Epithelial-Mesenchymal Transition Opens a Window of Toxicant Sensitivity

Abstract

BEAS-2B, an immortalized but non-malignant bronchial epithelial cell line, has been used extensively to model toxicant-induced carcinogenesis. There are currently inconsistencies in the literature as to which culture media is most appropriate for BEAS-2B, which has profound impacts on cellular phenotype. Previous work has suggested that culture media containing fetal bovine serum (FBS) results in a loss of epithelial identity. In the current study, we identify that culture media containing fetal bovine serum (FBS) induces epithelial-mesenchymal transition (EMT), a fundamental cellular process involved in embryonic development, maintenance of tissue extracellular matrix (ECM), and wound healing.. We show that FBS exposure decreases cellular epithelial markers and increases mesenchymal markers at the mRNA and protein levels. These changes begin to reverse following removal of FBS from the culture media. Additionally, exposure to FBS results in nuclear accumulation of transcription factors known to cause global gene expression changes driving EMT. We show that FBS also results in global gene expression patterns consistent with nuclear accumulation of these EMT-driving transcription factors. We also demonstrate similarity between FBS exposure and canonical activation of EMT by transforming growth factor-β1 (TGF-β1). At the cellular level, FBS-exposure also results in morphological and functional changes associated with EMT. These include an elongated spindle shape, increased cell invasiveness and anchorage-independent growth. Utilizing size-excluded media, we identify the EMT-inducing molecule or molecules within FBS to be larger than 30kD. Switching to a model of canonical EMT activation by TGF-β1, we were able to show that previously identified arsenite sensitivity in BEAS-2B in the presence of FBS is due to the induction of EMT. We showed that BEAS-2B and A549 cells stimulated to undergo EMT with TGF-β1 have deceased levels of cellular glutathione (GSH) and the catalytic subunit of glutamate-cysteine ligase (GCLC), the rate limiting enzyme in GSH synthesis. Both of these observations were reversible by removal of TGF-β1 as a stimulus. The reduction of GSH and GCLC resulted in sensitivity in both cell lines to arsenite and arsenic trioxide, which could be rescued by removal of TGF-β1 or supplementation with N-acetylcysteine, a precursor of GSH synthesis. This suggests the sensitivity is specific to the reduction of GSH following activation of EMT. We also show that EMT-stimulated BEAS-2B show greater markers of apoptosis and DNA double-strand breaks than unstimulated BEAS-2B at the same dose of toxicant. Ultimately, these studies show that the BEAS-2B cell line does not behave like an epithelial cell following culture in FBS. The culture conditions in which BEAS-2B are maintained should be scrutinized when interpreting published data in the literature. Also, these studies identify epithelial cells that have undergone EMT as a toxicant sensitive population, susceptible to both DNA damage and cell death following toxicant exposure. The cell type of origin for most lung cancers currently remains unknown. Our studies suggest that epithelial cells that have undergone EMT may be a candidate cell population for carcinogenesis.