Induction of mRNA for Chemokines and Chemokine Receptors by Prostaglandin F_2α Is Dependent upon Stage of the Porcine Corpus Luteum and Intraluteal Progesterone

Abstract

This study tested the hypotheses that prostaglandin (PG) F_2α increases expression of genes related to recruitment of leukocytes in mature but not early corpus luteum (CL) and that insensitivity to PGF_2α action in early CL is dependent on high intraluteal progesterone (P4) concentrations. Experiment 1 examined early (0.5 h) and late (10 h) in vivo effects of PGF_2α on mature (d 17 of pseudopregnancy) and early (d 9) porcine CL. Real-time PCR was used to measure mRNA for chemokines (IL8, CXCL2, CCL2, CCL8, CCL4, CCL11) and chemokine receptors (CCR1, CCR2, CXCR2, CCR5). Western blotting was used to measure protein expression and phosphorylation of nuclear factor-κB proteins. Treatment with PGF_2α for 10 h increased mRNA for almost all of these genes (all expect CXCL2 and CCL11) in d 17 CL but not d 9 CL. Treatment with PGF_2α also led to greater phosphorylation of nuclear factor-κB-1A protein in d 17 than d 9 CL. Experiment 2 had a 2 × 2 factorial design with d 9 gilts treated or not treated with epostane (3β-hydroxysteroid dehydrogenase inhibitor to suppress intraluteal P4) and treated or not treated with PGF_2α. Treatment with PGF_2α (10 h) or epostane alone did not induce expression of any of these genes in d 9 CL. However, PGF_2α + epostane increased expression of all of these genes except CCL11. In conclusion, PGF_2α increases mRNA for chemokines and chemokine receptors in mature CL with similar PGF_2α effects induced in early CL if intraluteal P4 is suppressed prior to PGF_2α treatment.

Mounting evidence supports a key role for the immune system in various aspects of luteal function but particularly in luteolysis. Increased accumulation of T lymphocytes and macrophages was observed during luteolysis (1, 2). There were also greater numbers of neutrophils in mature corpus luteum (CL) than early CL (3), and neutrophils were found to be critical for luteal regression in rats (4). Neutrophils can produce reactive oxygen species that have been associated with the luteolytic process (4, 5). Moreover, prostaglandin (PG) F_2α triggered infiltration of eosinophils into luteal tissues of ewes (6). In CL, activated T lymphocytes, macrophages, or other immune cells may exert luteolytic actions via inhibition of progesterone (P4) production, stimulation of PGF_2α secretion, stimulation of apoptosis in luteal cells, and phagocytosis of degenerating cells (7).

Consistent with the increase in immune cells, there is also an increase in chemokine expression during luteolysis. Expression of CCL2 [known as monocyte chemoattratant protein-1 (MCP)-1] was increased near luteolysis in all species that have been examined (1, 2). Expression of CCL2 is induced by PGF_2α (8), and chemokine C-C motif ligand (CCL) 2 can attract and activate monocytes/macrophages (9). Nevertheless, except for CCL2, few reports are available on expression of immunomodulatory molecules during luteolysis. Currently more than 50 chemokines and 18 chemokine receptors have been discovered (10). According to amino-terminal cysteine motifs, chemokines are classified into four families: the -CXC-, -CC-, -C-, and -CX3C- families. Chemokine receptors are also categorized into four families according to their corresponding ligand families (10): including CXCR (bind CXC chemokines), CCR (bind CC chemokines), XCR (bind C chemokines), and CX3CR (bind CX3C chemokines) (10). One notable characteristic of chemokine receptors is that many different ligands bind the same receptor and many ligands bind multiple receptors (11). Further investigation of different chemokines and their receptors in CL could provide insight into immune regulation and recruitment of leukocytes during luteolysis.

Furthermore, the nuclear factor-κB (NF-κB) family of transcription factors mediate transcriptional changes during immune responses, including regulation of expression of chemokines (12). The transcription factor NF-κB exists in an inactive form associated with the inhibitory protein inhibitor-κB (IκB) in cytoplasm of unstimulated cells. Upon stimulation, IκB is phosphorylated and degraded via the ubiquitin-proteasome pathway, with subsequent translocation of free NFκB to the nucleus and direct transcriptional regulation of genes by NF-κB. Thus, changes in chemokines may be associated with changes in NF-κB signaling during luteolysis.

