Unique footprints of HLA class I on clade C HIV-1 in a cohort of Ethiopian patients

Background: Despite the enormous genetic diversity of HIV, this virus follows predictable pathways of immune escape in response to immune selection pressure. Because there are substantial differences in HLA allele and haplotype frequency across different populations worldwide, differences in viral diversity at a population level may be seen. Here we characterized the HLA genotypes and associated HIV imprinting in a cohort of Ethiopian patients from Ethiopia and Israel. Methods: HIV infected, ARV-naive patients (n = 356) were recruited from clinics around Ethiopia and in Israel (Ethiopian Jewish immigrants). Sequence-based high resolution HLA typing was performed, as well as full-length viral sequencing using both Sanger and 454 deep pyrosequencing techniques. HLA-HIV sequence associations have been corrected for viral phylogenetic relatedness. To correct for multiple comparisons only associations with false discovery rate equal or lesser than 0.2 were considered. Results: There were 36 different HLA-A, 49 HLA-B, and 24 HLA-C alleles detected in this cohort. No major differences in allele frequencies were found between patients recruited in Ethiopia and Ethiopian Jews living in Israel. However, significant differences in HLA class I allele frequencies were found between these patients and the published frequencies in a cohort of clade C infected patients from South Africa. In this regard, alleles uniquely described in the past among Africans were much more frequent among South Africans compared with Ethiopians. Unusual high rates of HLA-B57 and HLA-B13 (population frequency of 18.2% and 12.3%, respectively), which are associated with disease control, were found among this Ethiopian cohort. Associations analysis revealed the highest associations density within HIV proteins rev and tat in contrast with the structural proteins as shown in other associations studies. The analysis reproduced many previously described HLA-associated HIV polymorphisms both in African and Caucasian cohorts, but new targets for selective pressure were found. Conclusions: This is the first work to characterize comprehensively and at a large scale an African cohort from Ethiopia, and the first to apply deep pyrosequencing to HLA-HIV association analysis. We found the unique HLA background of the Ethiopian cohort is associated with unique viral imprinting.