The development of the equine placenta involves a series of stage-specific events which ensure that the fetus is nourished throughout its 11 months of gestation. Initially, placental exchange to the developing embryo is histotrophic, via the yolk sac but, as the allantochorion develops and microcotyledons form, haemotrophic nutrition plays the major role in sustaining the increasing demands of the growing fetus. This review describes the development of the allantochorionic placenta of the mare and discusses some of the factors that influence its growth, size and functions and, hence, its control of fetal growth and maturation.
It may be reasonably argued that the single biggest factor influencing the development of the equine fetus in utero is the area of placental exchange at the fetomaternal interface (Rossdale 1966; Cottrill et al. 1991; Bracher et al. 1996; Allen et al. 2002a). That is, the total area of normal and healthy contact between the layer of fetal trophoblast covering the branched and sub-branched, highly vascularised, allantochorionic villi and the opposing layer of lumenal epithelium covering the equally convoluted and heavily vascularised maternal endometrium, all neatly arranged into multi-branched microcotyledons that cover almost the entire surface of the diffuse allantochorion and effect gaseous and haemotrophic maternofetal exchange. Add to this the myriad of discrete areolae situated between adjacent microcotyledons and above the mouths of endometrial glands where specialised, pseudostratified nonattached trophoblast cells continue to absorb endometrial gland secretions to provide additional histotrophic nutrition to the conceptus throughout gestation. Intimate and extensive contact and absorptive potential between the maternal and fetal epithelial layers of the placenta is of paramount importance to the growth and development of the equine fetus, and any physical or pathological process that limits this total area of noninvasive placental interchange will have adverse effects upon fetal growth and wellbeing, and is likely to delay its maturation and ‘readiness for birth’ (Rossdale and Silver 1982).
In this paper we examine the morphological changes and hormone- and growth factor- driven influences that lead to the initial development and establishment of the equine placenta, and we review briefly some physiological and noninfective pathological changes that effect placental, and hence fetal, development during equine pregnancy.
Differentiation of the fetal membranes and early placentation
The equine embryo does not enter the uterus until as late as Day 6 after ovulation (Battut et al. 1997), and it does so purely as a consequence of its secretion of significant quantities of prostaglandin E2 (PGE2) from Day 4. This relaxes the circular and contracts the longitudinal musculature of the oviduct (Webber et al. 1995), thereby propelling the embryo past the ampullary-isthmus constriction and into the uterus through the very tight uterotubal papilla (Webber et al. 1991, 1992). Here the trophectoderm of the late morula/early blastocyst secretes a highly glycosylated protein, which is moulded, by the continuing presence of the zona pellucida to form a tough, elastic glycocalyx that completely surrounds the embryo until about Day 21 of pregnancy (Oriol et al. 1993). This so-called ‘capsule’ (Betteridge 1989) maintains the embryo in a spherical configuration and provides it with sufficient toughness and elasticity to withstand the compressive forces of the strong myometrial contractions that propel it throughout the uterine lumen constantly from Day 6 to Day 16 after ovulation (Ginther 1983). These peristaltic myometrial contractions are stimulated by the embryo's continuing release of both PGE2 and PGF2α (Stout and Allen 2001, 2002), and its constant transuterine migration during this period is presumed to be important for the transmission of the embryonic ‘maternal recognition of pregnancy signal’ (Short 1969) that prolongs luteal function for maintenance of the pregnancy state (Ginther 1983; McDowell et al. 1988). Movement ceases abruptly on Day 16/17, when the size of the conceptus restricts its free passage through the uterus (Gastal et al. 1996) and a spasm-like increase in uterine tonicity ‘fixes’ the still-spherical conceptus at the base of one or other of the uterine horns (Griffin and Ginther 1991; Carnevale and Ginther 1992). Here it remains for the next 25 days, undergoing a series of morphological modifications of its component membranes and developing the embryo proper (Fig 1), all sustained by trophoblast uptake of the protein-rich exocrine secretions (histotroph) of the endometrial glands (Fig 2a; Amoroso 1952) but with, as yet, no attempt by the expanding allantochorion to establish a stable attachment to the maternal endometrium and commence the process of placentation. This only occurs from around Day 40 of gestation (Fig 1d; Samuel et al. 1974), some 5 days after the specialised, annulate, chorionic girdle portion of the fetal membranes (Allen and Moor 1972) has invaded the maternal endometrium to form the circle of ulcer-like endometrial cups (Fig 1d; Allen et al. 1973), which secrete considerable quantities of equine chorionic gonadotropin (eCG) into the maternal circulation for the