Review

Background

Nonspecific uterine infections reduce the reproductive efficiency of livestock [1–3]. They are called nonspecific because the initial colonizing bacterium is not known, and the specific bacteria causing the signs of infection are not known [2, 4–6]. Even though numerous bacteria, in a variety of combinations, have been isolated from infected uteri, Arcanobacterium pyogenes and Escherichia coli are usually associated with uterine infections in cattle and sheep [3–6].

Estimates of the annual incidence of uterine infections in postpartum animals in herds and flocks range from 10 to 50% of the dairy cattle [1, 2], 20 to 75% of the dairy buffaloes [7], and 5 to 10% of the dairy sheep [8]. Published estimates of the incidence of analogous uterine infections, as opposed to the mastitis, metritis, agalactica syndrome, in postpartum pigs are not available. However, swine herd managers and veterinarians often state that they are "common" (personal communications). Despite the published estimates for cattle, buffaloes, and sheep, the "true" incidence of uterine infections is not known for any livestock species because detection and diagnosis are often inaccurate and reporting is not mandatory.

In the United States, intramuscular (i.m.) injections of prostaglandin F2α (PGF2α) have become the preferred treatment for pyometra in cattle [2]. Pyometra is the most common type of uterine infection in dairy cattle, and it is the one most often associated with reduced reproductive performance [2]. Pyometra typically develops just after the onset of luteal function during the postpartum period [2]. Apparently, bacteria that are normally found in livestock environments are introduced during or after calving, particularly when cows with dystocia are assisted and when retained fetal membranes are removed manually [2]. In cows that later develop uterine infections, the bacteria seem to reside in the uterus, without proliferating into an infection, until luteal progesterone down-regulates immune functions. Then the bacteria proliferate and create a pathological condition. The infections usually persist until luteolysis. The resultant reduction in progesterone concentrations seems to permit the uterine immune system to up-regulate and resolve the infection. Thus, exogenous PGF2α is used to induce luteolysis and clear uterine infections, although the true mechanism of action of PGF2α in eliminating uterine infections is not known. In fact, PGF2α seems to have effects that are independent of its effects on corpus luteum function.

Even though exogenous PGF2α has become the preferred treatment, intrauterine and systemic antibiotic treatments are common in the United States and in other countries [2, 9]. General concerns about the relationship between antibiotic use in livestock and the potential for accelerating the evolution of antibiotic-resistant strains of bacteria have aimed our research at determining whether nonantibiotic, native compounds can be used to enhance host immunity and prevent or resolve uterine infections. Because of the pivotal role of progesterone in converting the uterus from resistant to susceptible to infections, understanding the mechanism of action of progesterone is essential for developing methods for enhancing the ability of the uterus to manage pathogenic bacteria. Thus, the purpose of this article is to briefly review the role of progesterone in making the uterus susceptible to infections and describe how eicosanoids may be used to override the immunosuppressive effects of progesterone.

Progesterone

Data from numerous studies that have been published over the last 50 years indicate that cattle, sheep, and pigs are resistant to uterine infections when progesterone concentrations are basal, and they are susceptible when progesterone concentrations are increased [10–15]. For example, spontaneous uterine infections in dairy cows do not usually develop until after formation of the first postpartum corpus luteum, although bacterial contamination can be sufficient to induce the onset of puerperal metritis very soon after calving when progesterone concentrations are basal [1, 2, 15]. Postpartum beef cows that received intrauterine infusions of A. pyogenes and E. coli when progesterone concentrations were basal did not develop uterine infections, whereas all cows developed uterine infections when the bacteria were infused after the onset of luteal function and progesterone concentrations had begun to increase [6]. The same was true for postpartum ewes (i.e., female sheep) with spontaneous onset of luteal function and for postpartum ewes that had been ovariectomized and treated i.m. with progesterone [16, 17]. In addition, none of the ewes or gilts (i.e., nulliparous pigs) that received intrauterine infusions of A. pyogenes and E. coli during estrus developed uterine infections, but all of the ewes and gilts that received A. pyogenes and E. coli infusions during the luteal phase of the estrous cycle developed uterine infections [18–20].

