The Story So Far
For this series we used the Small Molecule Antiviral Compound Collection (SMACC) library which is mined from the ChEMBL database. The SMACC library is a focused library that contains about 32,500 entries from the 2.4 million ChEMBL database, and the compounds have been specifically selected because they interact with viral targets. The SMACC library, and the conditions used to establish it, is described in a recent publication by Martin et al.
From the SMACC library, compounds active against Flaviviruses were selected, generating a focused library of approximately 2,000 compounds. This curated library was further diluted by the following criteria:
- Selectivity for DENV over other Flaviviruses
- Proper assay annotation in the original data
- The STOPLIGHT calculator
- Chemical clustering (selecting only one representative compound from each cluster)
- Additional manual inspection
- Commercial availability and cost
This generated a list of 73 compounds that were purchased and tested.
The selected compounds were screened in DENV2 and DENV4 phenotypic assays resulting in 10 compounds that displayed activity at 1 µL, and 50 compounds that showed activity at 10 µL. Further the library of 73 compounds was screened in a DENV2 Rdrp picogreen assay to determine %-inhibition. The 30 most promising compounds from these screens were selected and the IC50 was determined in the DENV Rdrp picogreen assay.
These 12 compounds showed the most promising DENV Rdrp IC50 and were evaluated further:
The table below summarizes the data collected in the initial assays. Notable compounds are RA-1904, RA-1936, and RA-1941 that fail to stop viral replication in the phenotypic whole cell assay (nLuc) but exhibit good inhibition of DENV2 Rdrp in the picogreen assay.
Kinetic solubility measurements identified five compounds with poor solubility as determined by the AViDD hit confirmation cascade (good kinetic solubility = 10×IC50): RA-1904, RA-1933, RA-1941, RA-1952, and RA-1954. Aggregation studies using [dynamic light scattering techniques] showed one compound with high % aggregation at 50 µM (RA-1933).
An orthogonal antiviral assay (using [Huh7] cells) was utilized to again determine the whole cell effect of the compounds. RA-1895, RA-1904, and RA1952 showed a very unfavorable IC50; RA-1936 and RA-1941 have not been evaluated in the assay.
Further studies on the Huh7 cells showed all the tested compounds displayed poor CC50 values, indicating possible toxicity issues. A gelbased assay (using similar conditions as described here showed that with the exception of RA-1895, RA-1904, and RA-1914, all compounds were able to block DENV Rdrp function. Unfortunately, they also showed effect on human mitochondrial RNA production, further adding to the picture that these compounds may not be suitable to explore due to toxicity.
The data from the additional assays are outlined below:
An expansion was done focusing on compounds around RA-1914 (10 analogues) and RA-1926 (5 analogues) along with 5 additional compounds from the SMACC library (three from the original screen above: RA-1925, RA-1926, RA-1947). Of these, 12 compounds showed an IC50 of >100 µM. Of the remaining 8 compounds one displayed assay interference, and the three from the previous tests had been excluded due to poor toxicity profiles.
This left four compounds that were submitted to the STOPLIGHT calculator and checked in an online PAINs remover program; these filters showed the remaining four compounds are likely false positives due to their reactive nature. The results are summarized below.
The 74 compounds submitted in the phenotypic assay for analysis were filtered through the STOPLIGHT calculator and checked in an online PAINs remover program:
- In the PAINs remover, 48 compounds passed the filter and 26 were identified as likely false positives.
- STOPLIGHT showed 68 compounds as yellow (may be unwise to progress) and only six compounds as green (likely no issues).
From these results all likely false positives were removed from the data. Three reference compounds had been included in the screen - EIDD-2749 (RA-1350) an Rdrp inhibitor with some reported activity on DENV, NITD-688 (RA-1551) a DENV pan-serotype inhibitor, and sinefugin (RA-1794) which showed no activity in the phenotypic assay.
Compounds which showed poor or no effect in the phenotypic (NanoGlo luciferase assay) of DENV2 and DENV4 were removed, this left 24 compounds showed good activity at 10 µM and 11 compounds with activity at 1 µM (n.b. low value = good activity; high value = low activity):
The top 11 compounds were investigated for further development:
