Placental Anatomy and Physiology

Overview of the Delivered Placenta

At term, the human placenta is usually a discoid organ, 15 to 20 cm in diameter, approximately 3 cm thick at the center, and weighing on average 450 g. There is considerable individual variation in both size and shape, and placentas are also influenced strongly by the mode of delivery, with loss of blood from the vascular spaces leading to deflation and thinning of the disc. Macroscopically the organ consists of two surfaces or plates: the chorionic plate, to which the umbilical cord is attached, and the basal plate , which abuts the maternal endometrium. Between the two plates is a cavity filled with maternal blood, delivered from the endometrial spiral arteries through openings in the basal plate ( Fig. 1.1 ). This cavity is bounded at the margins of the disc by the fusion of the chorionic and basal plates. The smooth chorion, or chorion laeve, extends from the rim to complete the chorionic sac. The placenta is incompletely divided into between 10 and 40 lobes by the presence of septa created by invaginations of the basal plate. The septa are thought to arise from differential resistance of the maternal tissues to trophoblast invasion and may help to compartmentalize, and hence direct, maternal blood flow through the organ. The fetal component of the placenta comprises a series of elaborately branched villous trees that arise from the inner surface of the chorionic plate and project into the cavity of the placenta. This arrangement is reminiscent of the fronds of a sea anemone wafting in the seawater of a rock pool. Most commonly, each villous tree originates from a single-stem villus that undergoes several generations of branching until the functional units of the placenta, the terminal villi , are created. These consist of an epithelial covering of trophoblast and a mesodermal core containing branches of the umbilical arteries and tributaries of the umbilical vein. Because of this repeated branching, the tree takes on the topology of an inverted wine glass, often referred to as a lobule , and two to three of these may be contained within a single placental lobe (see Fig. 1.1 ). As will be seen later, each lobule represents an individual maternal-fetal exchange unit. Near term, the continual elaboration of the villous trees almost fills the cavity of the placenta, which is reduced to a network of narrow spaces collectively referred to as the intervillous space (IVS) . The maternal blood percolates through this network of channels and exchanges gases and nutrients with the fetal blood, which circulates within the villi before draining through the basal plate into openings of the uterine veins. The human placenta is therefore classified in comparative mammalian terms as being of the villous hemochorial type, although, as we shall see, this arrangement pertains only to the second and third trimesters of pregnancy . Prior to that, the maternal-fetal relationship is best described as deciduochorial.

Placental Development

Development of the placenta is initiated morphologically at the time of implantation, when the embryonic pole of the blastocyst establishes contact with the uterine epithelium. At this stage, the wall of the blastocyst comprises an outer layer of unicellular epithelial cells, the trophectoderm, and an inner layer of extraembryonic mesodermal cells derived from the inner cell mass; together these layers constitute the chorion. The earliest events have never been observed in vivo for obvious ethical reasons, but it has been assumed that they are equivalent to those that take place in the rhesus monkey, which has a similar hemochorial placenta.

Attempts have also been made to replicate the situation by culturing in vitro fertilized human blastocysts on monolayers of endometrial cells. Although such reductionist systems cannot take into account the possibility of paracrine signals that emanate from the underlying endometrial stroma, the profound differences in trophoblast invasiveness displayed by various species are maintained. In the case of the human, the trophectoderm in contact with the endometrium undergoes differentiation to form two contrasting cell layers. In the outermost layer, fusion among neighboring cells generates the syncytiotrophoblast , a multinucleated, terminally differentiated, and nonproliferative tissue. The deeper cells remain as mononuclear proliferative progenitor cells, the cytotrophoblast cells . Tongues of syncytiotrophoblast begin to penetrate between the endometrial cells, but there is no evidence to suggest that cell death is induced as part of this process. Gradually, over the next few days, the conceptus embeds into the stratum compactum of the endometrium.

Recent ultrasound and comparative data indicate that upgrowth and encapsulation by the endometrium may be just as important as trophoblast invasion in this process. The earliest ex vivo specimens available for study are estimated to be around 7 days postfertilization; in these, the conceptus is almost entirely embedded. A plug of fibrin initially seals the defect in the uterine surface, but by days 10 to 12, the uterine epithelium is restored.

By the time implantation is complete, the conceptus is surrounded entirely by a mantle of syncytiotrophoblast ( Fig. 1.2A ), which tends to be thicker beneath the conceptus, in association with the embryonic pole. Vacuolar spaces appear within the mantle and gradually coalesce to form larger lacunae, the forerunners of the IVS. As the lacunae enlarge, the intervening syncytiotrophoblast is reduced in thickness and forms a complex lattice of trabeculae (see Fig. 1.2B ). Soon after, starting around day 12 after fertilization, the cytotrophoblast cells proliferate and penetrate into the trabeculae. On reaching their tips approximately 2 days later, the cells spread laterally and establish contact with those from other trabeculae to form a new layer interposed between the mantle and the endometrium, the cytotrophoblastic shell (see Fig. 1.2C ). The shell encapsulates the conceptus, initially sealing it off from the maternal tissues, and deficiencies in its development have been linked to miscarriage and other complications of pregnancy. Finally, at the start of the third week of development, mesodermal cells derived from the extraembryonic mesoderm invade the trabeculae, bringing with them the hemangioblasts from which the fetal vascular circulation differentiates. The mesoderm cells do not penetrate right to the tips of the trabeculae; these remain as an aggregation of cytotrophoblast cells—the cytotrophoblast cell columns, which may or may not have a covering of syncytiotrophoblast (see Fig. 1.2C ). Proliferation of the cells at the proximal ends of the columns and their subsequent differentiation contribute to expansion of the cytotrophoblastic shell. Toward the end of the third week, the rudiments of the placenta are therefore in place. The original wall of the blastocyst becomes the chorionic plate, the cytotrophoblastic shell is the precursor of the basal plate, and the lacunae form the IVS (see Fig. 1.2D ). The trabeculae are the forerunners of the villous trees, and repeated lateral branching gradually increases their complexity.

