Prostaglandins play an essential role during the perinatal period in the mare. Prostaglandin concentrations are low for the majority of pregnancy due to the regulatory action of progestagens on those enzymes responsible for metabolism of prostaglandins. Towards term, prostaglandin concentrations gradually increase, closely associated with upregulation of the fetal hypothalamo-pituitary-adrenal axis, stimulation of the prostaglandin synthesising enzyme PGHS-2 and changes in the ratio of progestagens and oestrogens. Recent evidence in the mare indicates that proinflammatory cytokines are key mediators of prostaglandin synthesis both at term parturition in healthy mares and at preterm parturition associated with placental infection. Prostaglandin concentrations rise substantially during active labour and decline after birth, associated with delivery of the placenta. During induced labour, prostaglandin concentrations are variable depending on the proximity to spontaneous parturition at term. Once the proinflammatory endocrine cascade is initiated, it is difficult to prevent active labour by administration of drugs that reduce prostaglandin concentrations in peripheral plasma. Further work is needed to establish the inter-relationships between prostaglandin production and other endocrine changes associated with labour at term and preterm in the mare.
Compared with other species, detailed information about the endocrinology of the perinatal period in the mare is sparse. However, since the reviews reported in the EVJ Perinatology Supplement in 1984 and the pioneering work of Peter Rossdale and colleagues, more information has emerged about the endocrinology of parturition, which indicates that the mare shows broadly similar hormone changes to women and other animals during late gestation and labour, albeit on a different time scale. Prostaglandins, produced by the uteroplacental tissues, are essential for parturition and are involved in the main physiological events i.e. cervical dilatation, rupture of the chorioallantois, contractility of the myometrium, placental separation and uterine involution (Olson 2003). There is increasing evidence from human obstetrics and research on experimental animals to show that the major release of prostaglandins during parturition, both before and at normal term, is stimulated by an inflammatory cascade within the uteroplacental tissues. This has been clearly demonstrated during preterm labour when uterine infection is present. But similar processes appear to take place, albeit at a lower level, during spontaneous, term labour when clinical infection is not present (Keelan et al. 2003; Lindström and Bennett 2005). For the majority of pregnancy, prostaglandin activity is suppressed by hormones, which are anti-inflammatory and prevent excess concentrations of prostaglandins reaching the myometrium. Towards parturition, this suppression is removed and tissue concentrations of prostaglandins are upregulated, mainly by glucocorticoids and activation of the fetal hypothalamo-pituitary-adrenal (HPA) axis. Glucocorticoids play a dual role in stimulating functional maturation of the fetal tissues in preparation for birth and the onset of labour by switching on proinflammatory pathways leading to prostaglandin release. Studies examining the hormone pathways during preterm labour in the mare as a consequence of placental infection have increased our understanding about the endocrinology of preterm birth. The current review summarises the role of prostaglandins during delivery in the pregnant mare in 3 different circumstances, during normal, spontaneous parturition at term, during spontaneous preterm labour associated with placental infection and when delivery is induced before or at full term.
Spontaneous parturition at term
Phases of uterine activity
Uterine relaxation and contractility are essential for maintenance of pregnancy and parturition. The activity of the gravid uterus has been divided into 4 periods (Challis et al. 2000; Fig 1). Prostaglandins stimulate uterine contractility and thus their actions are involved in each of these phases of uterine activity.
Phase 0 is a period of relative uterine quiescence and occurs for the majority of pregnancy. The gravid uterus is maintained under the influence of hormones that promote uterine relaxation or inhibit uterine activity and, under healthy conditions, is normally relatively unresponsive to physiological doses of uterotonic agents. In the mare, candidate hormones for maintaining uterine quiescence are progesterone and/or its related metabolites (such as 5α-pregnane-3,20-dione [5α-DHP], relaxin and parathyroid hormone related peptide [PTHrP]) (Holtan et al. 1991; Ousey et al. 2003; Fowden et al. 2008; Ryan et al. 2009).
Phase 1 is the activation phase that allows the transition from quiescence to coordinated muscular contractions within the myometrium. The uterus is stimulated partly by the mechanical forces of increased stretch caused by exponential fetal growth, and upregulation of fetal endocrine activity and other signalling pathways, including induction of prostaglandin and oxytocin receptors, activation of contractile associated proteins (CAPs), development of ion channels and tightening of gap junctions (Garfield et al. 1998; Mohan et al. 2004). Based on fetal and maternal hormone changes, this transition occurs over the last 1–3 weeks of equine pregnancy. Clinical signs include onset of mammary development and secretions, vulval and croup muscle relaxation, and the mare may spend longer periods recumbent and demonstrate periods of uterine activity, particularly at night (Haluska et al. 1987; McGlothlin et al. 2004) .
