Abstract 4410: Predictive model of palbociclib response reveals indications in which CDK4/6 inhibitor can be a potential combination partner

Abstract

CDK4/6 inhibitors are a promising class of targeted therapies, with three molecules (palbociclib, ribociclib, and abemaciclib) approved for advanced hormone-receptor positive (HR+) breast cancers. With hope of building on this success, clinical trials across multiple oncology indications have been initiated. The determinants of response to CDK4/6 inhibitors, beyond pRb-deficiency and Cyclin-E amplification, remain unclear as limited clinical data have been published in indications other than HR+ breast cancers. Here, we measured the response to palbociclib across a panel of tumor cell lines from multiple indications and were able to build a computational model predictive of drug response using the expression of a handful of transcripts. This model revealed cellular states that are related to palbociclib resistance, such as G1-S entry deregulation, or sensitivity, such as competent p53 pathway and high KRAS signaling. As validation of our model, we compared its predictions with dependency scores from Project Achilles and found good correlations between predicted sensitivity and dependence on CDK4 and CCND1. By applying the predictive model on TCGA data, our analysis uncovered indications, disease subtypes, and genomic alterations that are related to the predicted response to palbociclib in addition to Rb-deficiency and Cyclin-E amplification. As expected, HR+ breast cancer was predicted to be among the most sensitive diseases, along with prostate, thyroid, and some renal cancers. Among non-small cell lung cancers, squamous-cell tumors are predicted to be more resistant than adenocarcinomas. In adenocarcinomas, BRAF and KRAS mutations are predicted to be associated with increased responsiveness. Combining these results with published biomarkers of resistance to standard of care therapies reveals indications in which CDK4/6 inhibitors can be potential combination partners. These results may guide the design of new trials for CDK4/6 therapies and the evaluation of better stratification biomarkers. Citation Format: Marc Hafner, Nicholas Dompe, Wei Zhou, Eva Lin, Milena Durrbaum, Anneleen Daemen, Melissa R. Junttila, Ciara Metcalfe, Karl Merrick. Predictive model of palbociclib response reveals indications in which CDK4/6 inhibitor can be a potential combination partner [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 4410.