Diverse Functions of HBEGF During Pregnancy

SUMMARY

The establishment of pregnancy requires an intimate physical interaction and a molecular dialogue between the conceptus and the maternal reproductive tract that commences at implantation and continues until the placenta is formed and fully functional. Failure of the regulatory processes that ensure the fidelity of this relationship can precipitate a catastrophic pregnancy loss. One of the earliest identified molecular mediators of blastocyst implantation is heparin-binding epidermal growth factor (EGF)-like growth factor (HBEGF), which signals between the endometrium and implanting trophoblast cells to synchronize their corresponding developmental programs. HBEGF expression by trophoblast cells of the developing placenta appears to regulate extravillous differentiation and provide cytoprotection in a sometimes-hostile environment. This versatile member of the EGF signaling system will be examined in light of its associations with key events during early pregnancy.

INTRODUCTION

Each month, the uterus prepares to receive a viable blastocyst under the direction of estrogen (E2) and progesterone (P4). The sex steroid hormones induce many changes in the endometrium that enhance its receptivity to an implanting blastocyst (Dey et al., 2004). Among these changes is the expression of secreted and membrane-bound signaling factors that influence embryo adhesion necessary for attachment to the luminal lining and later for invasion into the decidualized stroma. The embryo is also a source of several important factors, some of which are produced to augment the preparation of the endometrium. The concurrent developmental programs of the embryo and endometrium influence each other to synchronize their progression and produce an optimal environment for implantation. One critical factor that appears around the time of implantation is heparin-binding epidermal growth factor (EGF)-like growth factor (HBEGF). Evidence supporting a central role for HBEGF during implantation in mice is abundant and was recently reviewed (Lim and Dey, 2009). The cellular processes and regulatory factors that mediate the developmental program during blastocyst implantation have also been reviewed previously (Armant et al., 2000; Armant, 2005). The focus of this article is on the many timely influences of HBEGF and the EGF signaling system upon trophoblast cells in the peri- and post-implantation environment as pregnancy is established.

HBEGF and other members of the EGF family are synthesized as transmembrane proteins (e.g., proHBEGF) that can signal to their receptors on adjacent cells (juxtacrine) or become secreted through the shedding activity of metallo-proteinases that cleave the extracellular domain (Riese and Stern, 1998). The secreted form (e.g., sHBEGF) is free to bind receptors on the same cell (autocrine) or distant cells (paracrine). Processing of proHBEGF to sHBEGF and subsequent signaling through its receptors is illustrated in Figure 1. EGF family growth factors, operating through four HER/ErbB receptor tyrosine kinases, induce receptor dimerization and autophosphorylation, leading to downstream signaling that involves an extensive array of pathways (Holbro and Hynes, 2004). HBEGF requires heparan sulfate proteoglycan (HSPG) as a cofactor for binding to its receptors, HER1 (a.k.a. EGF receptor) and HER4. Although HBEGF does not bind directly to HER2 or HER3, they can participate in downstream signaling through heterodimerization with other ligated HER proteins.

Figure 1.

HBEGF processing and signaling. In its transmembrane form, proHBEGF is proteolytically cleaved by metalloproteinases at the cell surface to remove (scissors) its pro-domain and secrete the extracellular domain (sHBEGF shedding). sHBEGF is then free to diffuse to nearby receptor tyrosine kinases where it can bind either HER1 (EGF receptor) or HER4. Receptor binding requires heparin sulfate proteoglycan (HSPG) as cofactor, and induces HER homo- or heterodimerization. Receptor autophosphorylation initiates down-stream intracellular signaling. Heparin sulfate proteoglycan (HSPG); human epidermal growth factor receptor family member (HER); matrix metalloproteinase (MMP); a disintegrin and metalloproteinase (ADAM).

HBEGF performs a variety of functions in different cell systems (Raab and Klagsbrun, 1997). In its transmembrane form, proHBEGF appears to bind HER1 orHER4 of adjacent cells, participating in cell attachment or eliciting further signaling (Higashiyama et al., 1995; Raab et al., 1996). HBEGF induces chemotaxis and mitosis of NIH 3T3 cells through HER1 signaling (Elenius et al., 1997). Using a variety of carcinoma cell lines, it was shown that p53 expression induced by DNA damage subsequently upregulates HBEGF protein, which utilizes both MEK/ERK and PI3K/Akt pathways to inhibit cellular apoptosis (Fang et al., 2001). It also inhibits apoptosis in ovarian cancer, breast cancer, gastric cancer, melanoma, and glioblastoma cells by activating HER1 and downstream ERK (Yotsumoto et al., 2008). Many of the diverse cellular activities regulated by HBEGF transpire in the course of embryonic development as the placental tissues emerge.