Lymphocyte proliferation in vitro has been used as a general measure of immune responsiveness, even though neutrophils, which are much more difficult to evaluate, probably have a greater role in the initial response to bacteria that enter the uterus [21–25]. Unstimulated, concanavalin A (Con A)-stimulated (stimulates T-cells), and lipopolysaccharide (LPS)-stimulated (stimulates B-cells) proliferation were greater for lymphocytes collected from postpartum ewes that had been ovariectomized before the onset of luteal function than was proliferation of lymphocytes collected from ewes with spontaneous luteal function and increased progesterone [17]. (In contexts, such as in the previous sentence, greater, increased, decreased, etc. refers to an event with P < 0.05). Unstimulated and Con-A stimulated lymphocyte proliferation were greater for sesame oil-treated than for progesterone-treated postpartum ewes [17]. Unstimulated, Con A-stimulated, and LPS-stimulated proliferation were greater for lymphocytes collected from estrous ewes than from luteal phase ewes, and unstimulated and Con A-stimulated lymphocyte proliferation were greater for estrous gilts than for luteal phase gilts [18, 20]. Indeed, several authors have reported that progesterone suppresses lymphocyte proliferation [26–29], and this effect of progesterone was associated with the inability of the uterus to prevent the development of infections.

The previous examples clearly support the idea that progesterone converts the uterus from an organ that is resistant to one that is susceptible to infections. Data from the studies selected for the examples, as well as reports from other authors, also support the idea that, when progesterone concentrations are basal and bacterial contamination is not severe enough to overwhelm uterine defenses (i.e., puerperal metritis is not an issue), the uterus "defaults" to resistant to infections. We have referred to this as a "protected period," although we acknowledge that some threshold, albeit unknown, load of bacteria will overwhelm uterine defenses. Basal progesterone and basal estradiol concentrations are two characteristics of the protected period, and we assume that "basal" immune responses are usually able to prevent bacteria from proliferating into a pathological condition. We have used postpartum and seasonally anestrous ewes to determine whether the uterus is, in fact, protected.

In two studies, none of the autumn-lambing ewes that were ovariectomized on day 9 or day 14 postpartum, before spontaneous onset of luteal function and increases in progesterone concentrations, and treated i.m. with canola oil or safflower oil developed infections after intrauterine infusion of A. pyogenes and E. coli [16, 17]. By contrast, all of the ewes in the two studies with spontaneous luteal function and all of the ewes treated i.m. with progesterone in canola oil or safflower oil developed infections after intrauterine infusion of A. pyogenes and E. coli [16, 17]. For the ewes ovariectomized on day 14, PGF2α concentrations in vena caval blood were greatest in ewes with the least vena caval progesterone, and PGF2α was least in the ewes with the greatest progesterone [17]. Vena caval blood was collected through catheters that were positioned just cranial to the entry of uteroovarian blood [30]. Indeed, 6α-methyl-17α-hydroxyprogesterone acetate (a progestogen commonly used to control the estrous cycle) reduced uterine secretion of PGF2α [31]. Ovariectomy increased and exogenous progesterone decreased Con A-stimulated lymphocyte proliferation, and ovariectomy increased LPS-stimulated lymphocyte proliferation [17]. The results were less clear for the ewes ovariectomized on day 9, but exogenous progesterone seemed to have some suppressive effects on lymphocyte proliferation [16]. Overall, suppressed PGF2α secretion, suppressed lymphocyte proliferation in vitro, and increased progesterone were associated with the inability of the uterus to prevent infections. When progesterone concentrations were basal, the "default" immune status of the ewes was adequate to prevent uterine infections.

