The components that have made HTS a smooth-running machine combined with the high-quality data that has been shown to support SAR have led to natural expansions of the HTS arena. Originally, an HTS operation simply identified active compounds in a mixture that was tested at a single concentration. The mixture was deconvoluted into its various components, and each component was tested for activity. This was considered a confirmed active. Most companies now test single compounds, so a second testing of the original solution is considered to be the confirmation. As the collections grew dramatically in size from 100,000 to 200,000 compounds to more than a million compounds, the downstream processes, including HTL, were flooded by thousands of active hits requiring further characterization. All HTS data sets have both false positives and false negatives. The first test to eliminate false positives in the hit set is to test the compounds for dose responsiveness. This testing step was reassigned from the therapeutic area to HTS when the hit sets went from handfuls of hits to thousands of hits being available at one time. The test for dose responsiveness progressed to the determination of the concentration that inhibits 50% of the response (IC50) values as the quality of screening assay data rose. This usually meant going from five-point concentration curves to, typically, ten points. Concentration curves can be made by diluting a stock concentration out from well to well across a 96-well plate with an eight-tip pipettor. A fast and efficient way to prepare dose-response samples for an HTS operation is to leverage the 96- or 384-tip pipettors and to make dose-response plates by diluting from plate to plate instead of from well to well within each plate (Figure 5.2). This method reduces the exposure of the DMSO solutions to the wet-air environment, and control wells on each plate are used to normalize for any DMSO concentration differences and plate-to-plate variability. Different groups will put controls either on both end columns or the first and last columns or the middle and last columns as they create 384-well plates from four of the 96-well plates that usually have only one column of controls. This plate-to-plate concept for compound dilutions has been fully validated against the well-to-well dilution scheme at BIPI. Today, the initial stock solutions are diluted in neat DMSO and plated out as the appropriate 2 ^L aliquots to be diluted into aqueous buffer immediately before testing. This procedure helps to keep the compounds in solution until they are diluted into the aqueous buffer and ensures that even if the highest concentrations come out of solution, the lower concentrations will have the best chance of being correct. Finally, compounds rarely exhibit sticking to pipettor tips and carryover when dilutions are done in DMSO, whereas
oooooooooooo oooooooooooo oooooooooooo oooooooooooo oooooooooooo oooooooooooo oooooooooooo oooooooooooo x y figure 5.2
Leveraging 96- or 384-tip array pipettors for efficient compound serial dilution. (A) typical intra-plate (x-y), serial dilutions accomplished using an 8- or 12-tip pipettor; (B) more efficient interplate (z), serial dilutions using an array pipettor.
fYYYYYYYTOmi oononoonooni oooonnonooni rrrrx-irrrinrxlL^ prx-rrinrnmoll onononnnonni ro^nonononni onononononn
this can be a serious problem for certain compounds diluted in assay buffer and can cause aberrant and misleading IC50 curves (Figure 5.3).
The project teams or therapeutic area teams using the IC50 values generated by the HTS assays often requested the HTS group to test all hits in a selectivity assay or some other assay that would qualify the hit set. Ideally, this counterscreen would also eliminate false positives caused by an assay artifact that, in some cases, can produce an authentic-looking dose-response curve. It was easier for HTS to test all of the hits at the same time rather than just the dose-responsive hits — the procedure usually followed by therapeutic-area laboratories. The testing of large hit sets in one assay on the same day under identical conditions will produce less error in the overall data because there are no additional errors such as those that result from day-to-day variation. In areas where there are closely related assays, the HTS group might perform several counterscreens to reduce the hit set, which may number to thousands of compounds, low hundreds, or even lower. After this qualification of the entire hit set, the HTL or therapeutic-area chemists evaluate all of the data to determine which compounds are of further interest. This step might include clustering of the hits (groups of like compounds based on their chemical structures and properties) to eliminate compounds that fail certain physical-property rules. This is also the first time the compounds are evaluated for SAR patterns. There could be hundreds of interesting compounds that have not been tested as part of the HTS library, and the chemist might want to confirm SAR patterns and check for potential false negatives. This next round of testing is performed by HTS to get all of these compounds tested in full dose-response curves under identical conditions again in both the original screen and any counterscreens. At BIPI, the HTS group now supports the primary assay and its counterscreens through the HTL process when the hit sets are sufficiently large. Today, automated chemistry is often used to expand hit sets if the hits are not
Was this article helpful?