Radiol Oncol 2003; 37(1): 29-35. The urokinase plasminogen activator and its inhibitors PAI-1 and PAI-2 in primary cutaneous melanoma Jasmina Markovič1, Borut Štabuc2 1Clinical Department of Anaesthesiology and Intensive Therapy, 2Clinical Department of Gastroenterology, University Clinical Centre Ljubljana, Ljubljana, Slovenia Background. We investigated the differences in urokinase plasminogen activator (uPA) and its inhibitors type1 and 2 (PAI-1/2) concentrations in clinically suspected nevi, primary cutaneous melanoma and normal skin and correlations with histopathological prognostic factors of primary melanoma. Patients and methods. Fifty-one patients were enrolled. The tissue concentrations of uPA, PAI-1 and PAI-2 were quantified by enzyme-linked immunosorbent assay (ELISA). Results. Mean uPA and PAI-1 concentrations in melanomas were higher than in normal surrounding skin (uPA: 1.08; vs. 0.48 ng/mgp; PAI-1: 14.07 vs. 2.07 ng/mgp; p < 0.001). uPA and PAI-1 concentrations were higher in melanomas than in nevi, and higher in nevi than in normal surrounding skin (uPA: p > 0.05; PAI-1: p = 0.02). PAI-2 concentration was higher in normal surrounding skin than in nevi and melanomas (p > 0.05). Melanoma uPA, PAI-1 and PAI-2 concentrations correlated significantly with normal skin (r = 0.73, 0.54, 0.38 respectively). PAI-1 was significantly lower in melanomas of Breslow thickness L 0.75 mm, Clark invasion of 0+I, without microscopic ulceration, without vascular invasion (p < 0.01) than in melanomas of Breslow thickness > 0.75 mm, Clark invasion > II, with ulceration and vascular invasion. Conclusions. Determination of uPA and PAI-1 can provide significant additional prognostic information for melanoma patients. Key words: skin neoplasms - melanoma; urokinase plasminogen activator; plasminogen activator inhibitor 1; plasminogen activator inhibitor 2; prognosis Introduction Received 24 January 2003 Accepted 14 February 2003 Correspondence: Jasmina Markovič, M.D., M.Sc., University Clinical Centre Ljubljana, Clinical Department of Anaesthesiology and Intensive Therapy, Zaloška 7, 1000 Ljubljana, Slovenia; Phone: +386 (0) 31626577; E-mail: jasmina.markovic1@kclj.si Malignant melanoma is one of the most aggressive human tumours. The prognosis of melanoma as well as other cancers is mainly dependent on its ability to invade and metas-tasise. Prognostic markers such as Breslow tumor thickness, Clark level of invasion, ul-ceration and vascular invasion are used for defining melanoma patients at risk and fol- 30 Markovič J et al. / uPA, PAI-1 and PAI-2 in melanoma lowing adjuvant therapy. The clinical impor-tance of other factors (clinical, biochemical-molecular) has not been established yet.1 Cancer invasion and metastasis are multistep events involving local invasion of the extra cellular matrix, angiogenesis, invasion of the blood vessel wall, survival of malignant cells in the vascular system, extravasation, and establishment of a secondary growth. During most of these steps, extra cellular matrix and basement membrane have to be de-graded.2 The breakdown of these barriers is catalysed by the proteolytic enzymes, which are released from the invading tumour. Four different proteases are mainly involved: met-alloproteinases (collagenases, stromelysin, gelatinases), cysteine proteinases (cathepsin B, H, L), aspartyl proteases (cathepsin D), ser-ine proteases (plasminogen activation sys-tem).3 The two known plasminogen activators (PAs) are tissue type (tPA) and urokinase type plasminogen activator (uPA). Proteolytic activity of tPA is important for the degradation of intravascular blood clots, while uPA contributes to extracellular proteolysis in a wide variety of physiological and pathological processes.2-4 uPA has many activities. It con-verts inactive plasminogen to plasmin, which