OC-03 - Modelisation of the procoagulant properties of cancer cells and their capacity to alter the antithrombotic efficiency of LMWHs and specific inhibitors of factor Xa.

INTRODUCTION: The risk of venous thromboembolism varies according to the histological type of cancer. The failure of antithrombotic treatment is more frequent in cancer patients as compared to non-cancer ones. AIM: We aimed to elucidate the mechanism of activation of blood coagulation induced by cancer, the impact of chemo-resistance phenotype on the capacity of cancer cells to trigger thrombin generation and the alterations of the efficiency of LMWHs and the specific inhibitors of factor Xa (fondaparinux and apixaban) in the presence of cancer cells. MATERIALS AND METHODS: Thrombin generation of human plasma was assessed in the presence of various cancer cell lines. The model of cancer-induced hypercoagulability was coupled to the research for the expression of procoagulant molecules by cancer cells. RESULTS: The pancreatic adenocarcinoma cells BXPC3 and the breast adenocarcinoma cells MCF7 were initially tested. The BXPC3 cells induce significantly higher thrombin generation as compared to the MCF7 cells. In the same line Marchetti et al. showed that malignant hematologic cells (NB4, HEL, and K562) and H69 small cell lung cells express different procoagulant potential on triggering thrombin generation of human plasma. The comparison of the procoagulant activity has been extended to cancer cell lines from various cancers (i.e. colon, ovarian and prostatic cancer) as well as to different cell lines of the same type of cancer. The differences of the cancer cell lines to trigger thrombin generation are mainly due to the expression of TF. The acquisition of chemoresistant phenotype by cancer cells is correlated with increased TF expression and enhancement of theit procoagulant activity. The ability of cancer cells to activate FXII is an alternative pathway of significant importance for some cancer cell lines (i.e. MCF7). Clinically relevant concentrations of LMWH and specific direct and indirect inhibitors of FXa (apixaban and fondaparinux) inhibit thrombin generation induced by cancer cells. The synergy between the anti-Xa and anti-IIa activities of LMWHs rather than the AT-dependent selective inhibition of FXa results in profound inhibition of thrombin generation induced by BXPC3 cells. This experimental model allowed the functional distinction between the two specific FXa inhibitors (apixaban and fondaparinux). CONCLUSIONS: The cancer cell-based model of hypercoagulability is suitable for the identification of the prothrombotic fingerprint of various cancer types. This experimental model allows to perform pharmacological studies for the evaluation of the efficiency of the antithrombotic drugs in cancer-induced hypercoagulability. It is suitable for the study of the impact of anticancer drugs on the procoagulant properties of the cancer cells.

Glycans are involved in RANTES binding to CCR5 positive as well as to CCR5 negative cells.

We show that cell surface glycans, sialic acid and mannose-containing species, are involved beside glycosaminoglycans (GAGs), heparan sulfate and chondroitin sulfate in the binding of full length (1--68) RANTES not only to CCR5 positive human primary lymphocytes or macrophages but also to CCR5 negative monocytic U937 cells. Pretreating the cells with neuraminidase, heparitinase, chondroitinase or adding soluble glycans such as mannan or GAGs (heparin or chondroitin sulfate), significantly inhibited RANTES binding. Such effects were not observed with truncated (10--68) RANTES. Heat-denaturation of (1--68) RANTES strongly decreased its binding to the cells, demonstrating involvement of the three-dimensional structure. Accordingly, full length, but not truncated (10--68) RANTES, specifically bound to soluble mannan as well as to mannose-divinylsulfone-agarose affinity matrix and to soluble heparin or chondroitin sulfate as well as to heparin-agarose. Soluble heparin exerts, depending on its concentration, inhibitory or enhancing effects on RANTES binding to mannose-divinylsulfone-agarose, which indicates that RANTES interaction with glycans is modulated by GAGs. These data demonstrate that full length RANTES, but not its (10--68) truncated counterpart, interacts with glycans and GAGs, in soluble forms or presented either by affinity matrices or CCR5 positive as well as CCR5 negative cells.

Human alpha1-acid glycoprotein binds to CCR5 expressed on the plasma membrane of human primary macrophages.

