Background:
Genome-wide association studies (GWAS) have shown that genetic variants in the MHC region are associated with spontaneous HIV-1 control. How these variants affect immune functioning is, however, poorly understood. We used a functional genomics approach in which we integrated GWAS with transcriptomic, plasma proteomic and functional immunological data to understand how genetics contribute to HIV control.
Methods:
1380 people living with HIV of European ancestry were included as part of the 2000HIV study (clinicaltrials.gov NTC03994835). To find genetic variants associated to control, a GWAS was performed in 67 HIV controllers (HIC) and 272 matched non-HIV controllers on ART (non-HIC) using a logistic regression model. To unravel how genetics affects phenotype, we performed quantitative trait locus (QTL) mapping with data from the 2000HIV cohort. A linear regression model was used to associate genetic dosages with: 1) cytokine production upon ex vivo PBMC stimulation with various stimuli (cQTLs), 2) levels of 2367 plasma proteins as measured by Olink technology (pQTLs), 3) bulk RNA expression in PBMCs (eQTLs) measured by RNA-sequencing in 1146, 1222 and 1208 individuals, respectively. All genome-wide significant (P < 5∙ 10-8) QTLs were intersected with suggestive GWAS loci.
Results:
In agreement with previous GWAS, the strongest association with HIV control was identified in the MHC locus (Fig A), where we found 8 independent suggestive SNPs (P < 1 ∙ 10-5, r2 < 0.5). SNP rs2524092-C (P = 2.44 ∙ 10-6, OR = 3.4) was associated with higher GZMA and KLRF1 and lower KIR2DL2, KIR2DL3 and MICA/MICB plasma levels (Fig B), lower HLA-C and higher HLA-H and PSORS1C3 gene expression in PBMCs. Additionally, rs2524092-C is in linkage disequilibrium (r2 = 0.52) with rs9264388-T (P = 1.13 ∙ 10-5, OR = 4.33), which was associated with lower IL-1β and TNF production upon PBMC stimulation with CMV (Fig B).
Conclusions:
We showed that genetic variants associated to HIV control are associated to plasma and gene expression levels of NK cell receptors, their ligands MICA/MICB and HLA-C, and effector molecule granzyme A. This suggests that these variants might contribute to HIV control by modulating NK cell function. The association of variants in the MHC locus with immune responses against CMV hints to a shared genetic basis of immune responses against HIV and CMV. Finally, we showed that our QTL mapping studies are valuable resources that can aid understanding how genetics contributes to disease outcome in PLHIV.
