Integrin expression in non-small cell carcinoma of the lung.

Alteration of integrin expression in a number of different malignant diseases has been recognized, with a trend of downregulation of collagen-laminin binding integrin expression in epithelial tumor types noted. This study evaluated the expression of a panel of integrin subunits that included subunits that form receptors that bind to collagen and laminin (alpha 2, alpha 3, alpha 6 beta 4) and subunits that form receptors that bind to fibronectin and fibrinogen (alpha 5, alpha V, beta 3, beta 6) in 51 specimens of non-small cell carcinoma (NSCCA) of the lung by use of immunohistochemistry. Integrin expression was then correlated with histologic type (squamous vs. adenocarcinoma), absence or presence of hilar or mediastinal nodal metastasis at resection, and cellular differentiation (well or poorly differentiated). In general, downregulation of the collagen-laminin binding subunits was noted in tumor cells of the NSCCA specimens when compared to the progenitor normal bronchial epithelium. No differences were noted in integrin expression between squamous cell and adenocarcinoma or between node-positive or node-negative tumors. However, downregulation of the integrin subunit alpha 3 was noted to be significantly more common in poorly differentiated tumors (p = 0.02) and several of the other collagen-laminin binding subunits also appeared to be more downregulated in poorly differentiated tumors. No upregulation was seen in the alpha 5 subunit of the fibronectin receptor or the beta 3 subunit of the vitronectin receptor, however, approximately 50% of tumors showed upregulation of the beta 6 subunit, the great majority of these being well-differentiated, node-negative tumors. Downregulation of the collagen-laminin integrins may thus be associated with differentiation of NSCCA, but not metastasis, and may serve as an adjunctive prognostic marker of disease. The beta 6 subunit appears to be associated with malignant transformation, but may serve as a positive prognostic factor.