Astrocytes sustain long-term productive HIV-1 infection without establishment of reactivable viral latency

The "shock and kill" HIV-1 cure strategy proposes eradication of stable cellular reservoirs by clinical treatment with latency-reversing agents (LRAs). Although resting CD4(+) T cells latently infected with HIV-1 constitute the main reservoir that is targeted by these approaches, their consequences on other reservoirs such as the central nervous system are still unknown and should be taken into consideration. We performed experiments aimed at defining the possible role of astrocytes in HIV-1 persistence in the brain and the effect of LRA treatments on this viral sanctuary. We first demonstrate that the diminished HIV-1 production in a proliferating astrocyte culture is due to a reduced proliferative capacity of virus-infected cells compared with uninfected astrocytes. In contrast, infection of non-proliferating astrocytes led to a robust HIV-1 infection that was sustained for over 60 days. To identify astrocytes latently infected with HIV-1, we designed a new dual-color reporter virus called NL4.3 eGFP-IRES-Crimson that is fully infectious and encodes for all viral proteins. Although we detected a small fraction of astrocytes carrying silent HIV-1 proviruses, we did not observe any reactivation using various LRAs and even strong inducers such as tumor necrosis factor, thus suggesting that these proviruses were either not transcriptionally competent or in a state of deep latency. Our findings imply that astrocytes might not constitute a latent reservoir per se but that relentless virus production by this brain cell population could contribute to the neurological disorders seen in HIV-1-infected persons subjected to combination antiretroviral therapy.