The decrease in luteal P4 production generally occurs earlier than structural changes associated with luteolysis. Thus, decreased P4 may be critical for allowing luteolytic agents, such as PGF_2α, to activate the full luteolytic cascade. Indeed, many researchers have suggested that high intraluteal P4 is protective of CL, and research has shown that blockade of luteal P4 action can allow apoptosis of luteal cells (13, 14). Interestingly, P4 is an effective immunosuppressive agent (15–17), leading us to hypothesize that the luteal protective effects of P4 may be mediated through suppressive actions of P4 on expression of luteal chemokines.

In many species including porcine and bovine species, PGF_2α is the primary luteolysin; however, treatment with PGF_2α during the early luteal phase will not regress CL, even though receptors for PGF_2α are already abundantly present in early CL (18, 19). In bovine, CL acquire capacity to regress in response to a single PGF_2α treatment, termed luteolytic capacity, around d 5–7 of the estrous cycle (20). The pig is a particularly interesting model for studying luteolytic capacity because pig CL acquire luteolytic capacity near d 13 of the estrous cycle, although full tissue mass and steroidogenic productivity are attained by d 9 (21). In this study, we hypothesized that PGF_2α will differentially regulate chemokines, chemokine receptors, and the NF-κB signaling pathway in early (without luteolytic capacity) compared with later (with luteolytic capacity) porcine CL. Specifically we hypothesized that expression of these molecules would only be induced in later CL and not in early CL, consistent with their postulated role in complete structural luteolysis, which will occur only after PGF_2α treatment of older CL. Furthermore, we hypothesized that intraluteal P4 prevents PGF_2α induction of these factors in early CL. Specifically we evaluated whether the suppression of P4 production in early CL would either induce these immunomodulatory molecules or allow early CL to be sensitive to PGF_2α induction of these molecules.

Materials and Methods

Animal procedures and experimental design

Cross-bred gilts (Cambrough × line 19) 6–8 months of age were obtained from the University of Wisconsin-Madison herd or purchased from Pig Improvement Co. (Franklin, KY). Animals were kept in individual pens with free access to water and fed a maintenance diet of corn and soybean meal. Animals were checked daily for standing estrus with a mature boar. First day of estrus was designated as d 0. Pseudopregnancy was induced in some gilts with daily injections of estradiol benzoate (2 mg im) on d 11–15. On the day ovaries were collected, anesthesia was induced with im injection of ketamine (15 mg/kg) and xylazine (0.3 mg/kg). Gilts were intubated and surgical plane of anesthesia maintained with halothane. Ovaries were collected via midventral laparotomy, and CL were dissected away from ovarian stroma and frozen in liquid nitrogen for further processing. The Research Animal Resource Center at the University of Wisconsin-Madison approved all animal procedures. Three different experiments were performed.

The first experiment examined acute (0.5 h) in vivo effects of PGF_2α on expression of mRNA for chemokines, chemokine receptors, and phosphorylation of proteins in the NF-κB pathway. A paired experimental design was used to ensure timing of CL collection because it would have been difficult to anesthetize and surgically collect CL within 30 min after PGF_2α treatment. On d 9 after estrus (n = 4) or d 17 of pseudopregnancy (n = 5), the gilts were anesthetized and one ovary was collected (control CL). After the removal of the control ovary, PGF_2α (500 μg of cloprostenol) was given im, and the other ovary was collected 0.5 h later (treated CL).

The second experiment examined later (10 h) effects of in vivo treatment with PGF_2α on expression of mRNA for chemokines and chemokine receptors. Animals were randomly assigned to one of four groups: d 9 saline (n = 5), d 9 PGF_2α (500 μg cloprostenol; n = 4), d 17 saline (n = 5), and d 17 PGF_2α (500 μg cloprostenol; n = 5). On d 9 of the estrous cycle or d 17 of pseudopregnancy, gilts received saline or PGF_2α and ovaries were removed 10 h later.