Initially, villi form over the entire chorionic sac, but toward the end of the first trimester, they regress from all except the deep pole, where they remain as the definitive discoid placenta. Abnormalities in this process may account for the persistence of villi at abnormal sites on the chorionic sac, hence the presence of accessory or succenturiate lobes. Also, excessive asymmetric regression may result in the umbilical cord being attached eccentrically to the placental disc.

Amnion and Yolk Sac

While these early stages of placental development are taking place, the inner cell mass differentiates and gives rise to the amnion, yolk sac, and bilaminar germ disc. The amnion, the yolk sac, and fluid compartment in which they lie play an important role in the physiology of early pregnancy; their development will be described. The formation of these sacs has been controversial over the years due mainly to the small number of specimens available for study, with folding and apoptosis both being proposed. Recent in vitro culture of human blastocysts has resolved the issue, demonstrating that the epiblast cells initially form a cluster of cells that transforms upon polarization into a rosette surrounding a small central cavity. The cells in contact with the hypoblast remain cuboidal and contribute to the embryo, whereas those in contact with the trophoblast become more squamous and are the forerunners of the amnion. Meanwhile, the primary yolk sac extends from the hypoblast layer around the inner surface of the trophoblast, separated from it by a loose reticulum of extraembryonic mesoderm. Over the next few days, considerable remodeling of the yolk sac occurs; this involves three closely interrelated processes. First, formation of the primitive streak in the germ disc and the subsequent differentiation of definitive endoderm lead to displacement of the original hypoblast cells into the more peripheral regions of the primary yolk sac. Second, the sac greatly reduces in size, either because the more peripheral portion is nipped off or because it breaks up into a number of vesicles. Third, the reticulum splits into two layers of mesoderm except at the future caudal end of the germ disc, where it persists as a mass; this is the connecting stalk that links the disc to the trophoblast. One layer lines the inner surface of the trophoblast, contributing to formation of the chorion, and the other covers the outer surfaces of the amnion and yolk sac. In between these layers is a large fluid-filled space, the exocoelomic cavity (ECC). The net result of this remodeling is the formation of a smaller secondary yolk sac (SYS); connected to the embryo by the vitelline duct, it floats within the ECC (see Fig. 1.2D ).

The ECC is a conspicuous feature ultrasonographically that can be clearly visualized using a transvaginal probe toward the end of the third week after fertilization (fifth week of gestational age). Between 5 and 9 weeks of pregnancy, it represents the largest anatomic space within the chorionic sac. The SYS is the first structure that can be detected ultrasonographically within that space; its diameter increases slightly between 6 and 10 weeks of gestation to reach a maximum of 6 to 7 mm, and then decreases slightly. Histologically, the SYS consists of an inner layer of endodermal cells linked by tight junctions at their apical surface and bearing a few short microvilli. Their cytoplasm contains numerous mitochondria, whorls of rough endoplasmic reticulum, Golgi bodies, and secretory droplets; this gives them the appearance of being highly active synthetic cells. With further development, the epithelium becomes folded to form a series of cystlike structures or tubules, only some of which communicate with the central cavity. The function of these spaces is not known, although it has been proposed that they serve as a primitive circulatory network in the earliest stages of development because they may contain nonnucleated erythrocytes. On its outer surface, the yolk sac is covered by a layer of mesothelium derived from the extraembryonic mesoderm. This epithelium bears a dense covering of microvilli, and the presence of numerous coated pits and pinocytotic vesicles gives it the appearance of an absorptive epithelium. Experiments in the rhesus monkey have revealed that the mesothelial layer readily engulfs horseradish peroxidase, and the proposed transport function is reinforced by the presence of a well-developed capillary plexus immediately beneath the epithelium that drains through the vitelline veins to the developing liver.

However, by week 9 of pregnancy, the SYS begins to exhibit morphologic evidence of a decline in function. This appears to be independent of the expansion of the amnion, which is gradually drawn around the ventral surface of the developing embryo. As it does this, it presses the yolk sac remnant against the connecting stalk, thus forming the umbilical cord. By the end of the third month, the amnion abuts the inner surface of the chorion and the ECC is obliterated. The fusion of the amnion and chorion and elimination of the ECC can be seen by ultrasound at around 15 weeks of gestation.

