Genome-wide association analysis of treatment resistant schizophrenia for variant discovery and polygenic assessment

Abstract

Background: Treatment resistant schizophrenia (TRS) is broadly defined as inadequate response to adequate treatment and is associated with a substantial increase in disease burden. Clozapine is the only approved treatment for TRS, showing superior clinical effect on overall symptomatology compared to other drugs, and is the prototype of atypical antipsychotics. Risperidone, another atypical antipsychotic with a more distinctive dopamine 2 antagonism, is commonly used in treatment of schizophrenia. Here, we conducted a genome-wide association study on patients treated with clozapine (TRS) vs. risperidone (non-TRS) and investigated whether single variants and/or polygenic risk score for schizophrenia are associated with TRS status. We hypothesized that patients who are treated with clozapine and risperidone might exhibit distinct neurobiological phenotypes that match pharmacological profiles of these drugs and can be explained by genetic differences. The study population (n = 1286) was recruited from a routine therapeutic drug monitoring (TDM) service between 2005 and 2022. History of a detectable serum concentration of clozapine and risperidone (without TDM history of clozapine) defined the TRS (n = 478) and non-TRS (n = 808) group, respectively.

Results: We identified a suggestive association between TRS and a common variant within the LINC00523 gene with a significance just below the genome-wide threshold (rs79229764 C > T, OR = 4.89; p = 1.8 × 10^-7). Polygenic risk score for schizophrenia was significantly associated with TRS (OR = 1.4, p = 2.1 × 10^-6). In a large post-mortem brain sample from schizophrenia donors (n = 214; CommonMind Consortium), gene expression analysis indicated that the rs79229764 variant allele might be involved in the regulation of GPR88 and PUDP, which plays a role in striatal neurotransmission and intellectual disability, respectively.

Conclusions: We report a suggestive genetic association at the rs79229764 locus with TRS and show that genetic liability for schizophrenia is positively associated with TRS. These results suggest a candidate locus for future follow-up studies to elucidate the molecular underpinnings of TRS. Our findings further demonstrate the value of both single variant and polygenic association analyses for TRS prediction.

Fig. 1

Manhattan plot showing the associations with risk of TRS. The − log10 transformed p-values for each SNP are shown on the y-axis and chromosomal positions are indicated on the x-axis. The dashed black line represents the standard GWAS p-value threshold of 5 × 10⁻⁸. The most significant SNP is highlighted by text

Fig. 2

Impact of rs79229764 C > T on brain tissue gene expression from CommonMind Consortium data. The 3 most significant gene expression associations were illustrated. Normalized expression results are shown on the y-axis, while rs79229764 C > T genotypes are indicated on the x-axis. Gene names and unadjusted p-values from association analyses are shown at the top part of the plots

Fig. 3

Forest plot showing the association between treatment resistant schizophrenia and schizophrenia polygenic risk score (PRS SCZ), smoking, age, and sex. Effects are reported as odds ratios (95% confidence interval)