Get access to all handy features included in the IVIS website
- Get unlimited access to books, proceedings and journals.
- Get access to a global catalogue of meetings, on-site and online courses, webinars and educational videos.
- Bookmark your favorite articles in My Library for future reading.
- Save future meetings and courses in My Calendar and My e-Learning.
- Ask authors questions and read what others have to say.
Domestic Horse (Equus caballus)
Get access to all handy features included in the IVIS website
- Get unlimited access to books, proceedings and journals.
- Get access to a global catalogue of meetings, on-site and online courses, webinars and educational videos.
- Bookmark your favorite articles in My Library for future reading.
- Save future meetings and courses in My Calendar and My e-Learning.
- Ask authors questions and read what others have to say.
1) General Zoological Data
The ancestry of horses from five-toed species via the small Eohippus of the American continent has been a prime topic of study for evolutionary biologists. It is best traced in two contributions: by Simpson (1951), and by Thenius & Hofer (1960). Coming from the American continent, the equid ancestors distributed to Eurasia while speciating into the extant members during their expansion to the West and into Africa. In the process, chromosomal fusions occurred in the equidae, causing karyotypes to evolve that range from 66 to 32 chromosomes, whilst preserving the same amount of DNA. The current domestic horse is thought to have derived from the Mongolian wild horse, Equus przewalskii. Other authors, however, have considered the extinct Tarpan to be the ancestral species of the domestic horses. A very large number of different horse breeds have been produced during their domestication, ranging from large tract animals to very small ponies.
Somewhat different views of the speciation of equidae are expressed in various chapters of the new book by Bowling & Ruvinsky (2000) that need to be consulted for details.
Norwegian Fjord Pony with shorn mane.
Przewalski's horse stallion with "standing mane" and "butterfly" belly, perhaps the ancestor of the modern domestic horse.
2) General Gestational Data
There is an abundance of gestational data on horse reproduction. They are summarized in many books and chapters, e.g. that by Samper (2000). Additionally, some information on equine placentas is contained in my chapters on Grevy's zebra and the Kiang. One of the most comprehensive books on the horse biology, however, is that edited by McKinnon & Voss (1992). It contains every aspect of equine reproduction and its ramifications. For specific details, I recommend the reader consult this large volume.
Gestation in horses is seasonal and lasts around 330 to 340 days. The length apparently depends largely on the specific breed. Because there is such enormous difference in the sizes and weights of these animals from different breeds, no data are given here of these features. The interested reader may find the extremes in the data supplied from the most interesting embryo transfer experiments by Allen et al. (2002a,b). One author (Caslick, 1932) found an average placental weight of 5,700g, while Prickett (1970) gives weights from 4,500 to 8,200 g. More detailed information on reproductive phenomena may be found also in chapters contained in the book by Bowles & Ruvinsky (2000).
Implantation and development of the placenta in the horse has probably been more extensively studied than that of most other domestic animal species. A really comprehensive description came originally from Amoroso (1961), with many other authors filling in finer details. Steven (1982) has written another most easily comprehended description of the horse placenta; it also reviews the long history preceding his detailed writings. Moreover, this description is accompanied by excellent drawings. The most recent comprehensive discussion on equine placentation is found by Wooding & Flint (1994). The same volume also contains abundant endocrine and fine structural information. Mossman (1987) has a surprisingly superficial description of equid membranes and it differs frequently from that found by other students. Furthermore, his drawing does not conform to what is now known of the equine placentation. Specifically, the site of cord insertion is inaccurate. Ramsey (1982), on the other hand, has a very concise chapter on horse placentation that is worth reading for initial orientation. The most recent comprehensive details of horse placentation come from writings by Allen & Stewart (2001) whose laboratory has perhaps contributed the most information to our understanding of this interesting organ.
The equine gestation is unusual in many ways. First of all, it has this unusually long preimplantation development of the massively enlarging blastocyst; it is suspended for days in a large amount of "uterine milk" (glandular secretion) that supports the trophoblast. This early spherical embryo is also unusual in that it possesses a hard secondary membrane beneath the zona pellucida (Betteridge et al., 1982) that lasts to about day 25. The conceptus is being shuttled back and forth in the uterine horns for at least a week before it finally becomes "fixed", usually in one horn, on day 15 to 17 (see Ginther, 1985). Here, the embryo comes to lie at an antimesometrial site and is contained in a recess that Enders & Liu (1991a) referred to as the "implantation chamber". Occasional "body implantation", in the middle of the uterus, occurs. The placenta ultimately extends into both uterine horns and is studded on its external surface by villi in the form of "microcotyledons". This term has become used customarily; it may be confusing to the novice, as the placenta looks externally, at least superficially, as though it is covered by a myriad of small villi. The villous projections interdigitate with the maternal endometrial folds and glands in a very complex manner. This, again, is diagrammed beautifully by Steven (1982). In contrast to the more uniform villous projections found in pig placentas, villous stems are produced in the horse ("microcotyledons" - 1 to 2 mm in size) that branch two or three times and which are separated from one another by intervillous stretches of "areolae or arcades or areolae". Vascularization of the endometrium in these areas also differs. The first microcotyledons can be recognized by day 61. The relation to maternal epithelium is epithelio-chorial without invasion in all these places. It is well shown in my chapter on Grevy zebra where an implanted placenta was available for study. It is generally similar to that of horses. Most gaseous and water exchange between mare and fetus occurs through the microcotyledonary trophoblast. Iron, and several proteins and some other substances are transferred by endometrial secretions in the areolae. Therefore, these regions also have a significantly different histologic appearance. This has been beautifully studied with histochemistry and fine structural details by Wooding et al. (2000).
