Deep placentation in human pregnancy is realised by deep invasion of the placental bed by the extravillous trophoblast, involving the decidua and the inner (junctional zone) myometrium. Interstitial invasion of the stroma and endovascular trophoblast invasion of the spiral arteries both occur. Deep endovascular trophoblast invasion into the myometrial segments of spiral arteries is important for proper placental functioning. Before this extended vascular invasion begins, decidua-associated vascular remodelling, which includes swelling and disorganisation of the vascular smooth muscle, occurs during a period of rising placental oxygen. This early remodelling step may accommodate the progressively increasing maternal blood flow to the developing placenta. The subsequent trophoblast-associated remodelling step enhances and stabilises the widening of the vessels, whereas the vascular smooth muscle and elastic lamina are replaced by a fibrinoid matrix with embedded trophoblast. Defective deep remodelling contributes to placental malfunctioning in complications of pregnancy.
According to Mossman’s definition, the placenta is an apposition of fetal and parental tissues for physiological exchange. This process has led to the elaboration of a variety of modifications of maternal–fetal tissue contacts and interaction during the evolution of the different mammalian orders. The haemochorial placenta represents the closest apposition of the maternal and fetal circulations, in which optimization of maternal blood supply is at least partly affected by the invasive activity of a subcategory of fetal trophoblasts. In humans, this ‘extravillous’ trophoblast invasion reaches greater depth than in most mammalian species, including most non-human primates. This invasion is associated with extensive vascular remodelling of myometrial and decidual segments of spiral arteries. Apart from this extended vascular remodelling, humans also differ in other features from primates with shallow placentation, such as interstitial blastocyst implantation, a high degree of decidual erosion, causing a shift of spiral artery outlets into the intervillous space, and possibly in the haemodynamical characteristics of intraplacental maternal flow. It should be evident that trophoblast invasion, with its associated spiral artery remodelling, is not an isolated phenomenon but forms an integrated part of the total placental development, which has been shaped in the course of our evolutionary history.
Invasion of the uterine wall begins with the implantation of the blastocyst. This initiating event proceeds to a deeper level in the human endometrium than in most other primates. This is because implantation is interstitial instead of superficial and results in the formation of a decidua capsularis. This feature is shared with the great apes, but not with Old World and New World monkeys. In all primates, implantation is initiated by the appearance of syncytial trophoblast in the blastocyst wall. Once interstitial implantation is complete, the blastocyst is completely surrounded by this ‘primitive’ syncytium. For some time, syncytial cells were regarded as primordial invasive trophoblasts in humans.
Villous trees and intervillous flow in primate placentae
The first step in maternal–fetal exchange in the developing placenta is the formation of maternal blood-filled lacunae within the thickening syncytial wall of the implanted human blastocyst. It is not clear whether this primordial ‘intervillous space’ merely represents extravasated maternal blood or is derived from engulfed parts of the subepithelial capillary network that have lost their endothelial lining. During the first weeks, these maternal blood spaces only communicate with decidual venules, but, at 8 weeks, connecting channels develop between the lacunae and spiral arteries. At this point, the anatomical basis for ‘intervillous’ maternal blood flow is established, but is contingent upon connecting channels and corresponding spiral artery outlets being clear of intravascular trophoblastic plugs.
Meanwhile, strands of mononuclear cytotrophoblast start to proliferate at the inner (fetal) side of the implanted blastocyst wall. The resulting cytotrophoblastic columns push themselves into the primitive syncytial mass, forming primary, secondary and, after the formation of fetal capillaries, tertiary villi. The most distal cytotrophoblasts subsequently break through the syncytium and spread laterally to form a shell of extravillous cytotrophoblasts, separating the placenta from the superficial decidua. In humans, this continuous shell is only a temporary structure, and begins to disintegrate early in pregnancy. Extravillous trophoblasts are still generated from the cell columns at the tips of anchoring villi, but most cells immediately start to migrate deeper into the maternal tissues. Fetal circulation starts at around 6 weeks as soon as the fetal capillary network has been connected with the fetal heart. Maternal flow starts somewhat later and gradually increases from ∼8 to 12 weeks, while intravascular plugs are dissolved and endovascular trophoblast migration into the spiral arteries begins. Only then can the placenta be considered as fully functional, with both fetal and maternal circulations being established. It has been argued that, because intra-arterial trophoblast invasion (and plugging) begins at the centre of the placental bed, maternal flow into the placenta will start from the most lateral non-plugged spiral arteries. Therefore, increasing oxidative stress in these lateral regions is responsible for the progressive regression of the ‘chorion frondosum’ to a ‘chorion laeve’.
