Human Parturition Involves Phosphorylation of Progesterone Receptor-A at Serine-345 in Myometrial Cells

Abstract

The hypothesis that phosphorylation of progesterone receptor (PR) isoforms, PR-A and PR-B, in myometrial cells affects progesterone action in the context of human parturition was tested. Immunodetection of phosphoserine (pSer) PR forms in term myometrium revealed that the onset of labor is associated with increased phosphorylation of PR-A at serine-345 (pSer345-PRA) and that pSer345-PRA localized to the nucleus of myometrial cells. In explant cultures of term myometrium generation of pSer345-PRA was induced by interleukin-1β and dependent on progesterone, suggesting that pSer345-PRA generation is induced by a proinflammatory stimulus. In the hTERT-HM^A/B human myometrial cell line, abundance of pSer345-PRA was induced by progesterone in a dose- (EC₅₀ ∼1 nM) and time-dependent manner. Prevention of pSer345 (by site-directed mutagenesis) abolished the capacity for PR-A to inhibit anti-inflammatory actions of progesterone mediated by PR-B but had no effect on the transrepressive activity of PR-A at a canonical progesterone response element. Taken together, the data show that human parturition involves the phosphorylation of PR-A at serine-345 in myometrial cells and that this process is ligand dependent and induced by a proinflammatory stimulus. We also found that in myometrial cells, pSer345 activates the capacity for PR-A to inhibit antiinflammatory actions of progesterone mediated by PR-B. Phosphorylation of PR-A at serine-345 may be an important functional link between tissue-level inflammation and PR-A-mediated functional progesterone withdrawal to trigger parturition.

The steroid hormone progesterone maintains pregnancy by promoting uterine quiescence and cervical closure and inhibiting the process of parturition (1–3). Withdrawal of the progesterone block to parturition is the principal physiological trigger for parturition, and in most species this occurs by a systemic decrease in maternal progesterone levels (3–5). Human parturition, which occurs without a systemic progesterone withdrawal (6–8), is instead thought to be triggered by a functional progesterone withdrawal, whereby uterine target cells, especially myometrial cells, become refractory to progestational actions of progesterone (9–11). This is thought to involve changes in progesterone receptor (PR) signaling in myometrial cells. Indeed, treatment with PR antagonists, such as mifepristone and onapristone, increases myometrial contractility and in most cases induces the full parturition cascade at all stages of pregnancy (12–14). Thus, it is generally considered that progesterone promotes human pregnancy via a nuclear PR antagonist-sensitive mechanism(s) and that parturition is triggered, in the absence of systemic progesterone withdrawal, by a physiologically controlled modulation of PR signaling in uterine target cells that causes a functional progesterone withdrawal. One mechanism for functional progesterone withdrawal is by changes in the relative levels and activities of the PR isoforms, PR-A and PR-B, in myometrial cells (9, 10).

The human nuclear PR exists as two major isoforms: the full-length PR-B and the truncated (by 164 N terminal amino acids) PR-A (15, 16). Both PRs function as ligand-activated transcription factors (17), and each can mediate distinct genomic actions of progesterone in a cell type- and context-specific manner (18–20). In general, responsiveness to progesterone is dependent on the net transcriptional activity of PR-A and PR-B. An important property of this dual receptor system is that in most cells PR-A decreases progesterone responsiveness (assessed by activity at a reporter containing the canonical progesterone response element [PRE]) by inhibiting the transcriptional activity of PR-B (10, 21–26). We have proposed that the transrepressive activity of PR-A in myometrial cells is a mechanism for functional progesterone withdrawal and a key trigger for human parturition (9–11). However, the control of myometrial cell PR transcriptional activity, and especially the transrepressive activity of PR-A, in the setting of human pregnancy and parturition, is not clearly understood. In the present study, we tested the hypothesis that PR function in human pregnancy myometrium is affected by site-specific serine phosphorylation.

