequencing (DIV80 – in order to identify differentially transcribed genes and isoforms contributing to pathogenicity). Please see schematic in Figure 2-5 for timeline of experiments performed in this study. The three control lines (two iPSC lines and one embryonic stem cell line) were obtained from various collaborators (see Table 8-22 for more information on the control lines). Where necessary they were validated alongside the self-generated clones.

Figure 2-5: Timeline of experiments

Timeline of experiments for the generation and study of human neuronal model of spinocerebellar ataxia type 15.

115

Skin biopsy, fibroblast culture and generation of iPSC lines Skin biopsies were taken from all three consented individuals. The biopsy area on the upper outer arm was cleaned aseptically and 1 - 2% lignocaine was injected subcutaneously and intra-dermally as local anaesthetic. Using a punch-biopsy-set and a medical scalpel, a 3-6 mm sized punch was carefully excised. The tissue was transferred to a flask holding biopsy medium immediately (Gibco® Roswell Park Memorial Institute medium (RPMI) 1640 (Life Technologies, 11875101) with 10% fetal bovine serum (FBS, CorningTM, 11-648-647) and 0.2% penicillin-streptomycin solution (Life Technologies, 15140-122)) and appropriate wound management was applied.

The biopsies were chopped under the hood and transferred into in vitro dishes where fibroblasts start growing out of the initial pieces of skin after several weeks (technicians’ facilities of Department of Molecular Neuroscience, IoN). These were passaged, RNA was harvested and fibroblasts were banked at several low-passage numbers individually as well as for the local and international biobank resource collaborations. Fibroblasts were cultured in Gibco® OptiMEM (Life Technologies, 31985062) plus 10% FBS (CorningTM, 11-648-647) medium. The following episomal plasmids published in 2011 (235) and originally obtained from Addgene were used for transfection: pCXLE hOct4 shp53, pCXLE hSK and pCXLE hUL (combination Y4 of the original Okita et al. paper (235)). Per line ~106 cells of early passage fibroblasts were counted and resuspended in nucleofector solution (Cell Line NucleofectorTM Kit, Lonza, VCA-1001). 1 µg per plasmid of the three plasmids described above was added per reaction (3 µg total) and cells were transfected using programme U-023 on a NucleofectorTM I device (Lonza, AAB-1001). Transfected cells were maintained in fibroblast media, media exchanged 24 hours post-electroporation and every third day subsequent. After one week, transfected cells were trypsinized, resuspended in normal fibroblast media on a feeder- free substrate (Geltrex-coated dishes, GeltrexTM: Life Technologies, A1413302) where they were fed daily with stem cell medium subsequently (E8TM plus supplement, Life Technologies, A1517001) and kept until emerging colonies were ready for manual picking, isolation, purification, expansion, banking and validation. I generated 10 to 20 iPSC clones per patient and banked all at low passages. Two clones per patient were cultured further and underwent all characterisation and validation steps described here.

Validation included proof of presence of original heterozygous ITPR1 deletion in the six self-derived patient lines and absence in the three lab-internal control lines, normal karyotyping, analysis of expression of endogenous pluripotency associated genes via ICC and quantitative PCR (qPCR), confirmation of downregulation of original fibroblast

116 genes (qPCR), confirmation of silencing of exogenous plasmid associated pluripotency factors (qPCR) and potential to differentiate into (neuronal and astrocytic) postmitotic tissue (cell culture). iPSC culture and neuronal differentiation iPSCs were maintained on Geltrex (Life Technologies, A1413302). Essential 8 media (Life Technologies, A1517001) was exchanged daily and cells were split using UltraPureTM 0.5mM EDTA (Life Technologies, 15575-020) in Dulbecco’s phosphate- buffered saline (DPBS) without magnesium and calcium (Life Technologies, 14190- 094) every 2 to 4 days depending on confluency and experimental plans. All procedures (splitting, freezing, thawing, pooling) were carried out in accordance with the lab’s internal regulations for sterile cell culture and the recommendations of iPSC culture (Life Technologies, Publication Number MAN0007035). See Appendix, Table 8 – 23 and following paragraphs for details on necessary reagents and procedures.

One day before neural induction was initiated, cells were split and pooled from multiple wells to grow to 100% confluency. The next morning when cells were confluent, neural induction was initated and this time-point was defined as Day zero/Day in vitro zero (D0/DIV0) of the experiment. Importantly, control and patient cells received the same treatment (e.g. neural induction, splitting of cells, replating for experimental readout) at the same time-points. Neuronal differentiation was initiated on a confluent sheet of iPSCs using dual SMAD inhibition (236) and the published Nature protocol for directed differentiation of cortical neurons was followed strictly for neural differentiation, important validation-steps, neuronal culturing and experimental planning (237). See Appendix, Section 5, “Neuronal cell culture” and Table 8-24 for culture condition details and neural maintenance media recipes. See Table 8-25 for details on extrinsic cues employed in neural induction, patterning and differentiation described in the following subchapters. gDNA extraction from cells and HumanOmniExpress BeadChip kit To check whether the derived patient iPSC lines still carried the genetic defect (heterozygous ITPR1 deletion) I extracted gDNA from iPSCs for genotyping: For each self-derived iPSC line and the three control lines media was removed from a confluent well of cells, cells scraped, collected in phosphate-buffered saline (PBS) and spun down shortly. PBS was removed and the pellet resuspended in 1 ml gDNA extraction buffer supplemented with Proteinase K. The sample was left in a heatblock at 55°C overnight for homogenisation. DNA was extracted the next day with a phenol/chloroform extraction once. Subsequently, 33 µl Na-Acetate and 2.5 ml 100%

117 ethanol was added per sample and frozen at -20°C for 1 hour, spun at top speed for 15 mins, before careful removal of ethanol. DNA was rinsed with 70% ethanol, spun at top speed for 15 mins and upon aspiration of ethanol, the pellet was left to air-dry and subsequently dissolved in TE. See Appendix, Section 5 for additional details of the protocol. DNA was quantified using nanodrop, diluted to 50 ng/µl in 15 µl and ran as a UCL service by Kerra Pearce on the HumanOmniExpress BeadChip-241v.1.1 (Illumina, WG-315-1101) at the UCL Genomics, Microarray and High Throughput Sequencing Facility. I analysed the obtained genotyping data using GenomeStudio Software 2.0 from Illumina.

Karyotyping In order to exclude chromosomal abnormalities in the self-derived iPSC lines, all six iPSC lines were grown to 80% confluency in a T25 flask, topped up with media and transported in styropor isolation boxes to Cell Guidance Systems Ltd (Moneta Building, Babraham Bioscience Campus, Cambridge CB22 3AT, UK) where the cells were processed for G-band karyotyping analysis (Giemsa stain of cells fixed in metaphase, following digestion of chromosomes with trypsin) as part of a paid service. The control cell lines had been karyotyped by the collaborators/Coriell previously.

RNA extraction, RT-PCR and qPCR To query specific gene expression at different time-points, RNA was extracted from fresh fibroblasts, iPSCs and neuronal cultures in TRIzol-Reagent (Life Technologies, 15596-018) and extraction, isolation and resuspension was performed according to the manufacturer’s recommendations (please see Appendix, Section 5, under RNA extraction where a detailed description of the protocol is stated). RNA was quantified using nanodrop and subsequently used to generate complementary DNA using SuperScript® III First Strand Synthesis System (Invitrogen,

18080051) – please see Appendix, Section 5, Tables 8 – 28, 8 – 29 and 8 – 30 for workflow, pipetting volumes and cycling conditions. Primers for the housekeeping genes GAPDH, Cyclophilin and beta-actin as well as individual target primers for the different cell lines (fibroblasts and iPSCs: OCT3/4, SOX2, NANOG, S100A4, ITPR1) were used together with Power SYBR® Green master mix (Applied BiosystemsTM, 4367659) and 10 µg of cDNA as input for amplification, detection and quantification of samples in 96-well format Mx3000P qPCR System (Stratagene. See Appendix, Section 5, Tables 8 – 26, 8 – 34 and 8 – 35 for workflow, pipetting volumes and cycling conditions).

118

Immunocytochemistry To confirm expression of target markers on the protein level, culture medium was aspirated and cells were washed with PBS briefly before being fixed in 4% paraformaldehyde for 10 min at RT. Cells were washed with PBS containing 0.3% Triton-100 (PBST, Sigma, T8787) three times for 5-10 min and blocked with 4% bovine serum albumin (BSA) in PBS for 20 min. Slides were incubated for 2 hours at RT or overnight at 4°C with the primary antibody/ies in 4% BSA in PBS (see Table 8-38 and Table 8-39 for lists of primary antibodies). Cells were washed with PBS containing 0.3% Triton-100 (Sigma, T8787) three times for 5 to 10 min again and incubated with the appropriate secondary antibodies (Alexa Fluor® dyes secondary antibodies, Life Technologies, see Table 8-40 for list of secondary antibodies) for 1 hour at RT (or overnight at 4°C) in the dark. They were washed again with PBST three times for 5 to 10 min and the first wash was combined with a counterstain with 4,6-Diamidino-2-phenylindole dihydrochloride (DAPI, Sigma, D9542, concentration 1:1000). Cells were mounted using mounting media (Sigma-Aldrich, F4680) or preserved using Anti-fade reagents (ProLong® Gold antifade reagent, Life Technologies, P36934) and visualised on a Zeiss fluorescence microscope.

Live cell imaging To investigate the cells’ calcium homeostasis and ability to handle calcium, iPSC- derived cortical neurons and glia cells at different days of maturation were compared with the corresponding control lines. All cells were grown on glass coverslips (22 mm, Bioptechs Delta TTM, 0420-0323-2), µ-slides (8-well, ibidi, Thistle Scienctific, IB-80826) or 24/96-well-imaging plates (BD Falcon Plates, BD Biosciences, 353226/353219).

When performing the live cell experiments, guidance and supervision was provided by

Dr. Z. Yao who - depending on timings/booking schedule, experiment equipment and location of experiments performed (e.g. automatic live cell imaging machine at UCL main Campus) – also kindly generated parts of the results and helped with interpretation of the findings.

Ratio-metric Ca2+ dye, acetoxy-methyl-ester Fura-2 (Fura-2, AM, Molecular ProbesTM), or non-ratiometric Ca2+-dye Fluo-4, AM (Thermo Scientific) were used to measure intracellular calcium levels. Fura-2 AM was excitated with 340 and 380 nm light and emitted fluorescence was recorded with a CCD camera through a 515 nm long-pass filter on an epiflourescence inverted microscope with a x20 objective. Fluo-4 AM images were obtained using either a Zeiss 710 confocal microscope with x63 objective

119 with excitation wavelength 488 nm and fluorescence measured from 505 to 550 nm or using ImageXpress high-throughput imaging system with x20 objective and 472/30 nm excitation and 520/35 nm emission filter set. For measuring responses to physiological stimuli, cells were first washed gently with PBS and loaded with Fura-2 or Fluo-4 with 0.005% pluronic acid for 30 mins in the dark. Subsequently, the loading dye was removed, cells were washed with HBSS (ThermoFisher Scientific, 14025092) twice. After taking a basal recording for several minutes, cells were stimulated with either ATP (50 µM), Glutamate (1 µM) or KCL (50 mM) to evaluate their neuronal and/or astrocytic response characteristics.

For measurements of ER calcium store and store operated calcium entry (SOCE), cells were washed with calcium-free medium (Gibco® Hank’s balanced salt solution (HBSS), no calcium, no magnesium, ThermoFisher Scientific, 14170112) after loading with either Fura-2 AM or Flou-4 AM dye. ER calcium store was measured at the increase of cytosolic calcium signal after thapsigargin (1 µM) addition in calcium free media and 1 mM CaCl2 was applied subsequently to measure store operated calcium entry. As a positive experimental control and for calibration purposes the ionophore ionomycin (1 µM) was added at the end of each experiment. Obtained data was analysed and plotted using Fiji/ImageJ and excel and statistical analyses performed using GraphPad Prism software. P-values < 0.01 were considered statistically significant.

Electrophysiology To assess their electrophysiologic properties, terminally differentiated cortical neurons were grown on glass coverslips (22 mm, Bioptechs Delta TTM, 0420-0323-2) and analysed on DIV80-DIV100. All recordings were conducted at room temperature and carried out by Dr. Sarah Crisp from the UCL Clinical and Experimental Epilepsy department after handing the cultured cells over to her at IoN. The patch pipette solution contained: 130 mM KCl, 10mM HEPES, 0.2mM EGTA, 2mM Mg-ATP, 0.3mM Na-GTP, 20 mM creatine phosphate (pH 7.3). The standard extracellular solution used in all experiments contained 125 mM NaCl, 2.5 mM KCl, 1 mM MgCl2, 2 mM CaCl2, 25 mM glucose and 10 mM HEPES (pH to 7.3 with NaOH). Patch pipettes had a resistance of 3 to 4 MΩ. Current-clamp recordings were obtained using an Axopatch 700B amplifier (Molecular Devices). Data were acquired and analysed using LabView software (V.10.0, National Instruments) with in-house acquisition programmes. Data were sampled at 20 kHz and filtered at 10 kHz. After setting the bridge balance and adjusting the holding current to keep the cell at −70 mV, 1-second-long current injections, in 10-20 pA incremental steps, were delivered. A hyperpolarising step was

120 used to determine passive properties. Depolarising steps were used to elicit action potentials. Parameters of action potentials were measured for the first action potential (defined as peak > 0 mV) elicited with each current injection. Graphs were plotted using Microsoft Excel 2010.

RNA sequencing In order to assess the transcriptional processes in iPSC-derived patient neurons as compared to controls, whole transcriptome shotgun sequencing/RNA sequencing was collaboratively performed.

Library preparation and sequencing I extracted RNA from neuronal cultures using the Maxwell RNA Isolation Instrument with the Maxwell® RSC simplyRNA Cells Kit (Promega AS1390) as described in more detail in Appendix, Section 5, RNA extraction. Obtained RNA was checked for integrity and quantity using RNA LabChip® kits on the Agilent 2100 bioanalyzer (Agilent Technologies). The following steps were then carried out by Claire Hall: Libraries were prepared using 1 µg of total RNA as input and the TruSeq® Stranded mRNA Kits (Illumina, RS-122-2103) by following the steps outlined in the manufacturer’s protocol (http://support.illumina.com/content/dam/illumina- support/documents/documentation/chemistry_documentation/samplepreps_truseq/trus eqrna/truseq-rna-sample-prep-v2-guide-15026495-f.pdf). Summarised briefly, in a first step mRNA was enriched via magnetic beads, purified and fragmented (ribosomal RNA is depleted during these first steps). Secondly, a cDNA library was generated by reverse-transcription with first strand cDNA synthesis (to prevent spurious DNA-dependent synthesis, Actinomycin D is added during this step) followed by second strand cDNA synthesis using dUTP to allow for strand specificity and RNAse H to digest any remaining RNA. 3’-5’ exonucleases and 5’-3’ polymerases in the end repair mix generated blunt-ended fragments where 5- phosphate groups are added during end repair to facilitate subsequent adapter ligation. The process of A-tailing (addition of a single 3’-A-overhang) made mRNA fragments compatible with the adapters and prevents self-ligation. Subsequent ligation of sequencing adapters with incorporated barcodes allowed multiplexing of samples and pooling of libraries. Finally, PCR amplification generated the libraries that were pooled and sequenced on a HiSeq 2500 (rapid run, 50 cycles).

121

More information about the illumina TruSeq® kits and the protocol can be found here: http://www.illumina.com/content/dam/illumina- marketing/documents/products/datasheets/datasheet_truseq_stranded_rna.pdf.

Bioinformatic analysis RNA sequencing data Bioinformatic analyses are led by Dr. Chris Sibley and data is analysed using his specialised inhouse RNAseq analysis pipeline. Briefly, reads derived from the sequencing are mapped against the human reference genome (version hg19) using Bowtie2 and TopHat software. Reads per gene are counted and normalized and subsequently used as input for statistical analysis where variance based on the replicates for each sample in the mutant versus control groups was calculated and differentially expressed genes extracted. The analyses are still ongoing and therefore, results are not shown in this thesis.

2.4.2 Methods Chapter 6.2

The study described in Chapter 6.2 set out to compare six different induction and six different patterning strategies (details explained in the next paragraph) with qPCR and ICC readouts at day in vitro 0 (D0/DIV0), four (D4/DIV4), seven (D7/DIV7) and fourteen (D14/DIV14) to find out which induction and patterning combination at which time-point meets the three goals mentioned in the introduction of Chapter 6.2.

Cell culture All experiments were conducted on the two control iPSC and one control embryonic stem cell line described in Chapter 6.1 and Appendix, Section 5, Table 8-22. iPSCs were cultured as described in Chapter 6.1 and Appendix, Section 5. One day before neural induction was initiated, cells were split and pooled from multiple wells to grow to 100% confluency. The next morning when cells were confluent, neural induction was initated and this time-point was defined as Day zero/Day in vitro zero (D0/DIV0) of the experiment.

Importantly, per experiment all different induction conditions were started on the same day and the different media was always exchanged 24 hours later (+/-1 hour) on all conditions simultaneously. In order to start neural induction, E8 media was removed and cells were washed with DPBS once. Neural maintenance media (N2B27, see Table 8 – 24 or details and recipe) containing the respective concentrations of different extrinsic cues (see Table 8 – 25 for their characteristics and presumed mechanism of

122 action) as stated below was added and subsequently exchanged with fresh medium after a gentle wash with DPBS every 24 hours (+/-1 hour).

The following six different neural induction conditions were examined in experiment 1 (please note that the nomination of the different conditions to “A”, “B”, “C”, “D”, “G” and “H” is entirely arbitrary, and was chosen here to distinguish this experiment and its conditions from previous experiments during the process):

A: Dual SMAD inhibition for all days (=dSMADi: 1 µM dorsomorphin (DM), 10 µM SB431542 (SB)) – representing the “default” and most widely utilized neural induction condition traditionally yielding forebrain neurons. B: 3 compounds neural induction for all days (=3cs: 1 µM DM, 2 µM SB, 3 µM CHIR99021) – representing a neural induction condition used in the generation of spinal cord motor neurons. C: 1 µM DM + 10 µM SB + 3 µM CHIR99021 for all days – representing a mixture of the forebrain and spinal cord induction cocktails A and B. D: staggered: dSMADi for 1 day, followed by 3cs (µM: 1,2,3 as above) for 3 and 6 days – representing a neural induction condition that permits the investigation of temporal events in neurogenesis when exposed to forebrain (1 day) followed by spinal cord motor neurons induction cocktails. G: staggered: dSMADi for 2 days, followed by 3cs (µM: 1,2,3 as above) for 2 and 5 days – representing a neural induction condition that permits the investigation of temporal events in neurogenesis when exposed to forebrain (2 days) followed by spinal cord motor neurons induction cocktails. H: staggered: dSMADi for 3 days, followed by 3cs (µM: 1,2,3 as above) for 1 and 4 days – representing a neural induction condition that permits the investigation of temporal events in neurogenesis when exposed to forebrain (3 days) followed by spinal cord motor neurons induction cocktails.

See Figure 2-6 for a graphic overview of investigated neural induction conditions in experiment 1.

123

Figure 2-6: Six different neural induction conditions, experiment 1

The upper part of the figure gives the timeline and character of experimental readouts and indicates the days of media exposure and the concentration of extrinsic cues for the six different induction conditions (A-H) investigated in experiment 1. Abbreviations: DM=dorsomorphin; SB=SB431542; CHIR=CHIR99021; DIV=day in vitro; ICC=immunocytochemistry; qPCR=quantitative PCR.

To investigate in a second step which extrinsic cues are necessary during patterning of successfully converted neurally induced cells from experiment 1, the following six different neural patterning conditions were examined in experiment 2 (please note that the nomination of the different conditions to “A”, “B”, “C”, “D”, “E” and “F” is entirely arbitrary, and was chosen here due to practicability and to distinguish this experiment and its conditions from previous experiments during the process):

A: 7 days dSMADi followed by 7 days no patterning (N2B27 only) – representing the “default” and most widely utilized neural induction condition followed by no patterning, traditionally yielding forebrain neurons. B: 7 days dSMADi followed by 7 days CHIR99021-patterning (3 µM) – representing the “default” and most widely utilized forebrain neural induction condition followed by patterning towards mid/hindbrain using a WNT agonist. C: 7 days dSMADi followed by 7 days FGF8-patterning (200 ng/ml) – representing the “default” and most widely utilized forebrain neural induction condition followed by patterning towards mid/hindbrain using FGF8.

124

D: 7 days of 3cs followed by 7 days no patterning (N2B27 only) – representing the spinal cord motor neuron induction condition followed by no patterning. E: 7 days of 3cs followed by 7 days CHIR99021-patterning (3 µM) – representing the spinal cord motor neuron induction condition followed by patterning towards mid/hindbrain using a WNT agonist. F: 7 days of 3cs followed by 7 days FGF8-patterning (200 ng/ml) – representing the spinal cord motor neuron induction condition followed b patterning towards mid/hindbrain using FGF8.

RNA extraction, RT-PCR and qPCR IPSCs and neuronal cultures at day in vitro 0, 4, 7 and 14 after neural induction were washed with PBS once. They were then gently lifted using incubation with UltraPureTM 0.5mM EDTA (Life Technologies, 15575-020) in DPBS without magnesium and calcium (Life Technologies, 14190-094) for 5 to 15 minutes until spontaneous lift up in one unperturbed sheet occurred. The cell sheet was collected in media, briefly spun down and the pellet immediately snapfrozen on dry ice. RNA was extracted from snapfrozen pellets using the Maxwell RNA Isolation Instrument with the Maxwell® RSC simplyRNA Cells Kit (Promega AS1390) as described in more detail in Methods Chapter 6.1 (RNA sequencing) and Appendix Section 5, RNA extraction. RNA was quantified using nanodrop. 2 µg of RNA were used to generate cDNA using RevertAid – Reverse Transcriptase from ThermoFisher Scientific (EP0441) and the RevertAid First Strand cDNA Synthesis Kit (KI621) – please see Appendix, Section 5, Tables 8 – 31, 8 – 32 and 8 – 33 for workflow, reagents, pipetting volumes and cycling conditions. 10 µg of cDNA were determined as input for qPCR using Fast SYBR® Green master mix (Applied Biosystems™, 4385612) for amplification, detection and quantification of samples in 384-well format QuantStudio™ 6 Flex Real-Time PCR System (Applied Biosystems™). See Table 8 – 27 for primers used in this project which are grouped by regional identity and/or germ layer origin. See Appendix, Section 5, Table 8 – 36 and Table 8 – 37 for workflow, pipetting volumes and cycling conditions. As control conditions for all experiments, one RT- sample (enzyme reverse transcriptase was replaced with water), and 1 sample where water instead of cDNA was added were run for each primer pair on each plate/qPCR run. Data was normalized to GAPDH expression, analysed and plotted in Excel and GraphPad Prism software.

Experiment 1 was performed and analysed using cell pellets from three independent inductions in each of the two different iPSC control lines (see Table 8-22) on all three time-points (D0, D4, D7), and confirmatory follow-up immunocytochemistry was obtained on two different time-points (D4 and D7) using three independent inductions

125 of one iPSC control line (Control 1). Patterning experiment 2 was performed using cell pellets at D14 from one independent induction in two independent control iPSC lines each.

Immunocytochemistry All cells were grown on glass coverslips or µ-slides as described in Methods Chapter 6.1, Live cell imaging. Cells were washed with DPBS once prior to fixation at day in vitro 4 and 7. ICC was performed as described in Methods Chapter 6.1 above. See Table 8-39 and Table 8-40 for lists of primary and secondary antibodies. OTX2-positive and DAPI-positive cells were counted. Data was analysed using ImageJ/Fiji software and plotted in Excel.

126

3 Chapter 3: Dissecting neurodegeneration with brain iron accumulation clinically, genetically and cellularly

Studying NBIA disorders via whole exome sequencing to identify variants in candidate genes and reduce the frequency of unexplained NBIA disorders

3.1 Introduction

NBIA describes a group of rare, inherited neurodegenerative diseases characterised by a combination of early-onset cognitive features and a movement disorder usually accompanied by typical MRI appearance (58). Their world-wide prevalence is estimated to be below 1/1000000, adding these disorders to the group of ultra-rare orphan diseases. Most NBIA disorders are inherited in autosomal recessive fashion, however X-linked and autosomal dominant inheritance has been described less frequently. Considerable progress has been made to identify underlying genes, however with the advent of next-generation sequencing techniques previously unknown expansions of the phenotype-genotype spectra have been discovered, establishing clarity, dynamics and confusion to similar degrees in this exciting field of neurodegeneration. To date, ten causal genes of different function with implication in iron, lipid metabolism and lysosomal and mitochondrial activity have been described and several additional genes are associated more loosely to the group of NBIA disorders (45); see Figure 3-1 for a frequency spectrum. There is significant phenotypic heterogeneity and clinical overlap to (atypical) parkinsonian syndromes, genetic forms of dystonia and chorea, Huntington’s disease like (HDL) disorders, leukodystrophies, hereditary spastic paraplegias, motor neuron disease, mitochondrial diseases, cerebellar ataxias and dementias (147, 151, 220, 359-366). Unexplored genotype- phenotype relationships are therefore likely to emerge as NGS approaches meet NBIA.

As outlined above, a specific proportion of NBIA cases cannot be explained by mutations in any of the associated genes, classifying them as idiopathic NBIA (58), see Figure 3-1, “Idiopathic“.

127

Figure 3-1: NBIA subtypes and frequency distribution

NBIA subtypes and frequency distribution. Data based on the North American database. Not NBIA: Patients with an NBIA-like phenotype, but non-NBIA mutations. Abbreviations: NBIA: neurodegeneration with brain iron accumulation, PKAN: pantothenate kinase- associated neurodegeneration, PLAN: phospholipase A2-associated neurodegeneration, INAD: infantile neuroaxonal dystrophy, MPAN: mitochondrial membrane protein- associated neurodegeneration, BPAN: beta-propeller protein-associated neurodegeneration, FAHN: fatty acid hydroxylase-associated neurodegeneration, CoPAN: Coenzyme A synthase protein-associated neurodegeneration, NF: neuroferritinopathy, KRS: Kufor-Rakeb syndrome, ACP: aceruloplasminemia. Figure and legend reproduced from (58).

Their existence suggests remaining gene discoveries are still to be made. However, this phenomenon simultaneously points towards the limitation of current genetic and non-genetic approaches which regularly fail to uncover the hidden contribution of noncoding DNA elements, epigenetic and environmental factors in disease pathogenesis; all presumably playing important roles and adding to the biological complexity of these rare degenerative disorders.

This study therefore set out to characterise a cohort of 91 familial and sporadic idiopathic NBIA patients using combined approaches of clinical investigations, exome sequencing and Sanger sequencing to identify underlying genetic causes and explore the associated clinical spectrum.

128

3.2 Results

Clinical results – Cohort For the clinical phenotype cohort analysis in this project, I only included those 91 patients where good quality exome sequencing data was generated in order to facilitate correlation of genetic and clinical data. From the clinical data, gender was unfortunately not reported in 8 cases and ethnicity was missing on 11 patients. The exact year of disease onset was unretrievable for 24 cases, however, categorisation into “since birth”, “early-onset” (< 20 years) or “late-onset” (> 20 years) was possible for 85 out of 91 cases. Family history was not clearly identifiable for 13 cases and detailed imaging data of sufficient quality or the neuroimaging specialist’s report was not retrievable for 9 patients. Table 3-1 summarises the analysis of the most important clinical hallmarks (gender, ethnicity, age at onset, family history, MRI findings and clinical symptoms) of the investigated cohort. Figure 3-2 plots the frequency of symptoms in the whole cohort. Figure 3-3 gives a visual account of the same data in a different way: It plots the frequency of clinical symptoms present in the investigated cohort of NBIA-like disorders as absolute symptom count, with the most common symptoms being parkinsonism, and dystonia (enrichment for oromandibular involvement in n=31 of 46 patients with dystonia). Please note that counts of each specific symptom for the whole cohort (not symptom percentages) are plotted here - due to double (or triple) presence of different symptoms in single patients the total symptom count for the whole cohort was 234.

129

Figure 3-2: Clinical symptoms NBIA cohort

Percentage calculated as absolute symptom count for each symptom onto the number of patients investigated (n=91).

Figure 3-3: Clinical symptoms NBIA cohort, absolute symptom counts

Number of absolute symptom counts as in NBIA cohort (total symptom count: 234).

130

Table 3-1: Clinical summary of NBIA cohort

N %# Overall cohort 91 n.a. Gender Female 41 49.4 Male 42 50.6 n.a. 8 n.a. Ethnicity European descenta 54 67.5 Asian descentb 19 23.75 African descentc 4 5 Other (mixed)d 3 3.75 n.a. 11 n.a. Age at onset [yrs]1 Mean: 25.7 Range: 0-65 Since birth 26 30.6 EO (<20yrs) 18 21.2 LO (>20yrs) 41 48.2 Family history2 Present 28 35.9 Absent 50 64.1 n.a. 13 n.a. MRI findings MRI normal 17 20.7 Iron deposits 37 45.1 -eye of the tiger 6 7.3 Brain atrophy 25 30.5 -cortical 11 13.4 -subcortical 4 4.9 -cerebellar 12 14.6 -not specified 4 4.9 White Matter Abnormalities 7 8.5 -WM lesions (not small vessel) 4 4.9 -TCC 5 6.1 n.a. 9 n.a. Clinical findings Parkinsonism 51 56.0/21.8 Dystonia 46 50.5/19.7 Intellectual disability/developmental delay 28 30.8/13.2 Cognitive decline/dementia 21 23.1/11 Psychiatric/behavioural disturbances 18 19.8/7.7 Myoclonus 17 18.7/7.3 Epilepsy 17 18.7/7.3 Chorea 11 12.1/4.7 Spasticity 9 9.9/3.8 Ataxia 8 8.8/3.4 Neuropathy 4 4.4/1.7 Tics 3 3.3/1.3 Retinopathy 1 1.1/0.4

Explanations: # Percentages calculated on total n with respective information available (different for each category) and rounded to 1 decimal place. 1 missing exact AAO in n=24, categorisation (Since birth, EO, LO) possible for n=85 (of 91). a 'European descent' subsumes British, Polish, Scottish, Sinti and Roma, Turkish, Greek, Hungarian nationalities and 'Caucasian not further specified'. b 'Asian descent' subsumes Iranian, Pakistani, Indian, Bangladeshi, Chinese, Saudi Arabian, United Arabian Emirates, Jordan, Chinese nationalities and 'Asian not further specified'. c 'African descent' subsumes

131

Black African and 'African not further specified'. d 'Other' subsumes mixed ancestry and Caribbean. 2additional consanguinity reported for 8 probands, 5 of those match criteria for familial, 3 for sporadic disease.

Abbreviations: n=absolute number, n.a.=not applicable/available, AAO=age at onset, EO=early-onset, LO=late-onset, yrs=years; FH=family history, WM=white matter, TCC=thin corpus callosum, MRI=magnetic resonance imaging.

Please note that percentages in MRI findings are calculated by counting each specific feature and calculating its percentage on the total of patients with detailed MRI data available (n=82) – due to double (or triple) presence of different MRI findings in single patients, these percentages therefore do not match up to 100%.

Please note that percentages in clinical findings are calculated by counting each specific symptom and calculating its percentage on the total of patients with phenotype data available (n=91) - due to double (or triple) presence of different symptoms in single patients, these percentages therefore do not match up to 100% (first percentage value). Second percentage value was calculated as count of specific symptom onto the total symptom count in the whole cohort (symptom count n=234). These percentages round up to slightly more than 100% due to rounding up to 1 decimal place of obtained values.

132

Genetic results - Cohort Out of 104 patients preliminarily selected for this study, 95 patients met the clinical inclusion criteria. However when analysing the genetic data, the exome sequencing data of four participants had to be discarded due to little amount and poor quality of DNA and subsequent libraries for sequencing and overall low exome coverage (also making required Sanger follow-up impossible) (n=2) and suspected sample mixup/duplication of another two cases implicated. Therefore, genetic cohort results (as in clinical cohort results chapter above) are reported on 91 patients. In 91 patients, 99,3% of the exome was covered with a depth of at least 2 reads (2x), 10x coverage was 95,2%, 20x coverage 84,9% and 69,9% of the exome were covered with at least 30 reads.

Figure 3-4: Coverage metrics NBIA cohort

Figure kindly generated by bioinformatician Dr. Jinhui Ding at NIH after a joint discussion.

Figure 3-4 gives a visual representation of the percentage of the exome covered 10 (blue) and 30 (red) times (y-axis, 0-1=0-100%), where each pair of red and blue line corresponds to one sample on the x-axis (S1-S95, as 95 exomes are illustrated here, however, four samples were excluded for the downstream analysis for reasons mentioned above).

In 91 analysed samples, 334087 coding variants were found in the group-vcf of the cohort, of which 76205 were nonsynonymous, stopgain, stoploss, frameshift, nonsense or splice site variants. Of these, 799 lay within the candidate genes (as to be found in Table 2-1). After further 1) filtering for frequency and presumed mode of inheritance in family and gene in question, 2) Sanger sequencing of interesting variants and 3)

133 segregation testing where applicable a highly likely genetic cause in 8.8% of patients was identified. See Figure 3-5 for variant filtering strategy.

Figure 3-5: Variant filtering strategy candidate genes

Filtering strategy and obtained variants for the candidate gene study.

1 See Table 2-1 for candidate genes investigated.

All variants within candidate genes were classified into the following three groups: 1) Highly likely pathogenic variants confirmed by Sanger sequencing: If variants were published as disease causing in convincing reports before, or were novel (absent in control databases) truncating/frameshi ft variants or novel missense variants in regions of the gene previously shown to harbour many disease causing variants with additionally convincing prediction scores and segregation in the family. See Table 3-2 for 10 variants thought to cause disease in 8 patients. 2) Less likely, but possibly pathogenic variants confirmed by Sanger sequencing: Novel missense variants where little information was available, frequencies were borderline, and/or predictions variable and/or the phenotype and/or inheritance mode of the patient did not fit the gene. See Table 3-3 for 23 variants less clearly convincing to cause disease in 21 patients.

134

3) Non pathogenic variants in known disease causing genes: Frequent in public control databases (see frequency filtering criteria in Methods for cut-off values). These were discarded from the analysis and not confirmed by Sanger sequencing.

Table 3-4 shows additional variants that looked interesting on manual curation, however, Sanger sequencing could not confirm their presence, most probably due to low coverage on the exome and/or false-positive variant calling or misannotation.

DNA of family members was successfully organised for the PLA2G6 and FA2H mutation carriers and results show perfect segregation throughout the family (see Figure 3-6 and Figure 3-7). For presumably compound heterozygous mutations in PANK2 no family members were available to test the mutation phase to be in trans.

135

Table 3-2: Highly likely pathogenic variants NBIA cohort 136

Abbreviations: M=male; F=female; n.a.=not available; S.=Saudi; EO=early-onset; AO=adult-onset; gen.=generalised; A=ataxia; DP=dystonia- parkinsonism; M=myoclonus; MD=myoclonus-dystonia; D=dystonia; ME=myoclonus epilepsy; LD=learning disability; S=seizures; PD=here Parkinsonism or atypical complex PD – outside this table PD=Parkinson’s disease; cHSP=complex hereditary spastic paraplegia; foc.=focal; gen.=generalized; orom.=oromandibular; het=heterozygous; hom=homozygous; coh=compound het; SNV=single nucleotide variant. Abbreviations for interpretation and naming of frequency and prediction annotations from ANNOVAR standards, see ANNOVAR documentation: http://annovar.openbioinformatics.org/en/latest/articles/VCF.

Table 3-3: Variants of unclear pathogenicity NBIA cohort 13 7

Abbreviations: M=male; F=female; EO=early-onset; AO=adult-onset; A=ataxia; DP=dystonia-parkinsonism; MD=myoclonus-dystonia; D=dystonia; C=chorea; PD=here Parkinsonism or atypical complex PD (outside this table PD=Parkinson’s disease); cHSP=complex hereditary spastic paraplegia; Psy.: psychiatric comorbidities; DD=developmental delay; cons.=consanguinity; sp.=spasticity; nsn=nonsynonymous; SNV=single nucleotide variant; het=heterozygous; hom=homozygous; hem=hemizygous; coh=compound het. Abbreviations for interpretation and naming of frequency and prediction annotations from ANNOVAR standards, see ANNOVAR documentation: http://annovar.openbioinformatics.org/en/latest/articles/VCF/. 138

Table 3-4: Interesting variants not confirmed by Sanger 139

These variants unfortunately could not be confirmed by Sanger sequencing (see red column “Sanger”) even though they had been detected on the vcf and partly on the BAM-file. Low coverage is the most likely explanation for most, see column BAM-file. Abbreviations: M=male; F=female; EO=early- onset; DP=dystonia-parkinsonism; auton.=autonomic features present; DD=developmental delay; LD=learning disability; sp.=spasticity; TCC=thin corpus callosum; cons.=consanguineous; M=myoclonus; D=dystonia; Psy.: psychiatric comorbidities; S=seizures; PD=here Parkinsonism or atypical complex PD – outside this table PD=Parkinson’s disease; het=heterozygous; hom=homozygous; coh=compound het; del=deletion; nsn=nonsynonymous; SNV=single nucleotide variant. Abbreviations for interpretation and naming of frequency and prediction annotations from ANNOVAR standards, see ANNOVAR documentation: http://annovar.openbioinformatics.org/en/latest/articles/VCF/.

Genetic results – Families Known NBIA genes

PLA2G6 mutation: Previously published mutation with an earlier age at onset and different radiographic presentation

Figure 3-6: Family tree, MRI and segregation status PLA2G6-case

A) Family tree of consanguineous PLA2G6-positive family. B) T2*-weighted MRI at age 26 of index case without evident basal ganglia abnormalities. C) Chromatograms of wildtype control sample (left panel, ctrl), heterozygous parents (left panel, I:1 and I:2) and homozygous index patient (right panel, II:1) and affected siblings (right panel, II:2 and II:3) for previously published pathogenic mutation c.2060G>A:p.R687Q. Abbreviations: wt=wildtype, mt=mutant, nt=not tested, ctrl=control sample.

A previously published homozygous mutation in PLA2G6 (p.R687Q) that segregated perfectly when testing further recruited family members (see Figure 3-6) was identified. The case described here (current age: 32 years) had an early-onset movement disorder with myoclonus, dystonia and parkinsonism with progressive cognitive decline since early infancy. The exact age at onset was difficult to establish: Parents reported normal pregnancy and normal early milestones. However when the patient tried to learn riding a bicycle in her early years, her movement disorder became more and apparent, and she never learnt to ride independently. Her young-onset progressive

140

parkinsonian syndrome with myoclonus and cognitive decline developed slowly further, however, her condition deteriorated significantly with pregnancy at age 25. The patient married her cousin and gave birth to a healthy son who is now 7 years old (for simplification reasons, exact relation between patient’s husband and herself not shown in family tree in Figure 3-6). On examination there were frequent oculogyric crises, action nystagmus, supranuclear gaze palsy, severe rigidity, tremor, bradykinesia and postural instability. Additionally, the patient shows signs of depression and anxiety, and progressive cognitive decline which render examination and interpretation of cognitive and mood findings more difficult. The patient is wheelchair-bound and urine incontinent, her relatives reported problems with mood instability. Brain MRI at age 26 was reported as normal, especially, no signal abnormalities indicating excess iron could be detected on T2-weighted MRI (see Figure 3-6, B).

141

Novel compound heterozygous FA2H mutations

Figure 3-7: Family tree and segregation status FA2H-positive case

A) Family tree of FA2H-positive family. B) Chromatograms showing mutation status for compound heterozygous mutations for heterozygous father (I:1, upper panel, left and right), compound heterozygous index (II:3, middle panel, left and right) and his affected sister (II:4, lower panel, left and right). Abbreviations: wt=wildtype, mt=mutant, nt=not tested.

I identified two novel, segregating compound heterozygous mutations in FA2H in one individual (see Figure 3-7 for mutation details, chromatograms and segregation).

Patient II:3 of family tree above (see Figure 3-7, A) presented with a picture of complex HSP. He had a congenital squint and developed absence seizures around the age of 2. Subsequently he developed problems of his lower limbs and only managed to walk on his toes at age 3. His stiffness progressively increased and he additionally developed a dystonic neck with severe retrocollis during his early teens. His absence seizures generalised into tonic-clonic seizures at the age of 11. Simultaneously, he developed dysarthric speech and slowly progressive problems with swallowing. He is now wheelchair-bound (since the age of 11) and bowel and urine incontinent with a baclofen pump since age 19. Since his teens, myoclonic elements complicated his movement disorder and his parents noted memory deficits.

142

The patient has two older, healthy siblings, and his parents are neither affected nor related. His younger sister suffers from a similar condition with severe spasticity and dystonic elements, however, her disease started at age 7 only and she does not have seizures. Unfortunately, the brain MRI sequences were not available to us, but the report mentioned cerebellar atrophy with some white matter changes, no thin corpus callosum and no iron accumulation.

Known NBIA genes: single case results

PANK2 Two putatively compound heterozygous, previously published mutations in PANK2 (see Figure 3-8) were identified in a female Caucasian patient with early-onset dystonia-parkinsonism and iron deposits on MRI. Unfortunately, this sample was a collaborator’s sample where neither additional family members could be contacted, nor the original MRI with further clinical details received.

Figure 3-8: PANK2 mutations

Previously published pathogenic PANK2 mutations p.R190X and p.G230R in an individual with iron accumulation in basal ganglia and an early-onset movement disorder.

143

FTL Two samples carried the c.460dupA:p.H153fs mutation (NM_000146) in FTL. Both patients had an adult-onset form of disease.

Patient 1 is a 43-year-old female who was born from non-consanguineous parents and had normal pregnancy and early milestones. At age 25, she developed slowly progressive generalised dystonia with severe oromandibular involvement and dystonic posturing of upper and lower limbs. Iron and copper studies were normal, no acanthocytes were detected on repeated blood films. The patient continued to develop severe dysarthria and features of ataxia and increased anxiety with mild to moderate cognitive dysfunction. An MRI at age 37 revealed bilateral, symmetrical signal abnormalities resembling the eye of the tiger sign with volume changes in the lentiform nuclei and hyperintense global pallidi (T1-weighted; hypointense on T2) incorporating additional areas hypointense on T1 and hyperintense on T2. Additional findings included increased susceptibility in the pallidostriatal pathway, red nuclei, dentate nuclei and to a lesser extent in the thalamic nuclei and the putamina and global mild loss of neuroparenchymal volume. These radiographic findings were progressive in the disease course as Figure 3-9 shows the prominent iron accumulation (see white arrows) and cortical volume loss at age 39. There is no reported family history for the disease: The patient is the only child of her biological father, and none of her four half- siblings is affected. She has four healthy children herself with age ranging from 18-25 years (see Figure 3-9, A).

144

Figure 3-9: Family tree, chromatogram and MRI of neuroferritinopathy case

A) Family tree, B) FTL mutation c.460dup A and C+D) MRI findings of neuroferritinopathy case (arrow points to iron accumulation in basal ganglia). Abbreviations: mt=mutant, wt=wildtype, het=heterozygous, ins=insertion, dup=duplication.

The second individual identified in this study harbouring the same mutation is a 60- year-old British female with normal history up to her late thirties when she developed progressive atypical adult-onset chorea of the hand and face. Within the following years, paroxysmal jerks and focal dystonia of the right hand complicated her clinical presentation. Her MRI showed increased signal in the basal ganglia including a possible eye of the tiger sign with symmetrical atrophy and signal abnormality in the globi pallidi, red nuclei and dentate that return high signal on T2-weighted imaging and to a lesser extent on T1-weighted sequences. Furthermore mild generalised widening of cerebral sulci beyond normal age-related atrophy was indicative of a global degenerative process. The patient’s mother died of “Parkinson’s“ and interestingly had had an abnormal CT scan, her sister also has an adult-onset movement disorder. Unfortunately, all efforts to receive additional clinical details, the MRI scan for further review and DNA of her relatives were to no avail. Figure 3-10 shows a screenshot of the BAM-file (A) and Sanger sequencing results of mutation c.460dupA (B).

145

Figure 3-10: BAM-file and chromatograms FTL mutation

A) Screenshot of BAM-file at insertion position. B) Sanger chromatograms of common c.460dupA mutation in FTL.

WDR45

A heterozygous insertion of three basepairs (ATA) at position c.1004_1005 of WDR45 (NM_001029896) was identified in a 35-year-old British female (see Figure 3-11, A). She was born to healthy parents after a normal pregnancy. Her developmental milestones were delayed, she never acquired language, did not learn to walk independently and suffers from severe learning disability and progredient spastic quadriplegia from birth. At age 3 she developed seizures with predominantly absent episodes and occasional tonic convulsions that are complicated by dystonic posturing, dysarthria and dysphagia and myoclonic jerks from age 15 on. Her slowly progressive movement disorder and her cognitive abilities deteriorate significantly in her twenties. The patient is wheelchair-bound since her early teens, and now fully dependent on her carers in the activities of daily living. An MRI at age 27 showed global volume loss and

146

signal changes in the globus pallidus and substantia nigra consistent with excess iron accumulation (see Figure 3-11, B and C, white arrows and red arrowheads). There is no family history for the disease; the patient has one healthy brother and no children herself.

Figure 3-11: WDR45 frameshift insertion and MRI findings

A) A heterozygous ATA insertion in exon 11 at position c.1004_1005 of WDR45 (NM_001029896) was confirmed using Sanger sequencing B) Axial T2-weighted MRI at age 27 showing iron accumulation in the basal ganglia (see white arrows) and C) global volume loss (red arrowheads) on coronal T1-weighted MRI.

Known dystonia, chorea, PD and HSP genes: single case results

FUCA1 One case carried a previously reported homozygous stopgain mutation in FUCA1 (Figure 3-12). This sample was submitted from external collaborators with early-onset dystonia-parkinsonism with evidence of excess iron accumulation on MRI. Unfortunately, all efforts to obtain further details regarding clinical course, family history, MRI data and further DNA samples from the family were to no avail.

147

Figure 3-12: Pathogenic stopgain mutation FUCA1

Chromatogram of homozygous stopgain mutation c.1138G>T, p.E380X in exon 6 of FUCA1 (NM_000147).

PDE10A In collaboration with my colleague Dr. N. Mencacci and team, a de novo heterozygous, novel, disease causing mutation in exon 11 of PDE10A (c.898T>C:p.F300L, NM_001130690, see Table 2-3) was identified in a female individual with early-onset chorea complicated by adult-onset parkinsonism with bilateral striatal hyperintensities and striatal atrophy on MRI. Details of the functional consequences of this mutation and a detailed clinical description can be found in the respective publication and in the discussion below, but will not be restated in detail here (patient represents ’Individual 3’ in the original paper) (367).

Variants of unclear pathogenicity LRRK2 Four heterozygous missense variants in LRRK2 in four unrelated individuals were identified: The patient carrying the variant c.2264C>T, p.P755L (NM_198578) is a 71- year-old Asian male with atypical parkinsonism consisting of a bradykinetic syndrome with marked frontal and psychiatric features (age at onset (AAO): 63 years). His mother who died at age 92 had developed “parkinsonism“ in her sixties, he reports 6 healthy siblings and no other neurological diseases in the family. His MRI at age 67 showed increased mineralisation exceeding what is normal for the patient's age and slight cerebellar, brainstem and midbrain atrophy, unfortunately, the original images were not at our display.

The variant c.3974G>A, p.R1325Q (NM_198578) was found in an 55-year-old Scottish male with a clinical history of progressive dystonia-parkinsonism since the age of 46.

148

The initial symptom was dystonic posturing of the right foot during walking which progressively spread to involve cervical segments. Around age 50 the patient additionally developed parkinsonian tremor. His family history for neurological disease is negative and no family members were available for segregation testing. An MRI at age 52 showed some small iron deposits in the cone of the globus pallidus on both sides. Additionally, a second variant of uncertain pathogenicity in ADCY5: c.1297G>A, p.V433M (NM_001199642) was found in this individual, see Table 3-3.

The patient carrying the c.4235G>A, p.R1412Q variant (NM_198578) is a 47-year-old Indian male with a complex movement disorder and reduced cerebral volume and an eye of the tiger sign on MRI. At age 40 he noticed a tremor at rest in his left index and ring fingers that progressed to involve the entire left hand within months. At 43 years of age he additionally developed involuntary tongue protrusion, stiffness in his left leg and occasional short temper and aggressiveness. There is no reported family history for neurological diseases: His parents are deceased. He is the youngest of 7 siblings (4 brothers and 3 sisters, aged 50-70 years, whereof the eldest brother died of a heart attack) who all reside in India. None of the family members have been examined clinically or genetically but by the patient’s account all are normal.

The fourth interesting LRRK2 variant, c.3716A>C, p.E1239A (NM_198578), was identified in a 48-year-old Turkish individual with complex HSP. He reported walking difficulties very early in life and never learnt to walk normally. His slowly progressive bilateral lower limb spasticity quickly caused him to use crutches permanently and be wheelchair dependent for longer distances. He additionally suffers from spondylolysthesis with bilateral sensory symptoms and has a mild learning disability since childhood complicated by adult-onset psychiatric disturbances with delusions and hallucinations (~age 36). His parents are first cousins. He has a younger brother with spasticity and more pronounced intellectual impairment. His other brother and sister are unaffected. The patient himself has 2 children, a healthy girl and a 23-year-old boy who is thought to suffer from skeletal dysplasia without evidence of HSP. A cerebral MRI at age 34 was normal.

NIPA1 A novel NIPA1 missense variant, c.661C>A, p.P221T (NM_144599), was identified in a 55-year-old female. This patient was reported to have suffered from a mild Rubella infection at age 4 after which she developed hearing impairment, symptoms of myoclonus epilepsy and obvious "clumsiness" with ataxia upon coordination. Her MRI was reported with cerebellar degeneration without further abnormalities. Her paternal

149

grandmother was thought to be clumsy from her sixties on and the patient’s own 10- year-old son has some hearing impairment and clumsiness. Both other sons are healthy and there is no further family history for neurological diseases. Unfortunately, none of the family members were available for clinical or genetic testing.

SLC33A1 A rare heterozygous missense variant (c.1284G>A, p.M428I) in SLC33A1 (NM_004733) was identified in a 23-year-old Greek female with early-onset oromandibular dystonia, eye of the tiger sign on MRI and negative family history for neurological disease.

L1CAM The rare hemizygous missense variant c.1849G>A, p.D617N in L1CAM (NM_001143963, see Figure 3-13, A) was identified in a 29-year-old Iranian male with a complex multisystem disorder consisting of learning disabilities, optic atrophy, dystonia and seizures. He was born to healthy parents after a normal pregnancy and delivery, but reported to be more irritable as a toddler and to display feeding difficulties, visual disturbance due to progressive pigmentary retinopathy and optic atrophy and a learning disability early on. As an infant he developed dystonic posturing, a remnant slight tremor of his hands and from the age of eleven focal seizures with secondary generalisation. His siblings are healthy, and there is no family history of neurological disease. He was tested negative on a CGH array, and negative for PANK2 and common mitochondrial mutations prior to inclusion into this study. His MRI at age 21 displayed atrophy of the superior cerebellar vermis (see Figure 3-13, B).

Figure 3-13: L1CAM variant and MRI

Chromatogram of hemizygous L1CAM variant (A) and severe atrophy of the superior cerebellar vermis on T1-weighted FLAIR-images (B).

150

KIF1A A rare heterozygous missense variant in KIF1A (c.5269G>A, p.A1757T, NM_001244008) was identified in a female Portuguese-British individual with complex hereditary spastic paraplegia and prominent ataxia. The patient suffered from infantile febrile convulsions and a mild developmental delay. She developed a bilateral upper limb tremor and more obvious difficulties with speech and walking at age 7 leading over to a gradual progressive cerebellar, pyramidal and extrapyramidal syndrome with cognitive features. Her MRI showed marked decreased signal with prominent susceptibility artefacts within the striatum, globus pallidus, SN, red and dentate nuclei bilaterally, hints of a thin corpus callosum and a very thin spinal cord. Her family relationships are complicated and the patient does not hold regular contact to any of her family members. Her mother suffered from psychological disturbances and was reported to be a heavy drinker. One maternal half-sister, one maternal half-brother and two maternal aunts suffer from schizophrenia. Her paternal grandfather and one paternal half-uncle were reported to have PD in their early sixties. None of these individuals were available for genetic or clinical examination.

SGCE Two heterozygous SGCE variants were identified in a 39-year-old female Bangladeshi individual with generalised dystonia (AAO: 22 years), see Table 3-3. Reportedly, she suffered from prominent orofacial and buccal lingual involvement of her dystonia and some parkinsonian features that developed later in the course of her disease. Her MRI showed low signal on T2- sequences in keeping with increased iron deposition in the pallidus (age of examination: 37 years). She was born to first cousin parents, and her family history for neurological disease was negative, as well as previous testing for mutations in PLA2G6, PANK2, c19orf12, FBXO7, FTL, DYT1 and DYT6. Unfortunately, this was a collaborator’s sample and it was finally impossible to obtain any DNA samples from her relatives, the original MRI images and further details on the clinical history.

THAP1 A novel heterozygous frameshift insertion c.211_212insG, p.L71fs (NM_018105, see Table 3-3) was identified in a male patient from Saudi-Arabia with early-onset bulbar dystonia and secondary generalisation whose sample was sent in by collaborators. Unfortunately, no further clinical details were available.

151

Further variants of unclear pathogenicity Table 3-3 gives an account of additional variants of unknown pathogenicity that cannot all be described in detail here. Especially, the VPS13B variant is interesting, and will be discussed with the other variants in the relevant section of the discussion below.

Further interesting variants picked up by exome sequencing, but not confirmed by Sanger sequencing Table 3-4 lists further interesting variants detected by WES but not validated by subsequent Sanger sequencing (e.g. homozygous FBXO7 variant, that was only confirmed to be present in heterozygous state and thereby a fairly frequent polymorphism, false-positive homozygous PINK1, ATP7B, ETHE1 variants, etc.).

3.3 Discussion

Cohort: Clinical and genetic results NBIA disorders are usually young-onset conditions. However, mean age at onset in the described cohort was 25.7 years. This bias is highly likely due to the fact that NHNN is an adult hospital, and therefore the early-onset paediatric cases with severe phenotypes and death within the first one to two decades are not routinely referred to this hospital. Since many patients were recruited through the NHNN outpatients’ clinic, there is a tendency to include patients with milder phenotypes and survival well into adulthood who often have a later and less clear age at onset. As known as one of the most unreliable items in clinical studies, age at onset inherently often suffers from inaccuracy due to recall bias of the patient and subjective differences in symptom awareness. The age at onset is additionally influenced by these characteristics and therefore should be interpreted with a level of caution. Even though patients where the majority of NBIA genes had previously been excluded were preferentially included in the study, 1 patient with presumably pathogenic compound heterozygous mutations in PANK2, 2 patients with the same heterozygous pathogenic FTL mutation (previously published pathogenic frameshift insertion in exon 4: p.H153fs), a compound heterozygous FA2H-case (p.D266E and p.H261fs) and one WDR45-positive patient (pathogenic insertion leading to a stopgain in exon 12: p.Y336delinsX (a deletion at the exact same location had been previously published in a BPAN patient) were found when looking for (known) pathogenic variants in the ten known and established causal NBIA genes and the wider associated genes (RAB39B, DCAF17, SCP2, GTPBP2). This is quite a high frequency (5.5%; 5/91), and is most likely due to the fact that the cohort was gathered from diverse sources and the

152

diagnostic exclusion of known genes had not been carried out homogeneously for each patient. Extending the search paradigm to genes causing dystonia-parkinsonism phenotypes, the most prevalent symptoms in the described cohort, several additional probably disease causing variants in associated genes were identified. Interestingly, the previously published homozygous mutation in FUCA1 was found in one of the collaborators’ cases and one pathogenic mutation in a novel gene, PDE10A, was identified by my colleague Dr. N. Mencacci and further described in his manuscript (367). These examples further exemplify the genetic and phenotypic heterogeneity that characterises NBIA. Additionally, two interesting heterozygous variants in SGCE were found, as well as a novel, heterozygous frameshift in THAP1, four LRRK2 variants of unclear pathogenicity and several interesting variants in HSP-genes. The variants of unclear pathogenicity are discussed in further detail below. Overall, a reliable clinical diagnosis with the highly likely pathogenic gene was established in ~8.8% (8/91) of patients. This yield suggests the presence of further causal NBIA genes yet to identify. Given that the overall exome coverage was not 100% (for none of the depths (2x, 10x, 20x, 30x)) it is even more likely that more disease causing variants in known and novel genes which were simply not covered by this approach were missed. This adds to bioinformatics “data loss” where normal error rates can cause misannotation with subsequent misfiltering. Additionally variants of potential pathogenicity were identified in ~21 of patients (additional 23.1% of cases), however, the final pathogenic nature remains unclear and variably strong arguments exist against pathogenicity in some of these cases (see Discussion below).