next 60–80 days (Cole and Hart 1930; Allen 1969). A microvillous junction between the apical surfaces of the lumenal epithelium of the endometrium and the noninvasive trophoblast of the allantochorion is established between Days 40 and 43 (Samuel et al. 1974) and, thereafter, during the next 40 days, the allantochorion expands slowly and steadily until it occupies the whole of the interior of the uterus by Days 85–90 (Allen and Wilsher 2009). Simultaneously, blunt villi of allantochorion interdigitate progressively with accommodating upgrowths of endometrium (sulci) over the entire surface of the expanding allantochorion. Extensive capillary networks develop on both maternal and fetal sides of this placental interdigitation (Fig 3; Abd-Elnaeim et al. 2006; Allen et al. 2007), and the allantochorionic villi and their apposed endometrial sulci become increasingly branched and sub-branched to create the complex microcotyledonary structure of the equine diffuse epitheliochorial placenta by about 150 days of gestation. Further development and sub-branching of the microcotyledons continues during the remainder of pregnancy (MacDonald et al. 2000) to maximise the microscopic area of fetomaternal epithelial contact at the interface and thereby accommodate the increasing gaseous and haemotrophic nutritional demands of the rapidly growing fetus. In addition, the specialised areola areas between the microcotyledons continue to imbibe histotroph as an important form of nutrition for the conceptus (Fig 2b).
During the second half of pregnancy, each multi-branched microcotyledon consists of highly convoluted tufts or networks of fetal capillaries covered by an ever-thinning layer of endometrial epithelium covering an equally convoluted network of maternal capillaries (Fig 3), to the extent that the breadth of tissue separating the maternal and fetal capillary networks is reduced to as little as 4–6 µm (Samuel et al. 1975). A significant number of hormonal and mitogenic stimuli must be involved in the formation of this amazing temporary organ which, at term, has developed a microscopic area of contact between maternal and fetal epithelial layers in excess of 40 m2 (Wilsher and Allen 2003).
Factors that may influence placental development and function
A number of noninfectious or ‘pseudophysiological’ factors have been shown to profoundly influence the differentiation, development and function of the placenta in the mare and thereby influence the rate of growth and maturation of the fetal foal.
Maternal body/uterine size
Maternal, and hence uterine, size markedly influences placental size and function and hence fetal growth in utero, by limiting the area of microscopic contact between trophoblast and endometrial epithelium at the placental interface. This was first illustrated dramatically by the classic experiment of Walton and Hammond (Walton & Hammond 1938) in Cambridge, UK, which employed artificial insemination to cross large Shire horses with much smaller Shetland ponies. The foal born from the Shire mare inseminated with Shetland Pony semen was far taller and heavier at birth than the foal from the reciprocal cross, and this great disparity in body size persisted into maturity. Tischner and his colleagues in Krakow, Poland, took the experiment an important stage further when they transferred Konik Pony embryos to the uteri of much larger draught-type recipient mares before re-mating the Konik Pony donors to the same Konik Pony stallion and allowing them to carry their foals to term (Tischner 1985). In all of the 3 sex-matched pairs of full sibling foals produced in this manner, those from the larger recipient mares weighed considerably more and were appreciably taller than their full siblings born from the genetic Konik Pony mothers. Also, as with the Shire x Shetland Pony crosses of Walton and Hammond (1938), these marked disparities to body size persisted into adulthood (Tischner 1987; Tischner and Klimczak 1989).
Similarly, Allen and co-workers in Newmarket, UK, used reciprocal embryo transfer between large Thoroughbred and small Welsh Pony mares to highlight the marked influences of maternal uterine size on placental and fetal growth. Foals resulting from the transfer of small pony embryos to the larger uteri of Thoroughbred recipient mares (‘luxurious’ uterine environment) were born approximately 15% larger (weight and height) than their Pony-in-Pony (P-in-P) controls, whereas Thoroughbred embryos transferred to smaller Pony uteri (‘deprived’ uterine environment) were born some 15% smaller than their Thoroughbred-in-Thoroughbred (Tb-in-Tb) counterparts (Allen et al. 2002a, 2004). These considerable bodyweight differences correlated closely with parallel differences in the weight and gross area of the respective allantochorions and with the total microscopic area of contact at the fetomaternal interface calculated stereologically (Fig 4a; Allen et al. 2002a). Also, as with the previous experiments of Walton and Hammond (1938) and Tischner (1987), differences in size at birth persisted to age 3 years, and although a degree of ‘catch-up growth’ occurred, the in utero‘deprived’ Thoroughbred-in-Pony (Tb-in-P) foals nevertheless remained appreciably smaller (approximately 5%) than their normal Tb-in-Tb controls (Allen et al. 2004).