The last experiment examined the effect of suppression of P4 production by epostane, an inhibitor of 3β-hydroxysteroid dehydrogenase, in d 9 CL on effects of PGF_2α (10 h after treatment) on expression of mRNA for chemokines and chemokine receptors. The first day of estrus was designated d 0. On d 7, gilts (n = 4/ group) were divided using a 2 × 2 factorial design. The four groups included the following: control, epostane, PGF_2α, and PGF_2α plus epostane. Epostane (10 mg/kg; a gift from Sanofi-Synthelabo Research, Malvern, PA) was fed in a mixture of corn and soybean meal every 12 h for 36 h. At 38 h animals received either vehicle or 25 mg PGF_2α. Ovaries were collected 10 h after injection (48 h).

Isolation of mRNA

mRNA was isolated from collected CL tissues using Magnetight oligo (deoxythymidine) magnetic beads (Novagen, Madison, WI) (18). Frozen CL tissue was ground to fine powder using mortar and pestle filled with liquid nitrogen. Approximately 15 mg of luteal powder were homogenized in 200 μl lysis buffer [4 m guanidine isothiocyanate, 0.1 m Tris (pH 8.0), 1% dithiothreitol, and 0.5% N-lauroylsarcosine] and neutralized with 400 μl binding buffer [100 mm Tris (pH 8.0), 400 mm NaCl, and 20 mm EDTA]. Samples were centrifuged at 16,000 × g for 5 min at 4 C to pellet cellular debris. The supernatant of each sample was transferred to tubes containing 200 μl of oligo (deoxythymidine) beads and allowed to hybridize for 10 min. Beads were then captured on a magnetic stand and washed three times with 500 μl wash buffer [10 mm Tris (pH 8.0), 150 mm NaCl, and 1 mm EDTA]. The mRNA was eluted by heating 15 μl of elution buffer (2 mm EDTA) to 65 C for 5 min and isolated mRNA was stored at −80 C.

Reverse transcription and real-time PCR

Reverse transcription (RT) was performed in 20 μl volume of RT master mix, which contained 1× RT reaction buffer, 5 μm random hexamer primers, 200 μm deoxynucleotide triphosphate, 40 U Moloney murine leukemia virus reverse transcriptase (Promega, Madison, WI), and 2 μl of mRNA. The RT reaction was carried out at 37 C for 1.5 h followed by heating to 95 C for 10 min in a programmable thermocycler (MJ Research, Watertown, MA).

Steady-state concentrations of investigated mRNA were quantified by real-time PCR using a GeneAmp 5700 sequence detection system (PE Applied Biosystems, Foster City, CA) with PCR products detected with SYBR Green I (Molecular Probes, Eugene, OR). Primers for amplification were designed using Primer Express (PE Applied Biosystems). Each PCR reaction mix (25 μl) contained 1× PCR buffer (Promega) with 1:20,000 dilution of SYBR Green I, 1.5 mm MgCl₂, 200 μm deoxynucleotide triphosphate, 250 nm forward primer, 250 nm reverse primer, 2 μl RT products, and 1.25 U GoToTaq polymerase (Promega) (22, 23). Thermal cycling conditions were 94 C for 30 sec, followed by 40 cycles at 94 C for 30 sec, 57 C for 30 sec, and 72 C for 30 sec, and finally 72 C for 10 min. Melting curve analyses and agarose gel electrophoresis were performed after real-time PCR reactions to monitor PCR product purity. Primers for investigated genes are listed in Supplemental Table 1, published on The Endocrine Society's Journals Online web site at http://endo.endojournals.org.

The threshold cycle numbers were determined for each amplified cDNA and for the housekeeping gene, ACTB (known as β-actin), in each unknown sample during real-time PCR. Relative quantification of investigated gene expression was evaluated using a standard curve method (24). Amount of investigated mRNA and ACTB mRNA was determined from the standard curve. Amount of investigated mRNA was then divided by amount of ACTB to obtain a normalized mRNA value.