Individual family and single case results: NBIA PLA2G6 The mutation p.R687Q (NM_0010004426) identified in one case has been published in the context of adult-onset dystonia-parkinsonism previously (corresponds to mutation p.R741Q (NM_003560) in Paisan-Ruiz et al.), initially establishing the association of PLA2G6 with adult-onset dystonia-parkinsonism (48). The affected individual in the Paisan-Ruiz et al. paper harbouring the same mutation however had an age at onset at 26 years with rapid cognitive decline and a movement disorder. MRI showed generalized cerebral atrophy and frontal white signal changes, but no excess basal ganglia iron on MRI. The case presented in this thesis however had an early-onset movement disorder consisting of myoclonus, dystonia and parkinsonism with progressive cognitive decline since early infancy. The clinical and radiographic presentation of the case from this cohort described in detail above is distinct from the one reported in Paisan-Ruiz et al. in regard to the following aspects: 1) Neither brain

153

atrophy nor white matter abnormalities were detected in the case from this cohort, however, MRI abnormalities could still develop with further disease progression. 2) Onset of disease was much earlier in the case from this cohort, even though similar features with the combination of prominent cognitive decline and a movement disorder were apparent. In summary, the finding from this cohort is an interesting extension to the spectrum. It illustrates, how the same mutation in independent families can present phenotypically with different ages of onset, different severity and diverse radiographic appearances. It represents another example of phenotypic heterogeneity, indicating that mutations always interact and exist in the context of the background genetic information, lifestyle and environmental influences. Tearing additional modifiers of genetic disease apart represents one of the biggest challenges of this and subsequent eras of research.

FA2H The clinical phenotype of both siblings with novel FA2H mutations is in concordance with previously described FA2H patients (147, 159, 368, 369). As in the case presented here, cerebellar atrophy and white matter lesions in FA2H positive patients have been described before, and thin corpus callosum seems to be more frequent than classic iron accumulation in FAHN – two characteristics that the patients described here did not reveal radiographically (159, 368, 370). Given the phenotypic overlap and the deleterious predictions of the mutation p.D266E plus the fact that mutation p.H261fs results in a frameshift, there is evidence that these novel mutations are the underlying genetic cause for the disease in this family. Patient II:3 was examined as a singleton as part of this cohort. When the novel compound heterozygous mutations described above (see Table 2-3 for details of mutations) were found, additional family members were recruited and tested (see Figure 3-7) for segregation. In the meantime of thesis submission, the older sister and older brother of the index patient have visited the outpatients’ clinics with their partners for family planning. Further carrier testing was performed and confirmed both siblings to be carrier of one but not both mutations, additionally supporting the interpretation of these mutations as the pathogenic ones in this family. Unfortunately, the mother’s DNA was not available for testing.

PANK2 Both mutations had been previously published as causative in homozygous states in patients with HARP and PKAN. The mutation c.568C>T:p.R190X (NM_024960) is identical to the C>T mutation at nucleotide 1111 in exon 5 that was found in a homozygous state in a patient with HARP syndrome by Ching et al. in 2002, establishing HARP as part of the PKAN spectrum (371). This mutation changes an

154

arginine codon to a stop codon at amino acid 190 (described as R371X in the original paper due to annotation to the longer transcript) and thereby shortens the PANK2 protein by 89 amino acids. The second heterozygous mutation identified in this individual (c.688G>A:p.G230R (NM_024960)) is located in exon 6 and identical to the common PANK2 mutation c.1561G>A:p.G521R/c.1231G>A:p.G411R which was identified frequently in homozygous and compound heterozygous fashion in numerous typical and atypical PKAN patients previously (103, 104, 372).

Unfortunately, this sample was a collaborator’s sample where no further family members could be contacted, and the original MRI with further clinical details could not be received. It was also impossible to receive more DNA (after confirmation of the mutations with Sanger sequencing (see Figure 3-8)) to perform allele specific amplification to prove mutations to be in trans. Given these facts, the finding of these presumably compound heterozygous mutations in a patient with early-onset NBIA phenotype is interesting. It is compelling to classify these mutations as pathogenic in this sample, however this remains non-finalised until further proof of compound heterozygosity, and of deleterious functional consequences in a compound heterozygous state for this combination of mutations becomes available.

FTL The common founder mutation 460InsA in FTL was identified twice in the described cohort. This mutation results in an altered reading frame and extends the peptide, thereby shown to disrupt the ferritin dodecahedron structure. Functionally, the abnormal structure goes along with pathological accumulation of ferritin and iron, primarily in central neurons (71, 74, 187, 189). Clinically, both presentations with progressive adult-onset dystonia/chorea and brain iron accumulation represent quite typical disease courses. The largest case series of

41 patients studied with this mutation revealed abnormal MRI in all of the studied patients, and focal dystonia/chorea as common presenting symptoms with a mean age at onset around 39 years (183). Unfortunately, the sparse clinical data available in case 2 hinders detailed comparison, however, from the level of detail known, the second case presented here appears more typical than the first case. The first case presented with a focal movement disorder, relatively early (25 years) compared to the mean age at onset, but within the range observed in the large case series. Additionally, patient 1 had normal serum ferritin levels at age of presentation (25 years), which is quite typical for premenopausal women with this mutation. However, interestingly, the negative family history is quite atypical in this case: Her four half-siblings, younger than herself, but beyond the age of disease presentation in her, are all reported to be healthy –

155

though with a range of disease onset up to 63 years it is possible that any of them are presymptomatic carriers, and the disease will still manifest, in case their shared mother was a rare, non-manifesting carrier. However, given the mother is reported to be asymptomatic, it is more likely that the mutation occurred de novo, or since the patient lost contact to her biological father, there is the possibility of the mutation being transmitted by him who is by now affected. Interestingly, all four children of the index patient are reported to be healthy, too. However, given their relatively young age, and neuroferritinopathy usually presenting in later adulthood, the disease can still manifest. Regrettably, none of the patient’s relatives were available for further testing or genetic counselling.

WDR45 The identified heterozygous mutation in WDR45 occurs at the same position as the mutation reported by Haack et al. in one of the two discovery papers. However, the mutation identified in this study is an insertion instead of the previously published deletion (76). This identified stopgain mutation leads to a premature stopcodon at amino acid 335 and subsequent early truncation of the protein. Full length WDR45 protein is implicated in early autophagy pathways and truncation of the protein is thought to result in defective autophagy, thereby exerting its pathogenic mechanism (76, 144-146). With the iPSC models currently set up in our laboratory, further insights into the pathogenic role of WDR45 mutations will be achieved.

WDR45 mutations have been established as the causal link in static encephalopathy of childhood with neurodegeneration in adulthood and NBIA, more specifically BPAN, previously (76, 135). The clinical presentation of the case identified here includes features of both SENDA (global developmental delay, and severe neurodegeneration in her twenties) and BPAN (iron accumulation on MRI, phenotypic spectrum) and is in line with previous reports of WDR45-positive patients. The unique occurrence in the index patient suggests a de novo event that is a common observation for this disease (76). However, given that DNA of further family members was not available, this remains speculation at this point.

Known dystonia, chorea, PD and HSP genes: single results FUCA1 The identified homozygous stopgain mutation c.1138G>T, p.E380X (NM_000147) was previously observed in a patient with fucosidosis, an early-onset autosomal recessive lysosomal storage disease (373). This mutation generates a premature stop codon 361 bp upstream of the innate stop codon and thereby is likely to act as a loss-of-function mutation. Fucosidosis belongs to the group of lysosomal storage disorders 156

representing a rare inborn error of metabolism where altered alpha fucosidase enzyme activity leads to build up of fucosylated compounds. Phenotypically, the defect usually leads to early-onset multisystem degeneration with variable presentation of facial dysmorphism, visceromegaly, angiokeratomas, movement disorder features, recurrent infections, dysostosis multiplex, progressive motor and mental retardation and premature death (373-378). MRI features include signal abnormalities in the basal ganglia (376, 377, 379) and radiographical signs reminiscent of the eye of the tiger sign have been reported (222). Unfortunately, the scarcity of clinical detail for this particular sample makes further interpretation and comparison with previous findings impossible.

PDE10A The finding of PDE10A mutations being causal for early-onset chorea established the link between this gene and early-onset chorea in humans for the first time (367). The striatum-enriched PDE10A protein belongs to a group of 21 different proteins within the cyclic nucleotide (cNMP) phosphodiesterase (PDE) family and regulates intracellular cNMP concentration mainly by degradation into monophosphate nucleotides (380, 381). Its regulatory N-terminal domain consists of a GAF-A and a GAF-B tail with GAF- B binding cAMP in order to regulate enzyme activity (382-384) - and importantly harbouring the pathogenic heterozygous mutations in all of our cases. Interestingly, our report was complemented simultaneously by findings of an independent group that established biallelic mutations in the GAF-A tail of PDE10A as causative for an autosomal recessive hyperkinetic movement disorder further supporting the importance of the intact enzyme in correctly regulating basal ganglia motor pathways (385). It furthermore points towards dual inheritance patterns (AR, AD) and differential pathogenic mechanisms for mutations in this gene depending on their location. Even though more research is certainly required, the PDE10A protein might potentially become an exciting drug target for hyperkinetic disorders due to its relatively well established functions, ongoing studies for the treatment of psychiatric disorders and existing pharmaceutical interest (386). Full details and discussion of these exciting findings however cannot be part of this thesis, and the interested reader is directed to the respective publications at this point (367, 385).

Variants of unclear pathogenicity LRRK2 The c.2264C>T, p.P755L (NM_198578) LRRK2 variant identified in the Asian patient with atypical PD was reported as pathogenic in Clinvar (CLNACC=RCV000032422.1; accessed: 13.04.2016), however it has been shown to be present in similar frequency in PD cases and healthy controls in Taiwanese, Japanese and Chinese populations as

157

well as present in sporadic PD patients (387-389). Even though the phenotype would be within the spectrum, this variant in this patient should therefore rather be interpreted as a probably rare polymorphism, especially given the individual originates from Asia, a population in which the frequency of this variant has been studied explicitly. It is however noteworthy that this variant had some deleterious in silico predictions, a GERP-score above 5 (GERP=genomic evolutionary rate profiling, see http://www.broadinstitute.org/~mgarber/GERP/documentation.pdf for further details) and a “pathogenic“ clinvar annotation. This example points towards the limitations of relying on in silico predictions and database comparisons again. It highlights the need of manual check of database provided information, segregation testing and ideally functional work for correct interpretation, which can be challenging to meet for each identified variant.

The variant c.3974G>A, p.R1325Q (NM_198578) identified in the Scottish patient with adult-onset dystonia-parkinsonism is a previously reported variant with unclear pathogenicity (390, 391). Additionally, an ADCY5 variant of uncertain significance was identified in the same individual. Unfortunately, no segregation analyses could be performed and final pathogenicity for any of the implicated variants cannot be shown.

The variants c.4235G>A, p.R1412Q and c.3716A>C, p.E1239A (NM_198578) were absent in all control databases except ExAC (frequency: 0.000008132 for p.R1412Q and 0.0000244 for p.E1239A). Both variants have to my best knowledge not been reported in patients before, had some deleterious in silico predictions and GERP- scores above 5. However, these criteria are not enough to show pathogenicity. Phenotypically, the Indian individual might fit an extended LRRK2-associated disease spectrum, whereas the complex HSP-phenotype of the Turkish individual, and his consanguineous background do argue against pathogenicity for the variant in this case. In summary, pathogenicity remains uncertain for the four identified LRRK2 variants.

NIPA1 The identified heterozygous missense variant c.661C>A, p.P221T (NM_144599) in NIPA1 is novel, was absent in all control databases except of ExAC (frequency: 0.00001626) and showed deleterious in silico predictions and high GERP- and CADD- scores (CADD=combined annotation dependent depletion, see (392) for more details). So far, there are few pathogenic missense mutations reported in NIPA1 causative of spastic paraplegia type 6 (SPG6, MIM #600363), a phenotype resembling pure spastic paraplegia in most cases, but rare additional complicating symptoms (peripheral

158

neuropathy, generalised idiopathic epilepsy) have been reported (393-396). With this finding, the phenotype identified here would add a very atypical presentation to the so far reported spectrum, however pathogenicity cannot be proven due to missing segregation analysis and functional investigation at this point.

SLC33A1

A heterozygous missense mutation in SLC33A1 (c.339T>G, p. S113R) has previously been shown to cause autosomal dominant spastic paraplegia type 42, a rare and uncomplicated spastic paraplegia subtype (SPG42, MIM #612539) and biallelic mutations in this gene have been associated with autosomal recessive congenital cataracts, hearing loss, and neurodegeneration (CCHLND, MIM #614482) (397, 398). The identified heterozygous missense variant in SLC33A1, the gene encoding an acetyl-CoA transporter, was absent in all control databases except for ExAC (frequency: 0.00000813) and had deleterious in silico predictions and moderately high CADD- and GERP-scores. Due to absence of further family members for genetic testing, a conclusion on its pathogenicity however cannot be made. The observed phenotype of early-onset dystonia with brain iron accumulation would be highly atypical of previously reported spastic phenotypes explained by heterozygous mutations within this gene. However, one cannot rule out the possibility of a second cryptic/missed or intronic mutation, or a copy number variant in this gene, contributing to dysfunction of the encoded acetyl-CoA transporter. Interestingly, dysfunction in the acetyl-CoA pathways has been shown to underlie other forms of NBIA (399).

L1CAM Mutations in the gene encoding the L1 cell adhesion molecule (L1CAM) have been associated with X-linked recessive spastic paraplegia type 1 (SPG1)/MASA syndrome (mental retardation, aphasia, shuffling gait, and adducted thumbs) and CRASH syndrome (corpus callosum hypoplasia, retardation, adducted thumbs, spastic paraplegia, and hydrocephalus) (MIM #303350) (400, 401). The patient with a rare variant in L1CAM suffered from learning disabilities, optic atrophy, dystonia and seizures and his MRI did not show any features resembling thin corpus callosum or hydrocephalus. Interestingly, cerebellar atrophy was observed. Given the unusual clinical features, the presence of this variant in ExAC and unavailability of segregation analyses one cannot conclude on potential pathogenicity for this variant.

KIF1A A rare heterozygous missense variant in KIF1A was observed in a female patient with spastic ataxic features. KIF1A encodes heavy chain kinesin member 1a, a member of

159

the kinesin superfamily motor proteins important for intracellular directional axonal transport of synaptic vesicles (402, 403). KIF1A-associated phenotypes range from autosomal dominant mental retardation type 9 (MIM #614255) and autosomal recessive spastic paraplegia type 30 (MIM #610357) to autosomal recessive hereditary sensory neuropathy 2C (HSN2C, MIM #614213) (404-406). Of these, the observed phenotype resembles most the SPG30 phenotype, however, upon close reinvestigation of the exome data still only one heterozygous variant in KIF1A could be identified. Due to absence of further family members, the association of this variant with the phenotype remains unclear.

SGCE Mutations in SGCE represent the major cause of the autosomal dominant myoclonus- dystonia syndrome (MDS, DYT11, MIM #159900) where – even though the disorder has reduced penetrance on maternal transmission - most SGCE-positive patients that manifest symptoms suffer from myoclonus followed by dystonia as the second most prevalent symptom (407-410). Large heterozygous deletions of SGCE are a frequent cause of MDS, however nonsense and missense mutations have been identified (411). The two identified heterozygous variants with in silico predictions that might argue in favour of an effect have to our best knowledge not been reported before and are absent in the control databases except for very low frequency in ExAC (see Table 3-3). However, given that the identification of two heterozygous variants without being able to prioritize and given the fact that the patient was a) born to a consangineous family (where initially one would expect a recessive disease), b) did not show any myoclonus, c) it was impossible to obtain more DNA of her healthy relatives, any other biomaterial of the affected individual to further study the pathogenicity, or further details on her clinical history (e.g. a recent examination, also in regard to potentially accompanying psychopathology, alcohol responsiveness of symptoms, etc.), the impact of the identified variants remains unclear at this point.

THAP1 The identified frameshift variant in exon 2 of THAP1 looks interesting and has to our best knowledge not been reported before. However, given the fact that clinical details were sparse on this sample, the early onset not particularly typical of autosomal dominant adolescent-onset isolated dystonia type 6 due to THAP1 mutations (412), and no further segregation studies could be performed, one cannot confidently assess the role of this variant for this sample.

160

Further variants of unclear pathogenicity Table 3-3 lists further variants that are not all discussed above. Final pathogenicity for these variants is unclear, the observed frequencies from control databases and in silico predictions were less convincing and the variants cannot be assessed properly due to missing DNA from further family members (e.g. variants in VPS35, ADCY5, ANO3, PNKD, KIF1C, SPAST). The lists further include a controversial variant in EIF4G1 (p.R1205H, corresponding to p.R1212H (NM_001194946) in the table of this thesis) that was initially published as pathogenic (413) but later on shown to be more frequent in controls than in cases (414); variants that were reassessed from potentially pathogenic but likely tolerable nature (390, 391) to the status of a potential rare risk factor (415) (e.g. LRRK2, NM_198578:c.3974G>A, p.R1325Q); and variants in genes where the inheritance pattern and/or phenotype did not match (e.g. LRRK2 variant with autosomal recessive complex HSP, KIF5A variant with predominant chorea, NIPA1 variant with myoclonus epilepsy phenotype). Particularly interesting is the identified heterozygous stopgain variant in VPS13B (NM_152564: c.11170G>T, p.E3724X) in a Turkish male individual with early-onset dystonia-parkinsonism and prominent psychiatric features. This exact same heterozygous variant has been published in an individual with Cohen syndrome previously (416). Cohen syndrome is an autosomal recessive multisystem disorder with frequent occurrence of postnatal microcephaly, mental retardation, pigmentary retinopathy, facial dysmorphism, myopia, and intermittent neutropenia, features that were not reported in the patient described here. The authors of this paper speculated cryptic copy number variation (CNV) of the other allele contributing to pathogenicity in their case. Careful search for additional rare, deleterious variants in VPS13B in this individual however only revealed frequent SNPs. One cannot however rule out the presence of large deletions or copy number variants missed by the isolated approach of exome sequencing, even though the phenotype in the patient is not typical for this disease. Due to lack of DNA from family members, segregation could not be assessed in this and the other individuals with variants of unknown pathogenicity for which the pathogenic nature remains obscure due to lack of evidence.

Interesting variants picked up by exome sequencing, but not confirmed by Sanger sequencing These 10 false positive variants (Table 3-4) are most likely due to false positive variant calling and/or low coverage on the respective parts of the exome. Exome sequencing technologies and data alignment algorithms improve constantly which will further reduce these false discoveries in the future.

161

Further remarks and novel candidate genes The low frequency of identified pathogenic variants and the high frequency of familial cases in the presented cohort suggests the cryptic presence of additional disease causing genes. Novel gene identification is more challenging and especially difficult in this study since it includes a relatively small sample size with high clinical heterogeneity and only few strong family histories where DNA from the large pedigree could be obtained. However, two ongoing approaches to identify biologically meaningful novel candidate genes involve looking for 1) rare homozygous stopgain variants with enriched expression in the brain, and 2) for rare, nonsynonymous variants with a known functional role in iron pathways in the brain. In parallel work, my colleague Dr. C. Bettencourt identified several genes coexpressed with other causal NBIA genes (417). These lists of enriched genes were used as seeds to analyse the exome data for potential novel disease causing genes. However, please note that the preliminary results of this work need further validation and therefore cannot yet be reported here (see section on Further ongoing work below).

Limitations Whole exome sequencing is a fast, reliable and comparably affordable technique that allows screening of known genes and identification of novel ones in disease cohorts as well as individuals or small families. However, there are some general caveats that need to be taken into account. Firstly, whole exome sequencing can reliably identify single nucleotide variants and small deletions, however it is not reliable yet in detecting copy number variants, repeats or large deletions/insertions which have been shown to be underlying disease in a considerable number of neurodegenerative disorders. Secondly, the exome is not evenly covered by the current exome capture technologies, which results in insufficient coverage for certain areas in which disease causing mutations can easily be missed. Thirdly, some neurological diseases are caused by tandem repeat expansions which can be extremely challenging to detect in exome sequencing reads due to the usage of short-read libraries (~250 bp) when generating the data. Finally, the gene harbouring the causative mutation for an individual may not (or insufficiently) be covered in the exome capture target kit definition. These are general concerns that apply to using whole exome sequencing in general. For the described cohort, 100% exome coverage was not available in particular and therefore the approach is likely to miss more variants, unnecessarily adding to the technical limitations at this time mentioned further above.

162

Three additional major points need to be specifically added to these rather technical caveats when it comes to the study outlined above: Firstly, the exome sequencing approach was not complemented with routine CGH array for each of the samples, or even for those with unknown genetic cause. This is a limitation of this study, since especially copy number variations are often enriched in patients with developmental delay, mental retardation and psychiatric phenotypes as present in the cohort investigated here. One therefore cannot exclude to have missed disease causing copy number variants in this cohort. Secondly, NBIA disorders are very heterogeneous and one limitation of the analysed cohort is that very few samples are pathologically proven NBIA cases. Thirdly, the sample size may be too small to detect novel pathogenic variants reliably or to conduct case-control studies and check for enrichment of pathogenic variants in the disease cohort. To address these points, the cases of this study were classified into clinical subgroups according to their main symptoms and mode of inheritance to create more homogeneous subgroups. However, this decreases sample size further for possible case-control studies but might make it easier to detect shared variants – if present - between a subset of patients. In collaboration with Dr. Boniface Mok from IoN and the bioinformatics team from Prof. Andy Singleton at NIH additional analysis strategies are developed to reduce the number of missed pathogenic variants.

Further ongoing work Even though it is very likely that a sample size of 91 patients with partly reduced exome coverage is not enough to detect any statistically significant effects, the whole cohort as well as the patient-subclasses will be queried for enrichment of novel and rare variants in comparison with controls via different burden testing methodology including sequence kernel association tests (SKAT) as has been shown efficient in small sample case-control studies (418). Of note: > 250 controls have been collected by Andrew Singleton’s laboratory and the exome sequencing data is readily available in the institute collaboratively. In joint work with Dr. Alan Pittman and Dr. Qiang Gang the necessary strict quality control and statistical settings are currently set up to perform case-control studies at the IoN.

In a parallel approach, rare novel candidates in this cohort are analysed. A combination of different filtering strategies are used to analyse the currently long lists of variants (essentially, one exome of an individual with Caucasian origins may show more or less 20000 different variants whereof only one or two, depending on the suspected inheritance mode, are considered pathogenic according to Mendelian criteria). As above, a variety of publicly available databases (ExAC, dbSNP, 1000g and NHLBI EVS

163

– each of which has their own advantages and disadvantages) are used to exclude variants based on their population frequency, similar to the candidate gene study presented above. When defining NBIA-like disorders as Mendelian diseases the target variant should be absent, or extremely rare within the general population (recessive disease alleles are usually “allowed” to be present at a minor allele frequency < 0.05 whereas for a dominant variant one usually only considers novel variants – however, these cut-offs are somewhat arbitrary). It has been shown in the past and needs to be taken into account that pathogenic mutations (at low frequencies) can nonetheless be present in commonly used population variant databases (e.g. example of LRRK2 G2019S variant). Therefore filtering for minor allele frequency always needs to be done with a certain degree of consideration which takes into account the expected frequency of the disease in the population, the age at onset in the disease population (e.g. for late-in-life manifesting diseases pathogenic variants are more likely to be present in public databases than for early-onset diseases) and the expected mode of inheritance. The clinical subgrouping as described above will allow to more flexibly adapt filtering strategies according to subgroups. Inheritance mode will further direct which variants to consider as possibly pathogenic (e.g. heterozygous or homozygous variants). Functional annotation of variants and subsequent ranking based on predictions of their deleteriousness will again be quite essential: Per se synonymous variants have been excluded from the analysis, as it is expected that causative alleles for Mendelian diseases will likely be frameshift, splice site, missense or nonsense variants. Using in silico tools to predict deleteriousness (e.g. SIFT (http://sift.jcvi.org/), Polyphen2 (http://genetics.bwh.harvard.edu/pph2/), Mutation Taster (http://www.mutationtaster.org/), and taking evolutionary conservation throughout species into account (phastCONS (http://compgen.cshl.edu/phast/help- pages/phastCons.txt), phyloP (http://compgen.cshl.edu/phast/help-pages/phyloP.txt) and GERP-rating (http://www.broadinstitute.org/~mgarber/GERP/documentation.pdf)) the list of potential candidates is currently analysed. Where DNA of other family members is available or can be obtained, the list of candidate genes will be further reduced by confirmation of interesting variants in the affected individual and segregation analysis in the family via Sanger sequencing. As a first subapproach to this all homozygous stopgain variants that are detected in this cohort (n=160) have been subsetted and filtered for frequency. Their expression patterns are currently analysed to find biologically meaningful genes with a role in neural maintenance and neural activity where stopgain or stoploss could provide a potential pathogenic mechanism causing complex and multifaceted phenotypes as in NBIA.

164

As a second strategy and briefly mentioned above, promising variants in genes that are coexpressed with known pathogenic NBIA genes are looked at. These module- separated gene lists have been generated by my colleague Dr. C. Bettencourt in the publication investigating weighted gene coexpression networks in NBIA (417). In a first step, all variants in the genes of the two most enriched modules were isolated, which yielded 4002 stopgain/stoploss, nonsynonymous or frameshift variants in the whole cohort, whereof, when filtered for novelty and homozygosity (if one assumes most NBIA disorders to be caused by homozygous/biallelic variants), 15 candidates were obtained which will need further follow-up regarding their implicated biological pathways and expression profiles. However, the analysis on these questions is not yet conclusively finished and therefore results cannot be reported here.

Thirdly, in collaboration with the NIH in Bethesda, in silico protein-protein-interaction analyses will generate a list of new interactors with genes/proteins that are already known to be implicated in NBIA. These will be used as seeds to find promising candidates and to mine down the exome data. Recently, Novarino et al. carried out similar work for the HSPs (28), and the NIH group is in the process of setting protein- protein-interaction-analyses up successfully.

Fourthly, PKAN and other (NBIA) disorders with a defined genetic mutation are studied in vitro via the pathophysiologic analysis of induced pluripotent stem cell derived neurons that carry the exact genetic information as the affected patients. For a collaborative project led by Dr. Selina Wray and Dr. Charlie Arber (Reta Lila Weston Institute, UCL), four patients with PANK2 mutations were biopsied, and fibroblasts from two patients with COASY mutations were collected from external collaborators. The reprogramming to iPSCs and their characterisation has been carried out successfully in all lines for both diseases and several batches of neurons from these iPSC clones of three PKAN patients and three controls were studied collaboratively at the Reta Lila Weston Institute. The awaited final results will give a better idea of the mechanisms and triggers of iron accumulation, transport deregulation, mitochondrial fitness including ATP-generation, mitochondrial membrane potential, ROS production as well as generation of tau-, and acetyl-CoA-protein and RNA-expression in human PANK2- defective neurons. Finally, these studies will lead to better understanding of disease mechanisms driving the cellular phenotype in this human neuronal model. They can help to understand clinical symptoms and represent a human neuronal model to explore ways to alter or halt neurodegeneration that ultimately could help with disease management in these severely affected patients.

165

Conclusions and summary In summary this study identified a range of novel and known mutations in known disease associated genes and identified a novel disease causing gene. It has identified a highly likely genetic cause for 8.8% of patients and variants of unclear pathogenicity in known neurodegenerative genes in an additional 23%. The overall observed clinical heterogeneity was high, again confirming the high clinical variability and common overlap to other neurodegenerative diseases that characterises NBIA-like disorders. Several extensions to known disease spectra could be described. It is clinically interesting and relevant for clinical practice that mutations in known NBIA genes were observed in patients without readily apparent MRI iron, and vice versa that patients with abnormal iron deposits were negative for all known NBIA genes.

Generally, cases were clinically well phenotyped and the current cohort size is quite large for an orphan disease like NBIA. However, some cases could not be returned to after identification of potential genetic candidates or family members were not available for segregation studies. This hindered efficient interpretation of genetic data and represents a general shortcoming of genetic cohort studies highlighting again the necessity of detailed clinical phenotyping, good collaborator reports and routine collection of genetic and clinical data of the whole pedigree in clinical practice where available.

The sequencing quality was good to average, compared to other published cohort studies. However, this study only employed one single sequencing technique (WES) with inherent limitations and could not yet submit the DNA of genetically unresolved cases to CGH-arrays or better whole genome sequencing pipelines (as e.g. offered by the 100K Genomes England project for patients from clinical practice within the NHS) which represents another current limitation. The fact that the analysis pipeline for the genetic data reported here needed to be established first was additionally time- consuming. Parents of here sequenced probands as well as more NBIA patients collected over time will be included in future study designs to complement this data, and help variant filtering and disease gene identification. Interesting variants (especially of the newly identified gene(s)) should be validated using a range of functional readouts, e.g. overexpression studies in simple model systems, study of transgenic mouse models or induced pluripotent stem cells as exemplified in the later chapters of this thesis.

166

The following chapter reports successful gene/mutation identification and detailed clinical characterisation in several pedigrees with overlapping symptoms of neurodegeneration.

167

4 Chapter 4: Genetics in families with rare neurodegenerative movement disorders: Examples from McLeod syndrome and hereditary ataxias

4.1 McLeod syndrome: Identification of a novel single base-pair deletion

4.1.1 Introduction

There is a significant clinical overlap of symptoms and pathogenesis between NBIA disorders and neuroacanthocytosis syndromes. Amongst the neuroacanthocytosis syndromes, McLeod syndrome (MCLDS, MIM #300842) is a very rare X-linked multisystem disease caused by mutations in the XK gene, which is thought to encode for a ubiquitously expressed membrane transporter (419) with as yet unidentified function. The protein product of XK is composed of 444 amino acids and has a calculated molecular weight of 50913 daltons.

MCLDS was originally described as a predominantly haematologic condition with weakly expressed Kell antigens on the red cell membrane as the pathognomonic feature (420). The following years appended haemolysis, acanthocytosis, hepatosplenomegaly and the X-linked inheritance to the clinical picture (421, 422). The wide recognition of neurological symptoms that finally led to the inclusion of this disease into the neuroacanthocytosis group came with delay (423-426). The neurologic abnormalities in patients affected by XK mutations are mirrored by high levels of XK expression in the brain in normal controls: Variably occurring striatal degeneration with the development of chorea might therefore be a consequence of the protein’s mislocalisation, dysfunction, reduced or missing activity in MCLDS. Furthermore, XK is highly expressed in skeletal and cardiac muscle which might explain the presence of late-onset muscular dystrophy and cardiomyopathy that are frequent features of this syndrome.

The rare McLeod syndrome forms the core of the neuroacanthocytosis syndromes together with autosomal recessive chorea-acanthocytosis due to biallelic mutations in VPS13A. Together, they feature an estimated prevalence of less than 1-5/1000000 world-wide (427). With an onset of neurological symptoms mostly around the 4th decade, MCLDS is a slowly progressive multisystem disease consisting of acanthocytosis, haemolysis, chronic anemia, hepatosplenomegaly, cardiomyopathy, myopathy, neuropathy, choreatic movement disorders, psychiatric manifestations and cognitive decline as the most common symptoms. Additional occurrence of rare 168

multisystem features such as chronic granulomatous disease (CGD) are possible (428). Not only the clinical phenotype of MCLDS but also the mutational spectrum of the XK gene shows heterogeneity, with gross deletions, spanning the entire or a significant part of the XK coding region, as well as small frameshift indels, and single nucleotide substitutions described as associated with the phenotype. The common ground of the diverse genetic variation is the prediction of an absent or truncated, non- functional gene product (428). The exact function of the XK gene in humans is only partially eluded and functional studies rare. Structurally, XK was found to share characteristics with membrane transport proteins of other pro- and eukaryotes (429).

Below, a patient with a novel XK frameshift deletion in exon 1, associated with a MCLDS without signs of CGD, was identified and clinicogenetically characterised.

4.1.2 Results

Patient presentation A 70-year-old Greek man was referred to the neurology department of the Papageorgiou Hospital in Thessaloniki by the hospital’s haematology department. The main initial findings were choreatiform tongue movements that had been noted by the physicians from the haematology unit. They had previously detected a weak expression of K red blood cell antigens, acanthocytosis, haemolysis, and a form of factor IX deficiency when screening the patient’s blood type prior to a blood transfusion needed for a scheduled routine surgical procedure. Those progressive tongue movements were noted recently and were attributed by the proband and his family to his ill-fitting denture. Upon closer inspection, the patient also had a slight generalised restlessness with frequent changes of posture but without any tic-like or choreiform movements of any particular part of his body apart from his tongue and the face.

Further neurological findings on examination besides chorea of the face and tongue included lower limb muscle weakness with muscle wasting. The patient’s past medical history includes severe progressive skeletal deformities after the age of 50, with osteoarthritis of the knees bilaterally that resulted in severe varus deformities of the knees. His family history reveals that the patient had a brother who died at the age of 60, and who presented with severe facial dyskinesias, chorea and schizophrenia-like psychosis. The patient’s sister, who died at the age of 70 after an acute ischemic stroke, was referred to as being healthy until that incident. The patient’s 38-year-old daughter and his 35-year-old son are both healthy (for the pedigree, see Figure 4-3, A). There is no additional family history regarding neuropsychiatric diseases.

169

Clinical investigations Laboratory investigation revealed blood type group A1, DAT negative (-), IAT positive (+). Antibody identification panel testing demonstrated panagglutination of the same intensity in all cellular lines in the LISS/Coombs phase. Enzyme phase, 4°C incubation phase and auto control (A/C) were negative. His detailed blood group profile was: C (+), c (+), D (+), E (-), e (+), Cw (-), K (-), k (-), Kpa (-), Kpb (-), P1 (+), Lea (-), Leb (+), Lua (-), Lub (+), Jka (+), Jkb (+), M (+), N (+), S (+), s (+), Fya (-), and Fyb (+). Peripheral blood smear revealed the existence of acanthocytes (see Figure 4-1) at a percentage of 19% of all red blood cells. Other abnormal laboratory values included: Factor IX 8% (normal values 60-150%), Activated Partial Thromboplastin Time (APTT) 58.1 sec (normal value 30.0 sec), Partial Thromboplastin Time Lupus Anticoagulans (PTT-LA) 61.3 sec (normal value <45 sec), Lactate dehydrogenase (LDH) 299 U/l (normal values 135-225 U/l), creatinephosphokinase 1400 U/l (normal range in males 24-195 U/l). All further routine blood tests were normal. Electrophysiological examination showed diffuse chronic sensorimotor peripheral neuropathy of axonal type. A brain MRI did not show any abnormal findings apart from mild cortical atrophy (see Figure 4-2), no striatal atrophy could be observed. Neuropsychological investigation and EEG were normal. The patient did not present any cardiac complications and had a normal echocardiogram and 24-hour monitoring.

Figure 4-1: Acanthocytes in peripheral blood

The smear shows frequent acanthocytes, representing ~19% of all red blood cells (Figure published in (430)).

170

Figure 4-2: T1- and T2-weighted MRI McLeod case

Proband’s T1- (left panel) and T2-weighted axial (right panel) MRI at time of presentation without major abnormalities apart from mild cortical atrophy (Figure published elsewhere and reproduced from (430)).

Genetic analysis Sequence analysis using Sanger sequencing of the three exons of the XK gene detected a hemizygous single base-pair frameshift deletion (see Figure 4-3, B) at exon 1 (c.229delC, p.Leu80fs), leading to a very premature stop codon (see Figure 4-3, C). Exons 2 and 3 revealed no further mutations in this gene. The deletion is absent in the patient’s healthy son and is present in his unaffected daughter in the heterozygous state (see Figure 4-3, B). To the best of my knowledge, the variant has not been previously reported and is not present in public databases, such as dbSNP, 1000 Genomes and Exome Variant Server. Due to the lack of further biomaterial (e.g. skin fibroblasts, liquor or fresh blood), no confirmatory analyses assessing the functional consequences of the deletion could be performed.

171

Figure 4-3: Family tree, segregation and conservation status of novel XK mutation

A) Family pedigree. The arrow indicates the proband. Black filled symbols indicate affected individuals, while half-filled indicate unaffected carriers of the mutation; B) Chromatograms depicting the mutation segregating in this family; C) Protein alignment showing conserved positions across species and the very premature stop codon (marked with the asterisk) caused by the c.229delC frameshift mutation (Abbreviations: WT – human wild-type allele; Mut – human mutant allele). This figure is published elsewhere and has been reproduced from (430).

4.1.3 Discussion

MCLDS belongs to the rare neuroacanthocytosis syndromes with approximately a few hundred cases world-wide (431). The disease distribution lacks obvious clusters and

McLeod syndrome has been described across Northern and Southern America, Europe, Japan and recently China (431, 432). It is heterogeneous in its clinical as well as its mutational spectrum. So far, gross deletions, including large parts of the XK gene or even nearby genes, small frameshift indels as well as point mutations have been described, all resulting in the absence or the truncation of a functional gene product (428). In the patient described here the neurological phenotype became obvious rather late in life and his haematological abnormalities were revealed coincidentally during routine blood group testing for a scheduled surgical procedure. The diagnosis of MCLDS is generally based on the haematological findings of weak Kell antigens. Therefore, carriers of this haematological phenotype could be diagnosed mainly in blood banks

172

(432). However, the majority of cases has been diagnosed by neurologists due to prominent neurological manifestations. The wide phenotypic spectrum of the disease and the insidious onset usually delays the diagnosis. Many more patients could be diagnosed properly if there was a thorough follow-up of all carriers of the haematological McLeod blood group phenotype that are discovered coincidentally, as in the patient described here.

The genetic testing in the described patient revealed a novel single base-pair frameshift deletion at exon 1 of the XK gene, which leads to a very premature stop codon. The resulting short mRNA product will most probably be subject to nonsense- mediated mRNA decay leading to loss-of-function as the most likely mechanism of pathogenicity in this patient. A second possibility would entail the short mRNA potentially escaping normal mechanisms of nonsense-mediated mRNA decay. However, even if some protein would be produced under these premises, it would lack more than 330 amino acids (normal size 444 amino acids) and therefore is highly unlikely to function normally. Despite lack of a formal functional confirmation, given the early location of the deletion in exon 1 and its introduction of a premature stop codon the deleteriousness of this frameshift deletion is highly probable. This is an illustration of a diagnostic success important for genetic counselling of both patient and his daughter especially, who was identified as a carrier and is still in her reproductive age.

Hereby the mutational spectrum of XK mutations leading to McLeod syndrome was successfully expanded.

Conclusions With this work, a novel XK frameshift mutation was identified in a 70-year-old male patient of Greek origin with choreatic movements of the tongue and face, lower limb muscle weakness, peripheral neuropathy, elevated creatinephosphokinase (CPK), acanthocytosis and haemolysis in the absence of Kell red blood cell antigens with an additional factor IX-deficiency and without signs of chronic granulomatous disease. The previously unreported hemizygous single base-pair frameshift deletion at exon 1 (c.229delC, p.Leu80fs) leading to a premature stop codon could be established as the causal link in this rare phenotype of a patient with McLeod syndrome that was discovered coincidentally during routine blood testing. This finding has important counselling consequences for both patient and his daughter who was identified as a carrier.

173

It reveals the importance of knowledge of these rare disorders, and the correct and thorough neurologic and multisystem investigations upon detection of weak Kell antigens.

174

4.2 SYNE1: Novel mutations and extension of the clinicogenetic spectrum

4.2.1 Introduction

The recessive spinocerebellar ataxias are an expandingly complex group of neurodegenerative conditions with significant genetic and phenotypic heterogeneity often delaying diagnosis (433). They are usually characterised by their early-onset multisystem features that can present with a range of neurological manifestations frequently including visual disturbance, pyramidal affection, epilepsy, motor and cognitive deterioration and neuropathy. To date there are over 70 genes that can cause recessive ataxia and this includes 21 SCAR genes (242). In 2007, SYNE1 was discovered as the first underlying genetic cause within the recessive ataxias associated with a pure cerebellar phenotype. It was termed recessive ataxia of Beauce (SCAR8 or ARCA1, MIM #610743), since it was originally identified in a number of French- Canadian families originating from the Beauce and Bas-St-Laurent regions of Quebec (434).

Subsequent investigations - before the recently published largest to-date multi-centre study by Synofzik et al. with 34 novel mutations (435) - had identified more than 11 biallelic truncating mutations and one missense mutation in individuals originating from Canada, France, Brazil and Japan (436). In 2013, a single case from Japan, and fairly recently in 2015, two families with motor neuron disease from Turkey with novel homozygous truncating mutations in SYNE1 have been reported, indicating an additional novel and at earlier times unrecognized association with motor neuron phenotypes for this previously relatively pure cerebellar ataxia (437, 438).

Furthermore, heterozygous missense mutations in SYNE1 have been identified in two unrelated probands with Emery-Dreifuss muscular dystrophy 4 (EDMD4, MIM #612998) where only the exons contributing to the muscle specific isoform of SYNE1 were investigated (439), and homozygous acceptor splice site mutations 2 basepairs 5- prime to exon 137 were identified in a consanguineous Palestinian pedigree with myogenic arthrogryposis multiplex congenita (440). In addition defects in SYNE1 were enriched in exome sequencing studies of mental retardation and autism (441, 442). SYNE1 encodes the spectrin repeat-containing nuclear envelope protein 1, a structural protein expressed in various tissues and believed to link the nucleoskeleton to the inner and outer nuclear membrane, to membranes of cell organelles, to the actin cytoskeleton, and to the sarcomere in muscle (439, 443, 444). Data from mice suggests a critical role in neurogenesis and neuronal migration for SYNE1 as an

175

organiser of nucleokinesis in the interplay with other complex proteins such as SYNE2, SUN1 and SUN2 (445). However, its direct functional role in the human CNS, and more particularly in the cerebellum, remains understudied.

In this study, an ethnically diverse UK cohort of recessive and sporadic cerebellar ataxia patients was investigated through a combination of targeted next-generation sequencing and exome sequencing, identifying four novel truncating mutations in a further four cases from a British, a Turkish and a Sri Lankan pedigree with pure cerebellar ataxia with mild cognitive impairment (all) and pyramidal tract signs in only the Turkish case and an axonal neuropathy only in the Sri Lankan pedigree.

4.2.2 Results

Family I Patient I:1 and I:4 are siblings from a non-consanguineous English family (see Figure 4-4, A). No complications during pregnancy or birth were mentioned on anamnesis. Both achieved normal motor milestones although unlike their two other siblings were noted to be poor at sports in school and had poor handwriting. There was however no concern about gait or balance until much later in life. Patient I:1 sustained a significant head injury at the age of 21 and was noted to have slightly slurred speech after his accident which was attributed at the time to multiple jaw fractures. There was no noticeable change in gait or coordination until the age of 40 when he noticed difficulty rising from a chair and sustained several falls when walking. He is currently 65 years of age and now has significant gait ataxia and dysarthria. His 49-year-old sister (I:4) did not notice any deterioration in her gait until age 32 after she also was involved in a road traffic accident. She has noticed a slowly progressive deterioration in her gait and speech since this time.

Examination of both siblings was very similar with both exhibiting clinical signs of cerebellar ataxia with broken ocular pursuits, dysarthria and limb and gait ataxia. Reflexes were normal and there was no clinical and electrophysiologic evidence of neuropathy. The available cognitive scores of the proband I:4 reflected cognitive underfunctioning. On focal cognitive tests, the main findings were weak arithmetic skills, poor visuospatial skills, evidence of executive dysfunction and reduced speed of information processing. Performance on tests of memory and naming were satisfactory, and visuoperceptual skills were intact.

176

Investigations showed both siblings had cerebellar atrophy on MRI (see Figure 4-4, B for patient I:4) with routine blood tests and electrophysiologic assessment being normal.

In both affected siblings, two novel compound heterozygous truncating mutations were identified in exon 18 (c.1849G>T:p.E617X) and exon 99 (c.18431G>A: p.W6144X) of SYNE1. One of the two unaffected siblings was available for testing and carried only the exon 99 mutation confirming the mutation phase to be trans (see Figure 4-4, C).

Figure 4-4: Clinical and genetic findings in family I

Family tree (A), sagittal T1-weighted MR-imaging (B) and sequencing chromatograms of two novel compound heterozygous truncating mutations segregating in family I. This figure is published elsewhere (1).

Family II Patient II:1 is one of 11 siblings from consanguineous parents (1st cousins) of Turkish origin. She was completely well until age 18 years when she developed progressive gait ataxia and dysarthria. At age 23 she was still mobilising without aid but had a high frequency of falls and required adaptations to her home to ensure safety. The patient is now 34 years old. She has 4 affected siblings who all developed symptoms with onset

177

in their late teens and were reported to be very similar to the index case although unfortunately neither clinical notes, nor DNA were available as none of the other family members are resident in the UK and contact has been interrupted.

Clinical examination of the index patient showed cognitive difficulties at the bedside but she was not formally assessed. There was broken pursuit eye movements, mild to moderate finger nose ataxia and marked gait ataxia. Reflexes were brisk in the lower limbs with sustained clonus at both ankles and extensor plantar reflexes. Brain MRI of the index case showed marked cerebellar atrophy. Nerve conduction studies showed no evidence of peripheral neuropathy. A novel homozygous variant in exon 108 of SYNE1 (c. 19897C>T p.Q6633X) was identified in the proband (see Figure 4-5 A, left panel). DNA for the parents and siblings was not available.

Figure 4-5: Genetic and MRI findings in patients II:1 and III:1

Chromatograms of two novel mutations identified in patient II:1 (left) and III:1 (right) (A) and sagittal T1-weighted MR-imaging of an unaffected control proband, and patient III:1 at age 32 (B). This figure is published elsewhere (1).

178

Family III The proband (III:1) is a 38-year-old Sri Lankan man who originally developed gait and balance problems at 22 years. He is the second of five siblings that are all well and without neurological problems. The patient’s parents were not known to be related. He was fit and active in sports as a child and has despite the obvious gait difficulties remained independent without usage of walking aids till present. Clinical examination showed hypermetric saccades, limb ataxia and extensor plantar reflexes although reflexes were reduced. There was no limb spasticity. His cognitive scores were mildly impaired on both the verbal and performance scales of the Wechsler Adult Intelligence Scale, edition three (WAIS-III). On a series of focal cognitive tests he presented with a poor performance on some tests of visual memory. In addition, processing speed was reduced and performance was poor on a test of selective attention. This patient was additionally videotaped for internal teaching and publication purposes.

Brain MRI showed cerebellar atrophy particularly over the vermis (see Figure 4-5, B). Nerve conduction studies showed an axonal neuropathy, other screening blood tests were normal.

A novel, truncating homozygous variant in SYNE1, exon 77 c.13429C>T:p.Q4477X, was identified (see Figure 4-5, A, right panel).

4.2.3 Discussion

To date, homozygous loss-of-function mutations in SYNE1 have been reported in French-Canadian, French, Japanese, Turkish and Brazilian individuals (436-438). With the first British and Sri Lankan cases detected in this study, the ethnic diversity underlying SYNE1-associated cerebellar ataxia was further extended. A summary of pheno- and genotype of these and other recently reported SYNE1 mutations (excluding the initial discovery paper 2007 and excluding the 34 novel truncating mutations published recently by Synofzik et al. (435)) can be seen in Table 4-1.

All novel cases reported in this study are associated with truncating mutations likely to result in production of a truncated protein or a complete loss or severe reduction of protein function. There appears to be some degree of negative correlation between the position of the stop codon in the reading frame and the severity or age at onset in the cases presented here: E.g. the Turkish case has the most 3’ prime mutation of all four, however this patient presented with the earliest age at onset (18 years) and additional pyramidal manifestations absent in the other three cases. Interestingly, the novel

179

mutation in the Turkish female individual has since been found once in the very recent, large multi-centre study (435). The age at onset, gender and ethnicity of the patients is the same. However, the clinical features are cerebellar ataxia plus motor neuron signs for the patient described in this thesis and cerebellar ataxia plus upper motor neuron affection with additional features including strabism divergens, fasciculations in the face, fibrillations in the tongue, reduced vibration sense, depression and CK elevation for their patient (435) which shows significant phenotypic differences.

Additionally, looking at cases I:1 and I:4 from the NHNN series, phenotypic heterogeneity can be observed within families. Even though based on very low numbers of patients, these observations are mirrored via the other recent reports in the literature (see Table 4-1 for the most recent ones, excluding Synofzik et al. (435)) and SYNE1-associated cerebellar ataxia can present as a mild and slowly progressive disease. The affected cases from family I are fairly typical of previously reported cases with a pure cerebellar phenotype without neuropathy or significant pyramidal signs but with signs of executive and cognitive dysfunction. The Sri Lankan case had a slowly progressive cerebellar syndrome, however with an additional axonal neuropathy. Cognitive decline was evident in all three families, which has been reported in the past in an extensive report on patients from Quebec (446).

The proband from family II however exhibited marked lower limb pyramidal signs with hyperreflexia and ankle clonus. Of note there were no variants in any other (spastic) ataxia or hereditary spastic paraparesis genes identified in the exome sequence data of this individual. Interestingly, the same mutation has since been identified in a patient with cerebellar ataxia, upper motor neuron and multisystem affection by Synofzik et al. in their multi-centre study (435). Previously, two other Turkish families had been described with affected members manifesting a spastic ataxia and motor neuron phenotype associated with two novel homozygous SYNE1 nonsense mutations (R7842X and Q7644X) (438).

See Figure 4-6 for locations of all reported mutations excluding the 34 novel mutations published by Synofzik et al. (435) that were reported just very recently. Mutations associated with motor neuron phenotype are represented by yellow filled boxes.

180

Figure 4-6: Recently reported and novel SYNE1 mutations, their location and phenotype

Locations of reported mutations of SYNE1-associated ataxic phenotypes. Mutations in black are from the discovery study, blue mutations are mutations reported since and red are the mutations identified in this study. Yellow filled boxes mark mutations associated with motor neuron phenotypes. Dotted lines connect the further 3’ mutation with their partner mutation in compound heterozygous (c. het) cases. Cave: sizes of black bars do not proportionally represent the sizes of the different 146 exons of SYNE1. Note: This diagram does not include the 34 novel mutations published by Synofzik et al. (435). This figure is published elsewhere (1).

Interestingly, all recent mutations presenting with significant pyramidal signs were located towards the 3’-end of the gene and seemed to be homozygous mutations and not compound heterozygous mutations. However, as with the age at onset and disease severity, this observation is interesting, but based on a very low number of cases. Furthermore, one of the homozygous mutations reported in the initial study in 2007 (p.Q7640X) is located far 3’ of the gene as well, but has been reported as pure cerebellar ataxia as part of the initial cohort (see Figure 4-6). Furthermore, the recent paper from Synofzik et al. suggests the picture to be more complicated as it identifies a wide range of neurological symptoms additional to motor neuron phenotypes associated with truncating SYNE1 mutations throughout the large gene and lacks to reveal a clear association of motor neuron phenotypes with 3’-mutation locations only.

181

Table 4-1: Recently reported SYNE1 mutations after discovery paper

Patient Patient AAO Clinical Ethnicity Mutation Ref. Nr ID (yrs) features NHNN- White 1 I:1 21 p.E617X + p.W6144X Pure ataxia British cohort NHNN- White 2 I:4 32 p.E617X + p.W6144X Pure ataxia British cohort Ataxia + NHNN- 3 II:1 Turkish 18 hom p.Q6633X pyramidal cohort tract signs Pure ataxia NHNN- 4 III:1 Sri Lankan 22 hom p.Q4477X + cohort neuropathy French- 5 II:10 30 p.R125X + p.W6620X Pure ataxia (436) Canadian French- 6 III:1 14 p.R2906X+ p.R7084X Pure ataxia (436) Canadian Ataxia - no 7 Case A Brazilian na hom p.Q1300X further (436) details hom c.10753- Ataxia - no 8 Case B French na 10757delCCAAG further (436) /predicted p.R3432Vfs*4 details Ataxia; Patient hom p.R7486fs7488X pyramidal 9 Japanese 6 (437) 1 (+ hom p.G185R) tract signs, MND Patient 10 Japanese 36 hom p.R3597X Pure ataxia (437) 2 Patient 11 Japanese 27 hom p.Y4534fs4539X Pure ataxia (437) 3 Ataxia; hom p.Q7644X pyramidal 12 SYNE1 Turkish 20 (=p.Q7573X for (438) tract signs, ENST00000423061) MND Ataxia; hom p.R7842X pyramidal 13 SYNE2 Turkish 26 (=p.Q7771X for (438) tract signs, ENST00000423061) MND

Demographic, genetic and clinical summary of SYNE1 mutations observed so far

(excluding Synofzik et al. (435)). Abbreviations: AAO=age at onset, hom=homozygous, na=not available, MND=motor neuron phenotype, Ref.=Reference, NHNN=National Hospital for Neurology and Neurosurgery. (all mutations for ENST00000423061/RefSeq protein NP_149062, unless otherwise stated as original paper referred to different transcript). This table is modified from (1).

It is likely that further genetic, epigenetic and environmental modifiers are contributing to the phenotypic variability and it is recommended therefore that SYNE1 mutations be considered also in the aetiology of complex as well as pure recessive ataxia. Further development and establishment of next-generation techniques in clinical diagnostics will reduce costs and involved timespans until diagnosis to enable SYNE1-associated cerebellar ataxia to be more readily identified in the future. However, given the large size of this gene (145 coding exons, 27436 basepairs, 8749 amino acids), it is likely to 182

inherently harbour a lot of genetic variation, and cautious interpretation of identified variants will be needed to infer pathogenicity and separate true signal from noise.

Conclusions Here, the ethnic and genetic diversity of SYNE1-associated cerebellar ataxia is expanded by identification of four novel truncating mutations in SYNE1 in pedigrees from British and Sri Lankan (for the first time), and Turkish origin, clinically confirming a rather pure cerebellar phenotype with less frequently observed additional pyramidal signs for SYNE1. Searching for a genotype-phenotype correlation, observations from this and several smaller studies before suggest our and other recently reported novel homozygous mutations located near the 3-prime end of the gene are more frequently associated with additional motor neuron involvement than compound heterozygous mutations and/or mutations near the 5-prime end. However, given the recent study from Synofzik et al, this picture may be even more complicated.

