UROKINASE-TYPE PLASMINOGEN-ACTIVATOR AND ITS INHIBITOR PAI-1 - PREDICTORS OF POOR RESPONSE TO TAMOXIFEN THERAPY IN RECURRENT BREAST-CANCER

Background: Urokinase-type plasminogen activator (uPA) is a proteolyti c enzyme thought to be involved in processes leading to tumor cell inv asion of surrounding tissues. Its activity during metastasis may be re gulated by an inhibitor, PAI-1. Previous work has shown that high leve ls of uPA and PAI-1 are associated with poor prognosis in primary brea st cancers. Purpose: In this pilot study, we explored possible associa tions between the expression levels of uPA or PAI-1 and the efficacy o f tamoxifen treatment in breast cancer patients with relapsed disease. Methods: Levels of uPA, PAI-1, estrogen receptor (ER), and progestero ne receptor (PgR) were assayed in cytosolic extracts derived from the primary breast tumors of 235 tamoxifen-naive patients who had recurren t disease. The extracts were classified as positive or negative for ea ch assayed factor. In some analyses, ER and PgR levels were evaluated together. In these analyses, three ER/PgR subsets were defined: low, i ntermediate, and high. All patients in the study received tamoxifen th erapy upon relapse (median follow-up, 57 months). Association of the t ested factors with the response to tamoxifen treatment was studied by logistic regression analysis. Association of the factors with progress ion-free and overall survival was further evaluated by Cox univariate and multivariate regression analyses. Results: Patients with uPA-negat ive tumors exhibited a better response (tumor regression or stable dis ease, maintained for more than 6 months) to tamoxifen treatment than t hose with uPA-positive tumors (51% versus 26% response; odds ratio [OR ] = 0.34; 95% confidence interval [CI] = 0.18-0.65). The response rate was also better for patients with PAI-1-negative tumors than for thos e with PAI-1-positive tumors (49% versus 35% response; OR = 0.57; 95% CI = 0.32-1.01). In addition, patients with uPA-positive or PAI-l-posi tive tumors showed shorter progression-free survival (P = .001 and P<. 05, respectively) and total survival after relapse (P = .005 and P<.00 5, respectively). When patients were stratified by ER/PgR status, the only statistically significant association between uPA levels and redu ced tamoxifen response was seen in the subset whose tumors possessed i ntermediate levels of ER/PgR (16% response in uPa-positive versus 60% response in uPA-negative tumors; OR = 0.13; 95% CI = 0.04-0.41). Overa ll, uPA status appeared independent of association with ER/PgR status in its ability to predict response to tamoxifen treatment. The associa tion of PAI-1 expression and the response to tamoxifen was less pronou nced when patients were stratified by ER/PgR status. Conclusion: Measu rement of primary breast tumor uPA levels may be useful in predicting the overall response of metastatic disease to tamoxifen therapy.