Western blot analysis

NF-κB-1A (NFKB1A; known as IκBα), phosphorylated (p) NFKB1A, RELA (known as NF-κBp65 subunit), and phosphorylated RELA (p-RELA) were analyzed by Western blotting. Frozen luteal tissue (20 mg) was homogenized in 600 μl of cold homogenization buffer. Lysate was centrifuged at 16000 × g for 10 min to obtain a clear lysate. Protein concentration in supernate was measured by BCA protein assay kit (Pierce, Rockford, IL). Protein samples (60 μg) were separated with 10% one-dimensional SDS-PAGE gel and transferred to polyvinylidene difluoride membrane using miniprotein II gel transfer system (Bio-Rad, Hercules, CA). After transfer, blots were incubated in blocking buffer (5% milk and 0.05% Tween 20 in TBS [10 mm Tris-Cl (pH 8.0), 150 mm NaCl]) for 1 h at room temperature, followed by four washes. Immunoblotting was performed by incubating blots with goat anti-p-NFKB1A, goat anti-RELA, rabbit anti-NFKB1A, or rabbit anti-p-RELA (Santa Cruz Biotechnology, Santa Cruz, CA) for 1 h, followed by four washes. To visualize Western blots for first antibodies produced in goats (anti-p-NFKB1A and anti-RELA), a donkey antigoat IgG conjugated to horseradish peroxidase (Santa Cruz Biotechnology) was incubated with blots for 1 h at room temperature. To visualize the blots that used first antibodies produced in rabbits (anti-NFKB1A, and anti-p-RELA), a donkey antirabbit IgG conjugated to horseradish peroxidase (Santa Cruz Biotechnology) was incubated with blots for 1 h at room temperature. After four washes, immunoreactive NFKB1A, p-NFKB1A, RELA, and p-RELA were detected with an enhanced chemiluminescent reagent (NEN Life Science Products, Boston, MA). Blots were exposed to x-ray film and quantified using Quantity One 1-D analysis software on a Bio-Rad imaging system.

Statistical analysis

A one-way ANOVA was used to analyze differences in mRNA and protein concentrations among treatment groups, followed by mean separation using Fisher's least significant difference test (SAS version 9.1.3; SAS Institute, Cary, NC). To compare two groups (control vs. treated), an unpaired t test was used, except in experiment 1 when a paired t test was used. In all experiments, P < 0.05 was considered to be statistically significant.

Results

Effects of PGF_2α on expression of chemokines

Expression of two -CXC- family members and four -CC- family members were investigated after PGF_2α treatment in CL with and without luteolytic capacity (Fig. 1) and in early CL with inhibited P4 production (Fig. 2). The two -CXC- family members that were investigated were IL8 (known as IL-8) and CXCL2 (known as growth related oncogene-b, GRO-β). Treatment with PGF_2α did not change expression of IL8 mRNA at 0.5 or 10 h after treatment in d 9 CL (Fig. 1A). However in d 17 CL, treatment with PGF_2α increased steady-state mRNA concentrations for IL8 at 10 h (2.5-fold; P < 0.01) but not 0.5 h after PGF_2α (Fig. 1A). In contrast to IL8, the mRNA encoding CXCL2 was increased at 0.5 h after PGF_2α in both d 9 and d 17 CL but with a greater increase in d 17 (4.8-fold) than d 9 (2.3-fold) CL (Fig. 1B). There was no effect of PGF_2α on CXCL2 mRNA at 10 h after PGF_2α in d 9 or d 17 CL (Fig. 1B).

Fig. 1.

Effects of PGF_2α treatment on expression of mRNA for IL8 (A), CXCL2 (B), CCL2 (C), CCL8 (D), CCL4 (E), and CCL11 (F) in CL at different stages. For each mRNA, the left figure shows results from d 9 CL and the right figure shows results from d 17 CL. The concentrations of each mRNA were normalized to ACTB mRNA. CTL, Control; PGF, PGF_2α. *, P < 0.05 compared with mRNA expression of corresponding control.

Fig. 2.

Effects of epostane with or without PGF_2α treatment on expression of mRNA for IL8 (A), CXCL2 (B), CCL2 (C), CCL8 (D), CCL4 (E), and CCL11 (F) in d 9 CL. The concentrations of each mRNA were normalized to ACTB mRNA. EPO, epostane. PGF, PGF_2α. *, P < 0.05 compared with mRNA expression of corresponding control.

