THE FIRST TWELVE WEEKS

2.1 DIFFERENTIATION OF THE TROPHOBLAST

2.1.1. Human trophoblasts in vivo: three differentiation pathways

Trophoblasts are unique cells derived from the outer cell layer of the blastocyst which mediate implantation and placentation (Hertig and Rock, 1956). Depending on their subsequent function in vivo, undifferentiated cytotrophoblasts can develop into 1) hormonally active villous syncytiotrophoblasts, 2) extravillous anchoring trophoblastic cell columns, or 3) invasive intermediate trophoblasts (Fig 1). Interestingly, within the villi of the human placenta-at all gestational ages-exist a population of cytotrophoblasts which remain undifferentiated. The purpose of this chapter is to review our knowledge about the various differentiated functions of human trophoblasts and to suggest control mechanisms for these differentiation pathways. Furthermore, by studying the differentiation behavior of villous cytotrophoblasts in culture, it is now possible to develop in vitro correlates for elucidating the biology of the human trophoblast.

2.1.1.1. Villous syncytiotrophoblast

The hormones secreted by the villous syncytiotrophoblast are critical for maintaining pregnancy (Conley and Mason, 1990; Petraglia et al. 1990a). Early in gestation, human chorionic gonadotropin (hCG) is essential to maintain corpus luteum progesterone production. Near the end of the first trimester, the mass of villous syncytiotrophoblast is large enough to make sufficient progesterone and estrogen to maintain the pregnancy. During the third trimester, large quantities of placental lactogen are produced, a hormone purported to have a role as a regulator of lipid and carbohydrate metabolism in the mother. Other syncytiotrophoblast products include: pregnancy specific ß1-glycoprotein (Kliman et al. 1986), plasminogen activator inhibitor type 2 (Feinberg et al. 1989a), growth hormone (Jara et al. 1989), collagenases (Moll et al. 1990), thrombomodulin (Maruyama et al. 1985; Ohtani et al. 1989), and growth factor receptors (Kawagoe et al. 1990; Posner, 1974; Uzumaki et al. 1989). The factors responsible for the regulated synthesis of these compounds has been the subject of a great deal of investigations, some of which will be reviewed below.

2.1.1.2. Anchoring trophoblasts

The premature loss of attachment of the developing conceptus or placenta to the uterus can terminate the gestation. Therefore, the anchoring trophoblast cell columns and the extracellular matrix proteins which mediate this attachment are critical to the developing pregnancy. It has been generally accepted that some form of cell-extracellular matrix interaction takes place at the attachment interface between these trophoblasts and the uterus. Some have considered Nitabuch's layer related to this function. In addition to the anchoring cell columns of the placenta, the extravillous trophoblasts of the external membranes (chorion laeve), play a critical role in maintaining attachment of the external membranes to the endometrial surface. Recently, a specific type of fibronectin-trophouteronectin-has been implicated as the protein responsible for the attachment of anchoring, extravillous trophoblasts to the uterus throughout gestation (Feinberg et al. 1991).

2.1.1.3. Invading trophoblasts

As human gestation progresses, invasive populations of extravillous trophoblasts attach to and interdigitate through the extracellular spaces of the endo and myometrium. The endpoint for this invasive behavior is penetration of maternal spiral arteries within the uterus (Pijnenborg, 1990). Histologically, trophoblast invasion of maternal blood vessels results in disruption of extracellular matrix components and development of dilated capacitance vessels within the uteroplacental vasculature. Biologically, trophoblast-mediated vascular remodeling within the placental bed allows for marked distensibility of the uteroplacental vessels, thus accommodating the increased blood flow needed during gestation. Abnormalities in this invasive process have been correlated with early and mid-trimester pregnancy loss, preeclampsia and eclampsia, and intrauterine growth retardation (Pijnenborg et al. 1981; Pijnenborg, 1990; Roberts et al. 1989).

2.1.2. In vitro model systems to study trophoblast differentiation

The most commonly utilized approaches for examining the regulation of hormone production by trophoblasts has come from in vitro studies. Model systems developed to study placental and trophoblast function have included placental organ and explant culture, trophoblast culture, chorion laeve culture, choriocarcinoma cell line culture, and placental perfusion studies. Recently, most investigators have turned to trophoblast cell culture since it eliminates the complications of more heterogeneous cell systems. Since the cytotrophoblast is the precursor of all other trophoblasts, a variety of methods have been proposed to purify this cell type from the human placenta (e.g., Kliman et al. 1986; Belisle et al. 1986; Loke et al. 1989; Yagel et al. 1989a; Bax et al. 1989; Truman et al. 1989; Branchaud et al. 1990; Fisher et al. 1990; Shorter et al. 1990; Dodeur et al. 1990; Loke 1990).

Where does one find cytotrophoblasts? Cytotrophoblasts are present in the human placenta throughout pregnancy as undifferentiated mononuclear cells (Boyd and Hamilton, 1970). Cytotrophoblasts can be found as early as the blastocyst, through all stages of placental development, and within the chorion laeve of the extraplacental membranes. Few investigators have utilized human blastocyst trophoblasts because of practical and ethical limitations, but all other starting material has been utilized to purify the trophoblast stem cell, the cytotrophoblast. For example, Fisher et al. (1989) have utilized first trimester placentae to yield 1x106 cells per gram of starting material. Yagel et al. (1989a) have also started with first trimester placentae to purportedly produce passable trophoblast cell lines. Shorter et al. (1990) have started with external term membranes and reportedly separated out trophoblasts using flow cytometry and cell sorting. As Loke has pointed out (Loke, 1983), a critical part of trophoblast purification is often overlooked-adequate markers to prove what cell has been purified. This continues to be a valid caution, and must be considered when evaluating the efficacy of any particular purification protocol and the claims made by the investigators about the cells they are examining.

Beginning with third trimester placentae, we have been able to obtain almost pure cultures of human cytotrophoblasts (Kliman et al. 1986; Fig. 2). Processing 60 g of villous tissue with yields of 2-5 x 106 cells per gram generates up to 300 million cells. Using similar methods, cytotrophoblasts can also be purified from first and second trimester placentae. Heeding Loke's caution, we have asked: are these cells trophoblasts? We have utilized a variety of immunochemical markers, electron microscopic examination, and biochemical assay of secreted products to substantiate our claim that these cells represent cytotrophoblasts which are capable of differentiating along multiple trophoblast pathways (Kliman et al. 1986; Feinman et al. 1986; Kliman et al. 1987; Kao et al. 1988; Feinberg et al. 1989a; Kliman and Feinberg, 1990; Kliman et al. 1991). We have demonstrated by time-lapse cinematography that when these mononuclear cytotrophoblasts are placed in Dulbecco's Modified Eagles' Medium (DMEM) containing 20% (v/v) heat-inactivated fetal calf serum (FCS), they flatten onto the culture surface within 3-12 h, migrate towards each other to form aggregates within the first 24 h, and over the next 24 h of culture, form syncytiotrophoblasts (Fig. 3). Concomitant with these morphologic changes, the trophoblasts synthesize and secrete a number of cell products, including protein hormones: chorionic gonadotropin, placental lactogen, and pregnancy specific ß1-glycoprotein (Kliman et al. 1986; Feinman et al. 1986) and steroid hormones: estrogen and progesterone (Kliman et al. 1986). We and others have utilized these cells to elucidate the products of trophoblast differentiation and to explore the mechanisms by which their synthesis and secretion is regulated.

