Thrombin induces tumor growth, metastasis, and angiogenesis: Evidence for a thrombin-regulated dormant tumor phenotype.

The association of idiopathic venous thrombosis with occult cancer is generally recognized. However, it has not been fully appreciated that thrombin generated during thrombosis can augment the malignant phenotype. Thrombin activates tumor cell adhesion to platelets, endothelial cells, and subendothelial matrix proteins; enhances tumor cell growth; increases tumor cell seeding and spontaneous metastasis; and stimulates tumor cell angiogenesis. These mechanisms are reviewed. Evidence is also presented to support the hypothesis that thrombin serves to preserve dormant tumor cells in individuals, preventing host eradication. It is proposed that tumor malignancy may be regulated by a procoagulant/anticoagulant axis.

Dissolution of arterial platelet thrombi in vivo with a bifunctional platelet GPIIIa49-66 ligand which specifically targets the platelet thrombus.

Patients with HIV-1 immune-related thrombocytopenia have a unique antibody (Ab) against integrin GPIIIa49-66 capable of inducing oxidative platelet fragmentation via Ab activation of platelet nicotinamide adenine dinucleotide phosphate oxidase and 12-lipoxygenase releasing reactive oxygen species. Using a phage display single-chain antibody (scFv) library, we developed a novel human monoclonal scFv Ab against GPIIIa49-66 (named A11) capable of inducing fragmentation of activated platelets. In this study, we investigated the in vivo use of A11. We show that A11 does not induce significant thrombocytopenia or inhibit platelet function. A11 can prevent the cessation of carotid artery flow produced by induced artery injury and dissolve the induced thrombus 2 hours after cessation of blood flow. In addition, A11 can prevent, as well as ameliorate, murine middle cerebral artery stroke, without thrombocytopenia or brain hemorrhage. To further optimize the antithrombotic activity of A11, we produced a bifunctional A11-plasminogen first kringle agent (SLK), which homes to newly deposited fibrin strands within and surrounding the platelet thrombus, reducing effects on nonactivated circulating platelets. Indeed, SLK is able to completely reopen occluded carotid vessels 4 hours after cessation of blood flow, whereas A11 had no effect at 4 hours. Thus, a new antithrombotic agent was developed for platelet thrombus clearance.

Role of molecular mimicry of hepatitis C virus protein with platelet GPIIIa in hepatitis C-related immunologic thrombocytopenia.

Patients with HIV-1 immune-related thrombocytopenia (HIV-1-ITP) have a unique Ab against platelet GPIIIa49-66 capable of inducing oxidative platelet fragmentation in the absence of complement. HIV-1-seropositive drug abusers are more prone to develop immune thrombocytopenia than non-drug abusers and have a higher coinfection with hepatitis C virus (HCV) than non-drug abusers (90% vs 30%). Molecular mimicry was sought by screening a phage peptide library with anti-GPIIIa49-66 antibody as bait for peptides sharing homology sequences with HCV. Several phage peptide clones had 70% homology with HCV protein. Sera from dually infected thrombocytopenic patients with HCV and HIV-ITP reacted strongly with 4 nonconserved peptides from HCV core envelope 1. Reactivity correlated inversely with platelet count (r(2) = 0.7, P < .01). Ab raised against peptide PHC09 in GPIIIa(-/-) mice induced thrombocytopenia in wild-type mice. Affinity-purified IgG against PHC09 induced oxidative platelet fragmentation in vitro. Drug abusers dually infected with HCV and HIV-1 had a greater incidence and severity of thrombocytopenia as well as titer of anti-GPIIIa49-66/PHC09 Ab. NZB/W F1 mice injected with recombinant core envelope 1 developed Ab versus PHC09 and significantly decreased their platelet count (P < .001). Thus, HCV core envelope 1 can induce thrombocytopenia by molecular mimicry with GPIIIa49-66.

C-terminal ADAMTS-18 fragment induces oxidative platelet fragmentation, dissolves platelet aggregates, and protects against carotid artery occlusion and cerebral stroke.

Anti-platelet integrin GPIIIa49-66 antibody (Ab) induces complement-independent platelet oxidative fragmentation and death by generation of platelet peroxide following NADPH oxidase activation. A C-terminal 385-amino acid fragment of ADAMTS-18 (a disintegrin metalloproteinase with thrombospondin motifs produced in endothelial cells) induces oxidative platelet fragmentation in an identical kinetic fashion as anti-GPIIIa49-66 Ab. Endothelial cell ADAMTS-18 secretion is enhanced by thrombin and activated by thrombin cleavage to fragment platelets. Platelet aggregates produced ex vivo with ADP or collagen and fibrinogen are destroyed by the C-terminal ADAMTS-18 fragment. Anti-ADAMTS-18 Ab shortens the tail vein bleeding time. The C-terminal fragment protects against FeCI3-induced carotid artery thrombosis as well as cerebral infarction in a postischemic stroke model. Thus, a new mechanism is proposed for platelet thrombus clearance, via platelet oxidative fragmentation induced by thrombin cleavage of ADAMTS-18.