Phase 2 is the stimulation phase when the primed uterus responds to uterotonic hormones (prostaglandins and oxytocin). It is the period of maximum uterine contractility or active labour and normally lasts <1 h in the mare. Because the uterus is primed and sensitive to uterotonins, it is highly unlikely that labour can be prevented once this phase is initiated (Garfield et al. 1998).
Phase 3 involves delivery of the placenta and the onset of uterine involution as the uterus becomes refractory to uterotonic agents. Again, prostaglandins and oxytocin are important for delivery of the placenta and uterine involution.
The role of prostaglandins during different phases of uterine activity
Prostaglandins are lipid-based hormones produced by many tissues of the body. They are synthesised from arachidonic acid, which is cleaved by different phospholipase A2 enzymes from phospholipids in cell membranes, cholesteryl esters and triglycerides (Olson 2003) (Fig 2). There is an abundance of arachidonic acid so the rate limiting step in prostaglandin synthesis is the cyclo-oxygenase-enzyme-1 and -2, or prostaglandin endoperoxide H synthase (PGHS)-1 and -2. A ubiquitous enzyme, PGHS-1, is found in many tissues, is constitutively expressed and has many other functions including platelet aggregation and regulation of blood flow. Whereas, PGHS-2 is an inducible enzyme and its activity is normally undetectable or low during pregnancy but it is rapidly activated by proinflammatory or mitogenic stimuli, such as interleukin-1β (IL-1β) or platelet-derived growth factor. The highly labile intermediary prostaglandin H2, is oxidised, reduced or isomerised by various synthases into different bioactive isoforms and also into thromboxanes. Prostaglandins are inactivated by 15-hydroxy prostaglandin dehydrogenase (PGDH), found primarily in the lung but also in many other tissues including the gravid uterus. Thus production of prostaglandins is balanced by the activity of PGHS-1 and -2 leading to increased synthesis, and catabolism by PGDH, which decreases concentrations of the active prostraglandins.
The major prostaglandins of clinical interest during equine pregnancy are PGE2, PGF2α and its stable major metabolite 13,14,dihydro-15-keto-prostaglandin F2α (PGFM). A vasodilator, PGE2, involved in membrane rupture and cervical remodelling/relaxation, while PGF2α stimulates myometrial contractions in all species studied including mares (Leadon et al. 1982; Wickland et al. 1982; Gyomorey et al. 2000; Olson 2003; Jenkin and Young 2004). In common with other species, prostaglandins are produced within the uteroplacental tissues of the mare, where they act locally in a paracrine or autocrine manner (Han et al. 1995). The main site of prostaglandin synthesis appears to be the allantochorion, rather than the amnion or endometrium, while prostaglandin degradation occurs in the adjacent maternal endometrium, but not in the underlying stroma or glandular tissue, based on tissue localisation of the PGHS-2 and PGDH enzymes, respectively (Han et al. 1995; Palm et al. 2010). Concentrations of PGE2 are higher than PGF2α in all fluid compartments. PGF2α and PGE2 are released into the amniotic/allantoic fluids, which contain the highest concentrations while fetal plasma concentrations are slightly lower. Only small quantities of these primary prostaglandins are released into the mare's peripheral circulation (Silver et al. 1979; Fowden et al. 1994) (Fig. 4). Instead, maternal plasma contains relatively high concentrations of PGFM. Thus, PGE2 and PGF2α are synthesised by the uteroplacental tissues and excreted into the fetal fluids and fetal circulation, while metabolism of PGs takes place in the uteroplacental tissues and the metabolites are primarily excreted into the peripheral circulation.
