| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Molecular, Cellular and Developmental Biology (S.L.S.-C.), Yale University, New Haven, Connecticut 06520; and Department of Obstetrics, Gynecology and Reproductive Sciences (V.M.A., G.M.), Yale University School of Medicine, New Haven, Connecticut 06520
Correspondence: Address all correspondence and requests for reprints to: Gil Mor, M.D., Ph.D., Department of Obstetrics, Gynecology and Reproductive Sciences, Reproductive Immunology Unit, Yale University School of Medicine, 333 Cedar Street FMB 301, New Haven, Connecticut 06520. E-mail: Gil.Mor{at}yale.edu
 |
Abstract |
|---|
|
I. Introduction |
|---|
Normal placental development is dependent upon the differentiation and invasion of the trophoblast, the main cellular component of the placenta that originates from the trophoectoderm of the blastocyst early in pregnancy. During this process of differentiation and invasion, trophoblast cells rapidly divide to form the interface between mother and embryo, while other trophoblast subpopulations invade the decidua (pregnant endometrium) to remodel the arterial blood vessels in the uterine wall, known as the spiral arteries, to accommodate the expansion of extraembryonic tissue and to increase blood flow to the placenta and developing fetus. As a developing organ, the placenta undergoes constant tissue remodeling, which is characterized by the functional loss of trophoblast cells by apoptosis. After proliferation and differentiation into specific cell subtypes, aging trophoblast cells are selectively removed and replaced by a younger population of trophoblasts without affecting neighboring cells (4).
Apoptosis, or programmed cell death, is an active process by which superfluous or dysfunctional cells are eliminated to maintain normal tissue function. Depending on the stimuli, apoptosis may be initiated by one of two known pathways: intrinsically by the mitochondrial pathway and extrinsically by either the death receptor-mediated pathway or in response to exogenous stimuli such as cytokines. The central executioners of apoptosis are the caspases, which are a family of cysteine proteases that cleave numerous vital cellular proteins to affect the apoptotic cascade (5). Several endogenous inhibitors, including flice-like inhibitory proteins (FLIPs), inhibitors of apoptosis (IAPs), and antiapoptotic Bcl-2 family members, inhibit caspase activation, thereby preventing further propagation of the death signal.
Whereas apoptosis is thought to be important for normal placental development, it may also be involved in the pathological conditions associated with this organ. Apoptotic cells have been identified in both the maternal and fetal compartments of the placenta during normal pregnancy, and the presence of these cells may be related to the stage of placental development, including the attachment and invasion of the trophoblast (6, 7), spiral artery transformation (8), trophoblast differentiation and turnover (9, 10, 11), and parturition (12, 13). Moreover, apoptosis has also been shown to play an important role in promoting maternal immune tolerance to paternal antigens expressed by trophoblast cells (6, 9, 14). In complicated pregnancies such as preeclampsia or IUGR, a greater incidence of trophoblast apoptosis has been observed (15, 16, 17, 18, 19, 20), suggesting that alterations in the regulation of trophoblast apoptosis may contribute to the pathophysiology of these diseases.
As cells undergo apoptosis, macrophages present at the maternal-fetal interface quickly remove apoptotic cells by phagocytosis to promote trophoblast survival and facilitate invasion and transformation of the spiral arteries (21). In placentas from complicated pregnancies, shallow trophoblast invasion and inefficient spiral artery transformation have been observed (22, 23), which may be partly due to the distribution and activation state of infiltrating macrophages.
In this review, we will discuss the role of apoptosis in the development and differentiation of the trophoblast, the molecular mechanisms by which trophoblast cell apoptosis can occur, how it is regulated during normal pregnancy, and the alterations that have been observed in pregnancy-related diseases associated with excessive trophoblast apoptosis. The role of macrophages in apoptotic cell clearance and trophoblast survival will also be discussed.
A. An overview of apoptotic pathways
1. The extrinsic pathway: characteristics of the TNF death receptor family.
Depending upon the stimuli, apoptosis can be initiated by one of two known pathways: the mitochondrial, or intrinsic, pathway and the death receptor-mediated, or extrinsic, pathway. If executed by the extrinsic pathway, apoptosis is initiated by members of the TNF death receptor family, which are part of the TNF-receptor (TNF-R) superfamily and have a C-terminal region of approximately 80 amino acids known as the death domain in common (24). To date, eight members of this family have been identified and include Fas (CD95/APO-1), TNF-R1 (CD120a), death receptor 3 (APO-3/WSL-1/TRAMP/LARD), TRAIL-R1 (death receptor 4), TRAIL-R2 (death receptor 5/TRICK2), death receptor 6, EDAR, and NGFR (25). Among the death receptors, Fas, TNF-R1, and TRAIL-R1/TRAIL-R2 are the most widely studied and best characterized. Several of these receptors and their corresponding ligand(s) have been shown to exist in at least two forms, a membranal form and a transmembrane-deficient soluble form, the latter of which may protect from death receptor-induced apoptosis under certain circumstances (26). A third, secreted form of Fas ligand (FasL), which is distinct from the soluble form, was recently described (27, 28). The interaction between a ligand and its membrane-bound death receptor results in the activation of preassociated death receptor trimers (29, 30). To transduce apoptotic signals, TNF death receptors have intracellular death domains, which mediate protein-protein interactions with other death domain-containing adaptor proteins such as the Fas-associated death domain (FADD; MORT1) and TNF-R-associated death domain (31). After FADD binds to the cytoplasmic tail of the death receptor either directly or, in the case of TNF-R1, indirectly through TNF-R-associated death domain (32), it recruits other cellular proteins, including procaspase-8 and procaspase-10 (33, 34) via death effector domains (DED) to form the death-inducing signaling complex (DISC), (35). This is the point at which the TNF death receptor pathways converge (Fig. 1
). Once the DISC is assembled, "initiator" caspase-8 and caspase-10 can become activated (34, 36); however, the mechanism by which this occurs is not fully understood.
|
). For additional information on TNF death receptor-induced apoptosis and how cell proliferation, survival, and apoptosis are regulated by members of the TNF superfamily, please refer to Refs.41 and 42 .
