Abstract
We have suggested in a previous in vitro study that tumor necrosis factor-α (TNFα) plays a role in the initiation of luteolysis in cattle. The aim of the present study was to examine the influence of different doses of TNFα on the estrous cycle in cattle by observing the standing behavior and measuring peripheral concentrations of progesterone (P4) during the estrous cycle. Moreover, we evaluated the secretion of P4, oxytocin (OT), nitric oxide (NO), and luteolytic (prostaglandin F2α [PGF2α] and leukotriene C4 [LTC4]) and luteotropic (PGE2) metabolites of arachidonic acid in peripheral blood plasma as parameters of TNFα actions. Mature Holstein/Polish black and white heifers (n = 36) were treated on Day 14 of the estrous cycle (Day 0 = estrus) by infusion into the aorta abdominalis of saline (n = 8), an analogue of PGF2α (cloprostenol, 100 μg; n = 3) or saline with TNFα at doses of 0.1 (n = 3), 1 (n = 8), 10 (n = 8), 25 (n = 3), or 50 μg (n = 3) per animal. Peripheral blood samples were collected frequently before, during, and up to 4 h after TNFα treatment. After Day 15 of the estrous cycle, blood was collected once daily until Day 22 following the first estrus. Lower doses of TNFα (0.1 and 1 μg) decreased the P4 level during the estrous cycle and consequently resulted in shortening of the estrous cycle (18.8 ± 0.9 and 18.0 ± 0.7 days, respectively) compared with the control (22.3 ± 0.3 days, P < 0.05). One microgram of TNFα increased the PGF2α (P < 0.001) and NO (P < 0.001) concentrations and decreased OT secretion (P < 0.01). Higher doses of TNFα (10, 25, 50 μg) stimulated synthesis of P4 (P < 0.001) and PGE2 (P < 0.001), inhibited LTC4 secreton (P < 0.05), and consequently resulted in prolongation of the estrous cycle (throughout 30 days, P < 0.05). Altogether, the results suggest that low concentrations of TNFα cause luteolysis, whereas high concentrations of TNFα activate corpus luteum function and prolong the estrous cycle in cattle.
Introduction
The mechanisms controlling the development, secretion, and regression of the bovine corpus luteum (CL) involve many factors produced both inside and outside the CL [1–3]. Prostaglandins (PGs), along with sex steroids, are among the most important factors involved in the regulation of the estrous cycle and pregnancy [3–5]. Progesterone (P4) produced by the CL is crucial for determining the duration of the estrous cycle and for achieving a successful pregnancy [1, 2]. PGF2α released from the uterus has been shown to cause regression of the ruminant CL, and the administration of PGF2α analogues terminates pregnancy in ruminants [4, 6]. However, the current concept that the regression of CL in cattle is directly brought about by PGF2α of uterine origin secreted in response to oxytocin (OT) released by the CL is inadequate to explain many events that actually occur at the time of regression [3–8].
In our previous study, although both OT and tumor necrosis factor-α (TNFα) affected endometrial PGF2α output at the follicular stage, only TNFα affected PGF2α output at the mid and late luteal stages by acting through specific binding sites present in the bovine endometrium during the whole estrous cycle [9]. TNFα stimulated PGF2α production by the endometrial stromal cells via activation of phospholipase A2 (PLA2) and nitric oxide (NO) synthase [10]. We recently showed that NO is a good candidate to serve as a mediator and/or modulator of PGF2α action on the bovine CL during luteolysis in cattle [11–13]. NO directly inhibits P4 secretion from bovine luteal cells and augments the action of extragonadal PGF2α on the CL [11]. The inhibition of NO production in the female reproductive tract counteracts both spontaneous [14] and PGF2α-induced [12, 13] luteolysis in cattle. Moreover, NO stimulates the secretion of leukotriene C4 (LTC4) in the bovine CL [13]. Several products of the lipoxygenase pathway of the arachidonic acid cascade, particularly leukotriene B4 and LTC4, have been demonstrated to play roles in luteolysis [15, 16]. An in vivo microdialysis study showed that LTC4 concentrations in perfusates from the CL rose before the decline in P4 during spontaneous luteolysis in heifers [17].
Complete structural regression of the bovine CL involves the action of products of immune cells [18–22]. At the time of luteolysis, immune cells invade into the bovine CL [18]. Cytokines produced by the immune cells, especially TNFα and interferon-γ (IFNγ), seem to participate in the regression of the bovine CL [19]. TNFα and its specific receptors (TNF-RI) are present in the bovine CL during luteolysis [20, 21]. TNFα in combination with IFNγ reduces P4 production and induces the apoptotic events that finally lead to functional and structural luteolysis [8, 21–23].