Uterine quiescence (Phase 0)
For the majority of pregnancy, concentrations of PGE2 and PGF2α are relatively low in equine fetal fluids and plasma, and increase only slightly over the second half of pregnancy (Silver et al. 1979) (Fig 3). These low levels are maintained by PGDH localised in the maternal epithelial cells and interstitial tissues; PGDH is present from 150 days of gestation onwards (Han et al. 1995). In pregnant animals and women, progesterone is the major regulator of prostaglandin metabolism and maintains uterine quiescence by multiple actions; it inhibits inflammatory response pathways, inhibits expression of uterine and cervical CAPs, inhibits calcium signalling pathways and blocks production of chemokines (Doualla-Bell et al. 1998; Loudon et al. 2003; Elovitz and Mrinalini 2004; Elovitz and Wang 2004; Gehrig-Burger et al. 2010). Progesterone acts through genomic and nongenomic pathways via 2 major isoforms of the progesterone receptor (PR), PR-A and PR-B. Progesterone controls prostaglandin production by maintaining PGDH activity, an effect that is enhanced by synthetic progesterone (medroxyprogesterone acetate) and diminished by cortisol (Patel et al. 1999). Moreover, high concentrations of progesterone attenuate cortisol stimulation of PGHS-2 (Guo et al. 2009). Therefore, during pregnancy, low circulating prostaglandin concentrations are maintained by the actions of progesterone on prostaglandin metabolising enzymes (Challis et al. 2002). In the pregnant mare, the biological role of progesterone is uncertain because it is virtually undetectable in the maternal plasma during the second half of pregnancy and does not prevent oxytocin-induced myometrial contractions in vitro (Holtan et al. 1991; Ousey et al. 2000). However, since progesterone is released in small amounts from the uteroplacental tissues into the umbilical circulation, particularly in late gestation, the actual uteroplacental tissue concentrations of progesterone may be of greater significance than maternal circulating concentrations, if progesterone acts by intracrine/paracrine, rather than endocrine, pathways (Ousey et al. 2003).
Progesterone is produced from the precursor pregnenolone (P5), which is derived from the fetal adrenal glands and excreted into the umbilical arteries (Holtan et al. 1991). The P5 is metabolised into progesterone by 3β-hydroxysteroid dehydrogenase (3β-HSD) in the fetal tissues of the allantochorion and then progesterone is metabolised into 5α-DHP by 5α-reductase located within the underlying maternal endometrial tissue (Hamon et al. 1991; Han et al. 1995). Thus progesterone is synthesised and metabolised within the uteroplacental tissues and is optimally located to regulate prostaglandins also produced and metabolised within the uteroplacental tissues, as occurs in other animals (Han et al. 1995; Challis et al. 2002). Therefore, the primary role for progesterone in the pregnant mare may be to inhibit prostaglandin production and, together with other hormones (see below), ensure that uterine electromyographic (EMG) activity remains low (Haluska et al. 1987). What is less certain is whether progesterone or its immediate metabolite, 5α-DHP, is more potent at the tissue level in regulating PGDH activity (Hamon et al. 1991; Fowden et al. 2008). Concentrations of fetal and maternal 5α-DHP are much greater than progesterone during the second half of pregnancy, and 5α-DHP binds more strongly than progesterone to the PR; moreover, recent evidence indicates that 5α-DHP is bioactive at least in early pregnancy in the mare (Holtan et al. 1991; Chavatte Palmer et al. 2000; Scholtz et al. 2010). Many other progestagen metabolites are also produced in abundance by the equine fetal and uteroplacental tissues and may have a biological activity in the pregnant mare (Holtan et al. 1991; Ousey et al. 2003).
In other species, uterine relaxation is also maintained by relaxin, PTHrP and inducible nitric oxide synthase (iNOS). These hormones are involved in calcium signalling pathways, which are important for gap junction electrical activity and uterine contractility (Garfield et al. 1998). Parathyroid hormone related peptide (PTHrP) is a major vasorelaxant while relaxin and iNOS increase intracellular nucleotides that inhibit calcium release. Relaxin and PTHrP are clearly important in pregnant mares. Relaxin is produced in high concentrations during normal equine pregnancy by the fetal trophoblast but concentrations are significantly reduced in mares at risk of preterm birth implying that relaxin plays a role in myometrial quiescence (Stewart et al. 1984; Ryan et al. 2009). A direct role for PTHrP in inhibiting myometrial contractions was demonstrated in vitro using strips of myometrium from one pregnant mare (Ousey et al. 2000). Therefore, these relaxatory hormones, together with progesterone, control contractile pathways within the gravid uterus and collectively ensure myometrial quiescence during most of pregnancy.