2. The intrinsic pathway: characteristics of the mitochondrial pathway and its role in death receptor-induced apoptosis.
Unlike the extrinsic pathway, which depends on death receptor signaling, in the intrinsic pathway, the apoptotic signal is initiated by, or directed to, the mitochondria. In response to cellular stresses such as DNA damage or growth factor deprivation, the mitochondrial pathway can be activated by p53, a tumor suppressor protein that transactivates proapoptotic Bcl-2 family members (43). However, the extrinsic and intrinsic pathways are not necessarily autonomous, because p53 can up-regulate the expression of certain death receptors, and the mitochondrial pathway may act to amplify signals triggered by the death receptor pathway, suggesting that crosstalk can occur between the two pathways (44, 45).
In addition to activating effector caspases, caspase-8 can also cleave Bid, a proapoptotic Bcl-2 family member, resulting in the translocation of truncated Bid to the mitochondria and the activation of the intrinsic pathway (46) (Fig. 1). Consequently, other proapoptotic Bcl-2 family members such as Bax and Bak increase the permeability of the outer mitochondrial membrane by a highly controversial mechanism (47, 48, 49) to release cytochrome c, apoptosis-inducing factor (50), and other proapoptotic factors (51, 52, 53). As a result, cytochrome c binds the adaptor protein, apoptotic protease activating factor-1, which together with ATP or dATP, subsequently recruits and activates "initiator" caspase-9, forming a macromolecular complex called an "apoptosome" (54). Analogous to the DISC in the extrinsic pathway, the apoptosome serves as the activation complex for caspase-9 in the intrinsic pathway. After recruitment, caspase-9 dimerizes and is subsequently activated by an allosteric mechanism similar to that by which initiator caspase-8 and caspase-10 are activated at the DISC. In turn, active caspase-9 activates "effector" caspase-3, caspase-6, and caspase-7, the point at which the mitochondrial and death receptor pathways overlap (40) (Fig. 1).
As effector caspases, caspase-3, caspase-6, and caspase-7 cleave a variety of vital cellular proteins, including DNA repair enzymes, nuclear lamins, and cytoskeletal proteins (5), which might explain the characteristic features of apoptosis such as nuclear condensation, membrane blebbing, and cell shrinkage. In addition, inhibitor of caspase-activated deoxyribonuclease is also cleaved by effector caspases, releasing caspase-activated deoxyribonuclease to nonspecifically cut the genomic DNA into approximately 200-bp fragments (55), eventually ending in apoptosis.
B. The role of apoptosis in normal pregnancy
1. Apposition and adhesion of the trophoectoderm.
Embryonic implantation consists of three consecutive phases: apposition, adhesion, and invasion, all of which have been shown to involve apoptosis. In the apposition, or preattachment stage, the developing blastocyst interacts with the epithelial cells of the endometrium in preparation for implantation. Using an in vitro model that mimics the apposition and attachment phases of implantation, Galan et al. (7) demonstrated that the presence of a human blastocyst reduced the percentage of human (h) endometrial epithelial cell (EEC) apoptosis in comparison to hEECs cultured without a blastocyst. The authors speculated that the embryo produces antiapoptotic factors, which promote hEEC survival and facilitate the attachment of the trophoectoderm to the uterine epithelium in the adhesion phase. Once the blastocyst had adhered to the hEECs, however, a massive increase in hEEC apoptosis was observed at and around the site of embryo attachment, suggesting that the trophoectoderm induces a paracrine apoptotic reaction in hEECs (7). Interestingly, similar findings have been reported in rodents (56, 57).
A potential mechanism by which the blastocyst may induce hEEC apoptosis is through the Fas/FasL system. Indeed, the majority of hEECs express Fas on their apical surface, whereas FasL is expressed by the trophoectoderm. As the blastocyst invades, the FasL expressed on the surface of the trophoectoderm may bind to Fas on the hEECs and induce apoptosis. In the presence of a blocking anti-Fas antibody, embryonic adhesion and outgrowth are significantly reduced in hEECs cultured with mouse blastocysts compared with the nontreated control (7). This suggests that the interaction between Fas expressed on the endometrial epithelium and FasL expressed on the trophoectoderm may enable the embryo to breach the epithelial barrier and invade the decidua during implantation. A more recent study has demonstrated that trophoblast outgrowth is also inhibited by suppressing the intracellular p38 MAPK pathway, but not the ERK pathway, suggesting that EEC apoptosis may be mediated by p38 MAPK (58).
2. Extravillous trophoblast invasion and spiral artery transformation.
The trophoblast can be divided into two populations, villous and extravillous trophoblast, each of which can be further subdivided. Both villous and extravillous trophoblast cells are originally derived from the trophoectoderm of the implanting blastocyst early in pregnancy. Villous trophoblast comprises proliferating mononuclear cytotrophoblast cells, which continuously fuse to form the multinucleated syncytiotrophoblast, the second layer of trophoblast, which is in direct contact with the maternal circulation (4). Extravillous trophoblasts are a specialized population of cytotrophoblast cells, which have the capacity to invade the decidua and remodel the uterine blood vessels to establish an adequate blood supply for the fetus. Analogous to villous trophoblasts, extravillous trophoblasts can be subdivided into two populations: interstitial and endovascular trophoblast cells. Whereas interstitial trophoblasts invade the interstitium, or wall of the uterus, endovascular trophoblast cells infiltrate and transform the spiral arteries, thereby increasing nutrients and maternal blood flow to the placenta (59).
Although the initial changes to the arterial blood vessels appear to be trophoblast independent and part of the maternal response to pregnancy (60), it is clear that, in the absence of trophoblast invasion, vessel remodeling is significantly reduced (8), suggesting that the trophoblast is important for spiral artery transformation. It is thought that extravillous trophoblasts infiltrate the spiral arteries and migrate along the lumen of the vessels, replacing the endothelium and musculoelastic tissue in the vessel walls. Indeed, using first-trimester decidual and villous explants, Dunk et al. (61) reported a disruption in the endothelial cell lining of the vessel and a complete loss of organized smooth muscle surrounding the vessel concomitant with trophoblast invasion.
