De novo mutations from sporadic schizophrenia cases highlight important signaling genes in an independent sample

Abstract

Schizophrenia is a debilitating syndrome with high heritability. Genomic studies reveal more than a hundred genetic variants, largely nonspecific and of small effect size, and not accounting for its high heritability. De novo mutations are one mechanism whereby disease related alleles may be introduced into the population, although these have not been leveraged to explore the disease in general samples. This paper describes a framework to find high impact genes for schizophrenia. This study consists of two different datasets. First, whole exome sequencing was conducted to identify disruptive de novo mutations in 14 complete parent–offspring trios with sporadic schizophrenia from Jerusalem, which identified 5 sporadic cases with de novo gene mutations in 5 different genes (PTPRG, TGM5, SLC39A13, BTK, CDKN3). Next, targeted exome capture of these genes was conducted in 48 well-characterized, unrelated, ethnically diverse schizophrenia cases, recruited and characterized by the same research team in New York (NY sample), which demonstrated extremely rare and potentially damaging variants in three of the five genes (MAF < 0.01) in 12/48 cases (25%); including PTPRG (5 cases), SCL39A13 (4 cases) and TGM5 (4 cases), a higher number than usually identified by whole exome sequencing. Cases differed in cognition and illness features based on which mutation-enriched gene they carried. Functional de novo mutations in protein-interaction domains in sporadic schizophrenia can illuminate risk genes that increase the propensity to develop schizophrenia across ethnicities.

Introduction

The heritability of schizophrenia is established from a century of family, twin and adoption studies, but genes of major effect for the devastating psychotic disorder remain elusive. Genome Wide Association Studies (GWAS) show a multitude of risk alleles of small effect sizes that are related to diverse psychopathologies (McCarroll et al., 2014). It is postulated that major genes for schizophrenia arise de novo to replenish the population risk and counteract the reduced reproductive fitness of the disease (Malaspina, 2001). If so, de novo mutations from sporadic cases may point to genes whose variation is more specifically associated with schizophrenia. There is a large body of evidence that de novo mutations are enriched in disease cases versus controls, as e.g., in autism (De Rubeis et al., 2014, Dong et al., 2014) and schizophrenia (Xu et al., 2011, Fromer et al., 2014, Purcell et al., 2014, Takata et al., 2014). Methodological rigor in examining family history and localization of these mutations within protein interaction networks as well as functional domains may enhance the probability that particular de novo mutations are relevant to the disease.

Penrose (1955) first proposed that advancing paternal age was the major source of de novo mutations, which was further explicated by Crow (2000). We hypothesized that gene variants associated with the risk for schizophrenia could be introduced into the population in association with paternal age de novo and then persist in the gene pool to contribute to familial illness, perhaps individually influencing the specific disease phenotype. We tested this hypothesis in a two-stage approach.

The first epidemiological demonstration that steadily increasing schizophrenia risk accompanies advancing paternal age was in the Jerusalem Perinatal Schizophrenia Study (JPSS) (Malaspina et al., 2001), wherein each decade increased schizophrenia risk by 1.4-fold, with the relative risk for offspring of fathers > 45 years being 3-fold greater than for 20–24 year-old fathers. After controlling for maternal age and other factors, advancing paternal age, beginning at 25 years, explained 26% of schizophrenia risk, comparable to estimates in other cohorts (Brown et al., 2002, Zammit et al., 2003, Sipos et al., 2004, Tsuchiya et al., 2005).

In our clinical research, we cast a net for genes associated with schizophrenia, sequencing 14 sporadic offspring–parent trios from the JPSS sample. Our findings identified a handful of de novo mutations in genes that we subsequently sequenced in 48 non-related individuals from our New York sample. These candidate genes all represent important central nervous system relevant signaling proteins, which may belong to pathways associated with schizophrenia. The samples are smaller than in many studies because our hypotheses are specific: genes showing functional de novo mutations in sporadic cases can shine a window on high impact genes for the illness per se.