Using an ELISA approach, we demonstrate that recombinant human IL-12 (rhIL-12) binds strongly to an immobilized heparin-BSA complex. This binding is completely displaceable with soluble heparin, IC50 approximately 0.1 microg/ml, corresponding to approximately 10 nM. By interpolation with our previous findings, this indicates an affinity for heparin greater than that of antithrombin III and comparable with that of FGF-2, two high-affinity heparin-binding proteins. Recombinant murine IL-12 also binds strongly to heparin. The binding of rhIL-12 to heparin shows specificity because chondroitin sulfates A and C fail to compete, whereas chondroitin B inhibits weakly. A highly sulfated heparan sulfate is a strong competitor, whereas other heparan sulfates show weak or no activity. Small heparin fragments inhibit binding, although activity decreases with size. An octasaccharide pool derived by cleavage of heparin with nitrous acid is a significantly stronger inhibitor than its heparinase I-derived counterpart, further indicating structural specificity in the interaction between rhIL-12 and heparin. The binding of recombinant p40 to heparin appears indistinguishable from that of the IL-12 heterodimer, implying that the heparin binding site is largely if not solely located in this subunit. These results show for the first time that IL-12 is a heparin-binding cytokine, a property common to the other Th1-response-inducing cytokines, IFN-gamma and IL-2. Our findings strongly suggest that IL-12 will tend to be retained close to its sites of secretion in the tissues by binding to heparin-like glycosaminoglycans, thus favoring a paracrine role for IL-12.

We have developed an enzyme-linked immunosorbent assay (ELISA) approach for the study of interactions between cytokines and glycosaminoglycans. This involves, as solid phase, a synthetic heparin-bovine serum albumin (BSA) complex in which the heparin is coupled via its reducing terminus to the protein using sodium cyanoborohydride. We have investigated the sensitivity and specificity of this experimental technique, employing antithrombin (AT III) and fibroblast growth factor 2 (FGF-2) as well-characterized heparin binding proteins. Using this ELISA method, we have established that human recombinant interleukin (IL-2) binds to heparin in a concentration-dependent manner. Soluble heparin competes for the binding of IL-2 to the complex with 50% inhibition at 5 microg/ml. This IC50 value provides an estimate of the binding constant of around 0.5 microM. This value is at least two orders of magnitude larger than that for the binding of IL-2 to its dimeric and trimeric cell surface receptors, but similar to that for binding to the IL-2 receptor beta polypeptide acting alone. Our ELISA shows that in addition to soluble heparin, fuciodan also competes for IL-2 binding, but chondroitin sulfate and dermatan sulfate are inactive. Of six heparan sulfates tested, only one highly sulfated preparation competed for IL-2. The interaction between IL-2 and heparin-like glycosaminoglycans is likely to be an important mechanism for retaining IL-2 close to its sites of secretion, thus giving rise to localized concentration gradients in the tissues.

An increasing number of polypeptide growth factors and cytokines are now known to bind to heparin and heparan sulphate glycosaminoglycans. Well studied examples are members of the chemokine and fibroblast growth factor families. However other growth factors including proinflammatory and haematopoietic cytokmes also bind to heparin [ 1-41, Such interaction with glycosaminoglycans of the extracellular matrix and cell surface may be important in protecting cytokines against degradation, and in localising them close to their sites of secretion. We have developed a novel ELISA approach for the investigation of heparin-cytokine interactions. For this heparin is covalently bound via its reducing terminus to a protein carrier, bovine serum albumin (BSA), using sodium cyanoborohydnde. This coupling method is intended to ensure that the heparin chains remain accessible, for subsequent engagement in protein-binding interactions. Antithrombin 111 (AT 111) and fibroblast growth factor-2 (basic-FGF), both proteins with well characterised heparinbindmg properties, bind to the heparin-BSA complex in our ELISA. In both cases, bindmg is a dose dependent and saturable over the 0-3Ong range. Free heparin competes with binding of both proteins to the complex, with an EC,, value of around 50Ong/ml. With AT III, we have found that an AT 111-binding heparin fiaction separated by aflimty chromatography on immobilised AT 111, is a strong competitor in our ELISA. Conversely the low affinity heparin fraction is a poor competitor. Thus our ELISA shows the anticipated specificity in heparin-protein interactions. We have now shown that recombinant human i n t e r l e h 2, rIL-2, binds to the heparin complex (see Fig 1). The binding is dose dependent, but over a range of 0-100ng does not reach saturation. As with AT 111 and FGF-2, free heparin displaces rIL-2 fiom the complex, but with rIL-2 the EC,, for mhibition of binding is an order of magnitude higher, at around 7pg/ml. Thus it would appear that the afhuty of IL-2 for heparin is lower than that of AT 111 and

The expression of lectins on the surface of the murine multi-myeloid progenitor cell line FDCP-Mix, and the human leukaemic cell line KG1, was assessed and compared to the pattern of lectin expression observed on human bone marrow mononuclear cells. Using flow cytometry, cell-surface lectins were identified by their ability to recognise fluorescein isothiocyanate-labelled neoglycoproteins. Both cell lines recognised neoglycoproteins expressing alpha-D-glucose and alpha-D-galactose residues. Inhibition studies suggested that recognition of these neoglycoproteins was via two independent receptors, each displaying characteristic sugar-binding properties. The CD34+ population of bone marrow mononuclear cells, identified by positive staining with the anti-CD34 antibody QBend10, were shown to interact with alpha-glucose-, alpha-galactose- and alpha-D-mannose-expressing neoglycoproteins. Similarly, binding of these probes to lymphocyte and monocyte sub-populations of CD34- bone marrow mononuclear cells was observed. In contrast, CD34- granulocytic cells did not appear to recognise these probes. It is suggested that the alpha-D-galactose binding activity observed for both cell lines and the alpha-D-galactose and alpha-D-mannose binding activity observed for bone marrow mononuclear cells represent expression of the component polypeptides of the previously reported galactosyl/mannosyl receptor. The glucosyl-specific receptor, observed on both cell lines and on bone marrow mononuclear cells, has not been reported previously. It is suggested that this receptor may mediate glucose transport or cell adhesion through recognition of glucosyl-containing compounds such as heparan sulphate.

We have investigated the binding of interleukin 7 (IL-7) to sulfated glycosaminoglycans and evaluated its biological consequences. IL-7 binds to heparin and heparan sulfate, to a lesser extent to dermatan sulfate and does not bind to chondroitin sulfate. It was eluted from heparin by 0.3-0.6 M NaCl and from heparan sulfate by < 0.3 M NaCl. We also measured the affinity of IL-7 for heparin using an affinity co-electrophoresis method and found an affinity of 25 nM. In spite of these findings, IL-7 does not bind to the S17 cell line which supports lymphopoiesis. However, addition of heparin to cultures of an IL-7-dependent pre-B cell line (2E8) inhibited IL-7-stimulated proliferation and IL-7 complexed with heparin was more resistant than free IL-7 to protease treatment. Taken together, these results suggest that heparin may act as a carrier for IL-7, blocking its interaction with target cells and protecting it from degradation during transit.