Ovarian cancer, the coagulation pathway, and inflammation

The history of a known association between coagulation and cancer dates back to 1865, when Armand Trousseau observed that patients who presented with idiopathic venous thromboembolism frequently had an underlying occult cancer and vice versa [9]. The association has only recently become more apparent. Previously, researchers focused on how tumors activate blood coagulation and how to overcome it. However, the underlying mechanism by which coagulation factors promote tumor cell growth, invasion, metastasis, and angiogenesis has become a hot topic in the field of cancer research. Some coagulation factors that displaya role in tumor progression have been reviewed [10]. The most frequent reports on coagulant proteins and cancer interactions include factor III (tissue factor; TF) [11], TF-factor VIIa [12, 13], factor Xa [14], factor IIa (thrombin)-factorII receptors (also called proease-activated receptors (PARs) [15], and factor XIIIa-factor Ia (fibrin) (Table 1) [16]. Two cascades, one intrinsic and the other extrinsic, lead to the formation of a fibrin clot. Although they are initiated by distinct events, the two cascades converge on a common pathway that includes thrombin, PARs, and fibrin and lead to clot formation. The intrinsic cascade, which is initiated when contact is made between blood and exposed endothelial cell (EC) surface, requires the clotting factors VIII, IX, XI, and XII. Also required are the proteins prekallikrein and high-molecular-weight kininogen and Ca²⁺ and phospholipids secreted from platelets. The extrinsic coagulation cascade is initiated at the site of injury in response to the release of TF. Activated factor X is the site at which the intrinsic and extrinsic coagulation cascades converge. TF and factor VIIa contribute to the extrinsic cascade and possibly to the development of cancer. Other factors from the intrinsic pathway, such as factors XI and XII, have not yet been directly implicated in cancer progression. Our recent analysis of the gene expression profile in EOC has revealed increased levels of factor XI and factor XII transcripts in the peritoneum of EOC patients [3]. These findings suggest that the intrinsic cascade might promote ovarian cancer cell metastasis in the peritoneal cavity. We also showed that the serine proteinase inhibitor D1 (SERPIND1), which inhibits plasmin, tissue kallikrein, and factor XIa, were downregulated [3].

Blood coagulation cascades can be activated by different mechanisms and to different levels in cancer patients. The alterations range from subtle abnormalities in laboratory tests to clinically overt thrombosis and disseminated intravascular coagulation [17]. Up to 50% of all cancer patients and 90% of those with metastases exhibit hemostatic abnormalities [17]. These abnormalities may be reflected in the dominance of the tumor cell-associated procoagulant pathway, which leads to thrombin generation and hypercoagulation. Similar observations were made using in vitro ovarian cancer cells for the coagulation process. Ovarian cancer cells may also express TFs and the other coagulation components that generate local thrombin, as indicated by the conversion of fibrinogen to fibrin [18]. Fibrin has been found on ECs and the surface of ovarian tumor cells and nodules [18]. Coagulation activation in malignancy may be triggered by direct or indirect mechanisms. Direct activation of blood coagulation by the induction of thrombin may occur through the activity of tumor cell procoagulation, whereas indirect activation may occur through the production of tumor-associated cytokines that trigger TF production by host macrophages (MAs) or ECs. The coagulation pathway components may contribute to tumor cell proliferation, invasion, and metastasis [10], although these alterations could also be a consequence of advanced disease [19].

The formation of new cellular masses at an organ site creates an obvious extra demand for oxygen, growth factors, and metabolites and increases the need for tumor-associated blood vessels. The stimulators and inhibitors of angiogenesis that can regulate new vessel formation include VEGF, acidic and basic FGF, hepatocyte growth factor, and TGF. Angiostatin, endostatin, and antiangiogenic antithrombin might function as inhibitors of angiogenesis [70].

