BIM Expression in Treatment-Naive Cancers Predicts Responsiveness to Kinase Inhibitors

Cancers with specific genetic mutations are susceptible to selective kinase inhibitors. However, there is a wide spectrum of benefit among cancers harboring the same sensitizing genetic mutations. Herein, we measured apoptotic rates among cell lines sharing the same driver oncogene following treatment with the corresponding kinase inhibitor. There was a wide range of kinase inhibitor-induced apoptosis despite comparable inhibition of the target and associated downstream signaling pathways. Surprisingly, pretreatment RNA levels of the BH3-only pro-apoptotic BIM strongly predicted the capacity of EGFR, HER2, and PI3K inhibitors to induce apoptosis in EGFR-mutant, HER2-amplified, and PIK3CA-mutant cancers, respectively, but BIM levels did not predict responsiveness to standard chemotherapies. Furthermore, BIM RNA levels in EGFR-mutant lung cancer specimens predicted response and duration of clinical benefit from EGFR inhibitors. These findings suggest assessment of BIM levels in treatment-naive tumor biopsies may indicate the degree of benefit from single-agent kinase inhibitors in multiple oncogene-addiction paradigms. SIGNIFICANCE: In several oncogene-addiction paradigms, assessment of BIM RNA levels identifies those cancers that fail to have substantial apoptotic responses to kinase inhibitors. BIM RNA levels may be assessed in diagnostic cancer specimens to predict which patients will receive less benefit from single-agent kinase inhibitors. Cancer Discovery: 1(4); 352-65. (C) 2011 AACR.