Uterine activation (Phase 1)
This is a priming phase during which the uterus demonstrates increased periods of activity and becomes more responsive to stimuli that lead to coordinated contractions. Concentrations of PGE2 and PGF2α in maternal plasma and fetal fluids rise in late gestation and increase again during labour in mares and other animals (Silver et al. 1979; Mitchell 1991; Olson 2003; Christiaens et al. 2008). In other species, prostaglandin synthesis is a direct consequence of increased expression and activity of PGHS-2, and increased metabolism by the respective PGF and PGE synthase enzymes (McLaren et al. 2000a; Palliser et al. 2004). Paradoxically, glucocorticoids are the main stimulators of PGHS-2 activity, despite these hormones more typically being associated with anti-inflammatory actions (Gibb and Lavoie 1990; Economopoulos et al. 1996; Sun et al. 2003) (Fig 4). In sheep, PGE2 concentrations rise about 15–20 days prepartum while PGF2α increases only during active labour (Gyomorey et al. 2000; Challis et al. 2002; Jenkin and Young 2004). The rise in PGE2 is concurrent with upregulation of fetal HPA activity and rise in fetal cortisol concentrations, which stimulate PGHS-2 and PGE2 production by the placenta. Moreover, PGE2 induces 17α-hydroxylase activity leading to production of oestradiol at term, which, in turn, stimulates endometrial PGHS-2 activity, production of PGF2α and CAPs culminating in labour (Whittle et al. 2000; Challis et al. 2002; Whittle et al. 2006). Glucocorticoids also appear to stimulate prostaglandin production in human pregnancy. However, the main source of glucocorticoids originates from the mother rather than the fetus, via transplacental transfer of cortisol associated with attenuated placental 11β-HSD-2 activity, upregulation of fetal membrane 11β-HSD-1 activity and from placental corticotrophin releasing hormone (CRH) production, leading to increased placental PGHS-2 and prostaglandin production (Riley et al. 1991; Economopoulos et al. 1996; Alfaidy et al. 2003; Murphy and Clifton 2003) (Fig 4).
Similar hormone changes take place in the mare. Fetal and maternal plasma PGFM concentrations gradually increase prior to spontaneous parturition at term in pony and Thoroughbred mares, indicating increasing prostaglandin metabolism (Fig 3) (Barnes et al. 1978; Silver et al. 1979). This rise is temporally correlated with increasing fetal adrenal activity. The fetal adrenal glands enlarge, and become more responsive to adrenocorticotropic hormone (ACTH) stimulation (Fowden and Silver 1995). Adrenal output of P5 increases over the last few weeks of gestation, but due to lack of 17α-hydroxylase activity, fetal cortisol concentrations remain low with little transplacental cortisol transfer from the mother due to cortisol inactivation by placental 11β-HSD (Chavatte et al. 1995a). Instead, the C21 pathway predominates, so there are large rises in all the major metabolites of progesterone, indicating increasing enzyme metabolic activity within the fetal and uteroplacental tissues (Holtan et al. 1991; Han et al. 1995; Fowden et al. 2008). Increased P5 and progesterone (progestagen) synthesis promotes PGDH activity and attenuates prostaglandin mediated uterine activity during the final weeks of gestation. However, over the last 6 days of pregnancy, the mare experiences increasing epochs of uterine EMG activity at night together with nocturnal increases in maternal plasma oestradiol-17β concentrations (Haluska et al. 1987; O'Donnell et al. 2003; McGlothlin et al. 2004). Oestrogens are essential for myometrial contractions in the mare because removal of the source of oestrogen precursors by fetal gonadectomy leads to weak contractions during delivery (Pashen and Allen 1979). Thus a close association probably exists between myometrial contractions and oestradiol-17β concentrations in the mare, which may be due to oestrogen stimulated utero-placental production of PGF2α as occurs in pregnant ewes (Whittle et al. 2000). Certainly, PGFM concentrations rise overnight in pregnant mares in association with fasting, a response that is correlated with the degree of fetal hypoglycaemia; this response increases in magnitude as pregnancy advances (Silver et al. 1979; Fowden et al. 1994). Moreover, induction of fetal hypoglycaemia by insulin administration stimulate a PGE production (Fowden et al. 1994). Therefore, the cumulative effects of reduced maternal feed intake at night, nocturnal physiological changes, increased fetal nutrient requirements and increased growth leading to increased uterine stretch, may facilitate prostaglandin synthesis in mares and the onset of myometrial contractions, consistent with 86% of foalings occurring during the night (Rossdale and Short 1967).
Labour (Phase 2)
By the time of active labour, the uterus is already primed and simply requires a switch in the balance from inhibitory to stimulatory factors in order for parturition to proceed. The concept of release from inhibition or ‘progesterone withdrawal’ was first suggested by Csapo (1956). However, in pregnant women and guinea pigs, progesterone levels are high as they enter labour and therefore progesterone withdrawal is functional rather than actual. Thus it is proposed that parturition is initiated by a series of biochemical actions that antagonise the ability of progesterone to maintain uterine quiescence via increased expression of progesterone metabolising enzymes and interactions involving different PR isoforms (Mendelson 2009). Uterine PRs are abundant in the pregnant mare but remain unchanged throughout gestation (Chavatte Palmer et al. 2000). Maternal progestagen concentrations increase in late gestation and only decline 24–48 h before parturition (Holtan et al. 1991). They do not approach zero until after parturition and, similar to women, mares begin labour with measurable quantities of circulating progestagens. It is proposed that these high levels of progestagens are necessary in order to inhibit expression of 3β-HSD, the enzyme responsible for conversion of P5 into progesterone (Chavatte et al. 1995b; Schutzer et al. 1996). These authors demonstrated that, at physiological concentrations, several progestagens inhibit placental 3β-HSD activity in vitro and in vivo, thus leading to a decline in production of progesterone and its metabolites. This, together with the decrease in the uteroplacental supply of P5 due to activation of adrenal cortisol production, will lead to ‘progesterone withdrawal’ within the uteroplacental tissues at term (Fowden et al. 2008). This decline in uteroplacental tissue concentrations of progesterone would reduce prostaglandin metabolism via PGDH with the consequence that PGHS activity in the allantochorion would predominate and increase availability of active PGF2α.