Angiogenesis can be mediated by either coagulation-dependent or-independent activation pathways [71]. Coagulation-dependent pathways always involve activation of the TF receptor to its ligand, followed by production of thrombin and ensuing clot formation. Coagulation-independent pathways appear to involve phosphorylation of the cytoplasmic domain of the TF receptor and subsequent downstream signaling events that occur independently of thrombin production or clot formation. Both pathways can contribute to angiogenesis and tumor progression. Increased expression of TF in tumors may contribute to angiogenesis in part by increasing VEGF protein expression and downregulating the expression of the antiangiogenic protein thrombospondin [11]. For example, TF-positive colorectal tumors have higher levels of microvessel density and VEGF expression than TF-negative colorectal tumors do [67]. Colocalization of TF and VEGF mRNA and protein has been demonstrated in breast cancer, malignant glioma, and adenocarcinoma of the lung [72]. In the coagulation-dependent pathway, the interaction of TF and factor VIIa induces Ca²⁺ oscillations and changes in gene expression, and the formation of the TF-factor VIIa complex leads to the activation of the mitogen-activated protein kinase (MAPK) pathway [73], which is a major inducer of VEGF expression. The catalytic activity of the TF-factor VIIa complex contributes to the activation of factor X to factor Xa and subsequently to thrombin. Factor Xa and thrombin) can also exert signaling activities through PAR-2 and PAR-1, respectively. Factor Xa activated by tumor cells may trigger PAR-2 expression as well as formation of TF-factor VIIa complexes on ECs and may induce intracellular signals; however, factor Xa exerts the most critical effect in the TF-factor VIIa-factor Xa complex [74].

In the clotting-independent pathway, the cytoplasmic domain of TF activates protein kinase C-dependent signaling, in contrast to the ligand-binding extraplasmatic tail of TF, which upregulates the synthesis of VEGF in response to different stimuli [75, 76]. The cytoplasmic tail of TF appears to regulate non-clotting-dependent mechanisms such as cytoskeletal reorganization, vascular remodeling, angiogenesis, and cellular metastasis. Results from a recent study provided an alternative explanation whereby TF can promote angiogenesis by a clotting-independent mechanism [77]: the genetic deletion of the TF cytoplasm domain involves the loss of negative regulatory control with PAR-2 upregulation. Expression of TF and PAR-2 and the function of TF-PAR-2 signaling pathway require the presence of both angiogenic growth factors and inflammatory cytokines. Inflammatory cytokines produced by monocytes (MO) are important for angiogenesis and collateral growth vessels. Colocalization of upregulated PAR-2 with phosphorylated TF occurs only in neovessels. The phosphorylation of TF appears to be the mechanism that switches off the negative regulatory control that promotes pathologic PAR-2-dependent angiogenesis [77].

However, TF-PAR-2 signaling selectively synergizes with the PDGF isoform BB but not with VEGF, basic FGF, and the PDGF isoform AA. PDGF-BB is available through release from activated platelets in the context of local coagulation or by synthesis from sprouting ECs. PDGF-BB also has a role in lymphoangiogenesis. PDGF-BB stimulated MAPK activity and cell motility of isolated lymphatic ECs in vitro and potently induced the growth of lymphatic vessels in vivo [78]. Expression of PDGF-BB in murine fibrosarcoma cells induced tumor lymphoangiogenesis, leading to enhanced metastasis in lymph nodes [78].

The cascade of TF-factor VIIa to factor Xa and subsequently to thrombin can be inhibited by TF pathway inhibitor (TFPI), which occurs when the TF-factor VIIa complex is combined with factor Xa and cell-bound TFPI. Once this stable quaternary complex is formed, it is translocated to and internalized in caveolae (i.e., small vesicles of the plasma membrane) [79]. Produced by ECs and tumor cells, most TFPI is expressed on the cell surface, although it can be detected peripherally in plasma [80]. In tumor cells, the diminished expression of TFPI could result in activated factor Xa and increase factor Xa-PAR-2 signaling [81]. Results from in vitro and in vivo studies have suggested that therapeutic strategies that target an increase in the expression of TFPI could inhibit tumor angiogenesis [82], growth [83], and metastasis [53, 84].