There were four -CC- family members that were investigated, CCL8 (known as MCP-2), CCL4 (known as macrophage inflammatory protein-1b, MIP-1b), CCL2, and CCL11 (known as eotaxin). Concentration of CCL2 mRNA increased at both 0.5 h (3.1-fold) and 10 h (3.6-fold) after PGF_2α in d 17 CL but increased at only 0.5 h (2.6-fold) after PGF_2α in d 9 CL (Fig. 1C). Treatment with PGF_2α increased CCL8 mRNA at both 0.5 h (1.7-fold) and 10 h (2.2-fold) after treatment in d 17 CL (Fig. 1D). However, PGF_2α did not increase expression of CCL8 in d 9 CL at any time (Fig. 1D). The mRNA concentrations for CCL4 were very low before treatments in both d 9 and d 17 CL. Treatment with PGF_2α increased CCL4 mRNA at 0.5 h after treatment in both d 9 (10.2-fold) and d 17 (13.7-fold) CL. However, at 10 h after PGF_2α treatment, mRNA for CCL4 remained elevated in d 17 (25.4-fold) but not d 9 CL (Fig. 1E). PGF_2α did not change CCL11 mRNA at any time in d 9 or d 17 CL (Fig. 1F).

We further investigated effects of inhibition of P4 production by epostane on PGF_2α-induced chemokine expression in d 9 CL. Treatment with only epostane for 48 h did not change expression of any of the examined chemokines in comparison with control (Fig. 2, A–F). Treatment with only PGF_2α for 10 h also did not alter expression of these chemokines (Fig. 2, A–F). Strikingly, the combination of epostane + PGF_2α (PGF_2α treatment at 38 h after beginning treatment with epostane) increased steady-state concentrations of mRNA for CXCL2 (1.9-fold), CCL2 (1.7-fold), and CCL8 (7.1-fold) and dramatically increased mRNA for IL8 (11.9-fold) and CCL4 (56.7-fold) in d 9 CL (Fig. 2, A–E).

Effects of PGF_2α on expression of chemokine receptors

The effect of PGF_2α on steady-state mRNA concentrations was evaluated for four chemokine receptors: CXCR2, CCR1, CCR2, and CCR5. These four chemokine receptors bind different chemokines: CXCR2 binds IL-8 and CXCL2, CCR1 binds CCL8, CCR2 binds CCL2 and CCL8, and CCR5 binds CCL4.

At 0.5 h after PGF_2α treatment, there was no change in mRNA concentrations for any of the four chemokine receptors in d 17 (Table 1) or d 9 (Supplemental Table 2) CL. At 10 h after PGF_2α treatment of d 17 CL (Table 1), there was an increase in mRNA for CXCR2 (1.8-fold), CCR1 (8.6-fold), CCR2 (2.1-fold), and CCR5 (1.9-fold). In contrast, PGF_2α treatment of d 9 CL produced only a small increase in mRNA for CXCR2 (1.3-fold) but did not change mRNA for CCR1, CCR2, or CCR5 (Supplemental Table 2).

Table 1.

Effects of PGF_2α treatment on expression of mRNA for CXCR2, CCR1, CCR2, and CCR5 in CL from gilts on d 17 of pseudopregnancy

mRNA	0.5 h		10 h
	Control	PGF2α	Control	PGF2α
CXCR2 (×10−3)	6.70 ± 1.43a	6.96 ± 2.05a	7.57 ± 0.87a	15.6 ± 1.62b
CCR1 (×10−3)	4.15 ± 0.42a	5.39 ± 0.57a	7.40 ± 1.61a	49.2 ± 6.90b
CCR2 (×10−2)	7.58 ± 2.43a	12.9 ± 3.18a	8.78 ± 2.18a	26.6 ± 3.73b
CCR5 (×10−1)	12.0 ± 1.97a	12.3 ± 3.66a	12.9 ± 1.69a	23.9 ± 3.37b

^a,bData within each row without common superscripts are significantly different between groups (P < 0.05).

Furthermore, we examined whether inhibition of P4 production by epostane altered the chemokine receptor response of d 9 CL to PGF_2α treatment. Concentrations of the four chemokine receptors were not increased by treatment with only epostane for 48 h (Table 2) or only PGF_2α for 10 h, except for CCR2 (Table 2). However, the combination of epostane plus PGF_2α treatment of d 9 CL (Table 2) increased mRNA for CXCR2 (2.4-fold), CCR1 (13.6-fold), CCR2 (4.4-fold), and CCR5 (4.4-fold).

