|Home | About | Journals | Submit | Contact Us | Français|
Uterine microbial disease affects half of all dairy cattle after parturition, causing infertility by disrupting uterine and ovarian function. Infection with Escherichia coli, Arcanobacterium pyogenes and bovine herpesvirus 4 causes endometrial tissue damage. Toll-like receptors (TLRs) on endometrial cells detect pathogen-associated molecules such as bacterial DNA, lipids and lipopolysaccharide (LPS), leading to secretion of cytokines, chemokines and anti-microbial peptides. Chemokines attract neutrophils and macrophages to eliminate the bacteria, although persistence of neutrophils is associated with subclinical endometritis and infertility. Cows with uterine infections are less likely to ovulate because they have slower growth of the postpartum dominant follicle in the ovary, lower peripheral plasma estradiol concentrations, and perturbation of hypothalamic and pituitary function. The follicular fluid of animals with endometritis contains LPS, which is detected by the TLR4/CD14/LY96 (MD2) receptor complex on granulosa cells leading to lower aromatase expression and reduced estradiol secretion. If cows with uterine disease ovulate, the peripheral plasma concentrations of progesterone are lower than in normal animals. However luteal phases are often extended in animals with uterine disease, probably because infection switches the endometrial epithelial secretion of prostaglandins from the F to the E series, by a phospholipase A2 mediated mechanism, which would disrupt luteolysis. The regulation of endometrial immunity depends on steroid hormones, somatotrophins and local regulatory proteins. Advances in knowledge about infection and immunity in the female genital tract should be exploited to develop new therapeutics for uterine disease.
Microbial disease of the female genital tract is most common and of greatest economic importance in humans and cattle amongst the mammals [1, 2]. Microbial infections of the genital tract cause infertility by disrupting uterine and ovarian function. Many of the mechanisms underlying the recognition of microbial pathogens by the innate immune system in vertebrates have been identified during the last ten years [3-5]. These mechanisms of innate immunity are not only important for classical immune cells such as neutrophils and macrophages but are also evident in the endometrial and ovarian cells of mammals [6-8]. As well as causing an immune and inflammatory response, microbes or pathogen-associated molecules disrupt endocrine function in the female reproductive tract of rodents and cattle [6, 7, 9, 10]. Here we outline advances in scientific knowledge about how infection and innate immunity affect the female reproductive tract to cause infertility in cattle.
Uterine disease within a week of parturition (metritis) is present in up to 40% of dairy cows (Fig. 1). Metritis incidence depends on the definition of disease (see below) but maximal herd rates for obvious clinical disease of 36% and 50% have been reported in large surveys [16, 17] and 18.5% to 21% of animals have metritis with signs of systemic illness such as pyrexia [18, 19]. Subsequently, 15% to 20% of cattle have clinical disease that persists beyond 3 weeks post partum (endometritis) and about 30% have chronic inflammation of the uterus without clinical signs of uterine disease (subclinical endometritis) [2, 15, 20, 21].
Metritis occurs within 21 days and is most common within 10 days of parturition. Metritis is characterized by an enlarged uterus and a watery red-brown fluid to viscous off-white purulent uterine discharge, which often has a fetid odour . The severity of disease is categorised by the signs of health. We propose that cows are classified as having grade 1 metritis if they have an abnormally enlarged uterus and a purulent uterine discharge without any systemic signs of ill-health. Animals with additional signs of systemic illness such as decreased milk yield, dullness and fever >39.5 °C, are classified as having grade 2 clinical metritis. Animals with signs of toxaemia such as inappetance, cold extremities, depression and/or collapse are classified as grade 3 metritis, which has a poor prognosis.
Clinical endometritis is defined in cattle as the presence of a purulent uterine discharge detectable in the vagina 21 days or more post partum, or mucopurulent discharge detectable in the vagina after 26 days post partum . A simple grading system based on the character of the vaginal mucus (Fig 2A) is readily used to evaluate cows with clinical endometritis . The endometritis grade correlates with the presence of pathogenic organisms associated with uterine disease (Fig. 2B) and is prognostic for the likely outcome of treatment (Fig. 2C) [11, 22].
