UCLA

Background: Persons living with HIV/AIDS (PLWHA) are at elevated risk for non-Hodgkin lymphoma relative to HIV-negative individuals. AIDS-related non-Hodgkin lymphoma (AIDS-NHL) etiology is characterized by B-cell activation and chronic inflammation in the context of HIV; prior work has highlighted the importance of inflammatory biomarkers in predicting subsequent AIDS-NHL diagnosis. General-population risk factors for NHL include prior family history, suggesting a heritable genetic component to NHL risk that may encompass inflammation-related genes. Previous genome-wide association studies (GWAS) have investigated NHL in the general population, but no AIDS-NHL GWAS has yet been published. We therefore investigate single-nucleotide polymorphisms for association with AIDS-NHL risk using a candidate-gene study in 700 HIV-positive men (178 AIDS-NHL cases), and a GWAS in 1,949 HIV-positive men (172 AIDS-NHL cases), from the Multicenter AIDS Cohort Study (MACS). Pathway analyses of GWAS results complement these two studies.

Methods: Candidate-gene study (30 SNPs; 24 genes): matched case-control design; conditional logistic regression, semi-Bayes correction and multiple imputation for missing covariate data in SAS. GWAS (n=4.86 million SNPs): imputation in MINIMAC3 for missing genotype data; logistic regression on genotype probabilities using SNPTEST v2.5.2 and adjustment for principal components from R package SNPRelate; regional plots using R package LocusExplorer. Pathway analyses: analysis of GWAS p-values using 13,094 gene sets in PASCAL and 6,212 gene sets in VEGAS2.

Results: Candidate-gene study: significant inverse association (dominant OR=0.68; 95%CI 0.47-0.99; log-additive OR=0.71, 95%CI 0.51-0.99) between NHL risk and SNP rs6815391 (3’ UTR, REX1/ZPF42, 4q35.2). GWAS: genome-wide-significant signal (4q33; top SNP rs2195807; p=1.48E-08; white-only p=1.93E-07) in the vicinity of uncharacterized noncoding variant LOC100506122. Pathway analyses: repeated occurrence of gene sets capturing inflammatory processes and muscle fiber/ cytoskeletal integrity at or near the top of most scenarios.

Conclusion: Each method yielded associations with different aspects of AIDS-NHL biology. The candidate-gene study suggests involvement of the REX1/ZPF42 SNP rs6815391 in HIV replication and production of anti-inflammatory cytokines; the GWAS points toward an as-yet-uncharacterized locus (LOC100506122); pathway analyses implicate myosin genes and pathways that may be involved in early stages of B-cell activation. Further candidate-gene, in vivo or in vitro studies may clarify the biological plausibility of observed associations.