Unexpectedly, TNFα also induces the production and output of luteotropic PGE2 [2, 24] in cultured bovine luteal [20] and stromal-endometrial cells and tissue [25], suggesting luteotropic and luteoprotective roles of TNFα. After fertilization, a relatively high proportion of luteotropic PGE2 relative to luteolytic PGF2α (PGE2 > PGF2α) is needed for proper embryonic development, recognition, and establishment of pregnancy [4, 5, 25, 26]. Thus, TNFα seems to play one or more roles in the regulation of the estrous cycle and pregnancy in cattle. However, the properties and mechanisms of the opposite TNFα actions (luteolytic versus luteotropic) during the estrous cycle have not been yet clarified. In the present study, therefore, we examined the influence of TNFα on the estrous cycle in conscious cattle by observing standing behavior and measuring peripheral concentrations of P4 during the estrous cycle. We evaluated the concentrations of P4, OT, stable metabolites of NO (nitrate/nitrite [NO2−/NO3−]), and luteolytic (PGF2α and LTC4) and luteotropic (PGE2) metabolites of arachidonic acid in the blood samples as indicators of TNFα action.
Materials and Methods
All animal procedures were approved by the Local Animal Care and Use Committee (agreement No. 4/2001/N).
Animals and Surgical Procedures
Normally cycling Holstein/Polish black and white (75% and 25%, respectively) heifers (18–20 months of age and final 400–450 kg body weight, n = 36) were used for the present study. Two to three weeks after weighing and choosing the animals for experiments (only animals weighing 390–400 kg were chosen), the estrus was synchronized using implants of a progesterone analogue (Crestar; Intervet, Boxmeer, Holland). Crestar consists of two components: a silicone ear implant containing the progestogen norgestomet (17α-acetoxy-11-β-methyl-19-norpregna-4-en-2.20-dione) and an injection containing norgestomet (3 mg/2 ml) with estradiol valerate (5 mg/2 ml). The estrus was synchronized using the standard procedure without any additional treatment with the analogue of PGF2α treatment, as recommended by the manufacturer for the estrus synchronization of maiden heifers. The Crestar injection was administered i.m., and at the same time the Crestar implant was inserted s.c. at the outer edge of the ear in all animals. After 10 days, the norgestomet implants were removed. The onset of estrus was confirmed by standing behavior and taken as Day 0 of the estrous cycle. The length of the estrous cycle was defined as the number of days between the first days of standing estrus.
For infusion of either saline, an analogue of PGF2α (aPGF2α, cloprostenol; Bioestrophan, Biowet, Gorzow Wielkopolski, Poland), or TNFα (recombinant human TNFα; HF-13; kindly donated by Dainippon Pharmaceutical Co., Ltd., Osaka, Japan) on Day 13 of the subsequent estrous cycle, a catheter was inserted into the posterior aorta abdominalis through the coccygeal artery, as described previously [12]. The animals were premedicated with i.m. xylazine at a dose of 25–30 mg per animal (Sedazin; Biowet, Pulawy, Poland), and local epidural anesthesia was induced by injecting 4 ml of 2% procaine hydrochloride (Polocainum Hydrochloricum; Biowet, Drwalew, Poland) between the first and second coccygeal vertebrae. The tip of the cannula was positioned in the aorta 65–70 cm ahead of the point of insertion, just cranial to the origin of the ovarian artery and caudal to the renal artery [12]. This placement allowed infused reagents to be transported by the bloodstream directly into the reproductive tract as was established in a previous study [27]. It has been demonstrated using adrenergic drugs [27, 28] that by means of such implanted cannula one can administer drugs into the female reproductive tract more specifically than by administration into the general circulation and that the required dose of experimental factors may be thereby markedly reduced. A second catheter was inserted into the jugular vein for frequent collection of blood samples.
Experiment 1 (Preliminary Study)
Twenty-one heifers were used to establish the effective dose of TNFα. On Day 14 of the estrous cycle, three heifers received the infusion of 20 ml of saline in the aorta abdominalis throughout a period of 30 min (control group). Another three animals were injected with 100 μg of cloprostenol (aPGF2α, the dose was chosen based on our previous data [6, 12]). The remaining 15 heifers were infused for 30 min with 0.1, 1, 10, 25, or 50 μg of TNFα (n =3 per dose) into the aorta abdominalis. Peripheral blood samples were collected from the jugular vein at 10-min intervals (beginning 1 h before and continuing until 3.5 h after the infusion). After Day 15 of the estrous cycle, blood was collected once daily until Day 22 following the first estrus. The blood plasma was separated by centrifugation (2000 × g, 10 min at 4°C) and stored at −20°C until P4 determinations were made.