degrades most substrates in the extracellular matrix (proteoglycans, laminin, fibronectin, vitronectin), and activates other proteases (procollagenases, uPA). uPA is activated also by kallikrein, trypsin, cathepsin B, L, ther-molysin and nerve growth factor.3,4 Beside these degradative functions uPA exerts other activities that may enable it to play a role in invasion and metastasis. These include stimulation of cellular proliferation, enhancement of cellular migration, alteration of cellular ad-hesive properties and activation of specific growth factors.3-5 Active uPA binds to a mem-brane-bound receptor known as u-PAR. uPA activity is controlled by two inhibitors, plas-minogen activator inhibitor type - 1 (PAI-1) and type - 2 (PAI-2). In addition to inhibiting Radiol Oncol 2003; 37(1): 29-35. uPA, PAI-1 modulates cellular adhesion and migration by its attachment to extracellular protein vitronectin. PAI-2 is important in inhibition of apoptosis.3,4 A strong impact of these proteolytic factors on prognosis of the cancer disease has been observed in a variety of malignancies. The strongest and most con-sistent evidence of a prognostic role exists with breast cancer. Elevated concentrations of uPA, PAI-1 and uPAR are associated with poor prognosis, high levels of PAI-2, on the other hand, correlate with good outcome.2,5 uPA is prognostic also in gastric, colorectal, oesophageal, renal, endometrial, and ovarian cancer.2,4,5 It is well known, that uPA, PAI-1 and PAI-2 expression in human melanoma cell lines correlated with a high metastatic ca-pacity in nude mice.6 However in human cu-taneous and uveal melanoma uPA, PAI-1 and PAI-2 had not been detected in early stage of melanoma but appeared frequently in advanced primary melanoma and melanoma metastatic lesions.7,8 Our study was aimed to find out the dif-ferences of uPA, PAI-1 and PAI-2 concentra-tions in clinically suspected nevi, primary melanoma and normal skin concentrations and their correlation to the most important histopathological prognostic factors: Breslow thickness, Clark invasion, ulceration and vas-cular invasion. Patients and methods Patients Fifty-one patients with clinical confirmed pri-mary cutaneous melanoma (27 women: mean age 49.7; range 21-84 years and 24 men: mean age 56.5; range 17-83 years) were enrolled in-to a prospective study between 1998-2000. Inclusion criteria were: absence of metastases and macroscopically and histological com-plete surgical removal of the primary cuta-neous melanoma (UICC pT1or T2N0M0, AJCC stage I and II).9 The Medical Ethics Markovič J et al. / uPA, PAI-1 and PAI-2 in melanoma 31 Committee at the Ministry of Health of the Republic of Slovenia approved the study pro-tocol. We totally excised melanoma lesions with safety edge of 1-2 cm. We excised 2 x 2 x 2 mm tissue specimens of the lesions and of the normal skin (at least 2 cm far away from the edge of tumours) for the quantification of uPA and PAI-1/2. They were snap-frozen in liquid nitrogen and stored at - 80°C. The remaining tissue was fixed in 10% for-malin and embedded in paraffin for histolog-ical examination. The histological diagnosis in 8 patients was dysplastic nevus and in 43 primary cutaneous malignant melanoma with clinical stage I (T1-2 N0 M0) less than 1.5 mm thick. The clinical and histopathological char-acteristics of primary tumours are shown in Table 1. Tissue extraction and ELISA for uPA, PAI-1 and PAI-2 The uPA concentrations were determined in 41 pairs of triton extracts, and the PAI-1 and PAI-2 concentrations in 51 pairs of cytosols prepared from tumour and adjacent normal tissue samples (matched pairs) weighing 50 mg, obtained at surgery. The still frozen cut sections were dipped into liquid nitrogen and then pulverized in a microdismembrator (Braun - Melsungen, Melsungen, Germany). For the