Relationship between infection, inflammation and premature parturition in mares with experimentally induced placentitis



Reasons for performing the study: Ascending placentitis results in premature birth and high foal mortality. By understanding how placentitis induces premature delivery, it may be possible to develop diagnostic markers and to delay premature delivery pharmacologically, thereby decreasing perinatal foal mortality.

Objective: To identify relationships between bacterial infection, inflammation and premature parturition in mares with experimentally induced placentitis.

Materials and methods: Experiment 1: Concentrations of allantoic fluid prostaglandins (PGs) F and E2 were measured in 8 mares after intracervical inoculation with Streptococcus equi ssp. zooepidemicus (at Days 285–291 of gestation) until parturition and compared with controls (n = 4). Experiment 2: mRNA expression of interleukin (IL)-1β, IL-6, tumour necrosis factor (TNF)-α and IL-8 in the chorioallantois from inoculated mares in Experiment 1 were compared with 7 mares that foaled normally.

Results: Bacterial inoculation resulted in 7 aborted fetuses and birth of one premature, viable foal. Infection was associated with inflammation of the chorioallantois in the region of the cervical star, isolation of bacteria and high concentrations of PGE2 and PGF in allantoic fluid obtained within 48 h of delivery (P = 0.04). Chorioallantois from all mares expressed mRNA for IL-8, TNF-α, IL-6 and IL-1β. Experimentally infected mares expressed more mRNA for IL-6 (P = 0.003) and IL-8 (P = 0.009) in the cervical star region and more mRNA for IL-6 (P = 0.004) in tissues from placental horns than control mares.

Conclusions and clinical relevance: Bacterial placentitis may result in liberation of cytokines from the chorioallantois and prostaglandin formation leading to abortion or birth of a precociously mature foal.


Premature delivery of a foal is devastating because foals born before the rise in fetal cortisol that typically occurs in the last 24–36 h of pregnancy usually die (Rossdale and Silver 1982; Silver 1990; Silver and Fowden 1994) due to lack of maturation of key organ systems needed to sustain life. The single most important cause of premature delivery of foals in the USA is placentitis. In a review of 3527 equine cases submitted to the Kentucky Diagnostic Laboratory over 5 years, placentitis was diagnosed as causing nearly a third of the premature births, still births and deaths in the first 24 h of life (Giles et al. 1993; Hong et al. 1993). In addition to inducing premature delivery, chronic placentitis may accelerate fetal maturation, resulting in the birth of precociously mature foals. For this reason, mares that exhibit clinical signs of placentitis, vaginal discharge or early udder development, are treated empirically in an attempt to delay parturition. If therapeutic strategies are to be developed for delaying premature delivery and the subsequent delivery of viable foals, the mechanism that induces premature labour needs to be identified.

In women and primates, intrauterine infection is also highly associated with idiopathic preterm labour. Bacteria ascend from the maternal vagina, infect maternal and fetal gestational tissues near the cervix and establish an inflammatory focus. It has been proposed that infection of human gestational tissues activates macrophages within the decidua resulting in production of proinflammatory cytokines and arachidonic acid metabolites by decidual and chorion cells. This inflammatory process initiates prostaglandin production by amnionic and/or decidual cells, stimulation of myometrial cell contractility and premature labour (Romero and Mazor 1988; Hillier et al. 1993; Dudley and Trautman 1994; Pollard and Mitchell 1996; Dudley 1997). Chronic chorioamnionitis has also been associated with activation of the fetal hypothalamic-pituitary-adrenal (HPA) axis resulting from stress-induced stimulation of corticotrophin releasing hormone, from direct action of proinflammatory cytokines on the fetal adrenal or from a combination of both (Besedovsky and del Rey 1996; Gravett et al. 2000).

It is not known what role, if any, proinflammatory cytokines have in placentitis in the mare. Data generated from our equine model of infection support the concept that proinflammatory cytokines are involved in equine placentitis because premature labour may occur without fetal infection (Calderwood-Mays et al. 2002) and infection may activate the fetal HPA axis and induce precocious maturation (Morris et al. 2007). The objectives of this study were to: 1) measure prostaglandin concentrations in allantoic fluid of mares with experimentally induced placentitis to determine if there was an association with premature delivery; and 2) determine if messenger ribonucleic acid (mRNA) expression of the proinflammatory cytokines interleukin (IL)-1β, IL-6, tumour necrosis factor (TNF)-α and IL-8 within the placenta was increased in mares with experimentally induced placentitis. It was hypothesised that bacterial invasion of the chorioallantois causes an acute inflammatory response in the tissue, and a rise in prostaglandins (PGs) F and E2 in allantoic fluid leading to premature delivery.