In another study, seasonally anestrous ewes were used to determine whether, in the long-term absence of ovarian progesterone, the uterus was resistant to infections [32]. The ewes in this study had not been detected in estrus for at least three months, had basal endogenous progesterone concentrations, and had no ovarian follicles large enough to be estrogen active. None of the control ewes developed infections after intrauterine infusion of A. pyogenes and E. coli, whereas all of the progesterone-treated ewes developed infections after intrauterine A. pyogenes and E. coli infusions [32]. Bacteria were infused into the uterus two days before commencement of progesterone injections to determine whether the uterus would eliminate the bacteria soon after they were introduced. The bacteria were apparently able to survive in the uterus until exogenous progesterone prevented the uterus from suppressing their proliferation. Results from these studies with postpartum and seasonally anestrous ewes support the hypothesis that, when progesterone and estradiol are basal, the uterus is resistant to infectious bacteria, although the uterus does not seem to eliminate bacteria as soon as they arrive. A direct effect of estradiol on the resistance or susceptibility of the uterus in cattle, sheep, and pigs to infections has not been established, but authors usually acknowledge the temporal associations between reductions in progesterone, increases in estradiol, and resistance to uterine infections [18, 20]. Thus, because the role of estradiol has not been established and because progesterone unequivocally down-regulates uterine immune functions, progesterone seems to be the ovarian steroid that primarily determines whether the uterus can prevent bacteria from proliferating into an infection.

Progesterone and Eicosanoids

Eicosanoids, including prostaglandins and leukotrienes, belong to a family of compounds that are synthesized from arachidonic acid through the cyclooxygenase and lipoxygenase pathways [33, 34]. The family is large; so only selected eicosanoids will be discussed.

Peripheral progesterone and 13,14-dihydro-15-keto-PGF2α (PGFM) concentrations were quantified in postpartum dairy cows to characterize the relationships among progesterone, PGFM, and onset of uterine infections [15, 35, 36]. 13,14-Dihydro-15-keto-PGF2α is a metabolite of PGF2α with a half-life of approximately 15 min, compared with approximately 1 min for PGF2α [37, 38]. During the postpartum period, jugular concentrations of PGFM are a close reflection of uterine secretion of PGF2α; however, that does not seem to be the case during the estrous cycle, when the uterus produces considerably less PGF2α [31, 37, 39]. The results of studies with postpartum dairy cows indicated that PGFM concentrations were less in cows that later developed uterine infections than they were in cows that did not develop uterine infections [15, 36]. Only small increases in progesterone, presumably luteal, were necessary to initiate the onset of uterine infections [15], and PGFM concentrations increased at the onset of uterine infections [35]. The increase in PGFM most likely reflected a uterine inflammatory response to the proliferation of bacteria and release of endotoxin [40, 41].

One interpretation of the results of the studies with dairy cows and sheep is that uterine PGF2α production is related to the ability of the uterus to prevent or resolve infections. Moreover, one may speculate that progesterone-PGF2α interactions modulate the ability of the uterus to prevent infections. Stated more directly, progesterone can down-regulate immune functions and prevent the uterus from resisting infections, whereas PGF2α, and probably other eicosanoids such as leukotriene B4 (LTB4), can up-regulate immune functions and override the effects of progesterone. In fact, in vitro experiments indicate that PGF2α, LTB4, 5-hydroxyeicosatetraenoic acid, 15-hydroxyeicosatetraenoic acid, and lipoxin B4 are chemoattractant to neutrophils [42]. Neutrophils are thought to be the initial defense against pathogens that enter the uterus, and suppressed neutrophil functions are thought to predispose the uterus to infections [22, 23, 25]. Prostaglandin E2 can suppress a number of immune functions, including neutrophil functions, and predispose cows to uterine infections [24]. Progesterone-eicosanoid interactions are clearly broader than just PGF2α interactions. However, because of the large number of compounds that can interact with progesterone and affect immune functions, our research has focused on progesterone-PGF2α interactions.