triton extracts the still frozen pulver was dispensed with Tris buffered saline (TBS) (0.02 M Tris-HCl, 0.125 M NaCl, pH 8.5) containing 1% non-ionic detergent Triton X-100 (Sigma, St. Louis, Missouri, U.S.A.). The suspension was gently shaken for 3 hours at 4° C. For the cytosol the still frozen pulver was dispensed in a phosphate buffer (5 mM Na2HPO4, 1.7 mM KH2PO4, 1 mM monothioglycerol, 10% (vv-) glycerol, pH 7.4). Both, the tissue extracts and cytosol suspension were subjected to ultracentrifugation (100 000 g/45 min at 4° C) to separate tissue debris. Supernatants were collected, divided into aliquots and stored at - 70°C until use. uPA, PAI-1 and PAI-2 concentrations were determined by commercially available ELISA kits (American Diagnostica, Inc., Greenwich, U.S.A.) for uPA, PAI-1 and PAI-2. Details of the kits are described elsewhere.10 Levels of the uPA, PAI-1 and PAI-2 are expressed in ng/mg proteins. Protein content was deter-mined with Bio Rad method. Statistical analysis Statistical analysis was performed using the SPSS for Windows program. Differences of uPA, PAI-1 and PAI-2 concentrations in melanomas, nevi and normal skin were analysed by the Wilcoxon test. Spearman’s Table 1. Clinical and histopathological characteristics of the tumours Characteristics Number Dysplastic nevus 8 Melanoma 43 Localization of primary lesion Head 4 Trunk 34 Extremities 13 Histopathological characteristics of N melanomas Breslow L 0,75 28 > 0,75 15 Clark 0+I 12 II + III 31 Ulceration Yes 7 No 36 Vascular invasion Yes 3 No 20 Undetermined 20 Histopathological type Lentigo maligna 2 Superficial spreading 35 Nodular 2 Unclassified 4 Radiol Oncol 2003; 37(1): 29-35. 32 Markovič J et al. / uPA, PAI-1 and PAI-2 in melanoma rank correlations were evaluated for the rela-tions between uPA, PAI-1 and PAI-2 concen-trations in melanomas and normal skin. Differences of uPA, PAI-1 and PAI-2 concen-trations in melanomas and histomorphologi-cal variables among various groups of pa-tients were analysed by the two-tailed t-test and the analysis of variance (ANOVA). The p-values L 0.05 were considered sig-nificant. Results uPA, PAI-1 and PAI-2 concentrations in melanomas, nevi and normal skin Among 51 tumour specimens (43 melanomas, 8 dysplastic nevi) and normal skin, the con-centration of u-PA was determined in 41 pairs of triton extracts (36 melanomas, 5 nevi), and of PAI-1 and PAI-2 in 51 pairs of cytosols (43 melanomas, 8 nevi) (Table 2). The mean uPA and PAI-1 concentrations in melanomas were significantly higher than in normal skin (p < 0.0001). The mean uPA con-centration was higher in melanomas than in dysplastic nevi and higher in dysplastic nevi than in normal skin, however statistically in-significant (p > 0.05). Mean PAI-1 concentra-tion was higher in melanomas than in dys-plastic nevi, and higher in dysplastic nevi than in normal skin with significant results Table 2. Differences of uPA, PAI-1 and PAI-2 concentrations between melanomas, nevi and normal skin uPA (ng/mgp) PAI-1 (ng/mgp) PAI-2 (ng/mgp) Melanoma 1,08 ± 0,58 14,07 ± 16,55*** 9,21 ± 18,32 Dysplastic nevi 0,99 ± 0,58 4,84 ± 8,32**** 13,44 ± 17,32 Normal skin around 0,48 ± 0,21* 2.07 ± 1,83** 13,05 ± 19,72 Melanoma Normal skin around 0,51 ± 0,36 0,89 ± 0,91 28,8 ± 39,55 Dysplastic nevi Levels are mean ± sd *, ** p < 0,0001, statistically significant lower uPA and PAI-1 normal skin levels according to melanoma levels ***, **** p = 0,021 statistically significant higher PAI-1 melanoma levels according to nevi and higher nevi levels according to normal skin levels Radiol Oncol 2003; 37(1): 29-35. (p = 0.02). Mean PAI-2 concentration was not significantly higher in normal skin than in dysplastic nevi and melanomas (p > 0.05). There was a significant positive correlation (rs = 0.45) between melanoma uPA and PAI-1 concentrations. Significant