Parturition occurs not only through withdrawal of inhibitory factors but also activation of stimulatory signals or hormones. For example in the mare, blockade of 3β-HSD with Epostane or Trilostane, does not initiate labour or parturition despite a decline in circulating progestagens, indicating that other regulatory hormones are also involved (Fowden and Silver 1987; Schutzer et al. 1996; Fowden et al. 2008). The major stimulatory pathway in all species studied including the mare, is fetal HPA activity and rising fetal cortisol concentrations (Jenkin and Young 2004). Glucocortiocoids not only stimulate fetal tissue differentiation and maturation, but in many species, they also initiate parturition by acting in a series of positive feed forward loops to alter placental steroidogenesis (Jenkin and Young 2004). In the equine fetus, both cortisol and ACTH increase during the last 48 h before parturition, indicating a similar positive feed forward loop but it is not certain whether this is the trigger for parturition in the mare (Cudd et al. 1995). Glucocorticoids also activate immune and inflammatory pathways, upregulate PGHS-2, attenuate PGDH, leading to prostaglandin synthesis and stimulate further glucocorticoid production (Whittle et al. 2000; Challis et al. 2002; Patel et al. 2003). It is clearly recognised in pregnant animals and women that labour is associated with an inflammatory response that culminates in release of prostaglandin (Christiaens et al. 2008). There is infiltration of neutrophils and macrophages into the myometrium, cervix and fetal membranes towards term. These cells then release cytokines (IL-1β, IL-6, tumour necrosis factor [TNFα]) and chemokines (IL-8) that activate nuclear factor kappa-B (NF-kB) and other proinflammatory transcription factors, such as activator protein-1 (Bowen et al. 2002; Mohan et al. 2004; Lindström and Bennett 2005). A transcription factor, NF-kB is classically activated in response to infection (lipopolysaccaride) but also by proinflammatory cytokines that contain NF-kB recognition elements within their promoter genes. Cytokine induced NF-kB therefore acts in a feed forward manner to activate further NF-kB and cytokine release. These cytokines also increase expression of genes promoting myometrial contractility including PGHS-2 and the PGF2α receptor (FP), oxytocin receptors (OR), and gap junction proteins (connexin-43) and iNOS (Mollace et al. 1998; Belt et al. 1999; Tonon and D'Andrea 2000; Zaragoza et al. 2006). Recent evidence has indicated increased expression of IL-6 and IL-8 in the equine endometrium and amnion, respectively, during spontaneous parturition at term, suggesting that upregulation of inflammatory proteins is a normal event in parturient mares (Palm et al. 2010).
Rupture of the cervical pole of the chorioallantois and release of allantoic fluid signals the start of second stage or active labour in the mare. As the mare enters active labour, maternal PGFM concentrations are already rising (up to 20 ng/ml) and increase rapidly (up to 150 ng/ml) following membrane rupture and onset of strong, coordinated uterine contractions (Fig 3) (Stewart et al. 1984; Vivrette et al. 2000). The increase in PGFM precedes the rise in oxytocin concentrations, which also increase exponentially during this expulsive phase of labour (Vivrette et al. 2000; Dudan-Hess et al. 2010). The high concentrations of prostaglandins and oxytocin are likely to be preceded by a rise in their respective uterine receptor numbers, as demonstrated in other species (Fuchs et al. 1984; Fuchs et al. 1996; Brodt-Eppley and Myatt 1999). Successful delivery of the foal requires softening and rupture of the chorioallantoic and amniotic membranes through remodelling of collagen and the extracellular matrix by many gelatinase and collagenase enzymes, collectively termed matrix metalloproteinases (MMP). In other species there is increased expression of MMP-2 and selective expression of MMP-9 by amnion and trophoblast/decidual cells, respectively, at term, and this is stimulated in part by prostaglandins (McLaren et al. 2000b; Ulug et al. 2001). In mares, MMPs are present in amniotic and allantoic fluids throughout pregnancy and at parturition. Amniotic concentrations of MMP-9 are high towards the end of gestation, but decline during labour, indicating that the tissue remodelling has already occurred by this time (Oddsdottir et al. 2011).