Experiment 2 (Influence of TNFα on the Estrous Cycle)
To test the hypothesis that TNFα differentially affects the duration of the estrous cycle and the secretory function of the CL and uterus, saline as a control (n = 5) and a possible luteolytic dose of TNFα (1 μg, n = 5) or a possible luteotropic dose of TNFα (10 μg, n = 5) in 20 ml of saline was infused throughout a period of 30 min into the aorta abdominalis on Day 14 of the estrous cycle. Peripheral blood samples were collected from the jugular vein at 10-min intervals (beginning 1 h before and continuing until 3.5 h after the infusion). After Day 15 of the estrous cycle, blood was collected once daily until Day 22 following the first estrus. The concentrations of P4, OT, PGE2, 3,14-dihydro,15-keto-PGF2α (PGFM), LTC4, and nitrite/nitrate in the plasma samples were measured. Standing behavior was checked every 12 h after TNFα or saline treatment to confirm the onset of estrus.
Hormone Determinations
Progesterone concentrations in plasma samples were assayed using a direct enzyme immunoassay (EIA) as described previously [12]. The P4 standard curve was produced for P4 concentrations ranging from 0.39 to 25 ng/ml, and the effective dose for 50% inhibition (ID50) in the assay was 2.85 ng/ml. The intra-assay and interassay coefficients of variation averaged 6.6% and 8.4%, respectively.
The EIA for OT was based on the second antibody method using the biotin-streptavidin-peroxidase technique as described previously [7]. The peptide was extracted from serum as described previously [28]. The coefficient of extraction averaged 89.7%. The standard curve was produced for OT concentrations ranging from 1.95 to 500 pg/ml, and the ID50 of the assay was 34.7 pg/ml. The intra-assay and interassay coefficients of variation were 7.8% and 11.7%, respectively.
The concentrations of PGFM in the plasma samples were determined with a direct EIA, as described previously [12]. The anti-PGFM serum (WS4468-5) was donated by Dr. W. J. Silvia, University of Kentucky, Lexington, KY. The PGFM standard curve was produced for PGFM concentrations ranging from 32.5 to 8000 pg/ml, and the ID50 of the assay was 315 pg/ml. The intra-assay and interassay coefficients of variation were on average 7.6% and 10.4%, respectively.
The concentrations of PGE2 were determined by a direct EIA test as described previously [29]. The anti-PGE2 serum was donated by Dr. S. Ito, Kansai Medical University in Osaka, Japan. Cross-reactivities of the anti-PGE2 serum, determined by measuring the inhibition of binding of peroxidase-labeled PGE2 to this antiserum, were as follows: PGE2, 100%; PGE1, 18%; PGJ2, 14%; PGA1, 10%; 15-keto PGE2, 8.8%; PGB2, 6.7%; PGA2, 4.6%; PGD2, 0.13% and PGF2α, 2.8%. The PGE2 standard curve was produced for PGE2 concentrations ranging from 0.07 to 20 ng/ml, and the ID50 of the assay was 1.25 ng/ml. The intra-assay and interassay coefficients of variation were on average 6.9% and 9.7%, respectively.
The concentrations of LTC4 were determined in plasma samples using a commercially available EIA kit (Cayman Chemical Co., Ann Arbor, MI) according to the instructions of the manufacturer. The intra-assay and interassay coefficients of variation were on average 4.9% and 7.4 %, respectively.
NO2−/NO3− Determination in Plasma
Plasma concentrations of NO2−/NO3−, the stable metabolites of NO, were measured by a colorimetric method using the Griess reaction as described by Green et al. [30] and adapted by us for serum samples [12]. The assay sensitivity was 0.065 μg/ml, and the standard curve was produced for NO2−/NO3− concentrations ranging from 0.05 to 6.9 μg/ml. The intra-assay and interassay coefficients of variation were on average 7.4% and 11.2%, respectively.
Statistical Analysis
Least squares means and SEMs were determined. The length of the estrous cycle and the total amount of released hormones (P4, OT), arachidonic acid metabolites (LTC4, PGE2, PGFM), and stable metabolites of NO represented as the area under the curve (relative units, Table 1) were analyzed using one-way ANOVA followed by the Bonferroni Multiple Comparison Test (ANOVA; GraphPAD PRISM Version 4.00; GraphPad Software, San Diego, CA). The analysis of hormones (P4, OT), arachidonic acid metabolites (LTC4, PGE2, PGFM), and stable metabolites of NO (NO2−/NO3−) in the jugular plasma samples, collected before, during, and after administration of different doses of TNFα and saline on Day 14 of the cycle, was performed using a repeated-measure design approach with treatments and time of sample collection (minutes) being fixed effects with all interactions included. All analyses were performed using repeated-measures ANOVA tests followed by the Bonferroni Multiple Comparison Test (GraphPAD PRISM; P < 0.05 was considered statistically significant).