The PR isoforms can be phosphorylated at multiple serine residues (at least 14) by a variety of protein kinases and hormonal and intracellular modulators (27). Serine phosphorylation affects PR transcriptional activity by modulating PR isoform stability, hormone sensitivity, nuclear localization, and promoter targeting (28, 29). We examined the presence of six phosphoserine-PR (pSer-PR) forms in human term myometrium by immunoblotting and found that PR-A phosphorylated at serine-345 (pSer345-PRA; number relative to amino acid 1 in PR-B) was readily detectable (Figure 1). This study expands on those findings by examining the abundance, localization, regulation, and function of pSer345-PRA in human myometrial cells. Our data suggest that phosphorylation of PR-A at serine-345 (and serine-344) is progesterone dependent and induced by proinflammatory stimuli in term myometrium, and that serine-344/345 phosphorylation is necessary for PR-A to inhibit PR-B-mediated antiinflammatory activity in the pregnancy myometrium. Because tissue-level inflammation is a causal factor in the human parturition process (30–34), activation of PR-A to repress PR-B through inflammation-induced site-specific serine phosphorylation may be a mechanism by which tissue-level inflammation induces functional progesterone withdrawal to trigger human parturition.

Figure 1.

pSer-PR form in human term myometrium. Immunoblot analysis of PRs, pSer-PRs, and GAPDH in 80 μg of whole-cell lysate from seven term myometrium specimens is shown. A strong immunoreactive band was detected with the pSer345-PR antibody. The T47D breast cancer cell line treated with progesterone (P4; 100 nM) for 16 hours was used as positive control for pSer-PR antibodies (see Supplemental Table 1 for antibody details).

Materials and Methods

Myometrial tissue

Full-thickness uterus (1–2 cm³) was excised from the upper margin of the lower uterine segment incision of women undergoing term (≥39 wk of gestation) cesarean section (c-section) delivery at MacDonald Women's Hospital, University Hospitals Cleveland Medical Center (Cleveland, Ohio). All subjects (n = 19) provided informed consent (University Hospitals Cleveland Medical Center Institutional Review Board approval number 11–06-04). Tissue was stratified into two clinical groups: 1) women not in labor (NIL) having an elective c-section without complications and with intact membranes, a closed cervix, and a quiescent uterus, and 2) women in active labor exhibiting regular and forceful contractions coupled with documented cervical change with effacement and dilation greater than 4 cm who required c-section for reasons not related to the labor process (eg, breech presentations, cephalopelvic disproportion, fetal distress). Tissue was excised after delivery of the placenta and was immediately washed in ice-cold PBS. Myometrium in ice-cold PBS was carefully dissected from connective tissue, perimetrium, and decidua by microscope-aided dissection, and tissue fragments (3–4 mm³) were either snap frozen in liquid nitrogen and stored at −80°C, placed into 4% paraformaldehyde in PBS, or processed for explant tissue culture.

Explant culture

Immediately after dissection, myometrium fragments (1–2 mm³) were incubated in phenol-red free DMEM at 37°C in a 5% CO₂ incubator for up to 6 hours in the presence or absence of test substances.

Cell lines and cell culture

The hTERT-HM^A/B cell line was derived from the hTERT-HM immortalized human myometrial cell line [provided by Dr William Rainey, University of Michigan, Ann Arbor, Michigan (35)] as previously described (11). The cells contain stably incorporated PR-A and PR-B transgenes controlled by independent inducers (doxycycline [DOX] for PR-A and diacylhydrazine [DAH] for PR-B). Cells were maintained at 37°C in a 5% CO₂ humidified incubator in DMEM/Ham's F12 (1:1) supplemented with 5% charcoal-stripped fetal bovine serum, 1% penicillin-streptomycin, 0.1 mg/mL geneticin, and 2 mM L-glutamine. The T47D breast cancer cell line was cultured in RPMI 1640 medium supplemented in 10% fetal bovine serum, 1% penicillin-streptomycin, and 2 mM L-glutamine (Life Technologies).