Placental delivery (Phase 3)
Once delivery is complete and the allantochorion begins to detach from the uterus, maternal prostaglandin concentrations and EMG activity decline (Haluska et al. 1987; Haluska and Currie 1988). Generally, prostaglandin concentrations are low within one hour of spontaneous foaling associated with placental delivery (Vivrette et al. 1995). In other species, reduced maternal prostaglandin concentrations at birth are associated with retained fetal membranes (Olson 2003). This relationship was not observed in one study of heavy draught mares (Ishii et al. 2008) but, when prostaglandins were blocked by administration of the PGHS-2 inhibitor, flunixin meglumine, prior to labour, placental delivery was delayed (Vivrette et al. 1995). Low oxytocin concentrations also have been reported in mares with delayed placental delivery (Ishii et al. 2008; F.E. Dudan and J.C. Ousey, unpublished observations). These results suggest that uterotonic hormones must be produced in sufficient quantities after birth in order to stimulate uterine contractions and expel the placenta. Because prostaglandins are produced within the placenta, which begins to detach during delivery, it seems likely that oxytocin also plays a key role in placental delivery in the mare.
Spontaneous premature parturition
Studying the endocrinology of premature parturition in the mare is difficult because of the precipitous nature of this event with few signs of impending delivery. Except in a few specialised referral practices, scanning of high risk pregnant mares in late gestation is not routine and, therefore, many cases of feto-placental compromise remain undiagnosed until the onset of clinical signs or until abortion or premature delivery occurs. There are multiple causes of equine premature parturition and perinatal deaths; umbilical cord vascular compromise (39–46%) and placental abnormalities (10–25%) predominate in Britain, whereas feto-placental infections are the leading cause of late gestation losses in the USA (Giles et al. 1993; Smith et al. 2003). Umbilical vascular problems (such as excessive cord twisting) tend not to stimulate premonitory signs in the mare, whereas feto-placental infections and other placental pathologies stimulate maternal (and fetal) endocrine responses, and premature lactation and vulval discharge are frequently observed (Macpherson and Bailey 2007). Therefore, placental infections have been induced experimentally in pony mares so that the aetiology of this condition can be investigated.
Intrauterine infections (IUI) are a significant cause of preterm birth and infant mortality in pregnant women (Goldenberg et al. 2008). Using clinical data and animal models, the endocrinology of IUI is relatively well understood (Elovitz and Mrinalini 2004; Adams Waldorf et al. 2011). Prostaglandins and their interactions with inflammatory mediators are key responses to IUI and initiate the onset of preterm labour. Bacteria gain access to the normally sterile amniotic cavity most commonly by ascending infection from the vagina and cervix but other potential routes include haematogenous dissemination through the placenta, accidental introduction via invasive procedures or by retrograde spread via the Fallopian tubes (Goldenberg et al. 2008). Once bacteria are in contact with the placenta, a typical inflammatory cascade is mediated, through recruitment of leucocytes, promoting release of proinflammatory cytokines and increasing concentrations of IL-1β, IL-6, IL-8 and TNFα in the fetal fluids, amnion, decidua and possibly placenta (Keelan et al. 1999). These in turn stimulate release of prostaglandins from the amnion, placenta, decidua and myometrium, via upregulation of NF-kB and PGHS-2 (Dudley et al. 1996; Belt et al. 1999). The rise in amniotic proinflammatory cytokines is elicited within hours of experimental infection (intra-amniotic infection with Group B Streptococcus in primates) followed within 12–24 h by a rise in PGE2, which is maintained until premature delivery some 36 h later (Gravett et al. 2007). Intrauterine infections (IUI) initiates fetal HPA activity and fetal exposure to glucocorticoids; amniotic fluid concentrations of cortisol and dehydropiandrosterone (DHEA) are high in pregnant women with IUI, and these stimulate placental CRH and oestrogen production, respectively (Gravett et al. 2000). Proinflammatory cytokines and prostaglandins stimulate placental 11β-HSD-1 and inhibit placental 11β-HSD-2, which activities collectively lead to transfer of high maternal cortisol concentrations across the placental barrier, which then act in a feed forward loop to stimulate fetal HPA activity (Alfaidy et al. 2001; Kossintseva et al. 2006). The rise in prostaglandins, IL-1β and IL-6 also acts in a series of feed forward loops to increase placental production of CRH, which stimulates further rises in amniotic fluid prostaglandins, fetal cortisol and DHEA the precursor for oestrogen production (Jones and Challis 1989; Petraglia et al. 1990; Smith et al. 1998; Gravett et al. 2000). Therefore, in the face of overwhelming endocrine changes that stimulate prostaglandin release and uterine contractility, IUI activation of the fetal HPA axis enables fetal tissue maturation and preparation for early delivery so that the infant can survive outside of the infected uterus despite its small size and truncated period of intrauterine development (Laatikainen et al. 1988).