During further placental development, intermediate and terminal villi branch off from the main stem villi, forming a system of villous trees, each consisting of a villous stem and side branches. Consequently, the placenta subdivides into functional units, the placental lobules. In casts of young placentae, each lobule seems to contain a relatively empty centre within the denser villous tissue. It was originally thought that the central open spaces would be created by the force of the incoming maternal blood, entering the intervillous space in the intralobular regions of the placental floor. In the human placenta at term, however, most spiral artery outlets tend to be distributed to interlobular areas. This contrasts sharply with monkey (i.e., rhesus monkey and baboon) placentae, in which spiral artery outlets are nearly always located in intralobular regions of the basal plate. This difference is likely to be caused by the more extensive decidual erosion occurring in human pregnancy, as a result of more intensive and deeper trophoblast invasion.
Intervillous maternal flow patterns also need to be considered. There is general agreement that, in most placentae, maternal and fetal gross flow directions are countercurrent, providing the most efficient system of exchange. This concept can be neatly illustrated by the organisation of rodent placentae. The bulk of maternal blood is delivered by spiral arteries and carried directly to the fetal side of the placenta via transplacental maternal channels; the main umbilical arterial branches carry the fetal blood to the maternal side of the labyrinth. Physiological exchange then occurs between fetal capillaries and the narrow maternal blood spaces of the labyrinth, while fetal and maternal blood flow back in grossly opposite directions. Such gross flow requirements differ from the anatomical arrangements of monkey and human placentae, as the spiral arteries open freely in the intervillous space at the basal plate. This poses the problem that oxygen-rich blood may remain concentrated around the spiral artery outlets, failing to spread through the whole thickness of the placenta. Radiological studies carried out by Ramsey et al. in rhesus monkey placentae revealed a haemodynamical solution to this problem. They showed that maternal blood leaves the spiral arteries in vigorous spurts, directed towards the chorionic plate. It has been assumed, but not completely confirmed, that a similar maternal bulk flow by spurts also occurs in the human placenta. Such maternal flow across the whole placental thickness requires a system of free-floating villi. Placentae of New World monkeys, however, are ‘pseudolabyrinthine’ showing numerous intervillous connections. The intervillous spreading of maternal blood by spiral artery ‘spurts’ would therefore be highly inefficient. Instead, the transplacental bulk transport of maternal blood in these animals is guided through maternal channels, which are comparable to analogous structures in labyrinthine rodent placentae.
Decidualisation, decidual erosion and the shaping of the basal plate
Around the time of implantation, a process of decidualisation begins spontaneously in the human endometrium. The process involves swelling of connective tissue cells and changes in the extracellular matrix. It also involves early remodelling changes in decidual spiral arteries (e.g., endothelial swelling and vacuolisation and vascular smooth muscle disorganisation). At the same time, the leucocytic content of the endometrium is drastically altered, mainly by infiltrations of macrophages and uterine natural killer cells, which may play an initiating role in early vascular remodelling. The whole complexity of decidual function is still poorly understood, and may not be limited to a nutritional, endocrine or immune role. Early decidual outgrowth may contribute to the interstitial type of implantation, which may depend on invasiveness of the conceptus and the capacity of the decidualising endometrial tissue to encapsulate the embryo, acting at the same time as a ‘biosensor’ for recognising (and eliminating) defective embryos. Optimal decidualisation may need proper ‘conditioning’ of the endometrium, which might require a succession of several menstrual cycles.