Maternal-Fetal Relationship During the First Trimester

For the placenta to function efficiently as an organ of exchange, it requires adequate and dependable access to the maternal circulation. Establishing that access is arguably one of the most critical aspects of placental development; over recent years, it has certainly been one of the most controversial. As the syncytiotrophoblast mantle enlarges, it soon comes in close proximity to superficial veins within the endometrium. These undergo dilation to form sinusoids, which are subsequently tapped into by the syncytium. As a result, maternal erythrocytes come to lie within the lacunae, and their presence has in the past been taken by embryologists as indicating the onset of the maternal circulation to the placenta. If this is a circulation, however, it is entirely one of venous ebb and flow, possibly influenced by uterine contractions and other forces. Numerous traditional histologic studies have demonstrated that arterial connections are not established with the lacunae until much later in pregnancy, although the exact timing was not known for many years. The advent of high-resolution ultrasound and Doppler imaging has appeared to answer this question, for in normal pregnancies most observers agree that moving echoes indicative of significant fluid flow cannot be detected within the IVS until 10 to 12 weeks of gestation.

It is now well accepted on the basis of evidence from a variety of techniques that a major change in the maternal circulation to the placenta takes place at the end of the first trimester. First, direct vision into the IVS during the first trimester with a hysteroscope reveals the cavity to be filled with a clear fluid rather than with maternal bloo d. Second, perfusion of pregnant hysterectomy specimens with radiopaque and other media demonstrates little flow into the IVS during the first trimester except perhaps at the margins of the placental disc. Third, the oxygen concentration within the IVS is low (<20 mm Hg) prior to 10 weeks of pregnancy and rises threefold between weeks 10 and 12 . This rise is matched by increases in the mRNA concentrations encoding and in activities of the principal antioxidant enzymes in the placental tissues that confirm a change in oxygenation at the cellular level . The mechanism that underlies this change in placental perfusion relates to the phenomenon of extravillous trophoblast invasion.

Extravillous Trophoblast Invasion and Physiologic Conversion of the Spiral Arteries

During the early weeks of pregnancy, a subpopulation of trophoblast cells differentiate from the deep surface of the cytotrophoblastic shell and migrate into the endometrium. Because these cells do not take part in the development of the definitive placenta, they are referred to as extravillous trophoblast . Their activities are, however, fundamental to the successful functioning of the placenta, for their presence in the endometrium is associated with the physiologic conversion of the maternal spiral arteries. The cytologic basis of this phenomenon is still not understood, but the net effect is the loss of the smooth muscle cells and elastic fibers from the media of the endometrial segments of the arteries and their subsequent replacement by fibrinoid. Some evidence suggests that this is a two-stage process. Very early in pregnancy, the arteries display endothelial basophilia and vacuolation, disorganization of the smooth muscle cells, and dilation. Because these changes are observed equally in both the decidua basalis and parietalis, and they are also seen within the uterus in ectopic pregnancies, they must be independent of local trophoblast invasion. Instead, it has been proposed that these changes are hormonally driven or the result of activation of decidual renin-angiotensin signaling. Slightly later, during the first few weeks of pregnancy, the invading extravillous trophoblasts become closely associated with the arteries and infiltrate their walls. Dilation ensues, and as a result, the arteries are converted from small-caliber vasoreactive vessels into funnel-shaped flaccid conduits.

The extravillous trophoblast population can be separated into two subgroups: the endovascular trophoblast migrates in a retrograde fashion down the lumens of the spiral arteries, temporarily replacing the endothelium; and the interstitial trophoblast migrates through the endometrial stroma. In early pregnancy, the volume of the migrating endovascular cells is sufficient to occlude, or plug, the terminal portions of the spiral arteries as they approach the basal plate ( Fig. 1.3 ). It is the dissipation of these plugs toward the end of the first trimester that establishes the maternal circulation to the placenta. The mechanism of unplugging of the arteries is unknown at present, but spaces appear in the plugs most likely through cell death. These spaces gradually coalesce to form channels that enlarge with advancing gestational age, establishing communication with the IVS. Trophoblast invasion is not equal across the implantation site; rather, it is greatest in the central region, where it has presumably been established the longest. It is to be expected, therefore, that the plugging of the spiral arteries will be most extensive in this region; this may account for the fact that maternal arterial blood flow is most often first detectable ultrasonographically in the peripheral regions of the placental disc. Associated with this blood flow is a high local level of oxidative stress, which can be considered physiologic because it occurs in all normal pregnancies. It has been proposed that this stress induces regression of the villi over the superficial pole of the chorionic sac, thus forming the chorion laeve ( Fig. 1.4 ).

Under normal conditions, the interstitial trophoblast cells invade as far as the inner third of the myometrium, where they fuse and form multinucleated giant cells. It is essential that the process is correctly regulated; excessive invasion can result in complete erosion of the endometrium and the condition known as placenta accreta (see Chapter 21 ). As they migrate, the trophoblast cells interact with cells of the maternal immune system present within the decidua, in particular macrophages and uterine natural killer (NK) cells. These interactions may play a physiologic role in regulation of the depth of invasion and in the conversion of the spiral arteries. Uterine NK cells accumulate in the endometrium during the secretory phase of the nonpregnant cycle and are particularly numerous surrounding the spiral arteries at the implantation site. Despite their name, no evidence suggests that they destroy trophoblast cells. On the contrary, their cytoplasm contains numerous granules with a diverse array of cytokines and growth factors. Extravillous trophoblast cells express the polymorphic human leukocyte C-antigen (HLA-C), which binds to killer cell immunoglobulin-like receptors (KIRs) on the NK cells. Recent evidence indicates that a degree of activation of the NK cells is necessary for successful pregnancy, most likely because of the release of cytokines, such as granulocyte-macrophage colony stimulating factor (GM-CSF), and proteases. GM-CSF increases the motility of extravillous trophoblast cells and so may mediate spiral artery remodeling. Hence, combinations of HLA-C and KIR subtypes that are generally inhibitory are associated with a high risk of pregnancy complications, which emphasizes the importance of immunologic interactions to reproductive success. These immunologic interactions influence birthweight across the entire range, including macrosomia.

