M172. POLYGENIC RISK SCORES ANALYSES IN ANTIPSYCHOTIC-INDUCED WEIGHT GAIN

Abstract

Background: Antipsychotic drugs are widely used for the treatment of schizophrenia (SCZ). Despite clinical efficacy, antipsychotics are associated with severe side effects including antipsychotic-induced weight gain (AIWG). Specifically, significant AIWG (i.e., >7% from baseline) is observed in more than 30% of individuals treated with antipsychotics and frequently leads to metabolic disturbances such as Type-2 diabetes (T2D) and cardiovascular diseases. Although twin and family studies have pointed to high heritability for AIWG (h2=0.6-0.8), there are currently no established predictors of AIWG given that it involves a combination of genes and non-genetic factors. Given the complex phenotypic nature of AIWG, polygenic risk scores (PRS) of certain morbidities, which combines thousands of common variants weighted by their effect size, may provide a measure of genetic liability for AIWG. Therefore, we aimed to investigate whether PRSs based on the genome-wide association studies (GWAS) for SCZ, body mass index (BMI) and diabetes (Type 1 & 2) were associated with AIWG. Method(s): For our analysis, we used two samples, including (1) a subset of individuals diagnosed with SCZ with AIWG phenotyping from the Clinical Antipsychotic Trials in Intervention Effectiveness (CATIE; n = 189, see Brandl et al., 2016), and (2) the Toronto cohort, a multi-study cohort of individuals diagnosed with SCZ with AIWG (n = 151, see Brandl et al., 2014). Within these two cohorts, our phenotypes of interest included the percentage of BMI/weight change from baseline to end-of-treatment, as well as the presence/absence of significant weight gain (>=7% weight change). We investigated associations between PRSs of SCZ, BMI, and diabetes and AIWG using regression models, corrected for age, sex, risk medication for AIWG, and duration of the study, using PRSice-2. We used the Psychiatric Genomics Consortium schizophrenia GWAS to calculate PRSs for SCZ. We used GWAS summary statistics from the GWAS Catalog of BMI and diabetes in individuals of European ancestry. For BMI, we used one dataset for BMI (i.e., GCST006900: 2,336,269 variants across up to 700,000). For Type-1 diabetes (T1D), we used one dataset from the GWAS catalog (ID: GCST005536) which included 123,130 variants across 6,683 cases, 12,173 controls, 2,601 affected sibling-pair families, and 69 trios. Likewise, we used three datasets forT2D (i.e., GCST006801: 8,404,432 variants across 4,040 cases and 113,735 controls, GCST007517: 133,871 variants across up to 48,286 cases and up to 250,617 controls, and GCST007518: 133,586 variants across up to 48,286 cases and up to 250,617 controls). Result(s): The demographics and characteristics of the subjects were as follows: 1) CATIE sample (N = 189) [age=41.4+/-12, male: 151 (79.9%), baseline BMI: 28.83+/-5.43, mediations: olanzapine n = 63; quetiapine n = 67; risperidone n = 59]; and 2) our local sample (N = 156) [age=36.7+/-11.2, male: 98 (63.8%), baseline weight: 78.97 +/-15.60, mediations: clozapine n = 68; risperidone n = 27; olanzapine n = 18; aripiprazole n = 16; others = 24]. We observed no significant associations of PRS for SCZ or BMI with AIWG. We observed significant associations of PRS for T1D with percentage BMI change from baseline at PT=0.0031 (R2=0.04, p = 0.01), as well as weight gain at PT=0.00015 (R2=0.09, p = 0.004) in the CATIE sample only. Furthermore, we observed significant associations of PRS for T2D status with percentage weight change from baseline at PT=0.49 (R2=0.03, p = 0.03), as well as weight gain at PT=0.0001 (R2=0.06, p = 0.02) in the Toronto cohort only. Conclusion(s): To the best of our knowledge, this is the first genetic association study evaluating whether PRSs for SCZ, BMI, or diabetes are associated with AIWG in patients with SCZ. We found that there was a genetic overlap between the risk of diabetes and risk for developing AIWG. Limitations of this study include a small sample size of the target data and that only patients of European ancestry were analyzed. Nonetheless, our suggestive results highlight the importance of studying metabolic-related disorders to unravel the mechanism of AIWG. Further studies with larger sample sizes and individuals of various ethnic ancestries are required.