Table 2.

Effects of treatment with epostane (EPO) and/or PGF_2α on expression of mRNA for CXCR2, CCR1, CCR2, and CCR5 in d 9 CL

mRNA	Control	EPO	PGF2α	PGF2α + EPO
CXCR2 (×10−3)	6.40 ± 1.62a	6.99 ± 2.22a	9.54 ± 2.37a	15.4 ± 1.96b
CCR1 (×10−3)	4.36 ± 0.50a	5.12 ± 1.34a	7.42 ± 1.89a	59.5 ± 27.6b
CCR2 (×10−2)	7.76 ± 1.03a	8.08 ± 2.03ab	15.2 ± 3.20b	34.3 ± 6.65c
CCR5 (×10−1)	9.64 ± 2.88a	8.50 ± 4.58a	22.0 ± 6.53ab	42.5 ± 13.2b

^a,b,cData within each row without common superscripts are significantly different between groups (P < 0.05).

Effects of PGF_2α on phosphorylation of proteins in NF-κB signaling pathway

Phosphorylated-NFKB1A and total NFKB1A were measured at 0.5 and 10 h after PGF_2α treatment of d 9 and 17 CL. At 0.5 h after PGF_2α, p-NFKB1A was detectable, whereas it was undetectable under control conditions (no PGF_2α) or at 10 h after PGF_2α treatment (Fig. 3A). There were no differences in total amounts of NFKB1A after PGF_2α treatment of d 17 CL (Fig. 3B). Furthermore, d 17 CL had greater concentrations of p-NFKB1A after treatment with PGF_2α for 0.5 h compared with CL on d 9 (920.2 vs. 353.0 intensity per square millimeter). However, similar to d 17 CL, p-NFKB1A was undetectable in d 9 CL under control conditions or at 10 h after PGF_2α treatment. There were no differences in total amounts of NFKB1A between control and PGF_2α-treated d 9 CL. In d 9 CL, total amounts of NFKB1A in control and at 0.5 h after PGF_2α were 904.7 intensity per square millimeter and 981.1 intensity per square millimeter, respectively. For the 10-h experiment, there was also no difference in NFKB1A in control (882.9 intensity per square millimeter) and 10 h PGF_2α treatment (989.6 intensity per square millimeter). We also measured p-RELA and RELA. Treatment with PGF_2α did not change concentrations of p-RELA or total RELA in d 9 (data not shown) or d 17 CL (Fig. 3).

Fig. 3.

Effects of PGF_2α treatment on phosphorylation (A) and total concentrations (B) of NFKB1A and phosphorylation (C) and total concentrations (D) of RELA in CL from gilts on d 17 of pseudopregnancy. CTL, Control; PGF, PGF_2α. *, P < 0.05 compared with the density of protein signal of corresponding control.

Discussion

Luteal regression involves a cascade of events including suppression of luteal P4 production and degeneration of luteal tissue. Immune system involvement in luteal regression is supported by an increasing body of evidence indicating that immune cells are not only involved in engulfing apoptotic luteal cells but also in directly causing loss of steroidogenesis and luteal cell death (7, 25). In the present study, treatment with PGF_2α increased mRNA for multiple chemokines from two different subfamiles in CL with luteolytic capacity (d 17) at both 0.5 and 10 h after PGF_2α. In contrast, effects of PGF_2α on chemokine expression in d 9 CL were observed only at early and not later times. Furthermore, treatment with PGF_2α induced a marked increase in mRNA for multiple chemokine receptors but only at the 10-h time point and only in d 17 and not d 9 CL. All of these data are consistent with the idea that PGF_2α treatment of CL attracts and activates immune cells through induction of chemokines and further that lack of immune system activation following PGF_2α treatment may allow the d 9 CL to escape the complete cascade of luteal regression.