Physiologic conversion of the spiral arteries is often attributed with ensuring an adequate maternal blood flow to the placenta, but such comments generally oversimplify the phenomenon. By itself, the process cannot increase the volume of blood flow to the placenta because it affects only the most distal portion of the spiral arteries. The most proximal part of the arteries, where they arise from the uterine arcuate arteries, always remains unconverted, and will act as the rate-limiting segment. These segments gradually dilate in conjunction with the rest of the uterine vasculature during early pregnancy, most probably under the effects of estrogen; as a result, the resistance of the uterine circulation falls and uterine blood flow increases from approximately 45 mL/min during the menstrual cycle to around 750 mL/min at term or 10% to 15% of maternal cardiac output. Studies in the mouse have demonstrated that the radial arteries account for approximately 90% of the total uteroplacental vascular resistance.

By contrast, the terminal dilation of the arteries will substantially reduce both the rate and pressure with which that maternal blood flows into the IVS. Mathematic modeling has demonstrated that physiologic conversion is associated with a reduction in velocity from 2 to 3 m/s in the nondilated section of a spiral artery to approximately 10 cm/s at its mouth. Reducing the velocity will ensure that the delicate villous trees are not damaged by the momentum of the inflowing blood. Slowing the rate of maternal blood flow across the villous trees will also facilitate exchange, whereas lowering the pressure in the IVS is important to prevent compression and collapse of the fetal capillary network within the villi. Measurements taken in the rhesus monkey indicate that the pressure at the mouth of a spiral artery is only 15 mm Hg and within the IVS is on average 10 mm Hg. The pressure within the fetal villous capillaries is estimated to be approximately 20 mm Hg, providing a pressure differential that favors their distention of 5 mm Hg.

Many complications of pregnancy are associated with defects in extravillous trophoblast invasion and failure to establish the maternal placental circulation correctly. In the most severe cases, the cytotrophoblastic shell is thin and fragmented; this is observed in approximately two-thirds of spontaneous miscarriages . As a result, plugging of the maternal spiral arteries is either incomplete or completely absent, leading to precocious and widespread onset of the maternal circulation throughout the developing placenta. Hemodynamic forces coupled with excessive oxidative stress within the placental tissues are likely to be major factors that contribute to loss of these pregnancies.

In milder cases, the pregnancy may continue but be complicated later by preeclampsia, intrauterine growth restriction (IUGR), or a combination of the two. The physiologic changes are either restricted to the superficial endometrial parts of the spiral arteries or are absent entirely (see Fig. 1.3 ). In the most severe cases of preeclampsia associated with major fetal growth restriction, only 10% of the arteries may be fully converted, compared with 96% in normal pregnancies . There is still debate as to whether this is due to an inability of the interstitial trophoblast to invade the endometrium successfully or, whether having invaded sufficiently deeply, the trophoblast cells fail to penetrate the walls of the arteries. These two possibilities are not mutually exclusive and may reflect different etiologies, including chromosomal aberrations, maternal thrombophilia, endometrial dysfunction, or other problems in the mother.

Whatever the causation, there are several potential consequences to incomplete conversion of the arteries. First, because of the absence of the distal dilation, maternal blood will enter the IVS with greater velocity than normal, forming jet-like spurts that can be detected ultrasonographically. The villous trees are often disrupted opposite these spurts, which leads to the formation of intervillous blood lakes, and the altered hemodynamics within the IVS result in thrombosis and excessive fibrin deposition. Second, incomplete conversion will allow the spiral arteries to maintain greater vasoreactivity than normal. Evidence from rhesus monkeys and humans shows that spiral arteries are not continuously patent but that they undergo periodic constriction independent of uterine contractions. It has been proposed that exaggeration of this phenomenon due to the retention of smooth muscle in the arterial walls may lead to a hypoxia-reoxygenation type of injury in the placenta, which culminates in the development of oxidative stress. Placental oxidative stress and related activation of the unfolded protein response signaling pathways are key factors in the pathogenesis of preeclampsia, in particular of the early-onset subtype. In vitro and clinical evidence suggests that fluctuations in oxygenation are a more physiologic stimulus for its generation than simply reduced uterine perfusion. The third consequence of incomplete conversion is that the distal segments of the arteries are frequently the site of acute atherotic changes. These are likely to be secondary changes, possibly induced by the involvement of these segments in the hypoxia-reoxygenation process or their abnormal hemodynamics; however, if the lesions become occlusive, they will further impair blood flow within the IVS, thus contributing to the growth restriction.

