Evaluation of urinary plasminogen activator, its receptor, matrix metalloproteinase-9, and von Willebrand factor in pancreatic cancer.

PURPOSE Pancreatic cancer remains a devastating problem with the majority of patients succumbing to death from this disease. A hallmark of pancreatic cancer is the loss of basement membrane that may be attributed to the action of urinary plasminogen activator (uPA) and matrix metalloproteinase-9 (MMP-9). These enzymes are also implicated in angiogenesis. uPA and microvessel density have been shown to be good prognostic indicators for breast and colon cancer. MMP-9 and microvessel density have not been investigated in pancreatic cancer. We have therefore investigated by immunohistochemistry: (a) frequency of uPA expression and its receptor uPAR and the site of synthesis of uPA by in situ hybridization (ISH); (b) MMP-9 and its coexpression with uPA; (c) microvessel density as determined by von Willebrand factor staining and its relationship to uPA and MMP-9 expression; and (d) correlation of these parameters with survival. EXPERIMENTAL DESIGN Archival paraffin sections of 27 pancreatic tumors were semiquantitatively investigated by immunohistochemistry using the following antibodies: (a) monoclonal antibodies (MAbs) uPA(1) and uPA(2) (3689 and 394, respectively); (b) MAb uPAR, (no. 3932); (c) MAb MMP-9 (no. 936); and (d) rabbit anti-F8RA/vWF. ISH was performed using a uPA cDNA. RESULTS Both uPA antibodies revealed overexpression of uPA (93%) often with uniform staining of tumor cells. uPAR and MMP-9 showed focal staining in only 52 and 37% of tumors, respectively. Morphologically normal appearing ductal cells in close proximity to tumors overexpressed uPA in contrast to distally located normal cells (P = <0.001). uPA staining was also investigated in pancreatic intraepithelial neoplasia (PanIN) lesions. PanIN 1A/B staining for uPA was seen in 8 cases (30%), that for PanIN 2 in 19 cases (70%), and for PanIN3 in 12 cases (44%). Lumen of microvessels in the tumor stroma also revealed staining of uPA in 10 cases (37%). ISH experiments revealed the presence of uPA mRNA not only in the cytoplasm of tumor cells but also in adjacent normal appearing ducts as well as in PanIN lesions. Patients with overexpression of uPA, uPAR, or MMP-9 had a trend toward poorer survival than those who did not express it. Microvessel density did not show any significant relationship with uPA, uPAR, and MMP-9 expression and survival. CONCLUSIONS We conclude that uPA and MMP-9 are potential prognostic indicators in pancreatic cancer, whereas microvessel density may not be one. This study confirms our previous observation that uPA is made by the tumor cells themselves. Presence of uPA in vessels of tumor stroma suggests that uPA is in circulation, and its measurement and that of MMP-9 in the blood of these patients may aid in prognosis. Patients showing overexpression of uPA and MMP-9 have a trend toward shorter survival time.