Tumor cell invasion of three-dimensional matrices of defined composition: evidence for a specific role for heparan sulfate in rodent cell lines

The abilities of rodent tumor cell lines; B16BL6, ND and LT dietary variants of B16BL6, +SA, RT7-4bs and RT7-4bLs to invade composite collagen I gels containing heparin, chemically modified heparins, heparan sulfate, chondroitin sulfate, hyaluronic acid, dextran, dextran sulfate, laminin and collagen IV were investigated, and compared to the invasion of plain collagen I gels. The presence of heparin or heparan sulfate most generally promoted tumor cell invasion of the gels, with more aggressive invasion being noted for the more metastatic variants examined. Of the chemically modified heparins tested, carboxyl-reduced heparin promoted matrix invasion by B16BL6 and +SA cells to the greatest degree. Hyaluronic acid marginally promoted invasion by +SA and RT7-4bs primary cells while, in these collagen I based gels laminin only promoted matrix invasion by primary +SA cells to a very limited degree. The tumor cell lines attached relatively poorly to heparan sulfate substrates compared to the other glycosaminoglycans tested, and the primary tumor cell lines also attached relatively poorly to collagen I. As expected, highly metastatic variants showed greater attachment to laminin than did their less metastatic counterparts. Apart from the negative correlation of cellular attachment to heparan sulfate substrates with invasiveness towards heparan sulfate containing gels, no other relationships emerged linking attachment rates with invasive activities for particular complex gel compositions. Our results suggest an important role for heparan sulfate, and possibly also tissue heparin, in promoting tumor cell invasion of extracellular matrices. Results from complex gels containing dextran or dextran sulfate failed to support the hypothesis that GAG sulfation is important to cellular invasion. The activity of the chemically modified heparins in promoting invasion, when present as components of these model matrices, suggests that part of the anti-metastatic activity of these compounds, when preincubated with tumor cells prior to intravenous inoculation, could result from interference with tumor cell extravasation.