Abstract
Latent HIV reservoirs persist in people living with HIV despite effective antiretroviral therapy and contribute to rebound viremia upon treatment interruption. Macrophages are an important reservoir cell type, but analysis of agents that modulate latency in macrophages is limited by lack of appropriate in vitro models. We therefore generated an experimental system to investigate this by purifying nonproductively infected human monocyte-derived macrophages (MDM) following in vitro infection with an M-tropic enhanced green fluorescent protein reporter HIV clone and quantified activation of HIV transcription using live-cell fluorescence microscopy. The proportion of HIV-infected MDM was quantified by qPCR detection of HIV DNA, and GFP expression was validated as a marker of productive HIV infection by colabeling of HIV Gag protein. HIV transcription spontaneously reactivated in latently infected MDM at a rate of 0.22% ± 0.04% cells per day (mean ± the standard error of the mean, n = 10 independent donors), producing infectious virions able to infect heterologous T cells in coculture experiments, and both T cells and TZM-bl cells in a cell-free infection system using MDM culture supernatants. Polarization to an M1 phenotype with gamma interferon plus tumor necrosis factor resulted in a 2.3-fold decrease in initial HIV infection of MDM (P < 0.001, n = 8) and a 1.4-fold decrease in spontaneous reactivation (P = 0.025, n = 6), whereas M2 polarization using interleukin-4 prior to infection led to a 1.6-fold decrease in HIV infectivity (P = 0.028, n = 8) but a 2.0-fold increase in the rate of HIV reactivation in latently infected MDM (P = 0.023, n = 6). The latency reversing agents bryostatin and vorinostat, but not panobinostat, significantly induced HIV reactivation in latently infected MDM (P = 0.031 and P = 0.038, respectively, n = 6). IMPORTANCE Agents which modulate latent HIV reservoirs in infected cells are of considerable interest to HIV cure strategies. The present study characterizes a robust, reproducible model enabling quantification of HIV reactivation in primary HIV-infected human MDM which is relatively insensitive to the monocyte donor source and hence suitable for evaluating latency modifiers in MDM. The rate of initial viral infection was greater than the rate of HIV reactivation, suggesting that different mechanisms regulate these processes. HIV reactivation was sensitive to macrophage polarization, suggesting that cellular and tissue environments influence HIV reactivation in different macrophage populations. Importantly, latently infected MDM showed different susceptibilities to certain latency-reversing agents known to be effective in T cells, indicating that dedicated strategies may be required to target latently infected macrophage populations in vivo.