The data presented here indicate SYNE1 mutations are not an uncommon cause of recessive ataxia with or without additional clinical features in patients from various ethnicities, and further analyses will likely reveal more mutations and clinical data towards possible genotype-phenotype correlations.

183

4.3 PNPLA6-associated disorders: Novel homozygous variants in a large consanguineous Parsi kindred with pure cerebellar ataxia

4.3.1 Introduction

Autosomal recessive cerebellar ataxia is an umbrella term unifying a heterogeneous group of neurodegenerative conditions inherited in autosomal recessive fashion with primary affection of the cerebellum. There are variably associated neurological signs such as spasticity, myopathy, intellectual disability, seizures, optic atrophy, neuropathy and cognitive impairment with an ever-growing number of causative genes identified (241, 447). With the advent of whole exome and whole genome sequencing (WGS) for research and diagnostics, gene discovery has advanced in many ways with new genes emerging on the scene, extension and dilution of phenotypic spectra and novel genotype-phenotype correlations (448-453).

Mutations in the patatin-like phospholipase domain-containing gene PNPLA6 were originally described in 2008 as causative for specific forms of complicated HSP (SPG39, MIM #612020) and have been associated with motor neuron disease since (454). However, six years later using a WES approach, Synofzik et al. identified PNPLA6 additionally as the most frequent unifying genetic cause of two distinct clinical syndromes with previously elusive genetic origin: Boucher-Neuhauser (BNHS, MIM #215470) and Gordon Holmes syndromes (GDHS, MIM #212840). In subsequent studies, mutations in this gene were shown to be causative of a broad spectrum of neurodegeneration, frequently including motor neuron disease, ataxia, chorioretinal dystrophy and hypogonadotropic hypogonadism (455) and more recently, PNPLA6 was independently identified as the genetic cause in several families with Laurence- Moon syndrome (LNMS, MIM #245800), Oliver-McFarlane syndrome (OMCS, MIM #275400) and Leber congenital amaurosis (LCA1, MIM #204000) (456, 457). See

Figure 4-7 (reproduced in a modified version from (455)) for frequently associated phenotypes.

184

Figure 4-7: PNPLA6-associated phenotypes

Continuum of phenotypes subsumed under PNPLA6-associated neurodegeneration. Figure modified and reproduced from (455).

PNPLA6 encodes a protein responsible for enzymatic deesterification of membrane phosphatidylcholine into fatty acids and glycerophosphocholine, with glycerophosphocholine being an essential precursor constituent of the neurotransmitter acetylcholine (458, 459). It is secondly also suggested to have catalyzing capabilities for the production of 2-arachidonoyl lysophosphatidylinositol and thereby links itself functionally to other spastic paraplegia genes (460) and more generally to the group of neurodegenerative disorders with deficits in lipid metabolism (461-463).

Here, a large consanguineous Parsi family from India with pure cerebellar ataxia in two affected cousins was described. Using homozygosity mapping and whole exome sequencing, two novel homozygous missense variants were identified in the PNPLA6 gene. To our knowledge, it is the first time that variants in this gene are associated with a rather pure form of cerebellar ataxia.

4.3.2 Results

Clinical history and examination The index patient (see Figure 4-8: IV – 7) started developing cerebellar symptoms at the age of 12, which progressed very slowly with balance and coordination problems affecting his gait and manual movements.

185

Figure 4-8: Family tree of large Parsi kindred

Family tree with segregation status of variants c.3847G>A, p.V1283M and c.3929A>T, p.D1310V (Abbreviations: wt – wildtype, mt – mutant, the arrow denotes the index case). This figure is published elsewhere (464).

It was reported that he rarely cried as an infant, had nystagmus and poor handwriting. He was a double graduate, though he needed a writer for the exams. In both the affected individuals, brisk lower limb reflexes and an ataxic gait suggested early-onset cerebellar ataxia with retained reflexes (EOCARR). Nerve conduction studies were within normal limits. EMG recordings were suggestive of bilateral S1 segment lesion at the proximal level for both patients and additional left L5 segment lesion (proximal level) for the affected cousin (IV – 3). There was no hypogonadotropic hypogonadism or chorioretinal dystrophy in both of them. The index case’s condition was significantly exacerbated by a stroke due to thrombosis in the left transverse sinus at age 60 preceded by headaches. In this context, an MRI was acquired that showed severe cerebellar atrophy and some degree of midbrain atrophy (see Figure 4-9). Polycythemia vera was detected at this time, which was controlled with medication and phlebotomy. It was only after the stroke that vision was impaired due to the ischaemic event and frequent falls led to parenchymal and subdural haematomas causing aphasia, making him disoriented and aggressive. He had a very strong grip, developed stereotypy of the hand and spasticity of the neck with the head tilting backwards. He was also severely constipated. The index patient died one year ago at the age of 70. Regarding the movement disorder, the paternal cousin of the index patient (see Figure

186

4-8: IV – 3) had similar symptoms with onset around puberty and pure cerebellar ataxia leading to frequent falls and leg fractures, for which she is now at the present age of 68 years wheelchair-bound. She had dysmetric slow saccades on an earlier investigation that have now diminished. Besides poor handwriting, she does not have any further complications or additional symptoms and is mentally alert, though speech is slow. In general, her disease seems milder than her cousin’s, however, this judgement is influenced by the index’s complicated course following stroke. In total, clinical data was available from 22 individuals of the family (Figure 4-8), whereof only the two reported cousins were found symptomatic with cerebellar ataxia but without hypogonadotropic hypogonadism or chorioretinal dystrophy.

Figure 4-9: MRI of index case IV-7

T1-weighted sagittal MRI of unaffected (left) and sagittal and coronal MRI of index case at age 60 depicting severe cerebellar atrophy circled in red and some additional degree of midbrain atrophy. This figure is published elsewhere (464).

Homozygosity mapping, whole exome sequencing, Sanger sequencing and in silico predictions Homozygosity mapping revealed one large region of shared homozygosity by the two affected cousins only on chromosome 19 (see Figure 4-10), spanning over 4 Mb (Chr19: 3630740 – 7759053).

187

Figure 4-10: Homozygous stretch shared by affected cousins

Blocks of homozygosity along chromosome 19. The arrow points to the homozygosity stretch shared by the affected cousins only (panel 2 and panel 4 from the top; panel 1 and 3 unaffected relatives). This figure is published elsewhere (464).

This region comprises 132 genes, from which two (ATCAY, PNPLA6) were particularly interesting candidates given the cerebellar phenotype of the patients. Whole exome sequencing revealed one synonymous change in SAFB2 and two non-synonymous homozygous missense variants in PNPLA6 at c.3847G>A (p.V1283M) and c.3929A>T (p.D1310V) in exon 32, transcript ENST00000414982 (see Figure 4-11 a, for variant filtering strategy). The PNPLA6 variants were confirmed by Sanger sequencing (see Figure 4-11, b) and segregated perfectly with the disease in all 20 unaffected and 2 affected individuals tested (see segregation status wt/wt, wt/mt, mt/mt for all 22 individuals in Figure 4-8 and exemplary Sanger sequencing results for each of the three observed genotypes in Figure 4-11, b). Nine heterozygous carriers were detected in four generations, but only the two affected cousins were homozygous. In addition, all exons of ATCAY as well as all exons of PNPLA6 were screened with Sanger sequencing and were ruled out to harbour any additional variants in the two candidate genes that could have been missed by exome sequencing. Additionally, screening of 40 British patients with a pure cerebellar phenotype revealed no further

188

mutations in PNPLA6. Furthermore, no additional mutations in known autosomal recessive cerebellar ataxia genes were detectable on the exome.

Figure 4-11: Filtering strategy, segregation and conservation status PNPLA6 variants

a) Filtering strategy for variants in the exome of the index case. b) Sanger sequencing of affected index case with homozygous changes at positions c.3847G>A, cDNA and c.3929A>T, gDNA (top panels), heterozygous unaffected (mid panel), and wildtype unaffected relatives (lower panel). No effects on splicing were observed for the c.3847G>A variant. c) Multiple sequence alignment showing conservation across species at the affected amino acid residues 1283 and 1310 (Abbreviations: WT – wildtype, Mut – mutant). This figure is published elsewhere (464).

The variants found in this consanguineous family p.V1283M and p.D1310V have, to our best knowledge, not been reported before. They are absent in over 13000 individual alleles listed in the Exome Sequencing Project variant server (NHLBI-ESP EVS), and also not found in dbSNP, ExAC, Complete Genomics 69 (cg69) and 1000 Genomes databases (see Table 4-2 for a summary of both variants). One of the identified PNPLA6 variants (c.3929A>T, p.D1310V) was conserved across species (see Figure 4-11, c), predicted to be disease causing by Mutation Taster (http://www.mutationtaster.org/), damaging by SIFT (http://sift.jcvi.org/) and possibly damaging by Polyphen2 (http://genetics.bwh.harvard.edu/pph2/). This variant locates downstream, but close to the recently identified mutational cluster within the c-terminal phospholipid esterase domain of the PNPLA6 encoded protein (455) further supporting

189

its putatively pathogenic role. The variant c.3847G>A, p.V1283M has non deleterious prediction scores and is likely to be a rare, but tolerated missense variant in linkage disequilibrium (LD). As it lies at the beginning of exon 32, an effect on splicing (e.g. exon skipping) could be possible, however, by Sanger sequencing cDNA from the affected cousins and their carrier and wildtype siblings potential differences in splicing due to this variant (see Figure 4-11, b) were ruled out. An additional impact of this predicted benign variant towards the phenotype cannot be excluded though and until further proof of functional pathogenicity becomes available the final interpretation of both novel variants remains unclear.

Table 4-2: Novel homozygous PNPLA6 variants identified

dbSNP, Position 1000g, GERP Poly- Chr Transcript Variant SIFTb Mutation Taster (hg19) ExAC, EVS, scorea Phen2 cg69

ENST00000 c.3847G>A, absent 19 7625900 2.79 T N B 414982 p.V1283M

ENST00000 c.3929A>T, 19 7625982 absent D D P 414982 p.D1310V 3.81

Frequency and in silico predictions of homozygous, novel PNPLA6 variants identified in large Parsi pedigree. This table has been published elsewhere (464).

Abbreviations: GERP=Genomic Evolutionary Rate Profiling; SIFT=’Sorting Tolerant From Intolerant; D=deleterious/damaging/disease causing; P=possibly damaging; N=Polymorphism; T=Tolerated; B=Benign; dbSNP=database of short genetic variation; 1000g=1000 Genomes project; ExAC=Exome Aggregation Consortium; EVS=exome variant server; cg69=Complete Genomics 69 genomes data. a Positive conservation scores represent a substitution deficit and indicate that a site may be under evolutionary constraint. Negative scores indicate that a site is probably evolving neutrally. Positive scores scale with the level of constraint, such that the greater the score, the greater the level of evolutionary constraint inferred to be acting on that site. b Using the ‘Sorting Tolerant From Intolerant’ algorithm (465) this tool predicts whether an amino acid substitution affects protein function based on the degree of conservation of amino acid residues in sequence alignments derived from closely related sequences.

4.3.3 Discussion

Here, the first PNPLA6 variants in a large kindred of Indian Zorastrians (Parsis) affected with a pure, autosomal recessive cerebellar syndrome were identified. The mutated gene in this family, the patatin-like phospholipase domain containing 6 gene on chromosome 19, encodes neuropathy target esterase (NTE), a phospholipase that produces glycerophosphocholine by deacetylation of intracellular phosphatidylcholine and therefore has important roles in membrane axonal integrity, phospholipid trafficking 190

and phosphatidylcholine-metabolism (458, 466, 467). NTE-activity can be impaired exogenously by exposure to neurotoxic organophosphorous (OP) compounds. Exposure can result in OP compound-induced delayed neuropathy (OPIDN) with further neurological symptoms (468-470). The clinical features of OPIDN are in part similar to the widespread symptoms one observes upon genetic impairment of NTE due to mutations in PNPLA6 (455-457). From what is known so far, spinocerebellar features can result from degeneration of long axons, which has been shown to correlate with increased phosphatidylcholine levels in the mouse brain (471-473). It is therefore likely, that toxic or genetic impairment of neuropathy target esterase activity can cause disruption of lipid-metabolism, impairment of axonal transport crucial to axons in a length-dependent manner and impaired acetylcholine production, leading via various mechanisms in different target tissues (retina, motor neurons, cerebellum) to broad clinical manifestations.

Recently, biallelic mutations in the PNPLA6 gene have been identified in more and more patients, further extending the phenotypic spectrum, frequency and geographic occurrence (474): compound heterozygous mutations have been reported in a sporadic BNS-patient with late-onset gait ataxia and mild retinal changes (475), the first two non-Caucasian PNPLA6 cases of Japanese origin have been published (476), and PNPLA6 was identified as the genetic cause in several families with Laurence-Moon syndrome, childhood blindness, Oliver-McFarlane syndrome and Leber congenital amaurosis (456, 457), see recent review (474). Here two novel homozygous variants associated for the first time with a pure cerebellar phenotype in a large kindred of Indian descent are presented. The more conserved and predicted deleterious variant identified here (c.3929A>T, p.D1310V) is the more likely genetic origin of the disease in this family. It lies between the previously published pathogenic mutations c.3365C>T, p.P1122L (located in the catalytically active phospholipid esterase domain of the phospholipase and shown to cause BNS) and the farthest downstream mutation towards the c-terminal site (c.4048C>G, p.R1362G, causing GHS) (455), further supporting its putative pathogenicity. It might therefore not have a direct diminishing effect within the catalytic centre, but an indirect steric inhibition of the protein’s original conformation hindering effective catalysis. An additional impact of the predicted benign change c.3847G>A, p.V1283M cannot be ruled out. Unfortunately, no further functional studies clarifying the role of the identified variants could be performed due to lack of further biomaterial from the affected patients, lack of established modified cell lines overexpressing the identified variants, and lack of biochemical readout assays in the laboratory. Until further functional evidence becomes available the final interpretation of the variants remains pending. However, criteria mentioned above (absence in public

191

databases, location close to mutational hotspot, deleterious in silico predictions, perfect segregation with the disease) argue in favour of a pathogenic implication of variant c.3929A>T, p.D1310V.

Conclusions In conclusion, this report further extends the clinical spectrum of PNPLA6-associated diseases to pure cerebellar ataxia without chorioretinal dystrophy or hypogonadotropic hypogonadism. Previous mutations in this gene have been associated with a more complex phenotype but the results here extend the associated disease spectrum. PNPLA6 should therefore be considered in cases of early-onset cerebellar ataxia despite the absence of chorioretinal dystrophy or hypogonadotropic hypogonadism that are regularly associated with mutations in this gene (455, 477).

The Zorastrian community declining as it is in India, is now spread world-wide and comprises the Parsis of India, the remaining small number in Iran, their diaspora who have re-migrated to various countries and possibly some Iranian followers of Islam, converted from Zoroastranism. All these belong to the same genetic pool increasing the possibility of further families with pure autosomal recessive cerebellar ataxia and associated homozygous variants in PNPLA6 presenting outside India in Iran and elsewhere.

After successful gene/mutation identification in the recessive families described in this chapter, the next chapter returns to cohort studies. It reports the identification of DNA repair variants associated with age at onset in trinucleotide repeat disorders.

192

5 Chapter 5: Trinucleotide disorders: The conundrum of modifiers of age at onset, somatic and repeat instability and repeat interruptions

5.1 DNA repair pathways underlie a common genetic mechanism modulating onset in polyglutamine diseases

Introductory remarks As exemplified in the previous results chapters of this thesis, gene identification and subsequent recognition of specific disease causing variation within these genes allows the creation of clinicogenetic subgroups and their thorough clinical characterisation. Once well characterised clinical cohorts where specific gene mutations have been identified are looked at in detail, it becomes obvious that the same mutation/genetic defect is associated with a range of phenotypic variation such as striking differences in age at onset, manifesting symptom, disease severity and disease progression, accompanying symptoms, etc. These differences might partly be due to environmental factors influenced by life style, nutrition, physical activity or psychological wellbeing and coping mechanisms. Differences in the epigenetic and the overall genetic makeup might explain other parts of the observed differences in disease manifestation. This chapter moves away from gene identification towards the investigation of genetic modifiers influencing disease course in genetically and clinically well characterised polyglutamine disorders.

5.1.1 Introduction

Around 50% of the human genome is repetitive DNA, of which 3% is composed of tandem repeat units, usually 1-13 bases long (478). These sequences are unstable and undergo dynamic mutation, changing in copy number meiotically (479). Though their functional significance is unclear, they can cause disease once expanded over a pathogenic threshold that is individual for each expansion locus. More than 30 human diseases are caused by expansion of these unstable microsatellite sequences (239).

Nine of these are characterised by repeats that encode glutamine, usually termed polyglutamine diseases (see Table 1-2). Even though they affect unrelated genes, they have common clinicogenetic and pathological hallmarks including autosomal dominant inheritance (except X-linked spinal and bulbar muscular atrophy), neuronal involvement, genetic anticipation, pathogenic repeat threshold (which is different for each disease), the inverse and differentially strong correlation of age at onset or disease severity with the number of repeats and the intracellular inclusion aggregates

193

containing the cognate polyglutamine protein. Together they represent the most common neurodegenerative genetic diseases of the nervous system. Phenotypic expression varies, as does the temporal and regional expression and the functional protein context of the disease causing expansions (238) (see Table 1-2). Though symptomatic treatment can bring relief, to date, there are no disease-modifying treatments for these devastating conditions largely reflecting the poor understanding of the true and early pathogenic molecular events leading to the relentless neurodegeneration observed.

In the polyglutamine diseases length of the CAG expansion is the main determinant of age at onset. Longer CAG repeat tracts lead to earlier AAO though this relationship varies between diseases and not all of the difference in age at onset is accounted for by CAG repeat length (see Table 1-2 for the individual diseases) (480, 481). In Huntington’s disease (481) and at least spinocerebellar ataxia types 2 and 3 (482), a substantial portion of this residual variance is heritable, suggesting the existence of additional undiscovered modifying factors within the genome. Recently, the Genetic Modifiers of Huntington’s Disease (GeM-HD) genome-wide association study (GeM-HD GWAS in the following) (356) found two genome-wide loci significantly associated with age at motor onset in HD on chromosomes 15 and 8, with two independent signals underlying the same locus on chromosome 15. For the SCAs, a few SCA genetic modifiers have been proposed (480, 482-485) but no GWAS has been reported in these diseases.

Genetic anticipation in these disorders occurs because the disease causing repeats are meiotically unstable and tend to expand over successive generations. Most repeat disorders also show tissue-specific somatic instability (486) and the repeats are interrupted in some, but not all of the diseases (see Table 1-2). Repeat interruption is thought to have a repeat stabilising function. In HD somatic instability is expansion- biased and age-dependent, with larger tracts more susceptible to expansion (487, 488). Somatic instability has been shown to occur in post-mitotic neurons and is prominent in striatum and cortex, tissues particularly affected in HD (489). It has been linked to disease onset and progression in both human (490) and mouse HD-studies (491) and reducing somatic expansion in an HD mouse model experimentally delays phenotype progression (492). Many of the principles of somatic instability in HD extend to SCAs (239, 486). However, the trinucleotide repeat SCAs are generally researched to a lesser extent given their lower frequency and genetic heterogeneity. Somatic instability has amongst others been attributed to the actions of DNA repair proteins (488, 493, 494), and besides the individually associated variants in DNA repair genes,

194

the GeM-HD GWAS found significant association between age at motor onset and several DNA repair pathways (356). These GeM-HD GWAS findings, along with evidence for somatic instability in other polyglutamine diseases (see Table 1-2) initiated the hypothesis that common variants in the genes encoding DNA repair machinery proteins not selected against in the general population might modulate AAO in all polyglutamine diseases by action through somatic expansion.

In the study described below, significant associations between variants in genes involved in DNA repair machinery pathways and the AAO of polyglutamine diseases as a group as well as with some of the polyglutamine diseases individually are reported.

5.1.2 Results

The primary analysis, which tested the overall effect of all 22 SNPs on AAO, revealed significant associations (after Bonferroni correction for eight tests) for HD+SCAs (p=1.43x10-5), HD (p=0.00194), all SCAs (p=0.00107), SCA2 (p=0.00350), and SCA6 (p=0.00162). The increased significance of these associations compared to an undirected test using two-sided SNP p-values (see Table 5-1) indicates concordance in the direction of effects across SNPs between these samples and GeM-HD GWAS (356). Specifically, the observed association with HD is a convincing replication of the GeM-HD study results in an independent sample.

195

Table 5-1: Results of combined analysis of SNPs

Disease GeM-HD P (High LD SNPs P (rs3512 P (All SNPs) Group concordance? removed) removed) ALL non directional 4.74x10-4 * 2.26x10-4 * 0.00492 * (HD+SCAs) Same as GeM-HD 1.43x10-5 * 6.98x10-6 * 2.26x10-4 * HD non directional 0.0226 0.00775 0.0364 Same as GeM-HD 0.00194 * 4.63x10-4 * 0.00394 * SCAs non directional 0.0188 0.0236 0.0771 Same as GeM-HD 0.00107 * 0.00142 * 0.00667 * SCA1 non directional 0.376 0.386 0.444 Same as GeM-HD 0.416 0.287 0.524 SCA2 non directional 0.0230 0.0629 0.0233 Same as GeM-HD 0.00350 * 0.0138 0.00442 * SCA3 non directional 0.176 0.114 0.355 Same as GeM-HD 0.0809 0.0381 0.205 SCA6 non directional 0.00588 * 0.0735 0.00506 * Same as GeM-HD 0.00162 * 0.0340 0.00163 * SCA7 non directional 0.155 0.217 0.297 Same as GeM-HD 0.0447 0.0885 0.113

P-values in this table obtained by combining single-SNP p-values using Brown’s method (358), allowing for LD between SNPs. Non-directional analysis combines two-sided p- values. “Same as GeM-HD” analyses combine one-sided p-values in the same direction as the SNP effects observed in GeM-HD study (356). In the “High LD SNPs removed” analysis, rs1037700, rs5893603 and rs16869352 were removed due to high LD (r2>0.8) with more significant SNPs in GeM-HD. * P-values that satisfy Bonferroni correction for 8 disease group tests. Note that SCA17 was included in the “HD+SCAs” and “All SCAs” grouped analyses, but was not tested independently due to small sample size. Abbreviations: HD – Huntington’s disease; SCA – spinocerebellar ataxia. This table is published elsewhere (2).

As a secondary analysis, individual SNP associations were examined. Three of these were significant after Bonferroni correction for eight disease combinations and 22 SNPs (see Table 5-2 below and Table 8-21 (in Appendix, Section 3)): rs3512 in FAN1 with all SCAs and HD+SCAs and rs1805323 in PMS2 with HD+SCAs. Each association was in the same direction as in the GeM-HD study (356). The most significant signal from the GeM-HD GWAS, rs146353869 (p = 4.30x10-20, associated with six years earlier age at motor onset of HD) could not be replicated. This is likely due to our sample size being much smaller than in the GeM-HD paper and thus less well powered to find associations with SNPs with relatively low frequency MAF such as rs146353869 (MAF=0.017). However, rs3512, the most significant individual SNP in this study, indexes the second significant chromosome 15 signal in the GeM-HD

196

GWAS (p=5.28x10-13, associated with 1.4 years later onset of HD), and is in the 3’UTR of FAN1. Three SNPs (rs1037700, rs5893603, rs16869352) were found to be in high LD (r2>0.8) in our sample with more significant SNPs from the GeM-HD analysis. Removing these SNPs reduced the significance of the multi-SNP associations with SCA2 and SCA6, although these remained nominally significant (see Table 5-1). Finally, all the significant multi-SNP associations from the primary analysis remained significant after removing the most significant single SNP (rs3512), suggesting that the signal enrichment is not being driven by a single SNP (Table 5-1).

Table 5-2: Individual SNPs significantly associated with AAO

Significant after Bonferroni for 22 SNPs and 8 disease groupings (p<2.84x10-4) Same Disease 2-sided p- SNP Gene direction as Group value GeM-HD? rs3512 FAN1 All (HD+SCAs) 1.52x10-5 Yes rs1805323 PMS2 All (HD+SCAs) 3.14x10-5 Yes rs3512 FAN1 All SCAs 2.22x10-4 Yes Significant after Bonferroni for 22 SNPs (p<2.27x10-3) rs1805323 PMS2 HD 3.14x10-4 Yes rs1805323 PMS2 SCA1 1.67x10-3 Yes rs1037699 RRM2B SCA6 4.86x10-4 Yes rs1037700 RRM2B SCA6 5.47x10-4 Yes rs5893603 RRM2B SCA6 2.13x10-3 Yes

Abbreviations: HD – Huntington’s disease; SCA – spinocerebellar ataxia. This table is published elsewhere (2).

To visualize the combined effect of our SNPs on residual AAO, a polygenic “age at onset score” was derived, defined as the sum of the number of minor alleles at each locus weighted by their effect size in the GeM-HD study (note that negative scores here correspond to earlier AAO). The residual AAO for each quartile of this risk score is plotted in Figure 5-1. As expected, there was a positive correlation between residual AAO in the data and increasing age at onset score, although the effect was small – the score accounts for approximately 1% of the variance of residual AAO.

197

Figure 5-1: Residual AAO boxplot

Boxplot of residual AAO (across all samples) by quartiles of polygenic age at onset score. Polygenic score calculated by summing the number of minor alleles (weighted by their effect on age at onset in the GeM-HD GWAS) across the 22 SNPs. Note that lower scores correspond to earlier than expected AAO, and thus smaller residuals. This figure is published elsewhere (2).

5.1.3 Discussion

For the first time, here presented data implicate a common mechanism by which genetic variation in DNA repair pathways underlies modification of age at disease onset in multiple polyglutamine diseases. Conceptually, alterations in DNA repair pathways determined by underlying genetic variation could predispose to earlier onset by interacting with polyglutamine etiology at various levels. It has been shown that rare loss-of-function variants in DNA repair genes cause multiple recessive ataxias (495). Examples of this entail ATM, mutated in ataxia-telangiectasia, which encodes a master regulator of DNA repair following double-strand breaks (496), PNPK which encodes a DNA-specific kinase that facilitates DNA repair and where mutations lead to recessive ataxia with oculomotor apraxia 4 (MIM #616267) (497), APTX encoding a protein that

198

interacts with PARP1 to mediate single-strand DNA breaks, implicated in ataxia with oculomotor apraxia 1 (MIM #208920) (498) and mutations in TDP1 which cause defects in single-strand break repair clinically resulting in ataxia with axonal neuropathy (SCAN1, MIM #607250) (499). The mechanisms by which ataxia and neurodegeneration result from these losses of function are not conclusively established and still under research. However, there is substantial evidence for the fine control exercised by ATM being critical in cell division and cell death pathways which could eventually lead to neuronal vulnerability and loss (500). However, it is notable that none of the genes associated with recessive ataxia syndromes were identified to contain significant HD-related variants in GeM-HD GWAS (356).

Mechanistically, several hypotheses linking DNA repair and somatic instability can be brought forward: It has been noted that repetitive DNA sequences are able to generate unusual secondary structures (501). In an attempt to repair and eliminate them, DNA mismatch repair proteins can bind to these. In the process of repair they can cause somatic instability of the CAG repeats (in the majority of events: expansion of repeats). A number of enzymes with the ability to nick DNA, and therefore necessitating DNA repair, are known to promote CAG expansion and both somatic instability/expansion, and HD-related phenotypes are ameliorated in mouse models by manipulating genes associated with DNA repair (491, 502-505). Critically, delay in phenotype onset in HD mice was recently demonstrated via suppression of somatic expansion by crossing HTT knock-in mice with Ogg1-/- mice, lacking the DNA cleaving 7,8-dihydro-8-oxo- guanine (8-oxo-G) glycosylase (492). Notably the single most significant SNP in the present study, rs3512, is in the 3’UTR of FAN1, which has DNA endo/exonuclease activity. Larger CAG repeats are associated with more severe pathology and earlier disease onset in affected patients, therefore somatic expansion provides a plausible mechanism by which the genetic variation identified here could alter AAO of disease (see Figure 5-2).

Limitations Several issues are highly likely to have reduced the power of the study presented here. Due to the inherent rarity of the diseases under study, the sample sizes for many of the SCAs were relatively small, and despite modelling the relationship of age at onset to CAG length separately for each disease, there is likely to be heterogeneity between diseases in this and potentially other respects not considered here. It could not be accounted for interruptions of pure CAG repeat tracts, which may stabilize repeat

199

instability (506), thus the power to detect any effects mediated by somatic instability may have potentially even been significantly reduced in this study.

Nevertheless it could be demonstrated that DNA repair genes as a group significantly modify AAO in the polyglutamine diseases taken together, in HD, in all SCAs, SCA2 and SCA6. Additionally, potential modifier SNPs in HD, SCA1 and SCA6 (Table 5-2 above and Table 8-21 in the Appendix) were identified for which functional elucidation will be exciting in further pathophysiologic studies. The effects of the identified SNPs on AAO are quite small, and it would be worth repeating the analysis with larger samples and more SNPs. Finally, one inherent limitation always will remain: The inaccuracy and recall bias when participants are asked to give their AAO. This is a limitation that most clinical studies in neurodegeneration have, since disease onset (and hence the patient’s age at onset) usually represents a slowly occuring process rather than a sharp defined event. Therefore, AAO can be subject to recall bias, estimation and hence inaccuracy when being reported during anamnesis.

Despite limitations: By suggesting common mechanisms for polyglutamine diseases here presented findings might offer novel therapeutic opportunities in multiple diseases along with the potential to improve clinical trial design by stratifying subject variability. Molecules targeting DNA repair have been developed and are used in the clinic to treat cancers (507, 508) and such therapeutics, along with others in development, may prove useful in some or even all of the polyglutamine diseases. Furthermore, these shared mechanisms may extend to diseases associated with non- CAG and non-translated repeats, most likely influencing those that show somatic instability via action through insufficient DNA repair which offers exciting and important novel areas of study.

200

Figure 5-2: Linking DNA repair with somatic instability

Hypothesized potential mechanism by which variants in DNA repair could influence somatic expansion of CAG repeats in polyglutamine diseases due to variation in genes encoding DNA repair proteins. The accessibility of repetitive DNA sequences during replication, transcription, etc., allows the formation of secondary DNA structures: SNPs in genes encoding DNA repair proteins may alter the kinetics or activity of DNA repair complexes (pink bobble). After endonuclease activity on the opposite strand (nick indicated by the grey arrow), such impaired repair may lead to further expansion of the repeat tracts by consequent gap filling synthesis by DNA polymerase (blue bobble). Please note: This figure is published elsewhere (2).

Conclusions The polyglutamine diseases, including HD and multiple SCAs, are amongst the commonest hereditary neurodegenerative diseases and are caused by trinucleotide expansions, encoding glutamine, in different genes in each disease. Even though longer CAG repeat tracts are associated with earlier disease onset, a large proportion of the difference remains unaccounted for, pointing towards additional genetic modifying factors in these diseases that need identification. By assessing the modifying

201

effects of variants in DNA repair genes across a range of polyglutamine diseases, significant association with age at onset when grouping all polyglutamine diseases (HD+SCAs) was found, furthermore in individual SNP analyses, significant associations for rs3512 in FAN1 with HD+SCAs and All SCAs and rs1805323 in PMS2 with HD+SCAs were identified.

For the first time it was hereby shown that DNA repair genes significantly modify the age at onset in HD and SCAs, suggesting a common pathogenic mechanism, most likely through the observed somatic expansion of repeats that can be modulated by manipulation of DNA repair in disease models.

202

5.2 The role of repeat interruptions in SCA6

5.2.1 Introduction

In the individual disease analysis of the study above significant association of combined DNA repair variants with AAO in SCA6 (see Table 5-1) was identified. One possibility for this lies in the exceptionally short CAG repeat tract causing disease in SCA6 since the pathogenic threshold in this disease is far lower than the pathogenic threshold for all other repeat disorders (see Table 1-2, pathogenic range). Shorter repeat tracts tend to be more stable in general, and in the rare cases studied, SCA6 repeats have been proven quite stable with small ranges of normal and pathogenic alleles. Therefore, it is possible that the effects of variants in DNA repair genes could play a much stronger role in this otherwise relatively stable nonexpanding condition.

Another important factor influencing repeat stability and thereby indirectly the results of the study above is the presence of interruptions within the expanded repeat tract. Interruptions reduce occurrence and dimension of meiotic and somatic expansion and therefore might be expected to ameliorate disease severity and age at onset. One major limitation of the analysis presented above is that this study did not account for interruptions. Since one could not sequence through the entire repeats for all disease groups, significance of yielded results could be expected to get masked in diseases with interrupted tracts. This led to the question whether there is a general presence or absence of interruptions in the SCA6 samples that showed strong individual disease association with variants in DNA repair genes since potential interruptions would very likely have influenced the obtained results.

SCA6 is one of the six polyglutamine SCAs. It is caused by expanded CAG repeats in exon 47 of CACNA1A, encoding the transmembrane pore-forming subunit of the P/Q- type or CaV2.1 voltage-gated calcium channel (509) and presents clinically with a late- onset slowly progressive mostly pure cerebellar ataxia. Voltage-dependent calcium channels are highly complex proteins with several subunits and implicated in a multitude of calcium dependent processes such as gene expression, neurotransmitter and hormone release, muscle contraction and action potential generation. Discovered in 1997, expansions of a CAG repeat in the c-terminal coding region of CACNA1A were shown to cause an autosomal dominant relatively slowly progressive and pure cerebellar ataxia (510). Whilst CACNA1A missense, nonsense or point mutations and single deletions can be found at various places of the gene and lead to two allelic disorders, namely Episodic Ataxia 2 (EA2, MIM #108500) and Familial Hemiplegic Migraine type 1 (FHM1, MIM #141500), SCA6 causing repeats are always

203

located in exon 47 near the region encoding the cytoplasmic c-terminal tail of the α1A subunit of the channel – for a recent review, see (511). This cytoplasmic c-terminus is known to play an important role in fast channel inactivation and modulation by intracellular signalling molecules (512). In SCA6, normal repeat alleles were shown to range from 4 to 18 CAG repeats whilst pathogenic expanded repeats range from 20 to 33 units (510, 513, 514). Intermediate allele size was determined with nineteen repeats. Intermediate repeat size has been shown to act as a susceptibility factor to disease with variable phenotypic expression and penetrance and predisposing to pathogenic expansion (515, 516). Interestingly, this pathogenic range in SCA6 is within the normal distribution of repeat sizes for all other polyglutamine diseases and far below the threshold for polyglutamine aggregation and phenotypic expression in these (usually 35-100 repeats, see Table 1-2), additionally distinguishing this disorder from the other polyglutamine diseases. However, several studies have shown an inverse correlation between AAO and repeat size (510, 517, 518) as typical for the group of polyglutamine disorders. In some SCA6 subcohorts, the inverse correlation seems to be even present for the sum of expanded pathogenic and normal allele in this disease with short pathogenic repeat alleles particularly (519) and the CACNA1A repeat seems to be more stable than other repeats within the group of polyglutamine diseases (510, 513). For example, an early analysis of CAG repeat sizes in Japanese autosomal dominant cerebellar ataxia pedigrees revealed expansions in 8 of 15 families, and more importantly absolute stability of the repeat size without anticipation in all identified affected families with SCA6 (517).

Interruptions of the expanded CAG-repeat tract have been thought to lend stability to the repeats and they have been observed in various other polyglutamine diseases like HD, SCA1, 2, 3 and 17 (520-522). However, up to now it remains undetermined whether the CAG repeats in SCA6 are interrupted or not: Even though no mentions of interruptions in SCA6 literature can be found, it remains unclear whether this has ever been formally assessed. Given the results implicating DNA repair variants as potential modifiers in SCA6 (and other polyglutamine diseases) and given the striking stability of repeats in SCA6 this study aimed to identify whether there are interruptions within the CAG repeat in SCA6 samples that might have stabilizing effect on repeats and thereby potentially even weakened the observed association presented in the previous subchapter.

204

5.2.2 Results

Recruited patients were representants of a UK SCA6 cohort of mixed ancestry and mainly singletons/index cases. No repeat interruptions were observed in 173 individuals with repeat expansions in CACNA1A. With the method described above, the mean repeat size of the unexpanded allele was determined in 173 patients at 12.3 +/- 1.3 (SD) with ranges from 7 to 15 repeats. Mean repeat size of the expanded allele in 173 patients was 23.4 +/- 1.2 (SD), ranging from 21 to 28 repeats.

AAO and gender was reliably available for 65 of these 173 patients. Within this subgroup of SCA6 patients, repeat size of the normal allele was 12.3 +/- 1.2 (SD; range: 7 to 15) and of the expanded allele 23.5 +/- 1.2 (SD; range: 21-27). Mean age at onset in 65 patients was 57.5 years +/- 10.1 (SD), ranging from 18 to 76 years. Thirty-nine patients of this SCA6 subgroup were male and 26 female. Table 5-3 summarises the clinicogenetic findings.

Table 5-3: Characteristics of SCA6 cohorts

Overall cohort AAO subcohort Patients (n) 173 65 Interruptions detected? N: O alleles N: O alleles

Mean normal allele size 12.3 (SD: 1.3) 12.3 (SD: 1.2) range: 7-15 range: 7-15 Mean expanded allele size 23.4 (SD: 1.2) 23.5 (SD: 1.2) range: 21-28 range: 21-27 57.5 (SD: 10.1) Age at Onset (yrs) n.a. range: 18-76 Gender n.a. M: 39 (60%), F: 26 (40%)

Abbreviations: yrs=years; n.a.=not available; SD=standard deviation; M=male; F=female; n=count; N=No.

Correlation of the expanded repeat size only and of the sum of expanded repeat allele size plus the normal allele size and AAO respectively revealed a weak negative correlation (-0.0244, and -0.0323 respectively) with a R2 of 0.04635 for the expanded allele only (see Figure 5-3) and R2=0.02339 for the sum allele size (expanded allele plus nonexpanded allele size) in the SCA6 subcohort (see Figure 5-4).

205

Figure 5-3: Inverse correlation of expanded allele size and AAO in SCA6 subcohort

Weak inverse correlation (R2=0.046) observed for AAO and repeat size in SCA6 subcohort.

Figure 5-4: Inverse correlation of sum allele size and AAO in SCA6 subcohort

Weak inverse correlation (R2=0.023) observed for AAO and sum allele size in SCA6 subcohort.

206

A scheme of expected chromatograms and example traces of obtained Sanger sequences of expanded repeat allele sequencing without presence of interruptions are shown in Figure 5-5 and Figure 5-6. It also illustrates how sometimes a discrepancy between the count on the forward (25 expanded repeats) and the reverse (24 expanded repeats) can occur with this method. In all cases where a discrepancy existed, the longer reading was taken forward for analysis.

Figure 5-5: CACNA1A repeat sequencing schematic and exemplary results unexpanded allele

A) Scheme of expected sequence of CACNA1A around the repeats. B) Forward sequence of sample with 14 uninterrupted CAG repeats on unexpanded allele and continuation of heterozygous expanded CAG-repeat allele on top of heterozygous CACNA1A forward WT sequence at the right end of the chromatogram and C) same sample, reverse sequence (Abbreviations: seq.=sequence, het=heterozygous, hom=homozygous, for=forward, rev=reverse.).

207

Figure 5-6: CACNA1A repeat sequencing schematic and exemplary results with expanded allele

A) Scheme of expected sequence of CACNA1A around the repeats with expanded and unexpanded alleles. B) Forward sequence of sample with 14 uninterrupted CAG repeats on unexpanded allele and continuation of in total 24/25 heterozygous expanded, uninterrupted CAG-repeats on other allele on top of heterozygous CACNA1A WT sequence forward (B) and C) same sample, reverse sequence (Abbreviations: seq.=sequence, het=heterozygous, hom=homozygous, for=forward, rev=reverse).

5.2.3 Discussion

The results presented here show that interruptions in SCA6 are not frequent and are therefore not likely to play a role in this disease and its striking inter-generational and intra-subject stability of repeat sizes. This finding distinguishes SCA6 from other polyglutamine diseases where repeat interruptions could be observed. It adds to other known features (e.g. small pathogenic repeat size and range, more frequent cytoplasmic aggregation and much rarer occurrence of intranuclear aggregation as compared to intranuclear inclusions being a hallmark of most other polyglutamine disorders) (523, 524), setting this particular polyglutamine disease apart. Importantly, this study minimises the possibility that repeat interruptions could have influenced the previously obtained association results between AAO and variants in DNA repair machinery. Interruptions were consequently ruled out in 65 out of the total of 69 SCA6 patients that contributed to the initial DNA repair association study

208

rendering it quite unlikely that the power to detect association in SCA6 was reduced by the presence of undetected repeat interruptions. This finding makes it quite interesting to pursue repeat sequencing in the other disease cohorts of the initial study to understand whether the presence of interruptions stabilising the repeats has potentially taken away statistical power to detect association in those diseases.

Interestingly, the observed inverse correlation between expanded allele size and AAO was considerably weaker than observed in previous studies, and no stronger correlation with sum of pathogenic and normal allele size than pathogenic expanded allele size alone was yielded, as had been reported for a subset of patients before (519). These discrepancies might be down to differences in sample size and genetic architecture with the cohort analysed here consisting of mainly index patients and a relatively small number (n=65) with AAO information and genetic data. To study the correlation between AAO and allele sizes in a larger cohort, a current follow up of all patients with missing AAO information via telephone interviews is on the way in order to increase sample size.

Limitations Expanded CAG repeat analysis using Sanger sequencing is not always informative and the manual analysis via visual inspection and separate reading of expanded and unexpanded alleles can be quite elaborate and time-consuming. Alternative techniques include separation of expanded and unexpanded repeat containing PCR products via gel electrophoresis, cutting from gel, elution and re-amplification and re-sequencing of the separated products, or PCR cloning and sequencing. Given that the SCA6 repeat is quite small in unexpanded and expanded form, it was decided to directly sequence expanded and unexpanded alleles jointly and to spend more time on thorough analysis of obtained readings. The quality of direct sequencing in this study was very high and interrupted repeats could be ruled out in cases mentioned above with high confidence. However, given that the gold standard of PCR cloning or separate sequencing of expanded and unexpanded alleles was not chosen, reported results have to be validated and interpreted with caution.

Additionally, in previous reports, a discrepancy between size length determination using PCR fragment analysis and direct sequencing has been noted: Direct sequencing has been shown to usually be more accurate (and therefore favourable, especially in size determination of intermediate allele size ranges) than fragment analysis which underestimated repeat size up to 3 to 4 repeats in previous studies (506, 525). In some DNA samples I found a discrepancy in the size of repeats between

209

forward and reverse of one repeat (for an example, see Figure 5-6). The longer reading was taken homogenously forward for analysis to reduce error within this study, however, the inaccuracy and discrepancy of sizes due to sizing technique could have additionally marginally contributed to the differences in correlation strength compared to previously published reports (519).

The results in the genetic chapters above highlight the beauty and the limitations of genetic studies equally. It is of paramount importance to not only understand disease causing genetic variation but also study disease modifying genetic variation. Here, novel genetic associations with age at onset were identified (Chapter 5.1) and an additional influence of repeat interruptions for SCA6 was excluded (Chapter 5.2). These findings might potentially indicate druggable pathways for therapeutic intervention and disease modification once the pathogenic effects of genetic variation in DNA repair pathways have been further elucidated. However, genetic findings can only be the starting point of further pathomechanistic research. Therefore, the next chapter tries to take a gene identification finding (causal heterozygous inositol-1, 4, 5,-trisphosphate-receptor-I (ITPR1) gene deletions in spinocerebellar ataxia type 15) forward in order to study the underlying cellular and molecular pathomechanism. It highlights the difficulties, necessities, limitations and small-scale progress when establishing necessary and novel disease models.

210

6 Chapter 6: Modelling cerebellar disease with human iPSCs

6.1 Generating a human iPSC model to study spinocerebellar ataxia type 15

6.1.1 Introduction

Introductory remarks The previous chapters of this thesis have investigated the underlying genetic disease causing mutations and the presence of genetic modifiers in a variety of movement disorders. These investigations have been able to reveal the genetic cause or pinpoint genetic variation that influences disease course for a significant subset of patients. However, these studies can only represent the starting point of further pathomechanistic research into disease initiating cascades and processes. Therefore, this last chapter describes pathomechanistic research into spinocerebellar ataxia type 15 using iPSC technology including the current limitations of available differentiation protocols to cerebellar neurons.

Spinocerebellar ataxias and SCA15: A brief overview The hereditary ataxias are genetic disorders with slowly progressive incoordination of gait, hands, speech, and eye movements and variable affection of additional neurological and non-neurological systems. Atrophy of the cerebellum is frequent, and dysfunction of the cerebellum and its associated systems is at the core of the clinical symptoms. They can be broadly classified into autosomal recessive and autosomal dominant ataxias with a growing number of associated genes and loci (243, 248). The prevalence of the autosomal dominant cerebellar ataxias is estimated to be approximately 1-5/100000 with differences between populations (244).

Amongst them, spinocerebellar ataxia type 15 is the most frequent non-trinucleotide repeat SCA in Central Europe (526). SCA15 patients present with an adult-onset, predominantly pure progressive cerebellar ataxia, however, pyramidal, extrapyramidal and cortical features have occasionally been reported (527, 528). MRI features include primarily vermal cerebellar atrophy as well as mild inferior parietal and temporal cortical volume loss (526, 529). Heterozygous deletions of up to 350 kilobases (kb) in the inositol 1,4,5-trisphosphate receptor gene resulting in haploinsufficiency of the receptor are the underlying genetic cause and have been assigned as causal of SCA15 only in 2007 (530).

211

In addition to pathogenic deletions in SCA15, heterozygous ITPR1 missense mutations have been associated with congenital nonprogressive spinocerebellar ataxia (531, 532), and more recently recessive truncating and dominant de novo mutations in the ion transport and channel domain of ITPR1 were shown to cause Gillespie syndrome (MIM #206700) characterised by aniridia, cerebellar ataxia and mental retardation (533, 534).

ITPR1 encodes the type 1 inositol 1,4,5-trisphosphate receptor (IP3R1), a ligand-gated calcium channel on the membrane of the endoplasmic reticulum (ER), highly expressed in the cerebellum as well as in the cortex (530, 535, 536). Inositol 1,4,5- trisphosphate receptors modulate cytoplasmic calcium concentration in nerve cells via calcium release from the intracellular stores, thereby affecting numerous physiological processes, including cell proliferation, differentiation and death, gene expression, neurotransmission and synaptic plasticity (537-543). Additional to its causal role in SCA15, a direct relation between dysfunction of IP3R1 and the development of Alzheimer’s disease, Huntington’s disease, and several other forms of spinocerebellar ataxias has been established (544-547).

However, most data stems from genetic association, post-mortem, computational, isolated model-system or animal studies (548-552) with respective limitations such as post-mortem tissue decay, homology differences between mice/animal models and human, reductionistic approaches in isolated biological or computational model- systems, and the exact underlying pathomechanism linking IP3R1s to cortical and cerebellar degeneration in humans remains incompletely studied due to the lack of a human neuronal model. iPSC technology and directed neural differentiation: A brief overview iPSC technology, a fairly recent technology still, has raised exciting new horizons in the generation and study of cellular disease models: As explained in more detail in the Introduction Chapter, one can generate induced pluripotent stem cells from patient skin fibroblasts or other somatically differentiated cells via gene delivery of four key transcription factors, a process called reprogramming, that undergoes constant improvement and has become a routine approach for many laboratories world-wide (264, 553). Similar to human embryonic cells, human iPSCs from patients are pluripotent, capable of self-renewal and carry the exact genetic information (causal defect plus background). Differentiation into neural cells has become a routine approach, as described in more detail in the Introduction Chapter. This allows the generation of faithful in vitro cell models of genetic diseases that possibly would not be amenable to investigation

212

otherwise. Even though this technology still has several shortcomings (e.g. low differentiation efficiency, recapitulation of (often late-onset) disease in a developmental and therefore highly dynamic and essentially fetal system (289, 554), etc.), an ever growing number of studies are finding cellular phenotypes of disease using this system and are therefore validating its utility in modelling adult-onset neurodegeneration (see (555) for a recent review). The neural differentiation protocol used in this study follows all steps strictly as described in detail step by step in Shi et al, Nature Protocols, 2012 (237). RNA sequencing, live cell imaging, electrophysiology and immunocytochemistry were used to validate neuronal properties and efficiency of differentiation as explained in the Introduction Chapter above.

Cellular phenotyping and disease modelling When creating a disease model, choice of phenotypic readout assays are equally important as the choice of the model system. In the analysis of iPSC-derived neurons for disease modelling, phenotypic assays can range from biochemical target readouts, cell viability assays and morphological studies to electrophysiology and live cell imaging, depending on the initial hypotheses and disease of study. Usually, a combination of hypothesis-driven and hypothesis-free explorative experiments are performed in order to assess the phenotype. This study chose chose a combination of ICC, qPCR and RNA sequencing to firstly assess the question whether the assumed pathogenicity in SCA15 is due to haploinsufficiency of ITPR1 (given the large size of the ITPR1 transcript and protein, the presence of various ITPR1 transcripts and its homologous regions shared with ITPR2 and ITPR3 as well as the difficulty to detect subtle dosage differences in expression levels within delicate structures such as iPSC-derived neurons this was expected to be difficult). Secondly, electrophysiology, ICC and live cell imaging were used to establish and validate neuronal properties of derived cells. Thirdly, live cell imaging (especially assays investigating thapsigargin induced calcium stores and store operated calcium entry) was employed to detect potential differences in the cells’ ability to handle calcium (see Figure 2-5 for neural differentiation protocol and a timeline of experimental readouts). Given the role of ITPR1 in the regulation of intracellular calcium, we hypothesized to detect differences in the overall thapsigargin induced calcium store size and store operated calcium entry of the mutant cells. Briefly, store operated calcium entry, short SOCE, is a phenomenon activated upon calcium depletion from the internal calcium stores. The endoplasmic reticulum and the mitochondria represent the major calcium storing intracellular organelles. Calcium depletion from the ER can occur upon activation of IP3R1, the ligand gated calcium

213

channel releasing ER calcium into the cytoplasm, and thought to be haploinsufficient in SCA15. This calcium store depletion gets sensed by calcium sensitive proteins (e.g. STIM1, STIM2) that can undergo conformational changes and translocation in order to activate the store-operated ORAI1 calcium ion channels on the plasma membrane, thereby refilling cellular plasma calcium first and organelle storage calcium secondly (556). We therefore hypothesized that a heterozygous deletion in ITPR1 might have an impact on the cell’s ability to handle, store and release intracellular calcium.

Given the shortcoming and limitations of previously reported models of SCA15- associated neurodegeneration, this chapter sets out to generate the first human neuronal iPSC-based model of SCA15 to study the cellular pathology of ITPR1 deletions in differentiated neurons.

6.1.2 Results

SCA15 patients Three individuals from two unrelated published British SCA15 families (529, 530) were contacted and clinically re-examined, videotaped and biopsied. Figure 6-1 shows their pedigrees and deletion sizes (for family 2, the exact breakpoint of the deletion could not be determined, but converged between exon 30 and 40 of ITPR1). Table 6 – 1 summarizes the clinical findings and Figure 6 – 2 shows the MRI of individual ST at age 37 with prominent global cerebellar atrophy, more pronounced over the superior vermis.

Figure 6-1: Pedigrees SCA15 families and ITPR1 deletion sizes

Arrows indicate the biopsied individuals and their respective ITPR1 deletion size. Figure rearranged from the two respective publications (529, 530) and Joyce van de Leemput’s doctoral thesis (UCL).

214

Table 6-1: Clinical findings of SCA15 patients

Individual CT Individual ST Individual MD

First symptoms at age Early twenties Age at onset 26 years 10 (~21 years)

Severe dysarthric Cerebellar ataxia Ataxias, speech, ataxic gait (le>ri), mild upper limb oscillopsia, Main clinical with wheelchair tremor, dysarthria – dysarthria, mild symptoms dependency due to note: slower cognitive increased risks of falls, progression and later executive mild cognitive decline onset than father CT problems Cerebellar Cerebellar and vermal atrophy, MRI No MRI available atrophy, see Figure 6- preferentially 2 involving the superior vermis Age at biopsy 72 40 46

Gender M M F

Figure 6-2: Sagittal MRI of individual ST at age 37

Sagittal T1-weighted MRI of individual ST at age 37 with pronounced cerebellar atrophy.

Generation and validation of iPSC lines Figure 6-3 illustrates the morphological changes of the cells going from initial skin biopsy-tissue (flask in upper left corner) to patient fibroblasts, via transfection to iPSCs and finally to emerging neurons.

215

Figure 6-3: Developmental stages: Punch biopsy, fibroblasts, iPSCs, neuronal precursors

Brightfield images of developmental steps of SCA15 fibroblasts in their process of becoming neurons (cell culture microscope Nikon Eclipse TS100, magnification of objective 10x, magnification of lens 10x/22).

Brightfield morphology of generated iPSC lines All six patient iPSC clones morphologically and functionally behaved like control iPSCs and ESCs in culture (see Figure 6-4).

216

Figure 6-4: Brightfield morphology iPSC clones

Brightfield images of six generated iPSC lines and 2 control iPSC and one control embryonic stem cell line (cell culture microscope Nikon Eclipse TS100, magnification of objective 10x, magnification of lens 10x/22). Clones are named and counted as derived from individuals. P after the clone name indicates passage number of iPSCs.

Genotyping of generated iPSC clones to confirm presence of deletion High-density genome-wide genotyping using the HumanOmniExpress BeadChip Kit which covers more than 713014 single-nucleotide polymorphisms reveals presence of heterozygous deletion of ITPR1 in all six clones and absence of a deletion in the three control lines (see Figure 6-5 for exemplary results on one control and one patient line).

As hypothesized, the B-allele frequency plots (respective upper panels of GenomeStudio Software view in A and B of Figure 6-5) show expected results (red boxes) with genotypes indicating heterozygosity for the control lines (blue dots at B- allele frequency 0.5, along the middle line) and loss of heterozygosity for the patient lines (no blue dots at B-allele frequency 0.5, along the middle line) over the SUMF1 and ITPR1 locus. However, in cases of loss of heterozygosity, the Log-R ratio (as a measure for copy number status, here depicted in the panel below B-allele frequency) is expected to go below zero at the deleted locus (see red line and red arrows in A and B of Figure 6-5). In the data reported here this could not be observed, however, when comparing Log-R ratios of controls versus patient lines, it became apparent that the Log-R ratios of the control lines went above zero and they stayed at baseline level for the patient lines instead (see red line and red arrows indicating the expected trend in B in Figure 6-5, and red line and red arrows in A in Figure 6-5 indicating the observed trend of upgoing Log-R ratio). 217

After consultation with different experts on genotyping, the conclusion was reached that this is most likely to do with the analysed sample architecture and the analysis settings in GenomeStudio: I analysed three control lines with presumed heterozygosity over the ITPR1/SUMF1 locus and six patient lines with presumed loss of heterozygosity. GenomeStudio Software (under my analysis settings) scanned the genotyping data and set the copy number status results of the majority of the analysed nine samples (hence the six patient samples with loss of heterozygosity) as “normal” baseline (it thereby assumes deletions and loss of heterozygosity over SUMF1/ITPR1 as default). It then displays the ones in minority (here: the three control lines with normal heterozygosity over SUMF1/ITPR1) in relation to the basal, suggesting that they have one allele more than the patient lines at this locus – which is essentially true. Therefore, in all likelihood the Log-R ratio line goes up in controls and stays at baseline levels in the patient lines in the analysis shown here.

