Background:  In ACTG A5197, a randomized, placebo-controlled study of a therapeutic rAd5 HIV-1 gag vaccine, vaccine recipients had marginally greater viral suppression during an analytic treatment interruption (ATI). We evaluated the durability and correlates of virologic control and characteristics of HIV sequence evolution among persons with initial viral suppression during the ATI.

Methods:  Plasma viral load (pVL) set point, a co-primary study endpoint, was defined as the mean pVL at ATI weeks 12 and 16. Participants with a set point of <3.0 log₁₀ copies/mL were categorized as initial virologic suppressors. HIV-1 gag and pol were sequenced from plasma virus obtained during the ATI at initial detectable viremia, ATI week 16, and ATI week 49. Immune responses were measured by flow cytometric analysis of lymphocyte phenotype and intracellular cytokine expression assays. Characteristics of participants with and without initial viral suppression were compared using Wilcoxon rank sum or Fisher’s exact tests.

Results:  Eleven of 104 participants (10.6%) were classified as initial virologic suppressors, 9 of whom had received the vaccine. Initial virologic suppressors had significantly less CD4⁺ cell decline by ATI week 16 as compared to non-suppressors (median 7 CD4⁺ cell gain vs 247 CD4⁺ cell loss, p = 0.04) and the ATI set point was significantly lower than pre-ARV therapy pVL among initial suppressors compared to non-suppressors (1.2 vs 0.6 log₁₀ copies/mL difference, p <0.01). Of the 10 initial virologic suppressors with a pVL at ATI week 49, only 3 maintained pVL <3.0 log₁₀ copies/mL. HIV-1 gag-specific CD4⁺ interferon-γ responses measured prior to the ATI were not associated with initial virologic suppression and no evidence of vaccine-driven HIV sequence evolution was detected. Compared to non-suppressors, initial virologic suppressors had a lower percentage of CD4⁺ CTLA-4⁺ cells prior to treatment interruption (median 8.9% vs 14.1%, p = 0.02), but a greater proportion of HIV-1 gag-reactive CD4⁺ TNF-α⁺ cells expressing either CTLA-4 (54% vs 33%, p = 0.01) or PD-1 (36% vs 14%, p = 0.04).

Conclusions:  In a rAd5 therapeutic HIV-1 gag vaccine trial, initial viral suppression was observed in a minority of patients, but this response was not sustained with longer follow up. The association between CTLA-4 and PD-1 expression on CD4⁺ T cells and virologic outcome warrants further study in trials of other therapeutic vaccines in development.