Role of tissue factor in feto-maternal development: a xiphos

Authors


Guillermina Girardi, Department of Biology, York College – CUNY, 94-20 Guy R. Brewer Blvd, Jamaica, NY, USA.
Tel.: + 1 718 262 5277; fax: + 1 718 262 2369.
E-mail: guillerminagirardi@gmail.com

Abstract

Summary.  In this review, the dual role of tissue factor (TF) in pregnancy is described. On the one hand, TF is required for embryonic and placental development in a successful pregnancy, and on the other hand, pathologic expression of TF can lead to serious pregnancy complications in humans and mice. Human studies show increased TF levels in plasma, amniotic fluid and and/or placentas of abnormal pregnancies affected by miscarriages, preterm birth, or pre-eclampsia. Interestingly, using two mouse models, we found that blood-borne TF plays a crucial role in the pathogenesis of pregnancy complications. TF on neutrophils and monocytes is a critical mediator in trophoblast injury and embryo damage in pregnancy loss induced by antiphospholipid antibodies and in the antibody-independent CBA/J × DBA/2 model of miscarriages. Blockade of TF or genetic diminution prevented pregnancy complications, suggesting that TF may be a good target for therapy in patients with recurrent miscarriages, pregnancy loss, and pre-eclampsia. In addition, statins, which downregulate TF, may constitute a good therapeutic option for women with pregnancy complications. Clinical trials should be conducted to confirm these observations in women.

The goal of writing this short review is to discuss the role of tissue factor (TF) in fetal and placental development and pathology. I will discuss the function of TF in normal placentation and embryonic development, but I will particularly focus on the crucial effects of TF in trophoblast injury and pregnancy complications, our understanding of which has shown significant progress in recent years. I have focused my comments on concepts that I believe are particularly important for understanding the hemostatic and non-hemostatic effects of TF in pregnancy. In recent years, research has shown that TF plays more than just the passive role of triggering the coagulation cascade. It is closely related to cell signaling processes involved in many cellular processes, such as growth, mobility, and cell activation. The complex of TF with factor VII is a key element in initiating the signaling events, either by directly signaling through TF or by activating nearby cell receptors.

Because of the dual role that TF plays in feto-maternal development, TF can be compared with a xiphos, a double-edged sword used by the ancient Greeks. TF is a crucial molecule in pregnancy, being required for normal placental and embryo development, but, at the same time, aberrant TF expression can cause placental injury and adverse pregnancy outcomes in mice and women. Mouse pregnancy is suitable for the study of mechanisms of abnormal human pregnancy, given the similarities in placental development [1]. Although the human and mouse placenta have certain differences, the association of TF expression with abnormal pregnancy outcomes is common to both species. Serious pregnancy complications, such as miscarriages, preterm delivery, and pre-eclampsia, have been associated with increased TF expression in humans and mice. In this review, recent insights gained from studies in mice will be presented and compared with the findings from human studies. I will particularly discuss the role of TF in inflammatory cells in placental injury and fetal death in different animal models of pregnancy complications, and I will discuss how these observations are correlated with the clinical scenarios.

TF and placental development

TF is a transmembrane glycoprotein and a member of the class II cytokine receptor superfamily [2]. TF, also called FIII or CD142, is not only the major cellular initiator of blood coagulation, but also contributes to inflammation [3]. Complexes of TF with FVIIa, FVIIa–FXa,, FXa and thrombin induce proinflammatory signals by activating protease-activated receptors (PARs) and inducing the expression of tumor necrosis factor (TNF)-α, interleukins, and adhesion molecules [4].

An incorrect synonym for TF is thromboplastin. Historically, thromboplastin was a reagent used in the laboratory to assay prothrombin times. Interestingly, this ‘reagent’ was extracted from human placenta. In 1945, Chargaff [5] demonstrated that human placenta has the ability to initiate the coagulation of blood because of ‘thromboplastic activity’, resulting predominantly from the presence of TF. In fact, the placenta is an organ rich in TF [6,7], with a concentration of 31 ng mg−1 (range: 25–37 ng mg−1) total protein, as compared with 0.004 ng mg−1 total protein measured in the plasma of the same women [8]. Interestingly, placental extracts have been used since ancient times to accelerate wound healing [9,10], and TF seems to be involved in this process, similarly to the wound-healing effects of TF in saliva [11,12].