In early pregnancy, a compact decidual layer directly faces the cytotrophoblastic shell. In Old World monkeys, such as the baboon, this remains largely intact throughout pregnancy. In humans, however, the shell disintegrates while the decidua is progressively invaded by the interstitial trophoblast. This invasion starts at the centre of the placental bed and progressively spreads towards the periphery in a ring-like fashion. Among these migrating cells, ‘streamers’ of bi-, tri- or multinuclear trophoblasts can be observed, indicating either a fragmentation of syncytial layers displaced during the formation of the cytotrophoblastic shell, or a gradual fusion of invading mononuclear trophoblasts. Such multinuclear giant cells may reach enormous sizes deeper in the decidua, and often lie in parallel rows facing the junction between the decidua and inner myometrium. It has been suggested that, because of their bulky size, penetration of the inner myometrial surface must indeed be difficult. Although the decidualising endometrium of early human pregnancy may literally be swamped by interstitial trophoblasts, this is not the case in the baboon, in which few, if any, interstitial trophoblasts are seen in the deeper decidual layers. This also seems to be the case in gibbons, the ‘lowest’ of the great apes, despite their interstitial mode of implantation, which might have raised the expectation that they share invasive features with the other great apes.
In contrast to baboon and rhesus monkey placentae, the borderline between the placental basal plate and the decidua is extremely irregular in humans. Furthermore, the more lateral spiral arteries in the human placental bed often show an almost parallel orientation to the basal plate. In such vessels, additional openings into the intervillous space may form, which are possibly responsible for necrotising changes within the decidual stretches containing the most distal spiral artery segments that had been cut short by the new outlets. This positional shift of the lateral spiral artery outlets is most likely the result of lateral expansive growth of the placental disk, causing stretching and flattening of the underlying maternal endometrial structures, as indicated by the tilted stacks of endometrial glands in the lateral areas of the placental bed. The additional openings in the spiral artery walls are most probably created by the action of the invasive interstitial trophoblast. One can also conceive that, during the progressing lateral outgrowth of the placenta, necrotising plaques of decidual tissue, still attached to anchoring villi, may be tilted towards the centre and protrude into the placental tissue. In this way, placental septa may be formed, as they are known to possess a core of decidual tissue. Decidual erosion and formation of placental septa are less obvious in monkey placentae because of the lower invasive activity of the interstitial trophoblast and a restricted placental expansion. This scenario of septa formation is also supported by the observation that most spiral artery openings into the intervillous space occur at the bases of protruding septa. Increased decidual erosion may explain the shift of arterial outlets from the intralobular to interlobular in humans, chimpanzees and gorillas, whereas, in Old World monkeys (baboons and rhesus monkeys), such outlets remain directed towards the lobular centres in late pregnancy. A restricted decidual erosion in monkeys was already revealed in 1786 by the very first description of a monkey placenta by John Hunter, who explicitly mentioned a much thicker decidua underlying the placenta as compared with humans.
The more prominent eroding action of the placenta on the underlying decidua in humans may be caused by an intenser invasiveness of the trophoblast, which might be supported by more extensive decidualisation. This concept contradicts the mere invasion-restricting function previously ascribed to the decidua in rodents. Indeed, there is increasing evidence, at least in humans, that decidualisation may initiate and even stimulate trophoblast invasion. A final point is that decidualisation in humans is no longer thought to be restricted to the endometrium itself, but that decidua-associated tissue alterations are extended to the inner myometrium, referred to as ‘junctional zone myometrium’. Changes in this ‘junctional zone’ include extracellular matrix changes and associated ‘loosening’ of the myometrial smooth muscle, as well as early remodelling of the myometrial spiral artery segments. This early decidua-associated transformation of the inner myometrium might provide an additional stimulus for the deeper trophoblast invasion beyond the decidual–myometrial junction.
Decidualisation, decidual erosion and the shaping of the basal plate
Around the time of implantation, a process of decidualisation begins spontaneously in the human endometrium. The process involves swelling of connective tissue cells and changes in the extracellular matrix. It also involves early remodelling changes in decidual spiral arteries (e.g., endothelial swelling and vacuolisation and vascular smooth muscle disorganisation). At the same time, the leucocytic content of the endometrium is drastically altered, mainly by infiltrations of macrophages and uterine natural killer cells, which may play an initiating role in early vascular remodelling. The whole complexity of decidual function is still poorly understood, and may not be limited to a nutritional, endocrine or immune role. Early decidual outgrowth may contribute to the interstitial type of implantation, which may depend on invasiveness of the conceptus and the capacity of the decidualising endometrial tissue to encapsulate the embryo, acting at the same time as a ‘biosensor’ for recognising (and eliminating) defective embryos. Optimal decidualisation may need proper ‘conditioning’ of the endometrium, which might require a succession of several menstrual cycles.
