Tumor Suppressors Condition Differential Responses to the Selective CDK2 Inhibitor BLU-222.

Abstract

Cyclin-dependent kinase 2 (CDK2) inhibitors have recently been developed and have entered clinical trials. Combination approaches can help broaden the use of therapeutic agents and establish more effective treatments. Here, we evaluated the selective CDK2 inhibitor BLU-222 for mechanisms of response in the context of ovarian and breast cancer models. Sensors of cellular CDK activity indicated that sensitivity to either CDK4/6 or CDK2 inhibition was related to the differential dependence on a single CDK for G1-S transition. Unlike CDK4/6 inhibitors, BLU-222 was able to robustly inhibit proliferation through cell-cycle inhibition in both G1 and G2 phases. However, it remained possible for cells to reenter the cell cycle upon drug withdrawal. The antiproliferative strength and impact on G1-S versus G2-M accumulation was mediated by the RB tumor suppressor. To broaden the sensitivity to CDK2 inhibition, combinatorial drug screens were performed that identified both synergistic (e.g., CDK4/6 inhibitors) and antagonistic (e.g., WEE1 inhibitors) relationships. Models that were exceptionally sensitive to CDK2 inhibition displayed coordinate expression of cyclin E1 and P16INK4A, an endogenous CDK4/6 inhibitor. Functional studies demonstrated that P16INK4A and CDK4/6 activity were key mediators of sensitivity to BLU-222. Clinical gene and protein expression analyses revealed a positive correlation between cyclin E1 and P16INK4A and identified that ∼25% of ovarian cancers exhibited coordinate expression of cyclin E, P16INK4A, and RB, indicative of strong sensitivity to CDK2 inhibition. Together, this work advances a precision strategy for the use of CDK2 inhibitors in the context of ovarian and breast cancers. Significance: The CDK2-specific inhibitor BLU-222 shows preclinical efficacy in breast and ovarian cancer with select determinants of response and holds promise in combinatorial strategies. See related article by House and colleagues, p. 1297.