RNA Dysregulation in Models of Frontotemporal Dementia and Amyotrophic Lateral Sclerosis
Total Page：16
File Type：pdf, Size：1020Kb
University College London RNA dysregulation in models of frontotemporal dementia and amyotrophic lateral sclerosis Doctoral Thesis Author: Jack Humphrey Supervisors: Prof Adrian Isaacs Dr Vincent Plagnol Dr Pietro Fratta UCL Institute of Neurology UCL Genetics Institute I 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. Jack Humphrey 3 Abstract Amyotrophic Lateral Sclerosis (ALS) and Frontotemporal dementia (FTD) are two rare but devastating neurodegenerative diseases that share pathological features and genetic factors. A central question in both diseases is the role of the RNA-binding proteins transactive response DNA-binding protein 43kDa (TDP-43) and fused in sarcoma (FUS). These proteins play a vital role in RNA regulation in all cells but in diseased neurons they alter their cellular localisation to form potentially pathogenic aggregates. This process can be linked to rare genetic mutations in the TARDBP and FUS genes, although most cases of ALS and FTD have no known genetic cause. My work uses the revolutionary technology of RNA sequencing to measure and compare gene expression and RNA splicing in different cellular and animal models of sporadic and genetic disease. Here I present the results of four studies that investigate the biology of TDP-43 and FUS, assessing both their normal cellular roles and the impact of rare disease-causing mutations. In these projects I analyse RNA sequencing data to discover novel gene expression and RNA splicing phenomena. This includes the repression of cryptic splicing by TDP-43 but not FUS, the progressive downregulation of mitochondrial and ribosomal transcripts in a mouse model of FUS ALS, a gain of splicing function by TDP-43 mutations affecting constitutive exon splicing, and widespread changes in intron retention caused by FUS knockout or aggressive FUS mutations. I also discover a novel mechanism for how FUS might regulate its own translation. This work expands on what is currently known about the roles in RNA regulation for TDP- 43 and FUS and provides new avenues for understanding both the causes and progression of ALS and FTD. 5 Impact Statement This thesis comprises four studies on TDP-43 and FUS, two proteins linked to the devas- tating neurodegenerative diseases Amyotrophic Lateral Sclerosis and Frontotemporal De- mentia. I analysed RNA sequencing data to develop insights in to the functions of these two proteins in different models of disease. This work has uncovered new roles for TDP-43 and FUS in specific types of RNA regulation. Additionally, I have discovered new mech- anisms to explain how disease-associated mutations in the two proteins affect these roles. Conclusions from my work are applicable to the RNA biology field as a whole as well as efforts to understand and treat these diseases. Insights into RNA-binding proteins and RNA regulation from this work can be transferred to many non-neurological diseases including muscular diseases, retinal diseases and certain cancers (Scotti and Swanson, 2015). Several of my chapters focus on developing new methods or adapting existing methods to analyse RNA splicing, a key mechanism of RNA regulation. RNA-sequencing is becoming a stan- dard experiment across all fields of biology and the use of sequencing data in analysis of RNA splicing is an area of extreme interest and method development. Three peer-reviewed papers have been published from this thesis. As of November 2018, “Quantitative analysis of cryptic splicing associated with TDP-43 depletion” (Humphrey et al., 2017) has been cited 11 times, “Humanized mutant FUS drives progressive motor neuron degeneration without aggregation in FUS Delta14 knockin mice” (Devoy et al., 2017) has been cited 11 times and “Mice with endogenous TDP-43 mutations exhibit gain of splicing function and characteristics of amyotrophic lateral sclerosis” (Fratta et al., 2018) has been cited 9 times and was the subject of a “News and Views” article in The EMBO Journal (Rouaux et al., 2018). These outcomes illustrate the importance of this work to the neurodegenerative disease and RNA biology fields. Three new mouse models of ALS/FTD have been generated for projects carried out in this thesis: the FUS ∆14 mouse, the TDP-43 RRM2mut mouse, and the TDP-43 LCDmut mouse. All three mice are available for labs around the world to use in further experiments. RNA-seq data created from these mice have been made publicly available and can be down- loaded from the Sequence Read Archive1. Several of my chapters rely on the re-analysis of published RNA sequencing data. I am pleased to see that sequencing data analysed in this thesis (chapter 5) has been reused in a further paper on TDP-43 (Sivakumar et al., 2018). I have tried to make all software written during this thesis publicly available where possible. This can be downloaded from the GitHub platform. This includes the RNA-seq analysis pipeline used in all chapters2, software developed to find and classify cryptic splicing (chapter 3)3, and software written to classify splicing events found in the TDP-43 LCDmut and RRM2mut mice (chapter 5)4. 