Growth-regulated oncogene is pivotal in thrombin-induced angiogenesis.

The mechanism of thrombin-induced angiogenesis is poorly understood. Using a gene chip array to investigate the pro-malignant phenotype of thrombin-stimulated cells, we observed that thrombin markedly up-regulates growth-regulated oncogene-alpha (GRO-alpha) in several tumor cell lines as well as endothelial cells by mRNA and protein analysis. Thrombin enhanced the secretion of GRO-alpha from tumor cells 25- to 64-fold. GRO-alpha is a CXC chemokine with tumor-associated angiogenic as well as oncogenic activation following ligation of its CXCR2 receptor. GRO-alpha enhanced angiogenesis in the chick chorioallantoic membrane assay 2.2-fold, providing direct evidence for GRO-alpha as an angiogenic growth factor. Anti-GRO-alpha antibody completely inhibited the 2.7-fold thrombin-induced up-regulation of angiogenesis, as well as the 1.5-fold thrombin-induced up-regulation of both endothelial cell cord formation in Matrigel and growth in vitro. Thrombin as well as its PAR-1 receptor activation peptide [thrombin receptor activation peptide (TRAP)] as well as GRO-alpha all markedly increased vascular regulatory proteins and growth factors: matrix metalloproteinase (MMP)-1, MMP-2, vascular endothelial growth factor (VEGF), angiopoietin-2 (Ang-2), CD31, and receptors KDR and CXCR2 in human umbilical vein endothelial cells. All of the thrombin/TRAP gene up-regulations were completely inhibited by anti-GRO-alpha antibody and unaffected by irrelevant antibody. Similar inhibition of gene up-regulation as well as thrombin-induced chemotaxis was noted with small interfering RNA (shRNA) GRO-alpha KD 4T1 breast tumor and B16F10 melanoma cells. In vivo tumor growth studies in wild-type mice with shRNA GRO-alpha KD cells revealed 2- to 4-fold impaired tumor growth, metastasis, and angiogenesis, which was not affected by endogenous thrombin. Thus, thrombin-induced angiogenesis requires the up-regulation of GRO-alpha. Thrombin up-regulation of GRO-alpha in tumor cells as well as endothelial cells contributes to tumor angiogenesis.

Complement-independent Ab-induced peroxide lysis of platelets requires 12-lipoxygenase and a platelet NADPH oxidase pathway.

Antiplatelet GPIIIa49-66 Ab of HIV-related thrombocytopenic patients induces thrombocytopenia and platelet fragmentation by the generation of peroxide and other reactive oxygen species (ROS). Here we report the presence of a functional platelet NADPH oxidase pathway that requires activation by the platelet 12-lipoxygenase (12-LO) pathway to fragment platelets. A new Ab-mediated mechanism is described in which the platelet 12-LO product, 12(S)-HETE activates the NADPH oxidase pathway to generate ROS.

Role of thrombin as a tumor growth factor.

The clinical observation that thrombosis in some patients heralds the onset of malignancy has been recognized for over a century. Thrombin the key terminal enzyme of coagulation also promotes angiogenesis and stimulates tumor-platelet adhesion, adhesion to endothelium, tumor implantation, tumor cell growth and metastasis. The thrombin receptor, a member of the protease-activated receptor family, is expressed on many tumor cell lines and on breast tumor biopsy specimens. In addition to mitogenic effects on fibroblast, smooth muscle cells and endothelial cells, thrombin also exerts direct effects on cancer cells by activation of the cell cycle through downregulation of p27(Kip1) and induction of Skp2, and cyclins D and A. MicroRNA 222, which inhibits p27(Kip1), is upregulated by thrombin. In the transgenic TRAMP mouse model of prostate cancer inhibition of endogenous thrombin by hirudin retards spontaneous tumor growth. Inhibition of thrombin may lead to tumor dormancy and could explain inhibition of tumor growth and metastasis by anticoagulants observed in animal models and a beneficial effect on survival observed in some clinical trials of anticoagulants in cancer patients.

Thrombin induces tumor cell cycle activation and spontaneous growth by down-regulation of p27Kip1, in association with the up-regulation of Skp2 and MiR-222.