In early pregnancy, a compact decidual layer directly faces the cytotrophoblastic shell. In Old World monkeys, such as the baboon, this remains largely intact throughout pregnancy. In humans, however, the shell disintegrates while the decidua is progressively invaded by the interstitial trophoblast. This invasion starts at the centre of the placental bed and progressively spreads towards the periphery in a ring-like fashion. Among these migrating cells, ‘streamers’ of bi-, tri- or multinuclear trophoblasts can be observed, indicating either a fragmentation of syncytial layers displaced during the formation of the cytotrophoblastic shell, or a gradual fusion of invading mononuclear trophoblasts. Such multinuclear giant cells may reach enormous sizes deeper in the decidua, and often lie in parallel rows facing the junction between the decidua and inner myometrium. It has been suggested that, because of their bulky size, penetration of the inner myometrial surface must indeed be difficult. Although the decidualising endometrium of early human pregnancy may literally be swamped by interstitial trophoblasts, this is not the case in the baboon, in which few, if any, interstitial trophoblasts are seen in the deeper decidual layers. This also seems to be the case in gibbons, the ‘lowest’ of the great apes, despite their interstitial mode of implantation, which might have raised the expectation that they share invasive features with the other great apes.
In contrast to baboon and rhesus monkey placentae, the borderline between the placental basal plate and the decidua is extremely irregular in humans. Furthermore, the more lateral spiral arteries in the human placental bed often show an almost parallel orientation to the basal plate. In such vessels, additional openings into the intervillous space may form, which are possibly responsible for necrotising changes within the decidual stretches containing the most distal spiral artery segments that had been cut short by the new outlets. This positional shift of the lateral spiral artery outlets is most likely the result of lateral expansive growth of the placental disk, causing stretching and flattening of the underlying maternal endometrial structures, as indicated by the tilted stacks of endometrial glands in the lateral areas of the placental bed. The additional openings in the spiral artery walls are most probably created by the action of the invasive interstitial trophoblast. One can also conceive that, during the progressing lateral outgrowth of the placenta, necrotising plaques of decidual tissue, still attached to anchoring villi, may be tilted towards the centre and protrude into the placental tissue. In this way, placental septa may be formed, as they are known to possess a core of decidual tissue. Decidual erosion and formation of placental septa are less obvious in monkey placentae because of the lower invasive activity of the interstitial trophoblast and a restricted placental expansion. This scenario of septa formation is also supported by the observation that most spiral artery openings into the intervillous space occur at the bases of protruding septa. Increased decidual erosion may explain the shift of arterial outlets from the intralobular to interlobular in humans, chimpanzees and gorillas, whereas, in Old World monkeys (baboons and rhesus monkeys), such outlets remain directed towards the lobular centres in late pregnancy. A restricted decidual erosion in monkeys was already revealed in 1786 by the very first description of a monkey placenta by John Hunter, who explicitly mentioned a much thicker decidua underlying the placenta as compared with humans.
The more prominent eroding action of the placenta on the underlying decidua in humans may be caused by an intenser invasiveness of the trophoblast, which might be supported by more extensive decidualisation. This concept contradicts the mere invasion-restricting function previously ascribed to the decidua in rodents. Indeed, there is increasing evidence, at least in humans, that decidualisation may initiate and even stimulate trophoblast invasion. A final point is that decidualisation in humans is no longer thought to be restricted to the endometrium itself, but that decidua-associated tissue alterations are extended to the inner myometrium, referred to as ‘junctional zone myometrium’. Changes in this ‘junctional zone’ include extracellular matrix changes and associated ‘loosening’ of the myometrial smooth muscle, as well as early remodelling of the myometrial spiral artery segments. This early decidua-associated transformation of the inner myometrium might provide an additional stimulus for the deeper trophoblast invasion beyond the decidual–myometrial junction.
