Investigating the Influence of the Extracellular Matrix on Triple Negative Breast Cancer in Diverse Patient Samples

Triple negative breast cancer (TNBC) is an aggressive subtype of breast cancer associated with poor prognosis, high rates of recurrence, and few targeted treatment options. The heterogeneity of TNBC and high metastatic potential complicates treatment strategies, making identifying the mechanisms that drive disease progression essential. African American (AA) women are disproportionately affected by TNBC and experience higher incidence rates, worse prognoses, and lower survival rates compared to their European American (EA) counterparts. Recent investigations have revealed ancestry-based differences in both molecular mechanisms and the tumor microenvironment, though most preclinical models and clinical trials still lack sufficient representation of global diversity. This dissertation explores molecular mechanisms of cancer and metastasis by utilizing in vitro and ex vivo samples from diverse patients and employs novel machine learning approaches to characterize gene and protein expression further. The epithelial to mesenchymal transition (EMT) is a dynamic process by which cancer cells convert from an epithelial to a mesenchymal phenotype, characterized by increased migration and invasiveness. Tumor biopsy samples from AA and EA patients revealed differential expression of 5 EMT-associated proteins quantified by a deep-learning AI model. Further, we analyzed gene expression from four regions across the tumor and surrounding extracellular matrix. Machine learning models highlighted the tumor/matrix border regions as the most distinctive between groups and identified over 20 differentially expressed EMT-associated genes. One gene of interest is S100A4, a protein that functions both intra- and extracellularly to promote tumor progression and metastasis. This study shows that S100A4 knockout via CRISPR-Cas9 decreases the migratory and invasive capacity of TNBC and non-tumorigenic mammary cells from diverse patients. Simulated co-culture with fibroblasts induced contrasting responses in cell lines from AA and EA patients, and mRNA expression analysis revealed time-dependent suppression of EMT gene pathways. This work addresses critical gaps in existing TNBC research and lays the foundation for future preclinical models that consider ancestry as a fundamental variable, advancing breast cancer treatment and outcomes.