IDENTIFICATION OF INTRA-TUMORAL HETEROGENEITY, DRUG-INDUCED CELLULAR SUBPOPULATION SHIFTS, AND NOVEL METABOLIC PROFILES IN TRIPLE NEGATIVE BREAST CANCER MOUSE MODELS

Introduction: Triple Negative Breast Cancer (TNBC) is an aggressive disease with a poor prognosis that accounts for 10-20% of breast cancer cases. Intra-tumoral heterogeneity and cell plasticity are thought to contribute to drug resistance in TNBCs. Further, unique metabolic differences in TNBCs may act as vulnerable pharmacological targets. This work aims to: 1) precisely identify cell state heterogeneity within TNBCs, 2) test whether drugs alter cellular subpopulations, and 3) further subdivide TNBCs by metabolic profiles. I hypothesize that treatment(s) induce cellular subpopulation shifts into resistant states. I also hypothesize that different TNBCs have a unique metabolic profile and are sensitive to glutamate and de novo pyrimidine metabolic pathway inhibitors. Methods: I used the TP53-/- Genetically Engineered Mouse Model (GEMM) syngeneic transplant tumors of the basal-like TNBC phenotype with the chemotherapeutic carboplatin/paclitaxel as well as inhibitors of the EGFR-MAPK pathway components EGFR and MEK, chromatin remodeling protein BRD4, and metabolic regulator dihydroorotate dehydrogenase (DHODH). To examine response to treatment, I performed both in vivo drug treatments and gene expression profiling using single cell RNA-sequencing (scRNA-seq). To identify metabolic phenotypes in TNBCs, I examined the different metabolites in patient breast cancer samples, patient-derived xenografts (PDX), and cell lines for both TNBC and ER+ cancers in vitro and in vivo by xenograft transplant. Results: I have identified clear intra-tumoral heterogeneity in several in vivo GEMMs. I performed 25 individual scRNA-seq experiments on the 2225L GEMM transplant line with the aforementioned treatments in triplicate and created an integrated dataset. In this dataset, different treatments gave rise to 2 rare subpopulations (i.e. 1-4%) that express genes consistent with previously described drug-tolerant persisters (DTP), annotated “Epithelial-DTP” and “Mesenchymal-DTP”. The Epithelial-DTP gene expression signature predicted poor patient outcomes in TNBC patients. In the metabolism study, analysis of metabolic changes across different breast cancer models led to the identification of two metabolic groups: “Nucleotide/Carbohydrate-Enriched” and “Lipid/Fatty Acid-Enriched”. This led to the generation of a metabolomics-informed gene expression signature that is predictive of overall survival in TNBC human datasets.Conclusion: Overall, the work in this dissertation provided new insights on TNBC heterogeneity, metabolism, and response to treatment.