Profiling of new antitumor natural products discovered using a genomics/cheminformatics-based drug discovery platform

1922 Analysis of bacterial genomes using the DECIPHER® database, a proprietary natural product database containing biosynthetic cluster gene sequence information linked to compound chemical structure, physicochemical properties and biological activities, allows for the efficient identification of New Chemical Entities (NCEs) of interest. Complex, custom-tailored, bioinformatics systems enable prediction of chemical structures based on gene sequence information. Use of this knowledge guides selection of microorganisms and fermentation conditions as well as purification and structure confirmation of the compounds produced. Among various natural products discovered using this platform, two compounds, ECO-3396 and ECO-7942, were predicted to have antitumor activity based on prediction and comparison of their structure with those of known bioactive compounds. ECO-3396 is produced by Micromonospora echinospora and is a highly oxygenated angular polycyclic ketide with an unusual peptide linkage to an aliphatic tripeptide. ECO-7942 is produced by a Streptomyces sp. and is a cyclic hexadepsipeptide composed of unusual amino acid units and bearing a C14 side chain. ECO-3396 and ECO-7942 exhibited potent antimicrobial activities against a broad panel of Gram+ pathogenic bacteria. Advanced profiling performed on these compounds revealed moderate in vitro cytotoxic activity for compound ECO-3396, with GI 50 values in the low micromolar range, and potent cytotoxic activity for compound ECO-7942, with GI 50 values in the low nanomolar range against several human tumor cell lines. In vivo toxicity of these compounds using various routes of administration was evaluated and maximum tolerated doses defined in a mouse system. ECO-3396 showed acceptable levels of toxicity whereas ECO-7942 displayed lower tolerability with MTD ≥ 2.5 mg/kg when dosed i.v. and MTD ≥ 5 mg/kg when dosed i.p. In vivo antitumor activity of these compounds was assessed in mouse xenograft human carcinoma models. ECO-3396 was tested in PC-3 human prostate carcinoma and in MX-1 human breast carcinoma mouse models. In both models, ECO-3396 displayed a modest antitumor activity when administered i.p. compared to vehicle control groups. ECO-7942 displayed significant antitumor potential in the mouse xenograft MX-1 model, when compared to control groups. This compound is currently under preclinical profiling to further define its biological activities and pharmacological properties.Genome scanning analysis performed on microorganisms and data evaluation using the DECIPHER® database is a powerful tool that can be applied to the discovery of natural product compounds with antitumor activity.