EARLY CYCLICAL REGULATION OF HBEGF IN THE ENDOMETRIUM

Rodents

It was first determined that HBEGF plays a central role in the establishment of pregnancy in experiments conducted using mice and rats. HBEGF expression is regulated in the mouse endometrium by leukemia inhibitory factor (LIF) (Song et al., 2000), a cytokine required for implantation (Stewart et al., 1992), as well as by P4 and E2 (Wang et al., 1994). LIF is necessary for the expression of HBEGF, amphiregulin, and epiregulin mRNA in the luminal epithelium of the mouse uterus surrounding a newly implanted blastocyst, as evaluated in LIF^−/− mice (Song et al., 2000). In both mice and rats, E2 upregulates HBEGF mRNA and protein levels in the luminal epithelium, whereas P4 and E2 together upregulate its expression in the stroma (Wang et al., 1994; Zhang et al., 1994a,b). This indicates that HBEGF mRNA is transcribed in vivo under conditions that normally stimulate cellular proliferation in the endometrium (Zhang et al., 1994b). It is important to note, however, that this hormonal regulation of HBEGF in the nonpregnant uterus does not necessarily trigger its expression during pregnancy, but may prepare the endometrium to respond to additional signals that induce HBEGF expression in specific cell populations.

The regulation of HBEGF in stromal cells suggests that it has a unique role in preparing the uterus for decidualization, in which stromal cells accumulate lipid and glycogen vacuoles and increase their secretion of a fibrous matrix. Near the implanting blastocyst, there is an increase in vascular permeability and the expression of genes such as Bmp-2, Fgf-2, and Wnt-4 (Farrar and Carson, 1992; Abrahamsohn and Zorn, 1993; Paria et al., 2001). Beads coated with HBEGF can induce local increases in vascular permeability and upregulate expression of Bmp-2, suggesting that HBEGF participates in preparing the uterus to receive an implanting blastocyst (Paria et al., 2001). In cultured mouse stromal cells, HBEGF induces decidualization-like changes, including the upregulation of cyclin D3 and the induction of stromal cell polyploidy (Tan et al., 2004).

Humans and Nonhuman Primates

HBEGF appears to regulate the endometrial cycle in other mammals, including humans and nonhuman primates. In contrast to rodents, in which sex steroids prepare the uterus for eventual HBEGF expression, E2 and P4 stimulate HBEGF expression in human endometrium prior to the appearance of a blastocyst (Lessey et al., 2002), and, in baboons, it is under the control of P4 (Leach et al., 2001). As a result of its regulation by steroids, HBEGF accumulates in a cyclical fashion. During the late proliferative phase until the early secretory phase of the menstrual cycle in humans, HBEGF is predominantly localized in the stromal compartment of the endometrium (Yoo et al., 1997; Leach et al., 1999; Chobotova et al., 2002a). Although transcript is present in the luminal and glandular epithelia, protein levels do not appreciably accumulate at this time (Yoo et al., 1997; Leach et al., 1999). These findings are consistent with evidence from baboons where protein and message levels rise in the stroma during the proliferative phase and decline by day 5 post-ovulation (Leach et al., 2001). The correlation of HBEGF with E2 levels is strengthened by ex vivo evidence demonstrating that HBEGF transcription is induced by E2 in stromal cells. Proliferative phase human stromal cells, when treated with either E2 or P4, upregulate transcription of HBEGF, though if treated with both steroids together, expression will appear in both stroma and the luminal epithelial cells (Lessey et al., 2002). When interpreted in light of the menstrual cycle, these data suggest that the high E2 levels prior to ovulation induce HBEGF expression in stromal cells.

In the early secretory phase, HBEGF expression is induced in the luminal epithelium by the combination of E2 and P4 present at that time. Consistent with this idea, HBEGF expression decreases in the stromal compartment during the early secretory phase, coincident with decreasing E2 levels, but increases in both luminal and glandular epithelial cells (Yoo et al., 1997; Leach et al., 1999). HBEGF mRNA levels peak in the luminal and glandular epithelia during the mid-secretory phase, just prior to the “window of implantation” (Yoo et al., 1997). Protein levels follow shortly thereafter, rising in the early secretory phase and peaking during the window of implantation (Days 19–22) when the uterus is optimally receptive to a blastocyst (Leach et al., 1999). Baboons display a similar pattern in which HBEGF protein levels are highest 5–10 days post-ovulation (Leach et al., 2001). In rhesus monkeys, HBEGF increases during the proliferative phase, peaks during the window of implantation, and declines thereafter (Yue et al., 2000). Its maximal expression during the mid-secretory phase supports the view that HBEGF is a key regulator of blastocyst implantation. Indeed, HBEGF is localized on the apical surface of human luminal epithelial cells and on the surface of pino-podes when it is at maximal levels in the human endometrium (Yoo et al., 1997; Stavreus-Evers et al., 2002).

