Half of all triple negative breast cancer (TNBC) patients harbor significant residual cancer burden following standard neoadjuvant chemotherapy treatment, resulting in distant metastasis and death for most of these patients. Intra-tumor heterogeneity (ITH) is pervasive in TNBC as is a barrier to development of effective therapeutic strategies, but the relative contributions of heterogeneous tumor cell populations to chemoresistance and metastasis are not well understood. To investigate the clonal dynamics that accompany chemotherapy treatment and metastasis, we employed orthotopic patient-derived xenograft (PDX) models of treatment-naïve TNBC, thus enabling experimentation with heterogeneous populations of human tumor cells that have undergone minimal manipulation.

To monitor the fates of PDX tumor cell lineages as they metastasized, we introduced a pooled lentiviral barcode library (Cellecta) into freshly dissociated PDX tumor cells which were orthotopically engrafted into recipient mice. Genomic analyses, including barcode enumeration, whole-exome sequencing, custom targeted DNA sequencing, and transcriptome sequencing, were conducted to characterize the clonal dynamics that accompanied metastasis. Of the thousands of diverse primary tumor clones, only ~2% harbored metastatic capacity. Of those, a rare population of the exact same clones predominated metastases in lung, liver, and brain, the three most common sites of human TNBC metastasis. These studies provide a quantitative map of the clonal architecture of multi-organ metastasis in TNBC and reveal that identical subclones can thrive in diverse secondary organ microenvironments.

NACT resistance leads to metastasis and death for most patients, yet the origins of chemoresistance in TNBC are unclear. We modeled NACT resistance in an array of PDX models derived from treatment-naïve TNBC biopsies in alignment with an ongoing neoadjuvant clinical trial (NCT02276443). Upon partial response to NACT, tumors entered a transient drug-tolerant state characterized by distinct histologic, proteomic, and transcriptomic features that were reverted as tumors regrew after cessation of treatment. Barcode-mediated lineage tracking and whole-exome sequencing revealed that the drug-tolerant state was not mediated by clonal selection. Based on transcriptomic and metabolic features of the drug-tolerant state, we conducted preclinical trials with an inhibitor of mitochondrial oxidative phosphorylation (IACS-010759), which significantly delayed the regrowth of residual tumors in PDX models. Together, these studies revealed that TNBCs can resist NACT through non-selective adaptation of a reversible phenotypic state, and that inhibition of oxidative phosphorylation may be a promising therapy in the neoadjuvant setting for TNBC.

Citation Format: Gloria V. Echeverria, Zhongqi Ge, Sahil Seth, Emily Powell, Xiaomei Zhang, Sabrina Jeter-Jones, Xinhui Zhou, Yan Jiang, Aaron McCoy, Shirong Cai, Yizheng Tu, Michael Peoples, Yuting Sun, Huan Qiu, Christopher Bristow, Alessandro Carugo, Jiansu Shao, Stacy L. Moulder, William F. Symmans, Timothy P. Heffernan, Jeffrey T. Chang, Helen M. Piwnica-Worms. Clonal dynamics and phenotypic evolution during chemoresistance and metastasis revealed by patient-derived xenograft models of triple negative breast cancer [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 2901.