Role of the Endometrium During the First Trimester

Signals from the uterine epithelium and secretions from the endometrial glands play a major role in regulating receptivity at the time of implantation, but the potential contribution of the glands to fetal development once implantation is complete has largely been ignored. This has been due to the general assumption that once the conceptus is embedded within the uterine wall, it no longer has access to the secretions in the uterine lumen. However, the fact that glycodelin-A, formerly referred to as PP14 or α ₂ -PEG, is derived from the glands and yet accumulates within the amniotic fluid with concentrations that peak at around 10 weeks’ gestation indicates that the placenta must be extensively exposed to glandular secretions throughout the first trimester. This exposure was confirmed by a review of archival placenta in situ hysterectomy specimens that demonstrated glands discharging their secretions into the IVS through openings in the basal plate throughout the first trimester (see Fig. 1.2 ). Their secretions, referred to as “uterine milk,” are a heterogeneous mix of maternal proteins; carbohydrates, including glycogen; and lipid droplets phagocytosed by the syncytiotrophoblast.

The secretions are not just a supply of nutrients but are also rich in growth factors such as leukemia inhibitory factor, GM-CSF, vascular endothelial growth factor (VEGF), epidermal growth factor (EGF), and transforming growth factor-β (TGF-β. Receptors for these factors are present on the villous tissues, and when EGF is applied exogenously to explants of villous tissues derived from early pregnancy, it stimulates division of the cytotrophoblast cells. In other species there is strong evidence that the activity of the glands and expression of the growth factors is stimulated during early pregnancy in response to lactogenic hormones secreted by the trophoblast. In this way, the placenta is able to stimulate its own development through what has been referred to as a servomechanism.

Based on human chorionic gonadotropin (hCG) and placental lactogen from the trophoblast and prolactin from the decidua, we have speculated that an equivalent signaling pathway operates in the human. There is no molecular evidence to support this hypothesis at present, but the gland cells do undergo characteristic hypersecretory morphologic changes, the so-called Arias-Stella reaction, during early pregnancy. In addition, the pattern of sialylation of the secretions changes between the late secretory phase of the nonpregnant cycle and early pregnancy. A loss of terminal sialylic acid caps occurs, which will render the secretions more easily degradable by the trophoblast following their phagocytic uptake. In addition, the change will ensure that any of the growth factors that gain access to the maternal circulation via the uterine veins will rapidly be cleared by asialoglycoprotein receptors in the maternal liver. Hence a unique proliferative microenvironment can be created within the IVS of the early placenta without placing the mother's tissues at risk of excessive stimulation. The recent derivation of organoid cultures of endometrial glands that are responsive to early pregnancy hormones opens new possibilities for elucidation of the signaling dialogue between the glands and the trophoblast during early pregnancy. Glycoproteins secreted by the glands, in particular glycodelin-A, have also been implicated in the regulation of placentation through their effects on trophoblast invasion, hormone secretion, and immune modulation at the maternal-fetal interface.

Ultrasonographic measurements suggest that an endometrial thickness of 8 mm or more is necessary for successful implantation , although not all studies have found such an association. Nonetheless, these measurements are in line with observations based on placenta-in-situ specimens, in which an endometrial thickness of over 5 mm was reported beneath the conceptus at 6 weeks of gestation. Gradually, over the remainder of the first trimester, the endometrium regresses, so that the thickness is reduced to 1 mm by 14 weeks of gestation. Histologically, there is also a transformation in the glandular epithelial cells over this period, from columnar to cuboidal. Secretory organelles become less prominent, although the lumens of the glands are still filled with secretions.

The overall picture is that the glands are most prolific and active during the early weeks of pregnancy and that their contribution gradually wanes during the first trimester. This would be consistent with a progressive switch from histotrophic to hemotrophic nutrition as the maternal arterial circulation to the placenta is established.

Increasing evidence indicates that some of the common complications of pregnancy, such as preeclampsia, have an endometrial rather than a primarily placental origin. However, attempts to correlate the functional activity of the glands with pregnancy outcome have met with mixed success. Thus reduced concentrations of mucin 1, glycodelin-A, and leukemia inhibitory factor have been reported in uterine flushings from women who have suffered repeated miscarriages. However, other studies have shown no significant association between the expression of these markers within the endometrium and outcome.

From the evidence available, it would appear that the functional importance of the endometrial glands to a successful pregnancy extends well beyond the time of implantation and that they play a key role in placental development during the first trimester.

Topology of the Villous Trees

One of the principal functions of the placenta is diffusional exchange, and the physical requirements for this impose the greatest influence on the structure of the organ. The rate of diffusion of an inert molecule is governed by Fick's law, so it is proportional to the surface area for exchange divided by the thickness of the tissue barrier. A large surface area will therefore facilitate exchange, which is achieved by repeated branching of the villous trees.

The villous trees arise from the trabeculae interposed between the lacunae (see Fig. 1.2 ) through a gradual process of remodeling and lateral branching. Initially the different branches have an almost uniform composition and the villi can be separated only by their relative size and position in the hierarchic branching pattern. At this stage, the mesodermal core is loosely packed; at the proximal end of the trees, it blends with the extraembryonic mesoderm that lines the ECC. The stromal cells possess sail-like processes that link together to form fluid-filled channels oriented parallel to the long axis of the villi. Macrophages are often seen within these channels, so it is possible that they function as a primitive circulatory system prior to vasculogenesis. In this way proteins derived from the uterine glands could freely pass into the coelomic fluid, and it is notable that the macrophages within the channels are strongly immunoreactive for maternal glycodelin-A secreted from the glands.

