Enrichment of somatic mutations in schizophrenia brain targets prenatally active transcription factor bindings sites

Abstract

Schizophrenia (SCZ) is a complex neuropsychiatric disorder in which both germline genetic mutations and maternal factors, such as infection and immune activation, have been implicated, but how these two strikingly different mechanisms might converge on the same phenotype is unknown. During development, cells accumulate somatic, mosaic mutations in ways that can be shaped by the cellular environment or endogenous processes, but these early developmental mutational patterns have not been studied in SCZ. Here we analyzed deep (267x) whole-genome sequencing (WGS) of DNA from cerebral cortical neurons isolated from 61 SCZ and 25 control postmortem brains to capture mutations occurring before or during fetal neurogenesis. SCZ cases showed a >15% increase in genome-wide sSNV compared to controls (p < 2e-10). Remarkably, mosaic T>G mutations and CpG transversions (CpG>GpG or CpG>ApG) were 79- and 62-fold enriched, respectively, at transcription factor binding sites (TFBS) in SCZ, but not in controls. The pattern of T>G mutations resembles mutational processes in cancer attributed to oxidative damage that is sterically blocked from DNA repair by transcription factors (TFs) bound to damaged DNA. The CpG transversions similarly suggest unfinished DNA demethylation resulting in abasic sites that can also be blocked from repair by bound TFs. Allele frequency analysis suggests that both localized mutational spikes occur in the first trimester. We call this prenatal mutational process skiagenesis (from the Greek skia, meaning shadow), as these mutations occur in the shadow of bound TFs. Skiagenesis reflects as-yet unidentified prenatal factors and is associated with SCZ risk in a subset (~13%) of cases. In turn, mutational disruption of key TFBS active in fetal brain is well positioned to create SCZ-specific gene dysregulation in concert with germline risk genes. Skiagenesis provides a fingerprint for exploring how epigenomic regulation and prenatal factors such as maternal infection or immune activation may shape the developmental mutational landscape of human brain.

Exploration of cell type–specific somatic mutations in schizophrenia and the impact of maternal immune activation on the somatic mutation profile in the brain

“…12 By the whole-genome sequencing (WGS) approach, a comprehensive study reported no significant difference in the somatic mutation burden between patients with SZ and controls (CTs), 13 while another study reported an increased number of sSNVs with altered mutation signatures in the neurons of patients with SZ. 14 Considering the immaturity of this field, more studies with different perspectives should be performed to understand the role of somatic mutations in the cause of SZ. In addition, the findings should be complemented with cellular and animal model studies, which have rarely been conducted thus far.…”

“…12 These inconsistencies may be due to differences in DNA sequencing and analytical procedures as well as differences in the tissues used in each study. In particular, one WGS study reported a 15% increase in sSNVs in neurons of an SZ group compared with a CT group, 14 indicating that somatic mutations can accumulate in the intragenic regions, which cannot be assessed by WES.…”

“…A whole‐exome sequencing (WES) study reported an increased number and variant allele fraction (VAF) of somatic single nucleotide variants (sSNVs) in neurons and nonneurons of the prefrontal cortex (PFC), 11 while another study reported no differences in the total number or VAF in the PFC but found that somatic mutations in patients were enriched in SZ‐related pathways 12 . By the whole‐genome sequencing (WGS) approach, a comprehensive study reported no significant difference in the somatic mutation burden between patients with SZ and controls (CTs), 13 while another study reported an increased number of sSNVs with altered mutation signatures in the neurons of patients with SZ 14 …”

The human brain through the lens of somatic mosaicism

“…Current estimates attribute 5-10% of the missing genetic heritability of more than 100 human disorders to somatic mutations (Yang et al, 2020). Somatic mutations are a known cause of, or have been implicated in several brain disorders from developmental neuropsychiatric disorders such as focal epilepsy, autism spectrum disorders (ASD) and schizophrenia (SCZ) to neurodegeneration and Alzheimer's disease (AD) (Lim et al, 2015;D'Gama et al, 2017;Lim et al, 2017;Winawer et al, 2018;Baldassari et al, 2019a,b;Rodin et al, 2021;Sherman et al, 2021;Maury et al, 2022;Miller et al, 2022).…”

“…A separate study also showed a significant contribution of large (>4 Mb) CNVs to ASD risk in 0.2% of probands (Sherman et al, 2021). Somatic mutation contribution to SCZ has been explored through deep WGS of purified neuronal populations, which identified two mechanisms contributing specifically to SCZ compared to control neurons, referred to as "skiagenesis" (Maury et al, 2022). A significant increase in sSNVs proximal (+/−1 Kb) to the midpoint of active transcription factor binding sites (TFBS) was observed, with enrichment seemingly related to fetal development but not brain-specific.…”