Background:  During natural killer (NK) cell development, NK receptors, such as KIR molecules, interact with their ligand to produce self-tolerant cells. These exhibit functional competence in situations where receptor-ligand binding is disrupted, such as occurs in HIV infection when human leukocyte antigen (HLA) -A and -B expression is down-regulated by Nef. Receptor-ligand interactions mediating strong inhibitory signals may eventually favor a more potent NK cell activation and better microbial or tumor control when inhibition is abrogated. In HIV infection, co-expression of different allelic combinations of KIR3DL1 subtypes with HLA-B influence disease outcome such that strong KIR-HLA interactions correlate with slower time to AIDS. The strongest influence on disease outcome in KIR3DL1 homozygote individuals is co-expression of HLA-B*57 and a KIR3DL1 genotype (3DL1*h/*y) lacking low-expressing allotypes at this locus. We hypothesize that co-expression of HLA-B*57 and 3DL1*h/*y not only affects HIV disease outcome, but also contributes to resistance to infection in individuals who are exposed repeatedly to HIV but remain uninfected (EU).

Methods:  The study population included 39 EU and 186 recently infected subjects enrolled in the Montreal Primary Infection cohort. All were KIR3DL1 homozygous. Inclusion criteria for EU were at least 3 years of follow-up in a high-risk cohort and at least 5 documented exposures to HIV. KIR3DL1 subtyping was performed by sequencing KIR3DL1 exons 3, 4, 5, 7, and 9. Single nucleotide polymorphisms corresponding to the KIR3DL1 subtypes KIR3DL1*001, *002, *004, *005, *007, *015, *019, and *020 were then identified. Sequence-specific oligonucleotide typing and sequencing were used for HLA-genotyping. We compared the genetic distribution of KIR3DL1 subtypes and of HLA-B*57 genotypes in EU vs primary infection subjects using a Fisher’s exact test.

Results:  The frequency of HLA-B*57 allele in the EU and primary infection population was 0.06 (5 of 78 alleles) and 0.02 (8 of 372 alleles), respectively (OR = 3.116, p = 0.0565), whereas KIR3DL1 *h/*y distribution was similar in both populations (EU, 26 of 39, 66.7%; primary infection, 106 of 186, 57.0%; p = 0.29). However, the combined B*57-3DL1*h/*y genotype was more frequent in EU (5 of 39, 12.8%) than primary infection subjects (5 of 186, 2.7%) (OR = 5.171, p = 0.0174).

Conclusions:  The 3DL1*h/*y-B*57 combined genotype not only plays a role in HIV disease progression, but may be related to protection from infection.