correlations be-tween melanoma and normal skin concentra-tions for uPA, PAI-1 and PAI-2 were found (rs = 0.73; 0.54; 0.38) (Figure 1). Figure 1. Correlation between melanoma and normal skin uPA concentrations in ng/mg proteins (n = 36, rs = 0,73, p < 0,0001). Markovič J et al. / uPA, PAI-1 and PAI-2 in melanoma 33 Correlation between melanoma uPA, PAI-1 and PAI-2 concentrations and relevant prognostic factors Melanoma uPA, PAI-1 and PAI-2 concentra-tions were compared to established histomor-phological factors (Table 3). In contrast to uPA and PAI-2, PAI-1 concentrations were significantly lower in melanomas of Breslow tumour thickness L 0.75 mm, Clark level of invasion 0+I, without microscopic ulceration on the tumour surface and absents vascular invasion of the tumours. Much higher positive correlations (rs = 0.67, 0.62 and 0.63 respectively; p L 0.02) be-tween u-PA and PAI-1 concentrations were found in the group of the melanomas of Breslow thickness > 0.75 mm, with the level of Clark invasion III and present vascular invasion. Discussion In the present study using specific ELISA per-formed on tumour extracts of primary melanomas less than 1.5 mm thick, higher uPA and PAI-1 and lower PAI-2 concentra-tions than in nevi and normal skin were found. The correlation between uPA, PAI-1 and PAI-2 concentrations in melanomas and normal skin was found to be positive. The prognostic value of uPA and PAI-1 in early stages primary malignant melanoma has not yet been investigated. However, a strong correlation between the metastases of the human melanoma cells in the nude mouse model as well as in the human cuta-neous melanoma and expression of uPA and PAI-1 was found.6 Dysplastic nevi are sup-posed to be an important risk factor for the development of the cutaneous melanoma. In 1979 Fräki et al. found out higher concentra-tions of plasminogen activators in primary melanomas and melanoma metastases com-paring to extracts of nevi.11 Recently De Vries et al reported that uPA, PAI-1, PAI-2 and uPAR appeared frequently in advanced pri-mary cutaneous and uveal melanoma and melanoma metastasis lesions.7,8,12 However uPA and PAI-1 accumulation was observed also in atypical nevocytes, whereas uPA pro-teolytic activity was detected only in melanomas.13 Table 3. Differences of mean concentrations of uPA, PAI-1 and PAI-2 in primary cutaneous melanomas with his-tomorphological prognostic factors Variable N uPA± SD (ng/mgp) pa N PAI-1± SD (ng/mgp) pa N PAI-2± SD (ng/mgp) pa Breslow L 0,75 mm > 0,75 mm 22 14 1,01± 0,54 1,2± 0,65 n.s. 28 15 9,68± 11,92 22,25± 20,92 0,016 28 15 10,89± 22,34 6,07± 5,35 n.s. Clark 0+I II+III+IV 9 27 0,89± 0,55 1,15± 0,59 n.s. 12 31 4,74± 6,45 17,67± 17,89 0,001 12 31 15,97± 32,36 6,59± 7,78 n.s. Ulceration Yes No 7 29 1,38± 0,74 1,01± 0,53 n.s. 7 36 27,76± 23,27 11,41± 13,82 0,015 7 36 3,2± 2,68 10,38± 19,82 n.s. Vascular invasion Yes No Undetermined 2 15 19 1,99± 1,04 1,02± 0,55 1,04± 0,52 n.s. 3 20 20 37,77± 23,01 9,5± 12,74 15,1± 16,6 0,017 3 20 20 5,13 ± 6,74 11,48± 25,62 7,55± 8,33 n.s. aAnalysis of variance, two-tailed t-test, p L 0,05; n.s. = not significant Radiol Oncol 2003; 37(1): 29-35. 34 Markovič J et al. / uPA, PAI-1 and PAI-2 in melanoma Previously, uPA and PAI-1 have been claimed to be of independent prognostic val-ue for disease free and overall survival in breast cancer patients.14,15 In addition to breast cancer, components of plasminogen activation system also have a prognostic val-ue in colorectal, gastric, oesophageal, blad-der, endometrian and ovarian cancer.2,5 Similarly Nekarda et al have found a stronger prognostic impact of PAI-1 than that of uPA in completely resected gastric cancer.16 The presence of the components of the PAs in malignant melanoma has earlier been studied using immunohistochemistry (IHC), in situ zymography