While an expanded substrate of diverse compounds is essential, it does not, in itself, ensure we focus our efforts on a higher percentage of winners. This is true because, in conventional programs, it is still the selection of an initial lead series that determines the ultimate outcome, as outlined in Figure 5.
The drive to focus one's efforts is a well established managerial concept and in itself can convey progress that is comforting to the chemist, to project team and to management. This happens early on in most programs - made easier by the fact that we most often start with a potency screen to pick the most active hits and, thereafter, a preferred lead series. At that point the serious work of Discovery usually begins in a multiyear, series of iterative steps as we seek acceptable levels of bioavailability, genetic toxicity, pharmaceutical properties, drug metabolism and so on for our chosen series. As has been elaborated by Lipinski and his coworkers, the rise in the observed molecular weight, increased lipophilicity and decreased solubility of initial screening leads has made the task of lead to drug conversion far more difficult (13). The provocative observation is that this degradation in the chemical starting point for recent Discovery programs coincides with our move away from pharmacological screens to in vitro molecular screens. Drugability has been the victim and increasing candidate selection criteria will not rescue a suboptimal early choice. This suggests that to change the productivity of the system we need to move beyond a focus on increasing the hurdle rate to fundamental changes in our processes for selecting and developing leads. If we do not change, it seems likely that even with vastly larger files we will still "select to fail."
Imagine that we have created an industry like screening file as indicated in Figure 5, and that we have access to all the leads previously represented by Companies A - D. What would we do? In general, the practice and the instinct is to focus in relatively quickly on one or two series based on initial probes to see which series most quickly delivers a strong SAR based on potency measures. Most typically after several months of work we would have a mid-nanomolar lead from one of these series. But is it the "right" one? Psychologically, there is great comfort from a nanomolar lead which makes it very difficult for any chemist to turn back to a micromolar hit even if it has superior drug potential as measured by molecular weight and other elements captured by the rule of five, In fact, many of us have learned that getting nanomolar potency is seductively easy compared to getting all the properties of drugability built in up front. We also often have all too convenient confidence in the ability of our colleagues in pharmaceutical sciences to rescue us downstream. The problem may also owe part of its origins to the compounds in our libraries and to the precision of our screening practices.
Historically the progress of the sciences has most often been linked to advances in the quantitative and qualitative precision of the analytical "eyes" we bring to bear on a problem. This is certainly true of the contribution of analytical chemistry to organic chemistry. There is a comparable and clear opportunity to push for the same end to end QA/QC if the screening and lead generation process. For example, by setting our screening criteria so that we achieve a manageable hit rate an unintended consequence is often that highly desirable 300 MW scaffold hits may never even reach our attention. Also because of the comfort that comes from large numbers we may not work as hard on the making sure that we understand the dynamics of our screening conditions. As one illustration, the impact of different procedures on hit identification is highlighted by the work of Sills and coworkers (17). Screening for tyrosine kinase inhibitors by three different methods they showed surprisingly little overlap in the compounds identified as leads. More investment in quality thinking and experimentation at the front end of the process is a shared chemistry and biology opportunity to improve the outcome.
The ultimate trap is the same one we experience with energy minimization procedures i.e. once you are in a well (or a hole) it is very hard to get out and detect the presence of other potentially superior minima and yet is hard not to keep digging deeper. If we express our scheme and sequenced series of hurdles in this way you can see a graphical representation of the initial selection of a series is determinative in an iterative, serial process (Figure 5).
Seeking the Most Drugable Lead
Was this article helpful?