Turning the screening scenario on its head
The pharmaceutical industry is under immense pressure to find better ways to bring new drugs to market, given the dwindling pipeline and the astronomical cost of late-stage failures. So, is the screening pool from which we select compounds just too small, or not chemically diverse enough?
Screening library subsets is designed to capture all the areas of the chemical space within a company’s full compound library, but by screening a much smaller number of compounds to allow the use of more complex phenotypic assays that provide data closer to the in vivo goal.
“However,” says TTP Labtech’s Dr Paul Wylie, “I believe that this approach is flawed as the subsets are selected on target-based information. Using data from historic target-based screens to select subsets for phenotypic screens runs the risk of missing good quality hits – or selecting poor compounds for next stage interrogation. Just because a compound is in a similar chemical space to another molecule, doesn’t necessarily mean that it will exert the same phenotypic response, in vivo.”
A good example is the catecholamines: dopamine, noradrenalin and adrenaline. Whilst they only differ by the addition of a single functional group, they exert very different physiological responses (see table 3).
“In my view,” says Dr Wylie, “it is time to turn the screening scenario on its head. Screening is about excluding the compounds that are not useful; so, instead of screening in a target-based way first, then using a phenotypic response to backup the data, we should screen phenotypically and follow up with target-based identification.”
What would be the requirements to successfully implement phenotypic screening at primary throughputs? “Target-based screening was adopted partly because it was easy to automate, fast, reproducible and relatively inexpensive to run on very large compound sets”, he explains. So, just how far off running a full deck phenotypic screen are we?
We would need:
- Automation to run assays in 1536-well plates, to achieve the required throughput. Many liquid handling systems are now able to do this very effectively.
- Manageable cell costs/cell numbers. Primary cells and iPS cell lines are more physiologically relevant but incur higher costs. HTS requires the use of around 100 cells per 1536-well plate to make these cost-effective.
- Consistent, fast scan and analysis times to fit into a fully automated HTS process. This is important because many HCA instruments have variable read times.
- Manageable data files from 2-3 million wells.
- Straightforward biology.
Dr Wylie believes that throughput and assay complexity are perceived to be the main barriers to running a full deck phenotypic HTS, but, he asks, “Is that really true?”
- The biology itself can be limiting, but only in certain projects. Phenotypic screening is robust. “Many phenotypic assays look at quite basic readouts: cell viability (live dead/apoptosis), reporter gene expression, cell cycle, angiogenesis formation. These are simple assays to screen,” he points out.
- For throughput: “it is true that many typical imaging systems do not fulfill the criteria for a HTS manager to run in a full-deck screen. So it is best not to use a standard microscope-based imaging system”, he concludes.
Laser-scanning imaging cytometers (LSICs), such as acumen Cellista (TTP Labtech), combine object recognition with bulk read speeds (typically under 5 minutes per 1536-well plate). It requires fewer cells (typically 100/well). It also produces very small file sizes, so the IT requirements are the same as for target-based screening. These features mean that LSICs like acumen Cellista are ideally placed to remove the compromises and offer practical high throughput, full deck phenotypic screening.
“The costs of phenotypic primary screens are higher than traditional target-based screening; but it is surely much cheaper than pursuing poor compounds which fail at later stage drug development,” explains Dr Wylie.
“I believe the only barrier to running full deck phenotypic screening today is the willingness and belief to accept it as a key area of drug screening and then to implement it.”