We have reported previously that human alpha(1)-acid glycoprotein (AGP) inhibits the infection of human monocyte-derived macrophages (MDM) by R5 HIV-1, and that a disulphide-bridged peptide mimicking the clade B HIV-1 gp120 consensus V3 domain (V3Cs) binds specifically to CCR5 (the major co-receptor of R5 HIV strains) on these cells [Seddiki, Rabehi, Benjouad, Saffar, Ferriere, Gluckman and Gattegno (1997) Glycobiology 7, 1229-1236]. The present study demonstrates that AGP binds specifically to MDM at high- and low-affinity binding sites with K(d) values of 16 nM and 4.9 microM respectively. The fact that heat denaturation of AGP only partly inhibited this binding (43%) suggests that protein-protein interactions are involved, as well as AGP glycans which are resistant to heat denaturation. Mannan, but not dextran, is a significant inhibitor (52%) of this binding, and sequential exoglycosidase treatment of AGP, which exposes penultimate mannose residues, has a strong stimulatory effect ( approximately 2.8-fold). Therefore AGP glycans (probably mannose residues) are involved, at least partly, in the binding of AGP to MDM. In addition, AGP inhibits the binding of V3Cs and macrophage inflammatory protein-1beta (MIP-1beta) to MDM. The anti-CCR5 monoclonal antibody 2D7, specific for the second extracellular loop of CCR5, also inhibited AGP binding (67%), whereas anti-CCR5 antibodies specific for the C-terminus of CCR5 region had no effect. Native AGP, like V3Cs (but not heat-denatured AGP), binds to 46 and 33-36 kDa electroblotted AGP-bound MDM membrane ligands, characterized as CCR5 by their interactions with anti-CCR5 antibodies and with MIP-1beta. Therefore both AGP glycans and MDM CCR5 are involved in the binding of AGP to MDM. This suggests that the inhibitory effect of AGP on the infection of human primary macrophages by R5 HIV-1 may be related to specific binding of AGP to a macrophage membrane lectin or lectin-like component and to CCR5.

Glycan and glycosaminoglycan binding properties of stromal cell-derived factor (SDF)-1alpha.

We show here that cell surface glycosaminoglycans (GAGs) are involved in the binding of stromal cell-derived factor (SDF)-1alpha to CD4(+)lymphoid CEM or monocytic U937 cells, inasmuch as pretreating the cells with heparitinase or chondroitinase inhibits SDF-1alpha binding by 40-41% and 31-35%, respectively. Soluble heparin or chondroitin sulfate partially but significantly inhibits SDF-1alpha binding to the cells by 45-52% and 42-56%, respectively, while dextran has no significant effect. Taken together, these results indicate the role of GAGs in SDF-1alpha attachment to the cells. However, the effects of heparitinase and chondroitinase as well as those of heparin and chondroitin sulfate are not additive, which suggests that SDF-1alpha may attach to the cells through different GAGs, and also through other ligands. Soluble mannan also inhibits SDF-1alpha binding to the cells by 30-33%. Additivity between this effect and that of heparin or chondroitin sulfate is observed. Therefore, beside GAGs, mannose-containing species may also be involved in SDF-1alpha attachment to the cells. Accordingly, SDF-1alpha specifically binds to heparin-agarose and mannose-divinylsulfone agarose affinity matrices, and these interactions are inhibited respectively by soluble heparin, chondroitin sulfate, and mannan. We have previously shown that gp120 of X4 strain HIV-1LAI presents specific carbohydrate-binding properties for mannosylated derivatives, including mannan, and for GAGs including heparin. The present data therefore indicate that, in the same manner as HIV-1 Env, SDF-1alpha can interact with GAGs and glycans at the cell surface.

Glycans and proteoglycans are involved in the interactions of human immunodeficiency virus type 1 envelope glycoprotein and of SDF-1alpha with membrane ligands of CD4(+) CXCR4(+) cells.