2.1.3. Trophoblasts as endocrine cells

Trophoblasts synthesize and secrete a vast array of endocrine products (for reviews see: Blay and Hollenberg 1989; Conley and Mason 1990; Petraglia et al. 1990a; Ringler and Strauss 1990; Sirinathsinghji and Heavens 1989).

2.1.3.1. Production and regulation of protein hormones

Chorionic gonadotropin. The most widely studied trophoblast product is chorionic gonadotropin. This glycoprotein is critical to pregnancy since it rescues the corpus luteum from involution, thus maintaining progesterone secretion by the ovarian granulosa cells. Its usefulness as a diagnostic marker of pregnancy stems from the fact that it may be one of the earliest secreted product of the conceptus. Ohlsson et al. (1989) have demonstrated by in situ hybridization that ß-hCG transcripts are present in human blastocyst trophoblasts prior to implantation. Placental production of hCG peaks during the tenth to the twelfth week of gestation, and tends to plateau at a lower level for the remainder of pregnancy. This difference in the rate of hCG secretion may be mimicked to some extent by trophoblasts cultured from first versus third trimester placentae. Kato and Braunstein (1990) have demonstrated that trophoblasts from first trimester placentae secrete greater amounts of hCG than trophoblasts purified from term placentae, suggesting that cultured trophoblasts may retain the regulatory effects of their in situ milieu even after several days of culture.

What regulates hCG synthesis and secretion in the trophoblast? Workers have attempted to answer this question by examining those factors that can regulate hCG synthesis and secretion in vitro. Table 1 summarizes our current knowledge of the regulatory factors that appear to modulate hCG secretion in cultured trophoblasts.

Human placental lactogen (hPL). This potent glycoprotein is made throughout gestation, increasing progressively until the 36th week, were it can be found in the maternal serum at a concentration of 5-15 µg/ml, the highest concentration of any known protein hormone. The major source of hPL appears to be the villous syncytiotrophoblasts, where it is made at a constant level throughout gestation (Sakbun et al. 1990a) In addition to the villous syncytiotrophoblast, hPL has been identified in invasive intermediate trophoblasts during the first trimester (Heyderman et al. 1981; Kurman et al. 1984). In addition to identifying hPL within trophoblasts in situ, cultured first trimester trophoblasts have been shown to secrete hPL in vitro (Dodeur et al. 1990). Sakbun et al. (1990a) have also identified hPL mRNAs in cultured trophoblasts. Hoshina et al. (1984), working with choriocarcinoma cell lines, have proposed that hPL gene expression occurs after a-hCG and ß-hCG gene expression, suggesting that hPL is a product of a more differentiated trophoblast. We have also shown that intracytoplasmic a-hCG appears prior to intracytoplasmic hPL in cultured term trophoblasts (Kliman et al. 1987).

The factors that regulate hPL synthesis and secretion are not as well studied as for hCG. Kato and Braunstein (1989) have demonstrated that the secretion of hCG and hPL are discordant during the first 5 days of term trophoblast culture, suggesting different regulatory pathways for these hormones. Dodeur et al. (1990) demonstrated that dibutyryl cAMP stimulated hPL secretion from cultured first trimester trophoblasts. Maruo et al. (1987) have shown that EGF, in addition to increasing hCG secretion by cultured human trophoblasts, also augments hPL secretion by these cells. Handwerger et al. (1987) showed that high density lipoproteins stimulate the release of hPL from human placental explants. Finally, Petit et al. (1989) have demonstrated that angiotensin II stimulates hPL release by cultured trophoblasts.

Prolactin. Al and Fox (1986) and Sakbun et al. (1987) have demonstrated by immunohistochemical studies that villous syncytiotrophoblasts contain prolactin. The significance of this finding is unclear at this time.

Relaxin. Sakbun et al. (1987), using anti-peptide antibodies, demonstrated immunoreactivity for the C-peptide and/or prorelaxin in villous cytotrophoblasts. More recently, Sakbun et al. (1990b) have demonstrated relaxin secretion by cultured trophoblasts. Trophoblast derived relaxin may, therefore, play an important role in maternal ECM modification as parturition approaches.

Chorionic adrenocorticotropin. An ACTH-like protein, lipotropin, and ß-endorphin have all been identified in placental extracts (Odagiri et al. 1979), presumably all derived from the common precursor pro-opiomelanocortin (Krieger, 1982). Liotta et al. (1977) demonstrated that ACTH is synthesized by cultured placental cells, and Al and Fox (1986) have demonstrated ACTH within villous syncytiotrophoblasts by immunohistochemistry. Mulder et al. (1986) demonstrated that isoproterenol stimulated ACTH secretion by placental explant cultures. The physiological role of placental ACTH is unclear. As with other placental hormones, it may represent a shift from maternal to placental control of this metabolic pathway.

2.1.3.2. Production and regulation of steroid hormones

Progesterone. The significance of placental elaboration of progesterone was revealed by Diczfalusy and Troen (1961), who showed that bilateral oophorectomy between 7 and 10 weeks of gestation had little impact on the conceptus or urinary pregnanediol levels. More recently, we have been able to demonstrate progesterone secretion by cultured term trophoblasts (Kliman et al. 1986). In addition we have identified various components of the steroidogenic machinery necessary for progesterone biosynthesis within cultured trophoblasts (Nulsen et al. 1989). Like hCG, progesterone synthesis and secretion seem to be upregulated by cAMP agonists (Feinman et al. 1986; Nulsen et al. 1989; Kliman et al. 1991).

Estrogen. The placenta does not have all the necessary enzymes to make estrogens from cholesterol, or even progesterone. Human trophoblasts lack 17a-hydroxylase and therefore can not convert C21-steroids to C19-steroids, the immediate precursors of estrogen. To bypass this deficit, dehydroisandrosterone sulfate from fetal adrenal is converted to estradiol-17ß by trophoblasts (Siiteri and MacDonald, 1966). Not surprisingly, trophoblasts contain the necessary enzymes to make this conversion (Conley and Mason, 1990), namely sulphatase, 3ß-hydroxysteroid dehydrogenase/Æ5Æ4-isomerase (3ßHSD), and aromatase. Lobo and Bellino (1989) have demonstrated that cultured trophoblasts synthesize aromatase, and that cAMP appears to stimulate aromatase production by these cells. Recently, Nestler (1990) demonstrated that insulin-like growth factor II inhibits aromatase in cultured human trophoblasts.

2.1.3.3. Production and regulation of hypothalamic-pituitary hormones

The placenta appears to produce a number of hypothalamic-pituitary hormones, including GnRH and CRH (for a recent review see Sirinathsinghji and Heavens, 1989). GnRH was first identified within villous cytotrophoblasts by immunochemical staining of intact placentae (Khodr and Siler-Khodr, 1978). More recently, Petraglia et al. (1990b) have demonstrated GnRH secretion by cultured trophoblasts and have shown that estrogen augments cAMP induction of trophoblast GnRH secretion. Corticotropin releasing hormone (CRH) is also made and secreted by cultured trophoblasts (Saijonmaa et al. 1988). Robinson et al. (1988) have demonstrated that glucocorticoids stimulate CRH release by cultured trophoblasts. Adding a further level of complexity to the regulatory signals impinging on the placenta, Petraglia et al. (1989b) have shown that neurotransmitters and peptides modulate the release of immunoreactive CRH and interleukin-1-ß increases both CRH and ACTH release from cultured human trophoblasts (Petraglia et al. 1990c). In addition to the hypothalamic factors GnRH and CRH, pituitary growth hormone is synthesized and secreted by first and third trimester cultured trophoblasts (Evain-Brion et al. 1990). It appears from these studies that the placenta, in addition to replacing much of the women's pituitary function during pregnancy, also replaces critical hypothalamic functions so as to maintain control and feedback loop mechanisms close to the conceptus.

