The mucosal barrier of the female genital tract is the first site for most sexual transmission of HIV and identification of mucosal factors predictive of HIV susceptibility would inform prevention strategies. Here we used unbiased mass spectrometry-based proteomics to characterize vaginal proteins that associate with HIV acquisition in women enrolled in the VOICE (MTN-003) trial.
VOICE was a Phase IIb randomized, placebo-controlled trial to assess HIV pre-exposure prophylaxis with daily oral tenofovir disoproxil fumarate (TDF), oral tenofovir-emtricitabine (TDF-FTC), or 1% tenofovir (TFV) vaginal gel conducted in South Africa, Uganda, and Zimbabwe. Proteomic analysis was performed on vaginal swabs collected at the last HIV seronegative timepoint of women who seroconverted (cases, n=121) and matched swabs from women who remained seronegative (controls, n=381). Differential protein expression was performed using t tests adjusting for multiple comparisons (Benjamini-Hochberg (BH)). IPA software and David Bioinformatics Resource were used for gene ontology and pathway analysis.
In this analysis 1,051 human proteins were identified across 502 participants (TDF n=106, TDF-FTC n=119, TFV n=18, oral placebo n=112, vaginal placebo n=147). Differential protein analysis did not identify any significant differences by study arm (TDF vs Placebo, 6.8%; TDF-FTC vs Placebo, 4.5%; TFV vs Placebo 2.1%, P<0.05). When comparing cases to controls, 284 (27%) proteins were differentially abundant (P<0.05), with 92 passing BH adjustment (32%). These proteins were from pathways related to epithelium development (p=6.8E-3) and innate immune response (p=7.5E-3). PLSDA using 3 proteins selected from LASSO distinguished cases from controls with 68% accuracy (95% CI 64-72%, p=7.9E-10). Cluster analysis grouped women based on expression of these biomarkers (p=5.9E-6) and associated with a 3-fold lower HIV infection risk (OR=3.0, p=5.9E-6). While some biomarkers associated with age or injectable contraceptive use, adjustment for these variables did not impact the relationship with HIV susceptibility.
This study identified a high-risk phenotype of reduced mucosal innate immunity and epithelial barrier integrity associated with increased risk of HIV acquisition. Understanding the underlying impact of these biomarkers to mucosal immunity is critical to improve HIV prevention options. Acknowledgements We thank the VOICE study participants and research team for their support for this study.