The only region of invasion that occurs in equine placentas takes place at the "chorionic girdle" with the formation of "endometrial cups", now shown to be the result of trophoblast modification. Their history and development have been extensively recapped by Allen (1982). The girdle is located at the interface between the expanding allantochorion and the regressing yolk sac, around the future site of insertion of the umbilical cord and towards the lower aspect of the pregnant horn. At that time, the placenta fills only one of the uterine horns, later it expands into both. These regions are the site of eCG (formerly PMSG) hormone production and they are temporary features. Enders & Liu (1991b) studied this process electronmicroscopically and later referred to the region as producing endocrine and exocrine secretions (a yellow viscid fluid) (Enders et al., 2000). At term, however, one is still able to see the remnants of this girdle around the insertion site of the cord as a thin band. The girdle is mature on day 35 and it's then binucleate, epithelioid trophoblastic cells begin to invade though the endometrial epithelium into the endometrial stroma. Here they engender a marked lymphocytic-plasmacellular and histiocytic reaction. This inflammatory reaction, however, is primarily expansive towards the end of the production phase of eCG. That the hormone production is located here was shown by extraction, by cytochemical reactions and was also proven from extraction of tissue-cultured cup cells. eCG persists for variable lengths. By day 120-130 the cups have involuted and then undergo necrosis. Thereafter, the necrotic debris may be present in small sacs or pouches that often project into the allantochorion. The pouches contain necrotic debris and must not be confused with hippomanes (Mossman, 1987). Much speculation is had as to the nature of the regression; is it "rejection" or are processes of apoptosis in play? Other processes have been invoked as well (lipid degeneration) to explain the means by which the reduced vascular supply leads to the degeneration (Enders et al., 2000). The fine structural study by Wooding et al. (2001) localized the eCG production to the heavily endowed cytoplasm of the cup cells and suggested that no hormonal stimulus causes the "constitutive" secretion of these unique cells.
Much interest has focused on the cup formation of different equines (smaller in donkeys) and the corresponding amount of eCG production. Not only has this been studied in various equidae, it has been examined especially well in embryo transfers and hybrids (see Allen, 1982; Allen et al., chapter 63 in McKinnon & Voss, 1993). As the girdle is large in horses than in donkeys, it was interpreted to be under genetic control of the paternal genotype. More recent studies of appropriate embryo transfer experiments carried out by Allen et al. (2002c) have been reported so far in abstract form only. They indicated an overriding influence of the maternal endometrial genotype in determining the size and development of the cups. Their final publication is awaited with great interest. Other factors now need also consideration, as the epigenetic aspects of imprinting that make reciprocal hybrids develop such markedly different phenotypes (Pennisi, 2001). The microscopic changes occurring at the girdle were well-studied in the fine structural analysis conducted by Enders et al. (1990). Much investigation has also focused on the relationship of this invasive trophoblast and the apparently immunologic reaction to the paternal MHC antigen composition, especially in the donkey-in-horse model (see Allen, 1982; Baker et al., 2001). The topic is still too unresolved and much too complex for a detailed consideration in this short review. It needs to be read in the original publications.
Although late in development, the allantoic sac becomes very large and contains fetal urine and hippomanes as characteristic components. A yolk sac develops early, is large at first but then involutes at the end of the first month of gestation. Its remnants can always be found in a much atrophied state within the insertion of the umbilical cord at term.
At the College of Agriculture in Lexington , Kentucky , a comprehensive workshop was held in 2003 on "The Equine Placenta" (Powell, 2004). The superb book that derived from this workshop is richly illustrated and can be recommended for many further details.
4) General Characterization of the Placenta
After all the mobility of the early conceptus and the very slow elongation of the large round conceptus, at last, by day 100, the allantochorion has filled both uterine horns. The outside is now studded with microcotyledons and villi that originate from the allantochorionic membrane. Next in the sequence of layers come the large allantoic blood vessels within the exocoelomic space and then, on the inside, is the allantoic cavity. The delivered placenta is usually quite dark red when it is inverted so it would assume its position in utero. The amnionic cavity more or less floats freely in the large allantoic sac that is filled with fetal urine. The allantoamnion (allantois outside, amnion inside) is also unusually vascularized when compared with other species. One has to assume that some fluid shifts may take place between the two cavities and perhaps that this is enabled by this extensive vasculature.
General situs of horse placenta towards mid-term.
Fetus in utero at about 3 months gestation. Mare died from "colic". Most of the umbilical vessels at this stage are within the allantois; fetus is in amnion.
Implanted placenta of Przewalski's horse. Mare died from coccidioidomycosis, fetus weighed 23,000 g.
Low power view of horse placenta with heavily vascularized allantochorion and arborized villi. The allantoic blood vessels lie in remnants of the exocoelomic space.
Horse placenta with vascular allantochorion above. This is one of the "microcotyledons".
The amount of vasculature of the allantoic membrane is remarkably irregularly distributed, here it is very rich.
Table 1. Weights and measurements of various equine placentas available
30.5 to 137.2 cm
4,500 - 8,200 g
36 to 84 cm
90 (30+60 cm)
23,000 g (demise)
40 cm (24+16 cm)
115 cm (53+62 cm)
49 cm (21+28)
4,840 g (in utero)
Somali wild ass
15,700 g (in utero)
11,475 g (stb.)
2,000 g (twin)
"Nipple" on amnion
(2,250 g actual placental weight, rest is membranes)
5) Details of fetal/maternal barrier
The surface of microcotyledons and their villous branches is a single layer of uninuclear trophoblast. Binucleate cells, as seen in ungulate placentas, are absent there; they are seen only at the girdle. There is no invasion of maternal tissue, rather, the trophoblast directly abuts the endometrial epithelium in a complex manner in the villous portions of the placenta. As pregnancy proceeds, the proximity of this exchange area increases and there is then a very close approximation of fetal and maternal capillaries. Only beneath the chorionic surface is the trophoblast more cylindrical and often vacuolated. Here also, some yellow pigmentary inclusions are found. They are iron-stain negative and may, as in ungulates, represent melanin, but other pigments cannot be excluded. We have not found this pigment further discussed in the equine literature except to note that, generally, it is assumed that iron is transported to the fetus with uteroferrin from endometrial secretions (Wooding et al., 2000). This cylindrical structure of the trophoblast is also apparent in between the microcotyledons, the areolae or arcades. It is assumed to have a histotrophic function for absorption of uterine secretions.