Materials and methods

Experiment 1: measurement of prostaglandins in allantoic fluid

Experimental design and animals: Twelve pony mares were bred naturally or by artificial insemination. Pregnancy was confirmed by ultrasonographic examination of the reproductive tract. Mares were divided into 2 groups: controls (n = 4); and experimentally induced ascending placentitis (inoculated, n = 8). Four of 8 inoculated mares were instrumented with allantoic catheters while 4 treatment and 4 control mares had allantoic fluid samples collected by allantocentesis. In preliminary trials, due to rapid blockage of allantoic fluid catheters with hippomane, catheters were modified in an attempt to improve patency. However, it was unlikely that catheters would remain patent for the 6–8 weeks needed to collect all samples in control mares. Therefore, we collected allantoic fluid in 4 treatment mares by allantocentesis to match the collection method chosen for the control mares. Mares were maintained at pasture during the day and in stalls at night except for instrumented mares, which were housed in stalls after surgery and hand walked 3 times daily. This project was approved by the Institutional Animal Care and Use Committee.

Surgical instrumentation and post surgical management: Presurgical preparation, anaesthetic technique, positioning on surgical table and surgical approach for instrumentation were as described previously (McGlothlin et al. 2004). Mares were fitted with allantoic fluid catheters between Days 265 and 275 of gestation (dGa 265–275). Catheter placement was through an incision on the dorsolateral surface of the pregnant horn below the attachment of the broad ligament. Two intrauterine catheters (PV 10)a, 30 cm in length, were inserted into the allantoic fluid. Catheters had been modified before surgery by creating 20 additional openings between the catheter tip and the 15 cm mark in an attempt to improve catheter patency. Catheters were marked with indelible ink at 5 cm increments to record the amount within the uterine lumen accurately and to estimate catheter movement at the body exit site. The placenta was over sewn to prevent bleeding and the incision closed with a Cushing pattern. The uterine catheters were tacked down onto the uterine serosa 5 and 15 cm from the uterine incision with No. 2-0 polydioxanone (PDS II)b. Allantoic fluid catheters were passed through the abdominal wall in the flank region at least 10 cm from the incision. The skin was undermined and the tubing passed through the subcutaneous space and brought out over the lumbar region. The intrauterine catheters were closed with a sterile nail. The low flank incision was closed using routine techniques. Catheters were wrapped in sterile gauze, placed in sterile plastic bags and inserted in a catheter pouch that was sewn to the mare's back. Mares were given potassium penicillin Gc (20,000 u/kg bwt), gentamicin (Gentocin)d (6 mg/kg bwt i.v.) and flunixin meglumin (Banamine)d (1 mg/kg bwt i.v.) through an indwelling jugular catheter 2 h before anaesthesia was induced. Altrenogest (Regu-Mate)e, an oral progestin was also given (0.0088 mg/kg bwt) for 7 days. Antimicrobials and flunixin meglumin were given at the appropriate intervals for 4–6 days after surgery, depending on the mare's response to instrumentation. No drugs were given to mares in the 72 h preceding the first collection of allantoic fluid or during subsequent collections.

Bacterial inoculation: A single stock solution of 1 × 109 colony forming units (CFU)/ml of Streptococcus equi, ssp. zooepidemicus (S. zooepidemicus) was divided into 2.5 ml aliquots, stored in vials at -80°C and used for all studies. The isolate was obtained from a clinical case of equine endometritis submitted to the Microbiology Laboratory of the College of Veterinary Medicine, University of Florida, and was identified as S. zooepidemicus based on standard microbiology techniques (beta haemolysis, Gram-positive, catalase negative, Lancefield group C and standard biochemical tests). Inocula were prepared by diluting the stock solution with sterile saline to yield either 1 × 108 CFU (instrumented mares) or 1 × 107 CFU (noninstrumented mares) in a volume of 1 ml (McGlothlin et al. 2004). Doses differed between groups as instrumented mares did not develop placentitis with the lower dose and noninstrumented mares aborted within 72 h when given the higher dose in preliminary trials.

