Viruses that persist despite seemingly effective antiretroviral treatment (ART) and can reinitiate infection if treatment is stopped preclude definitive treatment of HIV-1 infected individuals, requiring lifelong ART. Among strategies proposed for targeting these viral reservoirs, the premise of the "shock and kill" strategy is to induce expression of latent proviruses [for example with histone deacetylase inhibitors (HDACis)] resulting in elimination of the affected cells through viral cytolysis or immune clearance mechanisms. Yet, ex vivo studies reported that HDACis have variable efficacy for reactivating latent proviruses, and hinder immune functions. We developed a nonhuman primate model of post-treatment control of SIV through early and prolonged administration of ART and performed in vivo reactivation experiments in controller RMs, evaluating the ability of the HDACi romidepsin (RMD) to reactivate SIV and the impact of RMD treatment on SIV-specific T cell responses. Ten RMs were IV-infected with a SIVsmmFTq transmitted-founder infectious molecular clone. Four RMs received conventional ART for >9 months, starting from 65 days post-infection. SIVsmmFTq plasma viremia was robustly controlled to <10 SIV RNA copies/mL with ART, without viral blips. At ART cessation, initial rebound viremia to ~106 copies/mL was followed by a decline to < 10 copies/mL, suggesting effective immune control. Three post-treatment controller RMs received three doses of RMD every 35-50 days, followed by in vivo experimental depletion of CD8+ cells using monoclonal antibody M-T807R1. RMD was well-tolerated and resulted in a rapid and massive surge in T cell activation, as well as significant virus rebounds (~104 copies/ml) peaking at 5-12 days post-treatment. CD8+ cell depletion resulted in a more robust viral rebound (107 copies/ml) that was controlled upon CD8+ T cell recovery. Our results show that RMD can reactivate SIV in vivo in the setting of post-ART viral control. Comparison of the patterns of virus rebound after RMD administration and CD8+ cell depletion suggested that RMD impact on T cells is only transient and does not irreversibly alter the ability of SIV-specific T cells to control the reactivated virus.