DNA and RNA transfection

Cytomegalovirus promoter-driven expression plasmids for PR-A and PR-B and a PRE-controlled luciferase reporter plasmid (PRE-LUC) were provided by Dr Zafar Nawaz (University of Miami Sylvester; Braman Family Breast Cancer Institute, Miami, Florida). A constitutively active renilla luciferase expression plasmid (Promega) was used to normalize for transfection efficiency for luciferase assays. A short interfering RNA (siRNA) and scrambled control siRNA (Santa Cruz Biotechnology Inc) were used to knock down PR expression by RNA interference (RNAi). The PR siRNA was designed to induce degradation of all PR transcripts.

To examine the effect of pSer345 on PR-A function a site-directed mutagenesis kit (Agilent Technologies; catalog number 210518) was used to produce a PR-A expression plasmid in which serine-344 and serine-345 were converted to alanine, a serine-biomimetic amino acid that cannot be phosphorylated. Both serine residues were mutated because studies in breast cancer cells showed that phosphorylation of PRs at serine-345 is coupled with phosphorylation at the adjacent serine-344 (29), and in preliminary studies we found that alanine substitution at serine-345 failed to eliminate immunoreactivity with the pSer345 antibody (Cell Signaling Technology; catalog number 12783) in immunoblot assays, whereas substitution of both serine residues to alanine (^S344/345APR-A) eliminated pSer345 antibody reactivity (data not shown). Thus, the pSer345 antibody detects pSer344- and pSer345-PR, and as such it is possible that phosphorylation at both sites exists in the pregnancy myometrium.

hTERT-HM and hTERT-HM^A/B cells were transiently transfected with expression plasmids or siRNAs by nucleofection (Lonza Basal, Switzerland). Briefly, cells were harvested by trypsinization, centrifuged, and resuspended in smooth muscle nucleofection solution (Lonza; catalog number VPI-1004) to a concentration of 1 × 10⁶ cells per 100 μL containing DNA or siRNA to be transfected. The mixture was then transferred to an electroporation cuvette, placed in the Nucleofector device (Lonza), and subjected to program A33. Cells were then replated and allowed to stabilize for at least 16 hours.

Immunoblotting

Myometrium tissue specimens were homogenized using a bead mill (Bullet Blender STORM; Next Advance) in radioimmunoprecipitation assay buffer (Sigma), supplemented with protease (Roche) and phosphatase inhibitors (Sigma) and then centrifuged at 16 000 × g for 10 minutes at 4°C. Cultured hTERT-HM^A/B cells were washed in PBS, collected by scraping and centrifugation, and then lysed in radioimmunoprecipitation assay buffer containing protease and phosphatase inhibitors, and centrifuged at 16 000 × g for 10 minutes at 4°C. Supernatants were assayed for protein content using the BCA protein assay (Thermo Fisher Scientific). Lysates containing equal amounts of protein were diluted in gel-loading buffer (375 mM Tris-HCl; 6% sodium dodecyl sulfate; 48% glycerol; 9% betamercatoethanol; and 0.03% bromophenol blue, pH 6.8), heated for 5 minutes at 100°C, and subjected to denaturing SDS-PAGE on precast 4%–20% Tris-glycine polyacrylamide gels with the Novex electrophoresis system (Life Technologies). Proteins were then transferred to a polyvinylidene difluoride membrane (Millipore). For imumunodetection, membranes were first incubated in blocking buffer (5% nonfat milk in Tris buffered saline containing 0.1% Tween-20 [TBST]) at room temperature for 1 hour and then with primary antibodies (Supplemental Table 1) overnight at 4°C. The following day membranes were washed three times with TBST and incubated at room temperature for 1 hour with horseradish peroxidase-conjugated antimouse IgG or antirabbit IgG antibodies or infrared labeled antimouse and antirabbit secondary antibodies (for Li-Cor detection). After three washes with TBST, immunoreactive proteins were visualized using the HyGlo chemiluminescent horseradish peroxidase antibody detection reagent (Denville Scientific). Chemiluminescence was detected and quantified with the FluorChem E processor (ProteinSimple). Immunoreactivity in assays using Li-Cor infrared-labeled secondary antibodies was visualized using the Odyssey infrared imaging system (Li-Cor Biosciences).