Over the last 2 decades, a series of studies have investigated the endocrine events associated with naturally occurring and experimentally induced placental infections in mares. There are similarities between the endocrine pathways described by these equine studies and those in other animals described above. Early research in Thoroughbred mares with naturally occurring placental disturbances correlated precocious activation of the fetal HPA axis with infection and an early rise in maternal plasma progestagen concentrations and onset of mammary development before 310 days of gestation (Rossdale et al. 1991). Similar maternal changes were observed after inducing IUI experimentally by placental separation at the cervical pole in pregnant pony mares. Direct stimulation of the fetal adrenal with exogenous CRH or ACTH also elicits a rapid rise in maternal progestagens, presumably through enhanced production of adrenal P5, demonstrating a clear correlation between precocious fetal HPA activity and maternal progestagen rises (Rossdale et al. 1995; Fowden et al. 2008). Subsequently, specific models were developed to induce IUI by inoculating bacteria (Streptococcus equi ssp. zooepidemicus) into the cervical canal of pregnant mares of <300 days gestation to create ascending placentitis, and measuring maternal and placental responses (LeBlanc et al. 2002; Ryan et al. 2009; Bailey et al. 2010). Examination of maternal progestagens in the infected mares demonstrated a rise in progestagens within 8 days in those pregnancies with live fetuses, with some delivering live foals with precocious fetal maturation of the HPA and other tissues, while, in other mares, progestagen concentrations decreased in association with fetal death and abortion (Morris et al. 2007). Further investigations of proinflammatory cytokines and prostaglandin production in mares with placentitis (Fig 5) showed that allantoic fluid concentrations of PGE2 and PGF2α were approximately 10 times higher prior to abortion or preterm delivery than concentrations in control mares close to full term (LeBlanc et al. 2002). Silver et al. (1979) also reported elevated concentrations of PGE2 and PGF2α in both fetal plasma and allantoic fluid during the final 6 days prior to abortion in late gestation, while maternal PGFM concentrations were unchanged. Moreover, placental expression of numerous cytokines (IL-1β, IL-6, IL-8, IL-15, IL-18, interferon-γ) is greater in mares with placental infection than in noninfected mares (LeBlanc et al. 2002; Lyle et al. 2009). These results suggest that proinflammatory cytokines produced by the allantochorionic tisssues in response to infection and recruitment of leucocytes (Fig 5), stimulate uteroplacental production of PGE2 and PGF2α and result in increased prostaglandin levels, within the uteroplacental tissues, which remain elevated for some days, if left untreated. The rise in maternal progestagens indicates activation of the fetal HPA but, unlike normal term labour when prostaglandin production is suppressed by progestagens until close to delivery, both prostaglandin and progestagen concentrations were raised for several days in these infected pregnancies. Further investigations are needed to identify the endocrine changes within the uteroplacental tissues of abnormal pregnancies. The proinflammatory responses of infected uteroplacental tissues also appear to be associated with elevated fetal glucocorticoids as high allantoic cortisol concentrations are observed with significant expression of placental IL-1β in experimentally infected mares (Lyle et al. 2010a). This indicates that 17α-hydroxylase may be activated in the fetal adrenal gland some weeks prior to term in mares with IUI. Alternatively, there may be increased transplacental cortisol transfer from the mother as a result of impaired placental 11β-HSD activity, although preliminary evidence suggests no significant relationship between maternal and fetal cortisol concentrations in mares with in utero infections (Chavatte et al. 1995a; Lyle et al. 2010b). The decline in maternal progestagens and early fetal death in some mares indicates that the fetal glucocorticoid response is an essential element for fetal survival. Precocious activation of the fetal HPA appears to stimulate organ maturation, although not skeletal maturity and thus provides an ‘escape route’ for the infected foal, which will aid in its survival ex utero. However, these foals are often small with poor bone ossification (predominantly tarsus and carpus). Such detrimental effects may persist into later life and have consequences for the foals' future athletic potential. Therefore, treatments to prevent preterm delivery and to maintain pregnancies to term are currently being investigated (Bailey et al. 2010; Christiansen et al. 2010).