Effects of 30-min infusion of saline (control) or TNFα on total amounts of released P4, OT, PGFM, PGE2, LTC4, and stabile metabolites of NO in heifers on Day 14 of the estrous cycle.*
| Treatment | P4 | OT | PGFM | PGE2 | LTC4 | NO2–/NO3– |
|---|---|---|---|---|---|---|
| Control | 19.6 ± 1.3a | 43.8 ± 2.0a | 217.4 ± 1.7a | 14.2 ± 1.2a | 310.1 ± 15a | 18.9 ± 0.7a |
| TNFα (1 μg) | 22.0 ± 3.0a | 27.7 ± 2.8b | 769.9 ± 98b | 7.8 ± 1.3b | 306.9 ± 24a | 33.7 ± 3.9b |
| TNFα (10 μg) | 39.6 ± 1.7b | 42.4 ± 4.2a | 249.5 ± 42a | 30.1 ± 5.3c | 237.1 ± 9b | 24.7 ± 0.7a |
| Treatment | P4 | OT | PGFM | PGE2 | LTC4 | NO2–/NO3– |
|---|---|---|---|---|---|---|
| Control | 19.6 ± 1.3a | 43.8 ± 2.0a | 217.4 ± 1.7a | 14.2 ± 1.2a | 310.1 ± 15a | 18.9 ± 0.7a |
| TNFα (1 μg) | 22.0 ± 3.0a | 27.7 ± 2.8b | 769.9 ± 98b | 7.8 ± 1.3b | 306.9 ± 24a | 33.7 ± 3.9b |
| TNFα (10 μg) | 39.6 ± 1.7b | 42.4 ± 4.2a | 249.5 ± 42a | 30.1 ± 5.3c | 237.1 ± 9b | 24.7 ± 0.7a |
The area under the curve (relative units, mean ± SEM). The baseline was defined on the basis of data from the first hour of the experiment. The area under the curve was measured using data from the last 4 h of the experimental period (see Figures 4, 5, and 6).
Different superscript letters within a column indicate significant differences (P < 0.05) between treated groups.
Effects of 30-min infusion of saline (control) or TNFα on total amounts of released P4, OT, PGFM, PGE2, LTC4, and stabile metabolites of NO in heifers on Day 14 of the estrous cycle.*
| Treatment | P4 | OT | PGFM | PGE2 | LTC4 | NO2–/NO3– |
|---|---|---|---|---|---|---|
| Control | 19.6 ± 1.3a | 43.8 ± 2.0a | 217.4 ± 1.7a | 14.2 ± 1.2a | 310.1 ± 15a | 18.9 ± 0.7a |
| TNFα (1 μg) | 22.0 ± 3.0a | 27.7 ± 2.8b | 769.9 ± 98b | 7.8 ± 1.3b | 306.9 ± 24a | 33.7 ± 3.9b |
| TNFα (10 μg) | 39.6 ± 1.7b | 42.4 ± 4.2a | 249.5 ± 42a | 30.1 ± 5.3c | 237.1 ± 9b | 24.7 ± 0.7a |
| Treatment | P4 | OT | PGFM | PGE2 | LTC4 | NO2–/NO3– |
|---|---|---|---|---|---|---|
| Control | 19.6 ± 1.3a | 43.8 ± 2.0a | 217.4 ± 1.7a | 14.2 ± 1.2a | 310.1 ± 15a | 18.9 ± 0.7a |
| TNFα (1 μg) | 22.0 ± 3.0a | 27.7 ± 2.8b | 769.9 ± 98b | 7.8 ± 1.3b | 306.9 ± 24a | 33.7 ± 3.9b |
| TNFα (10 μg) | 39.6 ± 1.7b | 42.4 ± 4.2a | 249.5 ± 42a | 30.1 ± 5.3c | 237.1 ± 9b | 24.7 ± 0.7a |
The area under the curve (relative units, mean ± SEM). The baseline was defined on the basis of data from the first hour of the experiment. The area under the curve was measured using data from the last 4 h of the experimental period (see Figures 4, 5, and 6).
Different superscript letters within a column indicate significant differences (P < 0.05) between treated groups.