Recent evidence suggests that the transformation of the spiral arteries is mediated by the induction of apoptosis in the endothelial cells lining the lumen. Moreover, it appears that the proapoptotic signal inducing endothelial cell apoptosis may be delivered by the invading trophoblast. Using an in vitro coculture system consisting of fluorescently labeled first-trimester trophoblast cells and spiral arteries dissected from nonplacental bed biopsies obtained by cesarean section (62), Cartwright’s group demonstrated that only arteries with trophoblasts in the lumen of the vessels exhibited a loss in the endothelial layer, which stained positively for the caspase-mediated cleavage product of poly (ADP-ribose) polymerase (PARP), a marker of apoptosis. Similar results were obtained when endothelial cells were cocultured with a first-trimester extravillous trophoblast cell line. Furthermore, trophoblast-induced endothelial cell apoptosis was inhibited with the use of a pan caspase inhibitor (8).
Interestingly, the expression of Fas was detected on endothelial cells in spiral arteries and isolated primary endothelial cells, suggesting that the Fas/FasL system might represent the mechanism by which trophoblasts induce endothelial cell apoptosis. Indeed, trophoblast induction of endothelial cell apoptosis was abrogated in the presence of a blocking FasL antibody. This was confirmed using the spiral artery explant model, because apoptosis could not be detected in vessels when the blocking FasL antibody was added before trophoblast perfusion (8). Whether vascular smooth muscle cell loss is similarly regulated and endothelial cell apoptosis is required for changes in the smooth muscle layer remains to be determined. However, smooth muscle cells have also been shown to express Fas (8), and the Fas/FasL system was recently implicated as one mechanism by which trophoblasts may induce smooth muscle cell apoptosis (63, 64). Taken together, this suggests that the induction of apoptosis in endothelial cells, and possibly in smooth muscle cells, by the trophoblast is critical for spiral artery remodeling and to secure an adequate blood supply for the developing fetus. Defects in the signals delivered by the trophoblast, or in the response of the endothelium/vascular smooth muscle to the proapoptotic stimuli, may begin to explain the lack of vessel transformation observed in pregnancy complications such as preeclampsia.
3. Maternal immune tolerance.
In the early 1950s, Medawar (65) described the "fetal-allograft analogy," in which the fetus was viewed as a semiallogenic conceptus that evaded immune rejection. More recently, this analogy was redefined as maternal-placental tolerance rather than maternal-fetal tolerance because it is the trophoblast and not the fetus that is in direct contact with the maternal immune system. Because trophoblast cells express paternal antigens, which are considered antigenically foreign to the mother, the trophoblast should, therefore, be rejected by the maternal immune system. However, under normal conditions, the trophoblast avoids maternal immune rejection and benefits from an immune privilege state during pregnancy.
Several theories exist to account for the immune privilege status of the trophoblast, including the expression of nonclassical human leukocyte antigen molecules (66) and complement regulatory proteins (67), tryptophan catabolism by the indoleamine 2,3-dioxygenase enzyme (68), immunosuppression by regulatory T cells (69), and the clonal deletion of specific T cell clone types (70, 71). Indeed, apoptotic leukocytes, particularly T lymphocytes, have been localized to the maternal-fetal interface of normal placentas (72, 73) and antigen-specific T cell deletion has been observed during pregnancy (70, 71). This suggests that the trophoblast can respond to the presence of maternal leukocytes and that the maternal immune system may not be indifferent to the fetus as was once previously thought (65). In support of this, two studies have demonstrated that maternal T cells, which recognize paternal antigens on the trophoblast, are selectively abrogated during pregnancy in mice (70, 71). Interestingly, Tafuri et al. (70) also reported that T cell numbers and reactivity were restored after delivery, suggesting that maternal T cells acquire a tolerant state that is both transient and reversible. Further studies, however, are necessary to confirm these findings.
Because Fas and FasL were originally implicated in the removal of activated peripheral T cells after an immune response (74), it was suggested that one of the possible mechanisms by which maternal T cells may be tolerized to paternal alloantigens is by the Fas/FasL system. Indeed, trophoblast cells isolated from mouse placentas were shown to induce apoptosis in a mouse T cell line that expresses Fas, but not in a Fas-deficient T cell line. In contrast, trophoblast cells purified from the placentas of homozygous gld (generalized lymphoproliferative disease) mice, which do not express functional FasL, did not induce apoptosis in the Fas-expressing T cell line (71). Similarly, Kauma et al. (75) demonstrated that a FasL-expressing first-trimester trophoblast cell line was able to induce Fas-mediated apoptosis in activated lymphocytes in vitro. However, several in vivo studies have reported that cells expressing membranal FasL were rejected when transplanted into allogenic animals (76, 77), which brings into question the role of the Fas/FasL system in immune privilege. This may be due to the overexpression of FasL in transplanted cells engineered to express membranal FasL. Indeed, the overexpression of membranal FasL has been shown to be associated with inflammation, neutrophil activation, and rejection (78).
Analogous to the trophoblast, certain tumor cells have been shown to express FasL. It has been postulated that FasL-expressing tumor cells may protect themselves from immune surveillance by inducing Fas-mediated apoptosis in immune cells that infiltrate the tumor (79) during a process called the "Fas counterattack" (80). More recently, our group has demonstrated that FasL is expressed in the cytoplasm of both epithelial ovarian cancer cells and isolated first-trimester trophoblast cells, rather than on the membrane, and that these cells can secrete the full-length functional form of FasL via the release of microvesicles (14, 27). After microvesicle disruption, the secreted FasL is able to induce apoptosis in Fas-bearing immune cells, suggesting that the constitutive secretion of FasL may be one mechanism by which the trophoblast promotes maternal tolerance to paternal antigens and prevents fetal rejection during pregnancy (14).
