Abstract 1782: Human tumor explants are better predictors of clinical trial outcome than cell line xenografts for the KSP inhibitor ARRY-520

Abstract

Abstract Historically, human tumor cell line cultures and xenografts have served as standard models for identifying and optimizing anticancer drugs. Despite wide use, cell line models are poor predictors of clinical activity of drugs, particularly cytotoxics. Primary human tumor explant models have significant potential to inform clinical direction. These models have not undergone genetic drift or selection for growth in culture. Furthermore, low-passage explants retain the morphology and aspects of the tumor microenvironment reflecting the primary cancer in patients. These models further represent patients who have been treated with current chemotherapeutics. ARRY-520 is a novel kinesin spindle protein inhibitor which has been investigated in clinical studies in both solid and hematological cancers. In cell line xenografts, ARRY-520 has significant activity in solid tumors and even greater activity in hematological cell lines both in vitro and in vivo. In clinical studies, while ARRY-520 has demonstrated significant single-agent activity in multiple myeloma, activity in solid tumors has been modest. Thus, while preclinical data in hematological models has correlated with clinical activity in myeloma, the impressive solid tumor data has not translated to clinical activity in solid tumors. To probe the disparity of preclinical and clinical activity in solid tumors, we have investigated the ability of preclinical models to inform clinical success by retrospective analysis of the activity of ARRY-520 in solid tumor cell line and explant models, as compared to clinical data of this molecule in these indications. ARRY-520 shows potent activity in human tumor cell lines both in vitro, (IC50 0.5-14nM) and in vivo (50% tumor regression in 4/5 cell line xenografts CRC and PaCa including 100% regression in 2 models). By contrast, in 6 primary CRC and PaCa explants, ARRY-520 showed minimal activity (transient stable disease in 1/6 models and progressive disease in 5/6 models). Phase 1 solid tumor data on ARRY-520 was reported at ASCO 2010. In this study, 30 patients with solid tumors, treated at or above the MTD, were evaluable for response with best response of SD>12 weeks in 2 patients (7%). Within this study, there were 4 CRC and 3 PaCa patients, all of whom experienced progressive disease within 12 weeks of initiation. Further investigation of ARRY-520 in solid tumor explants is ongoing. Analysis of responsive and non-responsive models will be used to identify solid tumor types for further clinical development. In summary, in this study, human solid tumor explant xenografts more accurately predicted the actual clinical results from a Phase 1 solid tumor trial. These data suggest that solid tumor explant models may have significant utility in informing clinical decisions. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 1782. doi:1538-7445.AM2012-1782