OA04.05 LB
Background: The RV144 trial showed promise towards the development of an effective HIV vaccine. Antibodies to HIV Env V1V2 loop, Env-specific IgA and HLA class II DQB1*06 allele with high Env IgA were all significantly associated with risk of infection.
Methods: We used an integrated systems biology approach to identify novel correlates of immunogenicity and/or protection. Env-specific stimulated peripheral blood mononuclear cells from RV144 participants were used to test the hypothesis that innate pro-inflammatory responses would demarcate RV144 case-control groups.
Results: A case/control gene expression analysis using 133 controls and 27 cases yielded few significantly differentially expressed genes between cases and controls. To adjust for the heterogeneity and the case/control class imbalance, we employed a stratification strategy using a tree-based classification method with IgA, V1V2 and DQB1*06 as predictor variables and the infection status as the outcome. The obtained tree model showed a balanced accuracy of 61% and identified two different risk groups. Cases in the first group had low IgA titers and did not express the DQB1*06 allele. Transcriptional profiling showed downregulation of the type II interferon induced genes (Fisher enrichment test p value = 0.04) in these cases compared to controls. Cases in the second Group had high IgA and Low V1V2 titers. Transcriptional profiling showed, interestingly, up-regulation of several type I interferon genes including antiviral genes (Fisher enrichment test p value = 10−6) in these cases compared to controls.
Conclusions: Herein we have identified a gene expression signature, segregating case and control donors and which, counterintuitively, shows contrasting roles for innate immune response genes in conferring vaccine induced protection. These results have important implications for the development of more effective HIV vaccine strategies.
