Genetic E-Cadherin Loss Creates a Distinct Phenotype in an Early ILC Tumorigenesis Model

Abstract Text Invasive lobular carcinoma (ILC) is a common subtype of breast cancer, and despite being ∼95% estrogen receptor α positive (ER+) with other ‘low risk’ biomarkers, ILC has the highest risk of long-term recurrence among breast cancers. ILC are genetically distinct, with hallmark mutation/deletion of CDH1 leading to loss of cell adhesion protein E-cadherin. Genetic CDH1 loss is an early event in ILC tumorigenesis, yet the mechanisms by which E-cadherin acts as a tumor suppressor, with loss facilitating tumorigenesis, are not well understood and are likely highly context dependent since ILC are rarely ER- or HER2+, and nearly all are luminal-type (ER+). To study the role of E-cadherin loss as a driver of lobular tumorigenesis, we established the use of hyperplastic human mammary epithelial cells (HMECs, with CCND1 overexpression) with various modes of E-cadherin suppression: extracellular antibody inhibition; siRNA transfection (siCDH1); dox-inducible CRISPR/Cas9 CDH1 knockout. In RNAseq studies, using three independent HMEC models (HMEC 122, 153, 184), all forms of E-cadherin suppression were strongly associated with signatures of epithelial-mesenchymal transition (EMT) and estrogen response (adj. p values: 1x10-5-2x10-28). However, inhibiting cell to cell contact with the E-cadherin antibody induced EMT; conversely, genetic loss of E-cadherin via siRNA or CRISPR – i.e. as observed in ILC – repressed EMT. Importantly, though E-cadherin loss is a canonical part of EMT, ILC tumors present no mesenchymal features, supporting that the genetic loss of CDH1 is distinct from loss of E-cadherin cell-cell contacts. To further examine the epithelial gene increases upon genetic E-cadherin loss, we used a 600-gene signature of the two main epithelial cell types in the mammary gland, luminal and basal, which showed a particular increase in expression of luminal progenitor features specific to CDH1 siRNA/CRISPR. Using flow cytometry, we confirmed that E-cadherin loss increased the luminal progenitor population, observed by an increase in the EpCAMhi/CD49fhi population, and an increase in CD10 (MME) specifically in this sub-population, in all HMEC lines. Notably, CD10 did not increase in basal (EpCAMlow) cells. Together these results support that genetic E-cadherin loss shifts gene expression and cell phenotype toward a more luminal progenitor state, and differentially impacts luminal versus basal cells. An enrichment in luminal progenitor phenotype is consistent with the predominantly luminal (i.e. ER+) phenotype observed in ILC. Further understanding of the lineage specific effects of E-cadherin/CDH1 loss can elucidate the unique role of E-cadherin as a tumor suppressor, and identify key features, including ER, that may mediate progression of precancerous lesions to ILC specifically in the context of E-cadherin loss. Date of Presentation October 16, 2024