2.1.4. Trophoblasts as attachment cells

Examination of the junction between the human placenta and uterus reveals a population of trophoblasts grouped together in dense masses between the chorionic villi and the uterine stroma (Fig. 4). These cells appear to grow out of the cytotrophoblast layer of the nearby chorionic villi. In addition to being extravillous trophoblasts, they appear to function specifically as the junctional apparatus of the conceptus. In some places they form elongated structures, often referred to as 'cell columns.' Are these cells simply proliferating villous cytotrophoblasts, or possibly invasive intermediate trophoblasts in transit to penetrate the uterus? Recent work suggests that these are truly a unique form of trophoblast, with their own differentiation markers (Feinberg et al. 1991).

Previous studies have identified fibronectin staining in a variety of locations in pregnancy tissue, including ECM of the placental-uterine junction, uterine stroma, connective tissue core of placental villi, fetal membranes, and in walls of fetal blood vessels (Earl et al. 1990; Yamada et al. 1987; Vartio et al.1987). These immunolocalization studies, while specific for fibronectin, were performed with antibodies which may have reacted with more ubiquitous, less well-characterized fibronectin epitopes. The specificity of staining that we have identified with FDC-6 (Fig. 4), unlike other anti-fibronectin antibodies, implicates the oncofetal domain or closely adjacent portions of the type-III repeat connecting segment as critical moieties associated with implantation and trophoblast attachment (Feinberg et al. 1991).

During the differentiation process in vitro, cytotrophoblasts have the ability to attach to types I and IV collagen, fibronectin and laminin. We have shown that these associations can be specifically reversed with RGD containing peptides and antibodies to fibronectin, suggesting that human cytotrophoblasts have specific receptors to ECM proteins (Kao et al. 1988).

2.1.5. Trophoblasts as invasive cells

When purified human trophoblasts were co-incubated with endometrial explants in a suspension-culture system (Kliman et al. 1990b), the trophoblasts attached to the endometrial fragments, showed evidence of invasion into the tissue, and induced zones of necrosis at the points of contact with the exposed stromal surfaces. These results suggested to us that these trophoblasts possessed the machinery necessary for invasion. The ability to attach to ECM proteins is the first step in a series of steps necessary for cellular invasion (Liotta et al. 1983).

In order to better focus on trophoblast-ECM interactions, without the added complexity of human endometria, we cultured trophoblasts with the ECM material Matrigel®. Matrigel is a mixture of solubilized basement membrane components containing laminin, type IV collagen, heparan sulfate proteoglycan, and entactin from mouse Engelbreth-Holm-Swarm tumor (Kleinman et al. 1986). In our initial system (Kliman and Feinberg, 1990a), Matrigel was layered into a Millicell filter assembly, trophoblasts were added to the top chamber and incubated for up to 7 days. After 120 h, there was evidence of significant trophoblast penetration into the Matrigel layer.

2.1.5.1. Matribeach: a biologic assay of ECM degradation and invasion.

The Millicell-Matrigel invasion system revealed that after 120 h human trophoblasts could invade into ECM material, based on detailed histologic examination of the invading trophoblasts in cross section. Although these findings gave us the impetus to continue our invasion studies, we were not able to readily manipulate or quantify trophoblast invasive behavior by the Millicell-Matrigel invasion system. As Albini et al. (1987) have shown, undiluted Matrigel can form a significant barrier to invasion, even by aggressive tumor cells. Since we were also interested in examining the morphologic and degradative features of trophoblasts during the early stages of the invasive process, we developed a novel slope of undiluted Matrigel in order to examine the characteristics of ECM invasion as a function of time and Matrigel thickness (Kliman and Feinberg, 1990a). This system had the advantage of allowing us to quantify trophoblast-ECM interactions, morphology, and proteolytic activity during the invasion process in vitro at the light and electron microscopic level. Using this thickness gradient of Matrigel with an 8_ slope (Matribeach), we discovered that ECM thickness itself affected trophoblast morphology as well as the ability of these cells to degrade the ECM. In zone 2 of the Matribeach (see below), we observed cellular processes not observed in the Millicell-Matrigel invasion system.

Cytotrophoblasts from first or third trimester placentae, cultured on Matribeach, yielded a typical pattern of trophoblast morphology (Fig. 5; Kliman and Feinberg, 1990a). After 48 h, the trophoblasts on the glass surface had flattened, aggregated, and begun to form syncytia, as has been described previously (Kliman et al. 1986). On the thin part of the beach-zone 1 (defined as the Matrigel extending from the visible edge to 4 µm in thickness)-the cells flattened out in a fashion very similar to the flattening on glass. On zone 2 (defined as the Matrigel from 4 to 14 µm in thickness), the cells caused marked clearing of the Matrigel by creating pericellular zones of lysis around the trophoblast aggregates (Fig. 6c). Scanning electron microscopy (Figs. 6a, b) revealed that the trophoblast groups on zone 2 progressively eroded the Matrigel until the glass surface was exposed. Zone 2 lysis was particularly pronounced after the Matribeaches were immunocytochemically stained for mouse laminin, which resulted in multiple areas of negative staining around the cells. The same effect was observed when first trimester trophoblasts were used. The cells on zone 3-Matrigel thicker than 14 µm-remained spherical, mostly single, and exhibited little or no zones of lysis. These results were confirmed with multiple primary cultures of trophoblasts. Time course studies revealed that the basic separation of zones occurred by 24 h and was continued to be observed through 72 h of culture. The trophoblasts in all three zones were viable, as assayed by their positive staining for aand ß-hCG.

2.1.5.2. Effect of 8-bromo-cAMP on Matrigel degradation.

Cytotrophoblast responsiveness to 8-bromo-cAMP led us to examine the effect of this agent on trophoblast-Matribeach interactions. We found that 8-bromo-cAMP completely eliminated zone 2 matrix degradation by normal trophoblasts (Kliman and Feinberg, 1990a). The cells appeared uniformly spherical, with little interaction between other cells or the matrix at all thicknesses. In trophoblasts, 8-bromo-cAMP promotes differentiation towards a non-invasive villous syncytial trophoblast phenotype, with induction of the synthesis and secretion of hCG (Feinman et al. 1986; Kao et al. 1988). Additionally, this cyclic nucleotide dramatically down-regulates fibronectin production by cytotrophoblasts (Ulloa-Aguirre et al. 1987), thus potentially altering their interaction capabilities with the ECM. Interestingly, JEG-3 choriocarcinoma cells did not alter their morphology or degradative capabilities on zone 2 Matribeach when treated with 8-bromo-cAMP (Kliman and Feinberg, 1990a), suggesting that their invasive phenotype does not respond to regulatory agents in the same way as normal trophoblasts.