The effect of thrombin on tumor cell cycle activation and spontaneous growth was examined in synchronized serum-starved tumor cell lines and a model of spontaneous prostate cancer development in TRAMP mice. BrdUrd incorporation and propidium iodide staining of prostate LNCaP cells arrested in G(0) and treated with thrombin or serum revealed a 48- and 29-fold increase in S phase cells, respectively, at 8 hours. Similar results were obtained with TRAMP cells and a glioblastoma cell line, T98G. Cell cycle kinases and inhibitors in synchronized tumor cells revealed high levels of p27(Kip1) and low levels of Skp2 and cyclins D1 and A. Addition of thrombin, TFLLRN, or serum down-regulated p27(Kip1) with concomitant induction of Skp2, Cyclin D1, and Cyclin A with similar kinetics. LNCaP p27(Kip1)-transfected cells or Skp2 knockdown cells were refractory to thrombin-induced cell cycle activation. MicroRNA 222, an inhibitor of p27(Kip1), was robustly up-regulated by thrombin. The in vitro observations were tested in vivo with transgenic TRAMP mice. Repetitive thrombin injection enhanced prostate tumor volume 6- to 8-fold (P < 0.04). Repetitive hirudin, a specific potent antithrombin, decreased tumor volume 13- to 24-fold (P < 0.04). Thus, thrombin stimulates tumor cell growth in vivo by down-regulation of p27(Kip1).

Twist is required for thrombin-induced tumor angiogenesis and growth.

Twist, a master regulator of embryonic morphogenesis, induces functions that are also required for tumor invasion and metastasis. Because thrombin contributes to the malignant phenotype by up-regulating tumor metastasis, we examined its effect on Twist in five different tumor cell lines and two different endothelial cell lines. Thrombin up-regulated Twist mRNA and protein in all seven cell lines. Down-regulation of Twist in B16F10 tumor cell lines led to a approximately 3-fold decrease in tumor growth on a chorioallantoic membrane assay and approximately 2-fold decrease in syngeneic mice. Angiogenesis was decreased approximately 45% and 36%, respectively. The effect of Twist on angiogenesis was further examined and compared with the effect of thrombin. In studies using a Twist-inducible plasmid, several identical vascular growth factors and receptors were up-regulated approximately 2- to 3-fold in tumor cells as well as human umbilical vascular endothelial cells by both Twist as well as thrombin (vascular endothelial growth factor, KDR, Ang-2, matrix metalloproteinase 1, GRO-alpha, and CD31). Thrombin-induced endothelial cell chemotaxis and Matrigel endothelial cell tubule formation were similarly regulated by Twist. Thus, thrombin up-regulates Twist, which is required for thrombin-induced angiogenesis as measured by endothelial cell migration, Matrigel tubule formation, and tumor angiogenesis.

Thrombin up-regulates cathepsin D which enhances angiogenesis, growth, and metastasis.

Cathepsin D (CD) up-regulation has been associated with human malignancy and poor prognosis. Thrombin up-regulated CD mRNA and protein in eight tumor cell lines as well as in human umbilical vascular endothelial cells (HUVEC). Thrombin increased the secretion of CD by 3- to 8-fold and enhanced chemotaxis ( approximately 2-fold) in 4T1 murine mammary CA cells, which was completely inhibited with the knockdown of CD. Secreted 4T1 CD induced neoangiogenesis by 2.4-fold on a chick chorioallantoic membrane, which was blocked in CD-KD cells. The addition of pure CD (2 ng) to the chick chorioallantoic membrane increased angiogenesis by 2.1-fold, which was completely inhibited by Pepstatin A (Pep A). CD enhanced human HUVEC chemotaxis and Matrigel tube formation by 2-fold, which was then blocked by Pep A. CD enhanced HUVEC matrix metalloproteinase 9 (MMP-9) activity by approximately 2-fold, which was completely inhibited by Pep A as well as a generic MMP inhibitor, GM6001. The injection of CD-KD 4T1 cells into syngeneic mice inhibited tumor growth by 3- to 4-fold compared with empty vector (EV) cells. Hirudin, a specific thrombin inhibitor, inhibited the growth of wild-type and EV cells by 2- to 3-fold, compatible with thrombin up-regulation of CD. CD and thrombin also contributed to spontaneous pulmonary metastasis; 4-fold nodule inhibition with CD versus EV and 4.6-fold inhibition with hirudin versus EV (P < 0.02). Thus, thrombin-induced CD contributes to the malignant phenotype by inducing tumor cell migration, nodule growth, metastasis, and angiogenesis. CD-induced angiogenesis requires the proteolytic activation of MMP-9.

