Abstract LB-112: PLK1 inhibition in genetically characterized triple-negative breast cancer (TNBC) patient-derived xenograft (PDX) models results in robust antitumor activity including significant and durable tumor regression

Background: PLK1 plays multiple critical cell cycle roles and numerous preclinical studies have validated the impact of PLK1 inhibition in tumors. A Phase II trial of a PLK1 inhibitor in AML showed significantly increased ORR and OS (Döhner et al 2014), but results in solid tumors have been uniformly modest. However, recent studies indicate that PLK1 inhibition may yield greater responses in specific genetic subtypes. Maire et al (2012) reported impressive response and survival data in TNBC PDX models, and PLK1 has been implicated in promoting growth and survival in cooperation with MYC, PTEN-loss, the PI3K/mTOR pathway and AR signaling in prostate cancer and other models (Zhang et al 2014, Hou et al 2013, Liu et al 2011 and Tan et al 2013). We therefore tested whether the highly selective PLK1 inhibitor NMS-P937 could differentially inhibit the growth of a diverse panel of genetically-characterized TNBC PDX models, enabling selection of a potentially higher responding patient population. Methods: Ten genetically-characterized TNBC PDX models containing a broad set of mutations were tested for sensitivity to NMS-P937. After implantation and establishment in SCID mice (n = 3 per group), NMS-P937 was administered at different doses and schedules over a 24-day treatment period. Body weight, tumor growth and growth delay after cessation of treatment were measured. Results: Nine models were evaluable and significant tumor growth inhibition (50-90%) was noted in four. Mutations in the responding models have been proposed to enhance response to PLK1 inhibition in other tumor types, but not previously in TNBC. All lines with sensitivity-enhancing mutations responded while lines without did not. In the most sensitive model, NMS-P937 caused uniform, near-complete regressions that were stable for 21 days after dosing, with minimal body weight loss. This model harbors an amplified AR locus; though Hou et al proposed a connection between PLK1 and AR signaling, the extent and durability of the response was unanticipated. Further characterization of all models will be provided. Conclusion: This study appears to be the first report of a PLK1 inhibitor evaluation in a diverse panel of genetically-characterized models in any tumor type. Our results demonstrate that differentially sensitive molecular subtypes of TNBC can be identified and that underlying genetic profiles can be exploited to identify TNBC patients who may benefit most from PLK1 inhibitors. We are continuing t