Subclinical endometritis is characterised by inflammation of the endometrium that results in a significant reduction in reproductive performance in the absence of signs of clinical endometritis. The inflammation is presumably associated with recovery of the tissues after clinical endometritis, trauma or other non-microbial disease. Subclinical disease is defined by polymorphonuclear neutrophils (PMNs) exceeding between 5.5% of cells  and 10% of cells  in samples collected by flushing the uterine lumen or by endometrial cytobrush, in the absence of clinical endometritis about 5 weeks post partum. The incidence of subclinical endometritis is dependent on the cut-off for diagnosis and the time after parturition but is in the order of 37 to 74% of animals (Fig. 1) .
The placenta should be expelled within a few hours of parturition in cattle. During the first week post partum, the uterus contracts rapidly, and lochia is discharged containing remnants of fetal membranes and fluids. During the second to fourth week, any damaged endometrial tissue regenerates, a wave of ovarian follicles develop, a dominant follicle is selected and estradiol secretion leads to ovulation and formation of a corpus luteum (CL) to recommence ovarian cycles . The genital tract should have little evidence of the previous pregnancy by six weeks after calving and be capable of establishing the next pregnancy. However, about 50% of dairy cows have irregular ovarian cycles during the postpartum period and animals with abnormal vaginal discharge are more likely than normal animals to have delayed resumption of ovarian cycles after calving (anovulatory anestrus; odds ratio 4.5), or prolonged postpartum luteal phases (odds ratio 4.4) . Conception rates are about 20% lower for cows with endometritis, the median calving to conception interval 30 days longer and there are 3% more animals culled for failure to conceive [20, 21]. Cows are less fertile even after successful treatment of clinical endometritis than age-matched counterparts in the same herds that had no clinical uterine disease post partum ; this is probably because subclinical endometritis persists after the clinical signs have resolved. Animals with subclinical disease also have more days open, take longer to conceive, and have conception rates about half that of normal animals .
The financial impact of uterine disease is derived from infertility, increased culling for failure to conceive, reduced milk production, and the cost of treatment. The economic cost of a single case of metritis has been calculated to be about €292 . There are 24,146,000 dairy cows in the European Union (EU)  and 8,495,000 in the United States (US) . Using a conservative incidence rate of 20% for metritis [18, 19], we therefore calculate the annual cost of uterine disease in the EU is €1.411 billion and in the US is $650 million. The costs of endometritis are an additional burden on the dairy industry and need to be quantified.
During pregnancy the uterus is sterile but after parturition the uterine lumen is almost always contaminated with a wide range of bacteria (Fig. 1). However, development of clinical disease is dependent on the balance between host immunity and the pathogenicity of the bacteria. This balance can be tipped in favour of disease by risk factors such as retained placenta, dystocia, twins, and stillbirth [29, 30]. Unfortunately, these factors are not particularly amenable to intervention to reduce the incidence of disease and the factors that could be addressed, such as the cleanliness of the animal or environment, are less important .
Escherichia coli and Arcanobacterium pyogenes are the most prevalent bacteria isolated from the uterine lumen of cattle with uterine disease, and then a range of anaerobic bacteria such as Prevotella spp., Fusobacterium necrophorum, and F. nucleatum. [10-14]. Bacteria are also isolated from the uterus of animals that do not develop clinical disease. Indeed, the presence of coagulase negative Staphylococci and α-haemolytic Streptococci decrease the risk of endometritis , so probiotics may be considered in the future for prevention of disease. Infection of the uterus with E. coli appears to pave the way for subsequent infection with other bacteria or viruses [32-34]. Furthermore, E. coli infection during the first days or week after parturition is associated with negative effects on the ovary, hypothalamic-pituitary axis and general health, as well as uterine disease . However, the most severe endometrial lesions are caused by A. pyogenes . The strains of A. pyogenes isolated from the uterus all express the virulence gene plo , which encodes a cholesterol dependent cytotoxin called pyolysin . Cholesterol dependent cytotoxin molecules are attracted to cholesterol-rich domains in cell membranes, where they aggregate to form a pore leading to osmotic death of the cell , and pyolysin readily kills endometrial epithelial and stromal cells in vitro . Furthermore, A. pyogenes, F. necrophorum and Prevotella species act synergistically to enhance the likelihood and severity of uterine disease [38, 39]. F. necrophorum produces a leukotoxin, P. melaninogenicus produces a substance that inhibits phagocytosis, and A. pyogenes produces a growth factor for F. necrophorum. Presumably the necrotic lochia associated with retained placenta provides an excellent media for bacteria. Trauma to the tissues during parturition also likely facilitates adhesion and invasion of microbes. Finally, suppressed or dysregulated immune mechanisms around the time of parturition (discussed below) probably also perturb host defence against microbes.
