It has been challenging to identify latency reversing agents (LRAs) broadly active across multiple primary cell latency models and patient cells ex vivo. A Jurkat CD4+ T cell model of HIV latency was used to screen the Merck library of 2.9 million compounds. Unlike prior efforts, this screen was conducted with and without a suboptimal concentration of SAHA, a histone deacetylase inhibitor (HDACi), with the goal of sensitizing the latency model to allow the discovery of novel and potentially synergistic LRAs. While 34% of compounds active in the screen acted via known mechanisms of action (HDACis or Farnesyl Transferase (FT) inhibitors) the remaining 66% acted via unknown mechanisms of action (uMOA). Based on the activity of uMOA compounds in different Jurkat T- cell models of HIV latency, 144 compounds were chosen for additional detailed analysis in primary CD4 T-cell models of HIV latency.
The 144 compounds were analyzed in seven different primary CD4 T-cell models of HIV latency using up to a six-point dose titration, in the presence of suboptimal concentration of panobinostat or vorinostat, tailored to the HDACi responsiveness of each primary cell system. Latency reversal was quantified by changes in reporter expression (eGFP or luciferase), HIV Nef or Tat gene expression, or HIV mRNA, depending model used.
As expected based on previous studies of LRAs in these primary cell models (Spina, PLoS Path 2013) among the 144 selected uMOA compounds, no one tested was active in all models. However, 2 compounds displayed statistically significant activity in six of the seven cell models and 16 compounds exhibited activity in three or more systems. Of note, none of the compounds were overtly toxic at concentrations that exhibited LRA activity. Additionally, no compounds induced CD69 expression in naïve T-cells isolated from healthy volunteers, suggesting their LRA activity was not due to generalized T cell activation.
This novel screen was able to identify LRAs that lack generalized T cell activating properties and appear non-toxic, but display activity across multiple cell line and primary cell models of latency. The compounds appear to have mechanism(s) of action distinct from known LRAs and work in concert with HDACis, suggesting the opportunity for combination LRA therapy. Validation of these compounds in patient cells is now underway. Screening for LRAs in the presence of a sensitizing agent may be a broadly useful approach.