Another inhibitor of the TF-dependent pathway of blood coagulation that was recently identified is TFPI-2, which, in contrast to TFPI, inhibits the TF-factor VIIa complex but not factor Xa. TFPI-2 also inhibits various proteases [85]. Its expression was lower in laryngeal, breast, gastric, colon, pancreatic, renal, and endometrial tumor tissues and glial neoplasms than in normal counterpart tissue [85]. Some studies have provided evidence of the antiangiogenic and antimetastatic activity of TFPI-2 [86]. Epigenetic inactivation of TFPI-2 also contributes to the progression of pancreatic ductal adenocarcinoma. TFPI-2 mRNA was undetectable in many pancreatic cancer cell lines and in primary pancreatic ductal neoplasms. Hypermethylation of its promoter CpG island decreased TFPI-2 expression. However, expression of the TFPI-2 gene in pancreatic cancer cells has resulted in markedly suppressed proliferation, migration, and invasive potential in vitro [86]. TFPI-2 exerts antitumor effects through antiangiogenesis and antimetastatic mechanisms [87, 88].

The role of TF-factor VIIa in metastasis has been extensively investigated. Some recent studies have revealed that TF can promote the invasion and metastasis of the A2780 human ovarian cancer cell line through the TF-factor VIIa pathway. In turn, TF-factor VIIa can upregulate the transcription of uPA receptor expression and enhance tumor invasion and metastasis [63, 64]. Overexpression of TF of up to a 1,000-fold greater level than normal is characteristic of metastatic tumor cells relative to nonmetastatic cells [89]. Inhibition of TF through antibodies or active site-blocked factor XIIa interferes with experimental metastasis [90], but how TF on tumor cells contributes to tumor metastasis is unclear. TF-induced metastasis requires the participation of the cytoplasmic tail of TF and the assembly of an active TF-factor VIIa complex, which is thought to target G protein-coupled PAR-2. Factor VIIa induces the activation of p21^Ras and p42/p44 MAPK, which further induces the activation of the protein kinase B pathway in TF-expressing cells and increases protein synthesis [91]. Both the p42/p44 MAPK and protein kinase B pathways are able to inhibit apoptosis. Inhibition of apoptosis may be related to adhesion-independent survival and thus may be linked to metastasis. In one study, factor VIIa inhibited cell death and caspase 3 activation in Baby Hamster Kidney cells induced by serum-deprivation and loss of adhesion (lack of integrin signaling) when transfected with TF genes [14]. The effects of factor VIIa on caspase 3 activity are sensitive to inhibitors of the phosphatidylinositol 3'-kinase and P42/P44 MAPK pathways; thus, factor VIIa appears to be a survival factor for TF-expressing cells and seems to mediate the survival of tumor cells during metastasis [14, 92].

Another mechanism that links TF effects to metastasis is mediated by the overexpression of growth factors and proteins related to cellular reorganization. Interleukin (IL)-8, a member of the CXC chemokine (CXCL8) and a chemoattractant for neutrophils and lymphocytes, can act multifunctionally to induce tumor growth and metastasis, and IL-8 overexpression can occur through TF-FVIIa stimulation. In contrast, FXa and thrombin cannot upregulate IL-8 expression. IL-8 leads to increasing cell migration and invasion, and these effects are attenuated when binding of factor VIIa to TF is prevented [13]. In ovarian cancer studies, IL-8 has been suggested to be involved in the formation of ascites [93, 94]. In our peritoneal gene transcript profile, we showed that IL-8 connects a number of pathways that might be linked to the pathogenesis of cancer by cellspace analysis (Figure 1) [3]. IL-8 may induce other immune cell-related modulators, such as the IL-13 receptor, vascular cell adhesion molecule 1, chemokine receptor 1, and other molecules associated with inflammatory function [95]. It is known that the inflammatory response is involved in the progression of tumors and that IL-8 might amplify inflammation in the tumor microenvironment to promote tumor cell proliferation, angiogenesis, and metastasis.

Thrombin is a potent activator of angiogenesis and could be the underlying mechanism for the promotion of tumor progression after thrombosis. However, evidence suggests that angiogenesis induced via thrombin is mediated via PAR and is independent of fibrin formation, which can be modulated without interfering with blood coagulation [98]. Thrombin also stimulates the chemotaxis of inflammatory cells and thus possibly promotes tumor angiogenesis indirectly [98]. Furthermore, thrombin could induce several cellular effects on ECs, thereby contributing to the angiogenic cascade.