The high levels of TF in placenta relative to those in blood plasma indicate that TF may be clinically significant in obstetrics. In all other vascular beds, the blood vessel endothelium is the principal gatekeeper between TF and blood, preventing inappropriate activation of the clotting cascade, but the hemochorial type of placenta lacks this protective barrier. In these placentas, found in humans, many other primates, and rodents, maternal blood is in direct contact with the trophoblasts, which are rich in TF [6,7]. However, the coagulation cascade is not activated. Increased TF pathway inhibitor expression may account for the inhibition of TF procoagulant activity during pregnancy. However, conflicting results have been reported concerning TF pathway inhibitor (TFPI) synthesis by trophoblasts [13,14]. Thus, we also need to consider the possibility that the presence of TF in placenta may have biological effects besides contributing to the maintenance of local placental hemostasis.

Recent studies have suggested that TF plays a non-hemostatic role in the development of blood vessels, cell signaling, tumor metastasis, and embryogenesis [15–17]. In 1996, the murine TF gene was inactivated [18–20]. Mice heterozygous for the inactivated TF allele were phenotypically normal. However, 90% of homozygous TF−/− embryos died in utero [18]. Inactivation of the TF gene resulted in abnormal circulation from the yolk sac to the embryo beyond embryonic day 8.5, leading to embryonic death. In addition, viteline vessels from null mice lacked smooth-muscle α-actin-expressing cells, which participate in vessel wall organization [21]. In surviving embryos, an absence of TF in the placenta leads to structural abnormalities in the labyrinthine zone, the principal site of hemotrophic exchange between the mother and the fetus [22]. These studies performed in genetically modified mice emphasize the role of TF in placental and embryonic development.

Human studies also highlight the role of TF in embryonic development. TF appears to be the only protein in the coagulation pathway for which a congenital deficiency has not been reported, suggesting a role of TF in embryonic development [23]. Luther et al. [24] studied the role of TF expression in mammalian embryogenesis. In these studies, they found that, in early stages of human and mice organogenesis, TF was expressed in various tissues. In contrast, increased TF expression was not associated with presence of the TF ligand FVII. Considering that, at this early stage of pregnancy, no passage of coagulation factors through the placenta is detected, the widespread distribution of TF in embryos and the absence of FVII suggest that TF might have functions in addition to that of an initiator of the coagulation cascade. Interestingly, during embryogenesis, the cellular localization of TF is not restricted to the cell membrane. Intracellular staining for TF is observed in different cells. The striking early expression of TF in the cardiovascular system and central nervous system suggests a possible role for TF in the early development of these systems [21]. In conclusion, in both humans and mice, TF seems to be an important morphogenic factor in embryogenesis.

Aberrant TF expression is associated with bad pregnancy outcomes in mice and women

Although TF is required for embryonic and placental development, aberrant expression of TF is associated with abnormal pregnancies in mice and women. Human studies have demonstrated that changes in TF expression in the endometrium can result in various pathologies, including, infertility, miscarriages, pre-eclampsia, and preterm labor [25]. Endometriosis is a common cause of infertility. Aberrant expression of TF by endothelial cells was observed in endometriotic lesions in women [25]. Consistent with these observations in humans, a recent study performed in an athymic mouse model of endometriosis demonstrated that targeting aberrantly expressed endothelial TF causes regression of endometriosis [26].