Results
Experiment 1 (Preliminary Study)
Injection of aPGF2α shortened (17.7 ± 0.3 days) the estrous cycle length compared with that of control heifers injected with saline only (22.3 ± 0.3 days), whereas infusion of TNFα had a dose-dependent effect on the duration of the estrous cycle (Fig. 1, P < 0.05). Two low doses of TNFα (0.1 and 1 μg for 30 min) induced regression of CL (cycle durations, 18.8 ± 0.9 and 18.0 ± 0.7 days, respectively). The length of the estrous cycle was prolonged to more than 30 days by infusion of TNFα at a dose of 10 μg for 30 min compared with that in control heifers (22.3 ± 0.3 days, P < 0.05). Higher doses of TNFα (25 and 50 μg for 30 min) also lengthened the estrous cycle to 30 days (Fig. 1). During infusion of TNFα at these doses (25 and 50 μg for 30 min), clinical symptoms, including rapid increases in pulse and respiration rate, rise of blood pressure, and muscle contractions, were observed. Therefore, on the basis of this preliminary experiment, two doses of TNFα, 1 (luteolytic) and 10 (luteotropic) μg for 30 min, were chosen for further studies.
Concentration of progesterone in peripheral blood plasma of heifers infused with saline (n = 3) or with several doses of TNFα (each dose, n = 3) or cloprostenol (aPGF2α, 100 μg; n = 3) on Day 14 of the estrous cycle. The control group (control) was injected with saline only. Drugs were administered for 30 min into the aorta abdominalis. Different subscript letters indicate significant differences (P < 0.05) between treated groups
Experiment 2 (Influence of TNFα on the Estrous Cycle)
There were significant differences in the length of the estrous cycle between animals infused with 1 and 10 μg of TNFα (P > 0.05, Fig. 2). Infusion of 1 μg of TNFα shortened (17.5 ± 0.44 days) the cycle length compared with that of the group injected with saline (21.8 ± 0.65 days). In the heifers infused with 10 μg of TNFα, spontaneous luteolysis was prevented and the functional life of the CL was prolonged compared with those of the control group (more than 30 days versus 21.8 ± 0.65 days, P < 0.01).
Concentration of progesterone in peripheral blood plasma of heifers infused for 30 min with TNFα (1 or 10 μg per animal; each dose, n = 3) or saline (n = 3) on Day 14 of the estrous cycle. Saline or TNFα was infused into the aorta abdominalis. Different subscript letters indicate significant differences (P < 0.05) between treated groups
Although a lower dosage of TNFα (1 μg for 30 min) did not affect the P4 concentration in the blood (Fig. 3a , Table 1), it strongly reduced the OT concentration (P < 0.001; Fig. 3b, Table 1). The administration of 1 μg of TNFα induced PGF2α production, as shown by the increased PGFM concentration in the peripheral blood (P < 0.001, Fig. 4a) but did not affect PGE2 production (P > 0.05, Fig. 4b). Moreover, 1 μg of TNFα induced NO production, as shown by increase in the NO2−/NO3− level in the blood (P < 0.001; Fig. 5a , Table 1). The increase in the PGFM concentration showed a strong positive correlation with the NO2−/NO3− elevation (P < 0.001, r = 0.73). Furthermore, the elevation in the NO2−/NO3− concentration was inversely correlated with the OT concentration (P < 0.001, r = −0.71). One microgram of TNFα did not influence the LTC4 concentration in the blood during the experimental period (P > 0.05; Fig. 5b, Table 1).
Effect of TNFα (1 or 10 μg per animal; each dose, n = 3) or saline (n = 3) infusions on Day 14 of the estrous cycle on (a) progesterone and (b) oxytocin concentrations in peripheral blood plasma of heifers. Saline or TNFα was infused for 30 min into the aorta abdominalis. Different subscript letters indicate significant differences (P < 0.05) between treated groups
Effect of TNFα (1 or 10 μg per animal; each dose, n = 3) or saline (n = 3) infusions on Day 14 of the estrous cycle on (a) PGFM and (b) PGE2 concentrations in peripheral blood plasma of heifers. Saline or TNFα was infused for 30 min into the aorta abdominalis. Different subscript letters indicate significant differences (P < 0.05) between treated groups
Effect of TNFα (1 or 10 μg per animal; each dose, n = 3) or saline (n = 3) infusion on Day 14 of the estrous cycle on (a) stabile metabolites of NO (NO2−/NO3−) and (b) LTC4 concentrations in peripheral blood plasma of heifers. Saline or TNFα was infused for 30 min into the aorta abdominalis. Different subscript letters indicate significant differences (P < 0.05) between treated groups
Infusion of 10 μg of TNFα strongly elevated the concentration of P4 and PGE2 in peripheral blood (P < 0.001; Figs. 3a and 4b, Table 1) without any effect on the PGFM concentration (P > 0.05; Fig. 4a, Table 1). A strong correlation has been found between doses of TNFα and PGE2 levels (r = 0.86, P < 0.001). Although 10 μg of TNFα temporarily decreased the OT secretion at the time of the infusion (Fig. 3b), the OT concentration immediately returned to the basal level compared with the pretreatment value or the value in control animals infused with saline only (P > 0.05; Fig. 3b, Table 1). Although the higher dose of TNFα increased the concentration of NO2−/NO3− in the blood (P < 0.01, Fig. 5a), the effect was weaker than that in the lower dose of TNFα (1 μg) (P < 0.001, Table 1). Moreover, infusion of 10 μg of TNFα decreased the LTC4 concentration in the blood (P < 0.001; Fig. 5b, Table 1).