and in situ hybridisation. Both, ELISA and IHC have their specific ad-vantages. ELISA methods give an objective quantification of analyte levels, whereas IHC yields at best semi-quantitative information. ELISA and IHC may detect fractions of PA components with different efficiencies.17,18 At present, ELISAs measuring the levels of uPA and PAI-1 performed consistently well, at least by their more extensively proven clin-ical value and unequivocal interpretation as demonstrated by the Quality Assurance Center in Nijmegen and from the European Organisation for Research and Treatment of Cancer (EORTC).10 Breslow tumour thickness, Clark level of invasion, ulceration and vascular invasion are the most important histological variables pre-dicting melanoma outcome. The prognostic role of other possible markers like molecular and biochemical is not known yet. Our re-sults prove the presence of the correlation of uPA and PAI-1 concentrations with prognos-tic value of Breslow thickness, Clark invasion, ulceration and vascular invasion also in early melanoma stage I. We observed that higher PAI-1 concentrations were associated with melanomas of Breslow > 0.75 mm, Clark II+III+IV, present ulceration and vascular invasion, which are prognostic adversely. Significant positive correlation between uPA and PAI-1 in melanomas with poor prognosis Radiol Oncol 2003; 37(1): 29-35. proves that also uPA has a prognostic impact. It indicates that their role in melanoma growth is co-dependent. There are various speculative explanations of the role of excessive PAI -1 production in the tumour:14,16,19 a) PAI-1 is important for reimplantation of circulating tumour cells at distant loci; b) since PAI-1 is present in en-dothelial cells and platelets, increased PAI-1 levels may reflect a high degree of angiogene-sis, thus favouring tumour spread and metas-tasis; c) PAI-1 binds to adhesive glycoprotein vitronectin and influences cell adhesion and migration. The prognostic role of PAI-2 is not well known yet. In breast cancer patients with high uPA concentrations, PAI-2 correlated with good prognosis.20 In colorectal cancer, however, high PAI-2 concentrations were associated with aggressive disease.21 Our results prove that using ELISA it is possible to quantify the uPA, PAI-1 and PAI-2 concentration in very small specimens of early melanoma lesions. Analogous to breast cancer, uPA and PAI-1 could become an additional prognostic fac-tor for the progression of melanoma next to the established histological criteria. Further-more, the fundamental role of uPA and PAI-1 in tumour invasion and metastasising indi-cates that these factors should be explored as targets for tumour biology - oriented thera-pies.22 To conclude, our results indicate that us-ing ELISA uPA, PAI-1 and PAI-2 concentra-tions can be measured also in small samples of primary melanoma and that both uPA and PAI-1 might also be of prognostic importance in primary malignant melanoma. These find-ings need to be confirmed in further studies where relationship between uPA, PAI-1 and patient survival will be investigated. Markovič J et al. / uPA, PAI-1 and PAI-2 in melanoma 35 References 1. Slominski A, Ross J, Mihm MC. Cutaneous melanoma: pathology, relevant prognostic indica-tors and progression. Br Med Bull 1995; 51: 548-69. 2. Andreasen PA, Kjoller L, Christensen L, Duffy MJ. The urokinase-type plasminogen activator system in cancer metastasis: a review. Int J Cancer 1997; 72: 1-22. 3. Schmitt M, Harbeck N, Thomssen C, Wilhelm O, Magdolen V, Reuning U, et al. Clinical impact of the plasminogen activation system in tumor invasion and metastasis: prognostic relevance and target for therapy. Thromb Haemost 1997; 78: 285-96. 4. Reuning U, Magdolen V, Wilhelm O, Fischer K, Lutz V, Graeff H, Schmitt M. Multifunctional potential of the plasminogen activation system in tumor invasion and metastasis. Int J Oncol 1998; 13: 893-906. 