2.1.5.3. What is the role of the plasminogen activator (PA) system in human trophoblast invasion?

The biological repertoire necessary for the unique trophoblast-ECM interactions that occur during human gestation is unknown. Studies in non-human systems have proposed a critical role for trophoblast-secreted plasminogen activator (PA) during implantation and placentation. In mice, trophoblast production of PA correlates temporally with blastocyst invasion (Strickland et al. 1976; Sherman et al. 1976), and implantation-defective mouse embryos elaborate diminished amounts of PA (Axelrod, 1985). A role for PA in human nidation is suggested by the work of several investigators (Martin and Arias, 1982; Queenan et al. 1987), who demonstrated that trophoblasts in culture produce active PA. More recently, the results of Cajot et al. (1989) lend support to the concept that any cell type which produces active u-PA can harbor an invasive phenotype. These workers transfected non-invasive mouse L cells with a cosmid containing the complete human u-PA gene. Those cells which expressed human u-PA could both degrade and invade the ECM, suggesting that u-PA expression alone is sufficient to initiate these processes.

PA activity in vascular and extracellular spaces is modulated by PA inhibitors (PAIs), glycoproteins of the SERPIN (serine protease inhibitor) family that covalently bind to and inhibit u-PA (Loskutoff et al. 1986; Wun and Riech, 1987). Therefore, it is likely that PA-PAI interactions modulate trophoblast invasion in vivo and control fibrinolysis within the intervillous spaces of the placenta. Until recently, evidence for trophoblast elaboration of PAIs during pregnancy has been indirect. Two well characterized PAIs, PAI-1 and PAI-2, were isolated from total placental extracts (Kruithof et al. 1987; Wun and Riech, 1987; Ye et al. 1987). In addition, plasma PAI-1 and PAI-2 levels increase with advancing gestation, but decrease dramatically soon after delivery (Kruithof et al. 1987). Altered plasma levels of PAIs have been reported in preeclampsia of pregnancy (DeBoer et al. 1988; Gore et al. 1987), a disease in which abnormalities in fibrinolysis and trophoblast function often occur. Because of the potential role of the PAIs in controlling trophoblast invasion, we initiated studies to examine the synthesis and regulation of PAI-1 and PAI-2 in normal human cytotrophoblasts. PAI -1 and PAI-2 mRNA and protein are produced by cultured cytotrophoblasts, whereas only PAI-1 is found in JEG-3 cells, a malignant trophoblast cell line (Feinberg et al. 1989b; Feinberg et al. 1990). PAI-2 is localized by immunocytochemistry to villous syncytiotrophoblasts whereas PAI-1 is present primarily in invasive trophoblasts of implantation sites (Feinberg et al. 1989a). These findings suggest an important physiological role for these proteins in vivo, and may indicate that PAI-1 is specifically required for limiting the trophoblast invasive process.

2.1.5.4. Cyclic AMP and protein kinase C modulation of the PA system

Cyclic AMP agonists have profound effects on cytotrophoblasts in culture by affecting their intracellular morphology, endocrine function, and synthesis of fibronectin (Feinman et al. 1986; Ulloa-Aguirre et al. 1987). As cytotrophoblasts undergo morphologic and biochemical differentiation in culture, both hCG and progesterone production are dramatically stimulated by 8-bromo-cAMP and forskolin (Feinman et al. 1986; Kao et al. 1988). The increase in hCG secretion is the consequence of increased synthesis of aand ß-hCG protein subunits following increased transcription of their corresponding mRNAs (Ulloa-Aguirre et al. 1987). Cyclic AMP also affects u-PA gene transcription and proteolytic activity (Queenan et al. 1987) as well as PAI production in trophoblastic cells (Feinberg et al. 1988; Feinberg et al. 1990).

2.1.5.5. The plasminogen activator system in human trophoblasts.

We have utilized immunocytochemical, immunoblot, Northern blot, and zymographic analysis of cultured human trophoblasts as methods for studying PA and PAI production. We have identified immunoreactive u-PA in normal trophoblasts utilizing affinity purified rabbit anti-human u-PA antibody (Fig. 7) and have confirmed, by protein immunoblotting, the presence of u-PA antigen in trophoblast cultures. As in other cell systems, trophoblast invasion and remodeling of the ECM may be controlled by a balance between trophoblast elaboration of u-PA and PAIs. We recently determined that human trophoblasts in culture produce both PAI-1, the principal endothelial cell PAI, and PAI-2, a macrophage and placental-associated PAI (Feinberg et al. 1989a). Immunoblots of cellular protein extract and conditioned media indicate that freshly purified cytotrophoblasts contain barely detectable levels of u-PA, PAI-1, or PAI-2. As the cytotrophoblasts differentiate in culture, the maximal rate of production occurs in the first 24 h. Interestingly, u-PA production appears to peak at 24 h, whereas PAI-1 and PAI-2 production continues at a steady state (Fig. 8). Queenan et al. (1987), utilized trophoblasts prepared by the Kliman method and demonstrated secretion of a 50 kD active u-PA species on plasminogen-dependent zymography of conditioned culture media. These workers also showed maximal activity of this secreted u-PA after 24 h of culture and de novo synthesis of the protein. By pulse-labeling cultured cytotrophoblasts with 35S-methionine, followed by immunoprecipitation with specific anti-PAI antibodies, we have determined (unpublished results) that both PAI-1 and PAI-2 are synthesized de novo in our cell cultures.

Performing immunocytochemistry (using highly specific rabbit anti-human PAI-1 and PAI-2 antibodies provided by Dr. T.-C. Wun of Monsanto Co.), we have demonstrated (Feinberg et al. 1989a) that PAI-1 was localized to the trophoblast cell surface, as well as to the cytoplasm, whereas the cellular localization of PAI-2 was only cytoplasmic. Northern blot hybridization analysis using cDNA probes for PAI-1 and PAI-2 (obtained from Dr. T-C Wun [Wun and Reich, 1987; Ye et al. 1987]) revealed two trophoblast PAI-1 transcripts (2.3 and 3.0 kb) and one 2.3 kb PAI-2 transcript, consistent with findings in other cultured human cells (Feinberg et al. 1988; Feinberg et al. 1989a). Interestingly, the cAMP agonist 8-bromo-cAMP appears to coordinately regulate the plasminogen activator system in normal trophoblasts by down-regulating PAI mRNA levels two-fold (Feinberg et al. 1988) and by up-regulating u-PA mRNA three-fold in the first 24 h of culture. In zymographic analysis of serum-free trophoblast media we found secretion of an active u-PA at ~50 kD (unpublished results), in agreement with the findings of Queenan et al. (1987).

Although trophoblasts elaborate u-PA and PAIs in culture, correlation with trophoblast biology in vivo is critical. Therefore, we undertook studies to characterize PA and PAI expression by trophoblasts in situ within placental villi and implantation sites. Utilizing a monoclonal anti-human u-PA (gift of E. Barnathan, University of Pennsylvania) and the PAI-1 polyclonal antibody described above, we have been able to identify both PAI-1 and u-PA in the invasive trophoblasts of the placental bed (Feinberg et al. 1989a; Kliman and Feinberg, unpublished results). It appears, therefore, that both in vitro and in vivo, trophoblasts elaborate u-PA and the PAIs.