There are two possible mechanisms by which thrombin may promote angiogenesis: (1) Thrombin at physiologic concentrations can inhibit ECs' ability to adhere to collagen type IV or to laminin [99], even after short exposure of ECs to thrombin. The result is reversible and is mediated specifically through the thrombin receptor. The result is detachment and migration of ECs from basement membrane components, the initial step in the activation of normally quiescent ECs. (2) Thrombin upregulates VEGF expression via PAR activation on ECs or tumor cells [100] and through release of platelets [101]. Thrombin antagonizes any blood vessel-stabilizing effects of angiopoietin 1 by production of angiopoietin 2 [102, 103]. The angiogenic effects of thrombin include causing the release of basic FGF from ECs and the ECM [104]. Thrombin can also markedly stimulate expression of the VEGF receptor (kinase insert domain-containing receptor and flt-1). The activation of protein C kinase and MAPK signaling pathways was confirmed in the process of VEGF and VEGF receptor production [102]. Thrombin may thus contribute to new vessel formation by providing the pathologic microenvironment for tumor cell implants and proliferation; VEGF production through thrombin activation could facilitate this process. VEGF also increases the permeability of existing and new capillary vessels and results in plasma protein leakage and the development of a proangiogenic matrix [105]. A high expression level of VEGF allows the formation of malignant ascites. Thus, the thrombin-PAR system could play a role in the progression of ascites in EOC patients by inducing VEGF production.

In our peritoneal transcript profile study, we found that PAR-1 was upregulated in peritoneal tissue, which suggested that this receptor might be involved in EOC pathogenesis, invasion, and metastasis [3]. Overexpression of PAR-1 has been reported in malignant invasive melanoma [106] and breast cancer in vivo [96] and in breast cancer cell lines [107]. The increase in growth and metastatic potential of tumor cells may be partly attributable to the proangiogenic effects of thrombin. By mobilizing adhesion molecules, such as α_IIβ₃ integrin [108], P-selectin [109], and CD40 ligand [110], to the cell surface, thrombin enhances adhesion between tumor cells, platelets, ECs, and ECM and contributes to tumor progression. Thrombin also triggers the release of growth factors, chemokines, and ECM proteins that could promote the proliferation and migration of tumor cells. Facilitation of the metastatic activity of the thrombin-PAR complex has been demonstrated in vivo with experimental lung cancer [111], colon cancer [112], and breast cancer [15, 96]. The principal thrombin receptor, PAR-1, has been implicated in promoting these effects. Most of the cellular effects elicited by thrombin are mediated through the activation and subsequent signal transduction cascades of members of the PAR family. Likewise, thrombin and PAR inhibitors, such as argatroban [113], activating protein 1 [112], and small interfering RNAs for PAR-1 [114], have been shown to prevent tumor migration and metastasis. Thrombin can also facilitate trans-basement membrane migration by ECM remodeling after activation of collagen type IV-degrading enzyme, MMP-2, and α_vβ₃ integrin. However, an immunostaining study on EOC found that focal adhesion kinase (FAK), a major tyrosine phosphorylated protein of adherent cells, was abundantly expressed in invasive EOC but not in normal ovarian tissues, which is consistent with a high level of PAR1 expression [97]. Possibly, PAR1 in malignant ovarian carcinoma transmits signals leading to the phosphorylation of FAK and thereby to alterations in integrin expression. Increased expression of α_vβ₃ integrin was also detected in our peritoneal profile study [3]. Anticoagulants such as heparin and warfarin have been shown to significantly decrease experimentally induced pulmonary metastasis in vivo [115], and a similar decrease in tumor cell metastasis was observed in vivo with hirudin, a highly specific inhibitor for thrombin [15]. In another study, hirudin inhibited tumor implantation and growth of four different tumor cell lines (MDA-MB-231 human breast cancer, 4T1 human breast cancer, A549 human non-small cell lung cancer, and B16F10 murine melanoma) in nude and syngeneic mice by inhibiting spontaneous seeding of tumors through the blood to susceptible organs, including the lungs, liver, spleen, and lymph nodes [15]. Overall, the interaction of endogenously generated thrombin and host thrombin receptors enhances tumor angiogenesis by enhancing the expression of some genes that may be associated with an invasive profile, and this interaction involves activation of PAR and other tumor genes.

