Rare copy number variants and phenotypic variability in schizophrenia

One of the greatest barriers to clarifying the core biological processes that underlie schizophrenia (SCZ) and improving patient outcomes is the clinical and genetic heterogeneity of the disorder. For example, while psychosis onset is most common in early adulthood, some patients experience psychotic symptoms as early as childhood. Similarly, while cognitive functioning is impaired on average among SCZ patients compared to controls, there is considerable variability between patients. Determining whether severe phenotypes such as borderline intellectual functioning or child-onset psychosis can be explained by distinct genetic profiles, such as burden of known neuropsychiatric copy number variants (CNVs) or rare deletions affecting genes involved in early brain development is an important question with implications for personalized medicine. CNVs were called in a cohort of 618 patients with SCZ spectrum disorders, as well as 231 relatives of SCZ patients, and 586 unaffected controls. Logistic regression tested for differences in rates of neuropsychiatric CNVs among cases versus non-cases, and for associations between borderline intellectual functioning and child-onset psychosis and burden of known neuropsychiatric-associated CNVs or rare deletions affecting genes involved in distinct neurodevelopmental processes for SCZ spectrum patients. Neurodevelopmental gene-sets were defined previously via weighted gene co-expression network analysis of transcriptomic data from 1,061 brain samples from the developing human brain (Forsyth et al., 2020). Rates of SCZ-associated and broader neurodevelopmental disorder (NDD)-associated CNVs were significantly elevated in cases (up to 2.4%) compared to non-cases (up to 0.9%). SCZ- and NDD-associated CNVs were found at non-significantly higher rates in child-onset (up to 8.3%) compared to later-onset psychosis (up to 2.1%) and were associated with significantly higher likelihood of borderline intellectual functioning across SCZ spectrum patients (odds ratio up to 7.0). There were no associations between rare deletions affecting genes in any neurodevelopmental gene-set and child-onset psychosis; however, deletion of genes in a fetal gene regulatory gene-set and a glia maturation gene-set were associated with increased likelihood of borderline intellectual functioning. The association between deletion of fetal gene regulatory genes and borderline intellectual functioning remained nominally significant after excluding patients with known neuropsychiatric CNVs, and was also replicated among controls and relatives of patients with SCZ spectrum disorders. Results suggest that poor cognitive functioning in SCZ is associated with presence of known neuropsychiatric CNVs, and additionally with deletion of genes involved in regulating gene expression during early brain development. Associations between neurodevelopmental gene-sets and severe phenotypes in schizophrenia may offer opportunities to prioritize pathogenic genes outside known risk loci.