Erez et al. [27] reported a correlation of TF with preterm delivery, a serious pregnancy complication associated with high neonatal morbidity and mortality. High TF activity and decreased TFPI levels were observed in plasma from patients with preterm labor [27]. Similarly, these authors found increased TF levels and activity in amniotic fluid from patients with fetal demise [28]. TF levels were also found to be significantly increased in patients with pre-eclampsia [29–31]. Higher levels of TF were found in plasma and placentas from pre-eclamptic patients than in those with normal pregnancies [30]. Increased placental TF expression was correlated with increased TF mRNA placental levels, suggesting increased synthesis of TF by trophoblasts in pre-eclamptic placentas [31].

Most of these studies performed in humans reported increased levels of the protein TF, but no information regarding TF activity and/or activation of the coagulation cascade in these women is available. However, most of these studies concluded that increased plasma and placental TF levels may contribute towards the pathologic hypercoagulable state that leads to pregnancy failure. In addition, these studies do not provide information about TF being the cause or the consequence of bad pregnancy outcomes. Studies performed in mice identified a causative role for TF in pregnancy complications, and suggested that TF might also play a fibrin-independent proinflammatory role in the pathogenesis of pregnancy complications [32–34]. Recent studies showed increased PAR-2 expression in endothelial cells of women with pre-eclampsia [30]. Increased TF and PAR-2 expression suggest that coagulation-independent mechanisms may be involved in the pathogenesis of pre-eclampsia. Further studies should be performed in human samples in order to characterize the role of TF expression in pregnancy complications, as it is not clear whether TF is activating the coagulation cascade and/or triggering non-hemostatic signaling events.

In accord with the human studies, a strong correlation between TF expression and bad pregnancy outcomes was also observed in mice [32–34]. The fact that these mouse models recapitulate the findings observed in women indicate that they are appropriate tools with which to investigate the mechanisms responsible for bad pregnancy outcomes. Previous work in our laboratory has shown increased decidual TF expression in a mouse model of antiphospholipid antibody (APL)-induced pregnancy loss [32]. Furthermore, blockade of TF with a monoclonal antibody in wild-type mice or genetic reduction of TF prevented pregnancy complications in this model, highlighting the role of TF in fetal injury and pregnancy loss [32]. Increased placental TF expression was also described in CBA.J × DBA/2 mice [33]. In this model of antibody-independent pregnancy loss and growth restriction, blockade of TF with a monoclonal antibody also protected pregnancies [33]. Besides the similarities between human and mouse studies, some unique features were observed in mice. TF expression on inflammatory cells seems to play a crucial role in inflammation, trophoblast damage and fetal death in mice [33,34].

TF on inflammatory cells

Hemostasis is a much more complex process than the coagulation cascade described in the 1960s. The traditional notion of TF serving predominantly in hemostatic protection in the vascular bed has been challenged by the concept of blood-borne TF [35]. TF-containing neutrophils and monocytes in peripheral blood suggest that leukocytes constitute the main source of blood TF [35]. Monocytes and neutrophils can express TF in response to a variety of stimuli, including TNF, bacterial lipopolysaccharide, or complement component C5a [36,37]. TF is inherently thrombogenic, and may be involved in thrombus propagation at the site of vascular injury. However, TF expression on inflammatory cells has also been linked to cell activation. TF expression on inflammatory cells has been documented in chronic inflammatory conditions such as sepsis, atherosclerosis, and systemic lupus erythematosus [38–41]. TF on monocytes and synovial cells promotes leukocyte adhesion and transendothelial migration, potentiating inflammation in joints [42], whereas decreased TF activity abrogates the systemic expression of inflammatory mediators in animal models [43]. Our studies in mouse models of pregnancy complications highlight the importance of TF on inflammatory cells in trophoblast injury and fetal death [32–34].

