Identification of novel cellular factors involved in HIV-1 latency

Despite ongoing efforts, HIV-1 (the causative agent of AIDS) remains an unresolved health threat. Even though current therapy approaches efficiently block ongoing viral replication they cannot cure the infection due to the presence of a latent reservoir. It is of crucial importance to understand that this viral reservoir can fuel new rounds of viral replication and spread the infection. The viral reservoir is defined as cells (best-characterized are resting memory CD4+ T cells) harboring a replication-competent provirus while not producing new progeny virus, a state that can be reversed. Strategies to eradicate the viral reservoir include the so-called ‘shock and kill’ approach, which is a two-step process, aiming in the first step to reactivate the latent reservoir, leading to the production of new viruses. In a second step, denoted as ‚kill’, death of those virus-producing cells is induced by specific cells of the immune system. The identification of host factors involved in HIV-1 latency formation and maintenance is therefore a crucial step to support this and also other strategies. This current study aimed at validating novel host factors involved in HIV-1 latency. For this, the results of a genome-wide siRNA screen performed in HEK293T cells infected with a single- cycle HIV-1 reporter virus expressing luciferase (HIVluc), served as starting point. This model elucidates potential host factors important for HIV-1 transcriptional processes and is per se not a model for HIV-1 latency, as luciferase is constantly expressed from the viral promoter. Nevertheless, we hypothesize that there is a certain overlap in host factors, which not only have an influence on HIV-1 transcription but also on latency formation/maintenance. This assumption was confirmed by publications on two host factors, namely BRD4 and CYLD, which were reported to be important for the maintenance of HIV-1 latency but were also identified in the primary screen to influence HIV-1 luciferase expression. Two differential approaches were followed to choose and validate initial screening hits: For the first approach, follow-up hits were chosen based on their capability to bind and/or modify chromatin. This selection criterion is based on the importance of the chromatin environment for the formation and maintenance of HIV-1 latency. In a first set of experiments, the initial screening setup was recapitulated for a selection of hits and TRRAP, an adaptor protein found in multiprotein chromatin complexes and involved in epigenetic regulation of transcription, was reconfirmed as hit. Further experiments aimed to identify a potential role of TRRAP in HIV-1 latency. For this, J-Lat cells, a Jurkat derived latency cell model, were depleted of TRRAP by shRNA or siRNA and stimulated with various latency reversing agents (LRAs) to test for an additive effect on latency. Despite intensive testing, a clear involvement of TRRAP on latency could not be definitely determined. In the second approach, follow-up hits were chosen based on the availability of commercially available compounds, which target the respective primary hits. This approach is here referred to as druggability. Several compounds were chosen and tested for their potential action on latency reversal alone or in combination with LRAs. Auranofin was identified to enhance reversal of latency when applied in combination with all tested LRAs, i.e. TNFα, prostratin, SAHA or sodium butyrate in J-Lat cells and in combination with SAHA and sodium butyrate when tested in U1 cells. Following the identification of auranofin, compounds affecting the same proposed targets, i.e. PRDX5, and its upstream pathway member, i.e. TXNRD1, involved in the thioredoxin pathway, were tested. Whereas none of the additional compounds recapitulated the exact effect of auranofin, diminazene was identified to also have an effect on latency reversal, which was most effective when applied in combination with either TNFα or prostratin. This effect was much stronger than the one seen for auranofin, but was cell line specific as the effect was only seen in J-Lat but not in U1 cells. Verification of auranofin’s (and diminazene’s) proposed targets was tested by siRNA-mediated silencing, demonstrating a marginal increase in latency reversal, pointing towards the involvement of the associated system in latency reversal. Based on literature, auranofin targets the thioredoxin system involved in reactive oxygen species (ROS) detoxification and reduction of oxidized proteins. This process is crucial for the integrity of the cell. Diminazene, on the other hand, is likely to target the polyamine homeostasis, which participates in a plethora of processes and interactions. An imbalance in polyamines might also lead to a higher level of oxidized products, potentially linking it to the cellular redox state. Whether those pathways are interconnected needs to be determined in future studies. Nevertheless, both the redox system and the polyamine homeostasis are of crucial importance for the cell with a wide range of possible consequences upon their perturbation, which are likely to affect the proviral state. In summary, in an unbiased systems-level approach, we identified two compounds that mediate latency reversal in combination with other LRAs. We propose further investigation of the two systems targeted by those compounds, which have not been (extensively) studied for an involvement in HIV-1 latency as they could support the targeting of the latent reservoir.