In addition to changes in the morphology of the placenta, the function of the fetoplacental unit as an endocrine organ was also modified in this experimental model. For example, the Pony mares carrying the nutritionally deprived Thoroughbred foals showed much higher maternal plasma progestagen concentrations during the second half of pregnancy than the other 3 groups (Allen et al. 2002b), presumably as a result of fetal stress initiating premature enlargement and maturation of the fetal adrenal glands, which then secreted higher levels of pregnenolone for conversion to 5α-dihydroprogesterone and other 5α-reduced progestagen metabolites by the placental tissue (Holtan et al. 1991).
Perhaps more importantly, the growth-restricted or growth-enhanced intrauterine existence in the above animals resulted in alterations in cardiovascular function indicated by changes in post natal regulation of blood pressure and the circulating concentrations of cortisol (Giussani et al. 2003). Other metabolic systems were also affected and although no significant differences in insulin secretion were detected insulin sensitivity or glucose tolerance was discernible between the Tb-in-Tb vs. the Tb-in-P foals. The overgrown P-in-Tb foals showed higher basal insulin concentrations, greater beta cell responses to glucose and a steeper glucose-to-insulin relationship than the other groups (Forhead et al. 2004).
Following the elegant experiments carried out by Wallace et al. (1996, 1999, 2001) which demonstrated the profound influence of variations in maternal nutrition on placental, and hence fetal development in adolescent sheep, the current authors undertook a similar experiment in mares, which aimed to determine the effect of maternal nutrition on placental and fetal development in young primiparous Thoroughbred fillies that rapidly changed their body condition from a fit racehorse to that of a broodmare. Hence, 2 groups of maiden Tb fillies carrying half-sibling Tb foals were fed maintenance (n = 9) or ad libitum (2.5–3x maintenance diet; n = 11) throughout gestation. An unfortunate twist was added to the experiment when all the pregnant fillies suffered an unforeseen outbreak of Streptococcus equi infection (‘strangles’) between 90 and 130 days of gestation, which resulted in pyrexia, depression and inappetance for 7–10 days. The mares lost varying percentages of their bodyweight (0–19.5%) as a result of this episode and it was this degree of weight loss during the period of infection rather than the overall level of nutrition throughout gestation that most strongly influenced placental and fetal development at term. Hence, mares that suffered the larger percentage losses of bodyweight in mid-gestation had smaller allantochorions in terms of mass, volume and total microscopic surface area at term (Fig 4b; Wilsher and Allen 2006). However, a compensatory mechanism was observed for the smaller placentae in that placental efficiency (foal bodyweight/m2 microscopic surface area of the allantochorion) was higher in the mares that lost more weight and hence produced smaller placentae. Nevertheless, despite this adaptation, the foals born from all these infected mares were significantly lighter than those born from equivalent noninfected mares (44.9 ± 0.9 vs. 47.3 ± 0.8 kg, respectively). Other fetal growth parameters at birth were also modulated by the nutritional insult, including crown–rump length and ponderal index. Furthermore, foal plasma IGF-I levels 1 h post partum were negatively correlated with maternal weight loss (Wilsher and Allen 2006), perhaps indicating differences in fetal liver function, since maternal nutrient restriction in early-to-mid-gestation in sheep reduces IGF-I mRNA abundance in the fetal liver (Brameld et al. 2000).
Manipulation of dietary intake during pregnancy to optimise fetal growth and development is clearly important, particularly in the horse, which is primarily destined to be an athlete. Although many authors have looked at the influence of maternal weight gain or loss during pregnancy on foal birthweight (see (Hintz 1993)), placental development per se has not been assessed. Considering the ready ability to modulate placental and fetal growth in sheep through maternal diet (Robinson et al. 1999), it is surprising that more research in this area in the horse has not yet been undertaken.