At least six signal transduction mechanisms based on thrombin-PAR systems have been proposed to have a role in invasion and metastasis: (1) PAR may induce phosphorylation of FAK and integrin expression [97, 116], thereby contributing to the modulation of cytoskeletal reorganization. Changes in cell shape result from changes in actin stress fibers and focal adhesion and are mediated through the Rho family of GTPases [117]. Rho-Ras proteins downstream of PAR-1 are intermediates of the extracellular signal-regulated kinases 1 and 2 (ERK1/2)-MAPK pathway and are implicated in cancer progression [118]. (2) Loss of activator protein 2 (AP-2) results in the increased expression of PAR-1, which subsequently contributes to the metastatic phenotype of cancer by upregulating the expression of angiogenic molecules, proteases, and adhesion molecules, which are involved in tumor invasion and metastasis [119]. This upregulation is based on MAPK signaling. AP-2 is a 52-kDa protein transcription factor mapped to the short term of chromosome 6, which is a critical regulator of gene expression during mammalian development, differentiation, and carcinogenesis. Loss of AP-2 may contribute to the development and progression of many different types of cancers, including melanoma and prostate, breast, and colorectal carcinoma [119]. AP-2 is a negative regulator for PAR-1 signaling, and functional AP-2-binding elements have been identified in some genes, including c-erbB2, VEGF, TGF-β, IGFBP-2, E-cadherin, and c-myc [119]. (3) Overexpression of flotillin-2, a highly conserved caveolae/lipid raft-associated protein, is associated with upregulation of PAR-1. Binding of flotillin-2 to PAR-1 might stabilize it by blocking inactivation or degradation of an activated PAR-1. Cytoplasmic flotillin-2 might also modulate PAR-1 trafficking of the cell or stabilize PAR-1 mRNA, which could promote PAR-1-induced constitutive signaling through the MAPK or other signal transduction pathways [120]. (4) PAR-1 cooperates with α_vβ₅ integrin to promote cytoskeletal reorganization. Activation of PAR-1 increased the phosphorylation of FAK and paxillin and induced the formation of focal contact complexes. The activation of PAR-1 may foster tumor cell invasion via a mechanism involving cooperation with the α_vβ₅ integrin [116]. (5) The increased and persistent expression of PAR-1 is induced by deregulated defective receptor trafficking, which is correlated with the failure to efficiently downregulate activated PAR-1 by internalization and lysosomal sorting [121]. (6) Coexpression of PAR-1 and PAR-2 has been observed on some tumor cells and the cells of the tumor microenvironment, such as fibroblasts, platelets, and ECs. Simultaneous activation of PAR-1 and PAR-2 by thrombin is required to cause the chemokinetic effect. Thrombin-induced activation of PAR-2 could play a subtle role in signaling induced by the thrombin-PAR-1 complex and stimulate the motility of metastatic tumor cells [122].

Platelets can act as bridges between ECs and lymphocytes or MOs/MAs, and thrombin activation of platelets can enhance hematogenous metastasis [123]. In addition, platelet activation and fibrin formation are both important mechanisms by which tumor cell-related "procoagulant activity" promotes metastasis. Platelet activation induced by thrombin is another possible mechanism for metastasis, though some studies have shown that knockout of the PAR-1 and PAR-2 gene of mice does not affect thrombin signaling in mouse platelets but might attenuate signals in mouse ECs [124, 125].

Fibrin-induced neovasculation is based on clotting-related mechanisms that involve platelet activation and fibrin deposition. Cross-linked fibrin has been found in different human malignant tumors. It is present in the endothelium of angiogenic vessels of invasive cancer specimens but not in vessels of benign tumors [134]. Fibrin can bind to inflammatory cells or to tumor cells and is deposited around tumor cells as scaffolding that promotes further angiogenesis. Fibrinogen and fibrin fragments, such as fragment E, have been shown to stimulate angiogenesis both in vitro and in vivo [135]. The presence of the fibrin degradation product D-dimer is significantly associated with a poor prognosis in some cancer patients [136, 137].