Mares were inoculated between dGa 285 and 291. The inoculation procedure consisted of wrapping the tail, washing the perineum thoroughly with povidone iodine and drying the area thoroughly. A sterile artificial insemination pipette was introduced into the vagina by an operator whose hand and arm were covered with a sterile sleeve and glove. The cervix was located, the cervical plug that was in the caudal 2–3 cm of the cervical canal was removed manually and the inoculum deposited approximately 3 cm into the cervical canal.

Allantoic fluid collection and handling of samples: Allantoic fluid was collected once before and every 2–4 days after bacterial inoculation through allantoic fluid catheters in instrumented mares. Catheters were removed from the catheter pouch, scrubbed with betadyne for 5 min and rinsed with alcohol. The nail sealing the catheter was removed and fluid was collected through a sterile 3 ml syringe. The first 3 ml of fluid collected, the amount of fluid within the catheter, was discarded. A 4 ml sample of allantoic fluid was collected and processed as indicated below. After fluid collection, each catheter was infused with 3 ml of sterile saline and a sterile nail inserted into the catheter opening. Catheters were washed, wrapped in sterile gauze, placed in a new sterile plastic bag and returned to the catheter pouch. A minimum of 4 ml of allantoic fluid were collected every 4–6 days in noninstrumented, inoculated mares and every 7–10 days from dGa 280 to parturition in control mares by allantocentesis as described previously (Paccamonti et al. 1995). No samples collected at parturition were used in this study because of possible bacterial contamination with vaginal flora. Allantoic fluid was collected into a pyrogen free, sterile vial, immediately placed on ice and prepared as follows. For PGF and PGE2 assays, 2 ml of allantoic fluid were added to a chilled solution of 4.5 mmol/l EDTA and 2 mg/ml aspirin (a prostaglandin synthetase inhibitor) and placed on ice. To isolate aerobic bacteria, yeast or fungi, a drop of allantoic fluid was plated on blood agar and incubated for 72 h at 37°C. Growth on plates was identified.

Prostaglandin assays: PGE2 was measured by a competitive binding radioimmunoassay kit (NEK020)f. The cross-reactivity of the kit was 100% for PGE2, 30% for PGE1 and <1% for all other PGs. Inter- and intra-assay variation were 16.8% and 14.8%, respectively. Prostaglandin F was measured using an enzyme immunoassay kitg. Cross-reactivity with other prostaglandins was as follows: 100% PGF2α, 100% PGF, 100% PGF , 51% PGD2, 1% −6, 15-di keto-13,14-dihydro PGF,<1% for all other prostaglandins. Inter- and intra-assay coefficients of variation were 16.8% and 9.7%, respectively. See Supplementary Item S1 for additional details.

Experiment 2– mRNA expression of proinflammatory cytokines in placental tissue

Experimental design: The mRNA expression of the proinflammatory cytokines IL-8, IL-6, IL-1β and TNF-α were measured in the chorioallantois of the 8 mares with experimentally induced placentitis from Experiment 1 (4 instrumented and 4 not instrumented) and in 7 mares that delivered normal foals (3 from Experiment 1 and 4 additional mares) to determine if ascending placentitis was associated with an increase in the expression of cytokines in fetal tissues. One control mare was not included in Experiment 2 because she foaled unobserved and tissues for measurement by real-time polymerase chain reaction (PCR) had to be processed immediately. Four mares were added to the control group to increase sample size. These mares were maintained under similar conditions at the University of Florida but did not have any procedures performed on them other than routine preventative care. The chorioallantois was processed within 5 min of its passage. Sections of grossly abnormal and normal chorioallantois located at least 10 cm from the grossly abnormal areas were obtained. Samples were divided into 2 specimens, with one immediately submersed in liquid nitrogen until assayed for cytokine mRNA expression and the other placed in formalin for histological evaluation.

RNA isolation, DNAse treatment of RNA samples, complementary deoxyribonucleic acid (cDNA) synthesis and quantification of cytokine mRNA expression by real-time PCR: Total RNA was isolated from 50 mg of placental tissue using the RNeasy total RNA isolation kith. All RNA samples were treated with amplification grade DNAse Ii to remove any traces of genomic DNA. cDNA synthesis and relative quantification of IL-1β, IL-6, IL-8 and TNF-α mRNA expression were performed exactly as previously described (Garton et al. 2002). See Supplementary Item S1 for additional details.

