Schizophrenia, autism spectrum disorders and developmental disorders share specific disruptive coding mutations

Schizophrenia is a severe psychiatric disorder with high heritability and polygenic architecture. While common risk alleles contribute to a portion of the genetic liability, rare copy number variants (CNVs) and ultra-rare protein-coding variants also play a significant role. Exome sequencing studies have revealed an excess of damaging coding variants in genes under strong selective constraint. *SETD1A* is a notable example, showing rare coding variants associated with both schizophrenia and developmental disorders. Previous research, however, lacked the power to fully assess whether shared risk genes harbor the same functional classes of variants across schizophrenia and neurodevelopmental disorders (NDDs), or if the same specific mutations contribute to both. This study aims to address these questions using a large dataset from the Deciphering Developmental Disorders (DDD) study, which provides comprehensive de novo variant data. The hypothesis is that pleiotropic genes, impacting both schizophrenia and NDDs, will show a greater enrichment of severe variants in NDDs compared to schizophrenia, due to the typically more severe neurodevelopmental impairment observed in NDDs. The study will investigate both genic pleiotropy (same gene, different variants) and allelic pleiotropy (same gene, same variant).

Existing literature indicates a genetic overlap between schizophrenia and NDDs. Studies using CNVs have shown pleiotropic effects, but the multigenic nature of these CNVs hinders precise identification of causative genes. Exome sequencing studies have provided stronger evidence for genic pleiotropy by demonstrating an enrichment of ultra-rare coding variants in NDD-associated genes in individuals with schizophrenia. *SETD1A*, significantly associated with both schizophrenia and developmental disorders, showcases this overlap. However, previous research limitations prevented a comprehensive analysis of whether the mutation classes are consistent across disorders and whether the same specific mutations cause both conditions. This study bridges this gap by leveraging the extensive DDD study dataset, allowing for a deeper investigation into the nature of this shared genetic architecture.

This study analyzed de novo variants from 3444 schizophrenia trios sourced from 11 published studies and compared them to de novo variant data from 37,488 neurodevelopmental disorder trios (DDD study). The schizophrenia de novo variants were categorized as protein-truncating variants (PTVs) or missense variants, with the latter further classified using MPC scores (Missense badness, Polyphen-2, constraint). Two primary analyses were performed: 1) Genic pleiotropy analysis: Genes enriched for PTVs or missense variants (MPC ≥2) in NDDs were tested for de novo variant enrichment in the schizophrenia dataset. A two-sample Poisson rate ratio test was used to compare observed and expected variant counts. A multivariate Poisson regression was employed to examine the independent effects of NDD PTV and missense enrichment statistics on schizophrenia de novo variant enrichment, controlling for brain expression levels and gene constraint scores. 2) Allelic pleiotropy analysis: NDD variants were divided into a primary set (PTVs in loss-of-function intolerant genes and missense variants with MPC ≥2) and a comparator set (remaining variants). Schizophrenia de novo variants were tested for enrichment within each set using a two-tailed Poisson exact test. A case-control analysis using an independent Swedish schizophrenia dataset was performed to replicate the allelic pleiotropy findings. Phenotypic data was also collected for schizophrenia probands carrying variants from the NDD primary variant set to investigate clinical manifestations.

The study revealed strong evidence for both genic and allelic pleiotropy between schizophrenia and NDDs. In the genic pleiotropy analysis, genes enriched for PTVs in NDDs showed significant enrichment for PTVs in schizophrenia (rate ratio = 4.89, P = 2.14 × 10⁻⁸). Similarly, genes enriched for missense variants in NDDs showed enrichment for missense variants in schizophrenia (rate ratio = 1.86, P = 0.04). Multivariate Poisson regression confirmed the association between NDD PTV enrichment and schizophrenia PTV enrichment (β = 0.15, P = 3.3 × 10⁻¹¹), while similar association was found between missense enrichment statistics. The allelic pleiotropy analysis showed a significant enrichment of specific NDD variants (primary variant set) in schizophrenia (observed: 9, expected: 1.20, P = 5.0 × 10⁻⁶). This enrichment was significantly greater than the general enrichment of constrained de novo variants in schizophrenia. The case-control analysis replicated the association between the primary variant set and schizophrenia (P = 0.036). Three variants identified in schizophrenia probands were known to be pathogenic for syndromic disorders (Okur-Chung neurodevelopmental syndrome, Kabuki syndrome, Neurofibromatosis type 1). Phenotypic analysis revealed that while some schizophrenia probands with NDD-associated variants exhibited premorbid cognitive impairment, many did not.

The findings strongly support the hypothesis that shared genetic mechanisms contribute to both schizophrenia and NDDs. The observed congruence in variant classes across disorders suggests that similar functional disruptions underlie both conditions. The identification of specific NDD variants enriched in schizophrenia highlights the substantial overlap at the allelic level. The lack of a strong correlation between mutation type and neurodevelopmental outcome severity suggests that other genetic, environmental, or stochastic factors determine the clinical presentation. The inclusion of schizophrenia as a potential phenotype associated with mutations causing syndromic developmental disorders is supported. The observation of NDD variants in schizophrenia probands without evident neurodevelopmental comorbidities underscores the need for broader cognitive assessments in schizophrenia.

This study provides compelling evidence for shared molecular aetiology between schizophrenia and NDDs, supporting the continuum hypothesis. The identification of specific variants implicated in both disorders highlights targets for further investigation into pathophysiological mechanisms. The findings underscore the clinical importance of thorough cognitive assessment in schizophrenia and raise the possibility of re-evaluating diagnostic criteria to incorporate the spectrum of neurodevelopmental presentations.

The study relies on existing datasets, which might have biases in sample selection or ascertainment. The phenotypic characterization of schizophrenia probands carrying NDD variants was limited, potentially missing subtle neurodevelopmental features. Further studies with larger, more comprehensively phenotyped samples are needed to solidify these findings and address potential biases.