There is considerable variation in the finer structure and size of microcotyledons that has been assessed quantitatively by Wilsher & Allen (2003). Their extensive data show convincingly that there is a strong correlation of villous surface and fetal weight; primiparous mares (especially older primipara) had smaller foals, in part or perhaps wholly because of this smaller villous surface area. Various other quantitative data are given in this excellent contribution.
Jones et al. (2000) examined the pattern of glycosylation of the "interhemal barriers" of horse, donkey, and camel. This was similar between horse and donkey but much different in the camel. From this it was suggested that this may be important to prevent interspecific hybrids. Macdonald et al. (2000) studied the villous structures of the horse by scanning electron microscopy and published impressive, conclusive pictures of maternal and fetal aspects of the pony placentas. This suggested to them that the "microcotyledons" are the aggregates of encapsulation of 16 uterine crypt regions. With advancing gestation the villi kept branching until very close to term.
In the more columnar, vacuolated trophoblast beneath the chorion of microcotyledons, often contents of yellow inclusions are found (at arrows).
6) Umbilical cord
The umbilical cord of horses has been described in considerable detail by Whitwell (1975). It has a somewhat longer portion within the amnionic cavity and a shorter segment in the allantoic sac; but that varies. The membrane insertion delineates these two segments. There is a considerable variability in the length of the cord, with measurements from 36 to 84 cm in the thoroughbred foals (mean 55 cm) of her study. Excessive lengths have been considered to be the cause of abortion, of fetal strangulation, and of fetal demise (Whitwell, 1975; Caslick, 1932, others). The ultimate cause for this variability in cord lengths is unknown. Caslick measured lengths from 30.5 to 137.2 cm, with the variability perhaps being more distributed over the amnionic portion.
The length of the umbilical cord is surely determined in early gestation, as abortions with excessive lengths occur not uncommonly. Also, in contrast to many other species, the horse fetus has great mobility in early gestation. Similar variations in length occur in human placentation; and short cords are usually due to lack of fetal motion (Benirschke, 1994; Snider, 1997). One wonders whether mobility is perhaps in part the result of the "breed" of horse and notes that the variability in length of the horse cord is often in the amnionic portion. One can imagine that race horses might be more mobile, perhaps already in utero, than draught horses. Measurements comparing different "breeds" are outstanding but would be of great interest.
There is often considerable twisting of the horse umbilical cord. It has occasionally led to excessive torsion and strangulation, with impressions left on the stillborn foal (Whitwell, 1975; Hong et al.; Giles et al., 1993). Perhaps it is also the cause of some of the allantoic duct dilatations (6%) observed by Whitwell & Jeffcott (1975).
The latter authors found the cord to be inserted at the posterior aspect of the uterus, either in or near midline, or in the pregnant horn. The umbilical cord contains two arteries and one vein, in addition to a widely patent, thin-walled allantoic duct. It is accompanied by strands of smooth muscle fibers. While two major venous branches expand over the allantochorion, they fuse to become a single vein at the passing through the amnionic insertion around the cord. The allantoic portion of the umbilical cord is not a solid cord but a "web of vessels" (Whitwell, 1975). In addition to the three large vessels, numerous smaller blood vessels are also present in the umbilical cord, especially in the vicinity of the allantoic duct. Remnants of the vitelline duct may be present and Whitwell has identified remnants of the yolk sac in all placentas she has studied (personal communication). The cord may be significantly spiraled and, in the amnionic portion, it has many very small foci of sqamous metaplasia on its surface.
Absence of one umbilical artery was seen once in 145 cords with a fetus having renal anomalies. Furthermore, in excessively long umbilical cords and heavily spiraled ones, thrombosis and vascular calcification have been observed. The last kiang placenta observed had four umbilical vessels and no twists at all.
Situs of abdominal insertion of umbilical cord.
Umbilical cord of Onager (Hemionus onager) with allantoic duct at 5 o'clock, accompanied by many small vessels. There was diffuse subsurface acute funisitis.
Cord of term Przewalski's horse (Equus przewalskii) placenta with duct (top - All. D.) at the point of its entry into the allantoic sac. Note the bluish tree-like structure at bottom. It is shown in the next slide.
The surface focus of arborized squamous metaplasia from previous slide.
7) Uteroplacental circulation
Steven (1982) showed an excellent diagram of the vasculature in the uterus that supplies the microcotyledons and the arcades. It has been repeated in many subsequent publications on the horse placenta. He depicted the straight maternal arteries to proceed between the microcotyledons whose branches then enter the endometrium that correspond to the microcotyledons. The veins draining this tissue pursue a central flow from the centers of the microcotyledons. Moreover, towards term, there is more and more "invagination or infiltration" of the fetal capillaries into the trophoblast so that, at the end of gestation, fetal and maternal capillaries come to lie in very close physical proximity. This assures maximal possibility of nutrient/gaseous exchange.
8) Extraplacental membranes
A particularly detailed report on the membranes came from the study by Whitwell & Jeffcott (1975). They found that the majority of placentas are inverted at birth (villi inside), unless (in 24%) the nonpregnant horn was of the same size as that containing the fetus. In that case, the villi were usually external at delivery. These authors always identified a small tubular yolk sac remnant near the cord insertion and found hippomanes in the allantoic sac of all animals. In their experience, five regions of the placental surface lack villous endowment. These are: the regions of tubal origin, the "cervical star", and the area of the former chorionic girdle, where no villi grow.
The cervical star requires some further discussion. It is visible as a region without the red villi and has radiating, bare, white areas from its center. Whitwell (personal communication, 2002) made scanning EM pictures of the region. These showed a complete absence of villi and bare, thick folds of connective tissue. If ascending infection were to take place (the horse has a relatively short cervix and ascending infection is reasonably common), the organisms would have to traverse the region of the cervical star and would then come into very close contact with the allantoamnion. As the amnion expands and the fetus grows with later gestation (than that shown in the diagram above) the allantoamnion is closely pressed to the region of the cervical star, the region that ruptures during spontaneous delivery. Thus, infection of the amnionic sac can be explained reasonably easily.
