We have adopted an approach to kinase drug discovery that we believe is significantly more productive and efficient than the conventional process. This method entails screening an entire compound library against a full panel of kinase assays. A conventional high-throughput screen against a single target only yields hits against that one target. Only one parameter, potency, defines those hits. In contrast, a single library profiling screen reveals hits against virtually all biologically interesting kinase targets, and delivers two critical parameters, potency and selectivity, to define each of the hits. As a result, decisions about which targets to pursue can be made based on much more complete information. Projects may be chosen, and precious resources committed, based not only on the level of biological interest in a specific kinase target, but also on knowledge about the quality of chemical hits for each target.
Once medicinal chemistry optimization is initiated, we continue to screen compounds synthesized against the kinase assay panel. This approach accomplishes two objectives. First, the selectivity of each compound is established upfront, and becomes a driver for optimization along with usual parameters such as potency and pharmacokinetics. Secondly, kinase interaction patterns can change dramatically with relatively minor changes in chemical structure, and new compounds produced for an established program may represent a lead compound for a new target of interest. Using this approach, we have captured and continue to capture the entire kinase interaction for each compound in an expanding database. This database represents an extensive annotation of our collection of kinase-focused compounds and obviates the need to initiate each new project with a new screening campaign, thereby accelerating the drug discovery process. We have developed a quantitative description of kinase interaction patterns that provides a conceptual framework for analyzing the compounds within our database.
By utilizing a suite of proprietary computational tools, we are able to efficiently interrogate our database and identify the most promising lead compounds, providing a significant competitive advantage for rapidly exploiting new targets. Using this approach we have, in the last five years, advanced four drug candidates from discovery into clinical development. We have accumulated a library of kinase-focused compounds. Each compound has been screened using our accelerated approach, and the results are readily accessible within the database. Mining our database yielded the starting points for our current pipeline, including quizartinib, AC430, CEP-32496, and earlier stage programs.