Toward the end of the first trimester, the villi begin to differentiate into their principal types. The connections to the chorionic plate become remodeled to form stem villi, which represent the supporting framework of each villous tree. These progressively develop a compact fibrous stroma and contain branches of the chorionic arteries and accompanying veins. The arteries are centrally located and are surrounded by a cuff of smooth muscle cells. Although these have the appearance of resistance vessels, physiologic studies indicate that, under normal conditions, the fetal placental circulation operates under conditions of full vasodilation. Stem villi contain only a few small-caliber capillaries, so they play little role in placental exchange.

After several generations of branching, stem villi give rise to intermediate villi. These are longer and more slender in form and can be of two types: immature and mature. The former are seen predominantly in early pregnancy and represent a persistence of the nondifferentiated form, as indicated by the presence of stromal channels. Mature intermediate villi provide a distributing framework, and terminal villi arise at intervals from their surface. Within the core are arterioles and venules but also a significant number of nondilated capillaries, which suggest a limited capacity for exchange.

The main functional units of the villous tree are, however, the terminal villi. There is no strict definition as to where a terminal villus starts, but they are most often short, stubby branches up to 100 µm in length and approximately 80 µm in diameter that arise from the lateral aspects of intermediate villi ( Fig. 1.5 ). They are highly vascularized by capillaries alone and are highly adapted for diffusional exchange, as will be seen later.

This differentiation of the villi coincides temporally with the development of the lobular architecture, and the two processes are most likely interlinked. Lobules can be first identified during the early second trimester, following onset of the maternal circulation, when it is thought hemodynamic forces may shape the villous tree. Convincing radiographic and morphologic evidence shows that maternal blood is delivered into the center of the lobule and that it then disperses peripherally, as in the rhesus monkey placenta. Consequently it is to be expected that an oxygen gradient will exist across the lobule, and differences in the activities and expression of antioxidant enzymes within the villous tissues suggest strongly that this is the case. Other metabolic gradients (e.g., glucose concentration) may also exist, and together these may exert powerful influences on villous differentiation. Villi in the center of the lobule, where the oxygen concentration will be highest, display morphologic and enzymatic evidence of relative immaturity; therefore this is considered to be the germinative zone. By contrast, villi in the periphery of the lobule are better adapted for diffusional exchange.

Elaboration of the villous tree is a progressive event that continues at a steady pace throughout pregnancy. By term, the villi present a surface area of 10 to 14 m ² . This may be significantly reduced in cases of IUGR, although this principally reflects an overall reduction in placental volume rather than maldevelopment of the villous tree. In cases of preeclampsia alone, villous surface area is normal and is compromised only when associated with growth restriction. Attempts have recently been made to monitor placental growth longitudinally during pregnancy using ultrasound. Although the data show considerable individual variability, they indicate that in cases of growth restriction or macrosomia, placental volume is significantly reduced or increased, respectively, at 12 to 14 weeks. These findings suggest that ultimate placental size has its origins firmly in the first trimester.

Integrity of the Villous Membrane

One situation that may alter the balance of the two populations of nuclei is damage to the trophoblast layers and the requirement for repair. Isolated areas of syncytial damage, often referred to as sites of focal syncytial necrosis, are a feature of all placentas, although they are more common in those from pathologic pregnancies. Their origin remains obscure, but they could potentially arise from altered hemodynamics within the IVS or from physical interactions between villi. One striking example of the latter is the rupture of the syncytial bridges that form between adjacent villi and lead to circular defects on the surface 20 to 40 µm in diameter. Disruption of the microvillous surface leads to the activation of platelets and the deposition of a fibrin plaque on the trophoblastic basement membrane. Apoptosis of syncytial nuclei has been reported in the immediate vicinity of such plaques, but whether this reflects cause or effect has yet to be determined. With time, cytotrophoblast cells migrate over the plaque, differentiate, and fuse to form a new syncytiotrophoblastic layer. As a result, the plaque is internalized and the integrity of the villous surface is restored. In the interim, however, these sites are nonselectively permeable to creatinine and may represent a paracellular route for placental transfer.

In the past, more widespread apoptosis in the syncytiotrophoblast has been reported, with the interpretation that this reflects increased turnover of the trophoblast in pathologic conditions. However, recent research has clarified that although rates of apoptosis are increased in preeclampsia and IUGR, the cell death is confined to the cytotrophoblast cells.

Extensive damage to the syncytiotrophoblast is seen in cases of missed miscarriage, in which complete degeneration and sloughing of the layer can occur. Although apoptosis and necrosis are increased among the cytotrophoblast cells, the remaining cells differentiate and fuse to form a new and functional syncytial layer. A similar effect is observed when villi from either first-trimester or term placentas are maintained under ambient conditions in vitro.

Thus it is likely that considerable turnover of the syncytiotrophoblast takes place over the course of a pregnancy, although in the absence of longitudinal studies it is impossible to determine the extent of this phenomenon. Nonetheless it is clear that the villous membrane cannot be considered an intact physical barrier and that other elements of the villous trees may play important roles in regulating maternal-fetal transfer.