4. Villous trophoblast differentiation and turnover.
Another important function of apoptosis during pregnancy is related to the regulation of trophoblast growth and turnover in the villous part of the placenta. Morphological characteristics consistent with apoptosis have been observed in the villous trophoblast of normal placentas (10, 12, 13, 81), suggesting that villous trophoblast apoptosis is a physiological process that occurs during normal pregnancy. The villous trophoblast is an active tissue that is undergoing constant cell turnover and renewal by a well-orchestrated process of tissue remodeling involving apoptosis. Moreover, the percentage of apoptotic trophoblast cells has been shown to increase in villous placenta as gestation proceeds (12, 13), which suggests that villous trophoblast apoptosis is developmentally regulated. In pathological conditions such as preeclampsia and IUGR, a greater incidence of villous trophoblast apoptosis has been observed, suggesting that the appropriate regulation of trophoblast apoptosis is important for normal pregnancy (15, 16, 17, 18).
As mentioned previously, the villous trophoblast consists of two subpopulations: cytotrophoblast cells and the syncytiotrophoblast layer. The syncytiotrophoblast does not have the capacity to proliferate and is, therefore, dependent on the villous cytotrophoblast population for growth and renewal. Initially, proliferating cytotrophoblast cells begin to differentiate, leading to the fusion of these cells with the overlying syncytiotrophoblast. After syncytial fusion, a second differentiation process occurs in the outer syncytiotrophoblast layer, in which each aging nucleus is packed into a syncytial knot and extruded. Villous trophoblast turnover is completed once these syncytial knots are shed into the maternal circulation (4, 82).
Several components of the apoptotic cascade are differentially expressed in cytotrophoblast cells and the syncytiotrophoblast layer, and this is thought to be associated with the stage of apoptosis observed in the different compartments of placental villi. Activation of the apoptotic cascade is initiated in the villous cytotrophoblast and completed in the syncytiotrophoblast before the extrusion of apoptotic nuclei, suggesting that villous trophoblast apoptosis and differentiation are interdependent. However, the initial activation of the apoptotic cascade appears to be related to the differentiation of cytotrophoblast cells rather than trophoblast cell death. Therefore, apoptosis is not only involved in the removal of aging syncytiotrophoblasts, but it may also promote cytotrophoblast fusion and formation of the syncytial layer (9). Moreover, recent studies have demonstrated that caspases are important mediators of cytotrophoblast differentiation (83, 84, 85, 86). How caspases can promote cytotrophoblast differentiation without inducing apoptosis is an intriguing question, which will be discussed in a later section. For a recent review on the role of villous trophoblast apoptosis in placental morphology, please refer to Ref.82 .
To understand the differential effects of apoptosis in normal pregnancy and its potential role in pathological conditions, it is important to discuss the molecular mechanisms by which the above-described processes may be mediated. This is a rapidly expanding area of research as demonstrated by the increase in the number of publications in recent years. In the following section, we will discuss the current knowledge of the molecular mechanisms that regulate apoptosis in the trophoblast.
|
II. Trophoblast Expression and Function of the TNF-R Family and Their Respective Ligands |
|---|
Using different methods, our laboratory and others have demonstrated that isolated trophoblast cells predominantly express the full-length transmembrane form of Fas (88, 97) and that trophoblast Fas expression levels remain high throughout placental development (90). In contrast, the expression of FasL has been shown to decrease in villous trophoblasts at term (72, 89, 90, 93), but not in those collected before the onset of labor (6, 89), suggesting that a reduction in FasL expression may modulate the inflammatory processes associated with parturition (89). More recently, our group demonstrated that, contrary to previous notions, FasL is not expressed on the plasma membrane, but instead is associated with a specialized secretory lysosomal pathway in the cytoplasm of trophoblast cells and secreted via the release of microvesicles (14). This was recently confirmed in vivo by Mincheva-Nilsson’s group (98), who reported the presence of FasL-containing microvesicles in secretory lysosomes in first-trimester placentas by electromicroscopy. In addition, we also illustrated that this secreted form of FasL is able to induce apoptosis in activated immune cells and is distinct from the soluble form of FasL (14). The biological role of secreted FasL and how trophoblast secretion of FasL is regulated during pregnancy are areas that our group is currently investigating. Based on our in vitro studies, we postulate that FasL secretion is important for trophoblast immune privilege, which might explain why a decrease in FasL expression has been observed in villous trophoblast at term (89).
Although first-trimester trophoblast cells express both Fas and FasL, they are resistant to Fas-mediated apoptosis under normal conditions (88, 90, 99). Therefore, the expression of Fas and FasL by trophoblast cells does not necessarily correlate with susceptibility to Fas-induced apoptosis. Recent evidence from our laboratory suggests that trophoblast resistance to Fas-mediated apoptosis is due to the inhibition of the pathway downstream of Fas stimulation (99), which will be discussed in a later section. Interestingly, first-trimester trophoblast cells have been shown to actually proliferate rather than undergo apoptosis after Fas stimulation (99). This may be explained by studies illustrating that signals transmitted by Fas, as well as the other prototypical death receptors, TNF-R1 and TRAIL-R2, may also promote growth in certain cells through the activation of nuclear factor-
B (NF-B) and mitogen-activated kinases such as c-jun N-terminal kinase (100, 101, 102). Therefore, if the intracellular apoptotic cascade is blocked, the same receptors that induce apoptosis in trophoblast cells may also promote trophoblast cell proliferation under certain circumstances.
Interestingly, the soluble forms of Fas and FasL have been detected in the peripheral blood of normal pregnant women as well as in complications of pregnancy such as preeclampsia. Although the number of patients evaluated in these studies was relatively small, the levels of soluble Fas and FasL were shown to be higher in preeclamptic women than in the normotensive pregnant controls (103, 104, 105). The significance of these biological findings, however, is not fully understood.
2. TNF-R1 and TNF-
.
The expression of TNF-R1 and TNF-, the ligand for TNF-R1, has also been localized to villous (49, 106, 107, 108, 109, 110, 111, 112), as well as extravillous trophoblast cells (108, 113). Moreover, there is some evidence indicating that TNF-R1 and TNF- expression increases in placental tissues with gestational age (106, 110), suggesting that along with the Fas/FasL system, TNF-R1 and TNF- may also be involved in the process of parturition. However, in contrast to Fas, TNF- is a potent inducer of trophoblast apoptosis (88, 97, 108, 114). Because TNF-R1 and Fas have several components of the intracellular apoptotic cascade in common, we hypothesize that TNF- might activate an alternative pathway in trophoblast cells rather than the classical TNF death receptor pathway (99, 115).