The decidualization reaction in humans is completely under the control of the sex steroids (Kodaman and Taylor, 2004), as opposed to rodents that have an estrous cycle that depends upon cues from the embryo (Dey et al., 2004). The human menstrual cycle prepares the uterus for implantation prior to arrival of an embryo, beginning in the early secretory phase, and reaching prominence by the window of implantation. In humans, several stimuli that induce stromal cell decidualization appear to do so through upregulation of HBEGF. If stromal cells are treated with 8-Br-cAMP, which is known to artificially induce decidualization, the soluble form of HBEGF is upregulated, as are its receptors, HER1 and HER4 (Chobotova et al., 2005). HBEGF then induces stromal cell production of prolactin and insulin-like growth factor binding protein-1, which naturally induce decidualization (Chobotova et al., 2005). Important for decidualization, HBEGF stimulates stromal cell growth via HER1, though not through HER4 (Chobotova et al., 2002a). The proliferative activity of HBEGF could be the result of its upregulation of IL-11 secretion (Karpovich et al., 2003). When stromal cells are treated with transforming growth factor-β or tumor necrosis factor-α, HBEGF expression is induced, which subsequently protects against apoptosis (Chobotova et al., 2005). Both the transmembrane and soluble forms of HBEGF are induced by these treatments, suggesting that HBEGF could be acting in an autocrine, paracrine or juxtacrine fashion as a survival factor (Chobotova et al., 2005). In addition, tumor necrosis factor-α synergizes with HBEGF to stimulate cell growth by upregulating EGFR expression in stromal cell membranes (Chobotova et al., 2002a). More work is needed to determine which pathways are involved downstream of HBEGF signaling in stromal cells.

Decidualization and cytoprotection are only two of the important roles for HBEGF in the endometrium. HBEGF or conditioned medium from E2- and P4-treated stromal cells upregulates integrin β₃, LIF and HOXA10 mRNA in primary human luminal epithelial cells (Liu et al., 2007) and β₃ integrin mRNA in Ishikawa cells (Lessey et al., 2002). Each of the corresponding proteins contributes to a uterine luminal epithelial surface that is receptive to an implanting blastocyst. Integrin α_vβ₃ serves as an attachment factor for osteopontin (secreted phosphoprotein 1, SPP1). SPP1 and α_vβ₃ are localized to both the glandular and luminal epithelia during the mid- to late-secretory phase, presumably to mediate embryo attachment (Brown et al., 1992; Apparao et al., 2001; von Wolff et al., 2001; Carson et al., 2002; Kao et al., 2002; Nardo et al., 2002). LIF stimulates human embryo development to the blastocyst stage (Sargent et al., 1998). In mice, LIF expression in the uterus is essential for embryo implantation (Stewart et al., 1992), as is uterine expression of HOXA10 (Satokata et al., 1995). HOXA10 message and protein are both found in the human endometrium most abundantly during the mid-secretory phase (Taylor et al., 1998; Gui et al., 1999; Cermik et al., 2001; Li et al., 2002). It has been suggested that HOXA10 induction by P4 during the window of implantation leads to a block in the stromal cell cycle, facilitating decidualization (Qian et al., 2005). Thus, the stimulatory effects of HBEGF upon stromal cells appear to be important not only in decidualization, but also in the preparation of the uterine luminal epithelium for an attaching blastocyst.

THE ROLE OF HBEGF AT IMPLANTATION

Rodents

Expression of HBEGF in the pregnant mouse uterus is necessary for the timely and successful implantation of a blastocyst (Xie et al., 2007). Evidence suggests that HBEGF expression during pregnancy is not induced by sex steroids, as in the nonpregnant uterus (Wang et al., 1994), but by the presence of a blastocyst (Das et al., 1994b). HBEGF mRNA is expressed in the uterine epithelium exclusively at the site of implantation 6–7 hr before attachment occurs, followed by an accumulation of the protein during attachment (Das et al., 1994b). Physiological interactions between the endometrium and blastocyst can be examined by experimentally induced delayed implantation in ovariectomized mice treated with P4 (Yoshinaga and Adams, 1966). HBEGF expression does not occur during delayed implantation until the blastocyst is activated by administration of E2 (Das et al., 1994b), suggesting that the blastocyst induces HBEGF expression in the uterine luminal epithelium. Evidence from the hamster shows that HBEGF mRNA is expressed throughout the apical surface of the luminal epithelium prior to blastocyst hatching and implantation (Mishra and Seshagiri, 2000), but becomes localized to the epithelium surrounding an implanted blastocyst late on day 4 of gestation (Wang et al., 2002).