218

Figure 6-5: Loss of heterozygosity over SUMF1/ITPR1 locus

The lower panel (B) shows loss of heterozygosity over the SUMF1 and ITPR1 region for one exemplary clone, confirming the deletion is still present in the self-derived iPSC clones (see red boxes). Exemplary normal B-Allele-frequency and Log-R Ratio for one control line is shown in upper panel (A), excluding the presence of a deletion. Microarray processing was carried out by Kerra Pearce as a UCL service.

Normal karyotypes for all generated lines G-band karyotyping (CGS, Cambridge) revealed a normal karyotype and expected gender without major chromosomal abnormalities, deletions, rearrangements or translocations for all six generated iPSC lines (see Figure 6-6). The control cell lines had been karyotyped previously by collaborators and Coriell and therefore their normal karyotyping results are not shown here.

219

Figure 6-6: Normal G-band karyotyping results

Normal G-Band karyotyping results for all six generated patient lines (clones from top left to bottom right: CT1, CT2, ST1, ST2, MD3, MD5). Karyotyping carried out by Cell Guidance Systems, Babraham Research Campus, Cambridge.

Expression of pluripotency markers (qPCR) cDNA from all generated clones, as well as their original fibroblasts and the iPSC and ESC control lines was analysed on qPCR showing that all six generated iPSC lines express common pluripotency associated genes NANOG, SOX2 and OCT3/4 in a similar fashion to the embryonic stem cell control line and the iPSC controls. Conversely, cDNA taken from pooled fibroblasts of these patients do not express any of the pluripotency markers (see Figure 6-7).

Please note that these experiments were done on RNA harvested from each self- generated clone and the control lines only once (not as with other experiments at least three times) since they represent a complementary pluripotency confirming experiment in a range of different pluripotency readouts (ICC, exogenous and endogenous pluripotency expression, in vitro differentiation capacity).

220

Figure 6-7: Pluripotency marker expression iPSC lines

Pluripotency markers expressed in iPSC lines and not expressed in original fibroblasts (all results from 3 patient fibroblasts pooled here; all results from both iPSC controls pooled here) (ESC control normalized to 1). n=1. Abbreviations: a.u.=arbitrary units; norm.=normalised.

Expression of pluripotency markers (ICC) Immunocytochemistry for pluripotency markers SSEA4 and OCT4 and counterstaining with DAPI revealed SSEA4 and OCT4-positivity for all six generated lines and the ESC and iPSC control lines (see Figure 6-8 for exemplary ICC results on two patient and one control line). Clones MD5 (in the left lower part of the merged image) and CT1 (in the middle between the two depicted colonies) show some DAPI-positive cells that are negative for SSEA4 and OCT4, hereby functioning as internal negative control of the staining conditions. These cells most likely represent a carry-over of inefficiently and incompletely reprogrammed cells that remain fibroblastic or intermediate non- pluripotent cell types. The human embryonic stem cell control line functions as a positive control here, and ICC for pluripotency markers is shown as single and merged channel data for illustrative reasons in Figure 6-9, however, the majority of ICC data in the following will be presented as merged channel data for space saving reasons.

221

Figure 6-8: Pluripotency immunocytochemistry iPSC lines

Pluripotency immunolabelling showing SSEA4- and OCT4-positive exemplary iPSC colonies for two patient clones and one exemplary control. P indicates passage number. Scale bar (applies to all images): 100 µm (Please note: There is some slight, unfortunate degree of autofluorescence of DAPI stain due to plastic coverslips).

Figure 6-9: Pluripotency immunocytochemistry hESC line

Pluripotency immunolabelling showing exemplary SSEA4- and OCT4-positive hESC colonies as a positive control. Scale bar (applies to all images): 100 µm.

222

Downregulation of fibroblast marker S100A4 (qPCR) All six generated iPSC lines (as well as the embryonic stem cell line and the control iPSC lines) had silenced the fibroblast marker S100A4 compared to the original fibroblasts before transfection that highly express this marker (see Figure 6-10).

Please note that these experiments were done on RNA harvested from each self- generated clone and the control lines only once (not as with other experiments at least three times) since they represent a complementary confirming experiment in compliance with the behavioural and morphological changes that suggest these cells have lost their fibroblast identity.

Figure 6-10: S100A4 expression fibroblasts, iPSC lines

Fibroblast marker S100A4 retained in original patient fibroblasts but downregulated in ESC line and all self-generated as well as control iPSC lines. (all results from 3 patient fibroblasts pooled here; all results from both iPSC controls pooled here) (ESC control normalized to 1). n=1. Abbreviations: a.u.=arbitrary units; norm.=normalised.

Silencing of plasmid pluripotency transcripts after passage 10 (qPCR) cDNA from all generated clones between passage 10 and 12 showed robust expression of endogenous OCT4 (see Figure 6-11, A) and silencing of exogenous plasmid SOX2 and L-myc (see Figure 6-11, B and C). Five out of six generated lines had additionally silenced the exogenous OCT4. However, one clone (MD5) had not silenced it even after twelve passages in culture and therefore it is suspected to have integrated the plasmid carrying OCT4 into its genome (see Figure 6-11, D).

These results were confirmed in one independent round of RNA extraction and qPCR analysis.

223

This analysis was not done for the iPSC control lines since they were reprogrammed retrovirally.

Figure 6-11: Endogenous and exogenous pluripotency marker expression

Expression of endogenous and exogenous (plasmid) pluripotency factors in iPSC clones, shown as amplification plots on qPCR with number of cycles plotted on the x-axis and fluorescence intensity on the y-axis.

Differentiation into neural precursors and mature cortical neurons

Exemplary immunocytochemistry labelling of different patient lines at DIV15 and DIV34 shows neural rosettes express the forebrain marker OTX2 throughout their development (see positive expression on DIV15 and 34 shown in Figure 6-12), are positive for the early pan-neural marker PAX6 on day 15 and contain a proportion of mitotically active cells positive for Ki67 on day 34 (Figure 6-12, lower panels). Rosette formation and positivity for early neuronal markers could be achieved in all patient and control lines equally throughout differentiation.

224

Figure 6-12: Immunocytochemistry neural precursors

Exemplary immunocytochemistry for early neuronal markers OTX2 and PAX6 (upper two lanes, merge (left) and single (right)) and cycling marker Ki67 (lower panel, merge (left) and single (right)) in patient-derived neural precursors at DIV15 and DIV34. Scale bar (applies to all images): 100 µm.

Mature cortical neurons: Immunocytochemistry Immunocytochemistry for all six patient and control lines at DIV80 showed that all generated neurons are positive for the panneuronal marker TUJ1 and show typical neuronal morphology and synapse formation (Figure 6-14). However, please note how different confluency of each individual line potentially influences the strength of markers of cortical maturity and the staining quality: The postsynaptic density protein 95 (PSD95) which is highly expressed at the postsynaptic density of mature (glutamatergic) neurons, serves as an indirect marker for maturity of the generated neurons here: Very confluent patient neurons (e.g. MD5, see lowest panel to the right, Figure 6-14) and equally confluent control neurons (e.g. right panel, Figure 6-13) express PSD95 in a weaker fashion at some synapses. At the same time, highly confluent cultures are very difficult to image. The MD5 staining in the lowest panel to the right show DAPI-positive cells (most likely glia cells) that are negative for PSD95 and TUJ1, representing an internal negative control to confirm the validity of the staining. Single channel data is provided in Figure 6-15, however, the majority of ICC images shown here are presented as merged channels only, due to space saving reasons.

225

Figure 6-13: Immunocytochemistry cortical control neurons DIV80

TUJ1-, DAPI- and PSD95-staining for cortical control neurons at DIV80. Scale bar: 25 µm. x-magnification of objective on imaging-microscope indicated on each image.

Figure 6-14: Immunocytochemistry cortical patient neurons DIV80

TUJ1-, DAPI- and PSD95-staining for patients’ cortical neurons at DIV80 for clones CT1, CT2 and ST1 (left half) and clones ST2, MD3 and MD5 (right half). Please note that images of CT2 are different magnifications/focus of the same group of cells. Scale bar: 25 µm. x- magnification of objective on imaging-microscope indicated on each image.

226

Figure 6-15: Immunocytochemistry control neurons, single channel data

Single channel and merged PSD95-, TUJ1-, DAPI-staining for cortical control neurons at DIV80. Scale bar: 25 µm.

The analysis of ITPR1 expression via ICC, Western Blot and qPCR cannot be reported conclusively here due to difficulty with different ITPR1 antibodies (inspecificity on WB and ICC, large protein, frequent problems with protein transfer during WB) and ITPR1 primers in qPCR (different transcripts, shared homology between ITPR1, ITPR2 and ITPR3).

Neuronal astrocytic-coculture As reported in the Shi-protocol that was followed for cortical differentiation, astrocytes and other glia are prone to appear in neurally differentiated cultures around DIV40-50 (237) in coculture. The astrocytic population is increasing in proportion over time recapitulating developmental phenomena with gliogenesis following neurogenesis in this in vitro model. In the cultures an increase of glia cells (predominantly astrocytes) in some parts of the wells (especially in areas of the well where previous neurons had detached from) were observed around DIV100 (e.g. see ST1, ST2 in Figure 6-18), in other areas neurons had formed synapses with astrocytes and the proportion of healthy neurons was still the major populating cell type of the dish (e.g. see Figure 6-18, CT1, MD3 and MD5).

227

This was observed for control cultures equally (Figure 6-17). Morphologically, astrocytes were predominant, but radial glia could be observed (Figure 6-19).

Single channel data is provided in Figure 6-16, however, the majority of ICC images shown here are presented as merged channels only, due to space saving reasons.

Given the fact that astrocytes are dependent on ITPR1 to initiate calcium waves (557) additional live cell imaging trial experiments (under the supervision of Dr. Zhi Yao, IoN) were performed, investigating the store operated calcium entry of iPSC-derived astrocytes. However, no differences between patient and control astrocytes could be identified (data not shown).

Figure 6-16: Immunocytochemistry control neurons and glia, single channel data

GFAP-, TUJ1- and DAPI-staining for control neurons and glia at DIV100. Scale bar: 25 µm.

228

Figure 6-17: Immunocytochemistry control glia cells DIV100

GFAP-, DAPI- and TUJ1-staining for control neurons and glia at DIV100. Scale bar: 25 µm. x-magnification of objective on imaging-microscope indicated on each image.

Figure 6-18: Immunocytochemistry patient glia cells DIV100

GLAST/GFAP-, DAPI- and TUJ1-staining for patient neurons and glia cells at DIV100 for clones CT1, CT2 and ST1 (left half) and clones ST2, MD3 and MD5 (right half). Scale bar: 25 µm. x-magnification of objective on imaging-microscope indicated on each image.

229

Figure 6-19: Astrocyte and radial glia DIV100

GLAST/GFAP-, DAPI- and TUJ1-staining for astrocyte (left panel – please note, this astrocyte is a magnification from image ST1 in Figure 5-16) and radial glia (right panel) observed at DIV100. Scale bar: 25 µm.

Live cell imaging Conditions for live cell imaging in iPSC-derived neurons had to be optimized, but I generated good quality data with sufficiently powered n-numbers from two patient derived (MD3 and CT2) and one control line (Control 1) for statistical analyses. An exemplary dynamic trace of one of the performed experiments at DIV 59-62 investigating thapsigargin induced calcium release for MD3, CT2 and Control 1 is shown in Figure 6-20. No statistically significant differences in calcium pool in ER stores and store operated calcium entry at DIV59-62, DIV77 and DIV93 were detected (see Figure 6-21).

Equal density of cells was ensured by counting cells prior to equal seeding on coverslips, imaging plates and ibidi-plates for live cell imaging readouts. In case inherent biological differences between lines resulted in different growth speed up to the day of analysis, only density-matched wells for comparative analysis of control and patient iPSC-derived neurons were selected.

230

Figure 6-20: Exemplary dynamic traces, thapsigargin induced calcium stores

Exemplary dynamic trace of live cell imaging experiments showing Fura2-fluorescence ratios upon stimulation with thapsigargin and ionomycin. Abbreviations: Tg=thapsigargin.

231

Figure 6-21: ER calcium and store operated calcium entry in iPSC-derived neurons

Live cell imaging results at DIV59-62 (A and B), DIV77 (C and D) and DIV93 (E and F) indicating no significant differences in thapsigargin induced calcium peak (=ER store, A, C, E) or store operated calcium entry (=SOCE, B, D, F) between control and patient neurons. Data is collated from three independent inductions, and experiments were performed on >100 cells in each experiment, error bars represent standard error of the mean (SEM). Abbreviations: ER=endoplasmic reticulum, Tg=thapsigargin, DIV=day in vitro, ns=non-significant.

In cultures from DIV77 and DIV93 spontaneous calcium signalling could be regularly observed, indicating neuronal activity. Additionally, trial mitochondrial membrane potential and reactive oxygen species measurements were performed, but no differences between control and patient lines could be detected (data not shown).

Mature neurons: Electrophysiological properties of neurons at DIV80-100 Electrophysiological properties were investigated in collaboration with Prof. Dr. Dimitri Kullmann and Dr. Sarah Crisp. Patching techniques on iPSC-derived cortical neurons needed to be optimised and established and differentiated neurons are still not very easy to patch. I also experienced problems of frequently detaching neurons when transporting them from Wakefield Street (place of culture) to Queen Square House (place of analysis).

232

Due to the low n-numbers of patched neurons, formal statistical comparisons between control and patient lines were not possible.

However, the electrophysiological traces obtained indicated the presence of spontaneous as well as evoked electrophysiological potentials (cells fired spontaneously and generated single action potentials, short bursts of potentials and sustained firing upon current injection, indicating heterogeneous maturation status of cultures) in analysed cells (Figure 6 – 22 below). The observed activity can be compared to the electrophysiological activity of iPSC- derived neurons at DIV69 in the paper of Shi and colleagues that I follow for cortical differentiation (the latest evaluated time-point in their publication).

Table 6 – 2 gives the electrophysiological properties of one patched examplary control (Control 1) and one patient line (MD3).

Figure 6-22: Electrophysiological activity of iPSC-derived cortical neurons

Patching setup (A) and exemplary electrophysiological traces (B) of control (C) and patient lines (D) indicating inhomogeneous maturity within cultures on DIV80-100.

233

Table 6-2: Electrophysiological properties iPSC-derived neurons

Cell line RMP Take- Peak Amplitude 1/2 width 90-10 AHP (mV) off (mV) (mV) (ms) decay (mV) (mV) (ms) MD3 -36 -45.01 21.31 66.32 1.90 1.96 17.41 Control 1 -62 -35.52 14.99 50.51 2.69 2.29 20.62 Rheobase V Input Res Capacitance Cell line I (pA) Tau (ms) (mV) (Mohm) (pF) MD3 13.94 20 697 24.05 34.51 Control 1 13,8 10 1380 44.8 32.46

Abbreviations: RMP=resting membrane potential; mV=millivolt; ms=milliseconds; AHP=afterhyperpolarization; V=Voltage; I=electric current; pA=picoampere; Res=resistance; tau=R(esistor)C(apacitor) time constant.

6.1.3 Discussion

Generation of the model Firstly, six fully functional and thoroughly characterised iPSC clones were generated from skin samples of three different SCA15 patients that were clinically characterised and biopsied. Generated iPSC clones were able to differentiate to cortical neurons and astrocytes in vitro alongside their control counterparts. They expressed pluripotency factors with expected variation between individual clones potentially due to the mosaic expression within iPSC colonies at different time-points, retained the original genetic defect and genetic background of the patients biopsied, have silenced expression of their original fibroblast marker gene S100A4, and five out of six clones have silenced the exogenous plasmid DNA. However, the qPCR results amplifying plasmid pluripotency factors indicate that one clone (out of six) has most likely integrated the plasmid pCXLE-hOCT3/4-shp53 carrying OCT3/4 (one of the three plasmids of plasmid combination Y4 that I used for transfection, see (235)). Going back to this original paper that I utilised as a guide for the transfection, the authors report one clone out of seven with suspected integration of the plasmid into its genome (235). Even though unfortunate, the integration rate in this experiment is thereby comparable to the original experiment. The clone that showed integration was able to differentiate into a mature cortical neuron with positive TUJ1 and PSD95 staining (see Figure 6-14 for clone MD5), however, it was not utilised for further experimental readouts due to the uncertainty of the downstream effects of plasmid integration into the genome.

234

This point is clearly debatable and one could equally argue for inclusion of this clone into the planned and conducted experiments assuming that one would most probably find complete silencing of the likely integrated, exogenous pCXLE during neuronal differentiation and maturation. The iPSC validation employed here was done within the financial and technical restraints of the laboratory, and is lacking the traditional gold standard of iPSC validation showing the cells’ potential for teratoma formation in vivo in immunodeficient mice. Furthermore, this study only analysed a subset of targeted pluripotency markers instead of a more extensive gene expression profile approach as e.g. established within PluriTest (558).

Additionally, one could critically remark about the study presented here that findings of iPSC lines generated by transfection with episomal plasmids (self-generated lines) are compared with findings of control lines generated by retroviral transfection since in general, any biological differences obtained in the results might be mirroring in part the different reprogramming techniques or might be masking the inherent disease related difference. Though certainly not ideal, it has to be noted, that both techniques are valid, safe and widely accepted reprogramming techniques and that there is no scientific data yet to suggest that lines efficiently reprogrammed by different techniques behave differently. See the cited recent review article on the safety and efficacy of different reprogramming techniques (559).

Even though I generated a validated model that represents to our best knowledge the first human neuronal model of spinocerebellar ataxia type 15, no disease related difference could be observed so far. This is another argument that the theoretical critique above is less applicable to the study presented here. However, in the future, one should aim to utilise control cells that were generated by the same technique as the cells containing the disease related mutation or defect that is in the focus of the study. Choosing age matched control donors to generate control iPSCs seems to be less important since a recent study showed that once successfully reprogrammed to iPSCs, the iPSCs rejuvenate, and erase their age-related transcriptional and functional signature (560). Therefore, the fact that employed control cells were only partially age- matched to patient lines is likely to be of subordinate importance. The perfect gender match of control and patient lines in this study represents an advantage of the model increasing accuracy and limiting variance associated to non-disease related characteristics. Nowadays, the generation of isogenic lines via genome editing techniques has become an important alternative to collecting healthy control and disease cell lines from the population (561). However, these techniques were not as

235

established by the time when I started the SCA15 work in the iPSC field, and the large heterozygous deletion in ITPR1 causal in SCA15 makes the repair via genome editing techniques extremely challenging. The deletion disorders are among the most difficult diseases to model via iPSC technology given the difficulty to genetically correct the deficit once a phenotype has been identified and to reverse the observed phenotype by genetic correction of the underlying defect as a final proof.

After having shown that the patient lines are fully validated iPSC lines that might become an important tool in order to study the cellular phenomena underlying SCA15- associated neurodegeneration in humans, the study set out to investigate the cellular phenotype in differentiated neurons. Unfortunately, my inconclusive results from WB, qPCR and ICC (not shown here) could not conclusively answer the question whether haploinsufficiency of ITPR1 is the underlying pathogenic mechanism in our generated model and renders the correct interpretation of all acquired secondary functional data additionally more difficult.

However, I had no evidence that heterozygous ITPR1 deletions introduce a neurodevelopmental block in vitro, as all six patient lines were able to differentiate into cortical progenitors and finally into cortical neurons with convincing morphology, mature electrophysiological properties and spontaneous calcium signalling alongside their control counterparts. However, it has to be kept in mind that this is qualitative and descriptive as presented here and I did not specifically look for a developmental phenotype given the late-onset character of the disease in humans (526, 529, 530) and the recent finding that mice lacking IP3R1 specifically in Purkinje cells do not show morphological abnormalities in cortex and cerebellum during development but only display severe increase in spine length and spine density of Purkinje cells later in adulthood (535). In order to investigate the hypothesis of SCA15 as a developmental disorder, staining for the different cortical layers (TBR1, CUX1, CTIP2, SATB2, BRN2) at earlier and consecutive time-points should be performed in this model to check for equality of neural conversion and corticogenesis more systematically.

As mentioned, a developmental phenotype did not represent the focus of the hypothesis in the SCA15 phenotyping experiments. Given earlier findings from computational and animal work (535, 548, 562), it was rather hypothesized that differences in sensitivity and abundance of IP3R1 that could potentially result in disturbed calcium homeostasis in the mature, fully differentiated neuronal cells might be present, also in cortical neurons. Up to this point, this study could neither find any differences in the calcium stores from the endoplasmic reticulum elicited by live cell

236

stimulation with thapsigargin (an inhibitor of the sarco/endoplasmic reticulum Ca2+- ATPase, thereby giving an indirect account of the overall ER calcium store) nor any differences in store operated calcium entry even though it was initially hypothesized that a heterozygous deletion in ITPR1 might have an impact on the cell’s ability to handle, store and release intracellular calcium. A statistically non-significant trend of upregulated store operated calcium entry and ER calcium store size in patient lines at DIV77 was observed. However, and more importantly, the single (separated by time-points of live cell imaging) and overall results (all time-points together) were statistically not significant. The observation of apparent variation in results between the investigated time-points is most likely due to the differential maturation status of the cells at the different time-points. Ideally, it should therefore be matched with additional maturation assays at each reported time-point which this study fails to deliver.

Additionally, the fact that no significant calcium handling deficiency was apparent in the generated cells when analysing both patient lines and comparing them to the control line, might be down to various reasons: 1. This study has a major detriment that might be responsible for the absence of a phenotype: The unavailability of the right cell subtype (Purkinje cells) that is thought to be of major vulnerability in SCA15. The clinical presentation of SCA15 (a rather pure cerebellar syndrome with only infrequent cortical features) as well as the fact that ITPR1 deletions are assumed to have the biggest effect in Purkinje cells where ITPR1-dependent coincidence detection is crucial for synaptic plasticity (339) might speak in favour of this hypothesis. Secondly, due to difficulty with ITPR1 antibodies and qPCR results, this study fails to show important expression levels of ITPR1 in the presented model. 2. However, the absence of detection can, but does not necessarily mean that

there is no phenotype in the investigated cells: It might also be an indicator that we did not choose the right phenotypic assays, not the right time-points (e.g. the cells were not aged enough, but SCA15 is an adult-onset disease, etc.) or the model was not investigated with the right experimental and technical equipment (e.g. a problem of assay sensitivity, temporal or spatial resolution in employed techniques and the difficult inherent nature of iPSC-derived neuronal growth (e.g. confluency) that excludes certain techniques). For example, it could well be that SCA15 iPSC-derived cortical neurons might exhibit a calcium phenotype, e.g. if investigated with higher resolution imaging techniques such as recently established by Zheng et al. (563) that instead of looking at clear changes in calcium handling in the somae, would additionally allow to look at

237

very subtle calcium changes with much higher temporal and spatial resolution in individual cell processes. Addtionally, one could have used genetically encoded constructs to measure calcium, however, these techniques are difficult to achieve in iPSC lines. 3. Inherent lab, protocol and differentiation related difficulties might have contributed to the negative phenotyping data. These include a) Long generation time for cortical neurons until ready for experimental readout (up to DIV100). b) Transportation from WS (place of culture) to IoN (place of analysis) as a major risk factor for detachment, cellular stress or contamination. c) Choice of a rather restricted and challenging set of successful readouts (live cell imaging, ICC, electrophysiology), and unsuccessful or yet pending readouts (RNA sequencing, qPCR and Western Blot (WB), yielding results difficult to interpret due to computation-intense analysis (RNA sequencing) or unspecificity of primers (qPCR) and antibodies (WB) for ITPR1). d) Shortage of time at analysis facilities and dependence on collaborators might have additionally rendered thorough and patient investigation of the cellular phenotype more difficult in this model. e) Maybe the tendency of increased store operated calcium entry and ER calcium size in patient lines would have become more apparent if more control and patient iPSC-derived neurons had been analysed successfully and one had generated more data from the remaining patient and control lines.

Finally, a combination of all or some reasons stated above might explain the absence of an identified phenotype in this study.

Generally, it has to be critically appreciated that the pluripotency ICC results were not quantified systematically in this study. However, their qualitative pluripotency confirming character complemented the quantified qPCR pluripotency readouts and morphological and behavioural observation in culture and gave valid evidence of pluripotency. Furthermore, the neural precursor ICC results were not quantified systematically, since qualitative observations confirming their neural precursor identity were more important at this point of maturation. Finally, the ICC results of mature cortical neurons were not quantified systematically. Given the confluent nature of the cells (that also impacts negatively on the resolution of presented images) this would have been challenging, with a bias to distortions and would have certainly required

238

extensive imaging at the confocal microscope. Together with valid and extensive evidence from live cell imaging (neural responses to glutamate, not to ATP, spontaneous calcium signalling), qualitative ICC positive for neural markers and eletrophysiological activity however, the neural identitity of the investigated cells here was not to be questioned. However, the absence of quantification represents an important shortcoming of the approach and might negatively impact the interpretation of the functional data presented here.

Nonetheless, the results of this study led to do further work towards two major advancements: 1) the acceleration of the current protocols for neuronal differentiation to prevent detaching and unhealthy conditions due to long-term culture and 2) the generation of more region-specific neurons with hindbrain or cerebellar properties as the main pathology in SCA15 takes part in the cerebellum.

Parts of these efforts are reported in the next subchapter.

Future work Even though it has not identified a cellular phenotype/defect in the SCA15 neurons yet, the cortical model should not be abandoned. Several collaborative and non- collaborative investigations are still going on: As mentioned and described in more detail in the Methods Chapter RNA sequencing data from cortical patient and control neurons at DIV80-DIV100 was generated. Even though the data analysis is still ongoing, preliminary results suggest that patient neurons retain a transcriptional signature with lower ITPR1-expression and a number of differentially expressed downstream genes affecting various KEGG-pathways, which differentiates them transcriptionally from their control counterparts. These transcriptional differences will further be examined and might reform the hypothesis in order to design a live cell experiment confirming and functionally complementing the transcriptional abnormalities.

Even though there were problems with various antibodies for ITPR1 and so far generated data is difficult to interpret and therefore has been left out at this point, it needs to be assessed whether the so far assumed haploinsufficiency of ITPR1 as a basis of pathology holds true in the human neuronal model. It should be analysed using the already harvested cell pellets from different time-points of the above study, comparing ITPR1-transcript levels (qPCR), immunocytochemistry of ITPR1 and ITPR1

239

protein levels (WB) between control and mutant iPSC-derived neurons once more reliable antibodies for ITPR1 become available.

Even though iPSC-derived cortical neurons are tricky to patch, the study of the downstream effects of ITPR1 deletions with electrophysiology (Prof. Dimitri Kullmann and Dr. Sarah Crisp: differences in generation of action potentials and spontaneous firing properties) should be continued.

Lastly, even though no differences in trial experiments investigating SOCE and ER calcium stores in astrocytes could be detected here, it would be worth continuing astrocyte differentiation from the SCA15 iPSC lines in order to study possible glial cell autonomous effects contributing to pathology in spinocerebellar ataxia type 15. As IP3R1 is crucial for calcium-wave initiation in astrocytes (557) this could be an interesting phenomenon of study.

Conclusions Here, the first human neuronal model of spinocerebellar ataxia type 15 was established. The results in this subchapter highlight the difficulties, limitations and small-scale progress when establishing desperately needed, but novel disease models.

They furthermore show the importance of investigation of the right neuronal subpopulation and availability of efficient differentiation strategies towards cerebellar neurons when studying cerebellar diseases. The results shown here motivated the investigation of shorter protocols that direct cells towards hindbrain fates with the ultimate aim of generating a model to uncover the molecular events underlying selective vulnerability in cerebellar diseases.

240

6.2 Generating non-cortical neuronal derivatives to improve the accuracy in modelling cerebellar disease

6.2.1 Introduction

Given the fact that cortical neurons are only a surrogate that might lack a truthful recapitulation of disease phenotypes and cellular events underlying cerebellar neurodegeneration, this subsequent study set out to generate non-cortical neuronal derivatives.

This substudy wanted to explore the necessary extrinsic guidance and the duration of exposure of these cues that iPSCs need in order to firstly become efficiently neuralized and secondly adopt non-forebrain cell fates with the ultimate aim of establishing these cells for disease modelling of cerebellar disorders. Given that a validated SCA15 iPSC model is now available (see Chapter 6.1) and iPSCs of further cerebellar ataxia subtypes are following (see Future Work below, HipSci project) this would be of immense value for the future since cerebellar diseases still lack full benefit from iPSC technology (251).

Few cerebellar protocols are available, e.g. Erceg et al. (326) and Muguruma et al. (330), described in more detail in the general introduction (Chapter 1, last subchapter). However, they have the potential limitations such as administration of a multitude of different extrinsic cues, or uncontrolled intrinsic signalling due to 3D-based culture system. The reproduction of the Erceg et al. protocol (joint work with Dr. Abi Li, RLWI, WS; data not shown) has so far not been successful yet; and a modification and simplification of this protocol seemed necessary. In a second attempt, the team of Muguruma et al. at the RIKEN Institute in Japan was contacted with the aim of a potential collaboration on the SCA15 lines with generation of Purkinje cells under their guidance, however no feasible collaboration agreement was established. Consequently, this study therefore sets out to generate a novel 2D-protocol using the selective inhibitor of glycogen synthase kinase 3, CHIR99021, which promotes WNT- signalling and has been successfully utilized in differentiation of hiPSCs to dopaminergic neurons, spinal cord motor neurons and other non-forebrain cell types (312, 564-566). However, the exact influence of CHIR99021 on regional cell fate during early neural induction and neural patterning of iPSCs is largely unexplored and therefore was in the focus of this experiment. See Figure 6-23 for a schematic of experimental rationale.

241

Figure 6-23: Rationale of developmental experiments

Insights from the Erceg et al. protocol depicted in the left, and the 3D-culture protocol devised by Muguruma et al., led to the exploration of a 2D-culture based approach using WNT-agonist CHIR99021. Abbreviations: GSK3=Glycogen-synthase kinase 3. Figure collated from respective publications (326, 329, 330).

These facts led to the investigation of the “best” induction paradigm using CHIR99021 with a) cells surviving the induction phase b) pluripotency markers switched off after an ideally short amount of days (4 to 7 days) and c) cells preferentially expressing caudal markers or showing at least a reduction of their strong forebrain signature usually obtained with the conventional dual SMAD inhibition during neural induction.

6.2.2 Results

Experiment 1 yielded expression results at day 0, day 4 and day 7 for twelve different markers (Pluripotency: OCT4, NANOG, see Figure 6-24 A (OCT4) and B (NANOG). Pluripotency/early neural: SOX2 (data not shown). Mesendoderm: Brachyury (data not shown). Endoderm: GATA4 (see Figure 6-25, A). Mesectoderm: S100A4 (see Figure 6-25, B). Trophectoderm: CDX2 (data not shown). Early neural cells of different CNS regions: OTX2 and FOXG1 (forebrain in combination; OTX2 alone also midbrain), PAX6, Musashi, SOX1). It shows robust expression of pluripotency markers prior to neural induction (day 0) and efficient and homogenous silencing of pluripotency

242

markers as early as day 4 which remain efficiently silenced at day 7 for each of the investigated six conditions equally (see Figure 6-24).

Figure 6-24: Pluripotency marker expression timecourse for six different conditions

Expression timecourse (x-axis: arbitrary units of marker expression) for pluripotency markers NANOG and OCT4 for conditions A – H (left to right in each graph) at day 0 (D0), day 4 (D4) and day 7 (D7). Error bars indicate standard error of the mean (SEM) (when not visible, these were smaller than the datapoint depiction in the graph).

Markers indicating the different germ layers in pluripotent cells such as mesendoderm, endoderm, mesectoderm and trophectoderm (see GATA4 (endoderm) and S100A4 (mesectoderm) exemplarily shown in Figure 6-25) are highest expressed at day 0 and then efficiently downregulated at day 4 and 7 (even though with more variation than the pluripotency markers) indicating that an efficient fate change towards neuroectoderm is initiated at day 4 and 7 upon neural induction irrespective of the induction morphogens, that represses the other, non-neuroectoderm germ layers.

Figure 6-25: GATA4 and S100A4 expression timecourse for six different conditions

Expression timecourse (x-axis: arbitrary units of marker expression) for exemplary markers indicating endodermal (GATA4, A) or mesectodermal contamination (S100A4, B) for conditions A – H (left to right in each graph) at day 0 (D0), day 4 (D4) and day 7 (D7). Error bars indicate SEM (when not visible, these were smaller than the datapoint depiction in the graph).

Figure 6-26 gives an overview of the most informative timecourses of different neural markers for the six different conditions. It shows an efficient upregulation of forebrain

243

markers such as OTX2 and FOXG1 only for conditions A and less strongly for conditions G and H at day 4 and 7. Induction onditions B, C and D (all with prominent WNT-agonism) however repress the development of forebrain fates (see Figure 6-26).

Figure 6-26: Timecourse neural markers six different conditions

Expression timecourse (x-axis: arbitrary units of marker expression) for exemplary neural markers FOXG1 (A), OTX2 (B), PAX6 (C) and pluripotency/early neural marker SOX1 (D) for conditions A – H (left to right in each graph) at day 0 (D0), day 4 (D4) and day 7 (D7). Error bars indicate SEM (when not visible, these were smaller than the datapoint depiction in the graph).

Obtained results were confirmed on the protein level using targeted immunocytochemistry for the cortical marker OTX2 (see Figure 6-27). They show robust expression of forebrain marker OTX2 on day 4 and day 7 for the traditional forebrain yielding induction protocol A (dual SMAD-inhibition) which is completely suppressed as soon as WNT-agonist CHIR99021 is added during induction from day 0 or day 1 onwards (conditions B, C and D). The later CHIR99021 is added into the media, the less strong is this suppression of forebrain markers, indicating a sensitive time window for induction of non-forebrain fates.

244

Figure 6-27: Immunocytochemistry OTX2-positive cells DIV4 and 7, six different conditions

A: Representative immunocytochemistry images day in vitro 4 and 7 for six different conditions A-H. B: Quantification of OTX2-positive cells for day 4 (B) and day 7 (C), error bars indicate standard deviation. Scale bar in A: 25 µm.

Experiment 2 yielded qPCR expression results at day 14 for nine different regional markers. In the respective graphs for the six different conditions (A-F), these markers are clustered by their region into forebrain markers (OTX2, PAX6) to the left of each graph followed by midbrain markers (OTX2, EN1), hindbrain markers (HOXA1, WNT1, BMP7, GBX2) and spinal cord markers (HOXA4, HOXA5). Since OTX2 is expressed in fore- and midbrain its expression appears twice in each plot. See Figure 6-28 for expression results of different induction and patterning cues at day 14 which show an upregulation of mid- and hindbrain markers in conditions where CHIR99021 was added during the first seven days of neural induction which could even be consolidated and strengthened by usage of CHIR99021 or FGF8 during 7 days of patterning.

245

Figure 6-28: qPCR expression profile of different regional markers at DIV 14

Expression (normalized to GAPDH) in arbitrary units (a.u.) for different regional identity markers in neural precursor cells at DIV14. Error bars represent standard deviation. Please note extremely low expression of condition E.

6.2.3 Discussion

Experiment 1 showed comparable efficiency of all six different induction conditions for neural induction with pluripotency markers losing their expression at day four and remaining downregulated at day seven (see Figure 6-24). This is a good indicator that pluripotency markers are switched off earlier than after ten or twelve days as currently set as the standard induction time for the long cortical differentiation protocol from Shi et al. (237). These results – combined with an upregulation of neural markers as observed for some of the conditions shown in Figure 6-26 (e.g. condition A, G and H) – encourage an optimistic lookout that further shortening of neural induction times to achieve efficient neural conversion becomes possible. With shortening of time periods needed for efficient neural conversion one can save significant amounts of money as less media, less extrinsic cues, less tissue culture utilities and less workforce to induce a pluripotent stem cell towards neuronal fate will be needed. Shorter induction time will also reduce the overall experimental timeframes and significantly facilitates the generation of more replicates and hence more statistically powered (by higher n numbers) and trustworthy experimental findings.

Secondly, qPCR results of different non-neural germ line markers showed a downregulation for all examined germ line markers (mesendoderm: Brachyury; endoderm: GATA4; mesectoderm: S100A4; trophectoderm: CDX2) at day four and day

246

seven (see Figure 6-25). This further supports the efficient suppression of pluripotency and the initiated germ layer choice towards neuroectoderm by all neural induction paradigms.

However, differences in neural marker expression profile are apparent at day four and seven, depending on the induction condition. Especially forebrain markers such as FOXG1 and OTX2 are continuously and efficiently repressed upon presence of CHIR99021 in the media. If dual SMAD inhibition is exchanged for CHIR99021 containing induction media after day 1, 2, or 3, a time-dependent restoration of cortical markers could be observed both using qPCR and ICC (see Figure 6-26 and Figure 6-27). This finding is important and does not only reveal the potency of CHIR99021 in suppressing forebrain cell fate, but also stresses the importance of accurate and disciplined adherence to defined time-points and treatment of all cultures of comparisons at the same time, since “only” 24 hours did change the transcriptional regional profile of the investigated cells considerably.

Experiment 2 helps to work towards usage of different patterning cues necessary for cellular hindbrain like derivatives. This experiment showed that seven days of 3 compounds (Dorsomorphin at 1 µM, SB431542 at 2 µM and CHIR99021 at 3 µM concentration) induction followed by seven days of FGF8-patterning yielded enrichment of hindbrain regional markers at the expense of cortical or midbrain markers (see Figure 6-28). This is a particularly important finding given that early developmental studies in different animal organisms (mouse, chicken, drosophila) have shown expression of EN1/EN2, WNT1, FGF8 in the caudal mesencephalon with extension to the rostral rhombencephalon (567-569) indicating that regional identity of these cells might stem from these regions. A number of signalling pathway ligands and key transcription factors expressed at the midbrain-hindbrain border (EN1/2, GBX2,

OTX2, LIM1b, FGF8, WNT1 proteins) are critical regulators of MHB fate and isthmus activity. The midbrain–hindbrain border (or isthmus organizer region) is crucial for the faultless organization of the anterior hindbrain (rhombomeres 1 and 2) and for the proper induction of the midbrain and midbrain–hindbrain borders. Important convincing and growing experimental evidence suggests FGF8 (the extrinsic cue during patterning that yielded best results in the performed experiments) as the key isthmus organizer inducing protein at the midbrain-hindbrain border (570-575). However, the interaction of FGF8 and WNT1 is important since functional WNT1 activity seems to be prerequisite for correct FGF8 expression levels to stimulate proper MHB specification and isthmus organizer activity in vertebrates. Inter alia, the crucial role of WNT has

247

been shown in WNT1 knock-out mice where proper FGF8 and EN gene expression could not be maintained (576, 577).

Secondly, it has been suggested previously that WNT-signalling (together with BMP- pathway signals) plays a role in regulating granule cell precursor proliferation in the external germinal zone during cerebellar development (578). Here, cerebellar granule cells are suggested to require finely regulated levels of WNT-signalling to stabilise their differentiation and proliferation and an activation of WNT-signalling pathways has resulted in premature differentiation and inhibited proliferation of cerebellar granule precursor cells (578, 579). Therefore, the observed enrichment of hindbrain markers upon induction and patterning with these two extrinsic cues in particular (CHIR99021 as WNT-agonist and FGF8) in this novel 2D-adherent culture-based iPSC system is encouraging.

Showing that general developmental insights gained from the study of non-human model organisms can be generalized to a 2D human iPSC-based adherent culture system here is an important finding and will further inform strategies towards an efficient cerebellar cell differentiation from human iPSCs.

Interestingly, cells induced with 3 compounds as above and then continuously patterned with CHIR99021 for the next seven days had severe culture adherence problems, tendency to die throughout different independent inductions and looked unhealthy mirrored by their very low amplification of regional markers on qPCR (condition E, Figure 6-28).

Conclusions and future work Results yielded with the study described above will be taken further in various ways:

Especially, based on this data, one will be able to significantly shorten induction times in future protocols, saving precious time and money for expensive media and extrinsic cues.

Secondly, given that it yielded an important enrichment of midbrain/hindbrain markers, condition F (seven days of 3 compound induction followed by 7 days of FGF8- patterning) will be taken further to terminally differentiate the obtained precursors using notch-inhibitors, and investigate their regional profile using qPCR and their electrophysiological maturation. Once certain of obtained hindbrain mature neurons, one can start investigating the disease lines using this protocol and search for molecular and cellular events of neurodegeneration in spinocerebellar ataxias in the

248

right neuronal subtype, or at the least in a proxy hindbrain neuronal cell, using hiPSC technology.

249

7 Conclusions and future directions

This PhD split between genetics and stem cells has taken forward clinical data from patients and combined it with moleculargenetic investigations and explorative cellular functional models. It has combined a diverse range of clinical, genetic, bioinformatic and functional approaches with the underlying aim to improve clinicogenetic diagnosis, explore the role of different genetic variants and to provide mechanistic knowledge about the pathogenesis for exemplary genetic defects. It has investigated the role of variants in genes associated with neurodegeneration in a previously unexplored cohort of NBIA-like disorders in order to establish genetic diagnoses in families, help genetic counselling and refine geno- and phenotypes. Furthermore, it has discovered novel mutations and novel genotype-phenotype descriptions in the fields of hereditary movement disorders. Additionally, genetic variation influencing age at onset in trinucleotide repeat disorders and the absence of CACNA1A-repeat interruptions in SCA6 was discovered. Finally, a human neuronal model for spinocerebellar ataxia type 15 was generated. The limitations of currently available cerebellar differentiation protocols were realised and a novel protocol was devised with further ongoing work needed. By investigating neurodegeneration with a mix of genetic and stem cell technology, this PhD has answered some questions; it has however more substantially raised novel ones. A section describing the limitations as well as interesting and necessary future work for each of the results chapters discussed above can be found at the end of the discussion sections of the respective chapters. Of all techniques employed and fields touched in this PhD, the as of now least successful and potentially most difficult area – iPSC disease modelling for cerebellar ataxias – has been perhaps the most fascinating , and therefore anticipated future work in this region merits explicit mention at this point:

Future postdoc work: Unravelling the pathogenesis of cerebellar ataxias using induced pluripotent stem cells The work on human induced pluripotent stem cells derived from patients suffering from hereditary ataxias should continue to lead to investigations whether and how these cells can be harnessed in order to understand pathophysiology and develop therapeutics. Towards these aims, over the last year during their visit in the outpatients’ clinics several ataxia patients with defined mutations have been biopsied. Their fibroblasts already undergo biobanking and subsequent reprogramming to iPSCs and iPSC validation in a joint project with Sanger Institute at Cambridge where Prof. Wood, 250

Prof. Houlden, Dr. Bettencourt and myself recently won a grant to participate in the HipSci-project for rare diseases. This biobank allows researchers world-wide to request the iPSCs of any cerebellar ataxia they would like to study and where diagnosed patients have been biopsied. These efforts will greatly improve the poor translational yield of molecular research in cerebellar ataxias. However, compared to other neurodegenerative diseases, cerebellar ataxias have not yet taken full advantage of the breakthrough-discovery and usage of human induced pluripotent stem cells (251). This is in high proportion due to the poor differentiation efficiency, the intricate and complex needs of cerebellar cells, and the complicated nature of cerebellar differentiation protocols available.

A novel, and more effective adherent-culture based differentiation protocol for cerebellar derivatives should be developed. The important preliminary experiments for this have already been carried out (see Chapter 6.2). Once fully established, this protocol will be used to generate cerebellar derivatives of different spinocerebellar ataxia subtypes with a main focus on spinocerebellar ataxia type 15 where the iPSC lines and a cortical differentiation platform with preliminary results have been established (see Chapter 6.1). Please see Figure 7-1 for a graphical abstract of ongoing and future efforts. When comparing with cortical neurons (see Chapter 6.1) it might become possible to screen for the molecular events underlying cell-specific vulnerability (cerebellar versus cortical cells) and identify druggable pathways underlying neurodegeneration in spinocerebellar ataxias.

251

Figure 7-1: Future work iPSCs (SCA15 and cerebellar differentiation)

SCA15 iPSCs and differentiated counterparts will continue to undergo cellular phenotyping. Once the cerebellar differentiation protocol is established, various inherited cerebellar ataxias can be studied using the established platforms and protocols.

252

Final summary of individual result chapters This thesis has investigated the pathogenesis and genetic basis of rare hereditary movement disorders using a combination of induced pluripotent stem cell technology and classical genetics studies on DNA samples from cohorts of clinically characterised idiopathic and familial index patients, or small families with a focus on generating and analysing data from whole exome sequencing, Sanger sequencing and genotyping approaches.

The main findings from this work are summarised below.

Chapter 3: Clinicogenetic heterogeneity in NBIA

Here, the importance of thorough clinical investigation and the necessity of complete clinical and genetic data collection (all members of the affected families regardless of their affection status together with further biomaterial where possible) have been demonstrated. The unavailability of DNA or further biomaterial as well as clinical data in some of the families has rendered the interpretation of the genetic findings alone uncertain. However, the highly likely pathogenic mutation was identified in ~8.8% of patients from the NBIA cohort and potential pathogenic variants were found in further 23.1% of cases. With reported clinical and MRI findings classical confirmation of typical presentations (e.g. two neuroferritinopathy cases with classical disease and the common c.460dupA duplication in FTL) as well as potential extensions to the so far known clinicogenetic spectrum of the respective disease associated genes (e.g. PLA2G6, L1CAM, NIPA1, etc.) were presented. Further work would be required to prove pathogenicity in unclear cases, and where DNA from family members and further biomaterial from the patients was unavailable at least overexpression studies in cell or animal models would need to be performed to study functional consequences. Further variants in known and novel disease associated genes remain to be identified in the cohort. Given the limitations of the chosen approach of whole exome sequencing (reduced coverage for some areas of the exome increase the risk of missing disease causing variants; insensitivity of this technique towards larger insertions, deletions, pathogenic repeats and copy number variants present), the idiopathic cases of this cohort could be examined with CGH arrays or genotyping methods to rule out disease causing copy number variants. Finally, the exome data will be mined with novel annotation and bioinformatic tools as they are being developed (protein protein interactors, expression databases and more) to generate further hypotheses and

253

identify potential novel disease causing genes. The remaining unresolved cases might be submitted to whole genome sequencing projects.

Chapter 4: Genotypic spectrum, novel mutations and phenotype-genotype correlations in neuroacanthocytosis and hereditary ataxias

The work presented here widens the spectrum of clinical features associated with novel frameshift, truncating or missense mutations in XK-, SYNE1- and PNPLA6- associated disease. Firstly, a novel hemizygous frameshift deletion in exon 1 of XK (c.229delC, p.Leu80fs) was identified, leading to a premature stop codon in a male Greek individual. His choreatic movements of tongue and face and further neurological symptoms (peripheral neuropathy, lower limb muscle weakness) however were not the apparent symptoms that guided diagnosis from the start. His movement disorder together with a factor IX-deficiency and weak Kell antigens was detected on routine blood checks prior to a scheduled operation only, and caused the referral to the neurology department where thorough work up was then initiated. This case reveals the importance of knowledge of these rare disorders presenting with abnormal movement and haematological aberrations. It highlights the relevance of thorough neurologic, genetic and multisystem investigations upon detection of weak Kell antigens. Secondly, the ethnic, phenotypic and genetic diversity of SYNE1-associated cerebellar ataxia was expanded by investigating the NHNN-cohort of recessive and sporadic ataxia cases. Four novel truncating mutations in SYNE1 were identified for the first time in pedigrees from British and Sri Lankan origin and an additional Turkish case with pyramidal signs indicating motor neuron involvement was identified. Since, further studies have confirmed SYNE1 as a not uncommon cause of cerebellar ataxia associated with a range of non-cerebellar manifestations. Here presented data therefore confirm SYNE1 mutations are not an uncommon cause of recessive ataxia with or without additional clinical features in patients from various ethnicities.

Thirdly, novel homozygous variants in PNPLA6 were identified in a large multigenerational Parsi kindred with pure cerebellar ataxia for the first time. The most likely pathogenic identified homozygous variant c.3929A>T (p.D1310V) is located close to the mutational cluster in the catalytically active site of PNPLA6’s phospholipid esterase domain. It was absent from all control databases and showed high conservation and pathogenic in silico prediction scores. Even though the inaccessibility of further biomaterial rendered final confirmation of the putative pathogenic role of the

254

identified homozygous variant impossible, those findings suggest an extension of the clinical spectrum associated with PNPLA6 so far towards pure ataxias. It encourages genetic testing for PNPLA6 in cerebellar ataxias and spastic paraplegias with and without endocrine and ophthalmological features. In depth studies of the identified novel mutations could foster identification of factors that are responsible for the pleomorphic and extended clinical characteristics in patients carrying XK, SYNE1 and PNPLA6 mutations.

Chapter 5: DNA handling and repair pathways and interruption of repeat expansions potentially modulating age at onset in polyglutamine diseases

In this part of the PhD genetic modifiers for the clinically heterogeneous polyglutamine disorders were successfully identified. The identification of genetic modifiers influencing clinically heterogeneous Mendelian disorders is of paramount importance. It may unravel the underlying mechanisms modifying pathogenicity and phenotypic heterogeneity of known disease causing genes as well as point towards novel avenues for therapeutic intervention. Previously, a large genome-wide association study identified DNA handling and repair pathways to be associated with the age at onset in Huntington’s disease, however the implications of DNA repair variants in the other polyglutamine diseases were unclear. The genotyping study including samples of the important polyglutamine disorders (spinocerebellar ataxias 1, 2, 3, 6, 7, and 17, and DRPLA additional to HD) replicated findings from the GeM-HD GWAS for the first time in an independent cohort. It additionally suggests the universal and shared importance of variants in DNA repair and handling genes influencing age at onset in all polyglutamine diseases. In a second step, the occurrence of repeat interruptions in CACNA1A, the gene showing the shortest pathogenic repeat size (>19) with only limited repeat expansion observed and causal for spinocerebellar ataxia type 6, was investigated. No repeat stabilising interruptions could be identified in the NHNN SCA6 cohort. This suggests the results of the genotyping study not to be additionally influenced by masked repeat interruptions in SCA6.

Chapter 6.1: Creating the first human neuronal model for spinocerebellar ataxia type 15

Heterozygous ITPR1 deletions are responsible for spinocerebellar ataxia type 15, the most frequent non-trinucleotide repeat SCA in Caucasians. Despite the fact that 9

255

years have passed since its identification, the underlying mechanism leading to cerebellar degeneration with sporadically reported cortical features in humans remains uncertain. This work undertakes the challenge to create the first human neuronal model of spinocerebellar ataxia type 15 by generation of patient-derived iPSC lines, successful differentiation into cortical neurons and preliminary phenotypic study to reveal disease initiating molecular events. Even though in analyses presented here, no underlying and potentially treatable calcium dyshomeostasis or any other evident cell phenotype could be identified yet, and further studies are ongoing. These analyse the transcriptional signature of the cortical model, and secondly try to establish a cerebellar precursor cell differentiation protocol for further phenotypic study in non- cortical/cerebellar cells.

Chapter 6.2: Investigating the influence of WNT agonism during neural induction and neural patterning: Towards the generation of a non-cortical disease model

This part of the thesis describes the attempts to generate an efficient cell differentiation protocol for cerebellar disease. The downregulation of forebrain markers by usage of WNT agonism during the induction and patterning phase could be demonstrated, and transcriptional signatures could further be enriched for midbrain markers by FGF8 addition during the patterning phase. These results will need further confirmation and cells will need to be terminally differentiated and stained for hindbrain markers such as GBX2, or cerebellar cell subtype markers such as MATH1 and Calbindin to evaluate their usefulness.

In summary, this thesis has hereby been able to develop an overview of current themes and techniques in molecular neurodegeneration and contribute with novel insights to the existing literature. In the main projects the two young techniques of next-generation sequencing and iPSC culture were employed to a) characterise a NHNN cohort of NBIA-like disorders as well as further families and patient cohorts with rare inherited syndromes clinically and genetically (Chapter 3, 4 and 5) and b) functionally characterise the cellular pathology leading to neurodegeneration in SCA15 (Chapter 6.1) and the cellular events leading to a shift in differentiation capacity of iPSCs towards cerebellar derivatives (Chapter 6.2).

The projects described here have yielded important insights, such as the association of DNA repair variants with AAO in trinucleotide disorders (Chapter 5), the clinicogenetic 256

heterogeneity underlying hereditary ataxias and NBIAs (Chapter 3 and 4) or the importance of the availability of the right cellular subtype in the field of iPSC modelling of neurodegeneration (Chapter 6). Furthermore, the thesis has generated stimulating tasks and questions for further PhD and postdoctoral research that are mentioned at the end of the discussion of each (sub) chapter above.

Within its discussed restraints this PhD has achieved several important insights and generated resources, and findings will continue to be generated from the initial ideas and concepts developed here.

257

References

1. Wiethoff S, Hersheson J, Bettencourt C, Wood NW, Houlden H. Heterogeneity in clinical features and disease severity in ataxia-associated SYNE1 mutations. J Neurol. 2016. 2. Bettencourt C, Moss DH, Flower M, Wiethoff S, Brice A, Goizet C, et al. DNA repair pathways underlie a common genetic mechanism modulating onset in polyglutamine diseases. Ann Neurol. 2016. 3. Lander ES, Linton LM, Birren B, Nusbaum C, Zody MC, Baldwin J, et al. Initial sequencing and analysis of the human genome. Nature. 2001;409(6822):860-921. 4. Venter JC, Adams MD, Myers EW, Li PW, Mural RJ, Sutton GG, et al. The sequence of the human genome. Science. 2001;291(5507):1304-51. 5. International Human Genome Sequencing C. Finishing the euchromatic sequence of the human genome. Nature. 2004;431(7011):931-45. 6. Schmutz J, Wheeler J, Grimwood J, Dickson M, Yang J, Caoile C, et al. Quality assessment of the human genome sequence. Nature. 2004;429(6990):365-8. 7. Consortium EP, Birney E, Stamatoyannopoulos JA, Dutta A, Guigo R, Gingeras TR, et al. Identification and analysis of functional elements in 1% of the human genome by the ENCODE pilot project. Nature. 2007;447(7146):799-816. 8. Kidd JM, Cooper GM, Donahue WF, Hayden HS, Sampas N, Graves T, et al. Mapping and sequencing of structural variation from eight human genomes. Nature. 2008;453(7191):56-64. 9. Consortium UK, Walter K, Min JL, Huang J, Crooks L, Memari Y, et al. The UK10K project identifies rare variants in health and disease. Nature. 2015;526(7571):82-90. 10. Genomes Project C, Auton A, Brooks LD, Durbin RM, Garrison EP, Kang HM, et al. A global reference for human genetic variation. Nature. 2015;526(7571):68-74. 11. International HapMap C. A haplotype map of the human genome. Nature. 2005;437(7063):1299- 320. 12. International HapMap C, Altshuler DM, Gibbs RA, Peltonen L, Altshuler DM, Gibbs RA, et al. Integrating common and rare genetic variation in diverse human populations. Nature. 2010;467(7311):52- 8. 13. Sudmant PH, Rausch T, Gardner EJ, Handsaker RE, Abyzov A, Huddleston J, et al. An integrated map of structural variation in 2,504 human genomes. Nature. 2015;526(7571):75-81. 14. Wetterstrand KA. DNA Sequencing Costs: Data from the NHGRI Genome Sequencing Program (GSP). Available at: wwwgenomegov/sequencingcosts/ Accessed [23022016]. 2012. 15. Nash DB. Personalized medicine: are we there yet? American health & drug benefits. 2014;7(7):371-2. 16. Gusella JF, Wexler NS, Conneally PM, Naylor SL, Anderson MA, Tanzi RE, et al. A polymorphic DNA marker genetically linked to Huntington's disease. Nature. 1983;306(5940):234-8. 17. Bates GP. History of genetic disease: the molecular genetics of Huntington disease - a history. Nature reviews Genetics. 2005;6(10):766-73. 18. The Huntington's Disease Collaborative Research Group. A novel gene containing a trinucleotide repeat that is expanded and unstable on Huntington's disease chromosomes. The Huntington's Disease Collaborative Research Group. Cell. 1993;72(6):971-83. 19. Rommens JM, Iannuzzi MC, Kerem B, Drumm ML, Melmer G, Dean M, et al. Identification of the cystic fibrosis gene: chromosome walking and jumping. Science. 1989;245(4922):1059-65. 20. Hsiao K, Baker HF, Crow TJ, Poulter M, Owen F, Terwilliger JD, et al. Linkage of a prion protein missense variant to Gerstmann-Straussler syndrome. Nature. 1989;338(6213):342-5. 21. Goate A, Chartier-Harlin MC, Mullan M, Brown J, Crawford F, Fidani L, et al. Segregation of a missense mutation in the amyloid precursor protein gene with familial Alzheimer's disease. Nature. 1991;349(6311):704-6. 22. Orr HT, Chung MY, Banfi S, Kwiatkowski TJ, Jr., Servadio A, Beaudet AL, et al. Expansion of an unstable trinucleotide CAG repeat in spinocerebellar ataxia type 1. Nat Genet. 1993;4(3):221-6. 23. Sanger F, Nicklen S, Coulson AR. DNA sequencing with chain-terminating inhibitors. Proceedings of the National Academy of Sciences of the United States of America. 1977;74(12):5463-7. 24. Jacobi H, Bauer P, Giunti P, Labrum R, Sweeney MG, Charles P, et al. The natural history of spinocerebellar ataxia type 1, 2, 3, and 6: a 2-year follow-up study. Neurology. 2011;77(11):1035-41. 25. Jacobi H, du Montcel ST, Bauer P, Giunti P, Cook A, Labrum R, et al. Long-term disease progression in spinocerebellar ataxia types 1, 2, 3, and 6: a longitudinal cohort study. Lancet Neurol. 2015;14(11):1101-8. 26. Schule R, Wiethoff S, Martus P, Karle KN, Otto S, Klebe S, et al. Hereditary Spastic Paraplegia -clinico-genetic lessons from 608 patients. Ann Neurol. 2016. 27. Harding AE. Classification of the hereditary ataxias and paraplegias. Lancet. 1983;1(8334):1151-5.