Data analysis

Experiment 1: Allantoic fluid samples were divided into 4 periods. Period 1 included samples collected before inoculation and on dGa 280 in control mares (Table 2). Period 2 included allantoic samples collected 3–4 days after inoculation and samples from dGa 288 to 290 in control mares. Period 3 included samples collected 6–10 days after inoculation and samples from dGa 296 to 303 in control mares. Period 4 included allantoic samples collected within 48 h of delivery. There was no overlap of samples between periods. Samples included in Period 4 were from mares that delivered 10, 12, 20 and 27 days after inoculation. Data sets for Period 1 was complete. Data for Period 2 included 7 inoculated and 4 control mares; Period 3 included 5 inoculated and 4 control mares; and Period 4 included 4 inoculation and one control mare. A 2-way ANOVA with repeated measures was used to compare the effect of group (infected vs. controls), time (Period 1: pre-inoculation; Period 2: 3–4 days post inoculation; and Period 3: 6–10 days post inoculation), and interactions between groups and time on concentrations of PGs. No statistical comparisons were made between groups or within control group for Period 4 as data from only one control mare were available. When appropriate, multiple pairwise comparisons were done using the Fisher's PLSD Test.

Table 2. Concentrations of prostaglandin (PG)E2 and PGF in allantoic fluid of inoculated and control mares
Inoculated mares  Period 1 Basal Period 2 3–4 days post inoculation Period 3 6–10 days post inoculation Period 4 Within 48 h of delivery
  1. a,bNumbers in row with different superscript are significantly different (P<0.05). 1I-instrumented mare; NI-non-instrumented mare. 2Only one sample was collected within 48 h of delivery in control mares. Basal concentrations were obtained in inoculated mares before inoculation and on dGa 280 in control mares.

 No. of mares8754
2 I; 2 NI1
PGE2 (pg/ml) Mean ± s.d.78.1 ± 38.5a498 ± 360a2404 ± 114215, 783 ± 9104b
Median Range54.2179.520959011
PGF(pg/ml) Mean ± s.d.188.1 ± 49a348.9 ± 105.4a1062 ± 452.14021 ± 2159.8b
Median Range214288.7810182699
Control mares   dGa 280 dGA 288–290 dGa 296–303 Delivery
# of mares4441
PGE2 (pg/ml) Mean ± s.d.41.54 ± 4.7892.2 ± 47.984 ± 22.2426.32
PGF(pg/ml) Mean ± s.d.171.6 ± 30.7197.5 ± 29244.1 ± s23.34322

Experiment 2: Cytokine mRNA expression data followed a Gaussian distribution. Comparisons of cytokine mRNA expression between the cervical star region and placenta within the same animal were made using the Wilcoxon signed rank test. Comparisons of cytokine mRNA expression between control and placentitis groups were made using the Mann-Whitney U test. Significance was based on a P value of 0.05. Tendencies were reported if the P value was <0.10. The association between time from infection to delivery and cytokine mRNA expression was assessed using Spearman's coefficient of rank correlation.


Experiment 1

Seven of 8 inoculated mares aborted between 6 and 27 days after inoculation (Table 1) while one instrumented mare delivered a viable foal on dGa 309, 20 days after inoculation. Bacteria were isolated only from allantoic fluid samples collected in the 48 h preceding abortion (Period 4; n = 3) or from fetal stomach contents collected immediately after abortion (n = 7). S. zooepidemicus was isolated alone or in combination with another organism in inoculated mares that aborted. Three of 4 control mares delivered viable foals on dGa 318, 347 and dGa350, respectively. The fourth mare aborted 24 h after the third allantocentesis on dGa 296 from an iatrogenically induced Bacillus spp. infection.

Table 1. Outcomes of experimental mares that received a cervical inoculation of Streptococcus equi ssp. zooepidemicus (Strep. zoo)
MareInstrumented (I); Not instrumented (NI)Time from inoculation to deliverydGa mare delivered1Fetal outcomeBacteria isolated2
  1. *Ponies had complete data sets for evaluating temporal changes in allantoic fluid cytokines and prostaglandins. 1dGa = day of gestation. 2Bacteria were isolated from allantoic fluid sample collected within 48 h of delivery (n = 3) or from fetal stomach contents collected at necropsy (n = 7). 3Foals were subjected to euthanasia after delivery because they were recumbent and premature.