Endometrial degeneration (endometrosis)
The age-related degenerative changes that occur in the endometrium of mares (Kenny 1978; Ricketts and Alonso 1991) are nowadays termed endometrosis (Kenny 1993). They include untoward accumulations of leucocytes in the endometrial stroma, erosions of the luminal epithelium and deposition of excess fibrous tissue in the stroma, which causes blockage of the lymphatic drainage channels to form lymphatic lacunae in the stroma and lymph-filled cysts that protrude into the uterine lumen. But perhaps most significantly, fibrous tissue is laid down in concentric layers particularly around the fundic portions of the endometrial glands, ‘walling them off’ and thereby obliterating their exocrine secretory functions.
The typical changes wrought by endometrosis affect fetal and placental development in 2 ways. First, the reduced or blocked secretion of histotroph and growth factors by the degenerate endometrial glands (Gerstenberg et al. 1999; Stewart et al. 2000; Ellenberger et al. 2008; Hoffmann et al. 2009) deprives the young embryo, and probably also to varying degrees the later-stage fetus, of essential nutrients, thereby leading to reduced growth rates and fetal stress. Second, fibrous deposition in the endometrial stroma greatly reduces the ability of the endometrium to interdigitate closely and extensively with the allantochorion after Day 40 of gestation, also causing nutritional deprivation owing to the reduced total area of contact between the fetal and maternal epithelial layers at the placental interface.
These adverse effects were well illustrated by Bracher et al. (1996) when using scanning electron microscopy to compare the appearance of the maternal surface of the allantochorion at 120 days of gestation in 2 Thoroughbred mares mated to the same stallion, one young mare with a normal, healthy endometrium and the other an aged mare with severe endometrosis. The difference between the 2 placentae was stark, with the abnormal structure and great reduction in density of the microcotyledons in the aged mare being reflected by a 2-fold reduction in bodyweight of her fetus compared to that of the younger mare at the same stage of gestation; had her pregnancy been left undisturbed, she would probably have aborted before term as a result of fetal starvation.
Further elucidation of the deleterious effects of age and parity on the formation of the microcotyledons was provided by Wilsher and Allen (2003) who used stereological techniques to show that the plexiform nature of the microcotyledons (surface area/unit volume; Sv) is significantly reduced as parous mares become older. The same study also revealed that primiparous mares had significantly lower Sv values than young (5–9 years old) parous mares, despite their virginal endometrium, indicating that optimum development of the microcotyledons requires the ‘priming’ influence of a previous pregnancy.
In twin pregnancies the otherwise healthy placentae of the 2 conceptuses develop and interdigitate with the endometrium normally in early pregnancy but, as gestation advances, the 2 allantochorions compete for the limited area of endometrium with which they can develop useful microcotyledonary placental exchange. One conceptus inevitably overgrows the other to the point, usually around 7–9 months of gestation, when the area of attached allantochorion of the disadvantaged twin is insufficient to maintain its increasing nutritional demands and it literally starves to death, thereby initiating placental separation and abortion of both conceptuses (Fig 5; Jeffcott and Whitwell 1973).
However, owing to the judicious use of the ultrasound scanner in the veterinary management of mares at stud, the incidence of twin abortions has reduced sharply (Smith et al. 2003) as a result of manual reduction of twin conceptuses prior to Day 16 (Roberts 1982; McKinnon 2007). Previous reports had suggested that this course of action might result in a higher incidence of one type of vascular pattern on the allantochorion (K. Whitwell, personal communication; Rossdale and Ricketts 2002) and, although other authors have found no such association (Whitehead et al. 2005), it remains possible that fetal and placental development can both be influenced by events occurring prior to formation of the allantochorion (Wilsher et al. 2010). When twin reduction is undertaken later in gestation, the surviving twin may still be undersized at birth because its allantochorion was unable to initiate optimal villus formation over the area of endometrium previously occupied by its co-twin (Ball et al. 1993; Govaere et al. 2008). This suggests strongly that in order to develop normally in midgestation the allantochorion must have been exposed to stage-specific events much earlier in gestation.
The effects on placental and hence fetal development in equine pregnancy of a deficiency of 2 specific hormones of pregnancy, eCG secreted by the fetal endometrial cups in early gestation and phenolic and Ring-B unsaturated oestrogens secreted by the allantochorion in later pregnancy, are worthy of mention. Deficiencies of both hormones have been created experimentally.