TF in APL-induced placental damage and fetal loss: a proinflammatory molecule

TF expression is a characteristic feature associated with APL antibodies [44]. Mice that received APL exhibited strong TF staining throughout the decidua and on embryonic debris [32]. Because TF is the main initiator of the coagulation cascade, we expected to find increased fibrin deposition in deciduas from APL-treated mice. Surprisingly, neither an increase in fibrin staining nor thrombi were associated with increased TF staining, and abundant neutrophil infiltration was observed in deciduas from APL-treated mice [45,46]. Furthermore, anticoagulation with hirudin or fondaparinux was not sufficient to prevent pregnancy loss in this model [45]. Blockade of TF with monoclonal antibodies or genetic diminution of TF protected pregnancies in APL-treated mice, highlighting the causative and crucial role of TF in the pathogenesis of miscarriages induced by APL antibodies [32]. An important question needed to be answered. Is TF contributing to inflammation rather than coagulation in APL-induced pregnancy loss? To answer this question, we first decided to identify the source of TF in APL-induced pregnancy loss. To discern the role of trophoblast TF relative to that of TF from myeloid cells, we performed experiments in TFfloxed/floxed/LysM-Cre mice, which do not express TF on myeloid cells [32]. These mice were protected from APL-induced pregnancy loss, emphasizing the key role of TF in maternal myeloid cells [32]. Because monocytes and platelets are not required for APL-induced pregnancy loss (G. Girardi, unpublished observation), neutrophils are important mediators of fetal injury [46], and neutrophils from APL-treated TFfloxed/floxed/LysM-Cre mice do not express TF, we were able to conclude that TF expression on maternal neutrophils plays a causative and essential role in APL-induced fetal injury. We have previously demonstrated that neutrophils synthesize TF in response to complement activation [32]. These data are in accordance with studies published by Ritis et al. [37], who found TF mRNA in human neutrophils, but are in disagreement with the results published by Egorina et al. [47], who proposed that neutrophils do not synthesize TF but acquire it from other blood cells. We reported that, as a consequence of the interaction of complement split product C5a with its receptor (C5aR), TF synthesis and expression on neutrophils is increased [32]. The fact that TFfloxed/floxed/LysM-Cre mice treated with APL showed normal pregnancies and diminished decidual inflammation as compared with wild-type mice confirms that TF expression on neutrophils modulates the ability of the neutrophils to induce tissue injury. In fact, neutrophils from TFfloxed/floxed/LysM-Cre mice treated with APL showed lower generation of oxidants and less phagocytic capacity than neutrophils from APL-treated TFfloxed/floxed control mice [34]. Less free-radical-mediated oxidative damage in deciduas was also observed in TFfloxed/floxed/LysM-Cre mice treated with APL than in APL-treated TFfloxed/floxed control mice, which express TF, suggesting that TF modulates the oxidative burst in neutrophils and placental lipid peroxidation [34].

We then explored the mechanism by which TF contributes to inflammation and placental and fetal damage in APL-treated mice. Complexes of TF with FVIIa, FVIIa–FXa, FXa and thrombin induce proinflammatory signals by activating PARs [4,48]. PARs, notably PAR-2, are expressed by many cells that are crucially involved in innate and adaptive immunity, such as neutrophils and macrophages. Activation of PAR-2 leads to the production of various cytokines and chemokines that modulate the inflammatory response [49–51]. We hypothesized that TF-mediated signaling through PARs may promote inflammation, leading to trophoblast injury and pregnancy loss induced by APL antibodies. Increased TF and PAR-2 synthesis and expression were observed in neutrophils from APL-treated mice when compared with control antibody-treated mice [34]. To investigate whether PAR-2 signaling was involved in neutrophil activation, we studied neutrophils from APL-treated PAR-2-deficient (PAR-2−/−) mice. The genetic deletion of PAR-2 dramatically reduced neutrophil activation in APL-treated mice [34]. Reactive oxygen species (ROS) production and phagocytosis were significantly reduced in neutrophils from PAR-2−/− mice treated with APL as compared with APL-treated wild-type mice. We confirmed the role of TF–FVIIa–PAR-2 signaling in APL-induced neutrophil activation by using a specific monoclonal antibody, 10H10, that selectively blocks TF–VIIa signaling through PAR-2. Neither increased oxidative burst nor increased phagocytosis was observed in neutrophils from mice treated with APL and 10H10 [34]. Mice with a deletion in the cytoplasmic domain of TF that prevented signaling through PAR-2 were also protected from APL-induced placental damage [34]. Genetic deletion of PAR-2 and mutation in the cytoplasmic domain of TF reduced neutrophil activation and rescued pregnancies in APL-treated mice, demonstrating the importance of TF–FVIIa–PAR-2 interaction in neutrophil activation and fetal death in this model of miscarriage. Figure 1 summarizes the proposed mechanism for the role of TF in APL-induced pregnancy loss. APL antibodies induce complement activation [34]. C5a–C5aR interaction on neutrophils results in increased TF and PAR-2 expression, leading to neutrophil activation. Activated neutrophils release ROS and proteolytic enzymes, leading to decidual damage and fetal wastage. TF acts as an important proinflammatory molecule rather than a procoagulant mediator in APL-induced fetal injury [34]. Simvastatin and pravastatin reduce TF and PAR-2 synthesis and expression in neutrophils, diminish neutrophil activation and protect pregnancies in mice treated with APL antibodies [34].