The binding of ECs to fibrin with the involvement of the adhesion molecule vascular endothelial cadherin may be necessary for capillary tube formation, a critical step in angiogenesis. The fibrin matrix that develops around tumors provides a provisional proangiogenic environment that supports vessel formation and stimulates EC proliferation and migration [138]. The fibrin matrix can promote a proangiogenic response by upregulating the expression of α_vβ₃ integrin receptor to facilitate EC migration and capillary formation [139]. The α_vβ₃ integrin provides survival signals to ECs during their interaction with fibrin. Tumor-containing tissue has revealed increased deposition of fibrin stimulated by VEGF-induced vessel leakiness to sustain the proangiogeneic environment [7]. The fibrin matrix also stimulates the production of VEGF, basic FGF, IGF1, and IL-8 to promote an autocrine procoagulant loop by inducing TF expression in ECs [61]. The expression of IGF1 and α_vβ₃ integrin genes is upregulated in the peritoneum of EOC; therefore, fibrin may stimulate the production of these genes [3].

Two important regulatory proteins in the coagulation cascade, serine proteinase inhibitor D1 (Serpins, also called heparin cofactor II – HCII) and endothelial protein C receptor (EPCR), were first detected as an upexpression in a study we conducted [3], which suggests that there is value in studying them further. Serpins are a protein superfamily of which many members possess potent activity as serine proteinase inhibitors [140]. Another very important family of serpins is antithrombins. Both HCII and antithrombins can inhibit the blood coagulation proteinase thrombin.

HCII is a 480-amino-acid, single-chain glycoprotein with a molecular weight of about 66 kDa that is synthesized by the liver and circulates in the plasma. The human HCII gene is located on chromosome 22q11, spans 15.8 kb, and includes 5 exons [141]. When HCII interacts with heparin and dermatan sulfate, the potential inhibition of thrombin can be increased by more than 1,000-fold in vivo [142]. HCII inactivates thrombin by forming a stable, bimolecular complex in plasma. HCII does not inhibit other proteases involved in coagulation or fibrinolysis. It can also participate in wound healing by regulating the mitogenic and chemotactic activities of thrombin [143]. Two recent clinical research studies demonstrated that HCII is a very effective factor in combating heart and vascular disease, which it does by inhibiting the action of thrombin [144, 145]. HCII is inactivated after cleavage of its reactive site (Leu-444 through Ser-445) by elastase from leukocytes during the inflammation process without forming the stable thrombin-HCII complex [146].

HCII has a physiological role in the inflammation response. An active peptide from the amino-terminal region (corresponding to Asp-39 through Ile-66) with chemoattractant action for both neutrophils and MOs is proteolyzed by neutrophil elastase, and another leukocyte migration peptide is derived from dodecapeptide from Asp-49 through Tyr-60 of HCII. HCII-neutrophil proteinase products appear to have a role in the local inflammatory response [147]. Leukocytes and MOs/MAs are important cellular components of tissue injury, wound healing, and the microenvironment of EOC tumors [6]. The detection of upregulated HCII in the peritoneum and stroma of EOC patients is interesting because the degradation of HCII can produce a chemoattractant peptide for MAs [3]. MOs/MAs have multifunctional roles in EOC progression by releasing cytokines and chemokines, and HCII may help induce MA migration from the endothelium to extravascular tissue.

There is substantial evidence of infiltrating immune cells in EOC and the peritoneal environment. In a previous study, we showed that about 70% of T cells within EOC tumors were mononuclear cells [165]. Results from other researchers' experiments have suggested that the presence of these T cells is associated with an antigen-driven immune response [166‐168]. However, there is little evidence to suggest there is a chronic cellular adaptive immune response in vivo because interferon γ message is absent or present at low levels [169]. Expression of the CD3ζ chain was absent or poor in another study [170]. A recent study showed longer overall and progression-free survival among EOC patients with intratumoral T cells than among those lacking these cells [168]. In a more recent study, CD4+CD25+ regulatory T cells present in EOC tumors appeared to be associated with poor patient survival. Regulatory T cells preferentially move to and accumulate in tumors and ascites but rarely enter the draining lymph nodes in later cancer stages [171]. Large numbers of MOs/MAs are also present in ascites, where they may make up 50% or more of mononuclear leukocytes, whereas the proportion that consists of T lymphocytes is usually less than 40% [165]. In recent preliminary studies, we found that pelvic peritoneal tissue biopsied from patients with advanced-stage EOC had a higher proportion of MOs/MAs than T cells even in the absence of tumor involvement (unpublished data).