For the P4 concentration (Fig. 3a), two-way interactions were found between the 10-μg TNFα treatment and time of sample collection (P < 0.01). Three-way interactions were found among the 10-μg TNFα treatment, saline treatment, and time of sample collection (P < 0.01) and among the 10-μg TNFα treatment, the 1-μg TNFα treatment, and time of sample collection (P < 0.05). For the OT concentration (Fig. 3b), two-way interaction was found between the 1-μg TNFα treatment and time of sample collection (P < 0.05, Fig. 3b). Moreover, three-way interaction was found among TNFα (1 μg) treatment, saline treatment, and time of sample collection (P < 0.05). For the PGFM, concentrations (Fig. 4a), two-way interaction was found between the 1-μg TNFα treatment and time of sample collection (P < 0.001). For the PGE2 concentration (Fig. 3b), two-way interaction was found between the 1-μg TNFα treatment and time of sample collection (P < 0.001). Three-way interactions were found for both PGs among TNFα (1 and 10 μg) treatment, saline treatment, and time of sample collection (P < 0.001). For the NO2−/NO3− concentrations, two-way interactions were found between TNFα treatments (both 1 and 10 μg) and time of sample collection (P < 0.01). Three-way interaction was found among TNFα treatment, saline treatment, and time of sample collection (P < 0.01). For the LTC4 concentrations, three-way interaction was found among TNFα (10 μg), saline treatment, and time of sample collection (P < 0.05) and between both doses of TNFα and time of sample collection (P < 0.01).
Discussion
The present in vivo study showed that the effects of TNFα on the estrous cycle and the function of the CL in cattle varied dramatically, depending on the dose. At low doses (0.1 and 1 μg), TNFα induced luteolysis and shortened the estrous cycle, whereas at high doses it prevented luteal regression and prolonged the luteal lifespan. We did not measure the blood plasma concentration of TNFα achieved by the infusion of TNFα. However, based on the total amount of infused TNFα and the total amount of circulated blood, one could calculate that the lower doses of TNFα used in the present study only slightly raised the concentration of this cytokine in the peripheral blood, and the concentrations could be around the level previously observed during the estrous cycle in healthy cows [31]. On the other hand, the high doses of TNFα were calculated to lead to peripheral blood concentrations around those observed during inflammatory processes such as mastitis [32, 33]. Thus, the concentrations of TNFα used in the present study may not correspond to the local concentrations physiologically present in the female reproductive tract in cattle. However, the doses tested in the present study may be useful for possible applications of TNFα in the manipulation of luteal function either to promote or prevent luteolysis. Moreover, the data obtained in the present study may help explain some of the inconsistencies observed in in vitro studies [10, 20–23, 25, 34, 35].
We assume that the direct action of TNFα on the CL is one of the major mechanisms by which the effects of TNFα on the luteal lifespan vary depending on the dose. Shaw and Britt [31] used the in vivo microdialysis system to show that luteal concentrations of TNFα were detectable in the dialysates of cows exhibiting spontaneous or PGF2α-induced luteolysis only after the onset of P4 decline. These findings suggest that TNFα plays an important role in structural luteolysis. It has been shown that macrophages and other immune cells invade the bovine CL at the time of luteolysis [18, 34, 36, 37], and TNFα participates in the apoptotic events there [21–23]. Thus, lower doses of TNFα may directly induce apoptosis, resulting in regression of the CL and shortening of the estrous cycle.