Figure 1.

 Activation of neutrophils by the tissue factor (TF)–factor VIIa–protease-activated receptor (PAR-2) axis mediates placental damage and fetal death in mice treated with antiphospholipid (APL) antibodies. APL antibodies induce complement activation. C5a–C5aR interaction on neutrophils results in increased TF and PAR-2 expression, leading to neutrophil activation. Activated neutrophils release reactive oxygen species and proteolytic enzymes, leading to placental injury and fetal death.

TF increases the release of antiangiogenic molecules and affects placental development in mice

APL-induced pregnancy loss accounts for 25–30% of immunologically mediated miscarriages. We then decided to investigate whether TF might also be involved in the remaining majority of cases. We decided to study a model of miscarriage that did not require APL induction, but in which complement played a role [52]. We studied a mouse model of recurrent spontaneous miscarriage (CBA/J × DBA/2 mice) that shares features with human recurrent miscarriage and fetal growth restriction [33,52]. In this model, pregnancy loss is caused by the inability to correctly form placental blood vessels [33,52]. In response to complement split product C5a generation, macrophages infiltrate the uterus, releasing antiangiogenic factors that sequester vascular endothelial growth factor (VEGF) – a potent angiogenic molecule – and prevent the formation of the placental and embryonic blood vessels [33,52]. We found that TF is also an essential factor in placental and fetal injury in this model [33]. Increased TF expression and activity were found in placentas from CBA/J × DBA/2 mice as compared with control mice with normal pregnancies. Increased levels of the antiangiogenic factor sFlt-1, observed in CBA/J × DBA/2 mice, is commonly observed in patients with pre-eclampsia. Interestingly, we were able to translate the results obtained in mice to humans. We found increased TF staining in placental samples from patients with pre-eclampsia [33]. These studies are in accordance with human studies reported by other authors that showed increased TF expression in pre-eclamptic pregnancies [29–31].

In contrast to what we observed in the APL-induced pregnancy loss model, increased placental TF expression in CBA/J × DBA/2 mice was associated with increased fibrin deposition and diminished placental blood flow [33]. Increased thrombin–antithrombin complex levels were also observed in plasma from CBA/J × DBA/2 mice. Treatment with the anticoagulant hirudin, a direct inhibitor of thrombin, prevented placental damage and miscarriages, suggesting a role for thrombin in pregnancy complications in this model. The association of increased TF expression with increased thrombin generation and fibrin deposition suggests that activation of the coagulation cascade plays a role in the pathogenesis of placental damage in CBA/J × DBA/2 mice [33]. Further studies should be performed to study the possible interaction of thrombin with PAR-1.

In this model, we also found that blockade of TF protected pregnancies. Blockade of TF not only rescued pregnancies, but also diminished placental oxidative stress, suggesting that TF is a crucial effector in placental oxidative injury in CBA/J × DBA/2 mice. The fibrin deposition and diminished placental blood flow observed in these mice suggest that ischemia could be a possible source of ROS.

