Pharmacologic and genetic evidence converge on mechanisms of psychotic illness

Abstract

Idiopathic and substance-induced forms of psychotic illness afflict millions of people worldwide, and it is largely unknown whether these two forms emerge through the same molecular mechanisms. Though genetic studies have implicated thousands of genes in idiopathic psychotic illnesses (e.g., schizophrenia), consensus is lacking regarding which of these genes are most likely to treat psychotic illness when modulated pharmacologically and, as a result, antipsychotic medications targeting these genes have yet to be developed. Previous studies suggest that one way to determine if a candidate target gene is likely to lead to an effective treatment for a given illness is if the gene is implicated by multiple lines of evidence (e.g., genetic, pharmacologic). Here, pharmacologic, genetic, and clinical data were leveraged to determine if the idiopathic and substance-induced forms of psychotic illness are related to one another through a common set of genes. A set of medications that cause psychotic illness as a side effect ("propsychotics") were identified by analyzing 15 million medication side effects reports from over 100 countries. Gene products targeted by propsychotics overlapped significantly with those targeted by antipsychotics and for many of the overlapping targets propsychotics act through a mechanism that was qualitatively the opposite of the mechanism through which antipsychotics act (e.g., activation vs. inhibition). Propsychotic and antipsychotic target genes were significantly enriched for genes implicated in schizophrenia by rare loss-of-function genetic variation but not for genes implicated in schizophrenia by common genetic variation. Only one gene - GRIN2A, encoding the GluN2A subunit of the NMDA glutamate receptor - was implicated in psychotic illness by propsychotics, rare loss-of-function genetic variation, and common genetic variation. Mining genetic data from a diverse cohort of 30,000 adults treated in a New York City health system, a carrier of a rare loss-of-function variant in GRIN2A with severe psychotic illness was identified with a clinical course notable for psychotic symptoms and cognitive deficits that are not targeted by current antipsychotics. Altogether, this report shows how integrating pharmacologic, genetic, and clinical data from large cohorts can prioritize target genes for novel drug development and align the prioritized targets with specific clinical presentations.

Fig. 1. Summary of the workflow used to define propsychotics in VigiBase.

SE side effect.

Fig. 2. Proportion of propsychotics linked to MedDRA terms.

The vertical axis shows MedDRA psychosis side effect terms. The horizontal axis shows the proportion of propsychotics that were linked to each MedDRA psychosis side effect term. Only MedDRA psychosis side effect terms linked to greater than 1% of propsychotics are shown.

Fig. 3. Proportion of propsychotics in medication classes.

The vertical axis shows ATC Level 3 medication class, with bolded labels indicating drugs classified under the ATC Level 1 term “Nervous System”. The horizontal axis shows the proportion of propsychotics that were linked to each ATC Level 3 medication class. Only ATC classes linked to greater than 1% of propsychotics are shown.

Fig. 4. Comparative analysis of target gene mechanisms of action for antipsychotics and propsychotics.

A For every significant antipsychotic target gene mechanism of action there is an entry on the x-axis. The top panel shows on the y-axis the proportion of antipsychotics that exerted an activating mechanism of action on the target gene (blue) and the proportion of antipsychotics that exerted an inhibiting mechanism of action on the target gene (orange; represented as negative values), and only significant values are shown. The bottom panel shows on the y-axis the proportion of propsychotics that exerted an activating mechanism of action on the target gene (blue) and the proportion of antipsychotics that exerted an inhibiting mechanism of action on the target gene (orange; represented as negative values), and significant values are indicated by an asterisk (*). Target genes are grouped by neurotransmitter receptor class as indicated by the facet labels. B For every significant propsychotic mechanism of action that did not have a significant mechanism of action for antipsychotics there is an entry on the y-axis. On the x-axis is shown the proportion of propsychotics that exerted an activating mechanism of action on the target gene (blue) and the proportion of antipsychotics that exerted an inhibiting mechanism of action on the target gene (orange; represented as negative values), and only significant values are shown.