Several lines of investigation suggest that embryoderived HBEGF is at least one of the signals that direct peri-implantation expression of HBEGF in the uterus. Indeed, mouse and hamster blastocysts express HBEGF (Wang et al., 2002; Hamatani et al., 2004; Liu and Armant, 2004). Further confirmation comes from an experiment using beads coated with HBEGF that were placed in the uterus of pseudopregnant mice mated with vasectomized males (Paria et al., 2001). In this experiment, the HBEGF specifically induced endogenous uterine HBEGF gene expression, along with a local increase in vascular permeability and decidualization, comparable to changes arising from the presence of a blastocyst (Paria et al., 2001). Experiments using HBEGF-coated beads suggest that HBEGF produced by embryos induces the molecular and physiological changes in the uterus. These findings support the hypothesis that cross talk mediated temporally by HBEGF occurs between embryonic and maternal tissues in rodents. Its expression in the early preimplantation blastocyst (Wang et al., 2002; Liu and Armant, 2004) implies that trophectoderm-derived HBEGF could initially induce uterine HBEGF expression, along with other decidualization-related changes. Subsequently, uterine HBEGF secretion could advance trophoblast differentiation to an invasive phenotype and augment implantation. Additional work is required to distinguish the contribution of HBEGF from both of these sources and substantiate this hypothesis.

Evidence that HBEGF can advance the developmental program of the blastocyst is abundant and suggests that reciprocal signaling could exist between the embryo and uterus. Exogenous application of recombinant HBEGF matures rodent blastocysts in vitro, augmenting their ability to hatch from the zona pellucida, and accelerating trophoblast differentiation and adhesion competency (Das et al., 1994b; Wang et al., 2000). HBEGF increases the rate of hamster blastocyst hatching, adhesion competency, and trophoblast outgrowth (Mishra and Seshagiri, 2000; Seshagiri et al., 2002) and augments development of 8-cell rat embryos to the blastocyst stage (Tamada et al., 1999). Intraluminal injection of HBEGF induces implantation of rat embryos in a delayed implantation model (Tamada et al., 1999). Moreover, exposure of mouse embryos to HBEGF during culture before transferring blastocysts to the uterus significantly increases the number of implantation sites that eventually form (Lim et al., 2006).

In mice, HER4 trafficks to the trophectoderm surface late on day 4 and is a prerequisite for HBEGF signaling that accelerates trophoblast development (Wang et al., 2000). HBEGF can then bind HER1 and HER4 (preferentially to HER4) on the surface of mouse blastocysts in an HSPG-dependent fashion, leading to HER autophosphorylation (Das et al., 1994a; Paria et al., 1999; Wang et al., 2000; Lim et al., 2006). Lysophosphatidic acid also accelerates trophoblast outgrowth, and does so by transactivating HER1 and HER4 through Ca²⁺-dependent HBEGF trafficking and shedding (Liu and Armant, 2004). Cross-talk with other signaling pathways that advance blastocyst differentiation could similarly target HBEGF shedding to transactivate HER kinases. HBEGF signaling accelerates mouse blastocyst differentiation to an adhesive phenotype via Ca²⁺ influx through N-type voltage-gated channels and activation of protein kinase C and calmodulin (Wang et al., 2000). This signaling pathway induces trafficking of the integrin α₅β₁ subunit to the surface of trophectoderm cells where it mediates strong adhesion to fibronectin (Wang et al., 2000; Armant, 2005).

In addition to its ability to advance hatching and trophoblast differentiation, proHBEGF could directly mediate attachment of the trophectoderm to the luminal surface of the endometrium (Fig. 2). Cells engineered to express proHBEGF bind to day 4 mouse blastocysts, but not to delayed blastocysts (Raab et al., 1996), suggesting that proHBEGF, by binding to its receptors, is capable of supporting the attachment reaction during implantation. Binding between these cells and normal day 4 blastocysts is blocked by treating the blastocysts with heparinase or a synthetic peptide corresponding to the HSPG-binding domain of HBEGF (Raab et al., 1996). Furthermore, its potential binding partners HSPG, HER1 and HER4 are downregulated during delayed implantation, and are coordinately upregulated with HBEGF during activation with E2 (Paria et al., 1993, 1999; Smith et al., 1997). These reports provide intriguing information about the elusive mechanisms underlying diapause.

Figure 2.