Interestingly, TNF- can be proteolytically cleaved from the membrane to generate a soluble form of TNF-; however, only a small amount of TNF- could be detected in the culture supernatants from first-trimester villous explants (116) and isolated cytotrophoblast cells (117). This suggests that the trophoblast is not a primary source of soluble TNF- in placental tissues and that TNF- may instead be released from macrophages, natural killer cells, or other cells at the maternal-fetal interface (49, 106, 118). Nevertheless, because membranal TNF- and TNF-R1 colocalize to trophoblast subpopulations, other mechanisms must exist to prevent TNF--induced trophoblast apoptosis. Indeed, soluble TNF-Rs have been detected in the amniotic fluid and urine samples from pregnant women, suggesting that the secretion of soluble TNF-Rs may protect against the cytotoxic effects of TNF- expression (119). Moreover, Knofler et al. (108, 120) demonstrated that isolated villous trophoblast cells shed large quantities of soluble TNF-Rs, which suggests that the trophoblast may represent the source of the secreted soluble TNF-Rs.
3. TRAIL-R1/R2 and TRAIL.
In contrast to the TNF-R1/TNF- and the Fas/FasL system, TNF-related apoptosis-inducing ligand (TRAIL) and its receptors, TRAIL-R1 and TRAIL-R2, appear to be differentially expressed in villous placenta. Villous cytotrophoblast cells have been shown to express high levels of TRAIL-R1/TRAIL-R2, whereas strong TRAIL expression was detected in syncytiotrophoblasts from both trimesters of pregnancy (49, 121). This disparate localization of TRAIL-R1 and TRAIL-R2 to cytotrophoblasts and TRAIL to syncytiotrophoblasts may not only provide protection from maternal immune cells but may also prevent autocrine or paracrine killing by the same or neighboring trophoblast cell. Moreover, Chen et al. (121) revealed that the expression of TRAIL-R1 and TRAIL-R2 increased in villous trophoblasts with gestational age, suggesting that these receptors may cooperate with other TNF death receptors to increase trophoblast apoptosis toward the end of pregnancy.
With regard to the response of TRAIL-R1/TRAIL-R2 to TRAIL in trophoblast cells, the only study that has tested TRAIL sensitivity, of which we are aware, used trophoblast-derived choriocarcinoma cell lines, which were shown to be resistant to recombinant TRAIL (122). Whether normal trophoblast cells exhibit similar resistance to TRAIL-induced apoptosis is unknown.
4. Death receptors 3 and 6.
The TNF death receptors, death receptors 3 and 6, have also been shown to be expressed by isolated term trophoblast cells (49). However, trophoblast expression of the death receptor 3 ligand, TNF-like 1A (123), has not been reported, and a ligand for death receptor 6 has yet to be identified. Therefore, whether or not death receptors 3 and 6 have a functional role in trophoblast cells remains to be determined.
B. Non-death domain-containing TNF-Rs
In addition to the classical TNF death receptors, trophoblast cells have also been shown to express several non-death domain-containing TNF-Rs such as TNF-R2, lymphotoxin-ß receptor (LT-ßR), and herpes virus entry mediator (HVEM) and their respective ligands, TNF-, LT-/LT-ß, and LIGHT (49, 124). The expression of 4–1BB (CD137), another non-death domain-containing TNF-R, has not been detected, but the expression of its ligand, 4–1BBL, was observed in isolated term cytotrophoblast cells (49).
The precise role of non-death domain-containing TNF-Rs in trophoblast cell apoptosis and survival is not completely understood. Although capable of inducing apoptosis in other cell types (125), it was previously shown that TNF-R2 does not participate in TNF--induced trophoblast cell apoptosis and that apoptosis induced by TNF- is mediated almost entirely by TNF-R1 (109). This was supported by another study, which demonstrated that recombinant mouse TNF- decreased the viability of trophoblast cells isolated from TNF-R2 knockout mice, whereas similar treatment had no effect on TNF-R1-deficient mouse trophoblast cells (126). Recent evidence suggests that receptor interacting protein (RIP), a death domain-containing serine/threonine kinase, which is recruited to TNF-R1 upon receptor activation, may also affect TNF-R2 signaling. In the presence of RIP, TNF-R2 was able to trigger apoptosis, but when RIP was absent, TNF-R2 activated the NF-B pathway, suggesting that RIP may be important for TNF-R2-mediated apoptosis in certain cell types (101). Whether RIP is expressed by trophoblast cells and has a functional role in trophoblast TNF-R signaling remains to be elucidated.
Unlike TNF-R2, recent data from Gill and Hunt (127) suggest that HVEM and/or LT-ßR are capable of inducing LIGHT-mediated trophoblast apoptosis. Therefore, whether non-death domain-containing TNF-Rs promote trophoblast cell survival or induce apoptosis may depend on which receptors are expressed, the type or strength of stimulus, and the activation status of the intracellular apoptotic cascade and/or survival pathways in trophoblast cells.
C. Decoy receptors (DcRs)
A new area of research that has generated much interest in the field of apoptosis is the expression and function of DcRs. Interestingly, trophoblast cells have been shown to express several decoy receptors, which lack death domains and are unable to transduce an apoptotic signal. As the name implies, decoy receptors compete with the classical TNF death receptors, Fas and TRAIL-R1/R2, as well as the non-death domain-containing TNF-Rs, HVEM and LT-ßR, for ligand binding, thereby suppressing receptor-induced apoptosis (42).
Isolated term cytotrophoblasts have been shown to express DcR3 (TR6) (127), the decoy receptor for FasL and LIGHT, as well as DcR1 and DcR2 (49), both of which are capable of binding TRAIL. DcR1 and DcR2 are expressed in cytotrophoblast cells and the syncytiotrophoblast of placental tissues throughout gestation, however, the levels of DcR1 and DcR2 expression decrease in villous trophoblast with gestational age (121), suggesting that a reduction in decoy receptor expression may promote death receptor/ligand binding and allow trophoblast cell apoptosis to occur. Using a recombinant DcR3-Fc (rhDcR3-Fc) fragment to inhibit LIGHT binding to HVEM or LT-ßR, Gill and Hunt (127) also demonstrated that the rhDcR3-Fc fragment was able to restore cell viability in isolated cytotrophoblast cells treated with interferon-
(IFN-) and LIGHT, which suggests that DcR3 protects trophoblast cells from LIGHT-mediated apoptosis.
