Making an Impact on Hit Triage
Researchers are using phenotypic assays at earlier and earlier stages of drug discovery. Assays for use in hit triage must be exceptional.
In a previous post we’ve described the impact of phenotypic assays throughout drug discovery (see Phenotypic Platforms are Taking Over Drug Discovery). Now let’s talk about assay features important for use in hit selection and triage.
After running a high throughput screen, whether in a target-based or phenotypic assay, we may have 100s of active compounds. We will need to winnow these down to the most promising hits. Doing this early in the process saves time and money.
In target-based programs, it is common to run a phenotypic assay as a secondary screen for all active hits. Target-based drug discovery screens often use highly artificial systems (e.g., binding, enzymatic or reporter-based assays). Hits from these assays may be sensitive to degradation by metabolic enzymes or unable to reach the target. Running an efficacy related cell-based assay identifies hits that are both target-specific and active in cells. These are more likely to be active in vivo.
Phenotypic assays can also be used to get rid of hits with undesirable mechanisms. We’ve previously described how phenotypic assays can detect nuisance mechanisms (see Nuisance Compounds I have Known). In early discovery, however, when compound numbers are high, there is hesitancy to run phenotypic assays due to pressure to keep spending per compound low.
To overcome this hesitancy, assays used in early discovery must be economical and also have strong support for their value. To rely on an assay for decision-making, confidence in that assay must be high.
There are several considerations for phenotypic assays used in hit triage. These assays should be high throughput, short and simple enough to run lots of compounds in iterative cycles. Assays will also require good reproducibility from experiment to experiment. This can be managed somewhat by comparing hits to a reference control compound rather than relying on absolute assay measurement values.
To minimize assay costs, miniaturization and employing a tiered approach for hit triage can be employed. Less expensive assays can be run first to reduce the number of compounds advancing to later tiers.
Based on our experience testing compounds from 100s of programs over many years, we have proposed three phenotypic assays for consideration in hit triage screening funnels (see this article The Future of Phenotypic Drug Discovery).
These assays meet the criteria of having strong support for their value. They filter hits based on important biology of the vascular and immune systems and their ability to capture specific toxicity-associated mechanisms. The first assay measures cytotoxicity to human primary endothelial cells under inflammatory conditions. The second assay measures inhibition of proliferation of human primary endothelial cells and the third assay measures inhibition of T cell activation.
Applying these assays in a tiered fashion eliminates hits that have the potential to be acutely toxic, result in organ toxicity or cause immunosuppression. These assays were selected out of 100s of human primary cells-based assays for their ability to detect certain nuisance mechanisms as well as biology that impacts accurate interpretation of downstream efficacy studies.
Incorporating phenotypic assays that can detect important toxicity-associated mechanisms will pay off in the long term. Such assays are applicable to any program. Not only will the use of these assays in early discovery facilitate getting to “better chemistry faster”, but data from these assays can be leveraged for developing structure-based in silico models. A worthy target indeed.