As is evident to any student studying the very young horse placenta, the large allantoic vessels appear to lie free in a space which Enders & Liu (2000) showed to be the remains of the exocoelom. They also determined the nature of collagen development of the basement membranes in these young specimens.
Placental with fetus showing the "cervical star", site of future rupture of membranes.
Cervical star (white radiations) next to villous tissue (red) [Courtesy of Dr. K. Whitwell, Cambridge].
Region of cervical star in Przewalski's horse, from inside the placenta. The white area is the view of the thick fibrous membrane and endocervical mucus beneath.
Microscopic picture of the junction between endometrial implantation (left) and cervical star over cervix (right) in Przewalski's horse.
Microscopy of "cervical star" in Przewalski's horse. Below is the endocervical mucus, follows dense connective tissue, and above is thinned placental membrane.
The amnionic epithelium is a thin, single layer of cells with frequent areas of squamous metaplasia ("pustules, plaques") whose cells contain much glycogen.
The hippomanes are yellowish/green pasty concentration products of fetal urine and contain large numbers of birefringent crystals (some of which are oxalates). The material is acellular and probably mostly composed of these crystals, protein and urinary detritus cells. The last kiang placenta was accompanied by one large 35 g flattened piece of hippomanes.
Equines have an initially large yolk sac that remains functional for a month and then regresses completely, leaving a tiny remnant in the cord insertion.
Section of the "allantoamnion", with allantois (vascularized) above, amnion (avascular) below.
Hippomanes with birefringent crystals. Insert at higher magnification.
9) Trophoblast external to barrier
The only "invasion" of uterine tissue occurs at the chorionic girdle with the development of temporary "endometrial cups". There is an intense inflammatory response, probably because of this infiltration. This is further discussed above and also in the section on immunology.
Horse placentation is adeciduate (Ramsey, 1982). Thus, no true decidua forms and endometrial glands are being maintained, albeit modified, during the length of gestation. Importantly, there is extensive endometrial secretion of various nutrients in the areolar regions (between the microcotyledons (Wooding et al., 2000).
Vamanouchi et al. (1997) localized by hybridization studies of inhibin/activin in different gestational ages that the beta A-subunit binds to the endometrial epithelium and not to trophoblast. Thus, activin, rather than inhibin is produced by the endometrium. Its role in gestation is as yet poorly understood.
11) Various features
The placentation of most equidae is extremely similar and, thus, hybridization has been relatively often accomplished. The reader is also invited to see the chapters on the Kiang and Grevy's zebra. Acker et al. (2001) have described the morphology of equine embryos from days 17 to 40 after ovulation and "staged" them uniformly. They serially sectioned the embryos and found no anomalies, although there was a moderate degree of variation of the first appearance of various structural features.
An eminently readable general review of most endocrine parameters of equine gestations was provided by Sharp (1999). It includes the seasonal aspects, vernal transition and follicle maturation as well as pregnancy hormonal changes and their measurements. The changes of specific gonadotropic cells in the anterior pituitary during sexually active stages were defined by Tortonese et al. (2001). Human chorionic gonadotropin (hCG) and a GnRH analogue are used successfully in stimulating ovulation for artificial insemination (Samper, 2001). What is interesting in equine gestation is how the exact mechanism by which luteolysis (14 days after ovulation by uterine prostaglandins) is prevented in successful gestations; it is still under discussion. Sharp points out that some 20% failure rate exists at this time. The exact (?) product from the early (unimplanted!) conceptus that signals luteostasis, however, remains unknown. Because of the extensive mobility of the conceptus during the first week and because of the luteolysis that occurs when this is inhibited or restricted, a chemical signal appears to be likely.
The production of eCG (equine chorionic gonadotropin) is clearly the most unusual feature of the equine gestation. Allen & Stewart, in chapter 9 of McKinnon & Voss (1993) provided extensive details on the nature of this hormone in equine gestation. eCG is definitely now known to be produced by the trophoblast of the chorionic girdle. The cells can be cultured in vitro and are then shown to produce the hormone. Serum concentrations reach a peak around day 70 and then gradually fall. Nevertheless, the continued presence of the hormone in circulation after abortion often interferes with a new gestation. We assume also that this hormone may be responsible for the enormous and most unusual stimulation of the fetal gonads. It must be admitted, however, that this enlargement of gonads occurs afterthe peak of eCG levels and that there is regression of this interstitial component of the fetal gonads later in gestation. In addition, it must also be said that no eCG has been found in fetal fluids (Allen, 1982).
There is good evidence that this hormone is responsible for the recruitment of additional stimulation and development of follicles in the maternal ovaries, the so-called secondary corpora lutea. Alternatively, perhaps waves of pituitary FSH stimulate maternal follicular growth with eCG's luteotropic component causing their progesterone secretion. Many aspects of equine hormonal regulation are still under active investigation as some remain to be contradictory.
Allen et al. (2002b) studied eCG (formerly PMSG) and estradiol levels in thoroughbred horses and ponies, and in their reciprocal embryo transfers as well. The levels of eCG corresponded to the breed of the mare, rather than to the transferred foal. The quantity of estrogens produced, however, was greatest in the pregnancies with thoroughbred foals. The inference is that this depended on the large gonads that are so markedly stimulated in equine gestations. Progesterone levels also differed toward the end of pregnancy.
