The primary goal of the Douglass Lab’s Organic Program is to modify broad-spectrum chemotherapies like Taxol to overcome chemo-resistance mechanisms like Pgp mediated drug-efflux (Figure 1). To maximize clinical utility, this work focuses on the most commonly used drugs in the clinic (e.g. taxanes, anthracyclines, camptothecins, anti-metabolites, etc.) and drug-resistant genes that show the highest expression across patient tumors (ABCB1, ABCG2, etc.). Rationale design of therapies is based on careful definitions of:
- Historical meta-SAR analysis: to define accessible chemical space that maintains target-binding and pharmacokinetic properties using databases like ChEMBL.
- Drug-target binding structures: to maintain drug-binding to canonical targets (Figure 1 target in grey)
- Chemical-motifs that characterize MDR-substrates: to define the molecular determinants of drug-resistance (Figure 1 motif in red) to be removed or blocked (Figure 1 modification in green)
MDR “chemical-motifs” must be defined probabilistically, because of the broad substrate scope of most MDR-enzymes. This chemo-informatic work is modeled after the pioneering work of Dr. Anna Seelig who identified characteristic properties of Pgp substrates (Eur. J. Biochem., 1998, 251, 252) including the H-bond accepter motifs (Figure to the right) used to design Tesetaxel above. To derive these rules, Dr. Seelig manually curated a database of >100 different Pgp-substrates and defined several molecular features associated with Pgp-mediated drug-efflux. While innovative, this manual procedure is not scalable and does not take advantage of recent advances in machine learning. We plan to systematize and automate Dr. Seeligs work across all drugs and resistance genes by combining new chemical and genomic databases with modern methods in statistical learning.