PegI6 days294Dead Strep. zoo and E. coli
JillI6 days293Dead Strep. zoo
Que *I20 days309Viable foalNone
Nora *I27 days314Not viable3 Strep. zoo
714NI6 days296Dead Strep. zoo and E. coli
Tina *NI10 days296Dead Strep. zoo
62 *NI12 days298Dead Strep. zoo
KimNI15 days306Not viable3 Klebsiella and Strep. zoo

Thirty-seven allantoic fluid samples, 24 from inoculated mares and 13 from control mares were analysed for concentrations of PGF and PGE2 (Table 2). There was an effect of time on the concentrations of PGE2 and PGF in allantoic fluid collected from inoculated mares with Period 4 concentrations being higher than those of Periods 1 and 2 (P = 0.03). Concentrations of PGE2 and PGF in allantoic fluid collected in Periods 1 to 3 did not differ between or within groups. Allantoic fluid concentrations of PGE2 and PGF obtained from 2 inoculated mares, one that delivered a nonviable foal (a) and one that delivered a viable foal (b) and one control mare (c) are presented in Figure 1. Allantoic PG concentrations in control mares remained low and were relatively stable throughout the last 60 days of gestation.

Figure 1.

Allantoic fluid concentrations of prostaglandin (PG)E and PGF over time. a) is from an mare that delivered a nonviable bacteraemic foal 27 days after inoculation with Streptococcus equi ssp. zooepidemicus, b) is from the mare that delivered a viable, precociously mature foal on dGa 314, 20 days post inoculation and c) is from the control mare that delivered a viable healthy foal on dGa 347, 48 h after the last allantocentesis.

Experiment 2

The cervical star region of the chorioallantois from mares with experimentally induced placentitis was thickened, necrotic and covered with a purulent or brown exudate. Histopathological diagnosis for all inoculated mares was bacterial, suppurative, necrotising placentitis with oedema limited to the cervical star region. Chorioallantois collected from the 7 control mares and from tissues >10 cm from the cervical star region of inoculated mares were grossly and histologically normal. All fetal membranes expressed mRNA for IL-8, TNF-α, IL-6 and IL-1β; however, inoculated mares expressed significantly more mRNA for IL-6 (P = 0.003) and IL-8 (P = 0.009) in the cervical star region, and had significantly higher mRNA expression for IL-6 (P = 0.004) in grossly normal placenta than control mares (Fig 2). There was a tendency toward higher IL-1β mRNA expression in infected mares both at the cervical star (P = 0.09) and placental body (P = 0.06), and toward higher IL-8 mRNA expression at the body of the chorioallantois (Fig 2a). Within the same animal, there was a tendency toward higher IL-8 mRNA expression at the cervical star compared to the body of the chorioallantois in infected mares (P = 0.08; Fig 2c). Expression of mRNA IL-1β, IL-6 and TNF-α mRNA between the cervical star and horns within the same mare were not significantly different. There was no significant association between time from infection to delivery and IL-1b, IL-6, IL-8 or TNF-a mRNA expression.

Figure 2.

Interleukin (IL)-1β, IL-6, IL-8, and tumour necrosis factor (TNF)-α messenger ribonucleic acid (mRNA) expression at the cervical star and in the placental body of 8 mares with experimentally induced placentitis (interrupted line) and 7 reproductively normal mares (solid line). Numbers on the y axis represent the n-fold difference in cytokine mRNA expression above that of the sample with the lowest expression. The central box represents the values from the lower to upper quartile (25th–75th percentile). The middle line represents the median. The error bars extend from the minimum to the maximum value, excluding outliers (upper or lower quartile ± 3 times the interquartile range), which are displayed as separate points (open circles). *indicates a statistically significant difference in mRNA expression between mares with placentitis and reproductively normal mares (P<0.05).


