The benzoquinone ansamycin Hsp90 inhibitors 17-AAG and 17-DMAG are in Phase II trials but their use is limited by hepatotoxicity. In the last grant period we found that the active forms of the BQA Hsp90 inhibitors were their hydroquinone metabolites generated by the quinone reductase NQO1 rather than the parent quinones. We defined the molecular determinants of hepatotoxicity of BQA Hsp90 inhibitors and have focused on the development of entirely novel 19-substituted BQAs (19BQAs) which are metabolized by NQO1 to active hydroquinone ansamycin Hsp90 inhibitors. 19BQAs were rationally designed in our labs (DR and CJM) specifically to prevent off-target interactions with glutathione and protein thiols and to be markedly less toxic than the current BQA derivatives in clinical trials. Preliminary data validates the underlying rationale and shows that 19BQAs do not interact with cellular thiols and are not toxic in mouse hepatocyte systems. In this proposal, we will focus on testing the hypothesis that 19BQAs are potent Hsp90 inhibitors that have markedly less toxicity than their parent quinones using both in-vitro and in-vivo approaches. We will finalize the structure-activity relationship for 19BQAs in the geldanamycin, 17-AAG and 17-DMAG series and characterize their structure and binding to purified Hsp90 using X-ray and NMR analysis. Cellular studies will define a) the Hsp90 and growth inhibitory effects of 19BQAs in both human breast and pancreatic tumor cells b) metabolism of 19BQAs to their hydroquinones by NQO1 using isogenic breast and pancreatic tumor cell lines differing only in NQO1 and c) the effects of 19 BQAs on association of Hsp90 with co-chaperones which modulate Hsp90 function and deliver key client proteins essential for growth to Hsp90. The hepatotoxicity of 19BQAs in both mouse and human hepatocyte systems will be characterized and together with data from studies in tumor cell systems will allow us to advance compounds with the greatest therapeutic index to in-vivo testing in xenograft and explant systems where the use of an integrated PK-PD model will address the in-vivo therapeutic selectivity of 19BQAs. Using a mechanistic approach, these studies will validate the potential of an entirely novel class of Hsp90 inhibitors designed to be markedly less hepatotoxic than the BQA Hsp90 inhibitors currently in clinical trials.