Bovine herpesvirus 4 (BoHV-4) is the only virus consistently associated with uterine disease after parturition in cattle [40, 41]. Like other herpesviruses, BoHV-4 can establish latent infections in cattle, particularly in macrophages , and the viral infection is often identified concurrent with bacteria that cause uterine disease [43, 44]. So, the association between BoHV-4 infection and uterine disease has been hard to establish, although the contribution of BoHV-4 to uterine disease where the virus is endemic in cattle will become clear when a vaccine is developed. The virus is highly tropic for endometrial cells, rapidly replicating and killing epithelial or stromal cells [42, 45]. BoHV-4 replication is driven by host cellular factors transactivating the viral immediate early (IE2, also known as UL122) gene promoter. A luciferase reporter for the UL122 promoter was transactivated in a concentration dependent manner when transfected bovine stromal cells were treated with PGE, E. coli or its lipopolysaccharide (LPS; endotoxin), and PGE and LPS acted co-operatively . Furthermore, viral replication was reactivated in latently infected macrophages when co-cultured with stromal cells . We suggest that there may be a vicious circle comprised of bacterial endometritis leading to secretion of PGE, then PGE and LPS stimulating viral replication, which causes further endometrial tissue damage and inflammation (Fig. 3). Identifying the specific host cellular transcription factors that transactivate the BoHV-4 UL122 gene to drive viral replication will inform strategies to prevent herpesvirus genital tract disease in cattle and other species.
Mammalian pregnancy involves regulation of uterine immunity to facilitate implantation and survival of the semi-allogeneic fetus. The classical view is that immunity is suppressed during gestation, although it is emerging that some immune and inflammatory mechanisms are also critical for implantation in mammals . If the immunosuppressive mechanisms associated with pregnancy persist in the endometrium after parturition they would likely predispose to uterine disease. So, some of the roles of the uterine immune system during pregnancy may be at odds with the need to respond in a coordinated way to pathogenic organisms in the uterus after parturition.
During mid and late pregnancy, lymphocytes and macrophages are found in the intercaruncular endometrium, although not in the caruncular endometrium of cattle [52-54]. The subepithelial uterine stroma contains more CD4+ T cells, B cells, CD14+ macrophages and mast cells compared to other regions of the endometrium and the myometrium [55, 56]. Mast cells have a prominent sensor and effector function during bacterial infections in mammals , but their role in response to intrauterine bacterial contamination in cattle is not clear. Specialized uterine NK cells (uNK) are important for normal pregnancy in many mammals but uNK cells are not common at the end of gestation ; and are sparse in the bovine endometrium . Whether these immune cells are still present in the tissues after labour in cattle and whether they regulate the inflammatory process after pathogen contact is also largely unknown.
The initial defence of the mammalian endometrium against microbes is dependent on innate immune systems including Toll-like receptors, antimicrobial peptides, and acute phase proteins . Bacteria are detected by pattern recognition receptors on mammalian cells binding molecules specific to microbial organisms, often called pathogen associated molecular patterns (PAMPs) [3-5]. The most important group of such receptors are the toll-like receptors (TLRs), and 10 members of the receptor family are widely encoded in the mammalian genome and most often found in a broad range of immune cells [3, 4]. TLR1, TLR2, and TLR6 recognise bacterial lipids such as lipoteichoic acid (LTA), whereas TLR3, TLR7, TLR8, and TLR9 recognize nucleic acids, often from viruses. Lipopolysaccharide from Gram-negative bacteria such as E. coli is bound to LPS-binding protein and recognised by TLR4 in complex with CD14 and LY96 (MD2); TLR5 binds flagellin; and TLR9 also recognises bacterial DNA. Activation of TLRs initiates signalling cascades resulting in the synthesis and production of pro-inflammatory cytokines and chemokines, which mobilise and activate immune cells [4, 5], which in the case of bovine uterine disease is particularly associated with the influx of PMNs into the uterus .
