Effects of genetic polymorphism in MHC, KIR, and related loci on human disease

Division Of Basic Sciences - Nci

Investigators

Linked publications, trials & patents

Abstract

HLA class I (HLA-I) allotypes vary widely in their dependence on tapasin (TAPBP), an integral component of the peptide loading complex, to present peptides on the cell surface. We have determined tapasin dependence (TD) values for common HLA-I allotypes across worldwide populations (N = 250), which show a continuum of values for HLA-A, HLA-B and HLA-C. This variation is functionally relevant, as tapasin-independent allotypes were shown to present a broader array of peptides than tapasin-dependent allotypes, and tapasin independence conferred protection in HIV-1 disease. Given the role of tapasin in shaping the HLA-I peptide repertoire and its impact on disease outcome, we hypothesized that natural variation in its expression level could also affect disease pathogenesis and susceptibility. We have studied a large longitudinal malaria cohort collected in Uganda to estimate immunogenetic effects of variation on malarial disease outcome. One of these studies identified two SNPs, rs111686073 (G or C) and rs59097151 (A or G), that regulate TAPBP mRNA expression levels by altering AP-2alpha transcription factor and microRNA-4486 binding sites, respectively. The variants conferring high TAPBP mRNA level, rs111686073G and rs59097151A, associated with lower Plasmodium falciparum parasite prevalence and lower incidence of clinical malaria specifically among individuals carrying tapasin-dependent HLA class I allotypes. Tapasin-independent allotypes associated with relative protection regardless of the genotype at these two SNPs. These data suggest that higher tapasin expression levels can enhance peptide loading of tapasin-dependent allotypes to a similar extent displayed by tapasin-independent allotypes, resulting in better control of P. falciparum. Associations between HLA Class I alleles and HIV disease have long been known. Established associations in Blacks include protective associations for B5702, B5703, B5801, and B8101, and susceptible associations for B1801, B1802, B4501, and B5802. The Janssen vaccine phase 2 clinical trial studied many aspects of the immune response to a mosaic antigen vaccine; ELISA and ELISpot response to HIV antigens were predictive of protection against HIV challenge in the parallel macaque vaccine trial. Trial data with vaccine response reported includes 306 subjects, of which 172 are African Americans or Black Africans, thus we focused our analyses on this group. We investigated whether the association of HLA with HIV control also associated with differential gag/env response after vaccination. Response measures included the strength of ELISA responses, which measured antibody responses to five variant Env proteins, and ELISpot count, which measured T-cell responses to 11 Env, Gag, and Pol protein pools. In addition, we considered intra-cellular staining (ICS) antigen response levels elicited by both CD4 and CD8 T cells for 10 Env, Gag, and Pol protein pools, as these assess T cell functionality distinctly from ELISpot, and ICS responses were previously linked to protection in previous vaccine trials. A principal components analysis (PCA) was performed to reduce the degrees of freedom for the multiple HIV antigens tested in the ELISA and ELISpot analysis; the first principal component (PC) captured the strength of overall response, while the second PC captured differential response to different antigens. There were no significant associations for HLA class I or II in the first PC for ELISA or ELISpot responses, but the second PC for ELISpot response approached statistical significance for some class I alleles after correction for the number of comparisons tested (FDR less than 0.1). The second PC of the ELISpot data was primarily driven by the difference between the T cell responses to Gag vs. Env antigens. Responses to Gag are known to confer HIV control in natural history studies, whereas those to Env are less protective because of the extreme mutation rate of the gene encoding Env. Thus, we hypothesized that the HLA associations identified in the second PC reflected a differential response to Gag vs. Env by those alleles that showed significance. We therefore defined a "Gag - Env" variable for ELISpot and ICS by subtracting the mean Env response from the mean Gag response and tested for HLA class I associations with this variable. Several significant associations were observed. The strongest association was that of HLA-B5703 with the CD8 ICS Gag - Env response (FDR of 5E-07). Of note, HLA-B shows the strongest effects for control of HIV. We also observed a significant correlation between the effects of HLA-B alleles on HIV viral load control (measured as mean HIV viral load over time) and the Gag - Env variable. Thus, HLA-B alleles that confer protection in untreated HIV-1 infected subjects may also confer protective CD8 T cell responses to HIV vaccines that contain HIV Gag inserts as antigens. These data suggest that vaccine efficacy may apply to a proportion of subjects receiving the vaccine. HLA allelic diversity is thought to be maintained through overdominance, a selective advantage determined by allelic heterozygosity. Heterozygous advantage may be the result of a broader variety of peptides presented by HLA, resulting in a better immune surveillance than for homozygous individuals whose universe of peptides is determined by one allele only. We have previously reported this detrimental effect on HIV progression. A limitation of this analysis emanates from the blunt distinction between genetically homozygous vs. heterozygous individuals. A genotype composed of two alleles that are very closely related may for example, bind as limited a peptide repertoire as do true genetic homozygotes. Calculation of genetic distances between the pair of alleles composing a given genotype is one means for estimating the degree of functional similarity between pairs of alleles, assuming that as the genetic distance between alleles increases, so should the distinctiveness of the peptide repertoire. However, genetic distance, analyzed as a continuous variable, showed no effect on AIDS progression in our seroconverter cohort. Detailed studies of peptides eluted from various common HLA class I allotypes have been performed and made publicly available. These data provide the means to more directly quantitate the impact of allelic heterozygosity on the breadth of the HLA peptide repertoire. We focused on aggregate properties of the peptides presented by a given allotype rather than minor differences across peptides. We retained only those amino acids with a positional entropy lower than 0.1 (i.e. high selectivity for one amino acid at a given position) in order to reduce noise contributed at positions with high entropy. The resulting simplified logograms, termed submotifs, allow compact representation of the different families of peptides that a given allotype can present. This approach condensed a universe of 111,898 nonamer peptides presented by 95 different HLA class I allotypes into 382 distinct submotifs. We compared submotifs of the allotype pairs encoded by a given genotype at each locus, and calculated the proportion of submotifs that are not shared between that pair of allotypes in order to quantify the functional divergence between allotypes. Using the novel submotif metric, we compared, for each locus, individuals with the most functionally divergent genotypes to those with the least functionally divergent genotypes. This categorization at each locus improved the association with AIDS progression far beyond that observed for genetic zygosity at both HLA-A (HR 2.63, p 3.5E-05) and HLA-B (HR 2.68, p 2.5E-05). We plan to apply the metric, or related metrics to other human diseases including other infectious diseases, cancer and outcome to cancer immunotherapy and autoimmune diseases.