Cytokines, chemokines, adhesion molecules, and components of the ECM may contribute to a tissue environment that supports tumor proliferation and invasion. Chemokines and certain cytokines may facilitate the migration of immune cells, including T cells and MAs, into the tumor environment (Tables 3 and 4). A network of CC (cysteines with no intervening amino acid) and CXC chemokines has been found in solid tumors and in ascites [172]. For example, CCL2, one of the ligands for CCR2, localizes to epithelial areas of the tumor [173], and its expression appears to correlate with the numbers of lymphocytes and MAs at this site [174]. CCL5 localizes with tumor-infiltrating leukocytes, and CCL5 concentration may be associated with CD8+ T cell infiltration. Chemokines CCL2, CC8, CCL4, CCL5, CCL8, CCL22, CXCL2, and CXCL12 have all been found in increased amounts in ascitic fluid from EOC patients, and the presence of CCL5 in ascitic fluid has been associated with the number of T cells in ascites [175].

Adhesion molecules may also be involved in the implantation of cancers cells onto the peritoneal lining. For example, E-cadherins and P-cadherins are expressed during EOC progression and may facilitate peritoneal invasion because mesothelial cells express cadherin-binding catenins [4]. Upregulation of discoidin domain receptor 1, claudin 3, and epithelial cell adhesion molecule occurs early in the development of EOC [176]. Soluble intracellular adhesion molecule 1 concentrations are elevated in patients with ovarian malignant tumors but do not correlate with clinical status[177].

Depending on their degree of expression in the tissue microenvironment, coagulation and inflammation might influence each other. Furthermore, components generated from both the coagulation cascade and an inflammatory response might be involved in heart disease, as recently reviewed [178‐181]. Normal hemostasis is initiated when the blood vessels rupture, allowing blood cells to interact with extravascular cells and the ECM [182]. In one study, endotoxin, TNF-α, and IL-1α induced TF expression, primarily on MO/MA [58]. Blood clotting might thus be initiated when inflammatory cytokines and endotoxin induce the synthesis of TF on migrating leukocytes [183]. In this way, immune cells could help initiate the coagulation cascade through damaged tissue or cytokine production [184] (Table 5).

The complement system may also contribute to hemostasis by activating the membrane attack complement pathway, thereby contributing to phosphatidylserine expression on the outer membrane of cells [185]. Expression of phosphatidylserine on the outer cell membrane is necessary for the effective initiation and amplification of the coagulation cascade. Our microarray analysis of the peritoneum of EOC patients showed increased expression of the C2 component [3]. In contrast, expression of complement HF1, which blocks the membrane attack complement pathway, was increased and might have shifted the complement activation pathway toward a proinflammatory response with the recruitment of MO/MA. MO/MA are also an important source of complement components.

Chemokines can influence coagulation by activating platelets indirectly. Three main chemokines are involved in this process, CXCL12 (the ligand for CXCR4), CCL17 (the ligand for CCR4), and CCR22 (the ligand for CCR4) [186]. Platelets are an important link in the inflammation and coagulation processes. Inflammatory mediators, cytokines, and chemokines do not increase platelet production, but the platelets generated are more thrombogenic and thus are more sensitive to platelet agonists such as thrombin. Platelets contain high concentrations of the proinflammatory mediator CD40 [187] and certain chemokines [188]. When CD40 and these chemokines are released, they can induce TF synthesis and increase the production of inflammatory cytokines such as IL-6 and IL-8. Of the naturally occurring anticoagulants, protein C is the most negatively influenced by inflammation. Thrombomodulin and EPCR are both downregulated by inflammatory cytokines such as TNF-α [189] and IL-1 [190].

