A genome-wide association study of HIV-1 infection in African American and European American injection drug users. E. Johnson1, D. Hancock1, J. Levy2, G. Page3, S. Novak1, C. Glasheen1, N. Gaddis2, N. Saccone4, J. Rice5, Q. Wang6, M. Moreau6, K. Doheny7, J. Romm7, A. Brook6, L. Bierut5, A. Kral8 1) Behavioral Heallth Epidemiology, RTI International, Research Triangle Park, NC; 2) Research Computing Division, RTI International, Research Triangle Park, NC;; 3) Genomics, Statistical Genetics, and Environmental Research Program, RTI International, Atlanta, GA; 4) Department of Genetics, Washington University in St. Louis, MO; 5) Department of Psychiatry, Washington University in St. Louis, MO; 6) Rutgers University Cell and DNA Repository (RUCDR), Riscataway, NJ; 7) Center for Inherited Disease Research (CIDR), Johns Hopkins University, Baltimore, MD; 8) Urban Health Program, RTI International, San Francisco, CA.

   Nearly 40 million people worldwide are infected with HIV. Host genetic factors likely influence susceptibility to acquiring HIV upon exposure. However, the only validated genetic association with HIV-1 infection is the 32 base-pair deletion in the CCR5 gene, which is rare in the population. To identify and characterize novel genetic risk factors, we conducted GWA studies of HIV-1 infection in African American and European American injection drug users. Unlike the prior genetic studies, we focused on injection drug users whose transmission may be primarily parenteral rather the than mucosal or mother-to-infant, and used HIV-1 negative controls with individually assessed high probability of exposure. HIV-1 positive cases and negative controls were frequency-matched on behavioral and demographic risks of exposure. DNA was extracted from stored serum at RUCDR and treated with Illumina FFPE restoration prior to genotyping at CIDR on the Illumina Omni1-Quad array. Following quality control procedures for genotyped SNPs and participants, the analysis data set included ~700,000 autosomal SNPs in 2,010 African Americans (630 cases, 1,380 controls) and 1,149 European Americans (335 cases, 814 controls). In the separate ethnic groups, additive SNP genotypes were tested for association with HIV-1 infection using logistic regression models adjusted for age, gender, behavioral risk class, recruitment pre- or post- availability of antiretroviral therapy, and 10 principal component eigenvalues for population stratification. The ethnic-specific GWA results were then combined in a fixed-effects meta-analysis. The lowest P-value (meta-analysis P=1.88x10-6) was observed for a regulatory SNP on chromosome 9p13, located in a transcription factor binding site upstream of a gene essential to B-cell differentiation. B cells, a primary component of the adaptive immune system, may play a role in cell-to-cell transmission of the HIV virus and are known to be affected during disease progression (e.g., expansion of immature transitional B-cells and B-cell exhaustion). The SNP is associated with HIV-1 infection in both ethnic groups, but occurs more frequently in African Americans (MAF=21.9%) than in European Americans (MAF=1.3%). Analyses using 1000 Genomes imputed SNPs and genotyping of Hispanic American samples are underway to further characterize this biologically plausible gene region and others that may help determine the complex etiology of host response to HIV-1.

You may contact the first author (during and after the meeting) at