2.1.5.6. Role of other proteases in trophoblast invasion.

Matrigel clearing around the trophoblasts on zone 2 of Matribeach appears to be promoted by proteases which degrade both type IV collagen and laminin. Proteolysis of these proteins occurs as a function of time, with immunohistochemical analysis and scanning electron microscopy demonstrating progressively larger areas of negative staining around zone 2 cells from 24 to 72 h (Fig. 6; Kliman and Feinberg, 1990). The biochemical basis of this trophoblast-mediated proteolysis is unknown but, as in other cell systems, may be initiated by u-PA (Mignatti et al. 1986; Reich et al. 1988). However, the potential role of trophoblast-secreted collagenases (Fisher et al. 1989) or cathepsins (Lah et al. 1989) in this process is not known. Plasminogen activators are proposed to be activators of procollagenases (Mignatti et al. 1986; Reich et al. 1988), and therefore u-PA activation of collagenases may have a role in trophoblast proteolysis and invasion. In gelatin zymographic analysis of serum-free trophoblast media we found, in addition to an active u-PA, secretion of an active gelatinase in the 90 kD range (Kliman and Feinberg, unpublished results). This gelatinase was present in both the plasminogen-containing and plasminogen-free gels, in agreement with the findings of Fisher et al. (1989). However, when Matrigel was substituted for gelatin in our zymograms, only u-PA could be detected, and only in the plasminogen-containing gel. This suggests that the major secreted trophoblast protease capable of initiating Matrigel degradation is probably u-PA, and that this proteolysis is accomplished primarily through the localized activation of plasminogen to plasmin within the ECM.

2.1.6. The polyphonic hypothesis of trophoblast differentiation

We have noted a paradox about cultured trophoblasts. Cultured trophoblasts are capable of synthesizing and secreting a large array of placental products simultaneously, while trophoblasts in vivo seem to only make select products as defined by their location. For example, examination of tissue sections reveals that villous syncytiotrophoblasts immunochemically stain for hCG, hPL and pregnancy specific ß1-glycoprotein (Kliman et al. 1986), but do not stain for PAI-1 (Feinberg et al. 1989), trophouteronectin (Feinberg et al. 1991), or a 34 kD growth-factor (Roy-Choudhury et al. 1988). Yet, when cytotrophoblasts from term placentae are cultured, they can apparently synthesize all of these products simultaneously (Fig. 9). Double immunocytochemical staining for both hCG and hPL has in fact shown that cultured trophoblasts can contain at least these two hormones at the same time (Kliman et al. 1987). Since immunocytochemical staining of trophoblast cultures shows that the large majority of the cells stain for all of these products, it can be concluded that most, if not all, of the trophoblasts are synthesizing all of these products at the same time. Why do cultured trophoblasts make many products simultaneously, but trophoblasts in vivo only make a select group?

Our answer can be understood by an analogy. A piano is polyphonic: as each key is struck, a single tone is generated. If many keys are struck, many tones are generated. We imagine that the trophoblast is like a piano and that regulatory signals are like the keys. If one regulatory signal is struck, one product is produced. If many regulatory signals are struck, many products are produced. Following this logic, we have concluded that in culture, trophoblasts are stimulated by many simultaneous regulatory signals, hence many products. In vivo, on the other hand, the trophoblasts are stimulated by a discreet set of regulatory signals, hence a select group of products. The goal of future research, therefore, will be to define precisely how specific hormones (endocrine and paracrine), extracellular matrix proteins, and other factors harmonize the complex process of trophoblast differentiation.

REFERENCES

Al TA, Fox H (1986) Immunohistochemical localization of follicle-stimulating hormone, luteinizing hormone, growth hormone, adrenocorticotrophic hormone and prolactin in the human placenta. Placenta 7:163-72

Albini A, Iwamoto Y, Kleinman HK, Martin GR, Aaronson SA, Kozlowski JM, McEwan RN (1987) A rapid in vitro assay for quantitating the invasive potential of tumor cells. Cancer Res 47:3239-45

Axelrod HR (1985) Altered trophoblast functions in implantation-defective mouse embryos. Dev Biol 108:185-90

Bax CM, Ryder TA, Mobberley MA, Tyms AS, Taylor DL, Bloxam DL (1989) Ultrastructural changes and immunocytochemical analysis of human placental trophoblast during short-term culture. Placenta 10:179-94

Belisle S, Bellabarba D, Gallo PN, Lehoux JG, Guevin JF (1986) On the role of luteinizing hormone-releasing hormone in the in vitro synthesis of bioactive human chorionic gonadotropin in human pregnancies. Can J Physiol Pharmacol 64:1229-35

Belisle S, Petit A, Bellabarba D, Escher E, Lehoux JG, Gallo PN (1989) Ca2+, but not membrane lipid hydrolysis, mediates human chorionic gonadotropin production by luteinizing hormone-releasing hormone in human term placenta. J Clin Endocrinol Metab 69:117-21

Blay J, Hollenberg MD (1989) The nature and function of polypeptide growth factor receptors in the human placenta. J Dev Physiology 12:237-248

Branchaud C, Goodyer CG, Guyda HJ, Lefebvre Y (1990) A serum-free system for culturing human placental trophoblasts. In Vitro Cell Dev Biol 26: 865-870

Cajot JF, Schleuning WD, Medcalf RL, Bamat J, Testuz J, Liebermann L, Sordat B (1989) Mouse L cells expressing human prourokinase-type plasminogen activator: effects on extracellular matrix degradation and invasion. J Cell Biol 109:915-25

Conley AJ, Mason JI (1990) Placental steroid hormones. Baillieres Clin Endo Met 4:249-272

DeBoer K, Lecander I, ten Cate JW, Borm JJJ, Treffers PE (1988) Placental-type plasminogen activator inhibitor in preeclampsia. Am J Obstet Gynecol 158:518

Diczfalusy E and Troen P (1961) Endocrine functions of the human placenta. Vitam Horm 19:229-311

Dodeur M, Malassine A, Bellet D, Mensier A, Evain BD (1990) Characterization and differentiation of human first trimester placenta trophoblastic cells in culture. Reprod Nutr Dev 30:183-92

Earl U, Estlin C, Bulmer JN (1990) Fibronectin and laminin in the early human placenta. Placenta 11:223-231

Evain BD, Alsat E, Mirlesse V, Dodeur M, Scippo ML, Hennen G, Frankenne F (1990) Regulation of growth hormone secretion in human trophoblastic cells in culture. Horm Res 33:256-9

Feinberg R, Kao L-C, Ringler G, Murray S, Queenan J, Kliman H, Cines D, Wun T-C, Strauss JF3 (1988) Coordinate regulation of urokinase and plasminogen activator inhibitors in human cytotrophoblasts (Abst). Proc Soc Gyn Invest 35:231

Feinberg RF, Kao LC, Haimowitz JE, Queenan JTJ, Wun TC, Strauss JF3, Kliman HJ (1989a) Plasminogen activator inhibitor types 1 and 2 in human trophoblasts. PAI-1 is an immunocytochemical marker of invading trophoblasts. Lab Invest 61:20-6

Feinberg RF, Strauss JF3, Wun T-C, Kliman HJ (1989b) Plasminogen activators (PAs) and plasminogen activator inhibitors (PAIs) in human trophoblasts: markers of trophoblast invasion (Abst). Proc Soc Gyn Invest 36:487