Whole endometrium from normal non-pregnant cattle expresses TLRs 1 to 10 . Before and after parturition TLR2, TLR3, TLR4, TLR6 and TLR9 are expressed in the caruncular and intercaruncular endometrium, and TLR expression was greater in the caruncular than intercaruncular endometrium 4 to 6 h post partum . Purified populations of epithelial cells express TLRs 1 to 7 and 9, and stromal cells express TLRs 1 to 4, 6, 7, 9 and 10 . These TLRs appear to be functional as epithelial cells secreted prostaglandin E2 (PGE) in response to bacterial PAMPs. Pure populations of epithelial or stromal cells, not contaminated with leukocytes as determined by the lack of expression of the protein tyrosine phosphatase, receptor type, C (PTPRC; formerly CD45) pan-leukocyte marker, express the specific receptor complex comprising TLR4/CD14/LY96 (MD2), to bind LPS [6, 9]. Heat-killed E. coli or LPS provoke an inflammatory response by the endometrial cells, characterised by increased expression of transcripts for tumour necrosis factor (TNF), nitric oxide synthase (NOS2), and prostaglandin-endoperoxide synthase 2 (PTGS2; formerly COX-2) and the secretion of prostaglandins F2α (PGF) and PGE . Heat killed E. coli, LPS, A. pyogenes pyolysin, BoHV-4, bacterial DNA and lipids also influence endometrial cell prostaglandin secretion – particularly stimulating the secretion of PGE rather than PGF in cattle [9, 45, 46, 63]. This may explain why animals with uterine infection have higher concentrations than normal animals of LPS and PGE in the uterine lumen and peripheral plasma [32, 64]. Endometrial explants, and epithelial and stromal cells also secreted predominantly PGE in response to LPS and this effect was not reversed by oxytocin . This LPS-induced PGE secretion by endometrial cells is important for fertility because prostaglandins have multiple roles in endometrial function, and luteolysis is initiated by PGF from oxytocin-stimulated epithelial cells . In addition, PGE plays an important role in the mammalian immune response, acting through prostaglandin E receptors 2 and 4 (PTGER2 and PTGER4) to control inflammation . The bovine endometrial cells express the PTGER2 and PTGER4 receptors necessary to respond to PGE [9, 67]. The endometrial prostaglandin switch induced by LPS appears to be early in the prostaglandin synthetic pathway. Arachadonic acid is liberated from cell membranes by phospholipase 2 group IV and VI enzymes (PLA2G4 and PLA2G6), and converted to prostaglandin H and then PGE or PGF by synthase enzymes . Treatment of endometrial cells with LPS stimulated increased levels of PLA2G6 but not PLA24C protein in epithelial cells but did not change the levels of PGE or PGF synthase enzymes .
The antimicrobial peptides (AMPs) are an ancient component of the immune system and the defensins family are particularly important for mucosal immunity . Bovine uterine tissue expresses lingual antimicrobial peptide (LAP), tracheal antimicrobial peptide (TAP), bovine neutrophil β-defensins (BNBD4, DEFB5), and bovine β-defensins (BBD19, BBD123 and BBD124) . Furthermore, pure populations of endometrial epithelial cells express LAP, TAP, BNBD4 and DEFB5, and expression was increased when cells were treated with LPS . Mucin 1 (MUC1) is an epithelial cell glycosylated transmembrane protein that may also have a role in microbial defence of the endometrium in mammals . MUC1 is expressed by epithelial cells of the bovine endometrium and expression was increased when the cells were treated with LPS . Acute phase proteins are produced in the liver in response to pro-inflammatory cytokines and peripheral plasma concentrations are increased during the first few weeks post partum in cattle . However, no acute phase proteins were detected in bovine endometrial cells in vitro .