Monocytes are crucial cellular effectors of placental and fetal damage in this model [33,52]. In fact, monocyte depletion prevented pregnancy loss in the CBA/J × DBA/2 mice [33]. We also observed that TF is upregulated in circulating monocytes in the DBA/2-mated CBA/J females, raising an important question. Does TF expression on macrophages play a role in placental damage and pregnancy complications in the CBA/J × DBA/2 model? Our observation that TF is expressed in response to C5a and stimulates the release of sFlt-1 in monocytes answered this question. The antiangiogenic factor sFlt-1 released by macrophages sequesters VEGF, inducing inadequate placental development and perfusion, contributing to placental oxidative damage [33]. In addition, SM9-1 mouse trophoblasts incubated with sFlt-1 showed increased TF expression and oxidative stress [33]. Blockade of TF with a monoclonal antibody or inhibition of TF synthesis with pravastatin inhibited C5a-induced release of sFlt-1 in macrophages. TF inhibition diminished sFlt-1 release, restored placental blood flow, prevented placental oxidative damage, and rescued pregnancies, indicating the important role that TF plays in this model of immune-mediated pregnancy loss and suggesting that TF may be a good target for therapy [33]. TF plays a dual role in the pathogenesis of placental and fetal death in the CBA/J × DBA/2 model: by activating the coagulation cascade, and by activating inflammatory cells to release antiangiogenic factors [33]. The procoagulant and proinflammatory effects of TF in placental damage and fetal injury in the CBA/J × DBA/2 mice are illustrated in Fig. 2. TF initiates the coagulation cascade, compromising placental blood flow and inducing oxidative damage to trophoblasts. At the same time, TF activates the macrophages, inducing the release of sFlt-1. This potent antiangiogenic molecule prevents placental vessel formation, reduces blood flow, and induces oxidative damage. Increased placental oxidative damage initiated by the procoagulant and proinflammatory effects of TF is responsible for fetal injury and bad pregnancy outcomes in DBA/2-mated CBA/J females.

Figure 2.

 Mechanism of fetal death and growth restriction in CBA/J × DBA/2 mice. Tissue factor (TF) initiates the coagulation cascade, compromising placental blood flow and inducing trophoblast oxidative damage. At the same time, TF activates macrophages, inducing the release of sFlt-1. This potent antiangiogenic molecule prevents placental vessel formation, reduces blood flow, and induces oxidative damage. Increased placental oxidative injury initicated by the procoagulant and proinflammatory effects of TF is responsible for fetal injury and bad pregnancy outcomes in DBA/2-mated CBA/J females.

Conclusion

In this review, the dual role of TF in fetal and placental development is described. On the one hand, TF is required for a successful pregnancy, and on the other hand, pathologic expression of TF can lead to serious pregnancy complications in humans and mice. Human studies show increased TF in plasma, amniotic fluid and and/or placentas of abnormal pregnancies affected by miscarriages, preterm birth, or pre-eclampsia. Interestingly, using two mouse models, we found that blood-borne TF plays a crucial role in the pathogenesis of pregnancy complications. TF on neutrophils and monocytes is a critical mediator in trophoblast injury and embryonic damage in pregnancy loss induced by APL antibodies and in the antibody-independent CBA/J × DBA/2 model of miscarriages. Blockade of TF or genetic diminution prevented pregnancy complications, suggesting that TF may be a good target for therapy in patients with recurrent miscarriages, pregnancy loss, and pre-eclampsia. Targeting of TF might be associated with bleeding risks. Pravastatin downregulates TF and PAR-2, and thus targets the proinflammatory effects of TF without interfering with the coagulation cascade. This suggests that statins may constitute a good therapeutic option for women with pregnancy complications. A clinical trial should be conducted to confirm these observations in women.

Disclosure of Conflict of Interests

The author states there is no conflict of interest.

Ancillary