Even though exogenous PGF2α will induce luteolysis and allow the uterus to clear infections, authors have mentioned that exogenous PGF2α helped resolve uterine infections in cows that did not have any evidence of luteal function [6]. Clinical veterinarians often speculate that exogenous PGF2α stimulates uterine contractions that expel the bacteria (personal communications). But this ignores the issue that bacteria are likely to remain in the uterus and will be present when luteal function is reinitiated, and it ignores the fact that literature is not available to substantiate the idea that uterine contractions cleanse the uterus. Indeed, a study with mares indicates that uterine contractions may reduce the fluid associated with uterine infections, but uterine contractions do no seem to eliminate the bacteria [43]. However, one can reason, based on a good deal of literature, that exogenous PGF2α has direct effects on uterine immune functions.

A key point, from a mechanism-of-action perspective, is that the direct effects of exogenous PGF2α on uterine immune functions and the effects of PGF2α on luteal function and progesterone concentrations are completely confounded. In other words, is exogenous PGF2α effective because it stimulates immune functions, or is exogenous PGF2α effective because it removes progesterone? We have conducted studies with sheep and pigs to address these questions.

Prostaglandin F2α is not luteolytic in pigs until after approximately day 12 of the estrous cycle [44]. We used this information to determine whether PGF2α had direct effects on uterine immune functions. Sows (i.e., parous pigs) were assigned to a 2 × 2 factorial array of treatments (n = 6 sows/group); intrauterine infusion of A. pyogenes and E. coli (i.e., bacteria vs. phosphate-buffered saline [PBS]) and PGF2α (i.e., 10 mg of PGF2α vs. saline) were the two main effects. Bacteria or PBS was infused on day 7 of the estrous cycle, and PGF2α or saline was injected i.m. once on day 9 of the estrous cycle. Uteri were collected on day 11 of the same estrous cycle. Vena caval blood was collected throughout the experiment. Vena caval progesterone concentrations did not differ among groups and averaged 64 ng/mL during the study, which indicates that PGF2α did not affect luteal function. Vena caval estradiol-17β concentrations did not differ among groups and averaged 1 ng/mL, which is a typical luteal phase concentration. The PGF2α injection increased vena caval PGF2α concentrations, and this is consistent with the results of studies with sheep in which exogenous PGF2α stimulated uterine secretion of PGF2α [37]. Exogenous PGF2α also enhanced unstimulated and LPS-stimulated lymphocyte proliferation, and it increased vena caval PGE2 concentrations. None of the PBS-treated sows developed uterine infections, but all of the bacteria-treated sows developed uterine infections. Based on the volume of sediment, which is composed of leucocytes and bacteria, in the uterine flushings collected on day 11 and the ability to culture of A. pyogenes and E. coli from the flushings, the PGF2α-treated sows were resolving the uterine infections (sediment was 70% of total flushing volume for bacteria-saline, 30% for bacteria-PGF2α, and <5% for PBS-saline and PBS-PGF2α sows). Based on these results, PGF2α treatment allowed the uterus to begin resolving the infections, even though progesterone remained at luteal phase concentrations and estradiol concentrations were basal.