Juxtacrine Functions of HBEGF. Transmembrane proHBEGF is depicted mediating attachment between the trophoblast cell of a blastocyst and the surface of the uterine luminal epithelium. HBEGF can potentially bind HSPG, HER1, or HER4, to both mediate cell adhesion and elicit juxtacrine signaling downstream of HER kinases.

Humans

In contrast to the rodent endometrium, which may require stimulation from an embryo for HBEGF expression (Paria et al., 2001), HBEGF expression in the human endometrium precedes the appearance of the blastocyst (Leach et al., 1999). As in other species, HBEGF can simultaneously function during human implantation as an attachment factor and a growth factor. In humans, HBEGF, HER1, and HER4 are present in peri-implantation blastocysts with HER4 more prominent in the trophectoderm (Chobotova et al., 2002b). This suggests that HBEGF present on the endometrium might bind its receptors on an apposed blastocyst, and vice versa. Indeed, mature human blastocysts adhere to immobilized proHBEGF, an interaction that is competitively inhibited by addition of sHBEGF (Chobotova et al., 2002b). Similar results were obtained when human blastocysts were applied to a fixed monolayer of CHO cells overexpressing proHBEGF. As a growth factor, HBEGF accelerates development of human embryos to the blastocyst stage and their subsequent hatching from the zona pellucida (Martin et al., 1998; Sargent et al., 1998). While LIF can only induce development of embryos up to the blastocyst stage, HBEGF influences their maturation at least through the hatching stage (Sargent et al., 1998). Although these studies are instructive, it remains to be demonstrated whether these diverse functions of HBEGF are imperative in vivo.

THE ROLE OF HBEGF IN HUMAN PLACENTATION

HBEGF is present throughout the course of human gestation. It is abundant in 1st trimester decidua, and in placental tissue during all three trimesters (Birdsall et al., 1996), though its presence in the syncytiotrophoblast (STB) layer is weak in first trimester placentas (Yoo et al., 1997). HBEGF is also expressed in villous and extravillous cytotrophoblast (CTB) from weeks 14–35 in normal placentas and those delivered preterm (Leach et al., 1999, 2002). Several members of the EGF family and HER tyrosine kinases are expressed in placental trophoblast populations (Hofmann et al., 1992; Tanimura et al., 2004), indicating that the EGF signaling system is active during placental development.

Structure of the Early Placenta

As trophoblasts begin to invade into the endometrium, they differentiate along two paths (Morrish et al., 1998; Hunkapiller and Fisher, 2008). CTBs are highly proliferative, mononuclear cells that can fuse to form multinucleated STB cells. The STB initially invades into the endometrium. At the periphery of the conceptus, the chorionic villi emerge with a vascularized mesenchymal core and a stratified epithelium composed of CTB and an outer layer of STB. The STB transports nutrients, gases, and waste products between the maternal blood and fetal blood in capillaries of the chorionic villi. A few weeks after implantation, anchoring villi appear at the surface of the endometrium (decidua basalis) as CTBs penetrate through the STB layer and assemble into dense cellular columns. The most distal CTBs differentiate to an extravillous phenotype that invades interstitially. Upon reaching uterine blood vessels, the CTBs invade and remodel them, converting the spiral arteries into highly dilated vessels (Pijnenborg et al., 1983). Endovascular trophoblast cells produce proteins characteristic of endothelial cells in a process known as pseudovasculogenesis (Blankenship and Enders, 1997; Zhou et al., 1997a,b). Invasive CTBs remodel the maternal spiral arteries into permanently dilated structures as far as the first third of the myometrium (Pijnenborg et al., 1983). Maternal blood is thus directed to the inter-villous space at increased flow rates (Hunkapiller and Fisher, 2008). As trophoblasts differentiate to extravillous and endovascular phenotypes, they undergo a process known as integrin-switching in which differentiation is accompanied by changes in expression of adhesion molecules (Damsky et al., 1994; Zhou et al., 1997a,b). Different integrins mediate attachment, invasion, and vascular remodeling. For example, α₆β₄ is expressed by nonmotile villous CTBs, while invasive interstitial trophoblast cells express α₁β₁. CTBs of the columns in anchoring villi express α₅β₁ with differing levels of the other two integrins, depending on whether they have a proximal or distal position within the anchoring villi (Damsky et al., 1994). Primary CTB cultures from 1st trimester placentas undergo integrin switching during extravillous differentiation on Matrigel basement membrane, providing a useful experimental model.