One of the most studied and consistent observations of luteal chemokines is the increase in CCL2 near the time of luteolysis. For example, Townson et al. (2) reported an increase in CCL2 mRNA and protein and accumulation of monocytes/macrophages and T lymphocytes in bovine CL during the latter half of the estrous cycle. In rats, the concentrations of Ccl2 protein and the number of macrophages and apoptotic cells increased after the initiation of luteolysis by the proestrous prolactin surge in normally cycling rats (26) or by prolactin treatment of hypophysectomized rats (27). Treatment of sheep with PGF_2α increased CCL2 mRNA in ovine CL collected during the midluteal phase (8). Treatment of cattle with PGF_2α increased CCL2 in CL with luteolytic capacity but not in CL without luteolytic capacity (1).

Our study supports and extends these observations on CCL2. Clearly PGF_2α increased CCL2 mRNA in older CL at both 0.5 and 10 h after treatment. However, in CL without luteolytic capacity the early induction of CCL2 mRNA by PGF_2α was not followed by continued elevation at later times. We also observed a similar pattern for CCL4, with dramatic induction by 0.5 h after PGF_2α in CL with or without luteolytic capacity. However, only CL with luteolytic capacity had continued elevation of CCL4 mRNA at 10 h after PGF_2α. In contrast, CCL8 was elevated only in CL with luteolytic capacity at either 0.5 or 10 h after PGF_2α. Furthermore, CXCL2 was elevated at 0.5 h after PGF_2α in CL with and without luteolytic capacity, although the elevation was more dramatic in CL with luteolytic capacity. A pattern emerges from these somewhat diverse responses in PGF_2α-induced chemokine expression. Treatment with PGF_2α initiates many early chemokine aspects of luteolysis in CL with or without luteolytic capacity, such as induction of CCL2, CCL4, and CXCL2 mRNA. Nevertheless, later chemokine responses to luteolysis (increased IL8, CCL2, CCL8, and CCL4 mRNA) were observed only in CL with luteolytic capacity, suggesting that full PGF_2α-induced immune responses are not activated in CL without luteolytic capacity.

To date, the cell types that produce these chemokines in CL have not been determined except for CCL2. It is possible that multiple cell types are involved in secreting these chemokines in CL because endothelial cells, monocytes/macrophages, T lymphocytes, and neutrophils have been found to produce chemokines in other organs or systems (28, 29). Indeed, Liptak et al. (28) demonstrated that endothelial cells were an important source of CCL2 (MCP-1) in CL. In addition, peripheral blood mononuclear cells enhanced production of CCL2 by endothelial cells (28). We have also observed that after 8 d of luteinization of granulosa cells with forskolin, there are dramatic increases in IL8 and CXCL2 after PGF_2α, suggesting that large luteal cells may also be a source of chemokines from CL (30). Obviously further research is needed to completely define the contribution of each luteal cell types to production of intraluteal chemokines.

The two members of the -CXC- chemokine subfamily that were evaluated in this study, IL8 and CXCL2, are known to attract neutrophils (11). The CXCR2 binds both CXCL2 and IL8 and is primarily expressed on neutrophils (31). Treatment with PGF_2α increased CXCR2 mRNA only at 10 h after PGF_2α. The most dramatic induction was observed in d 17 CL, although there was a smaller (28%) increase in CXCR2 in d 9 CL after PGF_2α. The members of the -CC- family of chemokines are also potent modulators of immune cell function. CCL2 is a potent chemokine that recruits macrophages and increases migration of T lymphocytes (11). Both CCL2 and CCL8 bind and activate CCR2 found on monocytes and memory T cells (31). In addition, CCL8 binds to CCR1 found primarily on T cells, monocytes, and eosinophils (31). In contrast, CCL4 binds CCR5 primarily expressed on T cells and monocytes (31). CCL4 and CCL8 have been found to have similar actions to CCL2 on both macrophages and T lymphocytes. However, different from CCL2 and CCL4, CCL8 is also involved in infiltration of eosinophils (32). Thus, it is possible that CCL4, CCL2, and CCL8 increase accumulation of macrophages and T lymphocytes in CL; however, attraction of luteal eosinophils may involve increased CCL8. Treatment with PGF_2α increased mRNA for each of these chemokine receptors but only at 10 h after PGF_2α and only in CL with luteolytic capacity. There are at least two possible explanations for the PGF_2α-induced increase in chemokine receptor mRNA. First, PGF_2α could directly or indirectly increase mRNA concentrations by increasing expression or stability of mRNA for chemokine receptors in immune cells or other luteal cells. Another possibility, which seems more likely to us, is that PGF_2α treatment increased the numbers of immune cells containing chemokine receptors in CL with luteolytic capacity. Previous research has shown increased immune cells during luteolysis including macrophages/monocytes, neutrophils, and eosinophils in different species including cattle (4, 5), sheep (9, 11), rats (26), buffalo (6), and mares (33). Thus, the PGF_2α-induced increase in chemokine receptors may reflect only the increase in immune cells, such as macrophages, T lymphocytes, and eosinophils, that occurs during the luteolysis process.