Thrombin has a variety of effects on cells that can enhance the inflammatory response. For example, it augments leukocyte adhesion and activation, which amplifies the inflammatory response, and induces endothelium and platelet activation. The activation of platelets can produce cytokines (IL-6) and chemokines (IL-8), which are involved in inflammatory responses [186]. Thrombin also acts as a direct mitogen for fibroblasts, inducing PDGF and TGF from platelets and ECs. PDGF and TGF-β are important cytokines for stimulating neovessel formation, and TGF-β may contribute to an environment conducive to tumor escape from immune system surveillance. Notably, PDGF and TGF-β 3 receptor expression were increased in our peritoneal transcript profile study [3]; this finding provides additional evidence that thrombin could foster tumor cell growth and proliferation via PDGF and TGF release from platelets and ECs.

Coagulation factors have a profound effect on tumor cell behavior in both in vivo and in vitro studies. These factors could enhance tumor cell proliferation, invasion, angiogenesis, and metastasis. Hence, targeting activated coagulation factors might provide a viable cancer treatment strategy.

Heparins are the most extensively used anticoagulants in clinics. In blood coagulation, unfractionated heparin and low-molecular-weight heparins potentiate the activity of antithrombin III, thus inhibiting the activation of coagulation factors II and X [191]. They also release TFPI, a physiologic inhibitor of the TF pathway that prevents pulmonary embolism and is used to treat deep vein thrombosis [192]. Retrospective and meta-analytic studies of deep vein thrombosis treatment have shown longer survival among cancer patients with thrombosis who were treated with unfractionated heparin and low-molecular-weight heparins than among patients treated without heparin [193‐198]. Thus, the use of anticoagulants might allow them to live longer.

There is a higher incidence of venous thromboembolism in patients with melanoma or small cell lung cancer. In 2003, a pilot study of enoxaprin, a low-molecular-weight heparin for the treatment of advanced melanoma was reported [199], but the clinical outcome was not clear. Then, a phase II trial of combination chemotherapy and anticoagulant therapy (docetaxel plus enoxaparin) was performed in chemotherapy-naive patients with metastatic non-small cell lung cancer [200]. The median time to progression was 5 months, and the median survival time was 11 months. The most frequent toxic effects were neutropenia and asthenia; no clinically significant bleeding or thrombotic events were observed. Treatment was well tolerated in patients with advanced small cell lung cancer, and the results suggested that enoxaparin could prolong the time to disease progression [200]. Both the oral anticoagulant warfarin and unfractionated heparin have been shown to prolong survival time among patients with small cell carcinoma [201].

In 2004, a prospective, randomized, controlled trial enrolled 385 cancer patients into two arms, placebo and dalteparin (another low-molecular-weight heparin). Types of cancer included breast, colorectal, ovarian, and pancreatic [202]. Thirty-four percent of the dalteparin group and 31% of the placebo group received chemotherapy alone; 8% of each group received radiation therapy alone. Estimated overall survival at 1, 2, and 3 years did not differ significantly between groups overall. However, the estimated overall survival among patients with a better prognosis at enrollment (55 patients in the dalteparin group and 47 patients in the placebo group) was significantly longer in the dalteparin group at 2 years (78% and 60%; P = 0.03) and at 3 years (55% and 36%; P = 0.03). In 2005, another study confirmed these antineoplastic effects of dalteparin [203]. During the 12-month follow-up period of the study, 602 patients with solid tumors and venous thromboembolism were randomly assigned to a dalteparin or coumarin-derivative treatment group. Among patients without metastatic disease, the probability of death at 12 months was 20% in the dalteparin group compared with 36% in the oral anticoagulant group (P = .03). In patients with metastatic cancer, no difference was observed in mortality between the treatment groups (72% and 69%, P = .46). However, the observed effects of dalteparin on survival were statistically significantly different between patients with and without metastatic disease (P = .02). Clinical trials are warranted to investigate these findings. The exact mechanism by which heparin mediates antitumor or antimetastatic activity is unknown but also merits further study.