Altogether, these studies demonstrate that trophoblast cells possess multiple death receptors, including those that lack death domains, as well as several decoy receptors, but how these receptors function simultaneously to regulate trophoblast cell apoptosis and/or survival throughout pregnancy is only beginning to be understood.
|
III. The Role of Caspases in Trophoblast Differentiation |
|---|
Interestingly, it appears that the activity of initiator caspase-8 is more prevalent in first- (83) and third-trimester cytotrophoblasts (85) than in syncytiotrophoblasts. In addition, villous cytotrophoblast cells have been shown to exhibit annexin-V binding (9), which measures the externalization of the aminophospholipid, phosphatidylserine, from the inner to the outer leaflet of the plasma membrane in cells undergoing the initial stages of apoptosis. However, early apoptotic events, such as caspase-8 activation and phosphatidylserine externalization, may not be associated with trophoblast cell death, but instead may be required for the fusion of cytotrophoblast cells with the syncytiotrophoblast layer (83). Using an in vitro model consisting of trophoblast-derived choriocarcinoma cells, Rote and co-workers (131) demonstrated that treatment with forskolin induced cell fusion, whereas application of antibodies directed against phosphatidylserine prevented fusion in JAR cells, suggesting that phosphatidylserine externalization may be important for the induction of syncytial fusion in normal villous trophoblast.
Caspase activation and, in particular, the activation of caspase-8 (132, 133), are thought to be involved in the externalization of phosphatidylserine and, therefore, cytotrophoblast differentiation. Indeed, Huppertz et al. (83) reported the loss of fodrin, a cytoskeletal protein that is initially cleaved by active caspase-8, in villous cytotrophoblast before its differentiation into syncytiotrophoblast. Using antisense oligonucleotides and peptide inhibitors to inhibit caspase-8 expression and activity, respectively, the same group also demonstrated that the inhibition of caspase-8 reduces syncytial fusion, thereby preventing progression of the apoptotic cascade in syncytiotrophoblast and resulting in the accumulation of mononucleated cytotrophoblast cells (84). This suggests that the activation of caspase-8 is initiated in villous cytotrophoblast cells, which, in turn, promotes their fusion with the syncytiotrophoblast and formation of the syncytial layer.
Along with the results of these studies arose additional questions of how the trophoblast survives despite caspase activation. More specifically, 1) what prevents apoptotic signaling events downstream of caspase-8 activation in cytotrophoblast cells, 2) which other components of the apoptotic cascade are involved in this process, and 3) how trophoblast apoptosis is regulated in parallel with syncytial fusion during pregnancy, which will be addressed in the following section.
|
IV. Endogenous Regulators of Trophoblast Apoptosis |
|---|
). Several of these intracellular inhibitors are expressed in the placenta, including FLIPs, which share significant structural homology with caspase-8 in that they each contain two DEDs. This feature enables FLIPs to interfere with the recruitment and activation of caspase-8 at the DISC, thereby preventing death receptor-mediated apoptosis (134). Indeed, an increase in FLIP expression has been shown to protect first-trimester trophoblast cells from Fas-induced apoptosis (97). Moreover, FLIPs may reinforce the inhibition of the intracellular apoptotic cascade by activating survival pathways such as NF-B and promoting cell proliferation (135).
|
) thus far. Of the two FLIP isoforms observed in placental tissues, FLIPS appears to be the more potent inhibitor of apoptosis. In the presence of FLIPS, the recruitment and activation of caspase-8 are completely inhibited at the DISC (136). However, if FLIPS is such a potent caspase-8 inhibitor, it is unclear how caspase-8 can be activated and promote the differentiation of FLIPS-expressing cytotrophoblast cells. In contrast to FLIPS, FLIPL is unable to preclude caspase-8 recruitment to the DISC (136). It has been suggested that caspase-8 is activated by a two-step cleavage process, the first of which occurs autocatalytically and requires the presence of a caspase domain. In addition to DEDs, FLIPL also contains a caspase domain that is homologous to the catalytic domain in caspase-8. Although the caspase-like domain in FLIPL is inactive, it may be sufficient to induce the initial cleavage of caspase-8, resulting in the generation of the p41/43 caspase-8 cleavage products. During the second cleavage step, caspase-8 is cleaved transcatalytically, which requires a functional caspase domain. Because the caspase-like domain in FLIPL is catalytically inactive, this might explain why FLIPL is able to inhibit only the second caspase-8 cleavage step, whereas FLIPS, which lacks a caspase domain, prevents all caspase-8 processing (136). Therefore, the presence of FLIPL may provide an explanation to how caspase-8 can be activated and promote cytotrophoblast differentiation despite FLIPS expression in trophoblast cells.
|
B. The Bcl-2 family
Bcl-2 family members can modulate death signals either directed toward or initiated by the intrinsic pathway by differentially regulating the release of proapoptotic factors from the mitochondria. The Bcl-2 family is comprised of three functional groups, which are characterized by varying numbers of Bcl-2 homology (BH) domains. Members of the first group, such as Bcl-2 and Bcl-xL, contain four BH domains (BH1–BH4) and inhibit apoptosis, whereas Bax and Bak, which are part of the second group, are proapoptotic and lack a N-terminal BH4 domain. Whereas Bax and Bak increase the permeability of the mitochondrial membrane to cytochrome c release, Bcl-2 and Bcl-xL inhibit the release of cytochrome c and other proapoptotic factors from mitochondria, thereby preventing further propagation of the apoptotic signal (47, 139) (Fig. 2). The BH3-only proteins, Bid and Bik, are members of the third group and promote apoptosis by either activating proapoptotic members, such as Bax and Bak, or binding antiapoptotic Bcl-2 family members to inhibit their function. Thus, the outcome may depend on the relative abundance of pro- vs. antiapoptotic family members in the cell (140).