5. Duffy MJ, Maguire TM, McDermott EW, O’Higgins N. Urokinase plasminogen activator: a prognostic marker in multiple types of cancer. J Surg Oncol 1999; 71: 130-5. 6. Quax PHA, Van Muijen GNP, Weening - Verhoeff EJD, Lund LR, Dano K, Ruiter DJ, et al. Metastatic behaviour of human melanoma cell lines in nude mice correlates with urokinase-type plasminogen activator, its type-1 inhibitor and urokinase mediated matrix degradation. J Cell Biol 1992; 115: 191-9. 7. De Vries TJ, Quax PHA, Denijn M, Verrijp KN, Verheijen JH, Verspaget HW, et al. Plasminogen activators, their inhibitors, and urokinase receptor emerge in late stages of melanocytic tumor progression. Am J Path 1994; 144: 70-81. 8. De Vries TJ, Mooy CM, Van Balken MR, Luyten GPM, Quax PHA, Verspaget HW, et al. Components of the plasminogen activation system in uveal melanoma: a clinico - pathological study. J Pathol 1995; 175: 59-67. 9. American Joint Committee on Cancer: Manual for staging of cancer. Philadelphia: JB Lippincott Co.; 1997. 10. Benraad TJ, Geurts-Moespot J, Gro/ndahl-Hansen J, Schmitt M, Heuvel JJTM, de Witte JH, et al. Immunoassays (ELISA) of urokinase-type plas-minogen activator (uPA): report of an EORTC/ BIOMED-1 workshop. Eur J Cancer 1996; 32A (8): 1371-81. 11. Fräki JE, Nieminen S, Hopsu-Havu VK. Proteolytic enzymes and plasminogen activators in melanoma. J Cutan Pathol 1979; 6: 195-200. 12. De Vries TJ, Van Muijen GNP, Ruiter DJ. The plas-minogen activation system in melanoma cell lines and in melanocytic lesions. Mel Res 1996; 6: 79-88. 13. Delbaldo C, Masouye I, Saurat JH, Vassalli JD, Sappino AP. Plasminogen activation in melanocyt-ic neoplasia. Cancer Res 1994; 54: 4547-52. 14. Jänicke F, Schmitt M, Pache L, Ulm K, Harbeck N, Hofler H, Graeff H. Urokinase plasminogen acti-vator (u-PA) and its inhibitor PAI-1 are strong and independent prognostic factors in node-negative breast cancer. Breast Cancer Res Treat 1993; 24: 195-208. 15. Gro/ndahl-Hansen J, Christensen IJ, Rosenquist C, Brünner N, Mouridsen HT, Dano/ K, et al. High levels of urokinase-type plasminogen activator (uPA) and its inhibitor PAI-1 in cytosolic extracts of breast carcinomas are associated with poor prognosis. Cancer Res 1993; 53: 2513-21. 16. Nekarda H, Schmitt M, Ulm K, Wenninger A, Vogelsang H, Becker K, et al. Prognostic impact of urokinase-type plasminogen activator and its inhibitor PAI-1 in completely resected gastric can-cer. Cancer Res 1994; 54: 2900-7. 17. Ferrier CM, De Witte HH, Straatman H, van Tienoven DH, van Geloof WL, Rietveld FJR, et al. Comparison of immunohistochemistry with im-munoassay (ELISA) for the detection of compo-nents of the plasminogen activation system in human tumour tissue. Br J Cancer 1999; 79: 1534-41. 18. Ferrier CM, Suciu S, vanGeloof WL, Straatman H, Eggermont AM, Koops HS, et al. High tPA-expres-sion in primary melanoma of the limb correlates with good prognosis. Br.J.Cancer 2000; 83: 1351-9. 19. Brünner N, Nielsen HJ, Hamers M, Christensen IJ, Thorlacius-Ussing O, Stephens RW. The uroki-nase plasminogen activator receptor in blood from healthy individuals and patients with cancer. AP-MIS 1999; 107: 160-7. 20. Foekens JA, Buessecker F, Peters HA, Krainick U, van Putten WLJ, Look MP, et al. Plasminogen ac-tivator inhibitor-2: prognostic relevance in 1012 patients with primary breast cancer. Cancer Res 1995; 55: 1423-7. 21. Ganesh S, Sier CFM, Griffioen, Vloedgraven HJM, De Boer A, Welvaart K, et al. Prognostic relevance of plasminogen activators and their inhibitors in colorectal cancer. Cancer Res 1994; 54: 4065-71. 22. Schmitt M, Wilhelm OG, Reuning U, et al. The urokinase plasminogen activator system as a novel target tumor therapy. Fibrinol Proteol 2000; 14: 114-32. Radiol Oncol 2003; 37(1): 29-35.