Platelet fragmentation requires a specific structural conformation of human monoclonal antibody against beta3 integrin.

We have described an autoantibody against beta3 (GPIIIa49-66), a region of platelet integrin alphaIIbbeta3 that is unique. It induces platelet fragmentation in the absence of complement via antibody activation of platelet NADPH oxidase and 12-lipoxygenase to release reactive oxygen species, which destroy platelets. To study the mechanism of anti-GPIIIa antibody-induced platelet fragmentation, we screened a human single chain Fv antibody library with the GPIIIa49-66 peptide. Nine monoclonal antibodies were identified that were capable of binding to GPIIIa49-66. Surprisingly, binding avidity for GPIIIa49-66 did not correlate with activity of induction of platelet fragmentation. We therefore investigated the requirements for platelet fragmentation. Mutations were introduced into the heavy chain complementary-determining region-3 of clones 11, 43, and 54 by site-directed mutagenesis. The capability of these clones to induce platelet fragmentation or bind to GPIIIa49-66 subsequently changed. Molecular modeling of these clones with their mutants revealed that the ability to induce platelet fragmentation is affected by the side chain orientation of positively charged amino acids in the heavy chain of residues 99-102. Thus, a structural change in the conformation of anti-GPIIIa49-66 antibody contributes to its binding to the beta3 integrin and subsequent antibody-induced platelet fragmentation and aggregate dissolution.

Platelet particle formation by anti GPIIIa49-66 Ab, Ca2+ ionophore A23187, and phorbol myristate acetate is induced by reactive oxygen species and inhibited by dexamethasone blockade of platelet phospholipase A2, 12-lipoxygenase, and NADPH oxidase.

An HIV antibody (Ab) against platelet integrin GPIIIa49-66 induces complement-independent platelet particle formation by the elaboration of reactive oxygen species (ROS) downstream of the activation of the platelet NADPH oxidase by the 12-lipoxygenase (12-LO) product 12(S)-HETE. To determine whether other inducers of platelet particle formation also function via the induction of ROS, we examined the effects of the Ca(2+) ionophore A23187 and phorbol myristate acetate (PMA). Both agents induced oxidative platelet particle formation in an identical fashion as Ab, requiring Ca(2+) flux and 12(S)-HETE production as well as intact NADPH oxidase and 12-LO pathways. Since HIV-ITP patients with this Ab correct their platelet counts with dexamethasone (Dex), we examined the role of this steroid in this unique autoimmune disorder. Dex at therapeutic concentrations inhibited Ab-, A23187-, or PMA-induced platelet particle formation by inhibiting platelet PLA(2), 12-LO, and NADPH oxidase. The operational requirement of translocation of PLA(2), 12-LO, and NADPH oxidase components (p67 phox) from cytosol to membrane for induction of ROS was both inhibited and partially reversed by Dex in platelets. We conclude that (1) platelet particle formation can be induced by the generation of ROS; and (2) platelet PLA(2), 12-LO, NADPH oxidase, and cytosol membrane translocation, requirements for ROS production, are inhibited by Dex.

Hematopoiesis and stem cell renewal in long-term bone marrow cultures containing catalase.

Culturing mouse bone marrow in the presence of catalase dramatically alters hematopoiesis. Granulocyte output is initially increased 4- to 5-fold. This increase is transient and granulocyte production declines as immature (Sca-1+/LIN-) cells accumulate. One third of these immature cells have a phenotype (Sca-1+/c-Kit+) characteristic of hematopoietic stem cells. At 2 to 3 weeks there are greater than 200-fold more Sca-1+/c-Kit+/LIN- cells in treated cultures than in controls. This population contains functional stem cells with both short-term and long-term bone marrow repopulating activity. In addition to myeloid progenitors, this Sca-1+/LIN- population contains a large number of cells that express CD31 and CD34 and have an active Tie-2 promoter, indicating that they are in the endothelial lineage. After 3 to 4 weeks hematopoiesis in treated cultures wanes but if catalase is removed, hematopoiesis resumes. After 7 to 10 days the cultures are indistinguishable from untreated controls. Thus, protected from H2O2, hematopoietic progenitors multiply and become quiescent. This sequence resembles in vivo development in normal marrow. These results make it clear that peroxide-sensitive regulatory mechanisms play an important role in controlling hematopoiesis ex vivo and presumably in vivo as well. They also indicate that manipulation of the peroxide levels can be used to enhance the growth of hematopoietic stem cells in culture.

