Rare Copy Number Variation in Schizophrenia and Implications for Treatment

For over a decade, consortia efforts have been gainfully merging sources of genetic information—common and rare genetic variation—to elucidate the underlying biology of schizophrenia. Signal due to common variation on the genome, with millions of single variants of small effect measured in aggregate, is now enhanced by fine-mapping and the integration of rare variant signal.¹ Rare copy number variants (CNVs), reflecting large swathes of duplications or deletions that are penetrant and deleterious, contribute significantly to schizophrenia and to neurodiversity more broadly.^2,3 Common and rare variations, including these rare CNVs, could interact with myriad environmental exposures, and likely contribute to missing gene regulation yet to be characterized.⁴

Despite this progress, understanding the basic clinical implications of genetic variation has been slow going. Early genome-wide association studies (GWAS) and gene-based analyses mapped common genetic variation to broad and narrow definitions of schizophrenia⁵ and to variation in positive, negative, and thought disorder symptom domains of psychosis, informed by item-level symptom data.⁶ These studies observed that common variation accounted for some of the variability in diagnostic and symptom-level data. GWAS-based genetic correlation matrices and genetic structural equation models of the phenome,^7,8 and their related heatmaps and network graphs, have produced beautiful tapestries and mosaics for the psychiatric nosologist. And GWAS-based risk metrics (ie, polygenic scores), because they are based on millions of common variants, have become increasingly informative insofar as they provide statistically powerful models of the relationships of normally distributed genetic risks with important clinical data (for example, suicidal behavior, addiction, or other extreme outcomes) in new, independent cohorts.

But small effect sizes remain a limiting factor to clinical application—common variation alone can only “predict” low population base rate phenotypes within the context of case–control designs, and not in the clinic or the general population.

Rare CNVs, more difficult to detect and measure across cohorts, do not solve the problem of statistical power inherent to common variant studies of low base rates or severe clinical outcomes. However, there is good basis for studying rare CNVs conferring high risk as functionally and clinically informative to the conceptualization and treatment of psychotic symptoms. CNV-defined genetic subgroups may be uniquely informative. To the extent that rare CNVs can be identified in large cohorts of individuals with schizophrenia, their comparative association with symptoms and prognosis may be studied.

In this issue, Farrell et al. leverage data on rare, penetrant CNVs in a large cohort to help inform the clinical problem of treatment resistance.⁹ Antipsychotics fail to treat a large proportion of people with schizophrenia, and these preliminary data implicate copy number variation on chromosome 15 in a region of relatively high penetrance associated with schizophrenia¹⁰ (2%–9%). This report, building on a number of previous accomplishments in the study of rare variation in psychosis, provides some of the first evidence of a subtle yet potentially clinically relevant link between rare variation and persistent nonresponse to antipsychotic medications.

Importantly, to the extent that rare CNV and common and rare genome-wide variant effects together explain antipsychotic medication response, there is the hope of better understanding gene regulation in schizophrenia, and of informing clinical care for individuals with specific CNVs. Apart from the potential of this research to inform gene regulation, what makes this study exceptional (and what has remained elusive to the field) are tractable links between genetic variation and clinical intervention in schizophrenia.

Funding statement

This publication has been supported in part by NIH R01MH123489 (PI: Docherty) and by the Huntsman Mental Health Institute.

Conflict of Interest

The author has no conflicts of interest to report. The views expressed here are independent and do not reflect the views of the International Society for Psychiatric Genetics.