Syncytiotrophoblasts have been shown to express Bcl-2 at higher levels than villous cytotrophoblasts throughout pregnancy (9, 141, 142, 143, 144, 145, 146). Similarly, the expression of myeloid cell leukemia-1 (Mcl-1), another antiapoptotic Bcl-2-related family member, is more prominent in syncytiotrophoblasts compared with cytotrophoblast cells (9, 144). The expression of Bcl-2 and Mcl-1 in syncytiotrophoblasts was originally thought to be associated with the prevention of syncytiotrophoblast apoptosis after syncytial fusion. However, Bcl-2 and Mcl-1 prevent the activation of caspase-9 and do not have a direct effect on caspase-8 activation, which occurs during cytotrophoblast differentiation. In addition, caspase-8 can directly activate effector caspases without mitochondrial support (40), which might explain why apoptosis is still observed in Bcl-2- and Mcl-1-expressing syncytiotrophoblasts (12, 13, 147). Moreover, syncytiotrophoblasts have also been shown to express Bax (148) and Bak (145), suggesting that the expression of proapoptotic Bcl-2 family members may also promote syncytiotrophoblast apoptosis (142, 148). These findings also suggest that other intracellular inhibitors outside of the mitochondria may exist to prevent apoptosis from occurring in the syncytial layer.
C. IAPs
IAPs are unique in that they are capable of inhibiting both the mitochondrial and death receptor-mediated pathways (Fig. 2). To date, eight human IAPs have been identified and include X-linked IAP (XIAP; MIHA/ILP-1), ILP-2 (Ts-IAP), c-IAP1 (HIAP2/MIHB), c-IAP2 (HIAP1/MIHC), neuronal apoptosis inhibitory protein, Survivin (TIAP), Livin (KIAP/ML-IAP), and Apollon (Bruce). IAP family members are characterized by varying numbers of baculoviral IAP repeat (BIR) domains and, with the exception of neuronal apoptosis inhibitory protein, survivin, and Apollon, also contain a C-terminal really interesting new gene (RING)-zinc finger domain (149). The RING domain has E3 ubiquitin ligase activity, which enables IAPs to ubiquitinate and degrade themselves or other interacting proteins after certain apoptotic stimuli (150). In addition, c-IAP1 and c-IAP2 also contain a caspase-recruitment domain (151), the function of which in these proteins is still unknown (152). For a recent review on IAPs, please refer to Ref.153 .
XIAP, c-IAP1, c-IAP2, NAIP, Survivin, and Livin all have been shown to be expressed in villous trophoblasts and, except for c-IAP1, in extravillous trophoblast cells (90, 154, 155, 156). Among the IAPs, XIAP is the most potent and versatile member of the family because it is able to inhibit both the mitochondrial and death receptor-mediated pathways (157). XIAP contains three tandem BIR domains and a C-terminal RING domain, which have been shown to differentially inhibit initiator and effector caspases (158). Whereas the caspase-9-inhibitory activity of XIAP was localized to the BIR3-RING domain (159), the BIR2 domain, together with the linker region between the BIR1 and BIR2 domains of XIAP, has been shown to inhibit the activation of caspase-3 and caspase-7 (160). In response to certain apoptotic stimuli, XIAP can be cleaved into two distinct fragments, an N-terminal fragment containing BIR1–2 and a second fragment containing BIR3-RING. The BIR1–2 fragment has diminished ability to inhibit caspase-3 and may be susceptible to further caspase-mediated degradation, whereas the BIR3-RING fragment is more stable and retains the ability to inhibit caspase-9, but is unable to suppress Fas-induced apoptosis (150).
XIAP is highly expressed in villous cytotrophoblasts and syncytiotrophoblasts of first-trimester placentas, and its expression significantly decreases in third-trimester placentas (90), which correlates with the concomitant increase in placental apoptosis (12, 13). Similarly, a decrease in XIAP expression in primary trophoblast cells isolated from first-trimester placentas either by treatment with Phenoxodiol, which down-regulates XIAP, (99, 161) or by XIAP RNA interference (S. L. Straszewski-Chavez and G. Mor, unpublished observations) allows the release of caspase-9 and caspase-3 and the induction of trophoblast cell apoptosis. Furthermore, the down-regulation of XIAP renders first-trimester trophoblast cells sensitive to Fas-mediated apoptosis, as evidenced by the decrease in trophoblast cell viability and an increase in the activation of caspase-8, caspase-9, and caspase-3 (99) (Fig. 4). Therefore, we hypothesize that XIAP protects trophoblast cells from the proapoptotic effect of caspase-8 activity during cytotrophoblast differentiation by inhibiting caspase-9 and caspase-3 activation and promoting the formation of the syncytium. If the expression of XIAP decreases, as in villous cytotrophoblasts and syncytiotrophoblasts of third-trimester placentas (90), the protective effect of XIAP is removed and an increase in villous trophoblast apoptosis is observed (12, 13). Similar changes in XIAP expression and function may contribute to pathological conditions (S. L. Straszewski-Chavez and G. Mor, unpublished observations) and may be associated with the increase in trophoblast apoptosis observed in complicated pregnancies (15, 16, 17, 18, 19, 20).
|
and 5). The appearance of the XIAP p30 fragment correlates with an increase in caspase-9 and caspase-3 activity and may represent a dominant negative form of XIAP, which interferes with the function of the active, full-length form of XIAP (162, 163). How the expression of the XIAP p30 fragment and the full-length form of XIAP are regulated in trophoblast cells will be discussed in the next section.
|
In relation to the other members of the IAP family, Gill and Hunt (127) demonstrated that c-IAP2 expression levels decrease in cytotrophoblast cells after treatment with both IFN- and LIGHT, suggesting that c-IAP2 may protect trophoblast cells from LIGHT-mediated apoptosis, but this result was not definitive. Future studies are needed to determine whether the other members of the IAP family can inhibit trophoblast apoptosis and how each of the IAPs is regulated in trophoblast cells.