258

28. Novarino G, Fenstermaker AG, Zaki MS, Hofree M, Silhavy JL, Heiberg AD, et al. Exome sequencing links corticospinal motor neuron disease to common neurodegenerative disorders. Science. 2014;343(6170):506-11. 29. Bosch AM, Abeling NG, Ijlst L, Knoester H, van der Pol WL, Stroomer AE, et al. Brown- Vialetto-Van Laere and Fazio Londe syndrome is associated with a riboflavin transporter defect mimicking mild MADD: a new inborn error of metabolism with potential treatment. Journal of inherited metabolic disease. 2011;34(1):159-64. 30. Shendure J, Ji H. Next-generation DNA sequencing. Nature biotechnology. 2008;26(10):1135- 45. 31. Corder EH, Saunders AM, Strittmatter WJ, Schmechel DE, Gaskell PC, Small GW, et al. Gene dose of apolipoprotein E type 4 allele and the risk of Alzheimer's disease in late onset families. Science. 1993;261(5123):921-3. 32. Strittmatter WJ, Saunders AM, Schmechel D, Pericak-Vance M, Enghild J, Salvesen GS, et al. Apolipoprotein E: high-avidity binding to beta-amyloid and increased frequency of type 4 allele in late- onset familial Alzheimer disease. Proceedings of the National Academy of Sciences of the United States of America. 1993;90(5):1977-81. 33. Sekar A, Bialas AR, de Rivera H, Davis A, Hammond TR, Kamitaki N, et al. Schizophrenia risk from complex variation of complement component 4. Nature. 2016;530(7589):177-83. 34. Hirschhorn JN, Daly MJ. Genome-wide association studies for common diseases and complex traits. Nature reviews Genetics. 2005;6(2):95-108. 35. Cruchaga C, Haller G, Chakraverty S, Mayo K, Vallania FL, Mitra RD, et al. Rare variants in APP, PSEN1 and PSEN2 increase risk for AD in late-onset Alzheimer's disease families. PloS one. 2012;7(2):e31039. 36. Spataro N, Calafell F, Cervera-Carles L, Casals F, Pagonabarraga J, Pascual-Sedano B, et al. Mendelian genes for Parkinson's disease contribute to the sporadic forms of the disease. Hum Mol Genet. 2015;24(7):2023-34. 37. Manolio TA, Collins FS, Cox NJ, Goldstein DB, Hindorff LA, Hunter DJ, et al. Finding the missing heritability of complex diseases. Nature. 2009;461(7265):747-53. 38. Bansal V, Libiger O, Torkamani A, Schork NJ. Statistical analysis strategies for association studies involving rare variants. Nature reviews Genetics. 2010;11(11):773-85. 39. Sham PC, Purcell SM. Statistical power and significance testing in large-scale genetic studies. Nature reviews Genetics. 2014;15(5):335-46. 40. Wei WH, Hemani G, Haley CS. Detecting epistasis in human complex traits. Nature reviews Genetics. 2014;15(11):722-33. 41. Ebbert MT, Boehme KL, Wadsworth ME, Staley LA, Alzheimer's Disease Neuroimaging I, Alzheimer's Disease Genetics C, et al. Interaction between variants in CLU and MS4A4E modulates Alzheimer's disease risk. Alzheimer's & dementia : the journal of the Alzheimer's Association. 2016;12(2):121-9. 42. Elbaz A, Ross OA, Ioannidis JP, Soto-Ortolaza AI, Moisan F, Aasly J, et al. Independent and joint effects of the MAPT and SNCA genes in Parkinson disease. Ann Neurol. 2011;69(5):778-92. 43. Gusareva ES, Carrasquillo MM, Bellenguez C, Cuyvers E, Colon S, Graff-Radford NR, et al. Genome-wide association interaction analysis for Alzheimer's disease. Neurobiol Aging. 2014;35(11):2436-43. 44. Margolin DH, Kousi M, Chan YM, Lim ET, Schmahmann JD, Hadjivassiliou M, et al. Ataxia, dementia, and hypogonadotropism caused by disordered ubiquitination. N Engl J Med. 2013;368(21):1992-2003. 45. Wiethoff S, Bhatia, KP., Houlden H. Genetics of NBIA Disorders. . In: Bras J, Schneider, SA., editor. Movement Disorder Genetics. p. 263-291: Springer International Publishing; 2015. 46. Gregory A, Hayflick SJ. Genetics of neurodegeneration with brain iron accumulation. Curr Neurol Neurosci Rep. 2011;11(3):254-61. 47. Schneider SA, Bhatia KP. Syndromes of neurodegeneration with brain iron accumulation. Seminars in pediatric neurology. 2012;19(2):57-66. 48. Aggarwal A, Schneider SA, Houlden H, Silverdale M, Paudel R, Paisan-Ruiz C, et al. Indian- subcontinent NBIA: unusual phenotypes, novel PANK2 mutations, and undetermined genetic forms. Mov Disord. 2010;25(10):1424-31. 49. Hayflick SJ, Hartman M, Coryell J, Gitschier J, Rowley H. Brain MRI in neurodegeneration with brain iron accumulation with and without PANK2 mutations. AJNR American journal of neuroradiology. 2006;27(6):1230-3. 50. Paisan-Ruiz C, Bhatia KP, Li A, Hernandez D, Davis M, Wood NW, et al. Characterization of PLA2G6 as a locus for dystonia-parkinsonism. Ann Neurol. 2009;65(1):19-23. 51. Yoshino H, Tomiyama H, Tachibana N, Ogaki K, Li Y, Funayama M, et al. Phenotypic spectrum of patients with PLA2G6 mutation and PARK14-linked parkinsonism. Neurology. 2010;75(15):1356-61. 259

52. Kruer MC, Hiken M, Gregory A, Malandrini A, Clark D, Hogarth P, et al. Novel histopathologic findings in molecularly-confirmed pantothenate kinase-associated neurodegeneration. Brain. 2011;134(Pt 4):947-58. 53. McNeill A, Birchall D, Hayflick SJ, Gregory A, Schenk JF, Zimmerman EA, et al. T2* and FSE MRI distinguishes four subtypes of neurodegeneration with brain iron accumulation. Neurology. 2008;70(18):1614-9. 54. Stamelou M, Lai SC, Aggarwal A, Schneider SA, Houlden H, Yeh TH, et al. Dystonic opisthotonus: a "red flag" for neurodegeneration with brain iron accumulation syndromes? Mov Disord. 2013;28(10):1325-9. 55. Kruer MC, Boddaert N. Neurodegeneration with brain iron accumulation: a diagnostic algorithm. Seminars in pediatric neurology. 2012;19(2):67-74. 56. Amaral LL, Gaddikeri S, Chapman PR, Roy R, Gaddikeri RS, Marussi VH, et al. Neurodegeneration with Brain Iron Accumulation: Clinicoradiological Approach to Diagnosis. Journal of neuroimaging : official journal of the American Society of Neuroimaging. 2015;25(4):539-51. 57. Kruer MC, Boddaert N, Schneider SA, Houlden H, Bhatia KP, Gregory A, et al. Neuroimaging features of neurodegeneration with brain iron accumulation. AJNR American journal of neuroradiology. 2012;33(3):407-14. 58. Hogarth P. Neurodegeneration with brain iron accumulation: diagnosis and management. Journal of movement disorders. 2015;8(1):1-13. 59. Defendini R, Markesbery WR, Mastri AR, Duffy PE. Hallervorden-Spatz disease and infantile neuroaxonal dystrophy. Ultrastructural observations, anatomical pathology and nosology. J Neurol Sci. 1973;20(1):7-23. 60. Savoiardo M, Halliday WC, Nardocci N, Strada L, D'Incerti L, Angelini L, et al. Hallervorden- Spatz disease: MR and pathologic findings. AJNR American journal of neuroradiology. 1993;14(1):155- 62. 61. Li A, Wiethoff, S., Arber, C., Houlden, H., Revesz, T., Holton, J.L. Pathology and Genetics of Neuroaxonal Dystrophy/ Neurodegeneration with Brain Iron Accumulation. Review Article in Developmental Neuropathology, 2nd Edition, ISN Book Series. (in press).2016. 62. Wakabayashi K, Yoshimoto M, Fukushima T, Koide R, Horikawa Y, Morita T, et al. Widespread occurrence of alpha-synuclein/NACP-immunoreactive neuronal inclusions in juvenile and adult-onset Hallervorden-Spatz disease with Lewy bodies. Neuropathology and applied neurobiology. 1999;25(5):363-8. 63. Gregory A, Westaway SK, Holm IE, Kotzbauer PT, Hogarth P, Sonek S, et al. Neurodegeneration associated with genetic defects in phospholipase A(2). Neurology. 2008;71(18):1402- 9. 64. Haraguchi T, Terada S, Ishizu H, Yokota O, Yoshida H, Takeda N, et al. Coexistence of TDP-43 and tau pathology in neurodegeneration with brain iron accumulation type 1 (NBIA-1, formerly Hallervorden-Spatz syndrome). Neuropathology : official journal of the Japanese Society of Neuropathology. 2011;31(5):531-9. 65. Li A, Paudel R, Johnson R, Courtney R, Lees AJ, Holton JL, et al. Pantothenate kinase- associated neurodegeneration is not a synucleinopathy. Neuropathology and applied neurobiology. 2012. 66. Eidelberg D, Sotrel A, Joachim C, Selkoe D, Forman A, Pendlebury WW, et al. Adult onset Hallervorden-Spatz disease with neurofibrillary pathology. A discrete clinicopathological entity. Brain. 1987;110 ( Pt 4):993-1013. 67. Hartmann HA, White SK, Levine RL. Neuroaxonal dystrophy with neuromelanin deposition, neurofibrillary tangles, and neuronal loss. Light- and electron-microscopic changes in a 45-year-old woman with progressive psychomotor deterioration. Acta Neuropathol. 1983;61(3-4):169-72. 68. Keogh MJ, Chinnery PF. Current concepts and controversies in neurodegeneration with brain iron accumulation. Seminars in pediatric neurology. 2012;19(2):51-6. 69. Miyajima H, Nishimura Y, Mizoguchi K, Sakamoto M, Shimizu T, Honda N. Familial apoceruloplasmin deficiency associated with blepharospasm and retinal degeneration. Neurology. 1987;37(5):761-7. 70. McNeill A, Pandolfo M, Kuhn J, Shang H, Miyajima H. The neurological presentation of ceruloplasmin gene mutations. European neurology. 2008;60(4):200-5. 71. Curtis AR, Fey C, Morris CM, Bindoff LA, Ince PG, Chinnery PF, et al. Mutation in the gene encoding ferritin light polypeptide causes dominant adult-onset basal ganglia disease. Nat Genet. 2001;28(4):350-4. 72. Ohta E, Nagasaka T, Shindo K, Toma S, Nagasaka K, Ohta K, et al. Neuroferritinopathy in a Japanese family with a duplication in the ferritin light chain gene. Neurology. 2008;70(16 Pt 2):1493-4. 73. Ory-Magne F, Brefel-Courbon C, Payoux P, Debruxelles S, Sibon I, Goizet C, et al. Clinical phenotype and neuroimaging findings in a French family with hereditary ferritinopathy (FTL498- 499InsTC). Mov Disord. 2009;24(11):1676-83.

260

74. Vidal R, Ghetti B, Takao M, Brefel-Courbon C, Uro-Coste E, Glazier BS, et al. Intracellular ferritin accumulation in neural and extraneural tissue characterizes a neurodegenerative disease associated with a mutation in the ferritin light polypeptide gene. Journal of neuropathology and experimental neurology. 2004;63(4):363-80. 75. Khateeb S, Flusser H, Ofir R, Shelef I, Narkis G, Vardi G, et al. PLA2G6 mutation underlies infantile neuroaxonal dystrophy. Am J Hum Genet. 2006;79(5):942-8. 76. Haack TB, Hogarth P, Kruer MC, Gregory A, Wieland T, Schwarzmayr T, et al. Exome sequencing reveals de novo WDR45 mutations causing a phenotypically distinct, X-linked dominant form of NBIA. Am J Hum Genet. 2012;91(6):1144-9. 77. Ramirez A, Heimbach A, Grundemann J, Stiller B, Hampshire D, Cid LP, et al. Hereditary parkinsonism with dementia is caused by mutations in ATP13A2, encoding a lysosomal type 5 P-type ATPase. Nat Genet. 2006;38(10):1184-91. 78. Hayflick SJ, Westaway SK, Levinson B, Zhou B, Johnson MA, Ching KH, et al. Genetic, clinical, and radiographic delineation of Hallervorden-Spatz syndrome. N Engl J Med. 2003;348(1):33- 40. 79. Zeidman LA, Pandey DK. Declining use of the Hallervorden-Spatz disease eponym in the last two decades. Journal of child neurology. 2012;27(10):1310-5. 80. Hayflick SJ. Neurodegeneration with brain iron accumulation: from genes to pathogenesis. Seminars in pediatric neurology. 2006;13(3):182-5. 81. Schneider SA, Aggarwal A, Bhatt M, Dupont E, Tisch S, Limousin P, et al. Severe tongue protrusion dystonia: clinical syndromes and possible treatment. Neurology. 2006;67(6):940-3. 82. Marelli C, Piacentini S, Garavaglia B, Girotti F, Albanese A. Clinical and neuropsychological correlates in two brothers with pantothenate kinase-associated neurodegeneration. Mov Disord. 2005;20(2):208-12. 83. Egan RA, Weleber RG, Hogarth P, Gregory A, Coryell J, Westaway SK, et al. Neuro- ophthalmologic and electroretinographic findings in pantothenate kinase-associated neurodegeneration (formerly Hallervorden-Spatz syndrome). Am J Ophthalmol. 2005;140(2):267-74. 84. Chang CL, Lin CM. Eye-of-the-Tiger sign is not Pathognomonic of Pantothenate Kinase- Associated Neurodegeneration in Adult Cases. Brain and behavior. 2011;1(1):55-6. 85. Strecker K, Hesse S, Wegner F, Sabri O, Schwarz J, Schneider JP. Eye of the tiger sign in multiple system atrophy. Eur J Neurol. 2007;14(11):e1-2. 86. Delgado RF, Sanchez PR, Speckter H, Then EP, Jimenez R, Oviedo J, et al. Missense PANK2 mutation without "eye of the tiger" sign: MR findings in a large group of patients with pantothenate kinase-associated neurodegeneration (PKAN). Journal of magnetic resonance imaging : JMRI. 2012;35(4):788-94. 87. van den Bogaard SJ, Kruit MC, Dumas EM, Roos RA. Eye-of-the-tiger-sign in a 48 year healthy adult. J Neurol Sci. 2014;336(1-2):254-6. 88. Saito Y, Kawai M, Inoue K, Sasaki R, Arai H, Nanba E, et al. Widespread expression of alpha- synuclein and tau immunoreactivity in Hallervorden-Spatz syndrome with protracted clinical course. J Neurol Sci. 2000;177(1):48-59. 89. Neumann M, Adler S, Schluter O, Kremmer E, Benecke R, Kretzschmar HA. Alpha-synuclein accumulation in a case of neurodegeneration with brain iron accumulation type 1 (NBIA-1, formerly Hallervorden-Spatz syndrome) with widespread cortical and brainstem-type Lewy bodies. Acta Neuropathol. 2000;100(5):568-74. 90. Galvin JE, Giasson B, Hurtig HI, Lee VM, Trojanowski JQ. Neurodegeneration with brain iron accumulation, type 1 is characterized by alpha-, beta-, and gamma-synuclein neuropathology. The American journal of pathology. 2000;157(2):361-8. 91. Hortnagel K, Prokisch H, Meitinger T. An isoform of hPANK2, deficient in pantothenate kinase- associated neurodegeneration, localizes to mitochondria. Hum Mol Genet. 2003;12(3):321-7. 92. Kotzbauer PT, Truax AC, Trojanowski JQ, Lee VM. Altered neuronal mitochondrial coenzyme A synthesis in neurodegeneration with brain iron accumulation caused by abnormal processing, stability, and catalytic activity of mutant pantothenate kinase 2. J Neurosci. 2005;25(3):689-98. 93. Kuo YM, Duncan JL, Westaway SK, Yang H, Nune G, Xu EY, et al. Deficiency of pantothenate kinase 2 (Pank2) in mice leads to retinal degeneration and azoospermia. Hum Mol Genet. 2005;14(1):49- 57. 94. Brunetti D, Dusi S, Morbin M, Uggetti A, Moda F, D'Amato I, et al. Pantothenate kinase- associated neurodegeneration: altered mitochondria membrane potential and defective respiration in Pank2 knock-out mouse model. Hum Mol Genet. 2012;21(24):5294-305. 95. Kuo YM, Hayflick SJ, Gitschier J. Deprivation of pantothenic acid elicits a movement disorder and azoospermia in a mouse model of pantothenate kinase-associated neurodegeneration. Journal of inherited metabolic disease. 2007;30(3):310-7.

261

96. Brunetti D, Dusi S, Giordano C, Lamperti C, Morbin M, Fugnanesi V, et al. Pantethine treatment is effective in recovering the disease phenotype induced by ketogenic diet in a pantothenate kinase- associated neurodegeneration mouse model. Brain. 2014;137(Pt 1):57-68. 97. Bosveld F, Rana A, van der Wouden PE, Lemstra W, Ritsema M, Kampinga HH, et al. De novo CoA biosynthesis is required to maintain DNA integrity during development of the Drosophila nervous system. Hum Mol Genet. 2008;17(13):2058-69. 98. Wu Z, Li C, Lv S, Zhou B. Pantothenate kinase-associated neurodegeneration: insights from a Drosophila model. Hum Mol Genet. 2009;18(19):3659-72. 99. Yang Y, Wu Z, Kuo YM, Zhou B. Dietary rescue of fumble--a Drosophila model for pantothenate-kinase-associated neurodegeneration. Journal of inherited metabolic disease. 2005;28(6):1055-64. 100. Zizioli D, Tiso N, Guglielmi A, Saraceno C, Busolin G, Giuliani R, et al. Knock-down of pantothenate kinase 2 severely affects the development of the nervous and vascular system in zebrafish, providing new insights into PKAN disease. Neurobiology of disease. 2016;85:35-48. 101. Santambrogio P, Dusi S, Guaraldo M, Rotundo LI, Broccoli V, Garavaglia B, et al. Mitochondrial iron and energetic dysfunction distinguish fibroblasts and induced neurons from pantothenate kinase-associated neurodegeneration patients. Neurobiology of disease. 2015;81:144-53. 102. Hartig MB, Hortnagel K, Garavaglia B, Zorzi G, Kmiec T, Klopstock T, et al. Genotypic and phenotypic spectrum of PANK2 mutations in patients with neurodegeneration with brain iron accumulation. Ann Neurol. 2006;59(2):248-56. 103. Gregory A, Polster BJ, Hayflick SJ. Clinical and genetic delineation of neurodegeneration with brain iron accumulation. J Med Genet. 2009;46(2):73-80. 104. Gregory A, Hayflick SJ. Pantothenate Kinase-Associated Neurodegeneration. In: Pagon RA, Adam MP, Ardinger HH, Bird TD, Dolan CR, Fong CT, et al., editors. GeneReviews(R). Seattle (WA)1993. 105. Hong BS, Senisterra G, Rabeh WM, Vedadi M, Leonardi R, Zhang YM, et al. Crystal structures of human pantothenate kinases. Insights into allosteric regulation and mutations linked to a neurodegeneration disorder. J Biol Chem. 2007;282(38):27984-93. 106. Houlden H, Lincoln S, Farrer M, Cleland PG, Hardy J, Orrell RW. Compound heterozygous PANK2 mutations confirm HARP and Hallervorden-Spatz syndromes are allelic. Neurology. 2003;61(10):1423-6. 107. Zorzi G, Zibordi F, Chiapparini L, Bertini E, Russo L, Piga A, et al. Iron-related MRI images in patients with pantothenate kinase-associated neurodegeneration (PKAN) treated with deferiprone: results of a phase II pilot trial. Mov Disord. 2011;26(9):1756-9. 108. Cossu G, Abbruzzese G, Matta G, Murgia D, Melis M, Ricchi V, et al. Efficacy and safety of deferiprone for the treatment of pantothenate kinase-associated neurodegeneration (PKAN) and neurodegeneration with brain iron accumulation (NBIA): results from a four years follow-up. Parkinsonism & related disorders. 2014;20(6):651-4. 109. Timmermann L, Pauls KA, Wieland K, Jech R, Kurlemann G, Sharma N, et al. Dystonia in neurodegeneration with brain iron accumulation: outcome of bilateral pallidal stimulation. Brain. 2010;133(Pt 3):701-12. 110. Mikati MA, Yehya A, Darwish H, Karam P, Comair Y. Deep brain stimulation as a mode of treatment of early onset pantothenate kinase-associated neurodegeneration. European journal of paediatric neurology : EJPN : official journal of the European Paediatric Neurology Society. 2009;13(1):61-4. 111. Gregory A, Hayflick S. Neurodegeneration with Brain Iron Accumulation Disorders Overview. In: Pagon RA, Adam MP, Ardinger HH, Bird TD, Dolan CR, Fong CT, et al., editors. GeneReviews(R). Seattle (WA)1993. 112. Morgan NV, Westaway SK, Morton JE, Gregory A, Gissen P, Sonek S, et al. PLA2G6, encoding a phospholipase A2, is mutated in neurodegenerative disorders with high brain iron. Nat Genet. 2006;38(7):752-4. 113. Kurian MA, Morgan NV, MacPherson L, Foster K, Peake D, Gupta R, et al. Phenotypic spectrum of neurodegeneration associated with mutations in the PLA2G6 gene (PLAN). Neurology. 2008;70(18):1623-9. 114. Aicardi J, Castelein P. Infantile neuroaxonal dystrophy. Brain. 1979;102(4):727-48. 115. Nardocci N, Zorzi G, Farina L, Binelli S, Scaioli W, Ciano C, et al. Infantile neuroaxonal dystrophy: clinical spectrum and diagnostic criteria. Neurology. 1999;52(7):1472-8. 116. Gregory A, Kurian MA, Maher ER, Hogarth P, Hayflick SJ. PLA2G6-Associated Neurodegeneration. In: Pagon RA, Adam MP, Ardinger HH, Bird TD, Dolan CR, Fong CT, et al., editors. GeneReviews(R). Seattle (WA)1993. 117. Illingworth MA, Meyer E, Chong WK, Manzur AY, Carr LJ, Younis R, et al. PLA2G6- associated neurodegeneration (PLAN): Further expansion of the clinical, radiological and mutation spectrum associated with infantile and atypical childhood-onset disease. Molecular genetics and metabolism. 2014. 262

118. Kurian MA, McNeill A, Lin JP, Maher ER. Childhood disorders of neurodegeneration with brain iron accumulation (NBIA). Developmental medicine and child neurology. 2011;53(5):394-404. 119. Larsson PK, Claesson HE, Kennedy BP. Multiple splice variants of the human calcium- independent phospholipase A2 and their effect on enzyme activity. J Biol Chem. 1998;273(1):207-14. 120. Mubaidin A, Roberts E, Hampshire D, Dehyyat M, Shurbaji A, Mubaidien M, et al. Karak syndrome: a novel degenerative disorder of the basal ganglia and cerebellum. J Med Genet. 2003;40(7):543-6. 121. Shinzawa K, Sumi H, Ikawa M, Matsuoka Y, Okabe M, Sakoda S, et al. Neuroaxonal dystrophy caused by group VIA phospholipase A2 deficiency in mice: a model of human neurodegenerative disease. J Neurosci. 2008;28(9):2212-20. 122. Malik I, Turk J, Mancuso DJ, Montier L, Wohltmann M, Wozniak DF, et al. Disrupted membrane homeostasis and accumulation of ubiquitinated proteins in a mouse model of infantile neuroaxonal dystrophy caused by PLA2G6 mutations. The American journal of pathology. 2008;172(2):406-16. 123. Paisan-Ruiz C, Li A, Schneider SA, Holton JL, Johnson R, Kidd D, et al. Widespread Lewy body and tau accumulation in childhood and adult onset dystonia-parkinsonism cases with PLA2G6 mutations. Neurobiol Aging. 2012;33(4):814-23. 124. Crompton D, Rehal PK, MacPherson L, Foster K, Lunt P, Hughes I, et al. Multiplex ligation- dependent probe amplification (MLPA) analysis is an effective tool for the detection of novel intragenic PLA2G6 mutations: implications for molecular diagnosis. Molecular genetics and metabolism. 2010;100(2):207-12. 125. Engel LA, Jing Z, O'Brien DE, Sun M, Kotzbauer PT. Catalytic function of PLA2G6 is impaired by mutations associated with infantile neuroaxonal dystrophy but not dystonia-parkinsonism. PloS one. 2010;5(9):e12897. 126. Balsinde J, Balboa MA. Cellular regulation and proposed biological functions of group VIA calcium-independent phospholipase A2 in activated cells. Cellular signalling. 2005;17(9):1052-62. 127. Strokin M, Seburn KL, Cox GA, Martens KA, Reiser G. Severe disturbance in the Ca2+ signaling in astrocytes from mouse models of human infantile neuroaxonal dystrophy with mutated Pla2g6. Hum Mol Genet. 2012;21(12):2807-14. 128. Lands WE. Metabolism of glycerolipids. 2. The enzymatic acylation of lysolecithin. J Biol Chem. 1960;235:2233-7. 129. Perez R, Melero R, Balboa MA, Balsinde J. Role of group VIA calcium-independent phospholipase A2 in arachidonic acid release, phospholipid fatty acid incorporation, and apoptosis in U937 cells responding to hydrogen peroxide. J Biol Chem. 2004;279(39):40385-91. 130. Sumi-Akamaru H, Beck G, Kato S, Mochizuki H. Neuroaxonal dystrophy in PLA2G6 knockout mice. Neuropathology : official journal of the Japanese Society of Neuropathology. 2015;35(3):289-302. 131. Kurian MA, Hayflick SJ. Pantothenate kinase-associated neurodegeneration (PKAN) and PLA2G6-associated neurodegeneration (PLAN): review of two major neurodegeneration with brain iron accumulation (NBIA) phenotypes. International review of neurobiology. 2013;110:49-71. 132. Hartig MB, Iuso A, Haack T, Kmiec T, Jurkiewicz E, Heim K, et al. Absence of an orphan mitochondrial protein, c19orf12, causes a distinct clinical subtype of neurodegeneration with brain iron accumulation. Am J Hum Genet. 2011;89(4):543-50. 133. Hogarth P, Gregory A, Kruer MC, Sanford L, Wagoner W, Natowicz MR, et al. New NBIA subtype: genetic, clinical, pathologic, and radiographic features of MPAN. Neurology. 2013;80(3):268- 75. 134. Iuso A, Sibon OC, Gorza M, Heim K, Organisti C, Meitinger T, et al. Impairment of Drosophila orthologs of the human orphan protein C19orf12 induces bang sensitivity and neurodegeneration. PloS one. 2014;9(2):e89439. 135. Saitsu H, Nishimura T, Muramatsu K, Kodera H, Kumada S, Sugai K, et al. De novo mutations in the autophagy gene WDR45 cause static encephalopathy of childhood with neurodegeneration in adulthood. Nat Genet. 2013;45(4):445-9, 9e1. 136. Hayflick SJ, Kruer MC, Gregory A, Haack TB, Kurian MA, Houlden HH, et al. Beta-propeller protein-associated neurodegeneration: a new X-linked dominant disorder with brain iron accumulation. Brain. 2013;136(Pt 6):1708-17. 137. Paudel R, Li A, Wiethoff S, Bandopadhyay R, Bhatia K, de Silva R, et al. Neuropathology of Beta-propeller protein associated neurodegeneration (BPAN): a new tauopathy. Acta neuropathologica communications. 2015;3(1):39. 138. Takano K, Shiba N, Wakui K, Yamaguchi T, Aida N, Inaba Y, et al. Elevation of neuron specific enolase and brain iron deposition on susceptibility-weighted imaging as diagnostic clues for beta- propeller protein-associated neurodegeneration in early childhood: Additional case report and review of the literature. Am J Med Genet A. 2016;170(2):322-8.

263

139. Hoffjan S, Ibisler A, Tschentscher A, Dekomien G, Bidinost C, Rosa AL. WDR45 mutations in Rett (-like) syndrome and developmental delay: Case report and an appraisal of the literature. Molecular and cellular probes. 2016;30(1):44-9. 140. Nishioka K, Oyama G, Yoshino H, Li Y, Matsushima T, Takeuchi C, et al. High frequency of beta-propeller protein-associated neurodegeneration (BPAN) among patients with intellectual disability and young-onset parkinsonism. Neurobiol Aging. 2015;36(5):2004 e9- e15. 141. Van Goethem G, Livingston JH, Warren D, Oojageer AJ, Rice GI, Crow YJ. Basal ganglia calcification in a patient with beta-propeller protein-associated neurodegeneration. Pediatric neurology. 2014;51(6):843-5. 142. Xixis KI, Mikati MA. Epileptic spasms: a previously unreported manifestation of WDR45 gene mutation. Epileptic disorders : international epilepsy journal with videotape. 2015;17(4):467-72. 143. Yoganathan S, Arunachal G, Sudhakar SV, Rajaraman V, Thomas M, Danda S. Beta Propellar Protein-Associated Neurodegeneration: A Rare Cause of Infantile Autistic Regression and Intracranial Calcification. Neuropediatrics. 2016;47(2):123-7. 144. Lu Q, Yang P, Huang X, Hu W, Guo B, Wu F, et al. The WD40 repeat PtdIns(3)P-binding protein EPG-6 regulates progression of omegasomes to autophagosomes. Developmental cell. 2011;21(2):343-57. 145. Nakatogawa H, Suzuki K, Kamada Y, Ohsumi Y. Dynamics and diversity in autophagy mechanisms: lessons from yeast. Nature reviews Molecular cell biology. 2009;10(7):458-67. 146. Zhao YG, Sun L, Miao G, Ji C, Zhao H, Sun H, et al. The autophagy gene Wdr45/Wipi4 regulates learning and memory function and axonal homeostasis. Autophagy. 2015;11(6):881-90. 147. Dick KJ, Eckhardt M, Paisan-Ruiz C, Alshehhi AA, Proukakis C, Sibtain NA, et al. Mutation of FA2H underlies a complicated form of hereditary spastic paraplegia (SPG35). Hum Mutat. 2010;31(4):E1251-60. 148. Edvardson S, Hama H, Shaag A, Gomori JM, Berger I, Soffer D, et al. Mutations in the fatty acid 2-hydroxylase gene are associated with leukodystrophy with spastic paraparesis and dystonia. Am J Hum Genet. 2008;83(5):643-8. 149. Garone C, Pippucci T, Cordelli DM, Zuntini R, Castegnaro G, Marconi C, et al. FA2H-related disorders: a novel c.270+3A>T splice-site mutation leads to a complex neurodegenerative phenotype. Developmental medicine and child neurology. 2011;53(10):958-61. 150. Schneider SA, Bhatia KP. Three faces of the same gene: FA2H links neurodegeneration with brain iron accumulation, leukodystrophies, and hereditary spastic paraplegias. Ann Neurol. 2010;68(5):575-7. 151. Kruer MC, Paisan-Ruiz C, Boddaert N, Yoon MY, Hama H, Gregory A, et al. Defective FA2H leads to a novel form of neurodegeneration with brain iron accumulation (NBIA). Ann Neurol. 2010;68(5):611-8. 152. Stevanin G, Azzedine H, Denora P, Boukhris A, Tazir M, Lossos A, et al. Mutations in SPG11 are frequent in autosomal recessive spastic paraplegia with thin corpus callosum, cognitive decline and lower motor neuron degeneration. Brain. 2008;131(Pt 3):772-84. 153. Schule R, Schols L. Genetics of hereditary spastic paraplegias. Seminars in neurology. 2011;31(5):484-93. 154. Pierson TM, Simeonov DR, Sincan M, Adams DA, Markello T, Golas G, et al. Exome sequencing and SNP analysis detect novel compound heterozygosity in fatty acid hydroxylase-associated neurodegeneration. Eur J Hum Genet. 2012;20(4):476-9. 155. Alderson NL, Rembiesa BM, Walla MD, Bielawska A, Bielawski J, Hama H. The human FA2H gene encodes a fatty acid 2-hydroxylase. J Biol Chem. 2004;279(47):48562-8. 156. Zhu G, Koszelak-Rosenblum M, Connelly SM, Dumont ME, Malkowski MG. The Crystal Structure of an Integral Membrane Fatty Acid alpha-Hydroxylase. J Biol Chem. 2015;290(50):29820-33. 157. Liao X, Luo Y, Zhan Z, Du J, Hu Z, Wang J, et al. SPG35 contributes to the second common subtype of AR-HSP in China: frequency analysis and functional characterization of FA2H gene mutations. Clin Genet. 2013. 158. Kruer MC, Gregory A, Hayflick SJ. Fatty Acid Hydroxylase-Associated Neurodegeneration. In: Pagon RA, Adam MP, Ardinger HH, Bird TD, Dolan CR, Fong CT, et al., editors. GeneReviews(R). Seattle (WA)1993. 159. Aguirre-Rodriguez FJ, Lucenilla MI, Alvarez-Cubero MJ, Mata C, Entrala-Bernal C, Fernandez- Rosado F. Novel FA2H mutation in a girl with familial spastic paraplegia. J Neurol Sci. 2015;357(1- 2):332-4. 160. Zaki MS, Selim L, Mansour L, Mahmoud IG, Fenstermaker AG, Gabriel SB, et al. Mutations in FA2H in three Arab families with a clinical spectrum of neurodegeneration and hereditary spastic paraparesis. Clin Genet. 2015;88(1):95-7. 161. Potter KA, Kern MJ, Fullbright G, Bielawski J, Scherer SS, Yum SW, et al. Central nervous system dysfunction in a mouse model of FA2H deficiency. Glia. 2011;59(7):1009-21.

264

162. Zoller I, Meixner M, Hartmann D, Bussow H, Meyer R, Gieselmann V, et al. Absence of 2- hydroxylated sphingolipids is compatible with normal neural development but causes late-onset axon and myelin sheath degeneration. J Neurosci. 2008;28(39):9741-54. 163. Najim al-Din AS, Wriekat A, Mubaidin A, Dasouki M, Hiari M. Pallido-pyramidal degeneration, supranuclear upgaze paresis and dementia: Kufor-Rakeb syndrome. Acta neurologica Scandinavica. 1994;89(5):347-52. 164. Di Fonzo A, Chien HF, Socal M, Giraudo S, Tassorelli C, Iliceto G, et al. ATP13A2 missense mutations in juvenile parkinsonism and young onset Parkinson disease. Neurology. 2007;68(19):1557-62. 165. Williams DR, Hadeed A, al-Din AS, Wreikat AL, Lees AJ. Kufor Rakeb disease: autosomal recessive, levodopa-responsive parkinsonism with pyramidal degeneration, supranuclear gaze palsy, and dementia. Mov Disord. 2005;20(10):1264-71. 166. Behrens MI, Bruggemann N, Chana P, Venegas P, Kagi M, Parrao T, et al. Clinical spectrum of Kufor-Rakeb syndrome in the Chilean kindred with ATP13A2 mutations. Mov Disord. 2010;25(12):1929-37. 167. Machner B, Sprenger A, Behrens MI, Ramirez A, Bruggemann N, Klein C, et al. Eye movement disorders in ATP13A2 mutation carriers (PARK9). Mov Disord. 2010;25(15):2687-9. 168. Schneider SA, Paisan-Ruiz C, Quinn NP, Lees AJ, Houlden H, Hardy J, et al. ATP13A2 mutations (PARK9) cause neurodegeneration with brain iron accumulation. Mov Disord. 2010;25(8):979- 84. 169. Chien HF, Bonifati V, Barbosa ER. ATP13A2-related neurodegeneration (PARK9) without evidence of brain iron accumulation. Mov Disord. 2011;26(7):1364-5. 170. Crosiers D, Ceulemans B, Meeus B, Nuytemans K, Pals P, Van Broeckhoven C, et al. Juvenile dystonia-parkinsonism and dementia caused by a novel ATP13A2 frameshift mutation. Parkinsonism & related disorders. 2011;17(2):135-8. 171. Santoro L, Breedveld GJ, Manganelli F, Iodice R, Pisciotta C, Nolano M, et al. Novel ATP13A2 (PARK9) homozygous mutation in a family with marked phenotype variability. Neurogenetics. 2011;12(1):33-9. 172. Park JS, Mehta P, Cooper AA, Veivers D, Heimbach A, Stiller B, et al. Pathogenic effects of novel mutations in the P-type ATPase ATP13A2 (PARK9) causing Kufor-Rakeb syndrome, a form of early-onset parkinsonism. Hum Mutat. 2011;32(8):956-64. 173. Schultheis PJ, Hagen TT, O'Toole KK, Tachibana A, Burke CR, McGill DL, et al. Characterization of the P5 subfamily of P-type transport ATPases in mice. Biochem Biophys Res Commun. 2004;323(3):731-8. 174. Ramonet D, Podhajska A, Stafa K, Sonnay S, Trancikova A, Tsika E, et al. PARK9-associated ATP13A2 localizes to intracellular acidic vesicles and regulates cation homeostasis and neuronal integrity. Hum Mol Genet. 2012;21(8):1725-43. 175. Bras J, Verloes A, Schneider SA, Mole SE, Guerreiro RJ. Mutation of the parkinsonism gene ATP13A2 causes neuronal ceroid-lipofuscinosis. Hum Mol Genet. 2012;21(12):2646-50. 176. Paisan-Ruiz C, Guevara R, Federoff M, Hanagasi H, Sina F, Elahi E, et al. Early-onset L-dopa- responsive parkinsonism with pyramidal signs due to ATP13A2, PLA2G6, FBXO7 and spatacsin mutations. Mov Disord. 2010;25(12):1791-800. 177. Matsui H, Sato F, Sato S, Koike M, Taruno Y, Saiki S, et al. ATP13A2 deficiency induces a decrease in cathepsin D activity, fingerprint-like inclusion body formation, and selective degeneration of dopaminergic neurons. FEBS letters. 2013;587(9):1316-25. 178. Gusdon AM, Zhu J, Van Houten B, Chu CT. ATP13A2 regulates mitochondrial bioenergetics through macroautophagy. Neurobiology of disease. 2012;45(3):962-72. 179. Grunewald A, Arns B, Seibler P, Rakovic A, Munchau A, Ramirez A, et al. ATP13A2 mutations impair mitochondrial function in fibroblasts from patients with Kufor-Rakeb syndrome. Neurobiol Aging. 2012;33(8):1843 e1-7. 180. Kett LR, Stiller B, Bernath MM, Tasset I, Blesa J, Jackson-Lewis V, et al. alpha-Synuclein- independent histopathological and motor deficits in mice lacking the endolysosomal Parkinsonism protein Atp13a2. J Neurosci. 2015;35(14):5724-42. 181. Schultheis PJ, Fleming SM, Clippinger AK, Lewis J, Tsunemi T, Giasson B, et al. Atp13a2- deficient mice exhibit neuronal ceroid lipofuscinosis, limited alpha-synuclein accumulation and age- dependent sensorimotor deficits. Hum Mol Genet. 2013;22(10):2067-82. 182. Lopes da Fonseca T, Correia A, Hasselaar W, van der Linde HC, Willemsen R, Outeiro TF. The zebrafish homologue of Parkinson's disease ATP13A2 is essential for embryonic survival. Brain research bulletin. 2013;90:118-26. 183. Chinnery PF, Crompton DE, Birchall D, Jackson MJ, Coulthard A, Lombes A, et al. Clinical features and natural history of neuroferritinopathy caused by the FTL1 460InsA mutation. Brain. 2007;130(Pt 1):110-9. 184. McNeill A, Chinnery PF. Neuroferritinopathy: update on clinical features and pathogenesis. Current drug targets. 2012;13(9):1200-3. 265

185. Levi S, Rovida E. Neuroferritinopathy: From ferritin structure modification to pathogenetic mechanism. Neurobiology of disease. 2015;81:134-43. 186. Hautot D, Pankhurst QA, Morris CM, Curtis A, Burn J, Dobson J. Preliminary observation of elevated levels of nanocrystalline iron oxide in the basal ganglia of neuroferritinopathy patients. Biochim Biophys Acta. 2007;1772(1):21-5. 187. Mancuso M, Davidzon G, Kurlan RM, Tawil R, Bonilla E, Di Mauro S, et al. Hereditary ferritinopathy: a novel mutation, its cellular pathology, and pathogenetic insights. Journal of neuropathology and experimental neurology. 2005;64(4):280-94. 188. Keogh MJ, Jonas P, Coulthard A, Chinnery PF, Burn J. Neuroferritinopathy: a new inborn error of iron metabolism. Neurogenetics. 2012;13(1):93-6. 189. Maciel P, Cruz VT, Constante M, Iniesta I, Costa MC, Gallati S, et al. Neuroferritinopathy: missense mutation in FTL causing early-onset bilateral pallidal involvement. Neurology. 2005;65(4):603- 5. 190. Li W, Garringer HJ, Goodwin CB, Richine B, Acton A, VanDuyn N, et al. Systemic and cerebral iron homeostasis in ferritin knock-out mice. PloS one. 2015;10(1):e0117435. 191. Maccarinelli F, Pagani A, Cozzi A, Codazzi F, Di Giacomo G, Capoccia S, et al. A novel neuroferritinopathy mouse model (FTL 498InsTC) shows progressive brain iron dysregulation, morphological signs of early neurodegeneration and motor coordination deficits. Neurobiology of disease. 2015;81:119-33. 192. Vidal R, Miravalle L, Gao X, Barbeito AG, Baraibar MA, Hekmatyar SK, et al. Expression of a mutant form of the ferritin light chain gene induces neurodegeneration and iron overload in transgenic mice. J Neurosci. 2008;28(1):60-7. 193. Ogimoto M, Anzai K, Takenoshita H, Kogawa K, Akehi Y, Yoshida R, et al. Criteria for early identification of aceruloplasminemia. Internal medicine. 2011;50(13):1415-8. 194. Miyajima H, Takahashi Y, Kono S. Aceruloplasminemia, an inherited disorder of iron metabolism. Biometals. 2003;16(1):205-13. 195. Kono S, Miyajima H. Molecular and pathological basis of aceruloplasminemia. Biological research. 2006;39(1):15-23. 196. Miyajima H. Aceruloplasminemia, an iron metabolic disorder. Neuropathology : official journal of the Japanese Society of Neuropathology. 2003;23(4):345-50. 197. Hellman NE, Gitlin JD. Ceruloplasmin metabolism and function. Annual review of nutrition. 2002;22:439-58. 198. Gonzalez-Cuyar LF, Perry G, Miyajima H, Atwood CS, Riveros-Angel M, Lyons PF, et al. Redox active iron accumulation in aceruloplasminemia. Neuropathology : official journal of the Japanese Society of Neuropathology. 2008;28(5):466-71. 199. Oide T, Yoshida K, Kaneko K, Ohta M, Arima K. Iron overload and antioxidative role of perivascular astrocytes in aceruloplasminemia. Neuropathology and applied neurobiology. 2006;32(2):170-6. 200. Yoshida K, Kaneko K, Miyajima H, Tokuda T, Nakamura A, Kato M, et al. Increased lipid peroxidation in the brains of aceruloplasminemia patients. J Neurol Sci. 2000;175(2):91-5. 201. Patel BN, Dunn RJ, Jeong SY, Zhu Q, Julien JP, David S. Ceruloplasmin regulates iron levels in the CNS and prevents free radical injury. J Neurosci. 2002;22(15):6578-86. 202. Dusi S, Valletta L, Haack TB, Tsuchiya Y, Venco P, Pasqualato S, et al. Exome sequence reveals mutations in CoA synthase as a cause of neurodegeneration with brain iron accumulation. Am J Hum Genet. 2014;94(1):11-22. 203. Woodhouse NJ, Sakati NA. A syndrome of hypogonadism, alopecia, diabetes mellitus, mental retardation, deafness, and ECG abnormalities. J Med Genet. 1983;20(3):216-9. 204. Alazami AM, Al-Saif A, Al-Semari A, Bohlega S, Zlitni S, Alzahrani F, et al. Mutations in C2orf37, encoding a nucleolar protein, cause hypogonadism, alopecia, diabetes mellitus, mental retardation, and extrapyramidal syndrome. Am J Hum Genet. 2008;83(6):684-91. 205. Alazami AM, Schneider SA, Bonneau D, Pasquier L, Carecchio M, Kojovic M, et al. C2orf37 mutational spectrum in Woodhouse-Sakati syndrome patients. Clin Genet. 2010;78(6):585-90. 206. Ali RH, Shah K, Nasir A, Steyaert W, Coucke PJ, Ahmad W. Exome sequencing revealed a novel biallelic deletion in the DCAF17 gene underlying Woodhouse Sakati syndrome (WSS). Clin Genet. 2015. 207. Steindl K, Alazami AM, Bhatia KP, Wuerfel JT, Petersen D, Cartolari R, et al. A novel C2orf37 mutation causes the first Italian cases of Woodhouse Sakati syndrome. Clin Genet. 2010;78(6):594-7. 208. Di Fonzo A, Dekker MC, Montagna P, Baruzzi A, Yonova EH, Correia Guedes L, et al. FBXO7 mutations cause autosomal recessive, early-onset parkinsonian-pyramidal syndrome. Neurology. 2009;72(3):240-5. 209. Shojaee S, Sina F, Banihosseini SS, Kazemi MH, Kalhor R, Shahidi GA, et al. Genome-wide linkage analysis of a Parkinsonian-pyramidal syndrome pedigree by 500 K SNP arrays. Am J Hum Genet. 2008;82(6):1375-84. 266

210. Winston JT, Koepp DM, Zhu C, Elledge SJ, Harper JW. A family of mammalian F-box proteins. Curr Biol. 1999;9(20):1180-2. 211. Giannandrea M, Bianchi V, Mignogna ML, Sirri A, Carrabino S, D'Elia E, et al. Mutations in the small GTPase gene RAB39B are responsible for X-linked mental retardation associated with autism, epilepsy, and macrocephaly. Am J Hum Genet. 2010;86(2):185-95. 212. Wilson GR, Sim JC, McLean C, Giannandrea M, Galea CA, Riseley JR, et al. Mutations in RAB39B cause X-linked intellectual disability and early-onset Parkinson disease with alpha-synuclein pathology. Am J Hum Genet. 2014;95(6):729-35. 213. Jaberi E, Rohani M, Shahidi GA, Nafissi S, Arefian E, Soleimani M, et al. Identification of mutation in GTPBP2 in patients of a family with neurodegeneration accompanied by iron deposition in the brain. Neurobiol Aging. 2016;38:216 e11-8. 214. Burgetova A, Seidl Z, Krasensky J, Horakova D, Vaneckova M. Multiple sclerosis and the accumulation of iron in the Basal Ganglia: quantitative assessment of brain iron using MRI t(2) relaxometry. European neurology. 2010;63(3):136-43. 215. Al-Semari A, Bohlega S. Autosomal-recessive syndrome with alopecia, hypogonadism, progressive extra-pyramidal disorder, white matter disease, sensory neural deafness, diabetes mellitus, and low IGF1. Am J Med Genet A. 2007;143(2):149-60. 216. Muthane U, Chickabasaviah Y, Kaneski C, Shankar SK, Narayanappa G, Christopher R, et al. Clinical features of adult GM1 gangliosidosis: report of three Indian patients and review of 40 cases. Mov Disord. 2004;19(11):1334-41. 217. Tanaka R, Momoi T, Yoshida A, Okumura M, Yamakura S, Takasaki Y, et al. Type 3 GM1 gangliosidosis: clinical and neuroradiological findings in an 11-year-old girl. J Neurol. 1995;242(5):299- 303. 218. Zoons E, de Koning TJ, Abeling NG, Tijssen MA. Neurodegeneration with Brain Iron Accumulation on MRI: An Adult Case of alpha-Mannosidosis. JIMD reports. 2012;4:99-102. 219. Bartzokis G, Cummings J, Perlman S, Hance DB, Mintz J. Increased basal ganglia iron levels in Huntington disease. Arch Neurol. 1999;56(5):569-74. 220. Santillo AF, Skoglund L, Lindau M, Eeg-Olofsson KE, Tovi M, Engler H, et al. Frontotemporal dementia-amyotrophic lateral sclerosis complex is simulated by neurodegeneration with brain iron accumulation. Alzheimer disease and associated disorders. 2009;23(3):298-300. 221. Haba-Rubio J, Staner L, Petiau C, Erb G, Schunck T, Macher JP. Restless legs syndrome and low brain iron levels in patients with haemochromatosis. J Neurol Neurosurg Psychiatry. 2005;76(7):1009-10. 222. Gautschi M, Merlini L, Calza AM, Hayflick S, Nuoffer JM, Fluss J. Late diagnosis of fucosidosis in a child with progressive fixed dystonia, bilateral pallidal lesions and red spots on the skin. European journal of paediatric neurology : EJPN : official journal of the European Paediatric Neurology Society. 2014. 223. Altarescu G, Sun M, Moore DF, Smith JA, Wiggs EA, Solomon BI, et al. The neurogenetics of mucolipidosis type IV. Neurology. 2002;59(3):306-13. 224. Levy M, Turtzo C, Llinas RH. Superficial siderosis: a case report and review of the literature. Nature clinical practice Neurology. 2007;3(1):54-8; quiz 9. 225. Spengos K, Koutsis G, Tsivgoulis G, Panas M, Vemmos K, Vassilopoulos D. [Superficial siderosis of the CNS. Case report and literature review]. Der Nervenarzt. 2004;75(5):492-5. 226. Waldvogel D, van Gelderen P, Hallett M. Increased iron in the dentate nucleus of patients with Friedrich's ataxia. Ann Neurol. 1999;46(1):123-5. 227. Visanji NP, Collingwood JF, Finnegan ME, Tandon A, House E, Hazrati LN. Iron deficiency in parkinsonism: region-specific iron dysregulation in Parkinson's disease and multiple system atrophy. Journal of Parkinson's disease. 2013;3(4):523-37. 228. Berg D, Hochstrasser H. Iron metabolism in Parkinsonian syndromes. Mov Disord. 2006;21(9):1299-310. 229. Wang Y, Butros SR, Shuai X, Dai Y, Chen C, Liu M, et al. Different iron-deposition patterns of multiple system atrophy with predominant parkinsonism and idiopathetic Parkinson diseases demonstrated by phase-corrected susceptibility-weighted imaging. AJNR American journal of neuroradiology. 2012;33(2):266-73. 230. Davie CA, Barker GJ, Machado C, Miller DH, Lees AJ. Proton magnetic resonance spectroscopy in Steele-Richardson-Olszewski syndrome. Mov Disord. 1997;12(5):767-71. 231. Molinuevo JL, Munoz E, Valldeoriola F, Tolosa E. The eye of the tiger sign in cortical-basal ganglionic degeneration. Mov Disord. 1999;14(1):169-71. 232. Dusek P, Jankovic J, Le W. Iron dysregulation in movement disorders. Neurobiology of disease. 2012;46(1):1-18. 233. Ogretmen B, Hannun YA. Biologically active sphingolipids in cancer pathogenesis and treatment. Nature reviews Cancer. 2004;4(8):604-16.