Blood-derived PMNs are the main effector cells for removing bacteria from the uterus after calving. However, endocrine and metabolic changes around the time of parturition in cattle modulate PMN phagocytic function and gene expression [47, 73]. Further, blood PMNs obtained from cows with endometritis were significantly less phagocytic . The process of transmigration into the uterine lumen also modulates PMN function. For example, IL8-induced attraction of PMNs into the uterine lumen increased the generation of reactive oxygen species by these cells . However, when PMNs are in the uterine lumen their function is further modulated by soluble factors in lochial secretions. Whereas lochial secretions of healthy cows only moderately affected the function of PMNs, the secretions of infected cows severely depressed the generation of reactive oxygen species .
Changes in hormone concentrations around the time of parturition may influence the risk of peripartal infections . Progesterone and estrogen have immunomodulatory properties, changing the repertoire and expression density of hormone receptors in immune cells from cattle . In addition, estradiol and especially progesterone reduce the secretion of prostaglandins by epithelial or stromal cells stimulated with LPS . The somatotropic axis also influences the course of the bovine puerperium, mediated by changes in plasma and endometrial levels of insulin-like growth factor 1 (IGF1) [78, 79]. Indeed, IGF1 has immunomodulatory properties in addition to its growth-promoting function in mammals . Finally, there are several proteins found in the endometrium that could influence the immune response directly or affect the steroid or IGF1 pathways in endometrial cells. The uterine serpins are progesterone-induced members of the serpin superfamily of serine proteinase inhibitors and, at least in the sheep, inhibit lymphocyte proliferation to mediate the immunosuppressive effects of progesterone on uterine immune function . A family of glycan-binding proteins, the galectins, may also regulate uterine immunity by interacting with multiple galactose-ß1,4-N-acetylglucosamine units on cell surface glycoproteins [82, 83]. Lectin, galactoside-binding, soluble, 1 (galectin 1; LGALS1) controls mammalian cell proliferation, the survival of effector T-cells and neutrophils, and their extravasation in vivo [83-85]. One of the counter players of galectin 1 is lectin, galactoside-binding, soluble, 3 (galectin 3, LGALS3), which modulates the adhesion of T-cells to endothelial cells and the adhesion between T-cells and dendritic cells or macrophages . LGALS1 is expressed in the murine and human female reproductive tracts as well as by immune cells [87, 88]. In humans, LGALS1 expression is strongly enhanced in late phase endometrium and in the decidua , and LGALS1 is differentially expressed between normal and pathologically altered placentas [89, 90]. In cattle, LGALS3 is detected in the ovary, oviduct, uterus, and cervix and is postulated to be involved in mucosal defence . However, the role of galectins in postpartum uterine disease requires further exploration.
Cows with postpartum uterine infection, had slower growth of the first postpartum dominant follicle, lower peripheral plasma estradiol concentrations around the time of maximal follicle diameter, and in those animals that did ovulate, peripheral plasma progesterone concentrations were lower 5 to 7 days after ovulation (< 2 vs > 5 ng/ml) [10, 32]. These effects of uterine microbes on ovarian function could be caused by PAMPs or inflammatory mediators acting on the hypothalamus, pituitary or ovary.
Hypothalamic and pituitary function is critical for directing ovarian cycles. Follicle stimulating hormone (FSH) concentrations are not affected in animals with uterine disease, so follicle waves emerge in diseased as in normal animals . However, LPS suppresses hypothalamic release of GnRH, pituitary secretion of luteinising hormone (LH), and the sensitivity of the pituitary to gonadotrophin releasing hormone (GnRH) in sheep [49, 92]. The consequences of these changes would be that animals are less likely to ovulate, and this appears to be the case in cattle administered LPS . However, intrauterine infusion of a lower concentration of LPS in cattle did not disrupt LH secretion .