Feinberg RF, Kao LC, Wang C-L, Bui L, Kliman HJ, Strauss JF3 (1990) Plasminogen activator inhibitor (PAI) expression in normal and malignant human trophoblasts: regulation by 8-bromo-camp and phorbol esters (Abst). Proc Soc Gyn Invest 37:465

Feinberg RF, Kliman HJ, Lockwood CJ (1991) Oncofetal fibronectin: A trophoblast "glue" for human implantation? Am J Path 138:537-43

Feinman MA, Kliman HJ, Caltabiano S, Strauss JF3 (1986) 8-Bromo-3',5'-adenosine monophosphate stimulates the endocrine activity of human cytotrophoblasts in culture. J Clin Endocrinol Metab 63:1211-7

Fisher SJ, Sutherland A, Moss L, Hartman L, Crowley E, Bernfield M, Calarco P, Damsky C (1990) Adhesive interactions of murine and human trophoblast cells. Troph Res 4:115-138

Fisher SJ, Cui TY, Zhang L, Hartman L, Grahl K, Zhang GY, Tarpey J, Damsky CH (1989) Adhesive and degradative properties of human placental cytotrophoblast cells in vitro. J Cell Biol 109:891-902

Fournet DN, MacLusky NJ, Leranth CZ, Todd R, Mendelson CR, Simpson ER, Naftolin F (1987) Immunohistochemical localization of aromatase cytochrome P-450 and estradiol dehydrogenase in the syncytiotrophoblast of the human placenta. J Clin Endocrinol Metab 65:757-64

Gore M, Eldon S, Trofatter KF, Soong S-J, Pizzo SV (1987) Pregnancy-induced changes in the fibrinolytic balance: evidence for defective release of tissue plasminogen activator and increased levels of the fast-acting tissue plasminogen activator inhibitor. Am J Obstet Gynecol 156:674

Handwerger S, Quarfordt S, Barrett J, Harman I (1987) Apolipoproteins AI, AII, and CI stimulate placental lactogen release from human placental tissue. A novel action of high density lipoprotein apolipoproteins. J Clin Invest 79:625-8

Hertig AT, Rock J (1956) A description of 34 human ova within the first 17 days of development. Am J Anat 98:435-494

Jara CS, Salud AT, Bryantgreenwood GD, Pirens G, Hennen G, Frankenne F (1989) Immunocytochemical localization of the human growth hormone variant in the human Placenta. J Clin Endocrinol Metab 69:1069-1072

Kao LC, Caltabiano S, Wu S, Strauss JF3, Kliman HJ (1988) The human villous cytotrophoblast: interactions with extracellular matrix proteins, endocrine function, and cytoplasmic differentiation in the absence of syncytium formation. Dev Biol 130:693-702

Kato Y, Braunstein GD (1990) Purified 1st and 3rd trimester placental trophoblasts differ in in vitro hormone secretion. J Clin Endocrinol Metab 70:1187-1192

Kawagoe K, Akiyama J, Kawamoto T, Morishita Y, Mori S (1990) Immunohistochemical demonstration of epidermal growth factor (EGF) receptors in normal human placental villi. Placenta 11:7-15

Khodr GS, Siler KT (1978) Localization of luteinizing hormone-releasing factor in the human placenta. Fertil Steril 29:523-6

Kleinman HK, McGarvey ML, Hassell JR, Star VL, Cannon FB, Laurie GW, Martin GR (1986) Basement membrane complexes with biological activity. Biochemistry 25:312-8

Kliman HJ, Coutifaris C, Babalola GO, Soto EA, Kao LC, Queenan JTJ, Feinberg RF, Strauss JF3 (1989) The human trophoblast: homotypic and heterotypic cell-cell interactions. Prog Clin Biol Res 294:425-34

Kliman HJ, Feinberg RF (1990a) Human trophoblast-extracellular matrix (ECM) interactions in vitro: ECM thickness modulates morphology and proteolytic activity. Proc Natl Acad Sci U S A 87:3057-61

Kliman HJ, Feinberg RF, Haimowitz JE (1990b) Human trophoblast-endometrial interactions in an in vitro suspension culture system. Placenta 11:349-67

Kliman HJ, Feinman MA and Strauss JF3 (1987) Differentiation of human cytotrophoblasts into syncytiotrophoblasts in culture. Troph Res 2: 407-421

Kliman HJ, Nestler JE, Sermasi E, Sanger JM, Strauss JF3 (1986) Purification, characterization, and in vitro differentiation of cytotrophoblasts from human term placentae. Endocrinology 118:1567-82

Kliman HJ, Strauss JF3, Kao L-C, Caltabiano S, Wu S (1991) Cytoplasmic and biochemical differentiation of the human villous cytotrophoblast in the absence of syncytium formation. Troph Res, in press.

Krieger DT (1982) Placenta as a source of 'brain' and 'pituitary' hormones. Biol Reprod 26:55-71

Kruithof EKO, Tran-Thang C, Gudinchet A, Hauert J, Nicoloso G, Genton C, Welti H, Bachmann F (1987) Fibrinolysis in pregnancy: a study of plasminogen activator inhibitors. Blood 69:460-66

Lah TT, Buck MR, Honn KV, Crissman JD, Rao NC, Liotta LA, Sloane BF (1989) Degradation of laminin by human tumor cathepsin B. Clin Exp Metastasis 7:461-8

Liotta A, Osathanondh R, Ryan KJ, Krieger DT (1977) Presence of corticotropin in human placenta: demonstration of in vitro synthesis. Endocrinology 101:1552-8

Liotta LA, Rao CN, Barsky SH (1983) Tumor invasion and the extracellular matrix. Lab Invest 49:636-49

Lobo JO, Bellino FL. (1989) Estrogen synthetase (aromatase) activity in primary culture of human term placental cells: effects of cell preparation, growth medium, and serum on adenosine 3',5'-monophosphate response. J Clin Endocrinol Metab 69:868-74

Loke YW (1983) Human trophoblast in culture. In: Loke YW, Whyte A (eds) Biology of Trophoblast. Elsevier, Holland, pp 663-701

Loke YW (1990) New developments in human trophoblast cell culture. Colloque INSERM 199:10-16

Loke YW, Gardner L, Grabowska A (1989) Isolation of human extravillous trophoblast cells by attachment to laminin-coated magnetic beads. Placenta 10:407-15

Loskutoff DJ, Ny T, Sawdey M, Lawrence D (1986) Fibrinolytic system of cultured endothelial cells: regulation by plasminogen activator inhibitor. J Cell Biochem 32:273

Martin O, Arias F (1982) Plasminogen activator production by trophoblast cells in vitro: effect of steroid hormones and protein synthesis inhibitors. Am J Obstet Gynecol 142: 402-9

Maruo T, Matsuo H, Oishi T, Hayashi M, Nishino R, Mochizuki M (1987) Induction of differentiated trophoblast function by epidermal growth factor: relation of immunohistochemically detected cellular epidermal growth factor receptor levels. J Clin Endocrinol Metab 64:744-50

Maruyama I, Bell CE, Majerus PW (1985) Thrombomodulin is found on endothelium of arteries, veins, capillaries, and lymphatics, and on syncytiotrophoblast of human placenta. J Cell Biol 101:363-71