1https://www.ncbi.nlm.nih.gov/sra 2https://github.com/plagnollab/RNASeq_pipeline 3https://github.com/jackhump/CryptEx 4https://github.com/jackhump/Two_TDP-43_Mutant_Mice 7 Contents 1 Introduction 15 1.1 Amyotrophic Lateral Sclerosis and Frontotemporal Dementia comprise a spec- trum of disease . 15 1.2 RNA splicing is a key step for RNA regulation . 16 1.3 RNA-sequencing is a revolutionary technology to quantify gene expression and splicing . 20 1.4 RNA-binding proteins implicated in ALS/FTD . 20 1.5 Aims of my PhD . 25 1.6 Overview of chapters . 25 2 Methods 27 2.1 Library preparation and sequencing . 27 2.2 The Plagnol lab RNA-seq pipeline . 30 2.3 Differential splicing . 32 2.4 Functional analyses . 35 2.5 Conclusions . 38 3 Cryptic splicing occurs in published TDP-43 but not FUS depletion data 39 3.1 Overview . 39 3.2 Methods . 39 3.3 Results . 44 3.4 Discussion . 54 3.5 Summary . 57 4 FUS mutant mice show progressive changes in mitochondrial and riboso- mal transcripts 58 4.1 Overview . 58 4.2 Contributions . 58 4.3 Background . 58 4.4 Methods . 59 4.5 Results . 61 4.6 Discussion . 63 5 Loss and gain of TDP-43 splicing function in two mutant mouse lines 65 5.1 Overview . 65 5.2 Contributions . 65 5.3 Methods . 66 5.4 Results . 69 5.5 Discussion . 82 9 6 ALS-causative FUS mutations impair FUS autoregulation through intron retention 85 6.1 Overview . 85 6.2 Contributions . 85 6.3 Background . 86 6.4 Methods . 88 6.5 Results . 93 6.6 Discussion . 104 7 Conclusions 109 7.1 Issues arising . 110 7.2 Future directions . 113 8 Appendices 143 8.1 Appendices to chapter 3 . 144 8.2 Appendices to chapter 4 . 149 8.3 Appendices to chapter 5 . 151 8.4 Appendices to chapter 6 . 155 List of Figures 1.1 Splicing of a U2-dependent intron . 17 1.2 Protein domains of TDP-43 and FUS . 20 2.1 Pipeline for stranded RNA-seq preparation and analysis . 29 3.1 Schematic of the Cryptex pipeline . 41 3.2 Cryptic splicing discovered by the CryptEx pipeline . 45 3.3 Evidence of TDP-43 binding cryptic exons . 47 3.4 Cryptic exons in transposable elements . 48 3.5 Conservation and premature termination codon analysis . 49 3.6 Differential expression of cryptic exon genes . 50 3.7 Scoring cryptic splice sites against canonical splice sites . 51 3.8 Mining of ENCODE eCLIP data in K562 and HepG2 cells . 53 3.9 Cryptic exons found in different ENCODE shRNA knockdowns . 56 4.1 The FUS ∆14 model . 59 4.2 Differential gene expression analysis on the ∆14 mouse . 61 4.3 Z-score heatmap of the 12 month spinal cord dataset . 62 4.4 Gene ontology categories in the 12 month spinal cord samples . 63 5.1 The two TARDBP mutations and their location within the TDP-43 protein 69 10 5.2 Opposite effects on splicing CFTR exon 9 minigene . 70 5.3 Comparing exon usage between the two mutations and a TDP-43 knockdown 71 5.4 Direction of cassette exon splicing in the two mutant lines . 71 5.5 RNA maps visualise positional enrichment of iCLIP peaks and sequence motifs 72 5.6 RRM2mut leads to cryptic exon splicing . 73 5.7 RRM2mut gene expression has bias for long gene downregulation . 74 5.8 LCDmut leads to skiptic exon splicing . 75 5.9 Permuting sample order shows clear splicing difference in LCDmut mice . 75 5.10 RNA maps of skiptic and cryptic exons show direct binding by TDP-43 . 76 5.11 Functional analyses of the skiptic exons . 77 5.12 LCDmut and TDP-43 autoregulation . 79 5.13 Skiptic exon splicing in TARDBP ALS . 80 6.1 The structure of the FUS protein and known ALS mutations . 86 6.2 Joint differential expression increases power and adjusts effect sizes . 94 6.3 Overlapping the joint differential expression models shows that FUS muta- tions affect gene expression in generally the same direction . 95 6.4 Xlr genes are upregulated in both FUS KO and FUS MUT . 96 6.5 Overlapping genes are enriched in neuronal and RNA terms . 97 6.6 Overlapping genes are enriched in neuronal and RNA terms . 98 6.7 Splicing changes strongly overlap between KO and MUT joint models . 100 6.8 Intron retention events are highly conserved and occur in RNA binding proteins101 6.9 FUS intron retention is an NMD-insensitive autoregulation mechanism .