Challenges of the Maternal Environment

Blastocyst implantation takes place in a uterine environment low in oxygen (~18 mm Hg or 2%), a condition that appreciably changes (~60 mm Hg or 8%) only after the 10th week of gestation (Rodesch et al., 1992; Burton et al., 1999; Jauniaux et al., 2001; Burton and Jauniaux, 2004). Prior to the 10th week, invading endovascular trophoblasts occlude maternal blood vessels, leaving the intervillous space relatively hypoxic. Human trophoblast cells are programmed to function under these conditions, and, in fact, proliferate at higher rates when cultured at 2% O₂ (Genbacev et al., 1996, 1997). Furthermore, localized areas within the placenta are temporarily exposed to oxygenated maternal blood during this early developmental period, but they are able to survive these reoxygenation episodes (Hung and Burton, 2006). Perhaps more profound is the extensive introduction of fully oxygenated maternal blood that occurs around week 10 as trophoblastic plugs become dislodged. The placental unit normally survives this insult and trophoblast invasion continues into the second trimester (Norwitz et al., 2001). Elevation of oxygen leads to free radical production, and evidence suggests that excessive oxidative stress may precipitate placental pathologies such as preeclampsia (Hung and Burton, 2006). Free radicals are normally deactivated in the placenta by xanthine dehydrogenase/xanthine oxidase, which is elevated in placentas during labor, when there is an abundance of local oxygen fluctuations (Many and Roberts, 1997). Clearly, mechanisms exist that protect trophoblasts from erratic oxygen levels.

HBEGF Mediates Trophoblast Motility and Survival

HBEGF stimulates trophoblast invasion from first trimester chorionic villous explants cultured on Matrigel basement membrane, but has little effect on CTB cell proliferation (Leach et al., 2004). As gestation progresses, there is a decrease in trophoblast capacity to differentiate along the extravillous pathway (Librach et al., 1991; Damsky et al., 1994). EGF augments the invasiveness of cultured first trimester CTBs, but with less effectiveness when CTBs are obtained later in gestation (Bass et al., 1994). In HT-H cells (human embryonal carcinoma that spontaneously differentiate into trophoblasts), induction of HER1 phosphorylation by EGF or HBEGF requires derepression of the receptor by trophinin (Sugihara et al., 2007), a transmembrane protein expressed in both human trophoblasts and the uterine epithelium that is capable of mediating cell adhesion through homophilic binding (Suzuki et al., 1999). When ligated, trophinin interaction with the cytoplasmic protein bystin is disrupted, releasing bystin inhibition of HER autophosphorylation (Sugihara et al., 2007). Therefore, HBEGF can activate HER downstream signaling only when opposing trophinin proteins are engaged by opposing cells. The relationship between the trophinin and EGF signaling systems could be particularly important for trophoblast invasive differentiation during blastocyst implantation.

Cell lines generated from immortalized human cytotrophoblasts have proven to be highly useful for experimental investigation of implantation and early placentation. HTR-8/SVneo cells are an immortalized human CTB cell line originating from first trimester villous explants that characteristically secrete chorionic gonadotropin and become invasive in conjunction with appropriate integrin switching and upregulation of HLA-G (Graham et al., 1993; Kilburn et al., 2000). Therefore, HTR-8/SVneo cells closely resemble the primary CTB cells from which they are derived, providing a robust experimental model. HBEGF, EGF, and TGFA each induce integrin-switching in HTR-8/SVneo cells by signaling through HER1 or HER4 to increase the α₁ integrin subunit and decrease α₆ (Leach et al., 2004). As a result, there is an increase in trophoblast cell migration and invasive activity (Fig. 3). The extensive expression of HBEGF in trophoblasts, particularly within extravillous populations (Leach et al., 2002), could be vital for their invasive activities during the establishment of pregnancy and physiological conversion of spiral arteries. As gestation continues and trophoblasts loose their capacity for extravillous differentiation (Damsky et al., 1994; Librach et al., 1991), the involvement of HBEGF and the EGF signaling system in trophoblast invasion wanes (Bass et al., 1994). However, other roles for HBEGF could come into play late in gestation.

Figure 3.

HBEGF stimulation of trophoblast invasion. HBEGF activates HER kinases to enhance cytotrophoblast (CTB) motility and invasion. Other members of the EGF family, epidermal growth factor (EGF) and transforming growth factor alpha (TGFA) have similar capacities. Little is known about the responsible downstream signaling pathways, but there are associated changes in the expression of integrins α₆β₄ and α₁β₁ (integrin switching) by CTBs. Ligation of either HER1 or HER4 can induce integrin switching.