We demonstrate that human immunodeficiency virus HIV-1(LAI) envelope glycoprotein 120 (gp120(LAi)) specifically interacts with several membrane ligands on lymphoid CEM or monocytic U937 cells in addition to its previously identified receptor, CD4, and CXCR4, its coreceptor. In its native state, gp120(LAI) is able to elicit specific multimolecular complexes with these membrane ligands at the surface of the cells; most of the interactions are abolished by mannan or heparin but not by dextran. Similarly, stromal cell-derived factor (SDF)-1alpha interacts not only with CXCR4 expressed by CXCR4(+) CD4(+) U937, CEM, and HOS-CD4(+) CXCR4(+) cells but also with CD4 expressed by intact U937, CEM, and HOS-CD4(+) CXCR4(+/-) cells or electroblotted onto Immobilon. SDF-1alpha binding to CD4(+) CXCR4(+/-) cells, or soluble CD4 electroblotted onto Immobilon, is significantly inhibited by sCD4, whereas truncated sCD4 lacking D3 and D4 domains had no significant effect, which indicates that SDF-1 binds to CD4 but at regions different from the HIV-gp120-binding site. Heparin and mannan also inhibit SDF-1alpha binding to intact CD4(+) CXCR4(+/-) cells, and electroblotted soluble CD4. Heparitinase treatment of such cells reduced SDF-1alpha binding. These data demonstrate that glycans and glycosaminoglycans are directly or indirectly involved in the interactions of HIV-1 gp120(LAI) and of SDF-1alpha with membrane ligands of CD4(+) CXCR4(+) cells and thus could play a role both in HIV-1 infection and in the physiology of SDF-1alpha.

Antiviral activity of derivatized dextrans on HIV-1 infection of primary macrophages and blood lymphocytes.

The present study demonstrates at the molecular level that dextran derivatives carboxymethyl dextran benzylamine (CMDB) and carboxymethyl dextran benzylamine sulfonate (CMDBS), characterized by a statistical distribution of anionic carboxylic groups, hydrophobic benzylamide units, and/or sulfonate moieties, interact with HIV-1 LAI gp120 and V3 consensus clades B domain. Only limited interaction was observed with carboxy-methyl dextran (CMD) or dextran (D) under the same conditions. CMDBS and CMDB (1 microM) strongly inhibited HIV-1 infection of primary macrophages and primary CD4+ lymphocytes by macrophage-tropic and T lymphocyte-tropic strains, respectively, while D or CMD had more limited effects on M-tropic infection of primary macrophages and exert no inhibitory effect on M- or T-tropic infection of primary lymphocytes. CMDBS and CMDB (1 microM) had limited but significant effect on oligomerized soluble recombinant gp120 binding to primary macrophages while they clearly inhibit (> 50%) such binding to primary lymphocytes. In conclusion, the inhibitory effect of CMDB and the CMDBS, is observed for HIV M- and T-tropic strain infections of primary lymphocytes and macrophages which indicates that these compounds interfere with steps of HIV replicative cycle which neither depend on the virus nor on the cell.

Inhibition of human immunodeficiency virus infection of CD4+ cells by CD4-free glycopeptides from monocytic U937 cells.

We have previously demonstrated that human immunodeficiency virus (HIV) envelope glycoproteins have specific carbohydrate-binding properties for mannosyl/N-acetylglucosaminyl residues presented at high density on a carrier in vitro. Here, we investigated whether HIV envelope glycoprotein gp120 was able to interact with surface membrane carbohydrates of CD4+ cells by means of such lectin-carbohydrate interactions. CD4-free tryptic glycopeptides, prepared from the membrane of CD4+ monocytic U937 cells and partially purified by ConA-agarose affinity chromatography, could be eluted by mannan but not by methyl-alpha-mannose or methyl-alpha-glucose, which strongly suggests that they displayed oligomannosidic structures. These glycopeptides bound in a mannosyl-specific manner to radiolabeled recombinant gp120. Deglycosylation with N-glycanase which, as expected, strongly diminished binding of the glycopeptides to concanavalin A also abolished their interaction with gp120. In addition, the glycopeptides inhibited HIV infection of both U937 and CD4+ lymphoid CEM cells when preincubated with the virus. These findings indicate that, independently of the binding to CD4, mannosyl structures on CD4+ cells may play a role through lectin-carbohydrate interactions in envelope glycoprotein binding to a putative coreceptor(s) of HIV.

Interactions of HIV-1 envelope glycoproteins with derivatized dextrans.