Ablation of the endometrial cup reaction and a consequential absence of eCG from maternal blood was achieved by transferring donkey embryos (Equus asinus, 2n = 62) into horse mares (E. caballus, 2n = 64). In this xenogeneic D-in-H situation, but not in the reverse extraspecific pregnancy of horse embryos transferred to jennies (H-in-D), development of the donkey progenitor chorionic girdle was so retarded in the horse uterus that the girdle failed completely to invade the maternal endometrium during Days 36–38 of gestation. The conceptus nevertheless continued to develop normally to around Day 60, beyond which a clear division in outcome of the pregnancies occurred. In around 70% of the D-in-H pregnancies, the donkey allantochorion lay closely apposed to, but made no effort to interdigitate with, the surrogate horse endometrium (Fig 6). The unattached allantochorion degenerated and was actively attacked by maternal immune cells, which also accumulated in the underlying endometrial stroma (Allen 1982). In the remaining 30% of D-in-H pregnancies, however, implantation and placentation, albeit delayed and rather irregular, did occur during Days 40–100 of gestation, despite the absence of gonadotrophin and reduced concentrations of ovarian progestagens in maternal blood. Thereafter the pregnancies continued to term or thereabouts with the birth of live foals that ranged from well-grown and healthy to dysmature, small and barely viable as a result of varying degrees of inadequate placentation.
The second model of hormone deficiency in equine pregnancy involved the ablation of placental oestrogen secretion during the second half of gestation by performing bilateral fetal gonadectomy in 4 mares between 197 and 253 days of gestation, thereby removing the source of C-19 oestrogen precursors. All the gonadectomised fetuses continued to term despite the precipitous fall in maternal serum oestrogen concentrations following removal of their gonads, and they underwent spontaneous parturition at the expected time. However, parturition was characterised by weak myometrial contractions and an absence of the sharp rise in maternal serum prostaglandin F metabolite concentrations that normally accompanies foaling (Pashen and Allen 1979). The foals were mature, albeit severely lacking in body musculature, and the one foal that lived to maturity remained small, long-coated and unthrifty throughout her existence.
Although the authors did not ascertain whether fetal gonadectomy had induced any observable alterations in the placental structure, the severe disruption of oestrogen production in later gestation, although not abortifacient per se, clearly impeded the normal growth and development of the fetus thereafter, probably as a result of reduced vasculogenesis in the placenta. The absence of late-gestation oestrogens also interfered with the normal synthesis and storage of prostaglandins in the myometrium, thereby leading to abnormal parturition.
It is incredible that the tiny horse zygote, measuring only 180 µm in diameter at 12–18 h after fertilisation, has the inbuilt genetic ability over the 11 months of its intrauterine existence to drive the construction of a large, diffuse allantochorion which, in a Thoroughbred mare, weighs as much as 4 kg at term and is attached to the maternal endometrium over an area in excess of 40 m2 (Wilsher and Allen 2003). Equally astounding is the realisation that a great deal of the nutrient energy that powers this extensive building project derives as much from the exocrine secretions of the endometrial glands (histotroph) as from the haemotrophic exchange with the maternal blood supply. This is certainly the case during the first 40 days of gestation before stable microvillous attachment of the trophoblast to the endometrial epithelium occurs, and there is every reason to suppose that the close interdependency of allantochorionic development with an adequate supply of histotroph continues throughout pregnancy (Bracher et al. 1996; Ellenberger et al. 2008).
Hormones and growth factors are clearly important in the overall ‘construction project’ of equine pregnancy, with the fetal trophoblast supplying the steroid (oestrogens and progestagens) and gonadotrophic (eCG) hormones that play their important roles in ensuring the ‘health and safety’ aspects of the fetus's intrauterine residence. The maternal endometrial glands, with some additional help from the trophoblast, are especially responsible for producing the mitogenic growth factors to drive placental and thereby fetal growth and maturation throughout the gestation period. The extensive network of tightly coiled and multifunctional endometrial glands that exist throughout the uterus produce both the wide array of growth factors that stimulate development of the placenta and the copious quantities of histotroph that feed it. They are therefore pivotal as the essential bedrock foundation of the whole edifice of equine pregnancy.
Conflicts of interest
No conflicts of interest have been declared.
Sources of funding
Studies described in this work were kindly funded by The Horserace Betting Levy Board, The Havemeyer Foundation, New York, The New Hope Fertility Clinic, New York and Collingwood Neptune Ltd.