Kirkpatrick et al. (1990) developed progesterone assays that permit one to follow ovulation, and Czekala et al. (1990) evaluated estrone in urine and serum of equids during pregnancy. Other special hormones produced in equine gestation are equiline and equilenin; both are estrogens that are presumably produced by aromatization of fetal androgen precursors (Sharp, 1999). Allen & Stewart (2001) discussed the current status of much of our understanding of equine endocrine aspects in pregnancy. Marshall et al. (1999) identified numerous unconjugated steroids in larger quantities within umbilical arterial blood (fetus-derived) than the venous content. They discuss then pathways to C18 steroid and estrogen synthesis which appears to differ in equine substantially from other mammals.
It should further be noted that the original hypothesis by Walton & Hammond (1938), namely that fetal size may be determined by endocrine signals, has been disproved. They reciprocally crossed large Shire horses and small Shetland ponies with the resulting fetuses being intermediate-sized offspring. It was an ingenious experiment, but the recent study of embryo transfers with similar-sized animals indicated that it is the uterine size and the perfusion of uterus/placenta that regulate the fetal size, rather than hormones (Allen et al., 1993; 2002b).
Uterus and typically large ovaries in immature fetal donkey.
Testis of an abortus in Przewalski's horse. The large number of active interstitial cells surrounds few tubules with germ cells (arrows).
Fetal adrenal gland in domestic horse. A large medullary component is covered by a thin cortex that lacks a "fetal zone".
More recently, relaxin has been localized to the horse placenta (Klonisch et al., 1995). While it had been known for some time that relaxin was produced by the equine placenta, this is the first report to localize it to trophoblast and to establish an assay. The serum quantity shows marked differences among horse breeds and declines after parturition. Recent studies have suggested that the decline in serum relaxin levels may be useful for the prediction of placental problems (infection, impending abortion, and the like) and that corrective therapy (e.g. antimicrobials) may be administered to rescue the pregnancy (Ryan et al., 1999).
All domestic horses studied to date have had a chromosome number of 2n=64. The presumed ancestor, the Przewalski's horse (Equus przewalskii) has 2n=66, the donkeys have 2n=62 and zebras have 46, 44, and 32 chromosomes respectively - from North to South (Benirschke et al., 1965; 1967; Ryder et al., 1978). Hemiones have chromosomes in the 50s. An excellent study of the banding patterns, telomeres and a proposal for standardization of the donkey chromosomes comes from Raudsepp et al. (2000). Gadi & Ryder (1983) described nucleolus-organizing regions in most equidae and much more detailed modern genetic information was presented in chapters of the book by Bowling & Ruvinsky (2000). Lear (2001) determined the chromosomal location of telomere sequence, and Milenkovic et al. (2002) localized 136 genes in the horse map. Lear et al. (1999) reported an autosomal trisomy (# 31) in a colt with significant anomalies and reviewed the other four trisomies (of chromosomes # 28, 23, 26, 30) described in horses.
Hybrids among nearly all equidae have been produced and most, if not all, are sterile. The best known of these hybrids, the mule (and reciprocal hinny), have 2n=63 and are also nearly always infertile. At least all males have so far been azoospermic, and the very exceptional fertile female mule is noteworthy (Gray, 1972; Ryder et al., 1985; Benirschke & Ryder, 1985). These hybrids have been a "fertile" ground for research on imprinting and epigenetic control of development, as their phenotypes are so different (see for instance: Pennisi, 2001).
In recent years, gene localization is being studied in the horse, as for instance the α-globin gene complex was assigned to chromosome 13 (Oakenfull et al., 1993). Preliminary maps for chromosomes 1 and 10 were published by Kiguwa et al. (2000), mitochondrial control regions were defined for the Mongolian wild horse (Oakenfull & Ryder, 1998), and phylogenetic relations in equids were established with the globin genes (Oakenfull & Clegg, 1998). The exploration of this topic, however, has become too extensive for inclusion in this short chapter. George & Ryder (1986) compared mtDNA of various equidae to arrive at divergence rates. Holmes & Ellis (1999) investigated the MHC class I antigen diversity of perissodactyla and concluded that diversification in equids occurred after the split of horses from rhinoceroses.
Chromosomal errors have been described primarily in sterile mares, with 63,X being the most common karyotype. In contrast to the human Turner's syndrome (45,X) these mares are otherwise normal. Bowling & Hughes (chapter 30 in McKinnon & Voss, 1992) reviewed all cytogenetic errors then known as occurring in the horse. There are relatively few such chromosomal errors, and studies of abortion specimens have not shown chromosomal abnormalities to exist as the cause of abortion. Low-level mosaicism (XX/X) has been described in three sterile or subfertile mares by Wieczorek et al., 2001). Kent et al. (1986), reported XY sex reversal in the many horses and found it to be a heritable condition with considerable phenotypic heterogeneity. Mares with often marked masculinizing features possess XY chromosomes and pedigrees show its inheritance.
Twinning is relatively common in horses (1-2%), but it is not often very successful. Ginther (1982) stated that multiple ovulation occurs in Thoroughbreds (19%), Quarter Horses (9%), and in Appaloosa (8%). Because of their precarious outcome, most twins are eliminated by "s quashing" in early gestation (day 30). Spontaneous abortion of one twin is also common. Whitwell & Jeffcott (1975) classified the types of placentation in mares bearing twins. Survivors may have had a fused placenta with anastomoses and are then blood chimeric (Podliachuk et al., 1974). No good evidence of freemartinism in the females with male blood admixture had been forthcoming according to these authors. Nevertheless, at least two cases of freemartinism in horses due to chimerism and placental anastomoses were reviewed earlier (Benirschke, 1970). Furthermore, several cases of acardiac twinning have been recorded (see section of pathology).
G-banded karyotype of male domestic horse (Courtesy: Teri Lear).
Exceptional mule ("Krause") with offspring ("Blue Moon").
Many other genetic studies have been conducted in horses. Especially blood groups and serum proteins have been studied to assess the genetic variety and breadth of populations. This literature can be accessed from the paper by Cothran et al. (1998) who studied the polymorphism of the "Pantaneiro" horse of the Brazilian Pantanal. Despite a presumably small original stock, an ordinary degree of polymorphism was found and this is compared with many other breeds of horses.
