The prognostic value of E-cadherin and the cadherin-associated molecules alpha-, beta-, gamma-catenin and p120ctn in prostate cancer specific survival: a long-term follow-up study.

To determine the value of loss of expression of E-cadherin and cadherin associated molecules as prognostic markers for prostate cancer patients in a long-term follow-up study. Sixty-five prostate cancer specimens, obtained from patients with different stages of prostate cancer who underwent a radical prostatectomy or TUR-P between 1987 and 1991, were used for immunohistochemical analysis of the expression pattern of E-cadherin, alpha-, beta-, gamma-catenin and p120(ctn). Clinical records of these patients were studied for follow-up data and the prognostic value of expression of these adhesion molecules was determined by Kaplan-Meier survival analyses and multivariable proportional hazard regression analysis. Normal staining patterns were found in 36 cases (55.4%) for E-cadherin, 37 cases (56.9%) for alpha-catenin, 40 cases (61.5%) for beta-catenin, 25 cases (38.5%) for gamma-catenin, and 40 cases (61.5%) for p120(ctn). Overall, a strong correlation was found between the expression of E-cadherin and other cadherin-associated molecules. The 5-year survival rates for each staining were as follows: E-cadherin (normal 79.2%, aberrant 26.8%), alpha-catenin (normal 79.2%, aberrant 26.8%), beta-catenin (normal 73.1%, aberrant 27.3%), gamma-catenin (normal 86.4%, aberrant 37.1%), and p120(ctn) (normal 72.8%, aberrant 30.0%). There was a significant difference in survival between normal and aberrant expression in each staining (log rank P < 0.0001). The proportional hazard regression model including tumor stage and Gleason score revealed alpha-catenin expression as the best prognostic marker for patients with prostate cancer. Our data revealed a strong correlation between E-cadherin expression and other cadherin-associated molecules. Among these markers, alpha-catenin seems the best prognostic marker for prostate cancer specific survival. Larger studies are needed to confirm this result.