Background: HIV-1 specific cellular immunity contributes to control of HIV-1 replication. We attempted to stimulate HIV-1-specific cellular immunity in chronically infected individuals on effective ART using a replication defective adenovirus type 5 (Ad5) HIV-1 gag vaccine in a randomized, blinded prospective therapeutic vaccination study.

Methods: Consenting HIV-1-infected persons on effective ART with >500 CD4 cells at entry were randomized 2:1 to receive an Ad5 HIV-1 gag vaccine (10¹⁰ viral particles/dose) or placebo over a 26-week period. After vaccination, participants were offered a 16-week analytical treatment interruption. Viral rebound kinetics were compared between the 2 groups. The co-primary endpoints were analytical treatment interruption log₁₀ HIV-1 RNA copies (mean at week 12 and 16, set-point), and time averaged area under the curve (TA-AUC). Cytotoxic T lymphocyte (CTL) responses were measured using intracellular cytokine staining. Proliferative responses to p24 gag and whole AT-2 inactivated virus (MN) were assessed by carboxy-fluorescein diacetate, succinimidyl ester (CFSE) dye dilution.

Results:  Of 114 randomized, 110 (73 vaccine, 37 placebo) participated in the analytical treatment interruption. Baseline CD4 cells were higher in the vaccine than in the placebo arm (medians 853 vs 716 cells/mm³). Week 12 or 16 viral end-points were obtained on 107 participants. Vaccine benefit trends were seen for both primary endpoints, but they did not reach the protocol-specified level of significance, which required one-sided p≤ 0.0125. For vaccine benefit in TA-AUC and set-point, p = 0.024 and p = 0.059, respectively. The TA-AUC and set point were estimated 0.26 and 0.27 log₁₀ copies, respectively, lower in the vaccine arm than in the placeob arm. There is 97.5% confidence that in the vaccinated arm both the TA-AUC and set point were no more than 0.04 and 0.11 log₁₀ copies, respectively, greater than in the placebo arm.A secondary endpoint of analytical treatment interruption week 16 RNA was shifted an estimated 0.52 log₁₀ copies lower in the vaccine versus placebo arm (p = 0.020). Higher baseline CD4⁺ and CD8⁺ proliferative response to HIV MN correlated with lower analytical treatment interruption set-point viral load (Spearman's rank: –0.37, p <0.001 and –0.20, p = 0.05 respectively). Analyses of additional virus specific immune responses are in progress.

Conclusions: The Ad5 HIV-1 gag vaccine was generally safe and well tolerated. Although there was a trend in favor of viral suppression among vaccine recipients during an analytical treatment interruption, the differences in HIV-1 RNA levels did not meet protocol-specified levels of significance.