Discovery of Potent Non-nucleoside Inhibitors of Dengue Viral RNA-Dependent RNA Polymerase from Fragment Screening and Structure-Guided Design

Abstract

Flavivirus NS5 RNA-dependent RNA polymerase (RdRp) is an important drug target. Whilst a number of allosteric inhibitors have been described for Hepatitis C virus RdRp, few have been described for DENV RdRp. In addition, compound screening campaigns have not yielded suitable leads for this enzyme. Using fragment-based screening via X-ray crystallography, we identified a biphenyl acetic acid fragment that binds to a novel pocket of the dengue virus (DENV) RdRp, in the thumb/palm interface, close to its active site (termed “N pocket”). Structure-guided optimization yielded nanomolar inhibitors of the RdRp de novo initiation activity, with low micromolar EC₅₀ in DENV cell-based assays. Compound-resistant DENV replicons exhibited amino acid mutations that mapped to the N pocket. This is the first report of a class of pan-serotype and cell-active DENV RdRp inhibitors and provides a significant opportunity for rational design of novel therapeutics against this proven antiviral target.

Discussion of Chapter 14 in Dengue and Zika: Control and Antiviral Treatment Strategies

This discussion was held at the 2nd Advanced Study Week on Emerging Viral Diseases at Praia do Tofo, Mozambique.Transcribed by Hilgenfeld R and Vasudevan SG (Eds); approved by Dr. Siew Pheng Lim.

• Paul Young: Great work! A perfect example of a great collaboration between biology and chemistry. The work you describe seems to have hit a little bit the end of the road. You have identified and derived reasonably active sompounds but they do not appear to be effectively bioavailable. So what’s the next step?

• Siew Pheng Lim: So this is only one part of the story. Our goal is to design compounds that extend out of the pocket. In the parlance of chemists in the team these extension from the pocket are done to find the so-called sweet spot to further improve the potency. But despite a lot of effort, we did not find another sweet spot. I also talked a little about some similarity the N-pocket shares with HCV palm 1 site. However HCV palm 1 pocket has a another wall contributed by the C-terminal part of the protein which allows those inhibitors to bindto both sides. In the case of Dengue virus polymerase we appear to have only one wall, we are trying to find a way to reach a second wall and that has not been successful so far.

• David Jans: I just wanted to know what EC₅₀ you have been happy with. Several of the compounds have EC₅₀ of around 3 micromolar. Do you think that would be potent enough? In other words what should be the target for an effective drug?

• Siew Pheng Lim: If I compare against the phenotypic screens that my colleagues do, they typically find inhibitors that are double-digit nanomolar in EC₅₀. There they have a different challenge because they have issues with solubility and clearance as the compounds are too greasy. So I think for us if we can hit below one micromolar, We set ourselves a benchmark of less than one micromolar which we consider a good threshold. So you can see we have not quite reach that point yet.

• Subhash Vasudevan: In the case of HCV palm 1, the compound that targeted a similar pocket actually got to the clinic implying that it is possible to develop a non-nucleosidic inhibitor. The pocket that you are looking is also truly a hotspot for Dengue and other flaviviruses. So the question I guess is do we give up on that area based on your experience so far?

• Siew Pheng Lim: For now, this work has been put on hold, and we published the work so that other researchers in the field can expand on the effort and take a fresh look at approaches to target this important pocket. The search may include using other flaviviruses and find something we have not yet found to extend the chemistry. As David Jans alluded yesterday, we primarily work with DENV3 NS5 RdRp for which we have structural data. So it would be nice to have structural studies on other serotypes and other flaviviruses to see if something maybe different.

• Subhash Vasudevan: But your data clearly shows that Dengue 2 and Dengue 4 give broadly similar results. In the discovery process if you are trying to do DENV1 to 4 separately then that would be a huge challenge. But NS5 is a really important target and since the non-nucleoside approach is not quite ready yet for pan-serotype inhibitor, what about the nucleoside aproach?

• Siew Pheng Lim: Nucleoside approach is still ongoing. We started with KAB-344, the adenine analog (also known as NITD008) that eventually failed in the two-week animal toxicity studies and could not be progressed further. Since then we are working on other scaffolds. I think the issue with nucleosides is the unpredictable toxicity. To mitigate this we have implemented additional cytotoxicity assays trying to weed out toxicity early in the flowchart, such as mitochondrial assays –longer cytotoxicity assays – to capture those compounds that are potentially cytotoxic before we go into two-week toxicity studies in rats and dogs. Through this process, we have a few candidates that we are evaluating. The challenge with nucleoside inhibitors is the potentially lengthy scale up synthesis which can be expensive – so the way forward remains to be seen.