Background: MicroRNAs (miRNA) regulate gene expression through binding to the 3' untranslated region of target messenger RNA (mRNA) and have been implicated in lymphomagenesis. Only a few studies have been conducted to evaluate the effects of miRNA-related genes. In this study, we will test hypotheses to determine whether single nucleotide polymorphisms (SNPs) located within miRNA binding sites, miRNA coding regions, or miRNA processing genes, (collectively referred to as miRNA-related SNPs), are associated with susceptibility of AIDS-associated non-Hodgkin lymphoma (AIDS-NHL) among men from the Multicenter AIDS Cohort Study (MACS).
Methods: We genotyped 25 miRNA-related SNPs in 180 AIDS-NHL cases and 529 matched HIV-infected controls. Adjusted odds ratios (ORadj) and 95% confidence intervals (CIs) were calculated using conditional logistic regression, controlling for age, CD4+ T-cell count, HIV viral load, prior AIDS diagnosis, history of anti-retroviral therapy, race, and history of HCV infection, to estimate the associations between miRNA-associated SNPs and overall, systemic and central nervous system (CNS) AIDS-NHL. The semi-Bayes (SB) approach was used to correct for multiple comparisons. Through the incorporation of a normal prior of null-association (prior odds ratio, ORprior = 1.00; 95% prior limits: 0.25, 4.00), we proposed to pull observed associations toward the null, emphasizing our uncertainty. To give a comprehensive overview of our data and results, we presented both suggested (alpha=0.10) and statistically significant results (alpha=0.05), while noting those that remained evident after SB correction. We did place highlighted emphasis on significant results which remained evident after the SB approach as we were more confident in the report of these results (GEMIN3 rs197412, miR-196a rs11614913, and HIF1A rs2057482). Further, results were focused to those from analyses imputed for HIV viral load at set point and CD4+ T-cell count at reference date. Last, we estimated the association between SNPs within miRNA processing genes and mean loge miRNA serum levels among a subgroup of 77 MACS participants, comprised of 61 AIDS-NHL cases and 16 HIV-infected controls. Mean ratios (MRadj) and 95% CIs were calculated using linear regression, controlling for AIDS-NHL case-control status, CD4+ T-cell count at date of serum sample, and race.
Results: GEMIN3 rs197412, a non-synonymous SNP within a gene involved in miRNA processing and maturation, was associated with AIDS-NHL overall (ORadj= 1.35 per variant allele; 95% CI: 1.03-1.78). In subgroup analyses, an inverse association was observed between miR-196a rs11614913 and CNS HIV/AIDS NHL (CT/TT vs. CC ORadj= 0.43; 95% CI: 0.22-0.87). The variant allele of HIF1A rs2057482 was associated with increased risk of systemic AIDS-NHL (ORadj= 1.83 per variant allele; 95% CI: 1.16-2.90), whereas the same allele was associated with decreased risk of CNS AIDS-NHL (ORadj= 0.38 per variant allele; 95% CI: 0.17-0.87). These miRNA-related SNP associations with AIDS-NHL remained evident after SB correction. In the evaluation of genotype-phenotype associations, there was a suggested association between carriership of the variant allele (C) of GEMIN3 rs197412 with higher miRNA-222 serum levels compared to those with the referent genotype (MRadj= 1.21 per variant allele; 95% CI: 0.98-1.49), suggesting a potential functional effect of GEMIN3 rs197412 in relation to miRNA-222 expression.
Conclusion: These results suggest that miRNA-related SNPs are associated with AIDS-NHL susceptibility. Specifically, our study demonstrated that GEMIN3 rs197412 was associated with AIDS-NHL susceptibility, and that genetic variation within this miRNA processing gene may influence subsequent miRNA expression. Further, miR-196a rs11614913 and HIF1A rs2057482 AIDS-NHL were also associated with AIDS-NHL, highlighting the diverse array of biologic mechanisms underlying AIDS-lymphomagenesis.