This is the first report of relationships between inflammation of the chorioallantois, increased allantoic fluid prostaglandins, premature parturition and birth of a precociously mature foal in mares with experimentally induced ascending placentitis. Similar to women, inflammatory cytokines were increased in the chorioallantois in normal mares at parturition but levels were significantly higher in the chorioallantois of mares with placentitis (Dudley and Trautman 1994; Winkler 2003). Inflamed tissue from the cervical star region of infected mares expressed more mRNA for IL-8 and IL-6 than samples collected from normal mares. A chemokine, IL-8, contributes to the initiation of acute inflammation but is also critical to the propagation of inflammation by recruiting neutrophils to surrounding tissues. It induces cervical ripening, and influences the levels of IL-6 and IL-1β expressed within fetal membranes of women with chorioamnionitis (Romero and Mazor 1988; Dudley and Trautman 1994; Romero et al. 1998; Winkler 2003). An immunomodulatory cytokine, IL-6, has a multitude of functions including acting as a growth factor for activated B cells, promoting T-cell differentiation and stimulating production of acute-phase reagents by the liver. Studies in women have indicated a direct temporal relationship between increased IL-6 levels in fetal plasma and commencement of labour (Romero et al. 1998). Expression of TNF-α or IL-1β mRNA did not differ significantly between groups or tissue sites. In response to an inflammatory insult, TNF-α is secreted first, followed by secretion of IL-1β and then IL-6, which then inhibits secretion of TNFα and IL-1β (Chrousos 1995). It is possible that the lack of a difference in TNF-α or IL-1β mRNA between groups was due to minor increases, remission of these factors once the cytokine cascade had been initiated or variation within tissues collected from infected mares. We had previously reported that there were no differences in concentrations of IL-1, IL-6 or TNF-α in allantoic fluid between groups or over time (LeBlanc et al. 2002). Data were not included in this study because the methodologies used to measure fluid cytokines in that study lack specificity compared with those in use today. Using the 2002 findings may result in erroneous conclusions.

Concentrations of PGE2 and PGF rose significantly before premature parturition in infected mares and profiles of inoculated mares differed from the control mare (Fig 1). Prostaglandins began their rise 6–10 days post infection with mean PGE concentrations being 3.75 fold higher (median 3 fold higher) than PGF in samples collected within 48 h of delivery. Mean allantoic fluid concentrations of PGE and PGF of infected mares in Period 3 were 29 and 4.3 times higher (median 28.3 and 4.2 times higher) than control mares and were 37 and 9.3 times (median 21 and 6.2 times higher), respectively, higher than the control mare in Period 4 (Table 2). However, the wide variation in PGE and PGF in infected mares during Period 3 and low sample size may have precluded detection of differences. The source of the increased production of allantoic fluid PGE was not identified; however, we speculate that it originates from the chorioallantois in the region of the cervical star due to release of cytokines by activated macrophages and neutrophils in response to bacterial invasion of tissue. In women with chorioamnionitis, bacterial products stimulate production of IL-1, TNF and IL-6 in amnion and decidua, which then produce PGE, and these along with PGF are key mediators of uterine contractility (Romero and Mazor 1988; Bry et al. 1994; Dudley and Trautman 1994; Challis et al. 2002; Gibb and Challis 2002). Prostaglandin E concentration rises some days before labour commences while PGF concentrations rise substantially during labour in most species studied (Vivrette et al. 2000; Gibb and Challis 2002).

Silver et al. (1979) reported large rises in allantoic fluid concentrations of PGE and PGF 5–6 days before abortion in chronically catheterised fetuses with concentrations of PGE being higher than PGF. In contrast, concentrations of maternal plasma PGFM, PGE or PGF remained low and did not alter before abortion. In view of the low maternal concentrations of PGF compared with those in fetal fluids it appears that at least some of the PGF might be formed within the uterus (Silver et al. 1979).

Complete data sets of allantoic fluid were not obtained from all mares. Allantoic fluid catheters became obstructed by hippomane, which was identified at parturition in 2 instrumented mares. Allantoic fluid was not obtained by allantocensis in 4 mares in Period 4 because of length of time between sample collection, an abortion after the procedure conducted during Period 3, and an inability to detect imminent parturition in infected or control mares. Shortening the time from instrumentation to premature delivery would enable improved sample collection in instrumented mares. Lyle et al. (2009) inserted allantoic fluid catheters in standing ponies by laparoscopy and infected mares in utero 4–6 h later, with mares aborting after 8–18 days. In our study, mares were inoculated 7–10 days after surgery and aborted 6–27 days later.