Placental Vasculature

The development of the fetal vasculature begins during the third week after conception (the fifth week of pregnancy) with the de novo formation of capillaries within the villous stromal core. Hemangioblastic cell cords differentiate under the influence of growth factors such as basic fibroblast growth factor and VEGF. By the beginning of the fourth week, the cords have developed lumens and the endothelial cells become flattened. Surrounding mesenchymal cells become closely apposed to the tubes and differentiate to form pericytes. During the next few days connections form between neighboring tubes to form a plexus, which ultimately unites with the allantoic vessels developing in the connecting stalk to establish the fetal circulation to the placenta.

Exactly when an effective circulation is established through these vessels is difficult to determine. First, the connection between the corporeal and extracorporeal fetal circulations is initially particularly narrow, which suggests that there can be little flow. Second, the narrow caliber of the villous capillaries coupled with the fact that the fetal erythrocytes are nucleated during the first trimester and hence are not readily deformable will ensure that the circulation presents a high resistance to flow. This is reflected in the Doppler waveform obtained during the first trimester, and the resistance gradually falls as the vessels enlarge over the ensuing weeks.

Early in pregnancy there are relatively few pericytes; the capillary network is labile and undergoes considerable remodeling. Angiogenesis continues until term and results in the formation of capillary sprouts and loops. Both of these processes contribute to the elaboration of terminal villi . The caliber of the fetal capillaries is not constant within intermediate and terminal villi. Frequently, on the apex of a tight bend, the capillaries become greatly dilated and form sinusoids. These regions may help to reduce vascular resistance and facilitate distribution of fetal blood flow through the villous trees . Equally important is the fact that the dilations bring the outer wall of the capillaries into close juxtaposition with the overlying trophoblast. The trophoblast is locally thinned; as a result, the diffusion distance between the maternal and fetal circulations is reduced to a minimum (see Fig. 1.5 ). Because of their morphologic configuration, these specializations are referred to as vasculosyncytial membranes, and computational modeling reveals that they are the principal sites of gaseous and other diffusional exchange . The arrangement can be considered analogous to that in the alveoli of the lung, where the pulmonary capillaries indent into the alveolar epithelium in order to reduce the thickness of the air-blood diffusion barrier. Thinning of the syncytial layer will not only increase the rate of diffusion into the fetal capillaries but also reduce the amount of oxygen extracted by the trophoblast en route. The syncytiotrophoblast is highly active metabolically because of the high rates of protein synthesis and ionic pumping; however, by having an uneven distribution of the tissue around the villous surface, the oxygen demands of the fetus and the placenta can be separated to a large extent.

It is notable that the development of vasculosyncytial membranes is seen to its greatest extent in the peripheral regions of a placental lobule, where the oxygen concentration is lowest, and also in placentas from high altitudes. In both instances, it is associated with enlargement of the capillary sinusoids and may be viewed as an adaptive response aimed at increasing the diffusing capacity of the placental tissues. Conversely, an increase in the thickness of the villous membrane is often seen in cases of IUGR and in placentas from cigarette smokers. As mentioned earlier, the hydrostatic pressure differential across the villous membrane is an important determinant of the diameter of the capillary dilations and hence of the villous membrane's thickness. Raising the pressure in the IVS not only compresses the capillaries but also increases the resistance within the umbilical circulation. Both effects will impair diffusional exchange, which highlights the importance of full conversion of the spiral arteries.

Vascular changes are observed in many complications of pregnancy , where they may underpin changes in the topology of the villous tree. Increased branching of the vascular network is observed in placentas from high altitudes, which cause the terminal villi to be shorter and more clustered than normal. At present no experimental data indicate that this has any impact on placental exchange; in theory, however, this shortening of the arteriovenous pathway may lead to increased efficiency.

Physiology of the Secondary Yolk Sac and Exocoelomic Cavity

Now that development of the placenta and the extraembryonic membranes has been covered, we turn to their physiologic roles during pregnancy. Phylogenetically, the oldest membrane is the yolk sac, and the SYS plays a major role in the embryonic development of all mammals. The function of the yolk sac has been most extensively studied in laboratory rodents. It has been demonstrated that it is one of the initial sites of hematopoiesis, that it synthesizes a variety of proteins, and that it is involved in maternal-fetal transport.

The endodermal layer of the human SYS is known to synthesize several serum proteins in common with the fetal liver, such as α-fetoprotein (AFP), α ₁ -antitrypsin, albumin, prealbumin, and transferrin. With rare exceptions, the secretion of most of these proteins is confined to the embryonic compartments and the contribution of the SYS to the maternal protein pool is limited. This can explain why their concentrations are always higher in the ECC than in maternal serum. AFP is also produced by the embryonic liver from 6 weeks until delivery; it has a high molecular weight (±70 kDa) and, unlike hCG, is found in similar amounts on both sides of the amniotic membrane. Analysis of molecular variants of AFP that have an affinity for concanavalin A have demonstrated that AFP molecules within both the coelomic and amniotic fluids are mainly of yolk sac origin, whereas maternal serum AFP molecules are principally derived from the fetal liver. These results suggest that the SYS also has an excretory function and secretes AFP toward the embryonic and extraembryonic compartments. By contrast, AFP molecules of fetal liver origin are probably transferred from the fetal circulation to the maternal circulation, mainly across the placental villous membrane.

