Pharmacologic profiling of patient-derived xenograft models of primary treatment-na\xefve triple-negative breast cancer

Reid T Powell,Peter J Davies,Xinhui Zhou,Xiaomei Zhang,Yan Jiang,Jason B White,Virginia Giuliani,Lei Guo,Xiao Yan ,W Fraser Symmans,Jeffrey T Chang,Joseph R Marszalek,Xuan Liu,Christopher P Vellano,Shirong Cai,Ningping Feng,Timothy P Heffernan,Gloria V Echeverria ,Stacy L Moulder ,Abena B Redwood ,Yuan Qi,Yizheng Tu,Clifford Stephan,Helen Piwnica‐Worms

doi:10.1038/s41598-020-74882-4

Abstract

Triple-negative breast cancer (TNBC) accounts for 15–20% of breast cancer cases in the United States, lacks targeted therapeutic options, and is associated with a 40–80% risk of recurrence. Thus, identifying actionable targets in treatment-naïve and chemoresistant TNBC is a critical unmet medical need. To address this need, we performed high-throughput drug viability screens on human tumor cells isolated from 16 patient-derived xenograft models of treatment-naïve primary TNBC. The models span a range of TNBC subtypes and exhibit a diverse set of putative driver mutations, thus providing a unique patient-derived, molecularly annotated pharmacologic resource that is reflective of TNBC. We identified therapeutically actionable targets including kinesin spindle protein (KSP). The KSP inhibitor targets the mitotic spindle through mechanisms independent of microtubule stability and showed efficacy in models that were resistant to microtubule inhibitors used as part of the current standard of care for TNBC. We also observed subtype selectivity of Prima-1Met, which showed higher levels of efficacy in the mesenchymal subtype. Coupling pharmacologic data with genomic and transcriptomic information, we showed that Prima-1Met activity was independent of its canonical target, mutant p53, and was better associated with glutathione metabolism, providing an alternate molecularly defined biomarker for this drug.

Highlights

Triple-negative breast cancer (TNBC) accounts for 15–20% of breast cancer cases in the United States, lacks targeted therapeutic options, and is associated with a 40–80% risk of recurrence
We used a panel of orthotopic xenograft models generated from patients with newly diagnosed primary TNBC, who had not yet received neoadjuvant chemotherapy (NACT), as a resource for identifying targeted therapies with preclinical activity in primary TNBC
We established and validated a preclinical pharmaco-genomic resource generated from patient-derived xenograft (PDX) models from representative treatment-naïve tumors of patients with primary TNBC

Summary

Introduction

Triple-negative breast cancer (TNBC) accounts for 15–20% of breast cancer cases in the United States, lacks targeted therapeutic options, and is associated with a 40–80% risk of recurrence. Identifying actionable targets in treatment-naïve and chemoresistant TNBC is a critical unmet medical need To address this need, we performed high-throughput drug viability screens on human tumor cells isolated from 16 patient-derived xenograft models of treatment-naïve primary TNBC. The most heavily utilized models in HTS have been established cell lines[8,9], which may fail to recapitulate the behavior of primary cells and lack the heterogeneity seen in human tumors[10,11,12] For these reasons, many translational studies rely on patient-derived xenograft (PDX) models, which are thought to more accurately recapitulate the genomic heterogeneity of the patient tumors from which they were derived[11,13,14,15,16]. We used a panel of orthotopic xenograft models generated from patients with newly diagnosed primary TNBC, who had not yet received neoadjuvant chemotherapy (NACT), as a resource for identifying targeted therapies with preclinical activity in primary TNBC

Methods

Results

Conclusion