Placental function in development and disease

The placenta is the first of the fetal organs to develop and has several fascinating and critical functions. It mediates implantation and establishes the interface for nutrient and gas exchange between the maternal and fetal circulation, as well as initiating maternal recognition of pregnancy, altering the local immune environment and altering maternal cardiovascular and metabolic functions through the production of paracrine and endocrine hormones. Abnormalities in any one of these functions can be associated with poor pregnancy outcome, ranging from the mild (intrauterine growth restriction; IUGR) to the severe (implantation failure and embryonic, fetal or perinatal death). The present review describes recent insights into how placental development and function are regulated and the potential role for primary placental pathologies in explaining a variety of complications in human and animal pregnancies.

The gross anatomy of the placenta varies considerably among mammalian species (Perry 1981; Leiser and Kaufmann 1994; Wooding and Flint 1994). Given these differences, people often hesitate to extrapolate information from one species to another. This is unfortunate because, when one looks at the core functional cell types, several features are conserved across species as diverse as humans, rodents, equids and ruminants (Cross et al. 2003). At the most basic level, the fetal placenta is composed of an outer epithelial layer derived from the trophectoderm (trophoblast) and an inner vascular and stromal layer derived from the allantois (extra-embryonic mesoderm). The trophoblast layer goes on to subspecialise and, in general, fulfils two functions in every mammalian species. First, it generates the extensive surface area for nutrient exchange and provides the barrier between the maternal and fetal circulation. Second, trophoblast cells interact closely with the uterus and produce hormones that target maternal physiological systems and promote maternal blood flow to the implantation site.

In most mammals, the trophoblast layer becomes organised into highly branched villous structures, although how the villi are organised varies considerably (Wooding and Flint 1994). In pigs, the villi are spread diffusely across the placenta. In other species, they are clustered either into a band (horse) or into cotyledons. Rodents have a single cotyledon, ruminants (sheep, cow) have multiple cotyledons scattered across the placental surface and humans have multiple cotyledons that are grouped together resembling a single disc. In rodents and primates, the uterine epithelium is eroded such that maternal blood comes into direct contact with the trophoblast surface of the villi (haemochorial). Rodents have three trophoblast cell layers, including two syncytial layers and a single mononuclear cell type, whereas primates have a single syncytial layer plus an underlying mononuclear trophoblast cell layer. In contrast with these invasive species, the endometrium remains intact (epitheliochorial) in pigs, ruminants and horses. The invasive/endocrine trophoblast cell subtype varies considerably in the extent of uterine invasion. Pig trophoblast cells do not invade at all and ruminant binucleate cells migrate essentially only the distance of one cell diameter to enter the endometrium (Wooding and Flint 1994). In contrast, rodent trophoblast giant cells and glycogen trophoblast cells migrate several hundred microns into the uterus in association with spiral arteries and diffusely into the interstitium of the decidua, respectively (Adamson et al. 2002). Extravillous cytrophoblast cells in primates also invade in these two patterns, but much more deeply, and, indeed, normally as far as the myometrial cell layer of the uterus (Pijnenborg 1996).

The function of the placenta during early gestation is primarily to mediate implantation of the embryo into the uterus and, secondarily, to produce hormones that induce maternal recognition of pregnancy, the term used to describe the means by which the lifespan of the corpus luteum is extended, preventing the end of the ovarian cycle. The specific hormones used to induce maternal recognition of pregnancy vary among species. Pig embryos secrete oestrogen (Flint et al. 1978), whereas ruminant embryos produce an interferon (Spencer et al. 2004) to suppress uterine production of the luteolytic factor prostaglandin F_2α. Primate (Hearn et al. 1991) and horse (Allen 1978) embryos produce a chorionic gonadotrophin that directly stimulates luteal progesterone production. In rodents, mating alone produces a prolonged luteal phase (pseudopregnancy) and then placental lactogens, which are produced in the second half of gestation, have luteotrophic effects (Forsyth 1994).

After implantation, the major function of the placenta is to mediate, as well as to regulate, nutrient uptake from the mother to the fetus. It does so by forming the highly branched villous structure and establishing its own circulation connected to the umbilical cord (Watson and Cross 2005). Different nutrients transfer either passively or by active transporters (Hay 1994; Reik et al. 2003; Watson and Cross 2005). The placenta is also able to accumulate glycogen and this is most marked in rodents, in cells called glycogen trophoblast cells (Simmons and Cross 2005). In mice, glycogen accumulation is limited to the latter half of gestation and specifically occurs in a subset of cells in the spongiotrophoblast layer after embryonic day 12.5 that later invade the wall of the uterus (Adamson et al. 2002). Glycogen accumulation is increased in diabetic pregnancies (Barash et al. 1985; Padmanabhan and Shafiullah 2001), but is stimulated by insulin (Goltzsch et al. 1987). In contrast, glycogen content is reduced in insulin-like growth factor (IGF)-II mutant placentas (Lopez et al. 1996), suggesting that IGF-II may act in a paracrine manner within the placenta to regulate glucose uptake and the synthesis of glycogen, perhaps by acting through the insulin receptor. Glycogen deposits are a common pathological finding in different tissues, but because placental accumulation occurs in all pregnancies and it appears to be highly regulated, it may have a normal physiological function.

