The most important prognostic factor for oral squamous cell carcinoma (SCC) is the presence of lymph nodes metastasis, and the five-year survival rate of a patient with lymph node metastases drops to approximately 20%. Loss of the cell-cell adhesion protein E-cadherin is linked to tumor cell invasion, metastasis and poor prognosis in oral SCC. Therefore it is important to understand the regulation of this protein during tumor progression.

E-cadherin negative tumors often give rise to E-cadherin positive metastases, indicating a plasticity in the expression of E-cadherin throughout tumor progression. Since the different stages of tumor progression are linked to changes in the tumor microenvironment, it is possible that the tumor microenvironment may act to regulate the mechanisms of E-cadherin silencing, thereby influencing the plasticity of E-cadherin expression seen during tumor progression. I hypothesized that the transient expression of E-cadherin during tumor progression may be linked to reversible epigenetic regulation of the gene that is controlled by the microenvironment.

To test this hypothesis, oral SCC cells that were E-cadherin-deficient, due to methylation-mediated silencing, were grown as 3D tissues in a previously established model of stratified squamous epithelium. The formation of homotypic cell-cell interactions in the context of a 3D microenvironment resulted in the de-methylation of the E-cadherin promoter and re-expression of the protein. This induction of E-cadherin expression was found to be influenced by a low level of DNA methylation combined with histone H3K4 and H3K27 methylation. The mechanism of reversal was determined to involve the microenvironmental regulation of DNMT1 expression which inversely correlated with E-cadherin expression.

This data supports previous working indicating that H3K4 and H3K27 methylation is linked to a epigenetically plastic state of gene regulation in embryonic stem cells. This information combined with our findings suggests that an “epigenetic signature” of gene regulation may be identified to determine the strength of gene silencing during tumor progression. It also provides further evidence for a synergistic relationship between DNMTs and histone modification proteins in gene regulation.