HVTN 705 data repo: Unbiased cell clustering analysis of vaccine-induced T cell responses in the Imbokodo HIV-1 vaccine trial

<b>Background</b>HIV-1 T cell responses are associated with viral control and may be protective in a prophylactic vaccination setting. Traditional methods for analysing these responses might be biased towards specific functionalities or epitopes. This study presents an unsupervised and unbiased clustering analysis workflow, using the Leiden algorithm followed by selection of antigen-specific clusters using MIMOSA positivity calls, for high-dimensional flow cytometry data to identify distinct T cell populations associated with protection in the HVTN 705/HPX2008/Imbokodo HIV-1 vaccine efficacy trial.<b>Methods</b>Participants were vaccinated with Ad26.Mos4.HIV (M0, M3) and Ad26.Mos4.HIV + Clade C gp140 (M6, M12), and a validated 28-color intracellular cytokine staining assay was performed at M7, M13 and M24 in a pilot immunogenicity (n=60) and case-control cohort (n=283). 28-color phenotyping assays were also performed in PBMC from the case-control cohort (n=334).<b>Findings</b>Our clustering analysis workflow allowed identification of vaccine-induced subpopulations of both CD4+ and CD8+ T cells expressing combinations of the 28 markers. Durable HIV-1 antigen-specific CD4+ and CD8+ T cell responses were observed for up to 2 years, comprising mainly clusters of polyfunctional T cells expressing anti-viral cytokines and activation markers. Eight CD4+ and six CD8+ HIV-1 antigen-specific T cell clusters were induced by vaccination; only one CD4+ T cell cluster specific for Gag was mildly elevated in cases (acquiring HIV) compared to controls.<b>Interpretation</b>Our study introduces an innovative analysis approach to identify vaccine-induced T cell subpopulations in vaccine trials. Comparison of T cell clusters between cases and controls holds promise for improving the efficacy of future HIV-1 vaccination strategies.