The mechanisms leading to activation of these immune mechanisms only in CL with luteolytic capacity could not be fully elucidated in this study. Nevertheless, the acute phosphorylation of NFKB1A after PGF_2α only in CL with luteolytic capacity implicates the NF-κB pathway in this process. Previous studies have clearly shown pivotal roles for NF-κB in the regulation of inflammation (12). After stimuli that initiate inflammation, NFKB1A is phosphorylated, degraded in the cytoplasm, and free NF-κB translocates to the nucleus to regulate transcription of genes related to inflammation (12). The κB-element on which NF-κB binds was found in promoter regions of genes for chemokines and adhesion molecules (34–36). It was intriguing that PGF_2α treatment caused greater phosphorylation of NFKB1A in d 17 than d 9 CL, whereas RELA, a subunit of NF-κB, was phosphorylated in both d 9 and d 17 CL, and this phosphorylation was independent of PGF_2α treatment. Phosphorylation of RELA increases its transcriptional activity (37). Thus, differential activation of NF-κB by PGF_2α seems to primarily involve differential phosphorylation of NFKB1A and not RELA.

One particularly striking result from this study was the clear role that P4 played in inhibiting the action of PGF_2α in activation of these immune pathways in the d 9 CL. Clearly suppression of P4 alone or PGF_2α alone did not induce these inflammation-related genes. However after suppressing intraluteal P4 concentrations, PGF_2α was surprisingly potent in induction of essentially all inflammation-related genes evaluated in this study. This may indicate that suppression of P4 production may induce a physiological status in the early CL that resembles the later CL. In previous studies, P4 treatment decreased basal and cytokine-stimulated prostaglandin production by cultured luteal cells (38) and inhibited luteal cell apoptosis (13). Our findings extend these results to implicate intraluteal P4 in inhibiting the induction by PGF_2α of key molecules involved in the recruitment and activation of immune cells in the CL without luteolytic capacity. However, the intracellular mechanism(s) whereby intraluteal P4 inhibits expression of the chemokines and chemokine receptors has not yet been elucidated. In other tissues, P4 suppresses gene transcription via interaction with the NF-κB pathway (39). In human myometrial cells, inhibition by P4 of PTGS2 (known as prostaglandin-endoperoxide synthase-2) expression was associated with induction of mRNA and protein for NFKB1A to block NF-κB transactivation (40). In differentiated endometrial cancer cells, P4 induced expression of binding proteins that formed a complex that reduced NF-κB transcriptional activity (41). Thus, a reasonable speculation, based on the NF-κB pathway regulating expression of inflammation-related molecules (42), is that there is cross talk between P4 and NF-κB in regulating expression of chemokines in CL. Furthermore, the recruitment and activation of various types of immune cells into CL after PGF_2α may be critical for the decrease in luteal P4 production (4), increase in reactive oxygen species, and increases in TNF (known as TNF ligand superfamily member 2, TNF-α) from macrophages and IFNG (known as interferon-γ) from activated T-lymphocytes (43). TNF, IFNG, and reactive oxygen species have all been implicated in luteal cell death via apoptosis (5, 44).

In conclusion, our results are consistent with the idea that differential regulation of immune-related molecules may underlie at least some of the mechanisms involved in acquisition of luteolytic capacity. Furthermore, the NF-κB signaling pathway and intraluteal P4 seemed to be involved in regulating transcription of these molecules, suggesting a cascade of events that trigger activation of the immune system and complete luteolysis in the CL with luteolytic capacity.