An exploration of unsolved central nervous system white matter disorders with next-generation sequencing

Abstract

Leukodystrophies are a group of rare and devastating genetic neurological diseases that affect the brain white matter, which is composed chiefly of a substance known as myelin. Myelin ensheathes axons, and together these form long tracts that act as the information highway of the brain. Unfortunately, a large proportion of leukodystrophies remain without a diagnosis or “unsolved.” In 2010 it was reported that this may be as high as 50% for these disorders. This lack of understanding persists despite extensive diagnostic investigation, including advanced, comprehensive sequencing methods such as exome sequencing and in a limited number of cases, genome sequencing.Impatience with this statistic prompted clinicians and researchers to rapidly embrace the advent of next-generation sequencing methods, to allow unbiased, more comprehensive analyses of these affected individuals and families. However, our ability to more comprehensively analyse the genome comes with increasing challenges in understanding large- and small-scale genetic variation as well as defining the cohort of individuals with a more elusive cause of disease.In the chapters of this thesis, genome sequencing was first implemented on a cohort of 151 affected individuals and their biological parents in order to determine the potential value of genome sequencing as a diagnostic tool for this group of disorders (Chapter 2). Using this cohort, diagnoses were found in 58% of cases (n=85) and importantly were able to approximate the benefit of using genome sequencing to achieve a diagnosis through yielding additional analysis of intronic regions, GC-rich regions, and structural variant analysis in cases where exome sequencing had been previously unsuccessful.It was found that while genome sequencing offers us the highest likelihood in achieving a diagnosis, there were many lessons to be learned from this approach (Chapter 3). A large number of diagnoses were made in newly discovered white matter diseases, where gene-disease associations had not previously been made. One such example of this was the discovery of heterozygous variants affecting a relatively unknown gene, TMEM63A. Further, the identification of variants in GFPT1, not previously associated with central nervous system white matter disease, as the cause of leukoencephalopathy and episodic deterioration in two unique families, permitted disease-specific management and therapeutic benefit for the affected siblings.Based on the success of the genome sequencing approach, other ways to improve our diagnostic yield from genome sequencing studies were pursued. One avenue to increase this diagnostic rate was through improved identification and understanding of pathogenic variants that result in aberrant mRNA splicing, which can be difficult to identify by DNA analysis (Chapter 4). This approach was helpful in characterising targeted variants in GFAP, which are associated with Alexander Disease, but limits our ability to ambiguously study aberrant splicing in cases without targets of interest.The genome sequencing approach used in Chapter 2 enabled the ability to analyse genetic variants in intronic regions. However, this analysis is limited to solely determining if there are variants present or absent in these regions. What was found to be lacking was a functional interpretation, of how variants in these regions may disrupt molecular function. With this in mind, an RNA sequencing was pursued as an approach to assess aberrant splicing (Chapters 5-6) and confirmatory approaches based on solving transcript structure and relative quantification (Chapter 7). Using this approach on several cases, followed by targeted validation of amplicons using long-read sequencing on the Nanopore platform, strong candidates were found in a handful of cases with aberrant splicing and loss of expression caused by variants in proximal and distal intronic regions and allele-specific expression of damaging variants.Over the chapters presented in this thesis, it was demonstrated that genome sequencing is likely to confer the highest likelihood of achieving a diagnosis, potentially through a single test. For intractable cases, RNA sequencing approach and informatics pipeline prepared for adaptation for clinical use is presented and provides important considerations for analysis moving forward. The results found in this thesis have important implications for the diagnostic trajectory that leukodystrophy patients may endure in the future. While a few certain disorders have salient, recognisable characteristics and targeted biochemical testing that can be more rapidly provided and allow therapeutic intervention, for a large percentage of leukodystrophies, genome sequencing approaches provide the optimal path forwards to achieving a diagnosis.