The immunologic aspect of equine gestation has fascinated many investigators. While most of the fetal/maternal relation is epithelio-chorial and gives no reason for an immune response, the unusual development of "endometrial cups" at the chorionic girdle is different. Here, trophoblast invades the endometrium, destroys its epithelium and elicits an intense, yet transitory, lymphocellular and plasmacytic reaction whose raison d'être is still unclear, especially since it is transitory. Sharp (1999) suggested it to represent the "testing of the waters".
The major histocompatibility antigen complex (MHC) has been studied, especially often in relation to the invasive chorionic girdle cells and the attending inflammation (for instance, Baker et al., 2001 and numerous others cited there). This is a major, ongoing study from which much will be learned. It is helped by the availability of models such as the mule and the donkey-in-horse exchanges. So far, no genetic differences in horse and donkey MHC structure have been identified. Bacon et al. (2002) described MHC alloantigens in horse trophoblast. Hedrick et al. (1999) defined the MHC type II antigens in 14 Przewalski's horses and found very low variability. Baker et al. (1999) were surprised at findings in interspecific pregnancies (mares carrying mules). There was failure to down-regulate the usual cellular immune response. Transplanted girdle trophoblast (outside the uterus) differentiated normally and evaded maternal immune response (Adams & Antczak, 2001). Later, Baker et al. (2001) identified five donkey alloantigens of the MHC type I alleles by the use of interspecific equine pregnancies. They concluded that it was the relatively close relationship of horse and donkey that allowed the frequent success of hybrid pregnancies, despite the presence of alloantigens and focally invasive placentation, while in some other (cited) attempts a hybridization, most implantations fail.
15) Pathological features
Numerous pathological features of the reproductive organs have been summarized by McEntee (1990). They include endometrial atrophy, mucometra, pyometra, occasional tumors of the endometrium and some congenital anomalies. More important, some of these conditions have been the causes of reproductive failure and abortion (Platt, 1973a,b). Prickett (1970) reviewed the causes of abortion and placental lesions in mares. He considered that premature placental detachment ("red bag") was a prevalent cause of fetal demise and listed fetal diarrhea (meconium discharge) as an important event that may occur at any time in pregnancy. An important survey of a large number of spontaneous abortions and stillbirth at Newmarket has been produced by Smith et al. (2003). As was suggested earlier by Hong (v.i.), an excessively long hypertwisted (torsion) umbilical cord played a major role (38.8%). It is interesting to read that the cord may be so twisted that it obstructs the urine flow in the allantoic duct. Certainly human abortions and growth restriction in utero are frequent causes of abortion and stillbirth, but they generally are due restricted venous return from the placenta. Other major causes are ascending infection, usually most pronounced at the cervical star, herpes infection and sundry other inflammatory lesions. They refer to the apparently greater frequency of leptospirosis in Kentucky than occurring in the UK (v.i.).
Infections (ascending or hematogenous) were other listed causes of abortion with a variety of organisms the culprits. A more detailed examination of the causes of abortion and stillbirths is that by Giles et al. (1993) from their extensive experience in Kentucky. Bacterial infection was found to be the primary cause of fetal demise, followed by viral infection (equine herpes virus) and fungal infection next. "Placentitis" without recovered organisms was another major cause of demise. The most commonly recovered bacteria were streptococci, coliforms and leptospira. Other normal, commensal, oral organisms were also identified. In a Przewalski's horse at the San Diego Zoo we observed abortion in the first trimester of a foal due to placentitis and cultured Bacillus mycoides. The origin of this organism remained obscure. Interestingly, only some portions of the placenta showed intense inflammation, other regions remained normal. Patterson-Kane et al. (2002) found Rhodococcus equi as a cause of abortion in a thoroughbred mare. Leptospirosis and other organisms were identified by Donahue & Williams (2000) as causes of mare abortions. In reviewing the histopathology of many different types of intrauterine infection it is evident that the morphology differs from the common human infection, "chorioamnionitis". The inflammation is much more focal (other regions of the placenta may be entirely normal), is often associated with necrosis (obviously dependent on the organism) and the umbilical vessels are usually not inflamed, despite the fact that peripheral funisitis is common. One such abortion case at 9 months due to Leptospira pomona was supplied to me by Dr. P. Pesavento of Davis, CA. The organisms were identified with certainty only in the fetal kidneys, where they were abundant. There was much coagulative necrosis and associated inflammation in this placenta from a warmblood living on a pasture next to a cattle ranch. Leptospirosis abortion has a huge bibliography and is not an uncommon condition in equines. Hodgin et al. (1989) reported 4 cases of late abortion; Donahue et al. (1992), Poonacha et al. (1993), Giles et al. (1993) and Donahue et al. (1995) made huge contributions to our understanding this condition with their reports from Kentucky. Kinde et al. (1996) reported its occurrence following a flooding incident at Davis, and numerous other papers could be cited. The diagnosis is generally made by serology (Williams et al., 1994), Warthin Starry stains of kidneys, or by DNA fragment analysis (Lucchesi et al., 2002).
Fetal surface of mare with leptospirosis abortion. Note the ugly discoloration.
Higher magnification of the leptospirosis with thrombosed vessels.
Degeneration and inflammation in leptospirosis.
The markedly swollen, inflamed umbilical cord in leptospirosis.