Role of molecular mimicry to HIV-1 peptides in HIV-1-related immunologic thrombocytopenia.

Patients with early HIV-1 infection develop an autoimmune thrombocytopenia in which antibody is directed against an immunodominant epitope of the beta3 (glycoprotein IIIa [GPIIIa]) integrin, GPIIIa49-66. This antibody induces thrombocytopenia by a novel complement-independent mechanism in which platelets are fragmented by antibody-induced generation of H2O2 derived from the interaction of platelet nicotinamide adenine dinucleotide phosphate (NADPH) oxidase and 12-lipoxygenase. To examine whether sharing of epitope between host and parasite may be responsible for this immunodominant epitope, we screened for antibody-reactive peptides capable of inhibiting platelet lysis and oxidation in vitro, using a filamentous phage display 7-mer peptide library. Fourteen of these phage-peptide clones were identified. Five shared close sequence similarity with GPIIIa49-66, as expected. Ten were molecular mimics with close sequence similarity to HIV-1 proteins nef, gag, env, and pol. Seven were synthesized as 10-mers from their known HIV-1 sequence and found to inhibit anti-GPIIIa49-66-induced platelet oxidation/fragmentation in vitro. Three rabbit antibodies raised against these peptides induced platelet oxidation/fragmentation in vitro and thrombocytopenia in vivo when passively transferred into mice. One of the peptides shared a known epitope region with HIV-1 protein nef and was derived from a variant region of the protein. These data provide strong support for molecular mimicry in HIV-1-immunologic thrombocytopenia within polymorphic regions of HIV-1 proteins. A known epitope of nef is particularly incriminated.

Role of endogenous thrombin in tumor implantation, seeding, and spontaneous metastasis.

Tumor/host-generated thrombin (endogenous thrombin) was investigated with tumor growth and metastasis experiments in mice by the use of hirudin, a highly potent specific inhibitor of thrombin. Pretreatment with hirudin inhibited tumor implantation in nude or syngeneic mice, following subcutaneous injection of 2 human and 2 murine tumors. Hirudin induced a considerable lag period in the appearance of tumor growth, compared with phosphate-buffered saline (PBS) treatment, but had no effect on established tumor nodule growth in vivo or on tumor growth in vitro. Hirudin treatment induced central necrosis of the tumor nodule compared with no effect with PBS treatment. Greater protection was noted with longer duration of treatment. Tumor seeding into blood was examined with green fluorescent protein (GFP)-labeled tumor cells. Hirudin inhibited seeding into the blood as well as systemic organs which varied from complete protection to 15- to 32-fold in the blood and 17- to 395-fold in the lung. Hirudin inhibited spontaneous metastases from subcutaneously implanted tumor by reducing the number of tumor nodules in the lungs. Mouse survival in animals injected subcutaneously with highly aggressive 4T1 cells revealed 5 of 5 deaths of PBS-treated animals on day 40 compared with no deaths with hirudin treatment, with prolongation of survival with hirudin treatment of 16 days to more than 31 days. Thus, endogenous thrombin contributes to tumor implantation, seeding, and spontaneous metastasis. A potent antithrombin agent should be of clinical benefit to patients with cancer.

Thrombin induces increased expression and secretion of angiopoietin-2 from human umbilical vein endothelial cells.

Angiogenesis is required for tumor growth and metastasis. It has recently been suggested that thrombin is a potent promoter of angiogenesis. We therefore examined the possibility that thrombin could be inducing the expression of angiopoietin-2 (Ang-2), necessary for remodeling. Human umbilical vein endothelial cells were incubated with or without thrombin (1 U/mL) for 1 to 24 hours and then examined for messenger RNA (mRNA) by Northern analysis. Enhanced mRNA expression (about 4-fold over baseline) was noted at 4 hours. Enhanced expression of Ang-2 mRNA was secondary to enhanced transcription (about 4-fold), with no effect on stabilization. Enhanced Ang-2 mRNA transcription was inhibited by H7 and PD98059, indicating the requirement of serine/threonine kinases as well as the mitogen-activated protein kinase pathway. Up-regulation of mRNA was associated with enhanced Ang-2 protein synthesis and secretion as assayed by immunoblot. Thrombin-induced secreted Ang-2 inhibited the binding of recombinant (35)S-Ang-1 to its Tie-2-Fc receptor, demonstrating functionality. Hirudin reversed this effect, demonstrating thrombin specificity. Thus, thrombin-induced tumorigenesis and metastasis is associated with enhanced Ang-2 protein synthesis and secretion via enhanced transcription of Ang-2. This could help explain how thrombin promotes angiogenesis.