However, it is also known that TNFα is involved in regulating normal ovarian function, including the proliferation of the cells in the CL and follicles of various species [38, 39]. It has been shown that in the rat ovary TNFα has a mitogenic effect on theca-interstitial cells and preferentially increases the proportion of steroidogenic cells [40]. Thus, TNFα has a diverse spectrum of biological activities, including stimulation of cell proliferation and differentiation and induction of cell apoptotic death [40, 41]. It has recently been shown that the effects of TNFα on the rat ovary varied, depending on the dose; although lower doses of TNFα induced apoptosis in oocytes and follicles, higher concentrations showed no effect [41]. The ability of TNFα to exert a wide variety of effects is likely due to actions exerted via multiple signaling pathways involving two distinct receptors, i.e., TNF-RI (high affinity, responsible for the transduction of cell death signaling) and TNF-RII (low affinity, implicated in cell proliferation) [42–44]. In addition to TNF-RI mRNA [20], mRNA for TNF-RII is also highly expressed in the bovine CL (unpublished data). Therefore, it is possible that the lower concentrations of TNFα (0.1 and 1 μg) in the present study induced cytotoxicity in the bovine CL by preferential binding to TNF-RI, whereas higher concentrations of TNFα (10, 25, and 50 μg) bound to both receptors or bound preferentially to TNF-RII, stimulating a survival pathway.
The data obtained in the current in vivo study support and extend our previous in vitro observations [9, 10, 25] and suggest the new concept that TNFα may be one of the crucial factors involved in the regulation of luteolysis at the uterine level in cattle. Infusion of the lower dose of TNFα (1 μg) increased the production and output of two important luteolytic factors, PGF2α [3, 4] and NO [10, 12–14], resulting in shortening of the estrous cycle. TNFα, in contrast to OT, affected PGF2α output from the bovine endometrium not only at the follicular stage of the estrous cycle but also before the P4 decline at the mid and late luteal stages [9]. Interestingly, TNFα stimulates PGF2α synthesis in the stromal cells but not in the epithelial cells of bovine endometrium [10]. The density of the populations of leukocytes and macrophages increases markedly in luminal and glandular epithelia of the bovine uterus during the mid to late luteal phases [45]. Thus, TNFα may originate from the immune cells that infiltrate the bovine uterus during luteolysis. Moreover, we have recently found by an in situ hybridization study that TNFα mRNA is expressed in bovine endometrial tissue during the estrous cycle and that TNFα may be preferentially produced by the epithelial endometrial cells in cattle (unpublished data). These findings suggest that TNFα induces autoamplification of the PGF2α synthesis loop in the bovine endometrium [3, 4, 6] in an autocrine and/or paracrine manner and that uterine TNFα initiates a positive cascade between uterine PGF2α and various luteolytic factors to complete luteolysis [4, 6, 8, 9, 19, 21, 34, 38, 46].
NO is the most likely candidate for an important component of a luteolytic cascade induced by TNFα following uterine PGF2α. NO may mediate the luteolytic actions of both TNFα and uterine PGF2α [12–14, 46]. TNFα at a dose of 1 μg induced a two-phase pattern of NO output. The first peak of NO concentration was observed during TNFα infusion and was followed by a long-term increase in NO production, as measured by NO2−/NO3− concentrations in the blood. The NO production was highly correlated with the increase in PGFM concentration. Therefore, in addition to the direct effect of TNFα on the NO synthase activity in the uterus [10, 47], TNFα may have induced the increase of NO production in the ovary via effects on PGF2α [12, 46] in the present study. Although we expected that both TNFα and NO would stimulate the secretion of LTC4 [13], no stimulatory effect of the luteolytic dose (1 μg) of TNFα on LTC4 output was observed. In fact, peaks of luteolytic leukotrienes (types B4 and C4) [15, 16] have been demonstrated in bovine CL on Day 18 in heifers undergoing spontaneous luteolysis, and their frequency increased within the 12-h period during which the onset of P4 decline occurred [17]. Therefore, the release of LTC4 from the CL may be one of the last results of the activation of the luteolytic cascade induced by uterine TNFα and PGF2α.
TNFα at a luteolytic dose (1 μg) strongly decreased the concentration of OT in the peripheral blood. The inverse correlation between NO and OT concentrations after TNFα treatment suggests that inhibitory effects of TNFα on OT secretion may be mediated by NO. In support, NO donors strongly inhibited the OT production by cultured bovine luteal cells [11] and the OT secretion in microdialyzed bovine CL in vivo [48], but the local inhibition of NO in microdialyzed bovine CL in vivo stimulated OT secretion [14]. Altogether, these findings suggest that the luteal concentration of OT is down-regulated during TNFα- and NO-dependent luteolysis. In fact, the concentrations of OT in the blood [49] and in intact [47] and microdialyzed in vivo CL [17, 50] are extremely low at the time of spontaneous luteolysis. Moreover, the blockade of uterine OT receptors with a specific OT antagonist from Day 15 to Day 22 of the cycle affected neither luteolysis nor the duration of the estrous cycle in heifers [51]. These findings raise serious questions as to whether OT of luteal origin plays a significant role in initiating luteal regression in cattle, as suggested previously [4, 6, 9, 17, 50, 51]. Therefore, PGF2α secretion by the endometrium during luteolysis in cattle may be regulated not only by OT but also by one or more other factors, including TNFα [4, 9, 10].