A major outbreak of abortion on Kentucky horse farms occurred in 2001/2002 with hundreds of aborted or stillborn foals. It is now known as MRLS (Mare Reproductive Loss Syndrome) and the symposium has now been published (Powell et al., 2003). Placental infection was identified and commensal organisms were also frequent offenders in some of the aborted foals or of weak neonates. One of these organisms was Actinobacillus sp., a common oral species with significant cytolethality (DiRienzo et al., 2002). Stewart et al. (2002) found it to cause bacteremia in foals. Perhaps this and perhaps other organisms were the prime agents during this epidemic. Whether the "epidemic" was causally related to the contemporaneous population explosion of Eastern Tent Caterpillars (ETC) in the region has not been completely resolved. The placentas were often edematous. "Contracted foals" (arthrogryposis) and other congenital anomalies were found in 10% of cases of abortions and stillbirths. In general, however, ascending infection from the lower genital tract was the most common mode of infection, but hematogenous spread also occurred (Harrison et al., 2002). In addition, encephalitis due to Actinobacillusinfection was reported in three horses by Sebastian et al. (2002). Another thought of the cause of this unprecedented mortality is the ingestion or acquisition of cyanides from food sources, perhaps by inhalation (Burns, personal communication, 2002). Mandelonitrile (with cyanide production) from abundant black cherry tree leaves and ingested by ETC may have caused capillary damage that results in the syndrome. It is currently under active study.
Placental membranes of one of the foal deaths epidemic in Kentucky showing massive acute inflammation.
A similar case as the previous showing marked inflammation of an umbilical blood vessel.
Umbilical cord cross sections with edema from infected foals. Note greenish surface lesions, especially in right lower cord.
It is occasionally possible to absolutely prove the occurrence of maternal septicemia as the cause of fetal infection. This was for instance the case in the example of congenital coccidioidomycosis shown next. This infection has been a particular problem in San Diego where a number of Przewalski's horses have been lost to this fungal infection. The dry soils of this region and Arizona contain this organism commonly.
Composite of transplacental infection with Coccidioides immitis. A: chorion with granulomas; B: spherules in chorion; C: spherule in placental villus; D: fetal lung with numerous granulomas.
In addition to these reasons for abortion, an excessively long umbilical cord often caused fetal demise by "strangulation". This has been alluded to in the discussion on umbilical cord above and is also the topic of the publications by Vandeplassche & Lauwers (1986), and that by Hong et al. (1993). The latter authors found the normal cord length to be 52 cm; 4.5% of 1,211 abortion cases were due to excessively long cords with torsion and they were 72 cm long. The large survey of abortions and stillbirths by Smith et al. (2003) makes the same observations. What remains to be elucidated is why this is so relatively common in horses, as compared to its rare occurrence in most other species.
Endometritis is not infrequent and may be acquired by mating with an infected stallion (Platt et al., 1978), as well as easily created experimentally. Another commonly experienced condition that threatens gestations is "fescue toxicosis". Tall fescue (tough meadow grasses) is often heavily infected with fungi that produce ergot alkaloids and, when ingested, may lead to impaired placental performance and insufficiency with abortion (Hoveland, 1993; others). Most recently, West Nile fever virus infection was diagnosed in 30 horses of Florida (Wamsley et al., 2002) as well as in other Italian and American horses (Del Piero et al., 2002).
Equine specialists often speak of the "red bag" with which they mean a placenta presenting as an unusually dark red organ and they infer from this appearance that it had abrupted in utero, perhaps causing fetal demise.
Acardiac fetuses are the result of reversal of circulation in twins with placental anastomoses. They are not uncommon in human pregnancies where they generally represent one of monozygotic twins in which the normal twin "nourishes" the acardiac by reversal of blood flow. In horses, only some 4-5 cases have been described; more may occur but are overlooked. Höfliger (1974) reviewed some cases, and Crossman &Dickens (1974) illustrated such a 500 g "monster". We have seen a similar so-called acardiac amorphous ("globosus"), similarly attached to the insertion of the umbilical cord of a placenta. It was 12 cm in size, was also attached to the cord in its allantoic portion, contained blood and had a calcified shell. The soft tissue inside had fat, connective tissue and blood vessels. It has been suggested to us that this may not be an acardiac fetus but, rather, that this is the remains of a yolk sac. That is possible and bears future studies. In contrast to human acardiacs, where the karyotype is generally that of the co-twin and normal, no genetic studies have been reported in equine acardiacs. They would be of interest, as placental anastomoses exist occasionally between equine fraternal twins as well.
Equine acardiac ("globosus") fetus at arrows. It is close to the cord insertion.
Same acardiac fetus with arrow showing the two vascular connections to normal twin that are necessary for the reversed blood flow.
In addition to the acardiac twin, we have described a placental teratoma (Gurfield & Benirschke, 2003). Interestingly, the mare remained well later, thus precluding metastases from an ovarian teratoma. Most of the tissue had the appearance of dysgerminoma, but more mature elements (skin, thyroid, cartilage, etc.) were also found. It may therefore be of additional interest that we have identified two cases of benign cystic teratoma in the membranes of kiang placentas (see chapter on Kiang).
Since that publication, another placental teratocarcinoma with possible metastases to the foal was described by Allison et al. (2004). The foal died in 2½ months from metastases.
Teratoma in fetal surface of horse placenta (Gurfield & Benirschke, 2003).
The same teratoma in placental surface, here showing mostly immature, dysgerminoma-like cells in the chorioallantoic membrane.
More immature elements (cartilage, squamous epithelium, calcified degenerative material) in same placental teratoma.
Griner (1983) summarized a variety of pathologic features that he identified in exotic equidae of the San Diego Zoo. Trauma, neonatal deaths from omphalitis and intestinal helminths were the principal reasons for mortality. The next pictures are from a "papillary region" in the surface of the placenta that was initially interpreted as being squamous metaplasia. Higher magnification shows no desmosomes and the hypertrophied cells are continuous with the villous trophoblast. We have seen only one such lesion and believe that it represents a benign proliferation of trophoblast.
"Papillary" projection on outer placental surface.
Higher magnification of the previous lesion with allantochorionic vessels above and hypertrophied trophoblast below.
Higher magnification of the hypertrophied trophoblast of the "papillary" lesion.
