Evidence that platelet and tumour heparanases are similar enzymes

In order to enter tissues, blood-borne metastatic tumour cells and leucocytes need to extravasate through the vascular basal lamina (BL), a process which involves a battery of degradative enzymes. A key degradative enzyme is the endoglycosidase heparanase, which cleaves heparan sulphate (HS), an important structural component of the vascular BL. Previously, tumour-derived heparanase activity (which has been shown to be related to the metastatic potential of murine and human melanoma cell lines) was reported to cleave HS and be inhibited by heparin, as distinct from human platelet heparanase, which cleaved both substrates [Nakajima, Irimura and Nicolson (1988) J. Cell Biochem. 36, 157-167]. We recently reported the purification of human platelet heparanase and showed that the enzyme is a 50-kDa endoglucuronidase [Freeman and Parish (1998) Biochem. J. 330, 1341-1350]. We now report the purification and characterization of heparanase activity from highly metastatic rat 13762 MAT mammary adenocarcinoma and human HCT 116 colonic carcinoma cells and from rat liver using essentially the same procedure that was reported for purification of the human platelet enzyme. The rat 13762 MAT tumour enzyme, which has a native M-r of 45 kDa when analysed by gel-filtration chromatography and by SDS/PAGE, was observed to be an endoglucuronidase that degraded heparin and I-IS to fragments of the same sizes as the human platelet enzyme does. N-deglycosylation of both the human platelet and rat 13762 MAT tumour enzymes gave, in each case, a 41-kDa band by SDS/PAGE analysis, demonstrating that the observed difference in M-r between the platelet and tumour enzymes may have been due largely to differences in the relative amounts of N-glycosylation. Two peptides were isolated following Endoproteinase Lys-C digestion of both the human platelet and rat 13762 MAT tumour heparanases and were shown to be highly similar. Both the rat river and human colonic carcinoma heparanases also degraded both heparin and HS to fragments of the same sizes as the human platelet enzyme does. Western-blot analysis of an SDS/PAGE gel using antibodies raised against human platelet heparanase demonstrated that human platelet, human tumour and rat tumour heparanases were immunochemically crossreactive. In conclusion, because of the similarities in their sizes, substrate specificities, peptide sequences and immunoreactivities, we propose that heparanase activities present in human platelets, rat liver and in rat and human tumour cells are, in fact, mediated by a similar enzyme.