1453

Progression to oral cancer takes 10-20 years and the process occurs within 3-dimensional tissue architecture. The microenvironment of a lesion includes the spatial organization of cells (normal/progressing/stromal) coupled with interactions to their surroundings. We hypothesize that the microenvironment is important to and underlies carcinogenic progression. Epithelial cells rest on a basement membrane composed of extracellular matrix material (ECM); the biology of epithelial cells is known to be dependent on ECM and a lesion’s microenvironment includes ECM. To begin to model events that occur within a progressing lesion, we have characterized the effect of ECMs on integrin expression in a model for oral cancer progression. Primary cultures of normal oral epithelial (NOE) cells and a cell line from dysplastic leukoplakia (premalignant oral epithelial cell, POE) were cultured for 48 hrs on five surfaces (plastic, collagen 1, collagen 4, laminin, fibronectin and ECM-Gel) to generate steady-state conditions. Integrin expression was initially characterized by PCR and then cell surface expression quantified by flow cytometry. For integrins that were expressed by PCR, commercially available antibodies suitable for flow cytometry were available for α1, α2, α3, α5, α6, αv, β1, β2, and β3. With PCR, there was only one change greater than 2-fold and flow confirmed a down regulation of α2 on POE on laminin. Flow analyses demonstrated that there were reproducible modulations in cell surface expression induced by growth on ECMs (comparison of mean fluorescent intensity [MFI], MFI-ECM:MFI-plastic). In general, relative expressions varied up to 2-fold and recent data on HLA and T-cell receptor expression has shown that small changes in MFI, such as we have found, can have biological significance. In comparison to POE, NOE cells were more responsive to ECM-induced changes in expression. Broadly, expression was found to be both ECM and cell line dependent. Changes in integrin expression occurred both when cultured on appropriate ligands but also were found in an ECM/ligand independent manner. α3 and α6 are part of laminin receptors and were upregulated in both POE and NOE when grown on laminin. However, α6 expression, was increased on NOE cells cultured on each ECM, yet α6 is not part of the receptor for each ECM. Although α2 is part of a laminin receptor, its decreased expression on POE grown on laminin suggests dysregulation of this laminin receptor. In general, our results suggest that integrin expression on normal and premalignant oral cells can be modulated by the microenvironment on which they reside. If extrapolated to in vivo, a progressing lesion can be viewed as a dynamic entity; epithelial and stromal compartments may both synthesize and respond differently to products, such as ECMs, as a lesion progresses (Supported in part by NIH grant DE14395).

[Proc Amer Assoc Cancer Res, Volume 45, 2004]