The present study demonstrates that derivatized dextrans, such as carboxylmethyl dextran benzylamide and carboxymethyl dextran benzylamide sulfonate, specifically interact with HIV-1 envelope glycoproteins (rgp160 and rgp41) with significantly higher affinities than those observed for dextran sulfate (MW 8 kDa). These results suggest the possible involvement in HIV infectivity of surface membrane molecules which may bind the virus at pre or post-CD4 binding steps. They also suggest the possible use of these compounds in anti-HIV therapy.

Interactions of HIV-1 and HIV-2 envelope glycoproteins with sulphated polysaccharides and mannose-6-phosphate.

Envelope glycoproteins of human immunodeficiency viruses (HIV-1 and HIV-2) can interact with high-mannose glycans and with the mannosyl or N-acetylglucosaminyl core of complex-type oligosaccharidic structures. HIV-1 glycoproteins also specifically bind sulphated polysaccharides such as dextran sulphate (DS) and heparin. Here, we show that the latter property is shared by HIV-2 recombinant gp140 (rgp140) precursor glycoprotein. Binding of rgp140 and of corresponding rgp160 of HIV-1 to heparin- and DS-substituted (sulphated dextran beads; SDB) affinity matrices was inhibited by the soluble specific ligand and also by fetuin, asialofetuin or the anionic simple carbohydrate derivative mannose-6-phosphate (M6P). Interaction of HIV-1 rgp120 subunit with the two affinity matrices was also inhibited by M6P, but only rgp120 binding to heparin-agarose, and not that to SDB, was affected by fetuin and asialofetuin. These results suggest that HIV-1 and HIV-2 envelope glycoproteins presumably display different sulphated polysaccharide and carbohydrate recognition sites. Some of these may be common or in close proximity: with respect to rgp160, for example, the sites may be common on the gp41 moiety and/or in a region of gp120 which would be more accessible when expressed on rgp160 than on processed gp120, while they may be distinct on the cleaved gp120 subunit. Finally, because M6P is a marker of lysosomal enzymes, we verified that HIV-1 and HIV-2 envelope glycoproteins could specifically bind in a M6P-inhibitable manner to a representative lysosomal enzyme, bovine liver beta-glucuronidase coupled to agarose, suggesting that they may possibly interfere with lysosomal enzyme sorting in HIV-infected cells.

The interaction of a glycosaminoglycan, heparin, with HIV-1 major envelope glycoprotein.

We demonstrate in vitro the occurrence of a specific but low-affinity interaction between soluble tetrameric rgp160 or soluble monomeric or tetrameric rgp120 and heparin-agarose (HA). This interaction is saturable, pH and temperature-dependent, and can be inhibited by soluble heparin, but not by soluble dextran. In buffer supplemented with 10 mM CaCl2, the C50 of soluble heparin, i.e., the concentration of soluble heparin which leads to 50% inhibition of the binding of [125I]rgp160 or of [125I]rgp120 to HA, is 1.1 x 10(-4) disaccharidic molar concentration for rgp160 and 3.2 x 10(-4) dissacharidic molar concentration for rgp120, which indicates low-affinity interactions. Upon chromatography on HA, [125I]rgp160 is repeatedly eluted as a retarded fraction when compared to the elution volume of [125I]rgp160-soluble heparin complex. Under the same experimental conditions, [125I]rgp120 is also eluted, but as a less retarded fraction than [125I]rgp160. Taken together, these results suggest that, at least part of the described anti HIV-1 activity of heparin might be mediated by interaction with HIV-1 major envelope glycoprotein.

Molecular interaction between HIV-1 major envelope glycoprotein and dextran sulfate.

We investigated at the molecular level the interaction between, HIV-1 recombinant gp160 (rgp160) and low-molecular-weight dextran sulfate. We demonstrate the occurrence of a specific interaction between rgp160 and sulfated dextran beads, which is saturable, pH-dependent and inhibitable by soluble dextran sulfate but not by soluble dextran. This specific interaction has a low affinity, with an estimated Kd in the 10(-4) M range. In addition, the binding of rgp160 to soluble recombinant CD4 (sT4) can only be inhibited by the preincubation of rgp160, but not of sT4, with dextran sulfate. Taken together, these results demonstrate the occurrence of a low affinity, but specific interaction between dextran sulfate and rgp160. This may account, at least in part, for the anti-HIV-1 activity of dextran sulfate.