Because PGF2α is luteolytic in sheep after approximately day 4 of the estrous cycle, we used ovariectomized, progesterone-treated ewes to test the hypothesis that PGF2α has effects on uterine immunity that are independent of progesterone concentrations. Ewes were assigned to a 2 × 2 × 2 factorial array of treatments (n = 8 ewes/group); ovariectomy (i.e., ovariectomy vs. sham procedure), progesterone (5 mg of progesterone at 12-hour intervals vs. sesame oil diluent at the same times), and PGF2α (15 mg of PGF2α vs. saline) were main effects. On the day of estrus (i.e., day 0), ewes were either ovariectomized or a sham procedure was performed. Either progesterone or sesame oil was injected i.m. from day 0 through day 11. Catheters were positioned in the vena cava on day 5, and vena caval blood was collected from then until the end of the experiment. On day 6, all ewes received intrauterine infusions of A. pyogenes and E. coli. Prostaglandin F2α or saline was injected i.m. on day 9, and uteri were collected on day 12. Vena caval progesterone concentrations in all eight groups of ewes behaved as anticipated. For example, progesterone concentrations in sham ovariectomy-sesame oil-saline ewes were indicative of a "normal" luteal phase. Prostaglandin F2α induced luteolysis and reduced progesterone concentrations to basal values in ewes that did not receive exogenous progesterone. Ovariectomy reduced progesterone to basal concentrations, and exogenous progesterone maintained or increased (i.e., in sham ovariectomized, progesterone-treated ewes) progesterone concentrations. As expected, exogenous PGF2α increased vena caval PGF2α concentrations. All of the sham-ovariectomy ewes developed uterine infections. However, based on sediment volume and ability to culture A. pyogenes and E. coli from the uterine flushings, the sham-oil-saline ewes and the sham-oil-PGF2α ewes were resolving (sediment volume approximately 8%) the infections by day 12. The sham-progesterone-saline ewes had severe uterine infections on day 12 (sediment volume of 28%, which is considerably greater than usually seen for sheep with uterine infections), but the sham-progesterone-PGF2α ewes were resolving the infections on day 12 (sediment volume of approximately 15% vs. the 28%). The ovariectomy-oil-saline ewes did not have uterine infections on day 12 (sediment volume of 5%), but the ovariectomy-progesterone-saline ewes had typical infections (sediment volume of approximately 16%) on day 12. The ovariectomy-oil-PGF2α ewes did not have uterine infections on day 12 (sediment volume of approximately 2%), and the ovariectomy-progesterone-PGF2α ewes had nearly resolved the infections (sediment volume of approximately 4%) by day 12. Progesterone reduced unstimulated lymphocyte proliferation, whereas exogenous PGF2α increased unstimulated, Con A-stimulated, and LPS-stimulated lymphocyte proliferation. The results of this experiment indicate that exogenous PGF2α enhances the ability of the uterus to resolve infections, regardless of progesterone concentrations. The ewes in this experiment and the sows in the other experiment discussed received one injection of PGF2α, and this was enough to begin resolving or to nearly resolve the uterine infections. We have speculated that a second injection of PGF2α 12 to 24 hours after the first would resolve all of the induced infections, but that possibility has not yet been evaluated experimentally.

Based on current literature, the ability of the uterus to resist or to resolve uterine infections seems to be related to the ability of the uterus to secrete PGF2α. Exogenous PGF2α enables the uterus to resolve uterine infections, even when circulating progesterone concentrations are maintained at luteal phase concentrations or greater. Thus, PGF2α can up-regulate immune functions and override the down-regulatory effects of progesterone.

Mechanisms

Progesterone and eicosanoids have a variety of independent effects on immune cell functions. The negative effects of progesterone on immune functions have been reviewed extensively, especially the effects of progesterone on the synthesis of immunosuppressants and blocking factors, and will not be reviewed in this article [26–28]. Instead, this section will focus on relationships between progesterone and eicosanoids, because manipulating this relationship at the cellular or molecular level seems to have great potential for preventing or resolving uterine infections in livestock-production settings.

