A new model for post-integration latency in macroglial cells to study HIV-1 reservoirs of the brain.

Abstract

Macroglial cells like astrocytes are key targets for the formation of HIV-1 reservoirs in the brain. The 'shock-and-kill' HIV-1 cure strategy proposes eradication of reservoirs by clinical treatment with latency reversing agents (LRAs). However, virus activation may endanger the brain, due to limited cell turnover, viral neurotoxicity and poor penetration of antiretroviral drugs. Since the brain is not accessible to clinical sampling, we established an experimental model to investigate the LRA effects on HIV-1 latency in macroglial reservoirs. Human neural stem cells (HNSC.100) were used to generate a system that models HIV-1 transcriptional latency in proliferating progenitor, as well as differentiated macroglial cell populations and latency-modulating effects of LRAs and compounds targeting HIV-1 transcription were analysed. HNSCs were infected with pseudotyped Env-defective HIV-1 viruses. HIV-1 DNA and RNA levels were quantified by qPCR. Expression of latent GFP-reporter viruses was analysed by confocal microscopy and flow cytometry. NF-κB signalling was investigated by confocal microscopy and chromatin immunoprecipitation. Two of the eight well known LRAs (tumour necrosis factor-alpha, suberoylanilide hydroxamic acid) reactivated HIV-1 in latently infected HNSCs. Tumour necrosis factor-alpha reactivated HIV-1 in progenitor and differentiated populations, whereas suberoylanilide hydroxamic acid was more potent in progenitors. Pre-treatment with inhibitors of key HIV-1 transcription factors (NF-κB, Cdk9) suppressed HIV-1 reactivation. We conclude that latent HIV-1 in macroglial reservoirs can be activated by selected LRAs. Identification of small molecules that suppress HIV-1 reactivation supports functional cure strategies. We propose using the HNSC model to develop novel strategies to enforce provirus quiescence in the brain.