On the other hand, TNFα also induced the production and output of luteotropic PGE2 in cultured bovine luteal [20] and endometrial stromal cells and tissue in vitro [25]. Those in vitro observations were confirmed by the present in vivo data. A strong correlation has been found between doses of TNFα and PGE2 levels. The higher dose of TNFα (10 μg) stimulated P4 and PGE2 secretion and inhibited LTC4 secretion, resulting in the prolongation of the estrous cycle in cattle. These data suggest that TNFα has luteotropic and luteoprotective roles. In fact, TNFα gene expression and very high TNFα protein production have been found in embryo, placenta, and pregnant uterus and oviduct [38, 52]. It has recently been shown that TNFα stimulates PGE synthase mRNA expression in cultured bovine endometrial stromal cells [53]. Increased PGE synthase production and activity may change the PGE2/PGF2α ratio and contribute to establishing pregnancy [5, 25, 26, 54]. Since PGF2α and PGE2, which are synthesized by the bovine endometrium, are assumed to play opposite roles, the relative proportions of PGF2α and PGE2 synthesis may be more important than the absolute levels of each individual PG. The mechanisms and properties of features regulation of PGF2α and PGE2 synthesis in the bovine endometrium by different doses of TNFα are not known. However, it has been shown that TNFα induced a switch from the PGD2 to PGE2 synthesis pathway via the regulation of PGE synthase expression and activity in murine macrophages [55]. Both PGD2 and PGE2 may be converted enzymatically to PGF2α by PGF synthases (11-keto-PGD reductase, 9-keto-PGE reductase, and/or other aldo-keto-reductases) [25, 56]. However, Madore et al. [57] recently found that AKR1C family members (to which all the currently known PGFs belong) are not expressed in the bovine endometrium. Alternatively, an aldose reductase known for its 20α-hydroxysteroid dehydrogenase activity, AKR1B5, is a likely candidate enzyme for controlling the sufficient and timely production of PGF2α directly from PGH2 in the bovine endometrium [57]. Thus, modulation of the activity of PGF-converting enzymes may be a mechanism by which TNFα switches prostanoid metabolism in the bovine endometrium from production of luteolytic PGF2α to luteotropic PGE2. Further studies are needed to test these possibilities.
The mechanisms controlling the formation and maintenance of CL and, finally, luteolysis in cattle are known to be immune cell- and cytokine-dependent processes [8, 19, 21–23, 34]. The present in vivo study showed that TNFα inversely regulates the lifespan of CL in cattle, depending on its concentrations. These data together with results obtained from earlier in vitro studies suggest the concept that TNFα at low concentrations plays an important role in luteolysis, especially with regard to stimulation of PGF2a production in the uterus, and consequently leads to both functional and structural regression of CL in cattle. On the other hand, TNFα at high concentrations prolongs the estrous cycle in cattle by inducing a survival pathway in the CL and by contributing to P4 production.
Acknowledgments
We thank Dr. Genowefa Kotwica and Dr. Stanislaw Okrasa of the Warmia and Mazury Univeristy in Olsztyn, Poland, for OT and P4 antiserum, respectively; Dr. William Silvia of the University of Kentucky, Lexington, KY, for antiserum against PGFM; Dr. Seiji Ito of Kansai Medical University, Osaka, Japan, for PGE2 antiserum; and Dainippon Pharmaceutical Co., Ltd., Osaka, Japan for recombinant human TNFα (HF-13). The authors are grateful to Centrowet, Olsztyn, Poland, for the gifts of Crestar, cloprostenol, Sedazin, Polocainum Hydrochloricum, and other veterinary drugs used in the present study. The authors also thank Mrs. Hanna Kostuch and Mr. Jerzy Kostuch of the Animal Farm “Farmer” in Zalesie for their excellent cooperation and agreement to let us use the animals for the present experiments.
References
Author notes
This research was supported by the Grants-in Aid for Scientific Research (KBN 5P06K 003 21; JSPS 14360168) and Polish-Japanese Joint Research Project under the agreement between the Polish Academy of Sciences and the Japan Society for the Promotion of Sciences.