267

234. Bras J, Singleton A, Cookson MR, Hardy J. Emerging pathways in genetic Parkinson's disease: Potential role of ceramide metabolism in Lewy body disease. The FEBS journal. 2008;275(23):5767-73. 235. Okita K, Matsumura Y, Sato Y, Okada A, Morizane A, Okamoto S, et al. A more efficient method to generate integration-free human iPS cells. Nature methods. 2011;8(5):409-12. 236. Chambers SM, Fasano CA, Papapetrou EP, Tomishima M, Sadelain M, Studer L. Highly efficient neural conversion of human ES and iPS cells by dual inhibition of SMAD signaling. Nature biotechnology. 2009;27(3):275-80. 237. Shi Y, Kirwan P, Livesey FJ. Directed differentiation of human pluripotent stem cells to cerebral cortex neurons and neural networks. Nature protocols. 2012;7(10):1836-46. 238. Gatchel JR, Zoghbi HY. Diseases of unstable repeat expansion: mechanisms and common principles. Nature reviews Genetics. 2005;6(10):743-55. 239. McMurray CT. Mechanisms of trinucleotide repeat instability during human development. Nature reviews Genetics. 2010;11(11):786-99. 240. Shao J, Diamond MI. Polyglutamine diseases: emerging concepts in pathogenesis and therapy. Hum Mol Genet. 2007;16 Spec No. 2:R115-23. 241. Jayadev S, Bird TD. Hereditary ataxias: overview. Genetics in medicine : official journal of the American College of Medical Genetics. 2013;15(9):673-83. 242. http://neuromuscular.wustl.edu/ataxia/recatax.html. Recessive Ataxia. 2016. 243. Coutelier M, Stevanin G, Brice A. Genetic landscape remodelling in spinocerebellar ataxias: the influence of next-generation sequencing. J Neurol. 2015. 244. Ruano L, Melo C, Silva MC, Coutinho P. The global epidemiology of hereditary ataxia and spastic paraplegia: a systematic review of prevalence studies. Neuroepidemiology. 2014;42(3):174-83. 245. van de Warrenburg BP, Sinke RJ, Verschuuren-Bemelmans CC, Scheffer H, Brunt ER, Ippel PF, et al. Spinocerebellar ataxias in the Netherlands: prevalence and age at onset variance analysis. Neurology. 2002;58(5):702-8. 246. Zortea M, Armani M, Pastorello E, Nunez GF, Lombardi S, Tonello S, et al. Prevalence of inherited ataxias in the province of Padua, Italy. Neuroepidemiology. 2004;23(6):275-80. 247. Bird TD. Hereditary Ataxia Overview. In: Pagon RA, Adam MP, Ardinger HH, Wallace SE, Amemiya A, Bean LJH, et al., editors. GeneReviews(R). Seattle (WA)1993. 248. Hersheson J, Haworth A, Houlden H. The inherited ataxias: genetic heterogeneity, mutation databases, and future directions in research and clinical diagnostics. Hum Mutat. 2012;33(9):1324-32. 249. Sailer A, Houlden H. Recent advances in the genetics of cerebellar ataxias. Curr Neurol Neurosci Rep. 2012;12(3):227-36. 250. Schulz JB, Boesch S, Burk K, Durr A, Giunti P, Mariotti C, et al. Diagnosis and treatment of Friedreich ataxia: a European perspective. Nat Rev Neurol. 2009;5(4):222-34. 251. Wiethoff S, Arber C, Li A, Wray S, Houlden H, Patani R. Using human induced pluripotent stem cells to model cerebellar disease: Hope and hype. Journal of neurogenetics. 2015;29(2-3):95-102. 252. Manto M. The cerebellum, cerebellar disorders, and cerebellar research--two centuries of discoveries. Cerebellum. 2008;7(4):505-16. 253. Azevedo FA, Carvalho LR, Grinberg LT, Farfel JM, Ferretti RE, Leite RE, et al. Equal numbers of neuronal and nonneuronal cells make the human brain an isometrically scaled-up primate brain. The Journal of comparative neurology. 2009;513(5):532-41. 254. De Zeeuw CI, Berrebi AS. Postsynaptic targets of Purkinje cell terminals in the cerebellar and vestibular nuclei of the rat. The European journal of neuroscience. 1995;7(11):2322-33. 255. Ito M. The modifiable neuronal network of the cerebellum. The Japanese journal of physiology. 1984;34(5):781-92. 256. Voogd J, Gerrits NM, Ruigrok TJ. Organization of the vestibulocerebellum. Annals of the New York Academy of Sciences. 1996;781:553-79. 257. Ruigrok TJ. Ins and outs of cerebellar modules. Cerebellum. 2011;10(3):464-74. 258. Glickstein M, Strata P, Voogd J. Cerebellum: history. Neuroscience. 2009;162(3):549-59. 259. Schmahmann JD. Rediscovery of an early concept. International review of neurobiology. 1997;41:3-27. 260. Strata P. The Emotional Cerebellum. Cerebellum. 2015. 261. Strick PL, Dum RP, Fiez JA. Cerebellum and nonmotor function. Annual review of neuroscience. 2009;32:413-34. 262. Anand BK, Malhotra CL, Singh B, Dua S. Cerebellar projections to limbic system. J Neurophysiol. 1959;22(4):451-7. 263. Timmann D. [Contribution of the cerebellum to cognition]. Fortschritte der Neurologie- Psychiatrie. 2012;80(1):44-52. 264. Takahashi K, Tanabe K, Ohnuki M, Narita M, Ichisaka T, Tomoda K, et al. Induction of pluripotent stem cells from adult human fibroblasts by defined factors. Cell. 2007;131(5):861-72. 265. Tabar V, Studer L. Pluripotent stem cells in regenerative medicine: challenges and recent progress. Nature reviews Genetics. 2014;15(2):82-92. 268

266. Yamanaka S, Blau HM. Nuclear reprogramming to a pluripotent state by three approaches. Nature. 2010;465(7299):704-12. 267. Xie YZ, Zhang RX. Neurodegenerative diseases in a dish: the promise of iPSC technology in disease modeling and therapeutic discovery. Neurological sciences : official journal of the Italian Neurological Society and of the Italian Society of Clinical Neurophysiology. 2015;36(1):21-7. 268. Efthymiou A, Shaltouki A, Steiner JP, Jha B, Heman-Ackah SM, Swistowski A, et al. Functional screening assays with neurons generated from pluripotent stem cell-derived neural stem cells. Journal of biomolecular screening. 2014;19(1):32-43. 269. Malik N, Efthymiou AG, Mather K, Chester N, Wang X, Nath A, et al. Compounds with species and cell type specific toxicity identified in a 2000 compound drug screen of neural stem cells and rat mixed cortical neurons. Neurotoxicology. 2014;45:192-200. 270. Yang YM, Gupta SK, Kim KJ, Powers BE, Cerqueira A, Wainger BJ, et al. A small molecule screen in stem-cell-derived motor neurons identifies a kinase inhibitor as a candidate therapeutic for ALS. Cell stem cell. 2013;12(6):713-26. 271. Leiner HC. Solving the mystery of the human cerebellum. Neuropsychology review. 2010;20(3):229-35. 272. Voogd J. What we do not know about cerebellar systems neuroscience. Front Syst Neurosci. 2014; 2014 Dec 18;8:227. 273. Schlaeger TM, Daheron L, Brickler TR, Entwisle S, Chan K, Cianci A, et al. A comparison of non-integrating reprogramming methods. Nature biotechnology. 2015;33(1):58-63. 274. Yu J, Hu K, Smuga-Otto K, Tian S, Stewart R, Slukvin, II, et al. Human induced pluripotent stem cells free of vector and transgene sequences. Science. 2009;324(5928):797-801. 275. Jiang J, Lv W, Ye X, Wang L, Zhang M, Yang H, et al. Zscan4 promotes genomic stability during reprogramming and dramatically improves the quality of iPS cells as demonstrated by tetraploid complementation. Cell research. 2013;23(1):92-106. 276. Hou P, Li Y, Zhang X, Liu C, Guan J, Li H, et al. Pluripotent stem cells induced from mouse somatic cells by small-molecule compounds. Science. 2013;341(6146):651-4. 277. Warren L, Manos PD, Ahfeldt T, Loh YH, Li H, Lau F, et al. Highly efficient reprogramming to pluripotency and directed differentiation of human cells with synthetic modified mRNA. Cell stem cell. 2010;7(5):618-30. 278. Nakano-Okuno M, Borah BR, Nakano I. Ethics of iPSC-based clinical research for age-related macular degeneration: patient-centered risk-benefit analysis. Stem cell reviews. 2014;10(6):743-52. 279. Schwartz RE, Fleming HE, Khetani SR, Bhatia SN. Pluripotent stem cell-derived hepatocyte- like cells. Biotechnology advances. 2014;32(2):504-13. 280. Teng S, Liu C, Krettek C, Jagodzinski M. The application of induced pluripotent stem cells for bone regeneration: current progress and prospects. Tissue engineering Part B, Reviews. 2014;20(4):328- 39. 281. Dimos JT, Rodolfa KT, Niakan KK, Weisenthal LM, Mitsumoto H, Chung W, et al. Induced pluripotent stem cells generated from patients with ALS can be differentiated into motor neurons. Science. 2008;321(5893):1218-21. 282. Moretti A, Bellin M, Welling A, Jung CB, Lam JT, Bott-Flugel L, et al. Patient-specific induced pluripotent stem-cell models for long-QT syndrome. N Engl J Med. 2010;363(15):1397-409. 283. Reubinoff BE, Itsykson P, Turetsky T, Pera MF, Reinhartz E, Itzik A, et al. Neural progenitors from human embryonic stem cells. Nature biotechnology. 2001;19(12):1134-40. 284. Zhang SC, Wernig M, Duncan ID, Brustle O, Thomson JA. In vitro differentiation of transplantable neural precursors from human embryonic stem cells. Nature biotechnology. 2001;19(12):1129-33. 285. Zirra A, Wiethoff S, Patani R. Neural Conversion and Patterning of Human Pluripotent Stem Cells: A Developmental Perspective. Stem cells international. 2016;2016:8291260. 286. Miller JD, Ganat YM, Kishinevsky S, Bowman RL, Liu B, Tu EY, et al. Human iPSC-based modeling of late-onset disease via progerin-induced aging. Cell stem cell. 2013;13(6):691-705. 287. Lancaster MA, Renner M, Martin CA, Wenzel D, Bicknell LS, Hurles ME, et al. Cerebral organoids model human brain development and microcephaly. Nature. 2013;501(7467):373-9. 288. Marchetto MC, Carromeu C, Acab A, Yu D, Yeo GW, Mu Y, et al. A model for neural development and treatment of Rett syndrome using human induced pluripotent stem cells. Cell. 2010;143(4):527-39. 289. Patani R, Lewis PA, Trabzuni D, Puddifoot CA, Wyllie DJ, Walker R, et al. Investigating the utility of human embryonic stem cell-derived neurons to model ageing and neurodegenerative disease using whole-genome gene expression and splicing analysis. J Neurochem. 2012;122(4):738-51. 290. Israel MA, Yuan SH, Bardy C, Reyna SM, Mu Y, Herrera C, et al. Probing sporadic and familial Alzheimer's disease using induced pluripotent stem cells. Nature. 2012;482(7384):216-20.

269

291. Koch P, Breuer P, Peitz M, Jungverdorben J, Kesavan J, Poppe D, et al. Excitation-induced ataxin-3 aggregation in neurons from patients with Machado-Joseph disease. Nature. 2011;480(7378):543-6. 292. Lee G, Papapetrou EP, Kim H, Chambers SM, Tomishima MJ, Fasano CA, et al. Modelling pathogenesis and treatment of familial dysautonomia using patient-specific iPSCs. Nature. 2009;461(7262):402-6. 293. Sanchez-Danes A, Richaud-Patin Y, Carballo-Carbajal I, Jimenez-Delgado S, Caig C, Mora S, et al. Disease-specific phenotypes in dopamine neurons from human iPS-based models of genetic and sporadic Parkinson's disease. EMBO molecular medicine. 2012;4(5):380-95. 294. Schondorf DC, Aureli M, McAllister FE, Hindley CJ, Mayer F, Schmid B, et al. iPSC-derived neurons from GBA1-associated Parkinson's disease patients show autophagic defects and impaired calcium homeostasis. Nature communications. 2014;5:4028. 295. Woodard CM, Campos BA, Kuo SH, Nirenberg MJ, Nestor MW, Zimmer M, et al. iPSC- derived dopamine neurons reveal differences between monozygotic twins discordant for Parkinson's disease. Cell reports. 2014;9(4):1173-82. 296. Bellin M, Marchetto MC, Gage FH, Mummery CL. Induced pluripotent stem cells: the new patient? Nature reviews Molecular cell biology. 2012;13(11):713-26. 297. Cooper O, Seo H, Andrabi S, Guardia-Laguarta C, Graziotto J, Sundberg M, et al. Pharmacological rescue of mitochondrial deficits in iPSC-derived neural cells from patients with familial Parkinson's disease. Science translational medicine. 2012;4(141):141ra90. 298. Cao L, Tan L, Jiang T, Zhu XC, Yu JT. Induced Pluripotent Stem Cells for Disease Modeling and Drug Discovery in Neurodegenerative Diseases. Molecular neurobiology. 2014. 299. Ross CA, Akimov SS. Human-induced pluripotent stem cells: potential for neurodegenerative diseases. Hum Mol Genet. 2014;23(R1):R17-26. 300. Cyranoski D. Stem cells cruise to clinic. Nature. 2013;494(7438):413. 301. Grskovic M, Javaherian A, Strulovici B, Daley GQ. Induced pluripotent stem cells-- opportunities for disease modelling and drug discovery. Nature reviews Drug discovery. 2011;10(12):915-29. 302. Ausubel LJ, Lopez PM, Couture LA. GMP scale-up and banking of pluripotent stem cells for cellular therapy applications. Methods Mol Biol. 2011;767:147-59. 303. Frantz S. Embryonic stem cell pioneer Geron exits field, cuts losses. Nature biotechnology. 2012;30(1):12-3. 304. Schwartz SD, Hubschman JP, Heilwell G, Franco-Cardenas V, Pan CK, Ostrick RM, et al. Embryonic stem cell trials for macular degeneration: a preliminary report. Lancet. 2012;379(9817):713- 20. 305. Kordower JH, Chu Y, Hauser RA, Freeman TB, Olanow CW. Lewy body-like pathology in long-term embryonic nigral transplants in Parkinson's disease. Nature medicine. 2008;14(5):504-6. 306. Aubry L, Bugi A, Lefort N, Rousseau F, Peschanski M, Perrier AL. Striatal progenitors derived from human ES cells mature into DARPP32 neurons in vitro and in quinolinic acid-lesioned rats. Proceedings of the National Academy of Sciences of the United States of America. 2008;105(43):16707- 12. 307. Delli Carri A, Onorati M, Lelos MJ, Castiglioni V, Faedo A, Menon R, et al. Developmentally coordinated extrinsic signals drive human pluripotent stem cell differentiation toward authentic DARPP- 32+ medium-sized spiny neurons. Development. 2013;140(2):301-12. 308. Perrier A, Peschanski M. How can human pluripotent stem cells help decipher and cure Huntington's disease? Cell stem cell. 2012;11(2):153-61. 309. Jeon I, Lee N, Li JY, Park IH, Park KS, Moon J, et al. Neuronal properties, in vivo effects, and pathology of a Huntington's disease patient-derived induced pluripotent stem cells. Stem cells. 2012;30(9):2054-62. 310. Perrier AL, Tabar V, Barberi T, Rubio ME, Bruses J, Topf N, et al. Derivation of midbrain dopamine neurons from human embryonic stem cells. Proceedings of the National Academy of Sciences of the United States of America. 2004;101(34):12543-8. 311. Yan Y, Yang D, Zarnowska ED, Du Z, Werbel B, Valliere C, et al. Directed differentiation of dopaminergic neuronal subtypes from human embryonic stem cells. Stem cells. 2005;23(6):781-90. 312. Kriks S, Shim JW, Piao J, Ganat YM, Wakeman DR, Xie Z, et al. Dopamine neurons derived from human ES cells efficiently engraft in animal models of Parkinson's disease. Nature. 2011;480(7378):547-51. 313. Li XJ, Du ZW, Zarnowska ED, Pankratz M, Hansen LO, Pearce RA, et al. Specification of motoneurons from human embryonic stem cells. Nature biotechnology. 2005;23(2):215-21. 314. Hu BY, Zhang SC. Differentiation of spinal motor neurons from pluripotent human stem cells. Nature protocols. 2009;4(9):1295-304.

270

315. Li XJ, Hu BY, Jones SA, Zhang YS, Lavaute T, Du ZW, et al. Directed differentiation of ventral spinal progenitors and motor neurons from human embryonic stem cells by small molecules. Stem cells. 2008;26(4):886-93. 316. Patani R, Hollins AJ, Wishart TM, Puddifoot CA, Alvarez S, de Lera AR, et al. Retinoid- independent motor neurogenesis from human embryonic stem cells reveals a medial columnar ground state. Nature communications. 2011;2:214. 317. Corti S, Nizzardo M, Simone C, Falcone M, Nardini M, Ronchi D, et al. Genetic correction of human induced pluripotent stem cells from patients with spinal muscular atrophy. Science translational medicine. 2012;4(165):165ra2. 318. Devlin AC, Burr K, Borooah S, Foster JD, Cleary EM, Geti I, et al. Human iPSC-derived motoneurons harbouring TARDBP or C9ORF72 ALS mutations are dysfunctional despite maintaining viability. Nature communications. 2015;6:5999. 319. Valensi-Kurtz M, Lefler S, Cohen MA, Aharonowiz M, Cohen-Kupiec R, Sheinin A, et al. Enriched population of PNS neurons derived from human embryonic stem cells as a platform for studying peripheral neuropathies. PloS one. 2010;5(2):e9290. 320. Chambers SM, Qi Y, Mica Y, Lee G, Zhang XJ, Niu L, et al. Combined small-molecule inhibition accelerates developmental timing and converts human pluripotent stem cells into nociceptors. Nature biotechnology. 2012;30(7):715-20. 321. Lee KS, Zhou W, Scott-McKean JJ, Emmerling KL, Cai GY, Krah DL, et al. Human sensory neurons derived from induced pluripotent stem cells support varicella-zoster virus infection. PloS one. 2012;7(12):e53010. 322. Pomp O, Brokhman I, Ben-Dor I, Reubinoff B, Goldstein RS. Generation of peripheral sensory and sympathetic neurons and neural crest cells from human embryonic stem cells. Stem cells. 2005;23(7):923-30. 323. Eigentler A, Boesch S, Schneider R, Dechant G, Nat R. Induced pluripotent stem cells from friedreich ataxia patients fail to upregulate frataxin during in vitro differentiation to peripheral sensory neurons. Stem cells and development. 2013;22(24):3271-82. 324. Hick A, Wattenhofer-Donze M, Chintawar S, Tropel P, Simard JP, Vaucamps N, et al. Neurons and cardiomyocytes derived from induced pluripotent stem cells as a model for mitochondrial defects in Friedreich's ataxia. Disease models & mechanisms. 2013;6(3):608-21. 325. Bird MJ, Needham K, Frazier AE, van Rooijen J, Leung J, Hough S, et al. Functional characterization of Friedreich ataxia iPS-derived neuronal progenitors and their integration in the adult brain. PloS one. 2014;9(7):e101718. 326. Erceg S, Lukovic D, Moreno-Manzano V, Stojkovic M, Bhattacharya SS. Derivation of cerebellar neurons from human pluripotent stem cells. Current protocols in stem cell biology. 2012;Chapter 1:Unit 1H 5. 327. Salero E, Hatten ME. Differentiation of ES cells into cerebellar neurons. Proceedings of the National Academy of Sciences of the United States of America. 2007;104(8):2997-3002. 328. Su HL, Muguruma K, Matsuo-Takasaki M, Kengaku M, Watanabe K, Sasai Y. Generation of cerebellar neuron precursors from embryonic stem cells. Developmental biology. 2006;290(2):287-96. 329. Muguruma K, Nishiyama A, Ono Y, Miyawaki H, Mizuhara E, Hori S, et al. Ontogeny- recapitulating generation and tissue integration of ES cell-derived Purkinje cells. Nature neuroscience. 2010;13(10):1171-80. 330. Muguruma K, Nishiyama A, Kawakami H, Hashimoto K, Sasai Y. Self-Organization of Polarized Cerebellar Tissue in 3D Culture of Human Pluripotent Stem Cells. Cell reports. 2015. 331. Arseni C, Ciurea AV. Statistical survey of 276 cases of medulloblastoma (1935--1978). Acta neurochirurgica. 1981;57(3-4):159-62. 332. Koch P, Opitz T, Steinbeck JA, Ladewig J, Brustle O. A rosette-type, self-renewing human ES cell-derived neural stem cell with potential for in vitro instruction and synaptic integration. Proceedings of the National Academy of Sciences of the United States of America. 2009;106(9):3225-30. 333. Ku S, Soragni E, Campau E, Thomas EA, Altun G, Laurent LC, et al. Friedreich's ataxia induced pluripotent stem cells model intergenerational GAATTC triplet repeat instability. Cell stem cell. 2010;7(5):631-7. 334. Wurst W, Bally-Cuif L. Neural plate patterning: upstream and downstream of the isthmic organizer. Nat Rev Neurosci. 2001;2(2):99-108. 335. Ten Donkelaar HJ, Lammens M. Development of the human cerebellum and its disorders. Clinics in perinatology. 2009;36(3):513-30. 336. White JJ, Sillitoe RV. Development of the cerebellum: from gene expression patterns to circuit maps. Wiley interdisciplinary reviews Developmental biology. 2013;2(1):149-64. 337. Marzban H, Del Bigio MR, Alizadeh J, Ghavami S, Zachariah RM, Rastegar M. Cellular commitment in the developing cerebellum. Frontiers in cellular neuroscience. 2014;8:450. 338. Hekman KE, Gomez CM. The autosomal dominant spinocerebellar ataxias: emerging mechanistic themes suggest pervasive Purkinje cell vulnerability. J Neurol Neurosurg Psychiatry. 2014. 271

339. Schorge S, van de Leemput J, Singleton A, Houlden H, Hardy J. Human ataxias: a genetic dissection of inositol triphosphate receptor (ITPR1)-dependent signaling. Trends in neurosciences. 2010;33(5):211-9. 340. Buffo A, Rossi F. Origin, lineage and function of cerebellar glia. Prog Neurobiol. 2013;109:42- 63. 341. Hoshino M. Molecular machinery governing GABAergic neuron specification in the cerebellum. Cerebellum. 2006;5(3):193-8. 342. Hibi M, Shimizu T. Development of the cerebellum and cerebellar neural circuits. Developmental neurobiology. 2012;72(3):282-301. 343. Mendonca LS, Nobrega C, Hirai H, Kaspar BK, Pereira de Almeida L. Transplantation of cerebellar neural stem cells improves motor coordination and neuropathology in Machado-Joseph disease mice. Brain. 2015;138(Pt 2):320-35. 344. Wang S, Wang B, Pan N, Fu L, Wang C, Song G, et al. Differentiation of human induced pluripotent stem cells to mature functional Purkinje neurons. Scientific reports. 2015;5:9232. 345. Zeng X, Hunsberger JG, Simeonov A, Malik N, Pei Y, Rao M. Concise review: modeling central nervous system diseases using induced pluripotent stem cells. Stem cells translational medicine. 2014;3(12):1418-28. 346. Hunsberger J, Efthymiou AG, Malik N, Behl M, Mead IL, Zeng X, et al. Induced pluripotent stem cell models to enable in vitro models for screening in the CNS. Stem cells and development. 2015. 347. Kamao H, Mandai M, Okamoto S, Sakai N, Suga A, Sugita S, et al. Characterization of human induced pluripotent stem cell-derived retinal pigment epithelium cell sheets aiming for clinical application. Stem cell reports. 2014;2(2):205-18. 348. Soldner F, Laganiere J, Cheng AW, Hockemeyer D, Gao Q, Alagappan R, et al. Generation of isogenic pluripotent stem cells differing exclusively at two early onset Parkinson point mutations. Cell. 2011;146(2):318-31. 349. Hockemeyer D, Soldner F, Beard C, Gao Q, Mitalipova M, DeKelver RC, et al. Efficient targeting of expressed and silent genes in human ESCs and iPSCs using zinc-finger nucleases. Nature biotechnology. 2009;27(9):851-7. 350. Hockemeyer D, Wang H, Kiani S, Lai CS, Gao Q, Cassady JP, et al. Genetic engineering of human pluripotent cells using TALE nucleases. Nature biotechnology. 2011;29(8):731-4. 351. Ladewig J, Koch P, Brustle O. Leveling Waddington: the emergence of direct programming and the loss of cell fate hierarchies. Nature reviews Molecular cell biology. 2013;14(4):225-36. 352. Vierbuchen T, Wernig M. Molecular roadblocks for cellular reprogramming. Molecular cell. 2012;47(6):827-38. 353. Bamshad MJ, Ng SB, Bigham AW, Tabor HK, Emond MJ, Nickerson DA, et al. Exome sequencing as a tool for Mendelian disease gene discovery. Nature reviews Genetics. 2011;12(11):745- 55. 354. Van der Auwera GA, Carneiro MO, Hartl C, Poplin R, Del Angel G, Levy-Moonshine A, et al. From FastQ data to high confidence variant calls: the Genome Analysis Toolkit best practices pipeline. Current protocols in bioinformatics / editoral board, Andreas D Baxevanis [et al]. 2013;11(1110):11 0 1- 0 33. 355. Tabrizi SJ, Scahill RI, Owen G, Durr A, Leavitt BR, Roos RA, et al. Predictors of phenotypic progression and disease onset in premanifest and early-stage Huntington's disease in the TRACK-HD study: analysis of 36-month observational data. The Lancet Neurology. 2013;12(7):637-49. 356. Consortium GMoHsDG-H. Identification of Genetic Factors that Modify Clinical Onset of Huntington's Disease. Cell. 2015;162(3):516-26. 357. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. American journal of human genetics. 2007;81(3):559-75. 358. Brown MB. A method for combining non-independent, one-sided tests of significance. Biometrics. 1975:987-92. 359. Deschauer M, Gaul C, Behrmann C, Prokisch H, Zierz S, Haack TB. C19orf12 mutations in neurodegeneration with brain iron accumulation mimicking juvenile amyotrophic lateral sclerosis. J Neurol. 2012;259(11):2434-9. 360. Dziewulska D, Domitrz I, Domzal-Stryga A. Dementia means number of things - the overlap of neurodegeneration with brain iron accumulation (NBIA) and Alzheimer changes: an autopsy case. Folia neuropathologica / Association of Polish Neuropathologists and Medical Research Centre, Polish Academy of Sciences. 2010;48(2):129-33. 361. Fekete R. Late onset neurodegeneration with brain-iron accumulation presenting as parkinsonism. Case reports in neurological medicine. 2012;2012:387095. 362. Houlden H, Singleton AB. The genetics and neuropathology of Parkinson's disease. Acta Neuropathol. 2012;124(3):325-38.

272

363. Khalifa M, Naffaa L. Exome sequencing reveals a novel WDR45 frameshift mutation and inherited POLR3A heterozygous variants in a female with a complex phenotype and mixed brain MRI findings. European journal of medical genetics. 2015;58(8):381-6. 364. Martino D, Stamelou M, Bhatia KP. The differential diagnosis of Huntington's disease-like syndromes: 'red flags' for the clinician. J Neurol Neurosurg Psychiatry. 2013;84(6):650-6. 365. Schulte EC, Claussen MC, Jochim A, Haack T, Hartig M, Hempel M, et al. Mitochondrial membrane protein associated neurodegenration: a novel variant of neurodegeneration with brain iron accumulation. Mov Disord. 2013;28(2):224-7. 366. Swinnen B, Robberecht W. The phenotypic variability of amyotrophic lateral sclerosis. Nat Rev Neurol. 2014;10(11):661-70. 367. Mencacci NE, Kamsteeg EJ, Nakashima K, R'Bibo L, Lynch DS, Balint B, et al. De Novo Mutations in PDE10A Cause Childhood-Onset Chorea with Bilateral Striatal Lesions. Am J Hum Genet. 2016;98(4):763-71. 368. Pensato V, Castellotti B, Gellera C, Pareyson D, Ciano C, Nanetti L, et al. Overlapping phenotypes in complex spastic paraplegias SPG11, SPG15, SPG35 and SPG48. Brain. 2014;137(Pt 7):1907-20. 369. Rupps R, Hukin J, Balicki M, Mercimek-Mahmutoglu S, Rolfs A, Dias C. Novel Mutations in FA2H-Associated Neurodegeneration: An Underrecognized Condition? Journal of child neurology. 2013;28(11):1500-4. 370. Marelli C, Salih, M., Nguyen, K., Mallaret, M., Leboucq, N., Hassan, H., Drouot, N., Labauge, P., Koenig, M. Cerebral Iron Accumulation Is Not a Major Feature of FA2H/SPG35. Mov Disord Clin Pract. 2015;2(1):56-60. 371. Ching KH, Westaway SK, Gitschier J, Higgins JJ, Hayflick SJ. HARP syndrome is allelic with pantothenate kinase-associated neurodegeneration. Neurology. 2002;58(11):1673-4. 372. Zhou B, Westaway SK, Levinson B, Johnson MA, Gitschier J, Hayflick SJ. A novel pantothenate kinase gene (PANK2) is defective in Hallervorden-Spatz syndrome. Nat Genet. 2001;28(4):345-9. 373. Yang M, Allen H, DiCioccio RA. A mutation generating a stop codon in the alpha-L-fucosidase gene of a fucosidosis patient. Biochem Biophys Res Commun. 1992;189(2):1063-8. 374. Gautschi M, Merlini L, Calza AM, Hayflick S, Nuoffer JM, Fluss J. Late diagnosis of fucosidosis in a child with progressive fixed dystonia, bilateral pallidal lesions and red spots on the skin. European journal of paediatric neurology : EJPN : official journal of the European Paediatric Neurology Society. 2014;18(4):516-9. 375. Jain P, Sharma S, Kumar A, Aneja S. Hypomyelination with T2-hypointense globi pallidi in a child with fucosidosis. Journal of child neurology. 2014;29(7):988-9. 376. Zubarioglu T, Kiykim E, Zeybek CA, Cansever MS, Benbir G, Aydin A, et al. Clinical and neuroradiological approach to fucosidosis in a child with atypical presentation. Annals of Indian Academy of Neurology. 2015;18(4):471-4. 377. Ediz SS, Aralasmak A, Yilmaz TF, Toprak H, Yesil G, Alkan A. MRI and MRS findings in fucosidosis; a rare lysosomal storage disease. Brain & development. 2016;38(4):435-8. 378. Willems PJ, Gatti R, Darby JK, Romeo G, Durand P, Dumon JE, et al. Fucosidosis revisited: a review of 77 patients. Am J Med Genet. 1991;38(1):111-31. 379. Malatt C, Koning JL, Naheedy J. Skeletal and Brain Abnormalities in Fucosidosis, a Rare Lysosomal Storage Disorder. Journal of radiology case reports. 2015;9(5):30-8. 380. Lakics V, Karran EH, Boess FG. Quantitative comparison of phosphodiesterase mRNA distribution in human brain and peripheral tissues. Neuropharmacology. 2010;59(6):367-74. 381. Soderling SH, Beavo JA. Regulation of cAMP and cGMP signaling: new phosphodiesterases and new functions. Current opinion in cell biology. 2000;12(2):174-9. 382. Gross-Langenhoff M, Hofbauer K, Weber J, Schultz A, Schultz JE. cAMP is a ligand for the tandem GAF domain of human phosphodiesterase 10 and cGMP for the tandem GAF domain of phosphodiesterase 11. J Biol Chem. 2006;281(5):2841-6. 383. Heikaus CC, Pandit J, Klevit RE. Cyclic nucleotide binding GAF domains from phosphodiesterases: structural and mechanistic insights. Structure. 2009;17(12):1551-7. 384. Jager R, Russwurm C, Schwede F, Genieser HG, Koesling D, Russwurm M. Activation of PDE10 and PDE11 phosphodiesterases. J Biol Chem. 2012;287(2):1210-9. 385. Diggle CP, Sukoff Rizzo SJ, Popiolek M, Hinttala R, Schulke JP, Kurian MA, et al. Biallelic Mutations in PDE10A Lead to Loss of Striatal PDE10A and a Hyperkinetic Movement Disorder with Onset in Infancy. Am J Hum Genet. 2016;98(4):735-43. 386. Wilson LS, Brandon NJ. Emerging biology of PDE10A. Current pharmaceutical design. 2015;21(3):378-88. 387. Di Fonzo A, Wu-Chou YH, Lu CS, van Doeselaar M, Simons EJ, Rohe CF, et al. A common missense variant in the LRRK2 gene, Gly2385Arg, associated with Parkinson's disease risk in Taiwan. Neurogenetics. 2006;7(3):133-8. 273

388. Tan EK, Lim HQ, Yuen Y, Zhao Y. Pathogenicity of LRRK2 P755L variant in Parkinson's disease. Mov Disord. 2008;23(5):734-6. 389. Tomiyama H, Mizuta I, Li Y, Funayama M, Yoshino H, Li L, et al. LRRK2 P755L variant in sporadic Parkinson's disease. Journal of human genetics. 2008;53(11-12):1012-5. 390. Lesage S, Condroyer C, Lannuzel A, Lohmann E, Troiano A, Tison F, et al. Molecular analyses of the LRRK2 gene in European and North African autosomal dominant Parkinson's disease. J Med Genet. 2009;46(7):458-64. 391. Nuytemans K, Rademakers R, Theuns J, Pals P, Engelborghs S, Pickut B, et al. Founder mutation p.R1441C in the leucine-rich repeat kinase 2 gene in Belgian Parkinson's disease patients. Eur J Hum Genet. 2008;16(4):471-9. 392. Kircher M, Witten DM, Jain P, O'Roak BJ, Cooper GM, Shendure J. A general framework for estimating the relative pathogenicity of human genetic variants. Nat Genet. 2014;46(3):310-5. 393. Du J, Hu YC, Tang BS, Chen C, Luo YY, Zhan ZX, et al. Expansion of the phenotypic spectrum of SPG6 caused by mutation in NIPA1. Clin Neurol Neurosurg. 2011;113(6):480-2. 394. Rainier S, Chai JH, Tokarz D, Nicholls RD, Fink JK. NIPA1 gene mutations cause autosomal dominant hereditary spastic paraplegia (SPG6). Am J Hum Genet. 2003;73(4):967-71. 395. Reed JA, Wilkinson PA, Patel H, Simpson MA, Chatonnet A, Robay D, et al. A novel NIPA1 mutation associated with a pure form of autosomal dominant hereditary spastic paraplegia. Neurogenetics. 2005;6(2):79-84. 396. Svenstrup K, Moller RS, Christensen J, Budtz-Jorgensen E, Gilling M, Nielsen JE. NIPA1 mutation in complex hereditary spastic paraplegia with epilepsy. Eur J Neurol. 2011;18(9):1197-9. 397. Huppke P, Brendel C, Kalscheuer V, Korenke GC, Marquardt I, Freisinger P, et al. Mutations in SLC33A1 cause a lethal autosomal-recessive disorder with congenital cataracts, hearing loss, and low serum copper and ceruloplasmin. Am J Hum Genet. 2012;90(1):61-8. 398. Lin P, Li J, Liu Q, Mao F, Li J, Qiu R, et al. A missense mutation in SLC33A1, which encodes the acetyl-CoA transporter, causes autosomal-dominant spastic paraplegia (SPG42). Am J Hum Genet. 2008;83(6):752-9. 399. Venco P, Dusi S, Valletta L, Tiranti V. Alteration of the coenzyme A biosynthetic pathway in neurodegeneration with brain iron accumulation syndromes. Biochemical Society transactions. 2014;42(4):1069-74. 400. Gu SM, Orth U, Veske A, Enders H, Klunder K, Schlosser M, et al. Five novel mutations in the L1CAM gene in families with X linked hydrocephalus. J Med Genet. 1996;33(2):103-6. 401. Jouet M, Rosenthal A, Armstrong G, MacFarlane J, Stevenson R, Paterson J, et al. X-linked spastic paraplegia (SPG1), MASA syndrome and X-linked hydrocephalus result from mutations in the L1 gene. Nat Genet. 1994;7(3):402-7. 402. Hirokawa N, Noda Y, Tanaka Y, Niwa S. Kinesin superfamily motor proteins and intracellular transport. Nature reviews Molecular cell biology. 2009;10(10):682-96. 403. Lo KY, Kuzmin A, Unger SM, Petersen JD, Silverman MA. KIF1A is the primary anterograde motor protein required for the axonal transport of dense-core vesicles in cultured hippocampal neurons. Neuroscience letters. 2011;491(3):168-73. 404. Erlich Y, Edvardson S, Hodges E, Zenvirt S, Thekkat P, Shaag A, et al. Exome sequencing and disease-network analysis of a single family implicate a mutation in KIF1A in hereditary spastic paraparesis. Genome research. 2011;21(5):658-64. 405. Hamdan FF, Gauthier J, Araki Y, Lin DT, Yoshizawa Y, Higashi K, et al. Excess of de novo deleterious mutations in genes associated with glutamatergic systems in nonsyndromic intellectual disability. Am J Hum Genet. 2011;88(3):306-16. 406. Riviere JB, Ramalingam S, Lavastre V, Shekarabi M, Holbert S, Lafontaine J, et al. KIF1A, an axonal transporter of synaptic vesicles, is mutated in hereditary sensory and autonomic neuropathy type 2. Am J Hum Genet. 2011;89(2):219-30. 407. Gasser T. Inherited myoclonus-dystonia syndrome. Advances in neurology. 1998;78:325-34. 408. Gerrits MC, Foncke EM, de Haan R, Hedrich K, van de Leemput YL, Baas F, et al. Phenotype- genotype correlation in Dutch patients with myoclonus-dystonia. Neurology. 2006;66(5):759-61. 409. Lohmann K, Klein C. Genetics of dystonia: what's known? What's new? What's next? Mov Disord. 2013;28(7):899-905. 410. Zimprich A, Grabowski M, Asmus F, Naumann M, Berg D, Bertram M, et al. Mutations in the gene encoding epsilon-sarcoglycan cause myoclonus-dystonia syndrome. Nat Genet. 2001;29(1):66-9. 411. Grunewald A, Djarmati A, Lohmann-Hedrich K, Farrell K, Zeller JA, Allert N, et al. Myoclonus-dystonia: significance of large SGCE deletions. Hum Mutat. 2008;29(2):331-2. 412. Blanchard A, Ea V, Roubertie A, Martin M, Coquart C, Claustres M, et al. DYT6 dystonia: review of the literature and creation of the UMD Locus-Specific Database (LSDB) for mutations in the THAP1 gene. Hum Mutat. 2011;32(11):1213-24.

274

413. Chartier-Harlin MC, Dachsel JC, Vilarino-Guell C, Lincoln SJ, Lepretre F, Hulihan MM, et al. Translation initiator EIF4G1 mutations in familial Parkinson disease. Am J Hum Genet. 2011;89(3):398- 406. 414. Nichols N, Bras JM, Hernandez DG, Jansen IE, Lesage S, Lubbe S, et al. EIF4G1 mutations do not cause Parkinson's disease. Neurobiol Aging. 2015;36(8):2444 e1-4. 415. Ross OA, Soto-Ortolaza AI, Heckman MG, Aasly JO, Abahuni N, Annesi G, et al. Association of LRRK2 exonic variants with susceptibility to Parkinson's disease: a case-control study. Lancet Neurol. 2011;10(10):898-908. 416. Seifert W, Holder-Espinasse M, Spranger S, Hoeltzenbein M, Rossier E, Dollfus H, et al. Mutational spectrum of COH1 and clinical heterogeneity in Cohen syndrome. J Med Genet. 2006;43(5):e22. 417. Bettencourt C, Forabosco P, Wiethoff S, Heidari M, Johnstone DM, Botia JA, et al. Gene coexpression networks shed light into diseases of brain iron accumulation. Neurobiology of disease. 2016;87:59-68. 418. Lee S, Emond MJ, Bamshad MJ, Barnes KC, Rieder MJ, Nickerson DA, et al. Optimal unified approach for rare-variant association testing with application to small-sample case-control whole-exome sequencing studies. Am J Hum Genet. 2012;91(2):224-37. 419. Ho M, Chelly J, Carter N, Danek A, Crocker P, Monaco AP. Isolation of the gene for McLeod syndrome that encodes a novel membrane transport protein. Cell. 1994;77(6):869-80. 420. Allen FH, Jr., Krabbe SM, Corcoran PA. A new phenotype (McLeod) in the Kell blood-group system. Vox sanguinis. 1961;6:555-60. 421. Symmans WA, Shepherd CS, Marsh WL, Oyen R, Shohet SB, Linehan BJ. Hereditary acanthocytosis associated with the McLeod phenotype of the Kell blood group system. British journal of haematology. 1979;42(4):575-83. 422. Wimer BM, Marsh WL, Taswell HF, Galey WR. Haematological changes associated with the McLeod phenotype of the Kell blood group system. British journal of haematology. 1977;36(2):219-24. 423. Marsh WL. The Kell blood group, Kx antigen, and chronic granulomatous disease. Mayo Clinic proceedings Mayo Clinic. 1977;52(3):150-2. 424. Marsh WL, Marsh NJ, Moore A, Symmans WA, Johnson CL, Redman CM. Elevated serum creatine phosphokinase in subjects with McLeod syndrome. Vox sanguinis. 1981;40(6):403-11. 425. Swash M, Schwartz MS, Carter ND, Heath R, Leak M, Rogers KL. Benign X-linked myopathy with acanthocytes (McLeod syndrome). Its relationship to X-linked muscular dystrophy. Brain. 1983;106 (Pt 3):717-33. 426. Witt TN, Danek A, Reiter M, Heim MU, Dirschinger J, Olsen EG. McLeod syndrome: a distinct form of neuroacanthocytosis. Report of two cases and literature review with emphasis on neuromuscular manifestations. J Neurol. 1992;239(6):302-6. 427. Jung HH, Danek A, Walker RH. Neuroacanthocytosis syndromes. Orphanet journal of rare diseases. 2011;6:68. 428. Danek A, Rubio JP, Rampoldi L, Ho M, Dobson-Stone C, Tison F, et al. McLeod neuroacanthocytosis: genotype and phenotype. Ann Neurol. 2001;50(6):755-64. 429. Stanfield GM, Horvitz HR. The ced-8 gene controls the timing of programmed cell deaths in C. elegans. Molecular cell. 2000;5(3):423-33. 430. Wiethoff S, Xiromerisiou G, Bettencourt C, Kioumi A, Tsiptsios I, Tychalas A, et al. Novel single base-pair deletion in exon 1 of XK gene leading to McLeod syndrome with chorea, muscle wasting, peripheral neuropathy, acanthocytosis and haemolysis. J Neurol Sci. 2014;339(1-2):220-2. 431. Man BL, Yuen YP, Yip SF, Ng SH. The first case report of McLeod syndrome in a Chinese patient. BMJ case reports. 2013;2013. 432. Jung HH, Danek A, Frey BM. McLeod syndrome: a neurohaematological disorder. Vox sanguinis. 2007;93(2):112-21. 433. Anheim M, Tranchant C, Koenig M. The autosomal recessive cerebellar ataxias. N Engl J Med. 2012;366(7):636-46. 434. Gros-Louis F, Dupre N, Dion P, Fox MA, Laurent S, Verreault S, et al. Mutations in SYNE1 lead to a newly discovered form of autosomal recessive cerebellar ataxia. Nat Genet. 2007;39(1):80-5. 435. Synofzik M, Smets K, Mallaret M, Di Bella D, Gallenmuller C, Baets J, et al. SYNE1 ataxia is a common recessive ataxia with major non-cerebellar features: a large scale multi-centre study. Brain. 2016. 436. Noreau A, Bourassa CV, Szuto A, Levert A, Dobrzeniecka S, Gauthier J, et al. SYNE1 mutations in autosomal recessive cerebellar ataxia. JAMA neurology. 2013;70(10):1296-31. 437. Izumi Y, Miyamoto R, Morino H, Yoshizawa A, Nishinaka K, Udaka F, et al. Cerebellar ataxia with SYNE1 mutation accompanying motor neuron disease. Neurology. 2013;80(6):600-1. 438. Ozoguz A, Uyan O, Birdal G, Iskender C, Kartal E, Lahut S, et al. The distinct genetic pattern of ALS in Turkey and novel mutations. Neurobiol Aging. 2015;36(4):1764 e9-18.

275

439. Zhang Q, Bethmann C, Worth NF, Davies JD, Wasner C, Feuer A, et al. Nesprin-1 and -2 are involved in the pathogenesis of Emery Dreifuss muscular dystrophy and are critical for nuclear envelope integrity. Hum Mol Genet. 2007;16(23):2816-33. 440. Attali R, Warwar N, Israel A, Gurt I, McNally E, Puckelwartz M, et al. Mutation of SYNE-1, encoding an essential component of the nuclear lamina, is responsible for autosomal recessive arthrogryposis. Hum Mol Genet. 2009;18(18):3462-9. 441. Schuurs-Hoeijmakers JH, Vulto-van Silfhout AT, Vissers LE, van de V, II, van Bon BW, de Ligt J, et al. Identification of pathogenic gene variants in small families with intellectually disabled siblings by exome sequencing. J Med Genet. 2013;50(12):802-11. 442. Yu TW, Chahrour MH, Coulter ME, Jiralerspong S, Okamura-Ikeda K, Ataman B, et al. Using whole-exome sequencing to identify inherited causes of autism. Neuron. 2013;77(2):259-73. 443. Zhang Q, Ragnauth C, Greener MJ, Shanahan CM, Roberts RG. The nesprins are giant actin- binding proteins, orthologous to Drosophila melanogaster muscle protein MSP-300. Genomics. 2002;80(5):473-81. 444. Zhang Q, Skepper JN, Yang F, Davies JD, Hegyi L, Roberts RG, et al. Nesprins: a novel family of spectrin-repeat-containing proteins that localize to the nuclear membrane in multiple tissues. Journal of cell science. 2001;114(Pt 24):4485-98. 445. Zhang X, Lei K, Yuan X, Wu X, Zhuang Y, Xu T, et al. SUN1/2 and Syne/Nesprin-1/2 complexes connect centrosome to the nucleus during neurogenesis and neuronal migration in mice. Neuron. 2009;64(2):173-87. 446. Laforce R, Jr., Buteau JP, Bouchard JP, Rouleau GA, Bouchard RW, Dupre N. Cognitive impairment in ARCA-1, a newly discovered pure cerebellar ataxia syndrome. Cerebellum. 2010;9(3):443- 53. 447. Mancuso M, Orsucci D, Siciliano G, Bonuccelli U. The genetics of ataxia: through the labyrinth of the Minotaur, looking for Ariadne's thread. J Neurol. 2014;261 Suppl 2:S528-41. 448. Foo JN, Liu JJ, Tan EK. Whole-genome and whole-exome sequencing in neurological diseases. Nat Rev Neurol. 2012;8(9):508-17. 449. Jiang T, Tan MS, Tan L, Yu JT. Application of next-generation sequencing technologies in Neurology. Annals of translational medicine. 2014;2(12):125. 450. Pyle A, Smertenko T, Bargiela D, Griffin H, Duff J, Appleton M, et al. Exome sequencing in undiagnosed inherited and sporadic ataxias. Brain. 2015;138(Pt 2):276-83. 451. Rabbani B, Mahdieh N, Hosomichi K, Nakaoka H, Inoue I. Next-generation sequencing: impact of exome sequencing in characterizing Mendelian disorders. Journal of human genetics. 2012;57(10):621- 32. 452. Fogel BL, Lee H, Deignan JL, Strom SP, Kantarci S, Wang X, et al. Exome sequencing in the clinical diagnosis of sporadic or familial cerebellar ataxia. JAMA neurology. 2014;71(10):1237-46. 453. Nilius B, Voets T. The puzzle of TRPV4 channelopathies. EMBO reports. 2013;14(2):152-63. 454. Rainier S, Bui M, Mark E, Thomas D, Tokarz D, Ming L, et al. Neuropathy target esterase gene mutations cause motor neuron disease. Am J Hum Genet. 2008;82(3):780-5. 455. Synofzik M, Gonzalez MA, Lourenco CM, Coutelier M, Haack TB, Rebelo A, et al. PNPLA6 mutations cause Boucher-Neuhauser and Gordon Holmes syndromes as part of a broad neurodegenerative spectrum. Brain. 2014;137(Pt 1):69-77. 456. Kmoch S, Majewski J, Ramamurthy V, Cao S, Fahiminiya S, Ren H, et al. Mutations in PNPLA6 are linked to photoreceptor degeneration and various forms of childhood blindness. Nature communications. 2015;6:5614. 457. Hufnagel RB, Arno G, Hein ND, Hersheson J, Prasad M, Anderson Y, et al. Neuropathy target esterase impairments cause Oliver-McFarlane and Laurence-Moon syndromes. J Med Genet. 2015;52(2):85-94. 458. Atkins J, Luthjens LH, Hom ML, Glynn P. Monomers of the catalytic domain of human neuropathy target esterase are active in the presence of phospholipid. The Biochemical journal. 2002;361(Pt 1):119-23. 459. van Tienhoven M, Atkins J, Li Y, Glynn P. Human neuropathy target esterase catalyzes hydrolysis of membrane lipids. J Biol Chem. 2002;277(23):20942-8. 460. Tesson C, Nawara M, Salih MA, Rossignol R, Zaki MS, Al Balwi M, et al. Alteration of fatty- acid-metabolizing enzymes affects mitochondrial form and function in hereditary spastic paraplegia. Am J Hum Genet. 2012;91(6):1051-64. 461. Adibhatla RM, Hatcher JF. Altered lipid metabolism in brain injury and disorders. Sub-cellular biochemistry. 2008;49:241-68. 462. Yadav RS, Tiwari NK. Lipid integration in neurodegeneration: an overview of Alzheimer's disease. Molecular neurobiology. 2014;50(1):168-76. 463. Boukhris A, Schule R, Loureiro JL, Lourenco CM, Mundwiller E, Gonzalez MA, et al. Alteration of ganglioside biosynthesis responsible for complex hereditary spastic paraplegia. Am J Hum Genet. 2013;93(1):118-23. 276

464. Wiethoff S, Bettencourt C, Paudel R, Madon P, Liu YT, Hersheson J, et al. Pure Cerebellar Ataxia with Homozygous Mutations in the PNPLA6 Gene. Cerebellum. 2016. 465. Kumar P, Henikoff S, Ng PC. Predicting the effects of coding non-synonymous variants on protein function using the SIFT algorithm. Nature protocols. 2009;4(7):1073-81. 466. Strickland JA, Orr GL, Walsh TA. Inhibition of Diabrotica Larval Growth by Patatin, the Lipid Acyl Hydrolase from Potato Tubers. Plant physiology. 1995;109(2):667-74. 467. Zaccheo O, Dinsdale D, Meacock PA, Glynn P. Neuropathy target esterase and its yeast homologue degrade phosphatidylcholine to glycerophosphocholine in living cells. J Biol Chem. 2004;279(23):24024-33. 468. Morgan JP, Penovich P. Jamaica ginger paralysis. Forty-seven-year follow-up. Arch Neurol. 1978;35(8):530-2. 469. Sorokin M. Orthocresyl phosphate neuropathy: report of an outbreak in Fiji. The Medical journal of Australia. 1969;1(10):506-8. 470. Woolf AD. Ginger Jake and the blues: a tragic song of poisoning. Veterinary and human toxicology. 1995;37(3):252-4. 471. Hou WY, Long DX, Wu YJ. Effect of inhibition of neuropathy target esterase in mouse nervous tissues in vitro on phosphatidylcholine and lysophosphatidylcholine homeostasis. International journal of toxicology. 2009;28(5):417-24. 472. Richardson RJ, Hein ND, Wijeyesakere SJ, Fink JK, Makhaeva GF. Neuropathy target esterase (NTE): overview and future. Chemico-biological interactions. 2013;203(1):238-44. 473. Tanaka D, Jr., Bursian SJ, Lehning E. Selective axonal and terminal degeneration in the chicken brainstem and cerebellum following exposure to bis(1-methylethyl)phosphorofluoridate (DFP). Brain research. 1990;519(1-2):200-8. 474. Tarnutzer AA, Gerth-Kahlert C, Timmann D, Chang DI, Harmuth F, Bauer P, et al. Boucher- Neuhauser syndrome: cerebellar degeneration, chorioretinal dystrophy and hypogonadotropic hypogonadism: two novel cases and a review of 40 cases from the literature. J Neurol. 2015;262(1):194- 202. 475. Deik A, Johannes B, Rucker JC, Sanchez E, Brodie SE, Deegan E, et al. Compound heterozygous PNPLA6 mutations cause Boucher-Neuhauser syndrome with late-onset ataxia. J Neurol. 2014;261(12):2411-23. 476. Koh K, Kobayashi F, Miwa M, Shindo K, Isozaki E, Ishiura H, et al. Novel mutations in the PNPLA6 gene in Boucher-Neuhauser syndrome. Journal of human genetics. 2015. 477. Boucher BJ, Gibberd FB. Familial ataxia, hypogonadism and retinal degeneration. Acta neurologica Scandinavica. 1969;45(4):507-10. 478. Lander ES, Linton LM, Birren B, Nusbaum C, Zody MC, Baldwin J, et al. Initial sequencing and analysis of the human genome. Nature. 2001;409(6822):860-921. 479. Richards RI, Sutherland GR. Dynamic mutations: a new class of mutations causing human disease. Cell. 1992;70(5):709-12. 480. Tezenas du Montcel S, Durr A, Bauer P, Figueroa KP, Ichikawa Y, Brussino A, et al. Modulation of the age at onset in spinocerebellar ataxia by CAG tracts in various genes. Brain : a journal of neurology. 2014;137(Pt 9):2444-55. 481. Wexler NS, Lorimer J, Porter J, Gomez F, Moskowitz C, Shackell E, et al. Venezuelan kindreds reveal that genetic and environmental factors modulate Huntington's disease age of onset. Proceedings of the National Academy of Sciences of the United States of America. 101. United States2004. p. 3498-503. 482. van de Warrenburg BP, Hendriks H, Durr A, van Zuijlen MC, Stevanin G, Camuzat A, et al. Age at onset variance analysis in spinocerebellar ataxias: a study in a Dutch-French cohort. Annals of neurology. 2005;57(4):505-12. 483. Bettencourt C, Raposo M, Kazachkova N, Cymbron T, Santos C, Kay T, et al. The APOE epsilon2 allele increases the risk of earlier age at onset in Machado-Joseph disease. Archives of neurology. 2011;68(12):1580-3. 484. Raposo M, Ramos A, Bettencourt C, Lima M. Replicating studies of genetic modifiers in spinocerebellar ataxia type 3: can homogeneous cohorts aid? Brain : a journal of neurology. 2015. 485. Peng H, Wang C, Chen Z, Sun Z, Jiao B, Li K, et al. APOE epsilon2 allele may decrease the age at onset in patients with spinocerebellar ataxia type 3 or Machado-Joseph disease from the Chinese Han population. Neurobiology of aging. 2014;35(9):2179 e15-8. 486. Lopez Castel A, Cleary JD, Pearson CE. Repeat instability as the basis for human diseases and as a potential target for therapy. Nature reviews Molecular cell biology. 2010;11(3):165-70. 487. Iyer RR, Pluciennik A, Napierala M, Wells RD. DNA triplet repeat expansion and mismatch repair. Annual review of biochemistry. 2015;84:199-226. 488. Gomes-Pereira M, Monckton DG. Chemical modifiers of unstable expanded simple sequence repeats: what goes up, could come down. Mutation research. 2006;598(1-2):15-34.

