Key points are not available for this paper at this time.
Abstract Introduction: Drug action is mediated by a small molecule therapeutic binding to a specific structural motif on a protein. As structure determines function, it follows that unique structural motifs are enriched for sites of unique function and high therapeutic target value. We hypothesized that protein structure constitutes high value primary information that can drive the discovery of novel therapeutic agents. Method: We developed a unique physics-based protein structure analysis platform scaled to run on a supercomputer, allowing us to consider dynamic protein flexibility and the influence of solvent upon protein structure, and to probe large protein libraries as a primary source of structure information that can support up-front structure-based screens. We developed and then probed a high resolution protein x-ray crystallographic library, created and curated by us. From a 60 million compound library, we conducted virtual screens for docking to identified sites. 8-day colony formation assays were conducted on a panel of 8 human breast, prostate, lung and colon cancer cell lines, as well as 8- and 14-day trilineage hematopoietic colony formation assays on human cord blood stem cells. Findings: From the first 180 deposited protein structures, we screened for unique protein surface structure. Structures were deemed unique if they were not otherwise known to be associated with any particular functional roles and were not the binding site of known therapeutics. Further selection included those that were clefts and possessed physical characteristics compatible with binding small chemicals whose properties have been linked to effective therapeutics. For each of 8 sites on 6 different proteins, a suite of analytics probed docking solutions from the compound library, generating a ranked list of 100 compounds for each site. Based on factors inclusive of availability and low potential for toxicity, compounds were acquired and used in in-parallel selection and de-selection functional assays. Cancer cell growth inhibition was a positive selection criterion, with additional ranking based on greater potency and broader efficacy. Bone marrow toxicity is limiting for most drugs, and inhibition of bone marrow stem cell growth was as a deselection criterion. Multiple compounds clustered around 2 sites on 2 different proteins. One compound, Dxr2-017, exhibited very high relative efficacy against colon and breast cancer cells. Evaluation of Dxr2-017 in the NCI-60 panel corroborated these findings, and showed even higher efficacy against melanoma. We demonstrated that Dxr2-017 inhibited melanoma growth with low nanomolar efficacy, while concentrations of over 2300-fold higher had only limited inhibitory effects on bone marrow stem cells. Conclusions: This study provides proof-of-principle that protein structure constitutes high value primary information that can support identification of novel therapeutics. This approach is widely applicable and expandable. Citation Format: Fangfang Qiao, T. Andrew Binkowski, Wayne Anderson, Weining Chen, Gray Schiltz, Karl Scheidt, Amarnath Natarajan, Raymond Bergan. Protein structure inspired drug discovery abstract. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts) ; 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84 (6Suppl): Abstract nr 4493.
Qiao et al. (Fri,) studied this question.