In addition to having these functions in nutrient uptake, the placenta also plays an active role in regulating maternal physiology to the nutritional benefit of the fetus. First, the invasive trophoblast population produces angiogenic factors and vasodilators and, in some species, actually invades the arteries also to increase blood flow to the implantation site (Cross et al. 2002). Second, the placenta has been shown, in several species, to produce hormones that stimulate the pregnancy associated increase in maternal blood cell production and blood volume (Conrad et al. 1996; Kim et al. 2001; Klemcke et al. 2001; Zhou et al. 2002, 2005). Third, the placenta produces several metabolic hormones, such as placental lactogens and placental growth hormone, that can alter insulin production and promote insulin resistance in maternal tissues, resulting in increased glucose availability to the fetus (Soares et al. 1998; Linzer and Fisher 1999). The placenta also produces leptin (Pelleymounter et al. 1995) and ghrelin (Tschop et al. 2000), hormones that suppress and stimulate appetite, respectively.

The placenta is derived from two major cell types. The trophoblast cell lineage originates from the trophectoderm at the blastocyst stage. The stromal and vascular components of the placenta are derived from the allantois, which grows out from the embryo proper to make contact with the overlying trophoblast layer. Villous development initiates only at points of chorioallantoic attachment and this process is particularly vulnerable to perturbation by genetic (Watson and Cross 2005), epigenetic (Spindle et al. 1996) or nutritional (Wallace et al. 1999) influences. The cellular and molecular mechanisms underlying the development of the placenta are best understood in the mouse as a result of studies involving experimental embryology, culture trophoblast stem cells and analysis of mouse mutants that have defects in placental development. We now know of more than 100 genes that are critical for placental development and their functions have been reviewed in detail elsewhere (Hemberger and Cross 2001; Rossant and Cross 2001; Simmons and Cross 2005; Watson and Cross 2005). In addition to cell intrinsic mechanisms, placental development is also highly regulated by oxygen levels (Genbacev et al. 1997; Adelman et al. 2000; Caniggia et al. 2000) and the availability of nutrients in the maternal circulation (Cross and Mickelson 2005). This means that the basic pattern of placental development can be altered by environmental influences and, moreover, that placental development can attempt to compensate for other defects. A good example is the placental phenotype in Rb mouse mutants. The mutant placenta shows poor villous development, but the villi that do form are hypervascularised (Wu et al. 2003).

Attempts to clone most animal species using nuclear transfer technology have met with very high rates of embryonic, fetal and perinatal mortality (Sakai et al. 2005). Although the specific pathologies in the fetus vary, changes in the placenta are the most common finding. Initial studies had described only the placenta in those rare pregnancies that made it to term. In cattle and sheep, the term placenta in clones is characterised by a reduced number of cotyledons, hydroallantois and hypertrophy of remaining placentomes (Hill et al. 1999; De Sousa et al. 2001). Interestingly, the maternal caruncles are also underdeveloped, implying that primary fetoplacental defects indirectly affect the mother also (Hashizume et al. 2002). In more recent studies, examination of first trimester, cloned conceptuses showed that failure to establish cotyledons is common and likely accounts for the large number of conceptuses that fail during the first trimester (Hill et al. 2000; Hashizume et al. 2002). Presumably only those conceptuses that have a threshold minimum number of cotyledons are able to survive and make it to term. Because cotyledonary size can undergo a compensatory increase if the number of placentomes is reduced by carunclectomy (Anthony et al. 2003), the hypertrophy observed in placentas from clones is likely to be simply a secondary change. In mice, the term placenta in clones is enlarged owing to expansion of the spongiotrophoblast layer, but whether this is a primary or secondary change is unclear (Tanaka et al. 2001; Ogura et al. 2002; Singh et al. 2004).

The common finding of placental defects in cloned mice, cows and sheep suggests placental pathology as a major cause of pregnancy complications. This hypothesis has been supported by chimera studies in which tetraploid blastomeres are mixed with preimplantation cloned embryos. In these chimeras, the tetraploid cells can contribute to trophoblast derivatives, but not the embryos/fetus proper (Cross 2001). Tetraploid chimeras in both mice (Eggan et al. 2001) and cattle (Iwasaki et al. 2000) do better than regular clones, implying that trophoblast lineage defects in clones account for at least some of the pregnancy complications.