277

489. Gonitel R, Moffitt H, Sathasivam K, Woodman B, Detloff PJ, Faull RL, et al. DNA instability in postmitotic neurons. Proceedings of the National Academy of Sciences of the United States of America. 2008;105(9):3467-72. 490. Swami M, Hendricks AE, Gillis T, Massood T, Mysore J, Myers RH, et al. Somatic expansion of the Huntington's disease CAG repeat in the brain is associated with an earlier age of disease onset. Human molecular genetics. 2009;18(16):3039-47. 491. Dragileva E, Hendricks A, Teed A, Gillis T, Lopez ET, Friedberg EC, et al. Intergenerational and striatal CAG repeat instability in Huntington's disease knock-in mice involve different DNA repair genes. Neurobiol Dis. 2009;33(1):37-47. 492. Budworth H, Harris FR, Williams P, Lee do Y, Holt A, Pahnke J, et al. Suppression of Somatic Expansion Delays the Onset of Pathophysiology in a Mouse Model of Huntington's Disease. PLoS genetics. 2015;11(8):e1005267. 493. Mason AG, Tome S, Simard JP, Libby RT, Bammler TK, Beyer RP, et al. Expression levels of DNA replication and repair genes predict regional somatic repeat instability in the brain but are not altered by polyglutamine disease protein expression or age. Human molecular genetics. 2014;23(6):1606- 18. 494. Pearson CE, Nichol Edamura K, Cleary JD. Repeat instability: mechanisms of dynamic mutations. Nature reviews Genetics. 2005;6(10):729-42. 495. Storey E. Genetic cerebellar ataxias. Semin Neurol. 2014;34(3):280-92. 496. Ambrose M, Gatti RA. Pathogenesis of ataxia-telangiectasia: the next generation of ATM functions. Blood. 2013;121(20):4036-45. 497. Bras J, Alonso I, Barbot C, Costa MM, Darwent L, Orme T, et al. Mutations in PNKP cause recessive ataxia with oculomotor apraxia type 4. American journal of human genetics. 2015;96(3):474-9. 498. Harris JL, Jakob B, Taucher-Scholz G, Dianov GL, Becherel OJ, Lavin MF. Aprataxin, poly- ADP ribose polymerase 1 (PARP-1) and apurinic endonuclease 1 (APE1) function together to protect the genome against oxidative damage. Human molecular genetics. 2009;18(21):4102-17. 499. Takashima H, Boerkoel CF, John J, Saifi GM, Salih MA, Armstrong D, et al. Mutation of TDP1, encoding a topoisomerase I-dependent DNA damage repair enzyme, in spinocerebellar ataxia with axonal neuropathy. Nat Genet. 2002;32(2):267-72. 500. Shiloh Y, Ziv Y. The ATM protein kinase: regulating the cellular response to genotoxic stress, and more. Nature reviews Molecular cell biology. 2013;14(4):197-210. 501. Mirkin SM. Expandable DNA repeats and human disease. Nature. 2007;447(7147):932-40. 502. Wheeler VC, Lebel LA, Vrbanac V, Teed A, te Riele H, MacDonald ME. Mismatch repair gene Msh2 modifies the timing of early disease in Hdh(Q111) striatum. Human molecular genetics. 2003;12(3):273-81. 503. Tome S, Manley K, Simard JP, Clark GW, Slean MM, Swami M, et al. MSH3 polymorphisms and protein levels affect CAG repeat instability in Huntington's disease mice. PLoS Genet. 2013;9(2):e1003280. 504. Kovtun IV, Liu Y, Bjoras M, Klungland A, Wilson SH, McMurray CT. OGG1 initiates age- dependent CAG trinucleotide expansion in somatic cells. Nature. 2007;447(7143):447-52. 505. Goula AV, Berquist BR, Wilson DM, 3rd, Wheeler VC, Trottier Y, Merienne K. Stoichiometry of base excision repair proteins correlates with increased somatic CAG instability in striatum over cerebellum in Huntington's disease transgenic mice. PLoS Genet. 2009;5(12):e1000749. 506. Menon RP, Nethisinghe S, Faggiano S, Vannocci T, Rezaei H, Pemble S, et al. The role of interruptions in polyQ in the pathology of SCA1. Plos Genet. 2013;9(7):e1003648. 507. Farmer H, McCabe N, Lord CJ, Tutt AN, Johnson DA, Richardson TB, et al. Targeting the DNA repair defect in BRCA mutant cells as a therapeutic strategy. Nature. 2005;434(7035):917-21. 508. Jackson SE, Chester JD. Personalised cancer medicine. Int J Cancer. 2015;137(2):262-6. 509. Kordasiewicz HB, Thompson RM, Clark HB, Gomez CM. C-termini of P/Q-type Ca2+ channel alpha1A subunits translocate to nuclei and promote polyglutamine-mediated toxicity. Hum Mol Genet. 2006;15(10):1587-99. 510. Zhuchenko O, Bailey J, Bonnen P, Ashizawa T, Stockton DW, Amos C, et al. Autosomal dominant cerebellar ataxia (SCA6) associated with small polyglutamine expansions in the alpha 1A- voltage-dependent calcium channel. Nat Genet. 1997;15(1):62-9. 511. Giunti P, Mantuano E, Frontali M, Veneziano L. Molecular mechanism of Spinocerebellar Ataxia type 6: glutamine repeat disorder, channelopathy and transcriptional dysregulation. The multifaceted aspects of a single mutation. Frontiers in cellular neuroscience. 2015;9:36. 512. Catterall WA. Signaling complexes of voltage-gated sodium and calcium channels. Neuroscience letters. 2010;486(2):107-16. 513. Jodice C, Mantuano E, Veneziano L, Trettel F, Sabbadini G, Calandriello L, et al. Episodic ataxia type 2 (EA2) and spinocerebellar ataxia type 6 (SCA6) due to CAG repeat expansion in the CACNA1A gene on chromosome 19p. Hum Mol Genet. 1997;6(11):1973-8.

278

514. Yabe I, Sasaki H, Yamashita I, Takei A, Tashiro K. Clinical trial of acetazolamide in SCA6, with assessment using the Ataxia Rating Scale and body stabilometry. Acta neurologica Scandinavica. 2001;104(1):44-7. 515. Brenman LM. Spinocerebellar Ataxia type 6 (SCA6) phenotype in a patient with an intermediate mutation range CACNA1A allele. J Neurol Neurophysiol. 2013;4:144(doi: 10.4172/2155-9562.1000144). 516. Mariotti C, Gellera C, Grisoli M, Mineri R, Castucci A, Di Donato S. Pathogenic effect of an intermediate-size SCA-6 allele (CAG)(19) in a homozygous patient. Neurology. 2001;57(8):1502-4. 517. Ishikawa K, Tanaka H, Saito M, Ohkoshi N, Fujita T, Yoshizawa K, et al. Japanese families with autosomal dominant pure cerebellar ataxia map to chromosome 19p13.1-p13.2 and are strongly associated with mild CAG expansions in the spinocerebellar ataxia type 6 gene in chromosome 19p13.1. Am J Hum Genet. 1997;61(2):336-46. 518. Maruyama H, Izumi Y, Morino H, Oda M, Toji H, Nakamura S, et al. Difference in disease-free survival curve and regional distribution according to subtype of spinocerebellar ataxia: a study of 1,286 Japanese patients. Am J Med Genet. 2002;114(5):578-83. 519. Takahashi H, Ishikawa K, Tsutsumi T, Fujigasaki H, Kawata A, Okiyama R, et al. A clinical and genetic study in a large cohort of patients with spinocerebellar ataxia type 6. Journal of human genetics. 2004;49(5):256-64. 520. Chung MY, Ranum LP, Duvick LA, Servadio A, Zoghbi HY, Orr HT. Evidence for a mechanism predisposing to intergenerational CAG repeat instability in spinocerebellar ataxia type I. Nat Genet. 1993;5(3):254-8. 521. Matsuyama Z, Izumi Y, Kameyama M, Kawakami H, Nakamura S. The effect of CAT trinucleotide interruptions on the age at onset of spinocerebellar ataxia type 1 (SCA1). J Med Genet. 1999;36(7):546-8. 522. Sobczak K, Krzyzosiak WJ. Patterns of CAG repeat interruptions in SCA1 and SCA2 genes in relation to repeat instability. Hum Mutat. 2004;24(3):236-47. 523. Ishiguro T, Ishikawa K, Takahashi M, Obayashi M, Amino T, Sato N, et al. The carboxy- terminal fragment of alpha(1A) calcium channel preferentially aggregates in the cytoplasm of human spinocerebellar ataxia type 6 Purkinje cells. Acta Neuropathol. 2010;119(4):447-64. 524. Takahashi M, Obayashi M, Ishiguro T, Sato N, Niimi Y, Ozaki K, et al. Cytoplasmic location of alpha1A voltage-gated calcium channel C-terminal fragment (Cav2.1-CTF) aggregate is sufficient to cause cell death. PloS one. 2013;8(3):e50121. 525. Fratta P, Collins T, Pemble S, Nethisinghe S, Devoy A, Giunti P, et al. Sequencing analysis of the spinal bulbar muscular atrophy CAG expansion reveals absence of repeat interruptions. Neurobiol Aging. 2014;35(2):443 e1-3. 526. Synofzik M, Beetz C, Bauer C, Bonin M, Sanchez-Ferrero E, Schmitz-Hubsch T, et al. Spinocerebellar ataxia type 15: diagnostic assessment, frequency, and phenotypic features. J Med Genet. 2011;48(6):407-12. 527. Castrioto A, Prontera P, Di Gregorio E, Rossi V, Parnetti L, Rossi A, et al. A novel spinocerebellar ataxia type 15 family with involuntary movements and cognitive decline. Eur J Neurol. 2011;18(10):1263-5. 528. Di Gregorio E, Orsi L, Godani M, Vaula G, Jensen S, Salmon E, et al. Two Italian families with ITPR1 gene deletion presenting a broader phenotype of SCA15. Cerebellum. 2010;9(1):115-23. 529. Novak MJ, Sweeney MG, Li A, Treacy C, Chandrashekar HS, Giunti P, et al. An ITPR1 gene deletion causes spinocerebellar ataxia 15/16: a genetic, clinical and radiological description. Mov Disord. 2010;25(13):2176-82. 530. van de Leemput J, Chandran J, Knight MA, Holtzclaw LA, Scholz S, Cookson MR, et al. Deletion at ITPR1 underlies ataxia in mice and spinocerebellar ataxia 15 in humans. Plos Genet. 2007;3(6):e108. 531. Barresi S, Niceta M, Alfieri P, Brankovich V, Piccini G, Bruselles A, et al. Mutations in the IRBIT domain of ITPR1 are a frequent cause of autosomal dominant nonprogressive congenital ataxia. Clin Genet. 2016. 532. Huang L, Chardon JW, Carter MT, Friend KL, Dudding TE, Schwartzentruber J, et al. Missense mutations in ITPR1 cause autosomal dominant congenital nonprogressive spinocerebellar ataxia. Orphanet journal of rare diseases. 2012;7:67. 533. Gerber S, Alzayady KJ, Burglen L, Bremond-Gignac D, Marchesin V, Roche O, et al. Recessive and Dominant De Novo ITPR1 Mutations Cause Gillespie Syndrome. Am J Hum Genet. 2016. 534. McEntagart M, Williamson KA, Rainger JK, Wheeler A, Seawright A, De Baere E, et al. A Restricted Repertoire of De Novo Mutations in ITPR1 Cause Gillespie Syndrome with Evidence for Dominant-Negative Effect. Am J Hum Genet. 2016. 535. Sugawara T, Hisatsune C, Le TD, Hashikawa T, Hirono M, Hattori M, et al. Type 1 inositol trisphosphate receptor regulates cerebellar circuits by maintaining the spine morphology of purkinje cells in adult mice. J Neurosci. 2013;33(30):12186-96.

279

536. Taylor CW, Genazzani AA, Morris SA. Expression of inositol trisphosphate receptors. Cell calcium. 1999;26(6):237-51. 537. Berridge MJ. Neuronal calcium signaling. Neuron. 1998;21(1):13-26. 538. Berridge MJ. Inositol trisphosphate and calcium signalling mechanisms. Biochim Biophys Acta. 2009;1793(6):933-40. 539. Greer PL, Greenberg ME. From synapse to nucleus: calcium-dependent gene transcription in the control of synapse development and function. Neuron. 2008;59(6):846-60. 540. Parekh AB, Muallem S. Ca(2+) signalling and gene regulation. Cell calcium. 2011;49(5):279. 541. Pinton P, Giorgi C, Siviero R, Zecchini E, Rizzuto R. Calcium and apoptosis: ER-mitochondria Ca2+ transfer in the control of apoptosis. Oncogene. 2008;27(50):6407-18. 542. Verkhratsky A, Petersen OH. The endoplasmic reticulum as an integrating signalling organelle: from neuronal signalling to neuronal death. European journal of pharmacology. 2002;447(2-3):141-54. 543. Wiegert JS, Bading H. Activity-dependent calcium signaling and ERK-MAP kinases in neurons: a link to structural plasticity of the nucleus and gene transcription regulation. Cell calcium. 2011;49(5):296-305. 544. Kasumu A, Bezprozvanny I. Deranged calcium signaling in Purkinje cells and pathogenesis in spinocerebellar ataxia 2 (SCA2) and other ataxias. Cerebellum. 2012;11(3):630-9. 545. Bezprozvanny I. Role of inositol 1,4,5-trisphosphate receptors in pathogenesis of Huntington's disease and spinocerebellar ataxias. Neurochemical research. 2011;36(7):1186-97. 546. Supnet C, Bezprozvanny I. The dysregulation of intracellular calcium in Alzheimer disease. Cell calcium. 2010;47(2):183-9. 547. Mak DO, Cheung KH, Toglia P, Foskett JK, Ullah G. Analyzing and Quantifying the Gain-of- Function Enhancement of IP3 Receptor Gating by Familial Alzheimer's Disease-Causing Mutants in Presenilins. PLoS computational biology. 2015;11(10):e1004529. 548. Brown SA, Loew LM. Computational analysis of calcium signaling and membrane electrophysiology in cerebellar Purkinje neurons associated with ataxia. BMC systems biology. 2012;6:70. 549. Brown SA, Loew LM. Integration of modeling with experimental and clinical findings synthesizes and refines the central role of inositol 1,4,5-trisphosphate receptor 1 in spinocerebellar ataxia. Frontiers in neuroscience. 2014;8:453. 550. Matsumoto M, Nakagawa T, Inoue T, Nagata E, Tanaka K, Takano H, et al. Ataxia and epileptic seizures in mice lacking type 1 inositol 1,4,5-trisphosphate receptor. Nature. 1996;379(6561):168-71. 551. Ogura H, Matsumoto M, Mikoshiba K. Motor discoordination in mutant mice heterozygous for the type 1 inositol 1,4,5-trisphosphate receptor. Behavioural brain research. 2001;122(2):215-9. 552. Street VA, Bosma MM, Demas VP, Regan MR, Lin DD, Robinson LC, et al. The type 1 inositol 1,4,5-trisphosphate receptor gene is altered in the opisthotonos mouse. J Neurosci. 1997;17(2):635-45. 553. Deng XY, Wang H, Wang T, Fang XT, Zou LL, Li ZY, et al. Non-viral methods for generating integration-free, induced pluripotent stem cells. Current stem cell research & therapy. 2015;10(2):153-8. 554. van de Leemput J, Boles NC, Kiehl TR, Corneo B, Lederman P, Menon V, et al. CORTECON: a temporal transcriptome analysis of in vitro human cerebral cortex development from human embryonic stem cells. Neuron. 2014;83(1):51-68. 555. Wan W, Cao L, Kalionis B, Xia S, Tai X. Applications of Induced Pluripotent Stem Cells in Studying the Neurodegenerative Diseases. Stem cells international. 2015;2015:382530. 556. Smyth JT, Hwang SY, Tomita T, DeHaven WI, Mercer JC, Putney JW. Activation and regulation of store-operated calcium entry. Journal of cellular and molecular medicine. 2010;14(10):2337-49. 557. Scemes E. Components of astrocytic intercellular calcium signaling. Molecular neurobiology. 2000;22(1-3):167-79. 558. Muller FJ, Schuldt BM, Williams R, Mason D, Altun G, Papapetrou EP, et al. A bioinformatic assay for pluripotency in human cells. Nature methods. 2011;8(4):315-7. 559. Rao MS, Malik N. Assessing iPSC reprogramming methods for their suitability in translational medicine. Journal of cellular biochemistry. 2012;113(10):3061-8. 560. Mertens J, Paquola AC, Ku M, Hatch E, Bohnke L, Ladjevardi S, et al. Directly Reprogrammed Human Neurons Retain Aging-Associated Transcriptomic Signatures and Reveal Age-Related Nucleocytoplasmic Defects. Cell stem cell. 2015;17(6):705-18. 561. Hockemeyer D, Jaenisch R. Induced Pluripotent Stem Cells Meet Genome Editing. Cell stem cell. 2016;18(5):573-86. 562. Hisatsune C, Miyamoto H, Hirono M, Yamaguchi N, Sugawara T, Ogawa N, et al. IP3R1 deficiency in the cerebellum/brainstem causes basal ganglia-independent dystonia by triggering tonic Purkinje cell firings in mice. Frontiers in neural circuits. 2013;7:156. 563. Zheng K, Bard L, Reynolds JP, King C, Jensen TP, Gourine AV, et al. Time-Resolved Imaging Reveals Heterogeneous Landscapes of Nanomolar Ca(2+) in Neurons and Astroglia. Neuron. 2015;88(2):277-88. 280

564. Imaizumi K, Sone T, Ibata K, Fujimori K, Yuzaki M, Akamatsu W, et al. Controlling the Regional Identity of hPSC-Derived Neurons to Uncover Neuronal Subtype Specificity of Neurological Disease Phenotypes. Stem cell reports. 2015;5(6):1010-22. 565. Kirkeby A, Grealish S, Wolf DA, Nelander J, Wood J, Lundblad M, et al. Generation of regionally specified neural progenitors and functional neurons from human embryonic stem cells under defined conditions. Cell reports. 2012;1(6):703-14. 566. Maury Y, Come J, Piskorowski RA, Salah-Mohellibi N, Chevaleyre V, Peschanski M, et al. Combinatorial analysis of developmental cues efficiently converts human pluripotent stem cells into multiple neuronal subtypes. Nature biotechnology. 2015;33(1):89-96. 567. Bailey K, Rahimi Balaei M, Mehdizadeh M, Marzban H. Spatial and temporal expression of lysosomal acid phosphatase 2 (ACP2) reveals dynamic patterning of the mouse cerebellar cortex. Cerebellum. 2013;12(6):870-81. 568. Hallonet M, Alvarado-Mallart RM. The chick/quail chimeric system: a model for early cerebellar development. Perspectives on developmental neurobiology. 1997;5(1):17-31. 569. Millen KJ, Hui CC, Joyner AL. A role for En-2 and other murine homologues of Drosophila segment polarity genes in regulating positional information in the developing cerebellum. Development. 1995;121(12):3935-45. 570. Crossley PH, Martinez S, Martin GR. Midbrain development induced by FGF8 in the chick embryo. Nature. 1996;380(6569):66-8. 571. Lee SM, Danielian PS, Fritzsch B, McMahon AP. Evidence that FGF8 signalling from the midbrain-hindbrain junction regulates growth and polarity in the developing midbrain. Development. 1997;124(5):959-69. 572. Liu A, Losos K, Joyner AL. FGF8 can activate Gbx2 and transform regions of the rostral mouse brain into a hindbrain fate. Development. 1999;126(21):4827-38. 573. Martinez S, Crossley PH, Cobos I, Rubenstein JL, Martin GR. FGF8 induces formation of an ectopic isthmic organizer and isthmocerebellar development via a repressive effect on Otx2 expression. Development. 1999;126(6):1189-200. 574. Sato T, Araki I, Nakamura H. Inductive signal and tissue responsiveness defining the tectum and the cerebellum. Development. 2001;128(13):2461-9. 575. Sunmonu NA, Li K, Guo Q, Li JY. Gbx2 and Fgf8 are sequentially required for formation of the midbrain-hindbrain compartment boundary. Development. 2011;138(4):725-34. 576. Danielian PS, McMahon AP. Engrailed-1 as a target of the Wnt-1 signalling pathway in vertebrate midbrain development. Nature. 1996;383(6598):332-4. 577. Panhuysen M, Vogt Weisenhorn DM, Blanquet V, Brodski C, Heinzmann U, Beisker W, et al. Effects of Wnt1 signaling on proliferation in the developing mid-/hindbrain region. Molecular and cellular neurosciences. 2004;26(1):101-11. 578. Butts T, Green MJ, Wingate RJ. Development of the cerebellum: simple steps to make a 'little brain'. Development. 2014;141(21):4031-41. 579. Lorenz A, Deutschmann M, Ahlfeld J, Prix C, Koch A, Smits R, et al. Severe alterations of cerebellar cortical development after constitutive activation of Wnt signaling in granule neuron precursors. Molecular and cellular biology. 2011;31(16):3326-38.

281

8 APPENDIX

Table 8-0 Non-exhaustive list of important genes implicated in neurodegenerative Mendelian disorders and their discovery year

Disease Group Gene Inheritance Discovery Comments APP AD 1991 - FAD PSEN-1 AD 1995 Most common PSEN-2 AD 1995 - SNCA AD 1997 - LRRK2 AD 2004 Most common PARK2 AR 1998 -

282 PD PINK1 AR 2004 - DJ-1 AR 2003 -

ATP13A2 AR 2006 More complex PD PLA2G6 AR 2006 More complex PD FBXO7 AR 2008 More complex PD MAPT AD 1998 Pred. FTD GRN AD 2006 Pred. FTD C9orf72 AD 2011 Mixed – most common FUS AD 2009 Pred. ALS SOD1 AD>AR 1993 Pred. ALS TARDP AD 2008 Mixed FTD-ALS VCP AD 2010 Pred. ALS, allelic IBMPFD SQSTM1 AD 2011 Mixed UBQLN2 XLD 2011 Mixed MATR3 AD 2014 Pred. ALS OPTN AD 2010 Pred. ALS

PFN1 AD 2012 Pred. ALS FTD-ALS CHMP2B AD 2006 Pred. ALS CHCHD10 AD 2011+ FTD-ALS, SMAJ, mitochondrial myopathy HTT AD 1993 Most common HD PRNP AD 1989 HDL2 JPH3 AD 2001 HDL3 TBP1 AD 1999/2001 Allelic to SCA17; HDL4 ATXN1 AD 1993 Allelic to SCA1 and HDL ATXN2 AD 1996 Allelic to SCA2 ATNX3 AD 1994 Allelic to SCA3 ATN1 AD 1994 DRPLA

283 FTL AD 2001 - and hereditary TITF1/NKX2.1 AD 2002 Benign Hereditary Chorea

choreas ADCY5 AD 2012 FDFM, ADCY5-related chorea RNF216 AR 2013 RNF216-mediated neurodegeneration FRRS1L AR 2016 - Neuroacanthocytosis VPS13A AR 2001 - Wilson’s Disease ATP7B AR 1993 Hepatolenticular degeneration McLeod syndrome XK XL 2001 - PANK2 AR 2001 PKAN - most common PLA2G6 AR 2006 PLAN; PARK14 C19orf12 AR 2011 MPAN; SPG43 WDR45 XLD 2012 BPAN, previously SENDA syndrome FA2H AR 2010 FAHN; SPG35 NBIA ATP13A2 AR 2006 KRS; PARK9 FTL AD 2001 Hereditary Ferritinopathy CP AR 1995 Aceruloplasminemia, DM

COASY AR 2014 CoPAN NBIA DCAF17 AR 2008 Woodhouse-Sakati syndrome RAB39B XL 2010 - ATXN1 AD 1993 SCA1 ATXN2 AD 1996 SCA2 ATNX3 AD 1994 SCA3 – most common AD SPTBN2 AD>AR 2006/2014 SCA5 – cave, also SCAR14 (AR) CACNA1A AD 1997 SCA6 ATXN7 AD 1997 SCA7 ATXN8 AD 2006 SCA8 ATXN10 AD 2000 SCA10

284 TTBK2 AD 2007 SCA11 PPP2R2B AD 1999 SCA12

Ataxias KCNC3 AD 2006 SCA13 -autosomal dominant PRKCG AD 2003 SCA14 ITPR1 AD 2007 SCA15/16/29 TBP1 AD 1999/2001 SCA17; HDL4 KCND3 AD 2012 SCA19/22 TMEM240 AD 2012 SCA21 PDYN AD 2010 SCA23 EEF2 AD 2012 SCA26 ? FGF14 AD 2003 SCA27 AFG3L2 AD 2010 SCA28, cave: also SPAX5 (AR) BEAN1 AD 2009 SCA31 TGM6 AD 2010 SCA35 NOP56 AD 2011 SCA36 ELOVL5 AD 2014 SCA38

CCDC88C AD 2014 SCA40? ATN1 AD 1994 DRPLA Ataxias KCNA1 AD 1994 EA1 -autosomal dominant CACNA1A AD 1997 EA2, FHM (point mutations), also SCA6 (repeats) continued CACNB4 AD 2000 EA5 and EA2 SLC1A3 AD 2005 EA6, also seizures or hemiplegia FXN AR 1996 Friedreich’s ataxia - most common ATM AR 1995 Ataxia-telangiectasia, cancer, immunodeficiency TTPA AR 1995/1996 AVED APTX AR 2001 AOA1 SETX AR 2004 SCAR1/AOA2

285 C10orf2 AR/AD 2005/2001 IOSCA (AR), PEOA3 (AD) SIL1 AR 2005 Marinesco-Sjoegren syndrome

PEX7/PHYH AR 2002 Refsum-disease POLG1 AR 2001-2003 MIRAS, SANDO CYP27A1 AR 1991 Cerebrotendinous xanthomatosis GRID2 AR>AD 2013 GRID2-related ataxias, SCAR18 ANO10 AR 2010 SCAR10 PNPLA6 AR 2008/ Boucher-Neuhauser, Oliver-McFarlane, 2014 Laurence-Moon syndromes, SPG39 Ataxias SLC52A2 AR 2012 Brown-Vialetto-Van Laere syndrome -autosomal recessive WFS1 AR 1998/1999 Wolfram syndrome SNX14 AR 2014 SCAR20 SYNE1 AR 2007 SCAR8 ADCK3 AR 2008 SCAR9, Coenzyme Q deficiency ATCAY AR 2003 Cayman cerebellar ataxia ACO2 AR 2012 Infantile cerebellar degeneration

ABHD12 AR 2010 PHARC: PNP, hearing loss, ataxia, retinitis pigmentosa, cataract ATP8A2 AR 2013 CAMRQ4 STUB1 (CHIP) AR 2013 SCAR16 CLCN2 AR 2013 Leukoencephalopathy with ataxia CLN5 AR 1998 Neuronal ceroid lipofuscinosis CWF19L1 AR 2014 SCAR17 FLVCR1 AR 2010 Ataxia, posterior column with RP GOSR2 AR 2011 Progressive myoclonic epilepsy-6 GRM1 AR 2012 SCAR13 KCNJ10 AR 2009 SeSAME syndrome

LAMA1 AR 2014 PTBHS 286 Ataxias KIAA0226 AR 2010 SCAR15 -autosomal recessive PNKP AR 2015 AOA4 continued PTF1A AR 2004 Pancreatic and cerebellar agenesis RNF216 AR 2013 Gordon Holmes syndrome SLC9A1 AR 2015 Lichtenstein-Knorr syndrome SYT14 AR 2011 SCAR11 TDP1 AR 2002 SCAN1 TPP1 AR 2013 SCAR7/NCL2 VLDLR AR 2005 Cerebellar hypoplasia WWOX AR 2014 SCAR12 FMR1 XLD 1991 FXTAS – most common XLD ABCB7 XL 1999 Sideroblastic anemia and ataxia Ataxias CASK XLD 2008 CASK-related disorders -X-linked SLC9A6 XLD 2008 Mental retardation, syndromic OPHN1 XL 1998 Mental retardation, cerebellar hypoplasia, syndromic features

SACS AR 2000 ARSACS VAMP1 AD 2012 SPAX1 KIF1C AR 2014 SPAX2 Spastic ataxias MARS2 AR 2012 SPAX3, white matter changes MTPAP AR 2010 SPAX4 AFG3L2 AR 2011 SPAX5 (AR), SCA28 (AD) SPG7 AR/AD 1998 SPG7 ATL1 AD 2001 SPG3 SPAST AD 1999 SPG4 – most common NIPA1 AD 2003 SPG6 KIAA0196 AD 2007 SPG8

287 HSP KIF5A AD 2002 SPG10 -autosomal dominant RTN2 AD 2012 SPG12

HSPD1 AD/AR 2002/2008 SPG13; hypomyelinating leukodystrophy 4 (AR) BSCL2 AD 2004 SPG17, Silver syndrome; allelic to AD CMT type 2 REEP1 AD 2006 SPG31, AD HMN type 5B ZFYVE27 AD 2006 SPG33 SLC33A1 AD 2008 SPG42 REEP2 AD+AR 2014 SPG72 CYP7B1 AR 2008 SPG5 SPG7 AR/AD 1998 SPG7 SPG11 AR 2007 SPG11, TCC; CMT type 2X; ALS5 ZFYVE26 AR 2008 SPG15 HSP ERLIN2 AR 2011 SPG18 -autosomal recessive SPG20 AR 2002 SPG20, Troyer syndrome SPG21/ACP33 AR 2003 SPG21, Mast syndrome B4GALNT1 AR 2013 SPG26

DDHD1 AR 2012 SPG28 KIF1A AR 2011/2012 SPG30; HSN type 2C FA2H AR 2010 SPG35; FAHN PNPLA6 AR 2008/ SPG39, Boucher-Neuhauser, Oliver-McFarlane, 2014 Laurence-Moon Syndromes C19orf12 AR 2011 SPG43; MPAN GJC2 AR 2009 SPG44; hypomyelinating leukodystrophy GBA2 AR 2013 SPG46 AP4B1 AR 2011 SPG47, TCC, white matter changes AP5Z1 AR 2010 SPG48 TECPR2 AR 2012 SPG49

AP4M1 AR 2009 SPG50 288 AP4E1 AR 2011 SPG51 HSP AP4S1 AR 2011 SPG52 -autosomal recessive VPS37A AR 2012 SPG53 continued DDHD2 AR 2012 SPG54 C12orf65 AR 2012 SPG55 CYP2U1 AR 2012 SPG56 TFG AR 2013 SPG57; hereditary motor and sensory neuropathy Okinawa type KIF1C AR 2014 SPG58; SPAX2 USP8 AR 2014 SPG59 WDR48 AR 2014 SPG60 ARL6IP1 AR 2014 SPG61 ERLIN1 AR 2014 SPG62 AMPD2 AR 2014 SPG63 ENTPD1 AR 2014 SPG64 NT5C2 AR 2014 SPG65

ARSI AR 2014 SPG66 PGAP1 AR 2014 SPG67 HSP FLRT1 AR 2014 SPG68 -autosomal recessive RAB3GAP2 AR 2014 SPG69 continued MARS AR 2014 SPG70 ZFR AR 2014 SPG71 No SPG-designation GAD1 AR 2004 Spasticity and intellectual disability L1CAM XL 1994 SPG1 and MASA syndrome HSP PLP1 XL 1994 SPG2; Pelizaeus-Merzbacher disease -X-linked SLC16A2 XL 2004 SPG22; Allan-Herndon-Dudley syndrome GALC AR 2004 Krabbe disease

289 ABCD1 XL 1993 X-linked adrenoleukodystrophy/ adrenomyeloneuropathy GFAP AD 2001 Alexander disease LMNB1 AD 2006 - GBE1 AR 1996/ Adult polyglucosan body disease, glycogen 1998 storage disease ASPA AR 1993 Canavan disease CYP27A1 AR 1991 Cerebrotendinous xanthomatosis EIF2B1-5 AR 2002 Leukoencephalopathy with vanishing white matter Leukodystrophy SLC17A5 AR 1999 Sialic acid storage disorder, infantile FUCA1 AR 1993 Fucosidosis TUBB4A AD 2013 Hypomyelinating leukodystrophy-6; DYT4 FAM126A AR 2006 Hypomyelinating leukodystrophy-5 L2HGDH AR 2004 L2-hydroxyglutaric aciduria DARS2 AR 2007 LSBL EARS2 AR 2012 COXPD12

MLC1 AR 2001 MLC1 HEPACAM AR/AD 2011 MLC2A (AR); MLC2B (AD) ARSA AR 1991 MLD PSAP AR 1990 MLD; atypical Gaucher/Krabbe CSF1R AD 2012 HDLS SUMF1 AR 2003 Multiple sulfatase deficiency GJA1 AR/AD 2003 Oculodentodigital dysplasia GJC2 AR 2009 Hypomyelinating leukodystrophy-2; SPG44 PEX genes AR various Zellweger syndrome spectrum POLR3A AR 2011 Hypomyelinating leukodystrophy-7 Leukodystrophy POLR3B AR 2011 Hypomyelinating leukodystrophy-8

290 continued RNASET2 AR 2009 Cystic leukoencephalopathy without megalencephaly HSD17B4 AR 1997 D-bifunctional protein deficiency; Perrault syndrome1 ACOX1 AR 1994 Peroxismal acyl-CoA oxidase deficiency SCP2 AR 2006 Leukoencephalopathy with dystonia and motor neuropathy ALDH3A2 AR 1998 Sjogren-Larsson syndrome SOX10 AD 1998 SOX10-associated disorders TREX1 AR/AD 2006 Aicardi-Goutieres syndrome RNASEH2A AR 2006 Aicardi-Goutieres syndrome 4 RNASEH2B AR 2006 Aicardi-Goutieres syndrome 2 RNASEH2C AR 2006 Aicardi-Goutieres syndrome 3 SAMHD1 AR 2009 Aicardi-Goutieres syndrome ADAR AR 2012 Aicardi-Goutieres syndrome TOR1A AD 1997 DYT1- most common Monogenic dystonia TUBB4A AD 2013 DYT4; hypomyelinating leukodystrophy-6

THAP1 AD 2009 DYT6 MR1 AD 2004 DYT8; PNKD1 SLC2A1 AD 2011 DYT9, DYT18; GLUT1-deficiency Monogenic dystonia PRRT2 AD 2011 DYT10; PKD - -autosoautosomalmal dominantdominant SGCE AD 2001 DYT11, myoclonus dystonia continued ATP1A3 AD 2004 DYT12, RDP, alternating hemiplegia of childhood 2 CIZ1 AD 2012 DYT23, awaits confirmation ANO3 AD 2012 DYT24 GNAL AD 2013 DYT25 KCTD17 AD 2015 DYT26 Potential X-linked dystoniaTAF1 XL 2007 X-linked dystonia-parkinsonism, awaits confirmation

DRD, Segawa syndrome

291 GCH1 AD/AR 1994 Dopa responsive TH AR 1995 Segawa syndrome

dystonia SPR AR/AD? 2001 Sepiapterin reductase deficiency PTS AR 1994 Hyperphenylalaninemia Neuronal ceroid lipo- PPT1, TPP1, CLN3, GRN, AR/AD various - DNAJC5, CLN5, CLN6, fuscinoses MFSD8, CLN8, CTSD, ATP13A2, CTSF, KCTD7 SLC20A2 AD 2012 Fahr’s disease, IBGC1 Primary familial PDGFRB AD 2013 IBGC4 brain calcification PDGFB AD 2013 IBGC5

XPR1 AD 2015 IBGC6 PMP22 AD 1992 CMT1A, CMT1E, HNPP, DSS Selected hereditary SEPT9 AD 2005 HNA neuropathy with and SPTLC1 AD 2001 HSN1 without overlap to ATL3 AD 2014 HSN1F conditions above MPZ AD 1993 CMT1B, CMT2I, CMT2J, DSS, CHN LITAF AD 2003 CMT1C

EGR2 AD/AR 1998 CMT1D, DSS, CHN1, CMT4E NEFL AD/AR 2000 CMT2E, CMT1F KIF1B AD 2001 CMT2A1 MFN2 AD/AR 2004 CMT2A2, HMSN6A RAB7A AD 2003 CMT2B LMNA AD/AR 1999+ CMT2B1 (AR) and various others MED25 AR 2009 CMT2B2, BVSYS TRPV4 AD 2010 CMT2C, distal SMA, scapuloperoneal SMA GARS AD 2003 CMT2D, HMN5A HSPB1 AD 2004 CMT2F, HMN2B GDAP AD/AR 2001 CMT4A, CMT2K

292 HSPB8 AD 2004 CMT2L, dHMN2A AARS AD 2010 CMT2N, also EIEE29 (AR)

Selected hereditary DYNC1H1 AD 2010+ CMT20, de novo mental retardation, SMALED1 neuropathy with and LRSAM1 AD/AR 2010 CMT2P without overlap to DNAJB2 AR 2012/2014 CMT2T, DSMA5 conditions above MARS AD 2013 CMT2U, ILLD continued DNM2 AD 2005+ CMT2M, centronuclear myopathy, DI-CMTB (all AD), lethal congenital contracture

syndrome YARS AD 2006 DI-CMTC GNB4 AD 2013 DI-CMTF GDAP1 AR/AD 2002+ CMT2K, CMT4A, MTMR2 AR 2001 CMT4B1 SBF2 AR 2003 CMT4B2 SH3TC2 AR 2003 CMT4C2 NDRG1 AR 2000 CMT4D PRX AR/AD 2000+ CMT4F, DSS

FGD4 AR 2007 CMT4H FIG4 AR 2009+ CMT4J, YVS, ALS11 Selected hereditary GJB1/CX32 XLD 1993 CMTX1 neuropathy AIFM1 XL 2010+ CMTX4/Cowchock syndrome, COXPD6, X-linked deafness 5 continued PRPS1 XL 1995+ CMTX5, X-linked deafness 1, Arts syndrome, PRPS1-related gout, PRPS1-superactivity PDK3 XLD 2013 CMTX6 Several Mt-gene maternal - Mt-DNA-deletion syndromes, MELAS, various mutations, Mt- others Mt-DNA deletions GBA AR 1987 Homozygous: Gaucher disease,

293 Heterozygous: risk factor for PD Selected exemplary FOLR1 AR 2009 Neurodegeneration due to cerebral folate transport

others deficiency SLC33A1 AR/AD 2008/ Congenital cataracts, hearing loss, neurodegenera- 2012 tion; SPG42 (AD)

Abbreviations: FAD=familial Alzheimer’s disease, PD=Parkinson’s disease, FTD-ALS=frontotemporal dementia-amyotrophic lateral sclerosis spectrum, HD=Huntington’s disease, HDL=Huntington’s disease like disorders, NBIA=Neurodegeneration with brain iron accumulation, DRD=dopa responsive dystonia, HSP=hereditary spastic paraplegia, SPG=spastic paraplegia gene, IBMPFD=inclusion body myopathy and Paget’s disease of the bone, FDFM=Familial Dyskinesia with facial myokymia, DRPLA=dentatorubral-pallidoluysian atrophy, EA=episodic ataxia, FHM=familial hemiplegic migraine, AVED=ataxia with isolated vitamin E deficiency, AOA=ataxia-oculomotor-apraxia, SCA=autosomal dominant spinocerebellar ataxia, SCAR=autosomal recessive spinocerebellar ataxia, PEOA3=autosomal dominant progressive external ophthalmoplegia, IOSCA=infantile-onset spinocerebellar ataxia, ARSACS=autosomal recessive spastic ataxia of Charlevoix-Saguenay, MIRAS=mitochondrial recessive ataxia syndrome, SANDO=sensory ataxia, neuropathy, dysarthria and ophthalmoplegia, SCAN=spinocerebellar ataxia with axonal neuropathy, SPAX=spastic ataxia, PNP=polyneuropathy, CAMRQ4=cerebellar ataxia, mental retardation and dysequilibrium syndrome-4, RP=retinitis pigmentosa, SeSAME= seizures, sensorineural deafness, ataxia, mental retardation, and electrolyte imbalance, PTBHS= Poretti- Boltshauser syndrome, NCL= Neuronal ceroid lipofuscinosis, CMT=Charcot-Marie-Tooth disease, HMN=hereditary motor neuropathy, HSN=hereditary sensory neuropathy, ALS=amyotrophic lateral sclerosis, FAHN=Fatty acid hydroxylase-associated neurodegeneration, DM=Diabetes mellitus, PKAN=pantothenate kinase-associated neurodegeneration, PLAN=PLA2G6-associated neurodegeneration, MPAN=mitochondrial membrane protein-associated neurodegeneration, BPAN=beta-propeller protein-associated neurodegeneration,

SENDA=static encephalopathy of childhood with neurodegeneration in adulthood, KRS=Kufor-Rakeb syndrome, HMSN=hereditary motor sensory neuronopathy, MASA=mental retardation, aphasia, shuffling gait and adducted thumbs, TCC=thin corpus callosum, LSBL=leukoencephalopathy with brainstem and spinal cord involvement and lactate elevation, COXPD=Combined oxidative phosphorylation deficiency, MLC=megalencephalic leukoencephalopathy with subcortical cysts, MLD=metachromatic leukodystrophy, HDLS=hereditary diffuse leukoencephalopathy with spheroids, PNKD=paroxysmal nonkinesigenic dyskinesia, PKD=paroxysmal kinesigenic dyskinesia, RDP=rapid-onset dystonia-parkinsonism, IBGC=idiopathic basal ganglia calcification, HNPP=hereditary neuropathy with liability to pressure palsies, DSS=Dejerine-Sottas-syndrome, HNA=hereditary neuralgic amyotrophy, CHN=congenital hypomyelinating neuropathy, BVSYS=Basel-Vanagaite- Smirin-Yosef syndrome, SMA=spinal muscular atrophy, d=distal, EIEE29=early infantile epileptic encephalopathy 29, SMALED1=autosomal dominant lower extremity-predominant spinal muscular atrophy-1, DSMA=distal spinal muscular atrophy, ILLD=interstitial lung and liver disease, DI-CMTB/C/D/F=dominant intermediate-B/C/D/F-CMT, YVS= Yunis-Varon syndrome, SMAJ=late-onset spinal motor neuronopathy, Mt=mitochondrial, AD=autosomal dominant, AR=autosomal recessive, XLD=X-linked dominant, XL=X-linked, pred.=predominant.

Please note: This table is non-exhaustive. Some genes have multiple entries due to phenotypic overlap with several disease categories.

294

Section 1: Primer sequences for PCR and sequencing (genetics projects)

Chapter 3: Candidate genes:

Table 8-1: Primer sequences candidate genes

NBIA genes

Primer name Sequence (5’-3’) Amplicon size Conditions (bp) PANK2_x5_F TCTGTTGGGCTTTGTTGCTG PANK2_x5_R AACAAACCCCACCCCAAATG 299 TD65_55, D PANK2_x6_F ACATGGTGCTGTATTTGGGG PANK2_x6_R CCTCCACCCTCTACCTCAAG 295 TD65_55, D PLA2G6_x16_F GCTCCGAGAGTGCAGGG PLA2G6_x16_R AACAGAGCAGACCCTTGGG 273 TD65_55, D FA2H_5x_F AGTACCTCATCCACCGCTTC FA2H_5x_R CTTAGGTTCGGATGCCCAAC 270 TD65_55, D FA2H_6x_F CCTGGCTCACAAATGGGAAG FA2H_6x_R CAGGCTGTACAGGTAGGAGC 416 TD65_55, D FTL_4x_F TCTCTCAGCTCTGTGACTTCC FTL_4x_R TTGGTCCAAGGCTTGTTAGG 281 TD65_55, D WDR45_x12_F CCTAGGTCCTGAGATGCCTG WDR45_x12_R TTAATGCTTGCTGGCTGGTG 289 TD65_55, D

Dystonia/Chorea genes

Primer name Sequence (5’-3’) Amplicon size Conditions (bp) FUCA1_x6_F CAGACAGTAAGTTTGGGAGGC FUCA1_x6_R GATGGGGCACTGTTCTGTTC 229 TD65_55, D ETHE1_x1_F GAGGCTGTACTGAGGGTCG ETHE1_x1_R GAGTCCAGCCCTAAACCTCC 220 TD65_55, D NPC1_x4_F AAAATCGTTCTTGCTGGCCC 282 NPC1_x4_R TTGGCAAAACTCTGTCCGAC 239 TD65_55, D

NPC1_x13_F GTGATTGTGTCTGTCGCCTC NPC1_x13_R AGATGCTGAGCCCTGTGAG 243 TD65_55, D PNKD_x3_F AGTCTAGGGGAGCTAGGGAG PNKD_x3_R GCAGTCAGAGCTCACCATTG 187 TD65_55, D SGCE_x4_F GCTCTGAGTTCTCATTGCCC SGCE_x4_R AGGACTATCTGTTTGGCTTCCT 457 TD65_55, D SGCE_x7_F TTTGTTACATTCCCTACCTCCAA SGCE_x7_R TCCGTGTAATAGTCTCTGCTTTT 297 TD65_55, D THAP1_x2_F GTAAAGAATGGGAGGCAGCTG THAP1_x2_R AACTACAAGGTTCCAGGCAC 290 TD65_55, D ANO3_x2_F GAAACTTCGTTAAAACCGTCTCG ANO3_x2_R GCTCTGCTTTGTCAGTGGAG 195 TD65_55, D ANO3_x4_F AGGCTAAACTCTGGGAAAACA ANO3_x4_R CTCCAGCAGAAAACCAAATATGT 314 TD65_55, D GNAL_x2_F TGCGCACATTCCTAACTTCC GNAL_x2_R TGGGTAGCCTTGTAAGCCAG 172 TD65_55, D ADCY5_x11_F CCGTCTCCTCTCAGTACTCC ADCY5_x11_R AAACTTCTCTCCGTGCCTGA 242 TD65_55, D

295

PD genes PINK1_x1_F AAGTTTGTTGTGACCGGCG PINK1_x1_R GAGACGGTTAGGGAGCCC 278 TD65_55, D+D FBXO7_x1_F GGCGTCATCTGTGGGCTGPD genes

FBXO7_x1_R TCCCTGGTTCCGTTTACTGC 247 TD65_60, D LRRK2_x19_F AGTTTGATTTGCCAGTCTCCT LRRK2_x19_R GGCTGTCACCTTTCCCAATG 186 TD60_55, D LRRK2_x27_F GGTGGTTCAACTTCAGGCTC LRRK2_x27_R AATGGAAATTAAATTAAGTGACACATC 565 TD60_55, D+D LRRK2_x29_F TTTTACCAAACATTATCAACTACCC LRRK2_x29_R TCCCTGTTCCAAACAAATGG 496 TD60_55, D+D LRRK2_x30_F GGATTCTTGCCTGTCGTTTG LRRK2_x30_R ACTGAAGCAATTGTTTGCCC 289 TD65_60, D+D CHCHD2_x1_F GCTTAGCTCTTCGGTGGTTG CHCHD2_x1_R TCATTGCCCCAGTAGAGTCC 167 TD65_55, D VPS35_x17_F TGTGGTACGTGCTTGATCATG VPS35_x17_R CTTCATAAATTGGCCCCTCGG 210 TD65_55, D DNAJC13_x28_F TGGTGTCTCTTAGCCAGTGG DNAJC13_x28_R AGGTCAAAACAGCATTCCTCC 226 TD65_60, D HSP genes

AP4E1_x1_F GAACTTGTTCAGGTGGGACG AP4E1_ x1_R TACGTGCTTGGAGGTGAGG 300 TD65_55, D+D SPG7_x17_F GCCTGTTCTTTCTAGCTGGC SPG7_x17_R AGCCAACACCTCCCAACTAC 238 TD65_60, D KIF1C_x23_F TGAAGAAGGTGGTGAGGTCC KIF1C_x23_R CGTAGCTTGCAGTCGTGAG 288 TD65_55, D GJC2_x2_F CGGTAAGCTCCACGTCATTG GJC2_x2_R TGAACTTGGCCTGCTCGTC 311 TD65_60, D NIPA1_x5_F CAGTTTCACCGTGCCTTCC NIPA1_x5_R ATCCCCAAGAAGTCCACCAG 285 TD65_55, D KIF1A_x48_F TCAGCCTTCCCCTCTTTGTC KIF1A_x48_R AGGGCCCTATCAAATGGTCC 353 TD65_55, D HSPD1_x2_F TCACGTTGCTTGCTCTGTTG HSPD1 _x2_R CCCATTGTAACGGCCACAG 262 TD65_55, D SLC33A1 _x5_F CGGCCTGTATTGCTGTGATC SLC33A1 _x5_R TGTGAGGGGATCTACAAGCC 342 TD65_55, D KIAA0196 _x7_F AAGACCTTGGCAAGCAATGG KI AA0196_x7_R CTGTGGGTTAAAGGCCAAAAG 405 TD65_60, D BSCL2 _x1_F TGGTAGCATTGTGGACCTCC BSCL2 _x1_R AGAGTCAACAGAGGCTCGTC 414 TD65_60, D+D KIF5A _x25_F CAACTCAGTTCAACCCCAGC KIF5A _x25_R CCCCACCCTATGCCTTCTAG 316 TD65_55, D SPAST _x15_F CTGAGAGGCTGAG GTAGGAG SPAST _x15_R CGAGGCTGCAGTGAACTATG 398 TD65_60, D L1CAM _x14_F GAAGCCTCCACTCCAGAAGG L1CAM _x14_R CCGACAATGGAGTGATCAGC 414 TD65_55, D AMPD2 _x1_F CATGAGAAATCGTGGCCAGG TD65_60, D+D AMPD2 _x1_R TCCTTGTGACAGAGGACAGC 487

Abbreviations: x=exon; F=forward; R=reverse; TD=touch down PCR. For respective cycling programmes see Appendix Section 2; D=Default PCR master mix, see Table 8-7. D+D=Default+DMSO, see Table 8-8. D+D+S=Default+DMSO+SolutionQ, see Table 8-9. DMSO=dimethyl sulfoxide.

296

Chapter 4: XK gene, transcript ID ENST00000378616:

Table 8-2: Primer sequences XK

Primer name Sequence (5’-3’) Amplicon Conditions size (bp) XK_x1_F CCGCCACAGCCACACAGC McLeod-TD70_65, XK_x1_R GTGGGGAGATCAGGCTACAG 374 D+D XK_x2_ F TGAGCATGGAAAGTCAGAATCC XK_x2_R CCCCAAATGCAGGCAGAAAT 361 TD65_60, D XK_x3-PCR_F AGTGGAGGTGAGACTTGATGA XK_x3-PCR_R AACTTCCTGCCTTGTTCCCT 970 TD65_55long, D XK_x3-Seq1_F AGTGGAGGTGAGACTTGATGA XK_x3-Seq1_R GCCTTCTCTATGTTCTCAGGGA 343 s. Sequencing PCR XK_x3-Seq2_F CCTCTTTACCTCCGTCCTGA XK_x3-Seq2_R CAGTACCCAATGAGCAGCTG 399 s. Sequencing PCR XK_x3-Seq3_F TGTGCGCACCTCTGTTGG XK_x3-Seq3_R CTGTGCTCTGAGACCTTGGG 324 s. Sequencing PCR

Abbreviations: x=exon; F=forward; R=reverse; TD=touch down PCR. For respective cycling programmes see Appendix Section 2. D=Default PCR master mix, see Table 8-7. D+D=Default+DMSO, see Table 8-8. D+D+S=Default+DMSO+SolutionQ, see Table 8-9. DMSO=dimethyl sulfoxide.

SYNE1 gene, transcript ID ENST00000423061:

Table 8-3: Primer sequences SYNE1

Primer name Sequence (5’-3’) Amplicon size Conditions

(bp) SYNE1_x18_F TGAATGAAACCACCGCTCAG SYNE1_x18_R GCTACGCTGTAAAAGTCTCTCA 243 TD65_55, D SYNE1_x98_F TCCGGTGTGTAGCATGACTT SYNE1_x98_R TCAGGTTCTTCTCTGGGCAG 230 TD65_55, D SYNE1_x108_F GAGAGACTGTCAGGCACGTA SYNE1_x108_R AATGTGGCGGCTTGTAACAC 249 TD65_55, D SYNE1_x77_F AGCCAGCCTCAAGACTTACC SYNE1_x77_R TGAAGCTCAGAACTCTCATTCA 297 TD65_55, D

Abbreviations: x=exon; F=forward; R=reverse; TD=touch down PCR, for respective cycling programmes see Appendix Section 2; D=Default PCR master mix, see Table 8-7. D+D=Default+DMSO, see Table 8-8. D+D+S=Default+DMSO+SolutionQ, see Table 8-9. DMSO=dimethyl sulfoxide.

297

PNPLA6 gene, transcript ID ENST00000414982:

Table 8-4: Primer sequences PNPLA6

Primer name Sequence (5’-3’) Amplicon Cond. size (bp) PNPLA6_x1_F CCACTAATCCCAGCGTCTCT TD65_55, PNPLA6_x1_R GGAAGGCGTCTGATACCCAC 208 D+D+S PNPLA6_x2_F GGGTTCCTTCGACTCCTTGA PNPLA6_x2_R GAGTTGTAGTCCTCTTGGCG 158 TD65_55, D PNPLA6_x3_F GTTTCCCGGCATGCACTG PNPLA6_x3_R CACCATACAGCCTCCCAAGA 356 TD65_55, D PNPLA6_x4_F GCAGAGATGGGGATTTGCTG PNPLA6_x4_R TGTAAATAGGGAAGGGGCGC 248 TD65_55, D PNPLA6_x5_F GTCTGCCTCTTCTCCAAGGT PNPLA6_x5_R AAAGGCCCCTGAGATAACCC 211 TD65_55, D PNPLA6_x6_F TGGGGAAAGGGGTATGCG PNPLA6_x6_R GTAGTCACCCTGGCCCAG 399 TD65_55, D PNPLA6_x7_F GACCTCCAGCCTCTGTCG PNPLA6_x7_R CCACCACCCATCCTCCTG 242 TD65_55, D PNPLA6_x8_F ACATGCCAGTCACCAGGG PNPLA6_x8_R GGGACCCTGTGATTTCCTTC 219 TD65_55, D PNPLA6_x9_F CCTCTGCCCTTGTCTCTCTT PNPLA6_x9_R TCCCTTGGAGCCTGGAATG 244 TD65_55, D PNPLA6_x10+11_F CTTACCCCGCCCCATCTTAT PNPLA6_x10+11_R CTTCCTGGGAGATTGGGCTC 419 TD65_55, D PNPLA6_x12_F TCCCAACCTGCTAATCCTCC PNPLA6_x12_R CGGTGGACCTGGATTCAAAC 211 TD65_55, D PNPLA6_x13_F CCACTACACCTGGCTCATTT PNPLA6_x13_R TCCCTCCCAGCCACCTTC 244 TD65_55, D PNPLA6_x14+15_F CATCAGGAGGTCACAAGCCT PNPLA6_x14+15_R TGATCTCAGGTGGCATGGTT 397 TD65_55, D PNPLA6_x16_F GCATGTTGACCTGAGCTGG PNPLA6_x16_R TTCCAGGAGTGAGGTTGAGG 368 TD65_55, D 282 PNPLA6_x17_F TAGGACGGTTTTGAGGCCTT

PNPLA6_x17_R GATGTAAGTGCAGTCGGAGC 299 TD65_55, D PNPLA6_x18_F TACAGGTGCAGCTCCCAC PNPLA6_x18_R TTACCCGGGATGATGTGTGT 296 TD65_55, D PNPLA6_x19_F TCATTCCCCACCCAAGCAG PNPLA6_x19_R CTCTGTAAGCCCCGCCCA 249 TD60_55, D PNPLA6_x20_F ATAGGAGGGAGAGGTGGGAC PNPLA6_x20_R CTCCATTTCCCAGCATCCCT 250 TD65_55, D PNPLA6_x21_F AGAGGAGGAGGAGGGTTCAT PNPLA6_x21_R TCACTAATGTCCCACCCACC 267 TD60_55, D PNPLA6_x22_F CCTGTGGACCATGGTTCCC PNPLA6_x22_R CACTGCGAATCTCCCCATTG 247 TD60_55, D PNPLA6_x23_F TAAGTTCCTCCCAGCAACGG PNPLA6_x23_R CGAGGTGGGTGAGATCAGG 297 TD60_55, D PNPLA6_x24_F GATCAGGGACCCAGGTGTG PNPLA6_x24_R CAGCCACGCCCCTAGAAG 293 TD65_55, D PNPLA6_x25_F GCCCTCATGCTCCTGGGT PNPLA6_x25_R CCACGGTTTCTCCCAGGAA 233 TD65_55, D

298

PNPLA6_x26_F TTTGTTCCTTAGCAGTGCGG PNPLA6_x26_R TTTATGGCACTCCTGGGGTG 300 TD65_55, D PNPLA6_x27+28_F TGTAACTCCTATTTGACCCTGTC PNPLA6_x27+28_R GTCCCAGAAGCTCAGCGG 400 TD65_60, D PNPLA6_x29_F GCGTGTCTGTGCGTGTTT PNPLA6_x29_R GTCATCGAGCCTGTCTGTGT 250 TD65_55, D PNPLA6_x30_F TGTGGGACTGGGTTGAACAT PNPLA6_x30_R CATGCCACTCCTGCCCAC 397 TD65_60, D PNPLA6_x31gDNA_F AGAATGGGCAGACAGAGTGG PNPLA6_x31gDNA _R GGGGCTTGGGTTCTGTAATC 248 TD65_55, D x31+32+33cDNA_F CGACTGCTTCAAGACCATGG x31+32+33cDNA_R GTTGCCGGAGAATCGACTTC 398 (on cDNA TD65_60, D PNPLA6_x32_F AGAGAGTTCCCACCCTAGGA only) PNPLA6_x32_R CACTCCCCTCCCTCTTGG 249 TD65_55, D PNPLA6_x33_F GGGAGTGGCTGGAGATGG PNPLA6_x33_R CCTGCCTGACACTGTACTCA 241 TD65_60, D PNPLA6_x34_F CATCAGTGTCCCGTGCTG PNPLA6_x34_R GCTGGTCAAGTCATCAGTGC 294 TD65_60, D

Abbreviations: Cond.=Conditions; x=exon; F=forward; R=reverse; TD=touch down PCR. For respective cycling programmes see Appendix Section 2. D=Default PCR master mix, see Table 8-7. D+D=Default+DMSO, see Table 8-8. D+D+S=Default+DMSO+SolutionQ, see Table 8-9. DMSO=dimethyl sulfoxide.

