Convergent and Sex-Differential Genetic Underpinnings of Autism

Abstract

Autism is a highly heritable neurodevelopmental disorder characterized by restricted or repetitive behaviors and interests as well as challenges with social communication. Epidemiologically, autism is both common – one out of every thirty children in the United States is diagnosed with autism – and strikingly male biased, with three to four males diagnosed with autism for every one female. Twenty years of large-scale sequencing studies of autistic individuals have implicated many genetic variants across many different genetic variant classes in autism, demonstrating its substantial locus heterogeneity and complex genetic architecture. These sequencing studies have also consistently found that autistic females tend to carry a higher genetic load of autosomal autism-associated variation compared to autistic males. That females require more autosomal genetic risk to manifest autism than males suggests that females are somehow intrinsically genetically “protected” from autism, and this phenomenon has thus come to be known as the “female protective effect” (FPE). However, the genetic basis for the FPE remains unknown. In this dissertation, I address two major questions in autism genetics. First, how do the many diverse genes and genomic loci implicated in autism molecularly converge on this shared phenotype? Second, what mechanisms underlie the reduced genetic susceptibility of females to autism, as evidenced by the FPE? In Chapter 1, I discuss the genetic architecture of autism across three major axes (scale of variation, mode of inheritance, and allele frequency). I then summarize the evidence supporting the “convergence in autism” hypothesis, which proposes that the significant locus heterogeneity observed in autism can be explained by the convergent disruption of a core set of autism-associated genes and pathways by a wide array of distinct risk loci. Next, I discuss our current understanding of the origins of the male bias and FPE in autism. I emphasize that the male bias in autism cannot be fully explained by underdiagnosis of autistic females or X-linked causes of autism. Rather, the FPE suggests a generally decreased susceptibility of females to autosomal genetic risk for autism—a pattern I show extends to at least sixteen other male-biased congenital, developmental, or pediatric disorders. This suggests a more generalized FPE beyond autism. Finally, motivated by recent insights regarding the “inactive” X (Xi) and Y chromosomes – and particularly the recent realization that Xi and Y act in trans to modulate thousands of autosomal genes across the genome – I propose that a generalized FPE arises from genetic divergence between Xi and Y: the largest, oldest, and most common variation in the human genome. In Chapter 2, I demonstrate that genetic knockdown of autism risk genes in human neural progenitor cells and cortical organoids leads to convergent transcriptional changes due to shared disruption of a small set of key upstream regulators that coordinate tightly connected gene regulatory networks of autism genes. I further investigate convergence in autism through the creation and analysis of Consensus-ASD, a comprehensive and public-facing database integrating autism-associated variants across seven different genetic variant classes. Finally, I examine the role of X-linked genes, particularly those that are expressed from Xi, in driving both the male bias and FPE within autism. In Chapter 3, I analyze sequencing data from >400,000 healthy adults in two independent population cohorts to demonstrate that females in the general population harbor a significantly higher burden of deleterious autosomal variation, including both deleterious autosomal deletions and single-nucleotide variants, compared to males. Moreover, I demonstrate that this deleterious autosomal variation is concentrated in highly constrained and dosage- sensitive genes, suggesting that females may be more resilient to disruptions in autosomal gene dosage compared to males. This observation provides important context to the FPE in autism and further suggests a generalized FPE beyond autism, as hypothesized in Chapter 1. Finally, in the Appendix, I include three additional first-author works from my dissertation research that highlight the role of genetic variation in shaping phenotypes beyond autism, including Alzheimer’s disease, focal cortical dysplasia, and cardiac metabolism. Thus, while the main body of my thesis focuses on the genetics of autism, these additional works from my dissertation illustrate the broader application of genetic and genomic approaches to understanding complex human traits. In summary, my dissertation provides novel insights into how genetic variation contributes to both shared molecular features of autism and sex-specific patterns of genetic vulnerability.