The essential function of the Rev-RRE pathway makes it an attractive target for anti-HIV therapies. However, at this point, no therapies are clinically available, despite studies which have clearly shown that inhibitors of the pathway can be effective inhibitors of HIV replication (reviewed in reference 49
Previous screening assays for compounds that inhibit Rev function have, for the most part, involved in vitro assays that interfere with the high-affinity binding of Rev and the RRE (50
). However, while compounds discovered in this fashion often work well in vitro, many have shown considerable toxicity in cells or fail to specifically inhibit HIV replication.
In contrast, our previous efforts in this area used a cell-based screen that involved transient transfection with expression vectors producing Rev and Rev-dependent HIV GagPol (10
). p24 secretion into the medium was used as an end point to measure possible inhibition of Rev function. This screen identified a series of related compounds that inhibited HIV replication. However, the compounds did not inhibit Rev-RRE binding in vitro, and their mechanism of action was not studied further. Adapting this transient assay to more extensive screens proved cumbersome, and so a cell line that stably expresses HIV GagPol in a Rev-dependent context was created to allow higher throughput.
The present study utilized this stable cell line (5BD.1), which was made with the same plasmids used in the transient-transfection assay. Approximately 40,000 compounds were screened over a 4-week period. Eleven compounds were identified as primary hits, and in a secondary assay that more directly measured the inhibition of cellular Rev function, three emerged as Rev inhibitors. Thus, the data validate the use of the 5BD.1 cell line as a method for initially identifying inhibitors of Rev function. However, since the end point of the screening assay with the 5BD.1 cells was p24 release into the medium, inhibition at any of the other specific steps leading up to viral assembly and release would also score. It seems likely that some of these steps were the targets for some of the primary hits that did not score in the subsequent Rev assay. While we have not examined this issue in any detail, we believe that our results clearly support the utility of using the 5BD.1 cell line for further screening of compounds targeting many of the postintegration events in the HIV life cycle other than Rev function.
Since the compounds were identified using a broad-based cellular screening assay in which many specific steps (some of which are unknown) contributed to the readout of Rev function, further experimentation was needed to determine exactly how the compounds act. There is a dearth of knowledge about the specific steps in the Rev-RRE pathway, but it is generally accepted that the first step involves Rev-RRE recognition. Compounds 103833 and 104366 were tested in a Rev-RRE EMSA in vitro at concentrations up to 100 μM, which is at least 10-fold above their EC50
s, where they failed to show any inhibitory effect. While the conditions used in vitro may not directly reflect the in vivo situation, this result suggests that the compounds do not directly interfere with the primary Rev-RRE interaction. However, there are limitations to this interpretation, since the in vitro assay measures the binding of a bacterially made Rev protein to the viral RRE in the absence of all other components. Additionally, the bacterial protein is clearly lacking posttranslational modifications that might be present in the eukaryotic cells. The role that these factors might play in the in vivo binding step is largely not known, although it has been suggested that phosphorylation of Rev increases its RRE binding affinity (26
Since the compounds failed to inhibit Rev-RRE binding, some other step in the Rev pathway is likely to be inhibited, but our data do not allow us to distinguish if the compounds act directly on Rev or if they work by inhibiting a cellular factor that Rev requires. Inactivation of the function of such a factor could indirectly debilitate Rev, causing it to function less well in the presence of compound. Since our data do show that expression of protein from spliced mRNA or RNA whose export is driven by a CTE is largely unaffected by the compounds, it seems likely that, if a cellular factor is involved, its inhibition can be tolerated by the cell.
It is intriguing that both 103833 and 104366 share a great deal of structural similarity to bona fide kinase inhibitors, and there are some data to suggest that Rev function may be modulated by site-specific phosphorylation (26
). Compound 103833 (3-amino-5-ethyl-4,6-dimethylthieno[2,3-b
]pyridine-2-carboxamide) belongs to a class of compounds, thienopyridines, that have been described before as potent inhibitors of a variety of cellular kinases (44
). Compound 104366 [4-amino-6-methoxy-2-(trifluoromethyl)-3-quinolinecarbonitrile] shares some similarities with the anilino-3-quinolinecarbonitrile inhibitors of epidermal growth factor receptor and Src family kinases (5
). It will thus be interesting to examine the phosphorylation patterns on Rev in the presence or absence of compound. However, even without a clear definition of the mechanism of action of the present compounds, the present study has demonstrated that small-molecule targeting of Rev function can be a viable way to inhibit HIV replication.