Mignatti P, Robbins E, Rifkin DB (1986) Tumor invasion through the human amniotic membrane: requirement for a proteinase cascade. Cell 47:487-98

Moll UM, Lane BL (1990) Proteolytic activity of 1st trimester human placenta - localization of interstitial collagenase in villous and extravillous trophoblast. Histochemistry 94:555-560

Mulder GH, Maas R, Arts NF (1986) In vitro secretion of peptide hormones by the human placenta: I. ACTH. Placenta 7:143-53

Nishino E, Matsuzaki N, Masuhiro K, Kameda T, Taniguchi T, Takagi T, Saji F, Tanizawa O (1990) Trophoblast-derived interleukin-6 (IL-6) regulates human chorionic gonadotropin release through IL-6 receptor on human trophoblasts. J Clin Endocrinol Metab 71:436-441

Nulsen JC, Silavin SL, Kao LC, Ringler GE, Kliman HJ, Strauss JF3 (1989) Control of the steroidogenic machinery of the human trophoblast by cyclic AMP. J Reprod Fertil Suppl 37:147-53

Odagiri E, Sherrell BJ, Mount CD, Nicholson WE, Orth DN (1979) Human placental immunoreactive corticotropin, lipotropin, and beta-endorphin: evidence for a common precursor. Proc Natl Acad Sci U S A 76:2027-31

Ohlsson R, Larsson E, Nilsson O, Wahlstrom T, Sundstrom P (1989) Blastocyst implantation precedes induction of insulin-like growth factor II gene expression in human trophoblasts. Development 106:555-9

Ohtani H, Maruyama I, Yonezawa S (1989) Ultrastructural immunolocalization of thrombomodulin in human placenta with microwave fixation. Act Hist Cy 22:393-5

Oike N, Iwashita M, Muraki T, Nomoto T, Takeda Y, Sakamoto S (1990) Effect of adrenergic agonists on human chorionic gonadotropin release by human trophoblast cells obtained from 1st-trimester placenta. Horm Metab Res 22:188-191

Petraglia F, Sawchenko P, Lim AT, Rivier J, Vale W (1987) Localization, secretion, and action of inhibin in human placenta. Science 237:187-9

Petraglia F, Vaughan J, Vale W (1989a) Inhibin and activin modulate the release of gonadotropin-releasing hormone, human chorionic gonadotropin, and progesterone from cultured human placental cells. Proc Natl Acad Sci U S A 86:5114-7

Petraglia F, Sutton S, Vale W (1989b) Neurotransmitters and peptides modulate the release of immunoreactive corticotropin-releasing factor from cultured human placental cells. Am J Obstet Gynecol 160:247-51

Petraglia F, Calza L, Garuti GC, Giardino L, De RB, Angioni S (1990a) New aspects of placental endocrinology. J Endocrinol Invest 13:353-371

Petraglia F, Vaughan J, Vale W (1990b) Steroid hormones modulate the release of immunoreactive gonadotropin-releasing hormone from cultured human placental cells. J Clin Endocrinol Metab 70: 1173-1178

Petraglia F, Garuti GC, Deramundo B, Angioni S, Genazzani AR, Bilezikjian LM (1990c) Mechanism of action of interleukin-1-beta in increasing corticotropin-releasing factor and adrenocorticotropin hormone release from cultured human placental cells. American Journal of Obstetrics and Gynecology 163:1307-1312

Pijnenborg R (1990) Trophoblast invasion and placentation in the human-morphological aspects. Troph Res 4:33-47

Posner BI (1974) Insulin receptors in human and animal placental tissue. Diabetes 23:209-217

Queenan JT Jr, Kao L-C, Arboleda CE, Ulloa-Aguirre A, Golos TG, Cines DB, Strauss JF3 (1987) Regulation of urokinase-type plasminogen activator production by cultured human cytotrophoblasts. J Biol Chem 262:10903-6

Reich R, Thompson EW, Iwamoto Y, Martin GR, Deason JR, Fuller GC, Miskin R (1988) Effects of inhibitors of plasminogen activator, serine proteinases, and collagenase IV on the invasion of basement membranes by metastatic cells. Cancer Res 48:3307-12

Ren SG, Braunstein GD (1991) Decidua produces a protein that inhibits chorio-gonadotropin release from human trophoblasts. J Clin Invest 87:326-330

Ringler GE, Ulloa-Aguirre A, Kao L-C, Nulsen JC, Kallen CB, Kliman HJ, Strauss JF3 (1988) Control of chorionic gonadotropin (hCG) by cyclic AMP: lessons from primary cultures of cytotrophoblasts. In: Mochizuki M, Hussa R (eds) Placental protein hormones. Elsevier Science Publishers, Amsterdam, p 184

Ringler GE, Kallen CB, Strauss JF3 (1989) Regulation of human trophoblast function by glucocorticoids: dexamethasone promotes increased secretion of chorionic gonadotropin. Endocrinology 124:1625-31

Ringler GE, Strauss JF3 (1990) In vitro systems for the study of human placental endocrine function. Endocr Rev 11: 105-23

Robinson BG, Emanuel RL, Frim DM, Majzoub JA (1988) Glucocorticoid simulates corticotropin releasing hormone gene expression in human placenta. Proc Natl Acad Sci USA 85:5244-8

Roy CS, Sen MA, Murthy U, Mishra VS, Kliman HJ, Nestler JE, Strauss JF3, Das M (1988) Biosynthesis and turnover of a 34-kDa protein growth factor in human cytotrophoblasts. Eur J Biochem 172:777-83

Saijonmaa O, Laatikainen T, Wahlstrom T (1988) Corticotrophin-releasing factor in human placenta: localization, concentration and release in vitro. Placenta 9:373-85

Sakbun V, Koay ES, Bryant GGD (1987) Immunocytochemical localization of prolactin and relaxin C-peptide in human decidua and placenta. J Clin Endocrinol Metab 65:339-43

Sakbun V, Ali SM, Lee YA, Jara CS, Bryantgreenwood GD (1990a) Immunocytochemical localization and messenger ribonucleic acid concentrations for human placental lactogen in amnion, chorion, decidua, and placenta. Am J Obstet Gynecol 162:1310-1317

Sakbun V, Ali SM, Greenwood FC, Bryantgreenwood GD (1990b) Human relaxin in the amnion, chorion, decidua-parietalis, basal plate, and placental trophoblast by immunocytochemistry and northern analysis. J Clin Endocrinol Metab 70:508-514

Sherman MI, Strickland S, Reich E (1976) Differentiation of early mouse embryonic and teratocarcinoma cells in vitro: plasminogen activator production. Canc Res 36: 4208-4216

Sirinathsinghji DJ, Heavens RP (1989) Stress-related peptide hormones in the placenta: their possible physiological significance. J Endocrinol 122:435-7

Strickland S, Reich E, Sherman MI (1976) Plasminogen activator in early embryogenesis: enzyme production by trophoblast and parietal endoderm. Cell 9:231-40

Truman P, Ford HC (1986) The effect of substrate and epidermal growth factor on human placental trophoblast cells in culture. In Vitro Cell Dev Biol 22:525-8

Truman P, Pomare L, Ford HC (1989) Human placental cytotrophoblast cells: identification and culture. Arch Gynecol Obstet 246:39-49