The follicular fluid of cattle with uterine inflammation also contains LPS . Animals with clinical disease had concentrations of LPS that ranged up to 0.8 μg/ml; normal animals did not have measurable concentrations of LPS in their ovarian follicular fluid, while animals with subclinical disease had intermediate concentrations about 40 to 60 days after calving. Theca cells convert cholesterol to androstenedione, which then passes across the basement membrane of the ovarian follicle and is converted to estradiol by the granulosa cells. Treatment of bovine theca cells from any stage of follicle development with LPS did not affect androstenedione production or cell survival, but granulosa cells collected from growing or dominant follicles secreted less estradiol when treated with LPS . As with endometrial cells, LPS does not affect theca or granulosa cell survival. The effect of LPS on bovine granulosa cells appears to be a direct one as the granulosa cell cultures were free of contaminating leukocytes , and the granulosa cell compartment within the basement membrane of the ovarian follicle is devoid of immune cells in vivo, at least in mice . Furthermore, granulosa cells from cattle express the TLR4/CD14/LY96 (MD2) complex required for binding LPS . Aromatase transcript expression was reduced by LPS treatment of granulosa cells collected from dominant follicles . So, granulosa cells have a mechanism for direct action of LPS in the ovarian follicle to impair ovarian function and ovulation. The effect of uterine disease on follicular function may be further enhanced by cyokines released by the endometrial cells because granulosa cell steroidogenesis is also impaired by pro-inflammatory cytokines . If animals ovulate, the cytokines secreted by the infected endometrium may also partly explain the reduced progesterone secretion from the CL because bovine luteal cells are highly responsive to a range of cytokines and cytokines are also important in luteolysis [50, 96].
The extended luteal phases in some cows with uterine disease could be associated with effects on luteolysis or on luteal cell function. Certainly the switch in endometrial prostaglandin to PGE from PGF could disrupt the luteolytic mechanism . In ruminants, PGE is luteotropic, whilst PGF is luteolytic . Using endometrial explants the ratio of PGE:PGF concentration was 0.45 in response to oxytocin and 2.75 following LPS treatment . Furthermore, giving oxytocin after treatment of endometrial cells with LPS did not reverse the propensity to secrete PGE .
In conclusion, uterine infections are common after parturition in dairy cattle, causing infertility. The working model that links the mechanisms of infection and immunity with infertility are presented in Figure 3. Bacterial infection with E. coli precedes infection with other microbes that disrupt endometrial structure and function. The innate immune system is alerted to the presence of pathogens by endometrial cell TLRs detecting pathogen-associated molecules (PAMPs) such as LPS, DNA and bacterial lipids. The endometrial cells secrete cytokines and chemokines to direct the immune response, and increase the expression of anti-microbial peptides (AMPs). Chemokines attract PMNs and macrophages to eliminate the bacteria, although neutrophil function is often perturbed in postpartum dairy cows. Persistence of PMNs in the endometrium in the absence of bacteria is thought to be the primary characteristic of subclinical endometritis. Uterine disease also affects ovarian function. Cows with uterine bacterial infections have slower growth of dominant follicles in the ovary and lower peripheral plasma estradiol concentrations and so are less likely to ovulate. The release of gonadotrophin releasing hormone from the hypothalamus and LH from the pituitary can also be suppressed by LPS – further reducing the ability to ovulate a dominant follicle. Follicular fluid contains LPS in animals with endometritis, granulosa cells express the TLR4/CD14/LY96 (MD2) complex required to detect LPS, and LPS reduces estradiol secretion. If cows with uterine infections ovulate, the peripheral plasma concentrations of progesterone are lower than in normal fertile animals and luteal phases are often extended. Luteolysis is probably disrupted because bacteria switch the endometrial epithelial secretion of prostaglandins from the F to the E series. The regulation of endometrial immunity depends on steroid hormones, somatotrophins and possibly local regulatory proteins such as galectins. Advances in knowledge about infection and immunity in the female genital tract should now be exploited to develop new treatments and prevention strategies for uterine disease.
1Supported by: Martin Sheldon is a BBSRC Research Development Fellow (Grant No. BB/D02028X/1). James Cronin is funded through a DEFRA LINK award by Pfizer Animal Health and BBSRC (Grant No. F005121).
Summary sentence: Postpartum uterine infections with bacteria and viruses in cattle cause infertility principally mediated through innate immune mechanisms in endometrial and ovarian follicle cells that drive inflammation and change endocrine function.