Several investigators have described this "adenomatous hyperplasia" (dysplasia) of the equine allantois. (McEntee et al., 1988; Hong et al., 1993; Shivaprasad et al., 1994). It takes the form of numerous nodules on the allantoic membrane surface and is composed of glandular structures (that may contain secretions) embedded in loose connective tissue. Occasionally it is associated with neutrophilic infiltration but an infectious etiology probably does not always exist. McEntee et al. (1988) believed the lesion to be "dysplastic" and could not correlate it with abnormal fetal or neonatal growth. In contrast, however, Shivaprasad et al. (1994) found it to be correlated with abortion and fetal/placental infections. Grossly, the lesion appears as is shown in the next photographs, but it may also be even much more extensive.
"Adenomatous hyperplasia of allantois" in a somewhat weak horse neonate. (Courtesy, Manu Sebastian, Lexington, KY).
Another example of allantoic adenoma.
Meconium discharge in horses is conventionally referred to as "fetal diarrhea" A typical case is shown in the next photograph. It is not altogether clear to us whether the condition is similar to that observed in human gestations. Meconium discharge in humans is common (+/- 15%) and is usually the result of being post-dates (i.e. >40 weeks gestation). There is good reason to believe that this results from the elevated levels of fetal motilin concentration, rather than being secondary to "stress" (Benirschke, 2001). No such studies have been undertaken in horses.
The next lesion of a horse placenta was sent to me by Dr. Berrington of Langley, British Columbia. The foal was normal and survived; the lesions come from the lesser horn and were described as "abundant, scattered fatty masses" attached to the maternal surface. Despite the vast experience of the breeder and pathologist, such lesions had never been seen and the question was raised whether this may be the equivalent of molar degeneration of human placentas. A few of the lesions were quite large and grape-like. In my opinion the lesions are more like what in human pathology we refer to as "chorangiomatosis", a proliferation of fetal villus capillaries and occurring in chronic hypoxic conditions as well as for no apparent reason.
Swollen villus of unoccupied horn in the placenta of normal paint mare. There is vascular proliferation and an excess of trophoblast.
Higher magnification of previous photograph to show the abnormal trophoblast.
External placental surface in horse with "fetal diarrhea" (meconium passage).
16) Physiologic data
Cowell & Tyler (2001) provided a detailed review of the hematology and cytology of numerous fluids and tissues. The relationship of fetal size to maternal size and the "supply line" has been examined several times, but most decisively by Allen et al. (2002a). They found that ponies had a 325 day gestation, while the thoroughbreds had a length of 339 days. To further understand the finer reproductive regulations, embryo transfer experiments were conducted. These were also designed to verify the original observations by Walton & Hammond of crossbreeding Shire horse with Shetland ponies (1938; see above). When embryo transfers were performed between these two breeds then the gestational length was intermediate, as were the foals' weights. The ponies in thoroughbred mares were significantly heavier and correlated with weights of mare and placental sizes. Moreover, the density of microcotyledons was greater in the thoroughbreds, irrespective of the foal they carried. Thus, the placental contact area, rather than genetic background of the horses is important. Endocrine findings were more complex (Allen et al., 2002b, see section 12 above).
Wooding & Morgan (1988) defined the sites for calcium and glucose exchange in the areolar regions of the placenta. In subsequent studies, Wooding et al. (2000) showed iron transport to be achieved by the uteroferrin, a protein secreted by the endometrium in response to progesterone (Groothuis et al., 1997). Its presence can be identified in the microvesicular transport system, different from the facilitated transport of calcium. This protein, as well as many other components of uterine secretions at various cycle days were better defined by Zavy et al. (1982).
Ridderstrale et al. (1997) studied the presence of carbonic anhydrase at the interface in a variety of species. There was relatively little activity in horse placentas.
17) Other resources
Strains of fibroblast cells of most wild and domestic equidae are contained in the "Frozen Zoo" at CRES, the research division of the Zoological Society of San Diego. They can be made available on request to Dr. Oliver Ryder at firstname.lastname@example.org.
Thway et al. (2001) at Kansas Medical Center produced immortalized cell lines from girdle cells that exhibited a variety of characteristics, well-studied by these investigators; they were immortalized prior to their eCG productive ability, however.
18) Other remarks - What additional Information is needed?
As has been alluded to above, it would be of interest to know the lengths of umbilical cords in different breeds of horses. Also, the mechanism leading to the marked gonadal stimulation of the fetuses is of future interest.
The first successful cloning of equids was reported by Woods et al. (2003). A mule (cells from a 45 day embryo) was cloned into quarterhorse oocytes and the first live cloned equine foal was thus achieved in Iowa. The report suggests that births of additional horses can be anticipated. What may be of importance in this experiment is that the medium for culture of the cloned oocytes occurred in a 3-fold calcium concentration and that embryonic cells were employed. In August, 2003, a second normal foal was cloned in Italy (Galli, 2003). This report also presents new methodology of handling the newly constituted ovum with much improved success a sequel.
The first successful cloning of equids was reported by Woods et al. (2003). A mule (cells from a 45 day embryo) was cloned into quarterhorse oocytes and the first live cloned equine foal was thus achieved in Iowa. The report suggests that births of additional horses can be anticipated. What may be of importance in this experiment is that the medium for culture of the cloned oocytes occurred in a 3-fold calcium concentration and that embryonic cells were employed.
We are grateful to Teri Lear in Lexington for the banded horse karyotype. Special thanks also to Katherine Whitwell of Cambridge, UK for numerous suggestions and corrections. Dr. Allan J. Berrington of Langley, BC kindly submitted an abnormal placenta.
Acker, D.A., Curran, S., Bersu, E.T. and Ginther, O.J.: Morphologic stages of the equine embryo proper on days 17 to 40 after ovulation. Amer. J. Vet. Res. 62:1358-1364, 2001.
Adams, A.P. and Antczak, D.F.: Ectopic transplantation of equine invasive trophoblast. Biol. Reprod. 64:753-763, 2001....
How to reference this publication (Harvard system)?
Affiliation of the authors at the time of publication
1Livestock Disease Diagnostic Center, University of Kentucky, Lexington, USA. 2Department of Reproductive Medicine and Pathology, School of Medecine, University of California, San Diego, CA, USA.