Towards personalized medicine: Establishment and in vivo characterization of a low passage tumor xenograft tumor predictive of clinical response

B22 Human tumor xenograft models originating from high passage, commercially available cell lines are routinely used in anticancer drug development. While information from these studies is often useful in a preclinical setting, clinical relevance is often limited likely due to genetic alterations and adaptations from successive, long-term in vitro and in vivo passages. In an attempt to develop more clinically relevant preclinical models, we and others have successfully established low passage xenograft models from patient tumors engrafted into nude mice. However, whether these models are more predictive of clinical response is unclear. To address this question, we established a low passage tumor model from a non-small cell lung patient with an available clinical history which included platinum resistance. Based on these data, we carried out three in vivo studies evaluating past, current, and prospective therapies focused on targeted therapy combinations. Consistent with clinical results, the tumor model was insensitive to platinum regimens including cisplatin alone or an oxaliplatin/pemetrexed combination. However, cisplatin resistance was partially reversed with co-administration of the novel endoribonuclease, ranpirnase. Treatment with a sorafenib/bevacizumab regimen resulted in impressive tumor growth inhibition which was further enhanced with irinotecan including tumor regressions in half of the treated mice. Interestingly, the donor patient responded to this triple therapy with a 50% maximal remission and continued with stable disease on a sorafenib/bevacizumab regimen. Taken, together, these results demonstrate this model as predictive for clinical response of this patient. In addition, these data suggest low passage xenograft models may serve as a useful tool to better predict human response to test agents in a preclinical setting. Additional studies are currently underway to further validate and refine the use of these models in oncology drug development.