The mutant allele frequencies at SNPs -1082 and -592 in our cohort were 0.33 and 0.32 respectively. Chi square tests confirmed the allele frequencies to be in Hardy-Weinberg equilibrium for the entire cohort and for subgroups analyzed in this study. Linkage disequilibrium estimates between the two SNPs were calculated as r2
= 0.12 and D′= 0.70 using Haploview [14
]. The SNPs specified four haplotypes termed 1 (AC), 2 (AA), 3 (GC) and 4 (GA) which occurred at approximate frequencies of 37 %, 30 %, 29 % and 3% respectively.
Association between IL-10 genotype and susceptibility to HIV infection
There was no significant difference in the distribution of IL-10 genotypes for the -1082 position between the HIV-1 negative and HIV-1 positive groups (p= 0.1443) (). For the 222 participants enrolled when HIV negative and followed prospectively, Kaplan-Meier survival analysis of time to HIV-1 seroconversion for genotypes at the -1082 position () showed no significant differences in risk of HIV-1 acquisition between genotype -1082AA versus -1082GG or -1082AG. Using the Cox proportional hazards model, if an individual was -1082AA the hazard ratio of becoming HIV-1 infected was 2.14 (95% CI: 0.99-4.63, p=0.0543).
The influence of IL-10 promoter polymorphisms on HIV-1 susceptibility
However, there was a significant association between HIV-1 status and the IL-10 -592 genotype (), due to the large proportion of HIV-positive -592AA individuals as compared to the other genotypes (p=0.0033). In the analysis of high-risk HIV-1 negative individuals followed prospectively, individuals possessing the -592AA genotype were significantly more likely to become HIV infected compared to individuals who were -592CA or -592CC (). Using the Cox proportional hazards model, if an individual was -592AA they were at increased risk of becoming HIV infected with a hazard ratio of 3.0 (95% CI: 1.28-7.08, p=0.0117).
There were three haplotypes in this cohort with frequencies above 5%-haplotypes 1, 2 and 3. Only haplotypes 2 and 3 showed significant association with HIV-1 status. In normal logistic regression analysis, haplotype 2, predicted to result in low IL-10 production was associated with 1.8 times higher likelihood of being HIV-positive (95% CI 1.21-2.74, p=0.0038). In contrast, haplotype 3 (high IL-10 producer) was associated with 0.59 less likelihood of HIV acquisition (95% CI 0.36-0.96, p=0.032).
The effect of IL-10 promoter polymorphisms on viral load and CD4+ T cell counts
For the 64 individuals who became infected with HIV-1, we investigated whether there were differences in viral load and CD4+ T cell counts according to IL-10 genotypes between the phases of infection. We performed analysis at ≤ 3months post infection (acute phase) and 6-12 months post-infection (early chronic phase) in median viral load or CD4+ T cell count between the three genotypic groups for both SNPs at positions -1082 and -592 ().
The influence of IL-10 promoter polymorphisms on primary HIV-1 pathogenesis
Analysis of the overall change in viral load based on -1082 genotype (), shows that individuals possessing the -1082GG genotype (high IL-10 producer) had significantly higher median viral loads during the acute phase compared to -1082AA individuals (p=0.0002) and -1082AG individuals (p=0.0027). During the early chronic phase, however, there was no longer a significant difference between the three groups. Analysis of the overall change in CD4+ T cell count (), showed that during the acute phase, the -1082GG group had the lowest median CD4+ T cell count and this was significantly different from the -1082AA group (p= 0.0120). In comparison, during the early chronic phase of infection the -1082GG group was not significantly different from the other groups.
During the acute phase, the -592CC genotype (high IL-10 producer), did not have significantly different median viral load compared to -592CA or -592AA, p= 0.1292 and 0.0590 respectively (). During the early chronic phase, the -592AA group (low IL-10 producer) had a significantly higher median viral load compared to the -592CA and -592CC groups (p=0.0022 and 0.0075, respectively). However, there were no significant differences in CD4+ T cell counts between -592 variants at either the acute or early chronic phases of infection ().
During the acute phase of infection, haplotype 2 trended towards a negative effect on viral load (p= 0.0808), while haplotype 3 had a significantly positive effect on viral load (p=0.0002). However, during the early chronic phase haplotype 2 now had a significantly positive effect on viral load (p=<0.0001) and the positive effect on viral load of haplotype 3 was no longer significant (p=0.4930).