Epithelial-to-Mesenchymal Transition in Breast Cancer: A Multi-Level Study Integrating Patient Tissues, Sera, and Cellular Systems

The epithelial-to-mesenchymal transition (EMT) is a complex process pivotal in embryonic development, wound healing, and tumor progression, particularly in breast cancer (BC), a heterogeneous disease in which patients diagnosed at the same stage exhibit different clinical responses and survival outcomes. In BC progression, the EMT enhances cell survival and helps cancer cells acquire metastatic behaviors by losing epithelial phenotypic characteristics and acquiring mesenchymal traits. This thesis aims to better understand EMT landmarks in BC, which are useful for patient stratification and/or the development of targeted therapies.To this end, the protein expression patterns of key epithelial (E-cadherin, Cytokeratin-18) and mesenchymal (Vimentin, α-SMA) markers were first analyzed in a cohort of BC tissue samples by western blot. BC tissues displayed marked intra- and inter-patient heterogeneity and unexpected positive correlations between epithelial and mesenchymal markers. This idea was further supported by analyses of paired tumoral and adjacent normal tissues, revealing that EMT markers were frequently expressed at equal or higher levels in the healthy samples. Together, these findings highlight the complexity of EMT regulation in breast cancer and suggest that EMT-related features are not exclusively restricted to malignant epithelial compartments.Correlations between mesenchymal markers and tumor microenvironment components, analyzed using the TIMER database, revealed significant associations between EMT features and stromal and immune elements, supporting the idea that EMT dynamics are strongly shaped by microenvironmental cues rather than being purely tumor cell-intrinsic programs.To further explore the relationship between EMT markers and tumor aggressiveness, the analysis was extended to patient-derived sera. Sera from patients with primary or metastatic breast cancer consistently enhanced proliferation in both malignant (MDA-MB-231) and non-transformed (MCF10A) mammary epithelial cells, whereas sera from healthy donors showed a suppressive effect. These findings suggest that breast cancer sera convey systemic signals that promote pro-growth and pro-survival phenotypes. Furthermore, breast cancer sera significantly increased MDA-MB-231 cell migration in wound healing assays. In contrast, healthy sera reduced wound closure, suggesting a potential role for circulating tumor-associated factors in modulating migratory behavior. Moreover, western blot analyses showed high variability and positive correlation of epithelial and mesenchymal markers, further supporting a context-dependent modulation of EMT programs.To better capture tumor complexity beyond 2D systems, EMT regulation was investigated in 3D tumorspheres. Compared with two-dimensional monolayer conditions, MDA-MB-231 tumorspheres showed reduced proliferation (assessed by MKI67) and VIM expression, with a modest but non-significant increase in CDH1, consistent with enhanced cell–cell interactions. At the same time, these epithelial-associated features coexisted with the activation of EMT-related transcription factors (SNAI1, TWIST1), along with a trend toward increased CDH2 expression. Furthermore, SOX2 and FN1 were upregulated, reflecting a plastic, hybrid phenotype rather than a classical EMT program.When exposed to patient-derived sera, the 3D model was more responsive than 2D cultures. Indeed, while healthy sera disrupted tumorspheres’ integrity and downregulated proliferation- and EMT-related genes, breast cancer sera preserved spheres’ architecture and maintained gene expression near control levels. Although not all differences reached statistical significance, the overall transcriptional profile induced by breast cancer sera consistently diverged from that observed with healthy sera, supporting a biologically relevant, serum-driven effect.To further capture EMT complexity, a bioinformatic approach integrating EMTome, UALCAN, and Kaplan-Meier Plotter databases was applied, which identified a breast cancer-specific EMT signature. Enrichment analysis using STRING highlighted pathways related to extracellular matrix remodeling and cell-cell/cell-matrix interactions, reinforcing the link between EMT and microenvironmental dynamics. To define a highly stringent breast cancer-specific EMT gene signature, the analysis was restricted to genes with both diagnostic and prognostic significance, yielding 37 concordant genes across UALCAN and Kaplan-Meier platforms. These genes are associated with stemness programs, supporting the idea that heterogeneous microenvironmental signals sustain cancer stem cell traits and hybrid EMT states.Clinical validation using bcGeneXMiner and GOBO confirmed that poor-prognosis genes correlated with aggressive clinicopathological features. In contrast, favorable genes showed the opposite pattern, underscoring the strong association between EMT-related transcriptional programs and breast cancer aggressiveness.Finally, given the central role of metalloproteases in extracellular matrix degradation and EMT-associated invasion, gelatin zymography was used to evaluate the expression of their pro-forms as functional markers of EMT-related remodeling. Zymographic analysis revealed heterogeneous yet widespread gelatinolytic activity in breast cancer tissues, with complex correlations with several EMT markers. Notably, when performed using BC sera, zymographic analysis identified circulating pro-MMP-2 as a potential indicator of disease progression. Indeed, pro-MMP-2 levels decreased or remained stable in regressing tumors but increased in progressing disease. In contrast, pro-MMP-9 showed no comparable trend. Overall, these findings support the view that EMT in breast cancer is a dynamic, context-dependent, and highly plastic process. Rather than a linear, uniform program, it comprises a spectrum of hybrid states shaped by interactions between the tumor microenvironment and systemic circulating factors. Therefore, fully understanding EMT complexity requires integrative strategies combining tissue- and serum-based analyses, functional assays, and bioinformatic profiling to more accurately capture its biological relevance and clinical impact.