Somatic mutations in the human brain: Tracing the origins of cancer and schizophrenia

Abstract

Somatic mutations, present in a subset of cells in the body, provide a unique opportunity for mechanistic understanding of disease. Somatic mutations represent natural experiments in humans akin to conditional transgenic mouse models, where only a subset of cells is affected. These somatic mutations can occur at any time during development, producing a spectrum of mosaicism from being present in multiple organs to being limited to a single cell, elucidating cell type specific contributions to many diseases.

The Introduction chapter provides background to the field of genetics with a focus on how somatic mutations have been used to understand cancer evolution from normal tissues as well as current literature and models about the role of somatic mutations in neurological and neuropsychiatric disorders. Special attention is given to two broad classes of somatic variants, single nucleotide variants (SNV) and copy-number variants (CNV). Chapter 1 focuses on the analysis of clonal somatic mutations in non-diseased human brains, revealing a robust prevalence of oncogenic mutations, especially in the brain’s white matter. Targeted sequencing of 110 individuals for genes implicated in brain tumors detected oncogenic somatic mutations in 5.4% of brains, including the classic glioma mutation IDH1 R132H. The IDH1 R132H variant was highly enriched in glial cells, but surprisingly was also detected in neurons, suggesting that this mutation arose prenatally during brain development, and later expanded in glial cells through proliferation. This finding represents one of the first observation of the developmental origin of adult-onset variants in brain tumors in humans. Analysis of 1,816 healthy human brains, from fetal to old age, further revealed that mutational patterns of somatic mutations in normal human brain resembled those of gliomas, suggesting that the mutational processes of normal brain drive early glial oncogenesis. In addition, macro-clones carrying somatic mutations in the brain were surprisingly depleted in older individuals. This result suggests that the presence of somatic mutations affecting a significant portion of the cells in the brain might contribute to disease pathology throughout life.

Chapter 2 describes work on mosaic somatic copy number variants (sCNV), that not only suggests roles for them in schizophrenia (SCZ) causation, but also in response to treatment. While inherited and de novo CNVs are familiar causes of SCZ, this study represents the largest study of sCNVs. Studying 12,834 SCZ cases and 11,648 controls revealed that sCNVs were more common in SCZ than in controls. In addition, we observed recurrent, somatic deletions of exons 1-5 of the NRXN1 gene in 5 SCZ cases, with Hi-C analysis suggesting a specific rewiring of regulatory elements in response to this recurrent deletion. We also observed 5 recurrent intragenic deletions of ABCB11, a gene associated with anti-psychotic response, specifically in treatment-resistant SCZ cases. Taken together our results indicate an important role of sCNVs in both SCZ risk and treatment responsiveness.

Chapter 3 describes work studying somatic point mutations (sSNV) in ultra-deep (267x) whole genome sequencing data of neuronal DNA from SCZ cases and controls, uncovering two mutational footprints reflecting a mutational process we referred to a skiagenesis. We find that SCZ cases show significantly more sSNV compared to controls, suggesting a strong association to disease risk. This increased mutation burden was partly driven by very high mutation rates at active transcription factor binding sites (TFBS) in promoters of brain-expressed genes in a subset (13%) of SCZ cases. These mutations in TFBS show characteristic T>G substitutions and CpG transversion patterns as a consequence of the inability of DNA repair machinery to access and repair those transcription factor bound sites efficiently. Our results implicate a genetic footprint of prenatal events in SCZ that might synergize with environmental and other genetic liability to disease.

Chapter 4 ties together all the insights from the studies presented and develops the implications of their major findings, as well as potential future areas of research. Overall, this thesis presents a snapshot of how somatic mutations can be used to gain unique biological insights of neurological and neuropsychiatric disease.