Uzumaki H, Okabe T, Sasaki N, Hagiwara K, Takaku F, Tobita M, Yasukawa K, Ito S, Umezawa Y (1989) Identification and characterization of receptors for granulocyte colony-stimulating factor on human placenta and trophoblastic cells. Proc Natl Acad Sci U S A 86:9323-6

Vartio T, Laitinen L, Narvanen O, Cutolo M, Thornell L-E, Zardi L, Virtanen IJ (1987) Differential expression of the ED sequence-containing form of cellular fibronectin in embryonic and adult human tissues. J Cell Science 88:419-430

Wun TC, Reich E (1987) An inhibitor of plasminogen activation from human placenta. Purification and characterization. J Biol Chem 262:3646-53

Yagel S, Casper RF, Powell W, Parhar RS, Lala PK (1989a) Characterization of pure human first-trimester cytotrophoblast cells in long-term culture: growth pattern, markers, and hormone production. Am J Obstet Gynecol 160: 938-45

Yagel S, Lala PK, Powell WA, Casper RF (1989b) Interleukin-1 stimulates human chorionic gonadotropin secretion by first trimester human trophoblast. J Clin Endocrinol Metab 68: 992-5

Yamada T, Isemura M, Yamaguchi Y, Munakata H, Hayashi N, Kyogoku M (1987) Immunohistochemical localization of fibronectin in the human placentas at their different stages of maturation. Histochem 86:579-584

Ye RD, Wun TC, Sadler JE (1987) cDNA cloning and expression in Escherichia coli of a plasminogen activator inhibitor from human placenta. J Biol Chem 262:3718-25

Yuen BH, Moon YS, Shin DH (1986) Inhibition of human chorionic gonadotropin production by prolactin from term human trophoblast. Am J Obstet Gynecol 154:336-340

Table 1

Regulation of cultured trophoblast hCG secretion

Factor	Trophoblasts (Trimester)	Effect on hCG Secretion	References
cAMP	Term	Stimulates	Feinman et al. 1986
GnRH	Term	Stimulates	Belisle et al. 1989
ß-adrenergic agonists	First	Stimulates	Oike et al. 1990
Dexamethasone	Term	Stimulates	Ringler et al. 1989
Inhibin	Term	Inhibits	Petraglia et al. 1987
Activin	Term	Potentiates GnRH simulation of hCG secretion	Petraglia et al. 1989a
EGF	First, Term	Stimulates	Maruo et al. 1987
Interleukin-1.	First	Stimulates	Yagel et al. 1989b
Interleukin-6	First	Stimulates	Nishino et al. 1990
Basement Membrane	First	Stimulates	Truman and Ford 1986
Decidual Protein	Term	Inhibits	Ren and Braunstein 1991
Prolactin	Term	Inhibits	Yuen et al. 1986

FIGURES and FIGURE LEGENDS

Figure 1. Model of pathways of cytotrophoblast differentiation. All placental and extraplacental trophoblasts derive from undifferentiated cytotrophoblasts. The signals that direct cytotrophoblasts to the three major classes of differentiated trophoblast are not known. Autocrine, paracrine, extracellular matrix mediated, and release of inhibition mechanisms have been suggested.

Figure 2. Purification of human trophoblasts as described by Kliman et al. (1986). Minced villous tissue is subjected to trypsin-DNase digestion. Trypsin is inactivated by spinning the digest supernatants through a calf serum layer. The digest-supernatants are centrifuged on a 5-70% Percoll gradient, yielding a band of pure cytotrophoblasts. (Reproduced with permission from Ringler et al. 1988)

Figure 3. Formation of syncytiotrophoblasts in vitro. Initially, the isolated cytotrophoblasts are mononuclear single cells. At 24 h, the dominant form is the multicellular aggregate. After 24 h, increasing numbers of syncytia are seen until eventually they become the dominant form (Reproduced with permission from Kliman et al. 1986).

Figure 4. Trophouteronectin (TUN) immunohistochemistry. A Bouin's fixed, paraffin embedded tissue section from a 6-week gestation was immunohistochemically stained with monoclonal anti-human oncofetal fibronectin (FDC-6). The utero-placental junction exhibits a distinct band of TUN staining (arrow heads) at the zone of contact between the extravillous trophoblasts and the decidualized endometrium (D). Note the negatively stained chorionic villi (V). Bar represents 100 µm.

Figure 5. Trophoblast morphology on a Matrigel slope (Matribeach). Term human trophoblasts were plated onto Matribeach, cultured for 48 h, fixed and immunocytochemically stained for type IV collagen. Panoramic view of the Matribeach with clearly defined zones 1, 2, and 3. Glass surface (G). Bar represents 200 µm.

Figure 6. Trophoblasts degrade Matrigel. A) Scanning electron microscopy of an aggregate of trophoblasts (T) at 48 h with a circumferential area of Matrigel lysis (arrows). Thin cytoplasmic processes emanate from the cells and merge into the surrounding Matrigel. Note the flat filamentous nature of the undigested Matrigel (M). B) Now at 72 h, this trophoblast aggregate (T) has completely degraded the Matrigel (arrows) so that the surface of the underlying cover slip is exposed (G). Note again the flat surface of the undegraded Matrigel (M). C) Light microscopic appearance of an aggregate of trophoblasts on zone 2 of Matribeach after 48 h of culture showing an almost complete clearing of the pericellular Matrigel. Bars represent 10 µm. (Portions of this figure are reproduced with permission from Kliman and Feinberg, 1990a).

Figure 7. Urokinase immunocytochemistry. Cytotrophoblasts were cultured for 24-h, fixed for 15 m with Bouin's solution, and immunocytochemically stained with affinity purified rabbit anti-human u-PA antibody (a gift from Dr. R. Pittman, University of Pennsylvania). This low power field reveals a mixture of single flattened trophoblasts and forming aggregates. u-PA staining can be identified within the cytoplasm of all of the cells. In addition, fine surface staining can be identified at the contact points of several of the aggregated trophoblasts (arrow heads). Bar represents 20 µm.

Figure 8. Immunoblots demonstrating the presence of PAI types 1 (a-g) and 2 (a'-c') at 0 h (a,e,a'), 24 h (b,f,b'), 48 h (c,g,c'), and 72 h (d) in trophoblast cell extract (a-d, a'-c') and conditioned media (e-g). Arrow marks 46 kD.

Figure 9. Polyphonic hypothesis of trophoblast differentiation. In vivo, undifferentiated cytotrophoblasts can differentiate into either endocrine villous syncytiotrophoblasts (e.g., hCG secreting), into anchoring cell column trophoblasts [trophouteronectin (TUN) synthesizing], or into intra-myometrial invading intermediate trophoblasts (containing plasminogen activators and inhibitors, e.g., PAI-1). Specific products are only expressed by specific trophoblast types (i.e., TUN is not found in villous or invading trophoblasts), suggesting that the local environment dictates the pathway of trophoblast differentiation with a few select regulatory signals (represented by a few notes from the piano keyboard). Control may be mediated through autocrine, paracrine, extracellular matrix mediated, and release of inhibition mechanisms. In vitro, cultured cytotrophoblasts go through a series of morphologic changes, culminating in the production of syncytiotrophoblasts over a 96 h period. During this time, these cells contain and secrete a wide assortment of products simultaneously, suggesting that cultured trophoblasts are stimulated by a variety of regulatory signals at once (represented by a full piano keyboard).