Induced parturition (term and preterm)
Induction of delivery requires activation of uterotonic hormones, prostaglandins and oxytocin. Before 35 days of pregnancy, a single dose (250 µg) of exogenous PGF2α or its synthetic analogue (fluprostenol or chloprostenol) usually terminates pregnancy via their luteolytic action on the corpus luteum (Penzhorn et al. 1986). However, once pregnancy is dependent on placental production of progestagens, then prostaglandins are less effective at inducing parturition. During this period, single doses of prostaglandins or their analogues rarely induce delivery unless the mare is close to term; instead multiple, high doses of prostaglandins over hours or even days are usually required (Alm et al. 1975; Rossdale et al. 1979; Leadon et al. 1982). In mares that are induced in late gestation, maternal PGFM concentrations rise either rapidly (within 15 min) in mares that are close to foaling or more slowly (>60 min) if they are unready for birth (Barnes et al. 1978; Rossdale et al. 1979). If delivery does not occur or the foal is dead, then PGFM concentrations remain low (Silver et al. 1979; Leadon et al. 1982). Since prostaglandin production depends, in part, on oestrogen concentrations in parturient sheep and horses, PGFM concentrations are low and labour contractions are weak in the absence of raised oestrogen concentrations (Pashen and Allen 1979; Whittle et al. 2000). Thus, the late rise in oestradiol-17β concentrations in mares may provide the key to whether labour induction with prostaglandin is successful or not. Oxytocin appears to be more effective than prostaglandins at inducing delivery in the mare, although higher doses are required prior to term than those needed at term (Leadon et al. 1982). In women, oxytocin is a very potent uteronic agent and, when bound to its receptor, is capable of stimulating PGF2α production and upregulating PGHS-2 (Husslein et al. 1981; Terzidou et al. 2011). It may have similar actions in the mare and overcome the relaxant effects of progesterone and other relaxatory hormones.
Suppression of prostaglandins during delivery
Clearly, suppression of the production and action of the prostraglandins could be beneficial for the prevention of uterine contractions and preterm delivery in mares with IUI. However, reducing maternal prostaglandin concentrations does not prevent either spontaneous, or oxytocin-induced, labour in healthy mares. The PGHS inhibitor meclofenamic acid (500 mg per os) is effective at abolishing the rise in PGFM concentrations that occurs following surgery or during fasting in pony mares (Silver et al. 1979). However, administration of higher doses (1.5 g/day) to pregnant pony mares for 12 days before term does not prevent parturition or high PGFM concentrations during labour despite suppression of basal, prelabour PGFM levels in the mare (Pashen 1981). In mares given high doses of dexamethasone (100 mg i.m.) for 3 days daily, maternal plasma PGFM concentrations were suppressed but 2 mares foaled spontaneously within 48 h of their last dexamethasone treatment with low PGFM concentrations at delivery (Ousey et al. 2011). Moreover administration of the PGHS-2 inhibitor, flunixin meglumine to mares 8 and 1 h before induced parturition at term again suppressed PGFM concentrations but did not prevent labour (Vivrette et al. 1995). These data indicate that once the uterus is in an activated state and CAPs are initiated, it is difficult to reverse or halt myometrial activity.
Metabolism and synthesis of prostaglandins is central to the processes of uterine quiescence and activation, respectively in the mare. Increasing evidence suggests that proinflammatory cytokines and chemokines play an important role in activating prostaglandin production not only at normal term but also before term when feto-placental infection is present. Prepartum upregulation of these proinflammatory processes and prostaglandin synthesis is often closely associated with activation of the fetal HPA axis and with alterations in uteroplacental production of the progestagens and oestrogens involved in regulating myometrial contractility. This ensures that there is a temporal correlation between fetal maturation and onset of active labour. However, significantly more research is required in pregnant mares to establish the inter-relationships and temporal sequences of the endocrine and other events regulating prostaglandin production in both normal and compromised pregnancies.
Authors' declaration of interests
No conflicts of interest have been declared.
Source of funding
We are grateful to the Horserace Betting Levy Board for funding.
We would like to thank veterinarians, laboratory technicians, research assistants and students based at, or visiting, Rossdale and Partners Veterinary practice, Newmarket, and at the Department of Physiology, Development and Neuroscience, University of Cambridge, who have assisted with the equine studies over many years. We are grateful to the Horserace Betting Levy Board who have provided funding, and to the Equine Fertility Unit, Newmarket, the Horseracing Forensic Laboratories, Newmarket, and the Animal Health Trust, Kentford, who have also kindly provided research facilities. Finally we would like to thank Peter Rossdale for his inspiration, leadership and encouragement to investigate and study various aspects of perinatology in the mare.