During estrus, when progesterone concentrations are decreased and estradiol concentrations are increased, uterine production of PGF2α is increased, endometrial LTB4 production is increased, and the uterus is normally able to prevent infections from developing [45–48]. Uterine PGF2α and LTB4 production decrease to basal within a few days after estrus, when progesterone concentrations begin to increase and the uterus again becomes susceptible to infections [46, 47, 49]. Prostaglandin F2α enhanced neutrophil chemotaxis and the ability of neutrophils to ingest bacteria, and LTB4 enhanced chemotaxis, random migration, and antibody-independent cell-mediated cytotoxicity [42]. Prostaglandin F2α, which is considered a proinflammatory molecule, may stimulate the production of proinflammatory cytokines that enhance phagocytosis and lymphocyte functions [19, 50]. Leukotriene B4 may promote uterine involution and reduce the risk of uterine infections in cows [49]. Exogenous PGF2α increases uterine secretion of PGF2α and luteal production of LTB4 [37, 51]. Because definitive data are not available, one can only speculate that PGF2α enhances uterine LTB4 production. However, nordihydroguaiaretic acid, which inhibits lipoxygenase activity and LTB4 production, prolonged the luteal phase in cattle and sheep, and the uterus seems to have mediated the effect [52, 53]. The increase in uterine PGF2α production after exogenous PGF2α probably increases phospholipase A2 (PLA2) and cyclooxygenase 2 activities [54–56]. Increased PLA2 would increase the amount of free arachidonic acid that could be used to produce cyclooxygenase (e.g., PGF2α and PGE2) and lipoxygenase (e.g., LTB4) products. In addition, tumor-necrosis factor α (TNFα), which mediates inflammatory and cytotoxic responses [40], stimulates endometrial PGF2α production, and PLA2 seems to mediate this effect [57, 58]. Therefore, enhancing uterine secretion of PGF2α and LTB4 should up-regulate immune functions and enable the uterus to prevent or resolve infections. Exogenous fenprostalene (i.e., a long-acting PGF2α analogue), which should increase uterine PGF2α secretion, injected subcutaneously once sometime between days 7 and 10 postpartum reduced the incidence of endometritis in cows with dystocia and/or retained fetal membranes [36]. A single subcutaneous fenprostalene injection on the day endometritis was detected (i.e., sometime between days 14 and 28 postpartum) reduced the interval from parturition to conception in dairy cows [36]. However, studies to define the mechanisms of action of PGF2α in livestock have not been reported. This seems to be a fruitful area of study with considerable benefit to the livestock industries.

If the assertion that enhancing uterine secretion of PGF2α and LTB4 will enable the uterus to resolve infections is correct, it means that PGF2α and LTB4 must override the inhibitory effects of progesterone. In addition to its ability to stimulate the production of immunosuppressants and blocking factors, progesterone can stimulate prostaglandin E synthase activity and decrease the activity of a number of proinflammatory molecules [26–28, 59, 60]. For example, progesterone suppresses the production of IL-8, which promotes chemotaxis, superoxide release, and granule release from phagocytic cells, in reproductive tissues [61–65]. It may also suppress the production of IL-6, which promotes B-cell differentiation and production of acute-phase proteins [66]. In addition, progesterone inhibits the production of IL-12, which induces interferon-γ production and enhances natural killer-cell cytotoxicity, and PLA2, presumably via increased free arachidonic acid, seems to mediate the effects of IL-12 [29]. Even though progesterone has a variety of inhibitory effects, they are not always consistent among reports, much of the research has only been conducted with in vitro systems, and very little of the research has been focused on the relationship between production of various proinflammatory molecules and the ability of the uterus in livestock to resistant or resolve infections.

Conclusions

Progesterone seems to be the primary ovarian steroid that governs the ability of the uterus to resist infections. In livestock, progesterone typically down-regulates immune functions and makes the uterus susceptible to infections. Exogenous PGF2α increases uterine secretion of PGF2α, and probably LTB4, and these two eicosanoids are associated with enhanced uterine immune responses and resistance to infections. Even though progesterone and eicosanoids affect a variety of proinflammatory molecules that can alter uterine immune responses, the role of these molecules in determining whether the uterus in livestock is resistant or susceptible to infections has not been elucidated. Indeed, determining how uterine PGF2α is able to stimulate the uterus to resolve infections, even when progesterone concentrations are increased, should be important to scientists and clinicians working to understand the underlying causes of uterine infections. Information from this line of research should yield important new prevention and treatment strategies for uterine infections that do not rely on antibiotic and antimicrobial compounds.