Placental pathology has long been associated with IUGR and pre-eclampia in humans. However, it is important to note that the specific changes that are observed can be rather variable. The IUGR placenta most frequently has lesions within the villi, including inflammation and infarcts (Salafia et al. 1992, 1995). In severe, early onset disease, there is often a significant reduction in villous surface area (Macara et al. 1996), perhaps suggesting primary defects in villous formation. Pre-eclampsia is fundamentally a maternal disease characterised by hypertension, renal glomerulosclerosis leading to proteinuria, uricaemia and, in severe cases, seizures. The placenta in pre-eclampsia usually has a normal weight and surface area (Teasdale 1985). However, there is marked proliferation of villous cytotrophoblast cells (Fox 1964; Jones and Fox 1980; Arnholdt et al. 1991; Redline and Patterson 1995) and the syncytiotrophoblast shows focal necrosis (Jones and Fox 1980). In addition, trophoblast invasion of the uterine spiral arteries and their remodelling of the spiral arteries is frequently reduced (Brosens et al. 1972; Gerretsen et al. 1981; Khong et al. 1986). Intrauterine growth restriction is sometimes, although not always, observed in pre-eclampsia. When it is, the surface area of placental villi is reduced and is associated with a decrease in umbilical blood flow (Jackson et al. 1995). The underlying cause of pre-eclampsia is a matter of considerable debate, but recent studies in mice suggest that primary fetoplacental lesions are sufficient to initiate the disease (Cross 2003).

Studies in rats, sheep and guinea-pigs have shown that both feed restriction and overfeeding affect development of the placental villi. In rats, protein restriction (8% v. 20% dietary protein) from the time of mating leads to an increase in villous surface area without affecting overall placental volume, implying more extensive branching of the villi (Doherty et al. 2003). This implies that trophoblast cells can alter their development most likely in an attempt to compensate for the protein restriction, although the underlying mechanism is not clear. Importantly, however, the underlying vasculature volume does not undergo a proportional increase. The pups are mildly growth restricted (approximately 10% smaller by term), most likely because, in the absence of a proportional increase in vascularisation, the uptake efficiency of the placenta may not have improved despite the larger surface area available. Another model involves maternal iron restriction before and during pregnancy, which leads to anaemia. This perturbation also produces placentas with larger villous surface areas, indicating that trophoblast differentiation can be regulated by tissue oxygen levels (Lewis et al. 2001). This finding is consistent with observations in knockout mice lacking hypoxia-inducible factor (HIF; Adelman et al. 2000). Curiously, despite the fact that angiogenesis in most vascular beds is thought to be regulated by oxygen levels, maternal iron restriction in the rat does not result in a proportional increase in fetoplacental vascular volume (Lewis et al. 2001).

In sheep, both feed restriction and overfeeding can lead to IUGR (Wallace et al. 1999, 2004; Redmer et al. 2004; Reynolds et al. 2005). Cotyledon number is most affected by overnourishment during the first trimester when the cotyledons are being formed, whereas cotyledon size is most affected by nutritional status in the second and third trimesters (Wallace et al. 1999). Overfeeding during the first and second trimesters results in reduced trophoblast proliferation (Wallace et al. 2004) and expression of angiogenic factors (Redmer et al. 2004; Reynolds et al. 2005) and is associated with smaller cotyledons that are poorly vascularised. Any one of these manipulations is associated with IUGR. Collectively, these data indicate that, in rats and sheep, at least, nutrient status has major effects on the development of the villous structures in the placenta. The placenta can alter its development in an attempt to compensate but, beyond a certain point, the placenta can no longer compensate for altered nutritional status and IUGR results.

The effect of nutritional status on placental development has also been examined in guinea-pigs, an animal whose placenta has considerable structural similarity to that of humans. The most common guinea-pig model involves severe total feed intake, in which the pregnant dams are fed only 50–70% of the normal ad libidum intake, resulting in birthweights that are reduced by approximately 30%. By the end of gestation, the villous layer of the placenta (the labyrinth) is reduced in overall volume and surface area (Dwyer et al. 1992; Roberts et al. 2001a). Therefore, there is no obvious attempt of the guinea-pig placenta to try and compensate for undernutrition. At a mechanistic level, expression of IGF-I and -II, and the ratio of the IGFs to IGF binding proteins, is reduced in response to feed restriction (Sohlstrom et al. 1998; Roberts et al. 2001b, 2002; Olausson and Sohlstrom 2003). Insulin-like growth factor levels are correlated with the extent of villous branching and are inversely correlated with thickness of the trophoblast barrier (Roberts et al. 2001b, 2002), similar to changes seen in IGF-II-deficient mice (Constancia et al. 2002; Sibley et al. 2004). The response of the guinea-pig placenta to overall nutrient restriction is obviously rather different than that of the rat and sheep in response to protein and energy restriction, as discussed above. It is difficult to explain the differences at present. They could reflect a species difference or a difference in the effects of protein compared with total diet restriction.

We have learned, in the past few years, that the development of the placenta is highly regulated and is therefore susceptible to perturbation. Several genetic, epigenetic, nutritional and environmental factors have been identified and are associated with complications in pregnancy outcome ranging from embryonic lethality to fetal growth restriction. However, this work is rather underdeveloped and much more work needs to be performed to bridge insights from different animal models and humans. In addition, because the placenta is involved in mediating nutrient uptake as well as regulating maternal metabolic and cardiovascular function, much more work needs to be performed to understand how an abnormal placenta contributes to pregnancy complications common to humans and animals. Finally, it is clear that clinically useful markers of placental pathology are required in order to provide early diagnosis of pregnancy complications. The challenge to the field is that although many, many markers have been advocated over the years, there are still no clear-cut specific tests of placental function.