D. Inhibition of IAPs
In order for apoptosis to occur, the inhibitory effect of IAPs must be removed, which is mediated by a group of proteins that antagonize the anticaspase activity of the IAP family. Members of this group include second mitochondria-derived activator of caspase/direct IAP-binding protein with low-PI (Smac/DIABLO), high-temperature requirement protein A2 (OMI/HtrA2) and XIAP-associated factor 1 (XAF1) (Fig. 6). Whereas XAF1 is a proapoptotic nuclear protein (167), Smac/DIABLO and OMI/HtrA2 are mitochondrial-associated proteins that are released from mitochondria upon certain apoptotic stimuli to inhibit IAP function (51, 52, 53), albeit by different mechanisms. Both Smac/DIABLO and OMI/HtrA2 bind to the BIR domains within IAPs, thereby displacing caspases and allowing caspase activation, however, OMI/HtrA2 is also a serine protease that can cleave IAPs, rendering them inactive (168, 169).
|
). Although the expression of the full-length form of XIAP was reduced, the expression of the p30 XIAP cleavage product was higher in term placentas than in first-trimester placentas (Fig. 5) (99). In vitro experiments have shown that the overexpression of XAF1 in first-trimester trophoblast cells results in an increase in trophoblast cell apoptosis, characterized by the activation of caspase-9 and caspase-3 (170). Similar to XAF1 expression in term placentas, the overexpression of XAF1 in first-trimester trophoblast cells also correlated with the appearance of the p30 XIAP fragment of XIAP. Furthermore, treatment with TNF- induced the expression and translocation of XAF1 from the nucleus to the cytoplasm, resulting in XIAP cleavage and the activation of caspase-9 and caspase-3 in trophoblast cells (115). Therefore, the proapoptotic effect of XAF1 appears to depend on the cleavage of XIAP. However, XAF1 is not known to be catalytic, suggesting that XAF1 induces XIAP cleavage indirectly. Because XAF1 localizes to mitochondria before XIAP cleavage (Fig. 7), we hypothesize that XAF1 may induce the release of OMI/HtrA2 from mitochondria, resulting in the cleavage of XIAP. This suggests that XAF1, like p53, may represent a proapoptotic link between the nucleus and the mitochondrial pathway. Our group is currently focused on determining whether OMI/HtrA2 does play a role in XAF1-induced XIAP cleavage and whether the other IAP inhibitor, Smac/DIABLO, is important for trophoblast apoptosis.
|
|
V. Exogenous Regulation of Trophoblast Apoptosis |
|---|
A. T helper 1 (Th1)- and T helper 2 (Th2)-type cytokines
Several years ago Wegmann et al. (172) hypothesized that protection from maternal immune rejection may be due to the production of Th2-type cytokines at the maternal-fetal interface during pregnancy. Since this observation, numerous studies have been published in support of the notion that the predominance of Th2 over Th1 cytokines is essential for a successful pregnancy (173, 174, 175). This shift in cytokine profile not only promotes immune protection (176), but it may also directly regulate trophoblast survival (177). One of the possible mechanisms by which cytokines may affect trophoblast survival is by regulating the expression and/or function of components of the apoptotic cascade. Th1, proinflammatory cytokines such as TNF- and IFN-, may induce trophoblast apoptosis by up-regulating the expression of XAF1 (115) or Fas (97) in cytotrophoblast cells, thereby increasing caspase-3 activity (97, 108). In contrast, treatment with IL-10, a Th2, antiinflammatory cytokine, may inhibit the proapoptotic effect of TNF- and IFN- treatment by increasing FLIP expression in trophoblast cells (97). Interestingly, IL-10 has also been shown to up-regulate the expression of FasL in first-trimester trophoblast cells (97), which we postulate may increase FasL secretion and, therefore, promote trophoblast immune protection (Fig. 8A). Because first-trimester trophoblast cells are normally resistant to Fas-mediated apoptosis, IL-10-induced FLIP expression may protect against the autocrine or paracrine effects of FasL secretion. Moreover, an increase in FasL expression may further promote trophoblast survival by inducing Fas-mediated NF-B activation (100, 101, 102).
|
/IFN--induced apoptosis in cytotrophoblasts and syncytiotrophoblasts from third-trimester placentas, but the mechanism by which this occurs is still unclear (178). It is known, however, that treatment with EGF, as well as TNF- or IFN-, does not affect the expression of Bcl-2 (179) and that EGF protects from ceramide and acid sphingomyelinase (A Smase)-induced trophoblast apoptosis by decreasing the level of ceramide, a lipid-derived second messenger in the sphingomyelin pathway that is associated with TNF- and Fas-mediated apoptosis in trophoblast cells (180). In addition, basic fibroblast growth factor (bFGF), IGF-I, and platelet-derived growth factor have all been shown to provide partial protection against TNF- and IFN- treatment, but the mechanisms of action remain to be elucidated (181). Studies by Torry and associates (J. Arroyo, S. L. Straszewski-Chavez, R. J. Torry, G. Mor, and D. S. Torry, submitted) have demonstrated that placenta growth factor (PlGF) may protect trophoblast cells from serum deprivation by increasing the expression of several antiapoptotic proteins, including Bcl-2, Mcl-1, and XIAP, to prevent caspase activation, while down-regulating the expression of apoptosis-inducing factor, a mitochondrial-associated protein that translocates to the nucleus to cause caspase-independent chromatin condensation and DNA fragmentation. Furthermore, PlGF may also induce the activation of MAPK pathways such as c-Jun-N terminal kinase /stress-activated protein kinase, and p38 kinase, but not the ERK-1 and ERK-2 pathways in trophoblast cells in response to growth factor withdrawal (182). Altogether, this network of different cytokines and growth factors may promote or inhibit trophoblast apoptosis by influencing the expression of either pro- or antiapoptotic factors in the apoptotic cascade or by activating survival pathways. Understanding how these growth factor and cytokine networks are regulated during pregnancy may help elucidate the pathophysiology of pregnancy complications and provide new therapeutic approaches for the treatment of pregnancy-related diseases.
Top
Abstract
I. Introduction