Deep interstitial trophoblast invasion
From 8 weeks onwards, huge numbers of mononuclear extravillous trophoblasts appear within the inner (junctional zone) myometrium of the human placental bed, but not outside this region. Although immunohistochemical techniques were not readily available at the time of the initial studies, the outstanding morphological features of these cells (a striking basophilia and an oval shape) allowed easy distinction from the surrounding myometrial tissue. By studying the inner myometrium in conjunction with the overlying decidua, the continuity of the basophilic cells in the myometrium with the mononuclear trophoblasts invading the decidua basalis was obvious. Quantitation of the invading cells in different sectors of the placental bed indicated that the deep myometrial invasion begins at the centre of the placental bed, as in the decidua, but progressively expands to the more lateral areas. This invasion pattern is reflected by a quartic distribution on a complete placental bed section, indicating a ring-like spread towards the periphery. The position of the interstitial trophoblast in relation to the spiral arteries shifts from the decidua to the myometrium. In the decidua, the interstitial trophoblasts tend to cluster around the blood vessels; however, in the inner myometrial compartment, they become more dispersed. Individual mononuclear cells form clusters and fuse into multinuclear giant cells in the decidua and the myometrium, and their persistence in the latter compartment suggests that giant cells have lost their invasive properties. The trophoblastic nature of the invading mononuclear cells in the myometrium was confirmed in later studies, applying immunohistochemical staining techniques. Further, the progressive clustering and fusion of mononuclear trophoblasts could be immunohistochemically confirmed by the successive appearance and disappearance of cell adhesion molecules between clustering cytotrophoblastic cells.
The exact function of the numerous interstitial trophoblasts in the junctional zone myometrium remains unclear. A possible role in softening the uterine tissue has been suggested by Boyd and Hamilton, whereas Brosens has presented evidence for an endocrine function, according to histochemical localisation studies of steroid-converting enzymes. Another possible function has emerged from a study on early decidua-associated remodelling of myometrial spiral arteries, remodelling of myometrial spiral arteries, showing a correlation between the degree of vascular smooth muscle disorganisation and interstitial trophoblast volume density ( Fig. 1 ). The exact mechanism of this early remodelling has not been clarified, but a lytic effect of trophoblast-secreted matrix-degrading enzymes on vascular matrices is a likely early step in media disorganisation.
Decidua-associated spiral artery remodelling preceding deep endovascular trophoblast invasion
The emerging concept that the decidua acts as an initiator and stimulus to trophoblast invasion is also applicable to spiral arteries. In addition to alterations in the endometrial stroma, decidualisation also involves ‘decidua-associated remodelling’ of the endometrial spiral artery segments, consisting of endothelial swelling and vacuolisation, disorganisation of the vascular smooth muscle and swelling of individual smooth muscle cells. In the decidualising endometrium during the first trimester of pregnancy, these early vascular changes seem to be initiated by infiltrating macrophages and uterine natural killer cells, partly through the secretion of angiogenic growth factors, and also by their capacity to induce apoptosis in a proportion of vascular smooth muscle and endothelial cells. Such drastic vascular disorganisation is likely to result in reduced vasomotor activity, which may contribute to increased maternal blood flow to the placenta, even before the onset of endovascular trophoblast invasion.
This decidua-associated remodelling is extended into the inner (junctional zone) myometrium. The occurrence of perivascular oedema, disorganisation and swelling of vascular smooth muscle cells and thinning of the elastic lamina have been described in early studies of the placental bed ( Figs. 1 and 2 ). Recent imaging technology has confirmed drastic changes in vascular integrity in the inner myometrium, starting at 7 days after ovulation (i.e., around the implantation period). This provides additional support for considering the inner myometrial zone as a separate compartment of the uterine wall. It is still uncertain how far maternal infiltrating cells (mainly macrophages, as uterine natural killer cells are absent in the myometrium) may also act in this compartment as possible triggering agents for this early remodelling, as they do in the endometrium. We have already indicated in the previous section that the disintegration of the media and swelling of individual smooth muscle cells in the myometrial segments is intensified in the presence of the interstitial trophoblast. Further, orcein-staining of the elastica membrane is significantly weaker in inner myometrial areas containing interstitial trophoblast. As these invaded cells have been shown to contain angiogenic growth factors, it seems that they take over part of the activity of uterine natural killer cells that are restricted to the endometrium.