299

ATCAY gene, transcript ID ENST00000600960: Table 8-5: Primer sequences ATCAY

Primer name Sequence (5’-3’) Amplicon size Conditions (bp) ATCAY_x1_F ATTCCTCTGCGGCTTCCTTT ATCAY_x1_R GCGGTTCCCACTTAGAGAGA 240 TD65_55, D+D ATCAY_x2_F TGTTCTTGTCTGACTCGCCT ATCAY_x2_R GAAACCCGGAAACACTGAGC 220 TD60_55, D ATCAY_x3_F TCTCTTCCCATTCCTGCATGA ATCAY_x3_R ATCCAGAGACGGAAAGGTGG 399 TD65_55, D ATCAY_x4_F GGAGATATCCGGACTCTGGC ATCAY_x4_R GCGAGTTGACGGCTTAACAT 348 TD65_55, D ATCAY_x5_F CAGGGCTGGGAGAGGACT ATCAY_x5_R CTAGGGCCACAATGCAATCC 247 TD60_55, D ATCAY_x6_F CATGTTGGGTCCCTGCCTT ATCAY_x6_R CTGTGTGATTTTCATGCCCC 219 TD65_55, D ATCAY_x7_F ACGTGGAATGAGCAGAGTGA ATCAY_x7_R CACACTGAGGCCACTGGA 247 TD65_55, D ATCAY_x8_F CATGGCTCTGTCTGGACCTA ATCAY_x8_R CAGCAGCACGAAAACAGACT 229 TD65_55, D ATCAY_x9_F TCTCAGGGGAAAGGAAGGTT ATCAY_x9_R AGAAGGGCAGCGTCTTGAG 220 TD65_55, D ATCAY_x10_F CCAGAGAGGCCAGTACTAGC ATCAY_x10_R TCTTCCTCTCTAGCCTCCCC 239 TD65_55, D ATCAY_x11_F CCCAGCCAAGACCCTGTA ATCAY_x11_R TTCTGTCCTTCTAGAGAACCCT 250 TD65_55, D ATCAY_x12_F AGCTGACCCAAGCCTTTCTA ATCAY_x12_R GCCATGCAGAACTGTGAGTC 249 TD65_55, D exon13 UTR, not sequenced - -

Chapter 5: CACNA1A gene repeat sequencing, transcript ID ENST00000360228:

Table 8-6: Primer sequences CACNA1A

Primer name Sequence (5’-3’) Conditions SCA6_ F CACGTGTCCTATTCCCCTGTGATCC SCA6_R TGGGTACCTCCGAGGGCCGCTGGTG SCA6_PCR, SCA6 Abbreviations: F=forward; R=reverse; for SCA6_PCR cycling conditions see Appendix Section 2, Table 8-16. For SCA6 pipetting scheme see Table 8-10.

300

Section 2: PCR master mixes and cycling conditions, Workflow Sanger sequencing

PCR master mixes: Default PCR master mix: Table 8-7: PCR master mix, default PCR reaction Reagent Volume per reaction (µl) FastStart PCR Master, Roche* 7.5 Primer forward (c= 10 µM) 0.75 Primer reverse (c= 10 µM) 0.75 ddH2O 5 DNA 1 Total 15

*FastStart PCR Master, Roche, #04710452001

Default+DMSO (GC-rich or difficult to amplify for diverse reasons): Table 8-8: PCR master mix default + DMSO

Reagent Volume per reaction (µl) FastStart PCR Master, Roche* 7.5 Primer forward (c= 10 µM) 0.75 Primer reverse (c= 10 µM) 0.75 ddH2O 4.25 DMSO 0.75 DNA 1

Total 15

*FastStart PCR Master, Roche, #04710452001. DMSO=dimethyl sulfoxide.

301

Default+DMSO+SolutionQ (GC-rich or difficult to amplify for diverse reasons): Table 8-9: PCR master mix, default + DMSO + Solution Q

Reagent Volume per reaction (µl) FastStart PCR Master, Roche* 7.5 Primer forward (c= 10 µM) 0.75 Primer reverse (c= 10 µM) 0.75 ddH2O 1.25 DMSO 0.75 Solution Q 3 DNA 1 Total 15

*FastStart PCR Master, Roche, #04710452001. DMSO=dimethyl sulfoxide.

SCA6 PCR master mix (for repeat sequencing, Chapter 5): Table 8-10: SCA6 PCR master mix

Reagent Volume per reaction (µl) FastStart PCR Master, Roche* 5 Primer forward (c= 10 µM) 0.5 Primer reverse (c= 10 µM) 0.5 ddH2O 2 DMSO 1 DNA 1 Total 10

*FastStart PCR Master, Roche, #04710452001. DMSO=dimethyl sulfoxide.

302

PCR cycling conditions:

Table 8-11: McLeod-TD70_65 programme

McLeod-TD70_65 Temperature Time Nr (C°) (min) of cycles 94 10:00 1 94 00:30 70 00:30 8x 72 04:00 94 00:30 70* 00:30 16x 72 04:00 94 00:30 65 00:30 16x 72 04:00

72 05:00 1x 4 Hold -

*Reducing temperature in every cycle by ~0.3°.

Table 8-12: TD65_55long PCR programme

TD65_55long Temperature Time Nr (C°) (min) of cycles 94 10:00 1 94 00:30 65 00:30 8x 72 04:00 94 00:30

65* 00:30 16x 72 04:00 94 00:30 55 00:30 16x 72 04:00

72 05:00 1x 4 Hold -

*Reducing temperature in every cycle by ~0.63°.

303

Table 8-13: TD65_55 PCR programme

TD65_55 Temperature Time Nr (C°) (min) of cycles 94 10:00 1 94 00:30 65 00:30 8x 72 00:45 94 00:30 65* 00:30 16x 72 00:45 94 00:30 55 00:30 16x 72 00:45

72 05:00 1x 4 Hold -

*Reducing temperature in every cycle by ~0.63°.

Table 8-14: TD65_60 PCR programme

TD65_60 Temperature Time Nr (C°) (min) of cycles 94 10:00 1 94 00:30 65 00:30 8x 72 00:45 94 00:30 65* 00:30 16x 72 00:45 94 00:30 60 00:30 16x 72 00:45

72 05:00 1x 4 Hold -

*Reducing temperature in every cycle by ~0.3°.

304

Table 8-15: TD60_55 PCR programme

TD60_55 Temperature Time Nr (C°) (min) of cycles 94 10:00 1 94 00:30 60 00:30 8x 72 00:45 94 00:30 60* 00:30 16x 72 00:45 94 00:30 55 00:30 16x 72 00:45

72 05:00 1x 4 Hold -

*Reducing temperature in every cycle by ~0.3°.

Table 8-16: SCA6 PCR programme

SCA6_PCR Temperature Time Nr (C°) (min) of cycles 94 12:00 1 94 00:30 65* 00:30 25x 72 00:45 94 00:30 60 00:30 13x 72 00:45

72 10:00 1x 4 Hold -

*Reducing temperature in every cycle by ~0.2°.

PCR product clean up prior to sequencing (enzymatic Exosap method): Method for enzymatic removal of unincorporated primers (Exonuclease 1) and degradation of excess nucleotides (FastAPTm):

Protocol: -mix 5 µl of PCR product with 2 µl of Exosap mixture (Exonuclease I (EN0581) and FastAPTm (EF0654), Thermo Scientific) and incubate as indicated in Table 8-17

305

Table 8-17: Exosap PCR product cleanup

Exosap clean up Temperature Time Nr (C°) (min) of cycles 37 30:00 1 80 15:00 1 4 Hold -

The resulting product is ready for sequencing PCR using BigDye® and Sequencing Buffer:

Table 8-18: Pipetting scheme sequencing PCR

Reagent Volume per reaction (µl) Sequencing Buffer (5x)* 2 Primer for or rev (c= 10µM) 1 BigDye®* 0.5 ddH2O 3 Cleaned PCR product 3.5 Total 10

* BigDye® Terminator v1.1 and v3.1 5x Sequencing Buffer, #4336697, Thermo Fisher Scientific. Use foil for each step after addition of BigDye to protect sequences from degrading. For=forward; rev=reverse.

Sequencing PCR cycling conditions: Table 8-19: Sequencing PCR Sequencing Reaction Temperature Time Nr (C°) (min) of cycles 94 01:00 1 94 00:30 50 00:15 25x 60 04:00 4 Hold -

Sequencing PCR cleanup with prepared Sephadex plates: Protocol: 1) Preparation of Sephadex for each plate in a 50 mL tube: o 40 ml double distilled water (dddH20) (autoclaved, not Milli-Q)

o 2.9 grams Sephadex (Sephadex G-50 Bioreagent, for molecular biology, DNA grade, fine, S5897 Sigma) Mix well and allow the Sephadex to hydrate for at least 30 minutes at RT.

306

2) Preparation of purification plate: o add 350 µl of the well-mixed Sephadex to each well of a Corning glass plate (Corning® FiltrEX™ 96 well filter plates, 0.66 mm glass fibre filter, polystyrene, F5301 Sigma) o place the glass plate on an empty collection plate and centrifuge 3 minutes at 750 x g o place glass plate onto a new plate (which will go to the sequencer) and pipette the entire volume of the sequencing reaction (10 µl) from your Sequencing Plate onto the centre of Sephadex columns without touching the Sephadex o centrifuge for 5 minutes at 910 x g o check that there is liquid (min 10 µl) in all wells 3) Plates are now ready to go onto the sequencer (3730 DNA Analyzer system) and should be loaded together with the prepared sequencing template. If stored prior to sequencing, please use foil to protect from light.

307

Section 3: Supplementary tables for Chapter 5: DNA repair variants act as modifiers of AAO study

Table 8-20: Seed sense sequences for KASP assay design

Note that genotypes for SNPs in reverse orientation to chromosome given by our KASP assays (highlighted in red) will differ from those using HGVS nomenclature. Note: This table is published elsewhere in a modified version (2).

SNP to SNPs HGVS Names Seed sense sequences for KASP assay design Chromosome rs1800937 NC_000002.11:g.48025764C>T Forward TTGCCTGGCAGGTAGGCACAACTTA[C>T]GTAACAGATAAGAGTGAAGAAGATA rs4150407 NC_000002.11:g.128049631T>C Reverse AGTACACAATGGGAAGGTGGTCCAT[A>G]GACAAGAGCCTTCACCAGAAACTGA rs5742933 NC_000002.11:g.190649316G>C Forward GTAATTGCCTGCCTCGCGCTAGCAG[G>C]AAGGTAGTGTGGTGTGACTAACGGG rs1799977 NC_000003.11:g.37053568A>G Forward CTCAACCGTGGACAATATTCGCTCC[A>G]TCTTTGGAAATGCTGTTAGTCGGTA rs6151792 NC_000005.9:g.80056961C>T Forward TCACACAGCCATGTAAAATTAGGCC[C>T]GCAGACAATTCGAAGGAGGAGAAAA rs115109737 NC_000005.9:g.80102444G>A Forward GAATCACACAAGCTTATTTGCTATA[G>A]CATTATAATAACTTTTTACATCTGT rs71636247 NC_000005.9:g.80118976A>G Forward TGTATAAATATATGTGGAGAAAACC[A>G]TCTAGATAGAAGGCTTATTCCAAAA rs1805323 NC_000007.13:g.6026942G>T Reverse TCCAGTCACGGACCCAGTGACCCTA[C>A]GGACAGAGCGGAGGTGGAGAAGGAC rs12531179 NC_000007.13:g.6028687C>T Forward ATTTTTAGTAGAGACAGAGTTTCAC[C>T]GTGTTAGATAGTCTCGATCTCCTGA

308 rs3735721 NC_000008.10:g.103217695A>G Forward GCTGGGGCCAGCTTAGTTGTAAGAA[A>G]AACTATTATTGTATATAATTGGACA rs1037700 NC_000008.10:g.103250775G>C Reverse GGCCTCAGGCCGGGGTGAGACTTAC[C>G]CCTGCGTTTATCCGCCTCACGCTCT

rs5893603 NC_000008.10:g.103250839_103250840insG Forward TTGGCTGGCCCCGGGGCAGAGCAGC[->G]GAGCGGGACGCAAACCCAAAGTCAG rs1037699 NC_000008.10:g.103250930C>T Reverse AGGACAGGCCTGTCCGCCCGCCCTC[G>A]CCGCAGCCTGGCTTCGTCGTTGCGA rs16869352 NC_000008.10:g.103306033T>C Forward CAGCGTAAGGTAGCAATGCTTGGAA[T>C]ACACGCTTGCATTTTCCAATTGGCT rs61752302 NC_000008.10:g.103311153C>T Forward ACAATTTCAATATAAAATGAGCATT[C>T]GCCTTTCGATCCTTGGATTCTACTA rs72734283 NC_000014.8:g.75495059A>G Forward ATTATTTTATGATTTGACCTTGACA[A>G]CCCATCTAGCCAACTCCCATCCAGT rs175080 NC_000014.8:g.75513828G>A Forward GGTCATAGGACTTTCTCTCAAACTA[G>A]GCATCTGTTGTTCTAAACAATCTTC rs146353869 NC_000015.9:g.31126401C>A Forward AATGGTATGTATTAAAATGTGAATC[C>A]CAAGAGTGATGTGTCACTGTGCACT rs114136100 NC_000015.9:g.31197976C>T Forward GCTGCAATGGTCCTGGTCAAACAAC[C>T]GGTCATCCTTACTACCTTCGGAGTT rs150393409 NC_000015.9:g.31202961G>A Forward GCCTTTCTCAAATTGGCCAAACAGC[G>A]TTCAGTCTGCACTTGGGGCAAGAAT rs3512 NC_000015.9:g.31235005G>C Reverse ACAGAGAGCGTTAAAAGTAAAGGCA[C>G]TTCCAAGAGTAACACTGCTAATGCG rs20579 NC_000019.9:g.48668830G>A Reverse GCTGGACAGGAAGGGAGAATTCTGA[C>T]GCCAACATGCAGCGAAGTATCATGT

Table 8-21: Single SNP associations

A1 A2 MAF Beta P A1 A2 MAF Beta Beta Beta Beta Beta Beta Beta Beta P SNP Chr Pos (GeM-HD) (GeM-HD) (GeM-HD) (GeM-HD) (GeM-HD) (All) (All) (All) (All) P (All) (HD) P (HD) (SCA1) P (SCA1) (SCA2) P (SCA2) (SCA3) P (SCA3) (SCA6) P (SCA6) (SCA7) P (SCA7) (AllSCA) (AllSCA) rs1800937 2 48025764 T C 0.092 0.820 4.30E-03 T C 0.074 0.490 4.75E-01 0.520 6.21E-01 -0.571 6.51E-01 -0.459 8.18E-01 2.455 4.47E-02 0.614 8.25E-01 -10.050 5.34E-01 0.438 6.13E-01 rs4150407 2 128049631 C T 0.444 0.575 4.60E-04 G A 0.479 0.064 8.50E-01 -0.585 2.53E-01 -0.574 3.91E-01 1.384 1.03E-01 -0.013 9.85E-01 -2.129 2.55E-01 -2.702 3.83E-01 0.260 5.48E-01 rs5742933 2 190649316 C G 0.206 -0.699 9.49E-04 C G 0.205 -0.725 9.59E-02 -0.732 2.49E-01 1.102 2.19E-01 -2.333 3.69E-02 -1.005 2.19E-01 0.939 6.76E-01 0.551 8.77E-01 -0.714 2.03E-01 rs1799977 3 37053568 G A 0.319 0.847 7.16E-07 G A 0.280 -0.359 3.55E-01 0.531 3.39E-01 -0.241 7.58E-01 -2.555 2.20E-02 1.081 1.31E-01 -1.424 4.17E-01 -5.899 1.44E-01 -0.698 1.68E-01 rs6151792 5 80056961 T C 0.099 -1.049 2.09E-04 T C 0.117 -0.662 2.16E-01 -1.395 7.99E-02 -0.436 7.30E-01 0.495 6.79E-01 -1.347 2.21E-01 -1.577 5.07E-01 -0.116 9.77E-01 -0.350 6.09E-01 rs115109737 5 80102444 A G 0.060 -1.289 4.50E-04 A G 0.041 -2.095 1.81E-02 -3.014 2.10E-02 1.321 5.13E-01 -4.651 8.59E-02 0.100 9.50E-01 -5.197 1.41E-01 -5.756 2.87E-01 -1.726 1.28E-01 rs71636247 5 80118976 G A 0.054 -1.398 2.55E-04 G A 0.034 -2.208 2.63E-02 -1.917 1.89E-01 -1.974 4.39E-01 -6.400 4.86E-02 0.324 8.52E-01 -3.813 2.82E-01 -7.123 2.49E-01 -2.329 6.76E-02 rs1805323 7 6026942 T G 0.038 -0.950 3.04E-02 A C 0.043 -3.605 3.14E-05 -3.890 3.14E-04 -5.677 1.67E-03 -1.835 3.94E-01 -2.307 2.70E-01 -2.123 5.50E-01 -17.190 1.44E-01 -3.305 6.62E-03 rs12531179 7 6028687 T C 0.147 0.938 3.84E-05 T C 0.169 0.579 2.16E-01 1.070 1.23E-01 0.039 9.67E-01 1.137 3.08E-01 0.083 9.32E-01 -0.320 8.83E-01 -0.798 8.07E-01 0.367 5.39E-01 rs3735721 8 103217695 G A 0.085 -1.529 5.68E-07 G A 0.083 -0.389 5.25E-01 0.354 6.56E-01 0.692 5.13E-01 -3.278 6.32E-02 1.308 3.11E-01 -15.150 2.35E-03 -3.035 5.89E-01 -0.790 3.47E-01 rs1037700 8 103250775 C G 0.097 -1.541 5.03E-08 G C 0.094 -0.817 1.54E-01 -0.012 9.87E-01 1.046 2.46E-01 -4.132 2.11E-02 0.863 4.72E-01 -14.250 5.47E-04 -8.021 1.55E-01 -1.235 1.11E-01

309 rs5893603 8 103250839 G - 0.097 -1.548 4.28E-08 G - 0.093 -0.983 8.89E-02 -0.092 9.05E-01 0.914 3.13E-01 -4.189 1.84E-02 0.537 6.59E-01 -11.770 2.13E-03 -9.077 1.24E-01 -1.441 6.45E-02

rs1037699 8 103250930 T C 0.096 -1.570 2.70E-08 A G 0.094 -0.819 1.53E-01 -0.006 9.94E-01 0.758 4.13E-01 -3.519 3.97E-02 0.896 4.55E-01 -14.260 4.86E-04 -9.077 1.24E-01 -1.228 1.11E-01

rs16869352 8 103306033 C T 0.083 -1.528 4.01E-07 C T 0.080 -0.464 4.57E-01 0.691 3.98E-01 0.756 4.36E-01 -2.854 1.25E-01 0.681 6.27E-01 -10.850 3.24E-02 -7.745 1.64E-01 -1.067 2.09E-01 rs61752302 8 103311153 T C 0.023 -1.671 3.03E-03 T C 0.026 -0.150 8.92E-01 -0.520 7.46E-01 0.567 7.10E-01 -1.045 6.76E-01 4.882 1.18E-01 -8.015 1.69E-01 NA NA 0.019 9.89E-01 rs72734283 14 75495059 G A 0.099 0.858 4.32E-03 G A 0.089 0.898 1.40E-01 2.057 1.14E-02 1.585 1.82E-01 -1.099 5.41E-01 -0.650 5.88E-01 -1.686 5.59E-01 10.770 3.82E-02 0.318 6.98E-01 rs175080 14 75513828 A G 0.466 -0.434 7.72E-03 A G 0.435 -0.671 5.66E-02 -1.245 1.61E-02 0.279 7.16E-01 -0.090 9.23E-01 0.397 5.62E-01 -0.927 5.84E-01 -4.356 1.66E-01 -0.405 3.70E-01 rs146353869 15 31126401 A C 0.017 -6.107 4.30E-20 A C 0.017 -2.362 8.17E-02 -1.804 3.28E-01 1.980 5.64E-01 -8.999 3.81E-02 -1.537 4.94E-01 -3.496 5.52E-01 7.338 6.60E-01 -2.610 1.48E-01 rs114136100 15 31197976 T C 0.018 -5.073 8.49E-16 T C 0.019 -2.101 9.20E-02 -1.188 4.88E-01 1.609 6.00E-01 -1.168 7.89E-01 -3.519 8.25E-02 -3.464 5.55E-01 6.909 6.73E-01 -2.521 1.27E-01 rs150393409 15 31202961 A G 0.016 -5.765 9.34E-18 A G 0.013 -2.735 7.03E-02 -2.909 1.39E-01 -0.354 9.28E-01 -4.224 4.88E-01 -3.176 1.92E-01 -0.912 8.99E-01 7.443 6.57E-01 -2.551 2.17E-01 rs3512 15 31235005 C G 0.309 1.325 5.28E-13 G C 0.283 1.680 1.52E-05 1.297 2.94E-02 1.388 8.70E-02 1.020 3.03E-01 2.156 2.36E-03 0.886 6.37E-01 9.647 5.00E-03 1.809 2.22E-04 rs20579 19 48668830 A G 0.124 0.769 6.65E-03 T C 0.134 0.427 4.09E-01 0.119 8.82E-01 1.244 2.84E-01 0.412 7.55E-01 1.099 2.17E-01 -7.791 2.19E-02 -0.216 9.54E-01 0.515 4.28E-01

Beta denotes the effect size – that is, the number of years added to or subtracted from the expected age at onset for each copy of the minor allele (A1). MAF denotes the frequency of the minor allele in the GeM-HD study (Column F) and the present data (Column K). P-values coloured red satisfy Bonferroni correction for 8 disease groups and 22 SNPs. Note that for SNPs in reverse orientation to chromosome (rs4150407, rs1805323, rs1037700, rs1037699, rs3512, and rs20579) genotypes given by KASP assays (current study) are complementary to those obtained in the GeM- HD study, which uses HGVS nomenclature (see Table 8-20: Seed sense sequences for KASP assay design) but correspond to the same allele. This table is published elsewhere in a modified version (2).

Section 4: Bioinformatics

Example of developed pipeline/commands for candidate analysis using the Terminal on Mac (awk integrated), organised by different steps and commented on with explanation (#):

#important command to run on each txt-file that has been opened on a mac: perl -pi -e 's/\r/\n/g' known_genes_names.txt

#get the header: awk '/#CHROM/' /Users/Sarah/bin/annovar/NBIA_Clean_Filtered_Cohort.May.19.2015.vcf > vcf_header_NBIA_Clean_Filtered_Cohort.May.19.2015.txt

#convert to annovar: sarah_wiethoff:annovar Sarah$ perl convert2annovar.pl NBIA_Clean_Filtered_Cohort.May.19.2015.vcf - format vcf4old -includeinfo --outfile 1+NBIAfullCohortClean.annov

#keep track of the log (example follows): NOTICE: Read 317185 lines and wrote 276890 different variants at 317052 genomic positions (283396 SNPs and 33656 indels) NOTICE: Among 317052 different variants at 317052 positions, 35157 are heterozygotes, 241733 are homozygotes NOTICE: Among 283396 SNPs, 201406 are transitions, 81625 are transversions (ratio=2.47), 365 have more than 2 alleles

#annotate the annovar-format-transformed original vcf in a gene-based-fashion using humandb version hg19: perl annotate_variation.pl --geneanno 1+NBIAfullCohortClean.annov humandb/ --buildver hg19

#keep track of the log (example follows): NOTICE: Reading gene annotation from humandb/hg19_refGene.txt ... Done with 47495 transcripts (including 9716 without coding sequence annotation) for 25182 unique genes NOTICE: Reading FASTA sequences from humandb/hg19_refGeneMrna.fa ... Done with 33120 sequences NOTICE: Finished gene-based annotation on 317052 genetic variants in 1+NBIAfullCohortClean.annov NOTICE: Output files were written to 1+NBIAfullCohortClean.annov.variant_function, 1+NBIAfullCohortClean.annov.exonic_variant_function

#looks for all the homozygous stopgain-mutations in the annotated file with exonic variants and counts: awk '/stopgain/' 1+NBIAfullCohortClean.annov.exonic_variant_function | awk '/1\/1/' | wc -l 160

#looks for all the stopgain-stoploss-nonsynonymous AND the frameshift-mutations/frameshift insertions/deletions in the annotated file with exonic variants and amends them out of the annotated vcf into the file Clean_NBIA.FullCohort.codingEVF.frameshift: awk -F"\t" ' /stopgain|stoploss|nonsynonymous|frameshift|nonframeshift/ {{split($3,ar,"[/(/;/:/,]")}{OFS = "\t";print $4, $5, $6, $7, $8, ar[1], $9, $10, $11, $12, $13, $16, $2, $3, substr($0, index($0,$18))}}' 1+NBIAfullCohortClean.annov.exonic_variant_function > 1+Clean_NBIA.FullCohort.codingEVF.frameshift

#pulls out the splicing variants from the variant.function-file and puts them into the file together with above: awk -F"\t" ' /splicing/ {{split($2,ar,"[/(/;/:/,]")}{OFS = "\t";print $3, $4, $5, $6, $7,ar[1], $8, $9, $10, $11, $12, $15, $1, $2, substr($0, index($0,$17))}}' 1+NBIAfullCohortClean.annov.variant_function > 1+Clean_splicingNBIA

310

#looks for variants within genes from known-genes-list and sees whether present in the cohort/annotated vcf (done with all created PD/Dystonia/NBIA/IronMet-etc. lists separated by inheritances): awk -F "\t" 'NR==FNR{a[$1];next} $6 in a' /Users/Sarah/Desktop/US2014/Results_Genelists_Scripts_FinalNBIACohort_MinusOne/NBIA_FinalCoho rt_files/known_genes_names.txt /Users/Sarah/bin/annovar/1+Clean_NBIA.FullCohort.codingEVF.frameshift > 1+FullCohortClean.NBIA_known_genes

#looks for all splicing variants within genes from known-genes-list and sees whether present in the cohort/annotated vcf (done with all created PD/Dystonia/NBIA/IronMet-etc. lists separated by inheritances): awk -F "\t" 'NR==FNR{a[$1];next} $6 in a' /Users/Sarah/Desktop/US2014/Results_Genelists_Scripts_FinalNBIACohort_MinusOne/NBIA_FinalCoho rt_files/ known_genes_names.txt /Users/Sarah/bin/annovar/1+Clean_splicingNBIA > 1+FullCohortClean.AllSplicing

#additional to looking for variants in known genes, it is important to look at homozygous stopgain mutations. This command looks for homozygous (1/1) stopgains in the cohort and writes them into a file called CleanFullCohort_nbia_hom_stopgains.txt: awk '/stopgain/' /Users/Sarah/bin/annovar/1+Clean_NBIA.FullCohort.codingEVF.frameshift | awk '/1\/1/' > 1+CleanFullCohort_nbia_hom_stopgains.txt

#I did not do this for the candidate gene analysis, but in an exploratory fashion, I looked for intronic variants as well. This command looks for all the UTR3, UTR5, downstream, upstream, intergenic, ncRNA_exonic variants from the variant.function-file and prints out the important columns of the annotated vcf into the file NBIA.feb.25.2015.noncoding: awk -F"\t" ' /UTR3|UTR5|downstream|upstream|intergenic|ncRNA_exonic|intronic/ {{split($2,ar,"[/(/;/:/,]")}{OFS = "\t";print $3, $4, $5, $6, $7,ar[1], $8, $9, $10, $11, $12, $15, $1, $2, substr($0, index($0,$17))}}' 1+NBIAfullCohortClean.annov.variant_function > 1+NBIAfullCohortClean.noncoding

#This command then looks for variants in genes from known-genes-list within the cohort and prints them into a separate file: awk -F "\t" 'NR==FNR{a[$1];next} $6 in a' /Users/Sarah/Desktop/US2014/Cases/Excel- GenesForFiltering/txt-files/known_NBIA.txt 1+NBIAfullCohortClean.noncoding > 1+_10NBIA_inNONCODING_Clean awk -F "\t" 'NR==FNR{a[$1];next} $6 in a' /Users/Sarah/Desktop/US2014/Cases/Excel- GenesForFiltering/txt-files/known_NBIA_plusIRONMET.txt 1+NBIAfullCohortClean.noncoding > 1+_NBIAplus_inNONCODING_Clean

#Those readout-files I would then move to the cluster Kronos at IoN where most of the annotation databases, vcf- and bamtools and prediction tools and most recent version of annovar are. Here I would change the format of my files following annovar guidelines, and then create annotated files for downstream analysis using the following command: /data/kronos/NGS_Software/annovar_Nov2014/table_annovar.pl/data/kronos/swiethoff/1+FullCohort Clean.NBIA_known_genes /data/kronos/NGS_Software/annovar_Nov2014/humandb_hg19/ -buildver hg19 –protocol refGene, genomicSuperDups, 1000g2012apr_all, esp6500si_all, esp6500si_aa,esp6500_ea, snp129,snp137,cg69,exac02,ljb_all,clinvar_20140211,caddgt20 -remove - otherinfo -operation g,r,f,f,f,f,f,f,f,f,f,f,f -nastring . -csvout

311

Section 5: iPSC work related primers, detailed results and methods

Control iPSC and ESC lines

Table 8-22: Control iPSC and ESC lines

Age Repro- Name Alias iPSC/ Gender Source at gramming ESC biopsy method Control 1 JOM iPSC M ~71 retroviral collaborators

Control 2 ND iPSC M ~64 retroviral collaborators/ Coriell Control 3 Shef6 ESC F - - collaborators /ESC control

Note: I have not reprogrammed/generated these lines myself, but have received them from collaborators and validated them alongside my own lines. Information as provided by collaborators.

Cell culture and media reagents, stem cells:

Table 8-23: Stem cell culture and media reagents

Item Company Catalogue # Purpose Essential 8® media Life A1517001 iPSC/ESC feeding plus supplement Technologies EDTA, UltraPure™ Life iPSC splitting/neuronal 15575-020 0.5M EDTA, pH 8.0 Technologies sheet lifting DPBS without Life iPSC splitting/neuronal magnesium and 14190-094 Technologies sheet lifting calcium Life Coating of dishes/plates GeltrexTM A1413302 Technologies in iPSC culture

Cell culture and media reagents, neuronal cells:

Neuronal cells are maintained on laminin-coated dishes (Sigma Aldrich, L2020). Laminin is an extracellular matrix multidomain trimeric glycoprotein supporting the cells’ adhesion, proliferation and differentiation. It is prepared by diluting the manufacturer’s 1 mg/1ml stock 1:50 in sterile PBS (aliquotting of the original stock is preferred when only small quantities are used at once) and coating the surface of the dishes with subsequent incubation at 37°C for 4 hours before removal of the laminin and plating of neuronal cells.

312

Neuronal cells are split using incubation of cells at 37°C in dispase (Life Technologies, 17105) or accutase when cultures are older (Innovative Cell Technologies, AT104), until cells lift off spontaneously, and subsequent gentle spinning and washing in PBS before replating (see (237) for the detailed protocol).

Neural maintenance media/N2B27 makes up the base medium for initiating neural induction or patterning (see extrinsic cues below) or to maintain neurons as per “default”, meaning without presence of additional small molecules influencing differentiation. N2B27 is made up of equal quantities of N2 and B27.

To make N2B27 media, make up 500 ml stocks of N2 and B27 separately following below recipes and mix at equal quantities (all under sterile conditions) upon usage:

Table 8-24: N2B27 recipe and catalogue numbers

For 500 ml N2-media, mix:

Item Company Catalogue # Volume DMEM/F12 Life Technologies 31331-028 484 ml Glutamax N2 supplement Life Technologies 17502-048 5 ml Non essential Life Technologies 11140-050 5 ml amino acid B- Life Technologies 21985-023 1 ml mercaptoethanol Penicillin- Life Technologies 15140-122 5 ml Streptomycin Insulin Sigma Aldrich 19278 0.25 ml For 500 ml B27-media, mix: Item Company Catalogue # Volume Neurobasal Life Technologies 12348-017 485 ml Medium B27 supplement Life Technologies 17504-044 10 ml L-Glutamine Life Technologies 25030-024 5 ml

Abbreviations: DMEM/F12: Dulbecco’s modified eagle medium, nutrient mixture F12.

313

Extrinsic cues

Neural maintenance media (N2B27, see above) is supplemented with extrinsic cues below in respective concentrations indicated in the text and this table to make up neural induction media, or patterning media.

Table 8-25: Extrinsic cues for differentiation of iPSCs

Name/ Catalogue Concen- Company Mechanism Abbreviation # tration Potent, selective Dorsomorphin/ R & D AMPK inhibition and 3093/10 1 µM DM Systems BMP pathway inhibition Potent, selective TGF- SB 431542/ R & D 1614/10 10 µM/2µM beta, ALK4 and ALK7 SB Systems inhibition Selective small StemMACSTM Macs molecule GSK3 130-103- CHIR99021/ Miltenyl 3 µM inhibition and Wnt 926 CHIR Biotec signalling activation Heparin binding growth factor, FGF8b/ PeproTech 100-25 200 ng/ml promotion of cellular FGF8 proliferation and differentiation

Functional readouts: qPCR primer design criteria:

All qPCR primers were designed using primer blast under the following conditions: 1) Paste the target sequence into http://www.ncbi.nlm.nih.gov/tools/primer-blast/, ideally try and design intron spanning primers first with the intron being longer than 2000 bp (since the Taq polymerase in the Fast/Power SYBR® green master mix amplifies at a rate of 1 kb/15 s and 20 s cycles are used). If impossible, try to design exon junction spanning primers where it is important to have a maximum of 4 nt annealing to the 3' exon in order to prevent annealing of the primer only to the 3' exon. If this is impossible as well, design the primer to the exons, but carefully examine the qPCR result achieved with the control RT- condition where no reverse transcriptase enzyme was added during reverse transcription in order to check for gDNA contamination. 2) The conditions to ensure in primer blast during design are: - optimum length 20 nt, but allowed from 15-25 - GC content: 30-80% - Melting temperature: optimal 59, but allowed from 58-60 - for intron spanning primers: min intron length 2000, maximum 10000 nt - for exon junction spanning primers: max 5' end 5 nt, max 3' end 4 nt 314

3) Evaluate the suggested primer and its off-target effects. A primer is good if it has no off-target effect or if the off-targets are larger than 2000 bp 4) When ordered, check efficiency via efficiency curves in test runs.

315

qPCR primer sequences, SCA15 iPSC project (Chapter 6.1):

Table 8-26: qPCR primers SCA15 iPSC project

Primer name Sequence (5’-3’) Amplicon size (bp) HOUSEKEEPING GENES GAPDH_F ATGACATCAAGAAGGTGGTG GAPDH_R CATACCAGGAAATGAGCTTG 177 CYCLOPHILIN_F GGCAAATGCTGGACCAAACAC CYCLOPHILIN_R TTCCTGGACCCAAAACGCTC 147 B-ACTIN_F TCACCACCACGGCCGAGCG B-ACTIN_R TCTCCTTCTGCATCCTGTCG 351 FIBROBLAST GENES S100A4_F TCTTGGTTTGATCCTGAC 260 S100A4_R GGAAGACACAGTACTCTTGG END. PLURIPOTENCY MARKERS

OCT3/4_F TTCTGGCGCCGGTTACAGAACCA 218 OCT3/4_R GACAACAATGAAAATCTTCAGGAG NANOG_F AGCTTGCCTTGCTTTGAAGCA NANOG_R TTCTTGACTGGGACCTTGTC 256 SOX2_F CATGGCAATCAAAATGTCCA SOX2_R TTTCACGTTTGCAACTGTCC 119 END. + PLA. PLURIPOTENCY

OCT4 end_F CCCCAGGGCCCCATTTTGGTACC 143 OCT4 end_R ACCTCAGTTTGAATGCATGGG OCT4 pla_F CATTCAAACTGAGGTAAGGGAGA GC OCT4 pla_R TAGCGTAAAAGGAGCAACATAG 124 SOX2 end_F TTCACATGTCCCAGCACTACCAGA SOX2 end_R TCACATGTGTGAGAGGGGCAGTG 80 SOX2 pla_F TGCTTCACATGTCCCAGCACTACCAGA SOX2 pla_R TTTGTTTGACAGGAGCGACAAT 111 L-MYC end_F GCGAACCCAAGACCCAGGCCTGC 143 L-MYC end_R TCCCAGGGGGTCTGCTCGCACCGTGA L-MYC pla_F TGGGCTGAGAAGAGGATGGCTAC L-MYC pla_R TTTGTTTGACAGGAGCGACAAT 122 NEURONAL TARGET GENES

ITPR1_F1 AGAATGAAAGGAGGTCTGTA 100 ITPR1_R1 GAGAAGACAACTTCGAGAAC ITPR1_F2 CAAGTTTGACAACAAGACTG ITPR1_R2 TTCTTTCCTTGATCATTTCT 155 ITPR1_F3 TGGAGTGATAACAAAGACGA ITPR1_R3 CTGATCAGGGTCCACCTC 291 ITPR2_F1 TGGATCCAGAAATAGACATTA ITPR2_R1 GGTTGTACTTCCACAAATG 248 ITPR2_F2 CACGAATGAGAGTAAGAGATT ITPR2_R2 TCTATTTCTGGATCCATTTC 127 ITPR3_F1 AATGTGCTCAGCTACTACAG ITPR3_R1 GTTGAGGTTGGAATCATAGT 291

Abbreviations: end.=endogenous; pla.=plasmid; F=forward; R=reverse.

316

Primer sequences, developmental iPSC project (Chapter 6.2):

Table 8-27: qPCR primers developmental iPSC project

Primer name Sequence (5’-3’) Amplicon size (bp) HOUSEKEEPING GENE GAPDH_F AGAAGGCTGGGGCTCATTTG GAPDH_R AGGGGCCATCCACAGTCTTC 258 PLURIPOTENCY OCT4_F TTCTGGCGCCGGTTACAGAACCA OCT4_R GACAACAATGAAAATCTTCAGGAGA 218 NANOG_F GCTTGCCTTGCTTTGAAGCA NANOG_R TTCTTGACTGGGACCTTGTC 256 PLURIPOTENCY/EARLY NEURAL SOX2_F CATGGCAATCAAAATGTCCA SOX2_R TTTCACGTTTGCAACTGTCC 119 MES(END)ODERM BRACHYURY_F AATTAAGTACAATCCATTTGC BRACHYURY_R GTTGTCAGAATAGGTTGGAG 301 (MES)ECTODERM S100A4_F TCTTGGTTTGATCCTGAC S100A4_R GGAAGACACAGTACTCTTGG 260 ENDODERM GATA4_F CTGTCATCTCACTACGGGCA GATA4_R ATTTGAGGAGGGAAGAGGGA 245# TROPHECTODERM CDX2_F CAGCCAAGTGAAAACCAG CDX2_R TGATGTAGCGACTGTAGTGA 101 EARLY NEURAL, different regions OTX2_F GCTGAGTCTGACCACTTC OTX2_R CGATTCTTAAACCATACCTG 249 FOXG1_F AGGAGGGCGAGAAGAAGAAC FOXG1_R TCACGAAGCACTTGTTGAGG 213+ PAX6_F GTGTCCAACGGATGTGTGAG PAX6_R CTAGCCAGGTTGCGAAGAAC 254+ MUSASHI_F AAGAGATCCAGGGGTTTCGG MUSASHI_R GAAGGCCACCTTAGGGTCAA 117 SOX1_F TACAGCCCCATCTCCAACTC SOX1_R GCTCCGACTTCACCAGAGAG 199+ MIDBRAIN-HINDBRAIN BORDER, HOX-GENES EN1_F TGCACACGTTATTCGGATCG EN1_R TTGAGTCTCTGCAGCTGCTC 125 GBX2_F CAGGCTTCGCTCGTCGG GBX2_R GCTGTAGTCCACATCGCTCT 114 HOXA1_F CCCTACGCGTTAAATCAGGA HOXA1_R AAAAGTCTGCGCTGGAGAAG 85* BMP7_F GGAACGCTTCGACAATGAGAC BMP7_R GCAGGAAGAGATCCGATTCCC 86# WNT1_F CAACCGAGGCTGTCGAGAAA WNT1_R GTGCAGGATTCGATGGAACCT 105# HOXA4_F CCCTGGATGAAGAAGATCCA HOXA4_R GGTGTAGGCGGTTCGAGAG 84* HOXA5_F GCGCAAGCTGCACATAAGTC HOXA5_R CTTCTCCAGCTCCAGGGTCT 97* Abbreviations: x=exon; F=forward; R=reverse. + = primers from (237). # = primers from (330). * = primers from (316).

317

RNA extraction, Reverse Transcription, qPCR: RNA-extraction: RNA was extracted either by using TRIzol® reagent (Life Technologies, 15596-018) on living cells for work presented in Chapter 6.1 (work mainly done at QSBB) or the Maxwell RNA Isolation Instrument with the Maxwell® RSC simplyRNA Cells Kit (Promega AS1390) on snapfrozen cell pellets for work presented in Chapter 6.2 and the RNA sequencing of Chapter 6.1 (work mainly done at IoN).

In short, for extraction with TRIzol® reagent the following five steps were followed: 1) Prepare chloroform, isopropanol, 75% ethanol (in DEPC-treated water), RNase-free H20 and ensure to have a water bath or heat block (55–60°C) and a centrifuge at hand 2) Depending of density and type of cells and volume of dish/well to harvest, add ~125- 500 µl of chilled TRIzol® to the cultures upon prior removal of culture media and pipet up and down thoroughly for several times to harvest and lyse all cells, transfer to a labelled tube and leave at RT for ~5 mins for homogenisation. NB: TRIzol® volumes were kept the same per experiment/cell lines/replicate. 3) Add 200 µl of chloroform per 1 ml of TRIzol® reagent (for volumes mentioned above: 25-100 µl), shake vigorously by hand for 15 s and incubate for 2 to 3 mins at RT. 4) Centrifuge at 12000 x g for 15 mins at 4°C. (Steps 1-4=Phase separation) 5) Remove carefully and exclusively the aqueous phase to a new labelled tube. 6) Per 1 ml of TRIzol® reagent, add 500 µl of 100% isopropanol (for volumes mentioned above: 62.5-250 µl) and incubate at RT for 10 mins. 7) Centrifuge at 12000 x g for 10 mins at 4°C. (Steps 5-7=RNA precipitation) 8) Remove the supernatant and wash the RNA pellet with 1 ml of 75% ethanol per 1 ml of TRIzol® reagent (for volumes mentioned above: 125-500 µl). 9) Vortex briefly and centrifuge at 7500 x g for 5 mins at 4°C – discard the wash. 10) Vacuum or air dry the pellet (5 to 10 mins), then resuspend using RNAse free H20 (depending on size of pellet 10-20 µl) and incubate in a heat block or water bath (55- 60°C) for 10 to 15 mins, then proceed to quantification and quality check using nanodrop or Agilent 2100 bioanalyzer (RNA LabChip® kits, Agilent Technologies). (Steps 5-7=RNA wash and precipitation). (More information about the RNA extraction using TRIzol® can be found here: https://tools.thermofisher.com/content/sfs/manuals/trizol_reagent.pdf)

318

In short, for extraction with Maxwell RNA Isolation Instrument the following five steps were followed:

1) Set up of the machine: Selection of all samples that need extraction, placement of extraction columns and plungers, labelling of tubes and addition of 50 µl of elution- buffer (RNAse free H20) into each tube. 2) Preparation of homogenisation mix on ice: Add 20 µl of 1-Thioglycerol per 1 ml of homogenisation buffer provided in kit, mix well, keep on ice. 3) Add 200 µl of chilled homogenisation mix to the sample that was kept on dry ice until this step and resuspend the pellet via vortexing and pipetting. 4) Add 200 µl of lysis buffer to the sample, vortex for 15 secs and pipette the fully resuspended, homogenised and lysed sample into compartment 1 of the extraction columns. 5) Add 5 µl of chilled DNAse1 into compartment 4 and start the machine, put the extracted RNA immediately on ice when finished and proceed to quantification and quality check using nanodrop or Agilent 2100 bioanalyzer (RNA LabChip® kits, Agilent Technologies). (More information about the Maxwell RSC instrument and its usage can be found here: https://www.promega.co.uk/products/dna-and-rna-purification/rna-purification/maxwell- 16-system-rna-purification-kits/)

Reverse Transcription: All samples were handled on ice at all times. RNA was quantified and quality checked using nanodrop and/or Agilent 2100 bioanalyzer and stored away at -80°C or processed immediately for RNA sequencing or cDNA generation. To generate complementary DNA, reverse transcription with 1 µg RNA-input using SuperScript® III First Strand Synthesis System from Invitrogen (18080051) was performed for all experiments in Chapter 6.1. For all experiments described in Chapter 6.2, 2 µg RNA were used as input and reverse transcription was performed with the help of RevertAid – Reverse Transcriptase from ThermoFisher Scientific (EP0441) and the RevertAid First Strand cDNA Synthesis Kit (KI621). Briefly, for all experiments performed for Chapter 6.1, per reaction reagents from Table 8 – 28 were mixed and incubated for 5 mins at 70°C, then immediately put on ice for 5 mins, before components of Table 8 – 29 were added, the reactions spun down and cycling conditions detailed in Table 8 – 30 applied. From each batch of extractions, one separate reaction was added as a negative control where everything was done as with all other samples apart from water was added instead of the enzyme Reverse Transcriptase.

319

Table 8-28: Reverse transcription using SuperScript III step 1

Reagent Volume per reaction (µl) Random primers, 3 µg/µl (Invitrogen 48190011) 1 dNTPs, 10 mM 1 1 µg RNA + ddH2O, RNAse free 11 Total 13

Table 8-29: Reverse transcription using SuperScript III step 2

Reagent Volume per reaction (µl) Total from step 1 above 13 DTT (Invitrogen) 1 First Strand buffer 5x 4 RNase OUT (Invitrogen) 1 RT+: SuperScript III RT; or 1 RT-: ddH2O, RNAse free 1 Total 20

Table 8-30: RT-PCR cycling conditions SuperScript III

RT-PCR SuperScript III

Temperature Time Nr (C°) (min) of cycles 25 05:00 1 42 10:00 1 50 30:00 1 70 15:00 1 4 Hold -

Obtained 20 µl cDNA (concentration 50 ng/µl) was diluted 1/10 using 180 µl of RNAse free ddH2O to yield 200 µl cDNA (concentration 5 ng/µl) which was used as input for subsequent qPCR. To test primer efficiencies, obtained cDNA was diluted to four different concentrations to create efficiency curves (10/5/2.5/1.25 ng/µl).

Briefly, for all experiments performed for Chapter 6.2, per reaction reagents from Table 8 – 31 were mixed and incubated for 5 mins at 70°C, then held at 4°C before components of Table 8 – 32 were added, the reactions briefly spun down and cycling conditions of Table 8 – 33 were applied. From each batch of extractions, one separate reaction was added as a negative control where everything was done as with all other samples apart from water was added instead of the enzyme Reverse Transcriptase.

320

Table 8-31: Reverse transcription using RevertAid step 1

Reagent Volume per reaction (µl) Random primers 100 µM 1 dNTPs, 10 mM 1 2 µg RNA + ddH2O, RNAse free 13.5 Total 15.5

Table 8-32: Reverse transcription using RevertAid step 2

Reagent Volume per reaction (µl) Total from step 1 above 15.5 Buffer 5x 4 RT+: RevertAid enzyme; or 0.5 RT-: ddH2O, RNAse free 0.5 Total 20

Table 8-33: RT-PCR cycling conditions RevertAid

RT-PCR RevertAid

Temperature Time Nr (C°) (min) of cycles 25 10:00 1 42 60:00 1 45 10:00 1 70 10:00 1 8 Hold -

Obtained 20 µl cDNA (concentration 100 ng/µl) was diluted 1/10 using 180 µl of RNAse free ddH2O to yield 200 µl cDNA (concentration 10 ng/µl) which was used as input for subsequent qPCR. To test primer efficiencies, obtained cDNA was diluted to four different concentrations to create efficiency curves (20/10/5/2.5 ng/µl). Samples were either used for qPCR immediately or stored at -20°C for short-term and - 80°C for long-term storage. qPCR: For all experiments described in Chapter 6.1, 10 µg of cDNA was determined as input for qPCR using Power SYBR® Green master mix (Applied BiosystemsTM, 4367659) for amplification, detection and quantification of samples in 96-well format Mx3000P qPCR System (Stratagene).

321

For all experiments performed for Chapter 6.2, 10 µg of cDNA was determined as input for qPCR using Fast SYBR® Green master mix (Applied Biosystems™, 4385612) for amplification, detection and quantification of samples in 384-well format QuantStudio™ 6 Flex Real-Time PCR System (Applied Biosystems™). For all experiments, one RT- sample, and 1 sample where water is added instead of cDNA were run for each primer pair on each plate/qPCR run.

See pipetting and cycling conditions for both below:

Table 8-34: qPCR master mix, 96-well format, Mx3000P qPCR System

Reagent Volume per reaction (µl) Primer for (c= 10µM) 0.1 Primer rev (c= 10µM) 0.1 Power SYBR® green* 10 ddH2O, RNAse free 7.8 Input cDNA (5 ng/µl) 2 Total 20

*Power SYBR® green master mix (Applied BiosystemsTM, #4367659); for=forward; rev=reverse.

Table 8-35: qPCR cycling conditions on Mx3000P

qPCR on Mx3000P Temperature Time Nr (C°) (min) of cycles 95 10:00 1 95 00:30 60 01:00 40x 72 01:00 95 01:00 1 55 00:30 1 95 00:30 1

Table 8-36: qPCR master mix, 384-well format, QuantStudio 6 Flex Real-Time PCR System

Reagent Volume per reaction (µl) Primer for (c= 10µM) 0.4 Primer rev (c= 10µM) 0.4 Fast SYBR® green* 5 ddH2O, RNAse free 3.2 Input cDNA (10 ng/µl) 1 Total 10 *Fast SYBR® green master mix (Applied BiosystemsTM, #4385612); for=forward; rev=reverse. 322

Table 8-37: qPCR cycling conditions on QuantStudio

qPCR on QuantStudio Temperature Time Nr (C°) (min) of cycles 95 00:20 1 95 00:01 40x 60 00:20 95 00:15 1 60 00:15 1 95 00:15 1

Primary antibodies ICC, SCA15 iPSC project (Chapter 6.1):

Table 8-38: Primary antibodies Chapter 6.1

Antibody Species Company Catalogue # Dilution OCT3/4 goat Santa Cruz sc8628 1:500 SSEA4 mouse Abcam MC813/ab16287 1:200 KI67 rabbit Abcam ab15580 1:500 PAX6 mouse DSHB AB_528427 1:1000 OTX2 goat R&D systems AF1979 1:500 TUJ1 rabbit Covance MRB-435P 1:5000 TUJ1 mouse Covance MMS-435P 1:1000 PSD95 mouse Abcam ab2723 1:400 GFAP mouse Dako M 0761 1:200 GLAST rabbit Thermo Fisher PA5-34198 1:500

DAPI - Sigma Aldrich D9542 1:1000

Primary antibodies ICC, developmental iPSC project (Chapter 6.2):

Table 8-39: Primary antibodies Chapter 6.2

Antibody Species Company Catalogue # Dilution OTX2 rabbit R&D systems AF1979 1:500 DAPI - Sigma Aldrich D9542 1:1000

323

Secondary antibodies ICC, both projects (Chapter 6):

Table 8-40: Secondary antibodies ICC

Antibody Conjugate Dilution Goat anti mouse Alexa Fluor® 488 1:500 Goat anti rabbit Alexa Fluor® 488 1:500 Rabbit anti goat Alexa Fluor®- 488 1:500 Donkey anti mouse Alexa Fluor® 594 1:500 Goat anti rabbit Alexa Fluor® 594 1:500 Rabbit anti mouse Alexa Fluor® 594 1:500 Rabbit anti goat Alexa Fluor® 594 1:500

Alexa Fluor® dyes secondary antibodies, Life Technologies gDNA extraction protocol (from cells): For harvesting of gDNA from living cells, the following steps were followed: 1) Let cells grow confluent, and/or pool several wells to ensure a good yield. Remove all media, scrape cells off with cell-scraper, lyse and resuspend in PBS and transfer to labelled tube to spin down shortly and remove the PBS. 2) Resuspend cell pellet in 1 ml gDNA extraction buffer* with freshly added Proteinase K. 3) Put overnight into heatblock at 55°C. 4) Next day extract DNA with 1 ml phenol-chloroform once by adding 1 ml phenol- chloroform, mixing and centrifugation for 5 mins at max speed. 5) Take exclusively the upper phase very carefully and move it to a new labelled aliquot. 6) Add 33 µl Na-Acetate. Add 2.5 ml of 100% ethanol. Mix gently, but well, and leave at -20°C for an hour. 7) Spin at max speed for 15 mins, carefully aspirate ethanol. 8) Rinse with 70% ethanol. Spin at topspeed for 15 mins again, carefully aspirate ethanol. 9) Let the pellet air-dry and add 65 µl of 1 x TE buffer* and let dissolve in the fridge/on ice before quantification and quality check.

*Recipes: gDNA extraction buffer: 0.1 M Tris-HCl pH 7.5, 0.05 M EDTA pH 8.0, 1.25% SDS, 0.002 mg/ml RNase 1x TE buffer: 10 mM Tris-HCl pH 8, 1 mM EDTA.

324

Section 6: Web Resources The URLs for web resources frequently utilised in this thesis are as follows:

Online Mendelian Inheritance in Man (OMIM): http://www.omim.org/

Primer 3: http://primer3.ut.ee/

Homozygosity Mapper: http://www.homozygositymapper.org/

NHLBI Exome Variant Server EVS: http://evs.gs.washington.edu/EVS/

1000 Genomes project: www.1000genomes.org

Complete Genomics cg69 database: http://www.completegenomics.com/public- data/69-Genomes dbSNP: http://www.ncbi.nlm.nih.gov/projects/SNP

Exome Aggregation Consortium database: http://exac.broadinstitute.org/

MutationTaster: http://www.mutationtaster.org/

PolyPhen2: http://genetics.bwh.harvard.edu/pph2/

SIFT: http://sift.jcvi.org/

CADD: http://cadd.gs.washington.edu/home

GATK, Best Practices: https://www.broadinstitute.org/gatk/guide/best-practices.php

ANNOVAR: http://annovar.openbioinformatics.org/en/latest/

325