The purpose of this study was to investigate the association among matrix metalloproteinases (gelatinases A and B, stromelysin-3 (ST3) and matrilysin) mRNAs expressed in primary breast carcinomas and standard prognostic parameters and clinical outcome. mRNA levels were determined by Northern analysis in samples of 81 breast cancer patients (median follow-up, 40 months) and 27 samples of uninvolved adjacent breast tissue. Proteases were expressed by the majority of the tumors and normal breast tissues examined. ST3, gelatinase A and matrilysin mRNAs were more often expressed at high levels in carcinomatous than in normal breast tissues. Differences in the distribution of gelatinase B mRNA were not found. However, paired normal tissues generally produced weaker signals when compared to matched tumor samples. Univariate analysis showed no significant association of gelatinase A and matrilysin mRNAs with the classical prognostic markers (age, menopausal status, stage, size, nodal status, vascular infiltrate, necrosis, steroid receptors, metastasis and survival). Overexpression of ST3 was more frequently found in tumors of post-menopausal women (P < 0.022). Elevated expression of gel B mRNA was associated with the presence of vascular infiltrate (P < 0.026), necrosis (P < 0.039), PR negative tumors (P < 0.014) and inversely correlated to the number of survivors (P < 0.021). Multivariate analysis including 68 patients for whom all information was available indicated that neither stromelysin correlated significantly with pathological, clinical or biochemical features. High levels of gelatinase A and B mRNAs were inversely associated with the number of survivors. Our findings suggest that measurements of gelatinase A and B mRNAs expression in breast carcinoma may help to identify patients with an aggressive form of the disease.

Using Northern blot analysis we have measured the co-expression of the matrix metalloprotease MMP-9, plasminogen activator urokinase type (uPA) and its receptor (uPAR) mRNAs in 81 biopsies of breast carcinomas with the objective of analyzing the impact of these factors on the overall survival probability of the patients (median follow-up time: 4 years). Individual mRNA levels of either uPA or uPAR showed parallel variations with MMP-9 mRNA, suggesting a coordinate transcription of these markers. When median values were used as cutoff points to discriminate between high and low factor expression, no association was found with patient outcome and MMP-9 or uPA mRNA distribution. However, increased uPAR mRNA levels were associated with poor prognosis (p = 0.01). The combination of MMP-9 and uPAR mRNA measurements has not enhanced prognostic information compared to information supplied by the receptor alone (p = 0.01). The combination of MMP-9 and high levels of uPA mRNA led to a significant association with poor outcome (p = 0.03). Multivariate analysis supported the notion that increased uPAR mRNA production in primary breast cancer may be a predictor of overall survival.