The potential absorptive role of the yolk sac membrane has been evaluated by examining the distribution of proteins and enzymes between the ECC and SYS fluids and by comparing the synthesizing capacity of the SYS, fetal liver, and placenta for hCG and AFP. The distribution of the trophoblast-specific hCG in yolk sac and coelomic fluids together with the absence of hCG mRNA expression in yolk sac tissues provided the first biologic evidence of its absorptive function. Similarities in the composition of the SYS and coelomic fluids suggest that a free transfer for most molecules occurs between the two corresponding compartments. Conversely, an important concentration gradient exists for most proteins between the ECC and the amniotic cavity, indicating that transfer of molecules is limited at the level of the amniotic membrane. More recently, RNA-seq analysis of the SYS has been performed and the transcriptome compared with those of the murine and chicken yolk sacs, where an absorptive function has been well documented experimentally. Transcripts involved in lipid transport and cholesterol handling were among the most abundant (i.e., top 0.5%), and were matched by the high levels of apolipoproteins (ApoB, ApoA1, ApoA2, and ApoA4) in the coelomic fluid. Transcripts encoding various categories of transporter proteins were also well represented, including those for the adenosine triphosphate (ATP)-binding cassette family of efflux transporters.

These findings, and the high degree of conservation of transcripts among the three species, suggest that the human yolk sac membrane is an important zone of transfer between the extraembryonic and embryonic compartments and that the main flux of molecules occurs from outside the yolk sac—that is, from the ECC—in a direction toward its lumen and subsequently to the embryonic gut and circulation. The identification by immunohistochemistry of specific transfer proteins on the mesothelial covering and of the multifunctional endocytic receptors megalin and cubilin confirms the RNA-seq data and lends support to this concept. When, after 10 weeks of gestation, the cellular components of the wall of the SYS start to degenerate, this route of transfer is no longer functional, and most exchanges between the ECC and the fetal circulation must then take place at the level of the chorionic plate.

The development and physiologic roles of the ECC are intimately linked with that of the SYS, for which it provides a stable environment. The concentrations of hCG, estriol, and progesterone are higher in the coelomic fluid than in maternal serum and strongly suggest the presence of a direct pathway between the trophoblast and the ECC. Morphologically, this may be via the villous stromal channels and the loose mesenchymal tissue of the chorionic plate. Protein electrophoresis has also shown that the coelomic fluid results from an ultrafiltrate of maternal serum with the addition of specific placental and SYS bioproducts. For the duration of the first trimester, the coelomic fluid remains straw-colored and more viscous than the amniotic fluid, which is always clear. This is mainly due to the higher protein concentration in the coelomic fluid than in the amniotic cavity. The concentration of almost every protein is higher in coelomic fluid than in amniotic fluid, ranging from 2 to 50 times higher depending on the molecular weight of the protein investigated. The coelomic fluid has a very slow turnover, so the ECC may act as a reservoir for nutrients needed by the developing embryo. These findings suggest that the ECC is a physiologic liquid extension of the early placenta and an important interface in fetal nutritional pathways. Molecules such as vitamin B ₁₂ , prolactin, glucose and glycodelin-A are known to be mainly produced by the uterine decidua. This pathway may be pivotal in providing the developing embryo with sufficient nutrients before the intervillous circulation becomes established. We speculate that nutrients derived from the endometrial glands cross the early placenta into the ECC, where are then taken up and transported to the embryo by the SYS ( Fig. 1.6 ). In this way the early placenta functions physiologically as a choriovitelline placenta equivalent to that of most mammals during the initial stages of pregnancy, although the SYS never makes physical contact with the inner surface of the chorion .

Some analogies can be drawn between the ECC and the antrum within a developing graafian follicle. It has been suggested that the evolution of the latter was necessary to overcome the problem of oxygen delivery to an increasingly large mass of avascular cells. Because the contained fluid has no oxygen consumption, it will permit diffusion more freely than an equivalent thickness of cells. However, because neither follicular nor coelomic fluids contain an oxygen carrier, the total oxygen content must be low. An oxygen gradient will inevitably exist between the source and the target, whether it be an oocyte or an embryo. Measurements in human patients undergoing in vitro fertilization have demonstrated that the oxygen tension in follicular fluid falls as follicular diameter, assessed by ultrasound, increases. Thus diffusion across the ECC may be an important route of oxygen supply to the embryo before the development of a functional placental circulation, but it will maintain the early fetus in a low-oxygen environment. This may serve to protect the fetal tissues from damage by O ₂ free radicals and may prevent the disruption of signaling pathways during the crucial stages of embryogenesis and organogenesis. The presence in the ECC of molecules with a well-established antioxidant role—such as taurine, transferrin, vitamins A and E, and selenium—supports this hypothesis. Associated with this, the low-oxygen environment may also favor the maintenance of “stemness” in embryonic and placental stem cells. It is notable that the proliferative capacity of the placenta rapidly declines at the end of the first trimester, which may reflect the loss of growth factor stimulation from the endometrial glands or the rise in intraplacental oxygen concentration.