One foal was born precociously mature on dGa 309, 20 days after inoculation (Fig 1). Our previous work showed that her dam's plasma progestagen concentrations rose steadily from 3.72 ng/ml at inoculation to a high of 30.5 ng/ml 48 h before delivery. In contrast, the mare that carried her foal for 27 days (Fig 1) delivered a nonviable foal on dGa 314 and did not exhibit a significant rise in progestagens. Concentrations at inoculation were 4.16 ng/ml, with levels fluctuating from 4.1 to 11 early in infection, decreasing to 5 ng/ml in mid and late infection only rising to 9.9 ng/ml in the last 3 days before delivery (Morris et al. 2007). Early work by Rossdale et al. (1991), supported by work of Ousey (Ousey et al. 2005) showed that naturally occurring placental disturbances in Thoroughbred mares 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 dGa 310 (Rossdale et al. 1991, 1995). Direct stimulation of the fetal adrenal with exogenous corticopin-releasing hormone (CRH) or adrenocorticopic hormone (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). Our findings indicate that placentitis and the ensuing immunological effects may have stimulated the equine fetal HPA axis as one foal was born precociously mature, her dam developed mammary secretions prematurely and had a significant rise in plasma progestagens after inoculation (Morris et al. 2007). The signalling pathways responsible for release of cortisol from the equine fetal adrenal glands subsequent to intrauterine infection are unknown but data from other species suggest that exposure to proinflammatory cytokines is one likely mechanism resulting in fetal HPA axis activation (Chrousos 1995; Turnbull and Rivier 1995). Lyle et al. (Lyle et al. 2009, 2010) recently reported that cortisol concentrations in equine fetal fluids increase before spontaneous abortion in an in utero model of fetal infection. Mares had a >25-fold expression change in IL-1β at the cervical star. They also showed that fetal equine adrenal glands collected at dGA 293 supplemented with ACTH, IL-1β or both produced cortisol (Lyle et al. 2009, 2010).

Bacteria were isolated from only 3 allantoic fluid samples collected in the 48 h preceding abortion. However, bacteria were isolated from the fetal stomach contents of all mares. This difference may be related to dilution of a small amount of bacteria within the large fluid volume of the allantoic space. Once bacteria enter the amnionic space, contaminated amnionic fluid enters the fetus passively through the oropharynx and oesophagus, resulting in colonisation of the fetal stomach and in some cases fetal lungs (Calderwood-Mays et al. 2002). Allantoic fluid concentrations of PGE2 and PGF rose before bacteria were isolated from allantoic fluids. Upregulation of proinflammatory cytokines may be more of a stimulus for these PGs than bacterial concentrations in the allantoic fluid. Three mares had mixed infections, a common finding in mares with placentitis (Calderwood-Mays et al. 2002; Bailey et al. 2010). Vaginal flora contains both Gram-negative and Gram-positive bacteria (Hinrichs et al. 1988). During ascending placentitis the cervix opens allowing pus and inflammatory by-products to drain out of the uterus into the vagina. Because of the mare's pelvic and perineal anatomy, vaginal fluids can reflux back into the cervix, enabling normal commensal inhabitants of the vagina to enter the uterus through a breach in the placental barrier.


The data support the theory that premature parturition associated with ascending placentitis is due to an abnormally regulated cytokine response to an infectious stimuli and prostaglandin formation in the mare.

Authors' declaration of interests

No conflicts of interest have been declared

Source of funding

This work was supported by the Dorothy Havemeyer Foundation, the State of Florida pari-mutuel wagering trust and the Grayson Jockey Club.


The authors thank Dr Kevin Anderson for technical assistance, Dr Maron Calderwood-Mays for histological interpretation of fetal membranes and Barbara Sheerin for performing prostaglandin assays.

Manufacturers' addresses

a Scientific Commodities Inc, Lake Havasu City, Arizona, USA.

b Ethicon, Somerville, New Jersey, USA.

c Marsam Pharmaceuticals, Inc, Cherry Hill, New Jersey, USA.

d Schering-Plough, Kenilworth, New Jersey, USA.

e DPT Laboratories, San Antonio, Texas, USA.

f NEN Life Science Products, Boston, Massachusetts, USA.

g Cayman Chemical Co, Ann Arbor, Michigan, USA.

h Qiagen Inc., Valencia, California, USA.

i Gibco BRL, Rockville, Maryland, USA.