The plasminogen activation system plays a role in cancer progression via extracellular matrix degradation and tumor cell migration . Numerous research groups have demonstrated that the antigen content of uPA and PAI-1 in primary breast cancer tissue correlates with disease aggressiveness and has a strong prognostic impact on primary breast cancer [1–6, 21, 22]. All these data have been obtained at the protein level using quantitative enzyme-linked immunosorbent assay (ELISA)-based methods applied to cytosol. The EORTC Receptor and Biomarker Study Group initiated quality control programs for this type of assay . However, the use of this method encounters major limitations in routine clinical practice. Although being robust, reproducible and quality-assured, the ELISA requires a substantial amount of frozen tissue, which compromises its use in small tumors (< 1 cm) and requires adequate logistics for the storage of frozen tumor samples. Furthermore, the biochemical methods have been abandoned to the detriment of immunohistochemical methods for hormone receptor assessment in an increasing number of pathology laboratories. Unfortunately, immunohistochemical assays of the uPA/PAI-1 system have provided unsatisfactory results mainly due to an absence of consensus regarding uPA and PAI-1 cellular localization . The measurement of uPA - and PAI-1 -expression at the mRNA level using molecular biology techniques could thus constitute an alternative to the immunochemical assays currently being used. Indeed, real-time quantitative NASBA or RT-PCR needs very low quantities of material (< 100 ng of total RNA) and thus allows for the analysis of smaller tumors. Moreover, the availability of aqueous tissue storage reagents that rapidly permeate tissue to stabilize and protect cellular RNA in unfrozen specimens will facilitate routine laboratory use of those techniques.
Previous studies focused on messengers of the components of the plasminogen activation system in human breast cancer by mainly comparing normal, benign and malignant breast tissues or by examining the cellular localization of uPA and PAI-1 [25–30]. More recently, Castello et al.  developed a quantitative real-time RT-PCR assay and showed that uPA and PAI-1 mRNA expression increased with tumor severity in breast cancer, thereby confirming previous results obtained by Northern blotting . Moreover, it has been demonstrated that high uPA and PAI-1 mRNA expression was significantly associated with shorter disease-free survival in a population of 130 primary breast cancers independent of the hormone receptor and the lymph node status .
To our knowledge, we are the first to evaluate the prognostic impact of uPA and PAI-1 at the mRNA level in a specific group of lymph node positive- and hormone receptor-positive breast cancers. In the present study, we show that PAI-1 mRNA, as measured by quantitative RT-PCR in the primary tumors, has the strongest prognostic value for MFS and BCS. We also observed that the prognostic value of PAI-1 is stronger that of uPA. Increased uPA messenger level was also associated with metastasis-free survival and breast cancer specific survival in univariate analysis, but did not represent a statistically significant independent prognostic factor. The fact that histological grade, one of the most important prognostic factors in hormone receptor-positive breast cancer did not emerge as an independent factor in our analysis was not expected. A major reason might be that we selected a specific population of hormone receptor- and node-positive breast cancer with mainly large tumors. This selection bias might, besides the quite small number of patients, explain these results.
Our results are in accordance with previous studies showing that PAI-1 protein displayed stronger prognostic impact than uPA in lymph node-positive patients and that this marker remained a strong prognostic factor after long-term follow-up both for primary breast cancer and after the first relapse [4, 34]. Moreover, it has recently been shown that PAI-1 mRNA expression increased with colorectal cancer and the oesophageal squamous carcinoma stage and was associated with poor prognosis suggesting that the gene expression of this marker may serve as a new prognostic factor in these two types of cancer [35, 36].