In vivo, placentation initiates in a relatively low oxygen environment. Although CTBs proliferate faster at low O₂ concentrations, invasion is more aggressive at higher O₂ (Genbacev et al., 1996, 1997). Therefore, pseudovasculogenesis is initially moderate, and rapid CTB growth leads to occlusion of the spiral arteries. It has not been clear how trophoblast tissues survive during this hypoxic phase. We have observed that HBEGF, but not other EGF family members, is upregulated in the HTR-8/SVneo CTB cell line when exposed to2% O₂ (Armant et al., 2006). Using specific antagonists of HBEGF signaling, it can be demonstrated that HBEGF inhibits apoptosis due to hypoxia, but has no effect on proliferation rates enhanced by low O₂. This study also established that cytoprotection requires metalloproteinase shedding of HBEGF and binding to either HER1 or HER4 (Fig. 4A). Quantification of mRNA and protein levels indicated that HBEGF increases 100-fold through a post-transcriptional mechanism. Antagonizing HBEGF signaling prevents its upregulation, suggesting that low levels of resident proHBEGF are cleaved through activation of metalloproteinases during hypoxia to initiate its synthesis downstream of HER kinases. A recent study suggests that ADAM-17 may be responsible for initiating this signaling event in vivo (Hung et al., 2008). Further studies are needed to determine which metalloproteinases are directly responsible for HBEGF shedding in trophoblast cells. In contrast to 1st trimester CTBs, trophoblasts in villous explants from term placentas do not elevate HBEGF levels in response to low oxygen and their survival is compromised by hypoxia (Imudia et al., 2008). However, addition of HBEGF to term villous explants cultured at 2% O₂ inhibits trophoblast apoptosis, demonstrating its persistent role as a survival factor. These findings highlight the importance of HBEGF in the success of early placentation and in protecting trophoblast cells from the damaging effects of hypoxia encountered throughout gestation.

Figure 4.

HBEGF and changing oxygen levels in the placenta. In (A), HBEGF is secreted by first trimester CTBs cultured at low O₂. Evidence suggests that low O₂ activates metalloproteinases that cleave proHBEGF, releasing sHBEGF to signal through its receptors, HER1 or HER4. Downstream signaling activates a post-transcriptional mechanism to increase HBEGF synthesis. In (B), the cytoprotective activity of HBEGF in CTBs is depicted. Stress due to low O₂, reoxygenation injury, or ethanol exposure increases CTB apoptosis. Accumulation of endogenous sHBEGF at low O₂ (as in A), or induction of the EGF signaling system by exogenous HBEGF, EGF or TGFA activates signaling downstream of HER kinases that inhibits apoptosis.

Using placental villous explants, Hung et al. demonstrated that reactive oxygen species are generated in an in vitro reoxygenation injury model (Hung et al., 2001), leading to apoptosis of the trophoblasts (Hung et al., 2002). HBEGF can prevent apoptosis induced by reoxygenation injury in HTR-8/SVneo cells (Fig. 4B) by signaling through its receptors, HER1 and HER4 (Leach et al., 2008). Although HBEGF is downregulated within 30 min of elevating the O₂ concentration from2% to 20%, accumulated sHBEGF generated by trophoblasts at low oxygen may protect against sudden exposure to oxygenated maternal blood in the uterine environment. Ethanol, a teratogenic substance with oxidative qualities, also induces apoptosis in HTR-8/SVneo cells (Wolff et al., 2007). When HTR-8/SVneo cells are exposed to ethanol and HBEGF together, apoptosis is inhibited in a HER1- or HER4-dependent fashion. During both reoxygenation injury and ethanol exposure, other members of the EGF family are cytoprotective to varying degrees (Wolff et al., 2007; Leach et al., 2008). In contrast, only HBEGF inhibits apoptosis in CTBs cultured at low oxygen (Armant et al., 2006). It has been proposed that insults such as reoxygenation injury are not always tolerated and can lead to placental pathologies (Hung and Burton, 2006). The activity of the EGF signaling system could be a critical variable that keeps oxidative stress in check.

Clinical Considerations

Heparin has long been used to treat pregnant women with certain thrombophilias (Greer, 2003), but its impact on trophoblasts has only recently been investigated. Since heparan sulfate is an obligate cofactor for HBEGF, clinical application of heparin during pregnancy could influence the activity of this growth factor. Indeed, unfractionated heparin inhibits apoptosis induced by diverse signals in 1st trimester cytotrophoblast cells (Hills et al., 2006). The same down-stream signaling pathways activated by heparin were also induced by treating the cells with HBEGF. It could be speculated that addition of its cofactor activates HBEGF in trophoblasts to enhance survival. On the other hand, fractionated heparin, also used as an anticoagulant, suppresses the invasive capacity of primary extravillous trophoblasts (Ganapathy et al., 2007). This suggests that there could be clinical advantages to using unfractionated heparin over fractionated heparin.

