LAUSR

Background The relationship between schizophrenia (SZ) and human immunodeficiency virus (HIV) infection is complicated epidemiologically and genetically. Although observational studies have described their co-occurrence and their joint relationship with risky sexual behavior (RSB), their genetic correlations are not well studied. Methods We performed an extensive search for genetic factors common to SZ and HIV using publically accessible summary statistics from genome-wide association studies of HIV infection and schizophrenia. To study the relationship of these disorders with risky sexual behavior (RSB), 2379 European Americans were genotyped and assessed for RSB score using the Semi-Structured Assessment for Drug Dependence and Alcoholism (SSADDA). Genetic relationships between traits were analyzed in three ways: linkage disequilibrium (LD) score regression to estimate genetic correlation; GPA (Genetic analysis incorporating Pleiotropy and Annotation) to test pleiotropy and identify pleiotropic loci; and polygenic risk scores (PRS) for SZ and HIV to predict RSB using linear regression. Bipolar disorder, height and body mass index (BMI) were tested as negative controls. Results We found that SZ and HIV have a positive genetic correlation, which is highly significant both with (cor=0.2, p=0.001) and without (cor=0.17, p=0.002) inclusion of the MHC region. Consistently, our pleiotropy analysis (p=5.31E-28) showed that a majority of the shared genetic variants have the same effect direction (shared-effect pleiotropic SNPs), meaning that the same allele tends to increase or decrease the risk of both SZ and HIV. Pleiotropic SNPs that influence schizophrenia and HIV infection with the same effect direction (shared) were enriched for chromatin assembly (p=1.3E-9) nucleosome organization (p=4.4E-9), and protein-DNA complex assembly (p=1.8E-8). SNPs with opposite effect directions (antagonistic) were enriched for the regulation of synaptic transmission (p=8.2E-6) and neurotransmitter transport (p=5.5E-5), neuronal differentiation (p=3.8E-5), and the regulation of nerve impulse transmission (p=2.4E-5). SZ PRS computed with antagonistically pleiotropic SNPs significantly predicted RSB score (p=0.019), but SZ PRS based on either shared pleiotropic SNPs or all SNPs did not predict RSB. Discussion The epidemiologic correlation between schizophrenia and HIV can partly be explained by overlapping genetic risk factors, which are related to risky sexual behavior. These findings call for further study of these SNP sets, and when available, using genetic risk scores from them to test other phenotypes that could yield important insights into these very serious disorders and behaviors associated with them.