A pathological reduction of HBEGF expression is observed in placentas of patients with the hypertensive syndrome preeclampsia (Leach et al., 2002), which is associated with decreased trophoblast invasion (Brosens et al., 1972) and increased levels of trophoblast apoptosis (DiFederico et al., 1999; Allaire et al., 2000). This finding led us to hypothesize that dysregulation of HBEGF and the resulting deficiencies in trophoblast function could contribute to preeclampsia, although it remains to be determined when during the course of gestation HBEGF levels are first reduced in preeclamptic pregnancies. Trophoblast invasion is shallow in preeclampsia possibly due to a lack of HBEGF-induced cell migration and a rise in apoptosis, exacerbated by reduced cytoprotection. The increased oxidative damage to trophoblast cells (Redman and Sargent, 2005) and the paucity of HBEGF in preeclamptic placentas is consistent with the hypothesis that HBEGF performs an important role as a survival factor throughout gestation. Patients delivering preterm without hypertensive disorder produce placentas with normal levels of HBEGF (Leach et al., 2002). However, there was an intermediate reduction of HBEGF in extravillous trophoblasts for those women who delivered small for gestational age infants. Intrauterine growth restriction, like preeclampsia, has been linked to aberrant trophoblast invasion (Khong et al., 1986) and elevated apoptosis (Ishihara et al., 2002), suggesting shared elements in the etiology of both disorders. Premature placental perfusion in preeclamptic pregnancies (Jauniaux et al., 2003) could induce apoptosis by stressing trophoblast cells before HBEGF has sufficiently accumulated during the early hypoxic period. Because there are varying levels of severity of placental reoxygenation, it has been hypothesized that preeclampsia may be the result of sub-lethal levels of reoxygenation, leading to abnormal placentation, whereas more pronounced reoxygenation will terminate the pregnancy (Leach et al., 2008). More work is needed to determine if the etiology of this disease involves a lack of HBEGF production by trophoblasts or the activity of factors that suppress HBEGF signaling.

LOOKING AHEAD

HBEGF performs numerous functions during pregnancy that are conserved across mammalian species with divergent reproductive physiologies. In addition to preparing both the preimplantation embryo and uterus for their mutual interaction, HBEGF appears to directly facilitate the process of implantation. As development proceeds, HBEGF provides a stimulus for trophoblast invasion and serves as a critical survival factor in an environment subject to wide swings in oxygenation. Due to its roles in trophoblast cell differentiation and survival, dysregulation of HBEGF could have profound consequences, from complete pregnancy failure to the onset of obstetrical disorders. The association of its deficiency with preeclampsia is consistent with a critical role for HBEGF in human placentation. New insights into the contribution of HBEGF to placentation could be obtained with the mouse model, which has not been exploited. HBEGF-deficient transgenic embryos develop to term with heart defects (Iwamoto et al., 2003; Jackson et al., 2003), but this system has not been examined in depth for errors in placentation due to the absence of HBEGF. Additionally, the immune system undergoes unique changes during pregnancy that include tolerance of the fetal–maternal allograft and the recruitment of uterine natural killer cells, macrophages and dendritic cells to the implantation site. It is currently thought that these cells ensure proper implantation and that trophoblast cells help to orchestrate their immune function (Mor, 2008). However, the role of HBEGF in the trophoblast–immune relationship during pregnancy remains unexplored.

Tissue-specific differences in HBEGF function could stem from the microenvironment near targeted cells. Expression of its cofactor, HSPG, varies among cells and with differentiation. Other growth factors, such as LIF, can profoundly affect HBEGF expression, as seen in the peri-implantation uterus. HBEGF forms complexes with other transmembrane proteins, including CD9 and integrin α₃β₁ (Raab and Klagsbrun, 1997), but the relevance of these interactions for implantation and placentation has not been investigated. There is a need for better understanding of the molecular interactions that regulate HBEGF expression, as well as the genes induced by HBEGF in reproductive tissues.

The ability of HBEGF to regulate diverse outcomes is not well understood. Autocrine regulation of HBEGF is post-transcriptional in CTBs cultured at reduced O₂ levels (Armant et al., 2006). One possibility is that HBEGF signaling releases its own message from translational repression. Maintenance of a dormant, stable HBEGF mRNA pool could provide a reserve for rapid mobilization during episodes of hypoxia without a high cost in energy. Although the exact nature of this post-transcriptional regulation is not yet understood, it would be useful to know whether HBEGF signaling similarly targets other proteins for synthesis. Finally, an examination of the intracellular circuitry downstream of the HER kinases activated by HBEGF will be important for clarifying the molecular basis of its varied functions.