Immunohistochemistry (IHC)

An immunoperoxidase-based IHC kit (immunoperoxidase secondary detection system; catalog number DAB150; Millipore Inc) was used on formalin-fixed, paraffin-embedded sections (5 μm) of term myometrium and hTERT-HM^A/B cells grown on glass coverslips. hTERT-HM^A/B cells were treated with DOX and DAH to induce expression of PR-A and PR-B and exposed to progesterone or vehicle overnight. After washing with PBS, cells were fixed with 4% paraformaldehyde. Tissue sections were deparaffinized and rehydrated in graded ethanol. Cells and tissue sections were subjected to antigen unmasking by incubation in 10 mM sodium citrate buffer (pH 6.0) for 30 minutes at 100°C, allowed to cool to room temperature, rinsed in Tris-buffered saline (TBS), and incubated in blocking solution (5% BSA in TBST) for 1 hour. Cells and tissue sections were then incubated with primary antibody diluted in signal stain antibody diluted in diluent buffer (catalog number 8112; Cell Signaling Technology) overnight at 4°C in a humidified chamber. The next day cells and sections were washed (3 × 5 min) in TBS with agitation. Cells and tissue sections were then incubated with biotin-conjugated secondary antibody for 30 minutes at room temperature, washed in TBS, incubated with streptavidin horseradish peroxidase solution for 30 minutes, washed in TBS, and incubated in 3,3′-diaminobenzidine for 5 minutes and then washed in TBS, mounted, and examined by light microscopy for immunoreactive signal indicated by brown 3,3′-diaminobenzidine staining.

RNA extraction and quantitative RT-PCR (qRT-PCR)

Total RNA was extracted from hTERT-HM and hTERT-HM^A/B cells using a NucleoSpin RNA kit (Macherey-Nagel). RNA integrity was determined using a TapeStation analyzer (Agilent Technologies), and samples with an RNA integrity number greater than 8 were used for downstream assays. RNA (300 ng) was reverse transcribed with random primers using Superscript IV reverse transcriptase (Thermo Fisher Scientific). Primers for IL-8 (forward, 5′-TGGCAGCCTTCCTGATTTCT-3′; reverse, 5′-TTAGCACTCCTTGGCAAAACTG-3′) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH; forward, 5′-TTGCCATCAATGACCCCTTCA-3′; reverse, 5′-CGCCCCACTTGATTTTGGA-3′) were designed using the Primer Express software (Applied Biosystems). qRT-PCR was performed using SYBR Green (Applied Biosystems) with a StepOnePlus real-time PCR system (Applied Biosystems). Abundance of mRNA relative to GAPDH was calculated using the δ cycle threshold (Ct) method (relative mRNA abundance = 2^{−(Ct gene of interest − Ct GAPDH)}).

Luciferase assay

Luciferase activity was measured in whole-cell lysate with the dual-luciferase reporter assay kit (Promega) on a GloMax 20/20 luminometer (Promega). Data were normalized to the renilla luciferase activity.

Statistical analyses

All experiments were performed in triplicate or greater. Data were subjected to a normalcy test, and groups were compared by the Wilcoxon rank-sum test. Normally distributed data were compared using the Student t test. Differences were considered statistically significant when the value was P < .05.

Results

pSer-PRs in the human pregnancy myometrium

Abundance of six pSer-PRs in term myometrium (seven samples from elective c-section deliveries at > 39 wk gestation) was determined by immunoblotting with site-specific pSer-PR antibodies (Figure 1 and Supplemental Table 1). Only anti-pSer345-PR (catalog number 12783; Cell Signaling Technology) produced a robust signal in protein lysate from term myometrium. The other pSer-PR antibodies produced a weak or barely detectable signal at the migration positions for PR-A and PR-B. Positive control assays using protein lysate from progesterone-treated T47D breast cancer cells produced the expected immunoreactive bands at the PR-A and PR-B position for each of the pSer-PR antibodies. In all tissues, antibodies reactive with PR-A/B (catalog number M3568; Dako) and PR-B (catalog number 3157; Cell Signaling Technology) produced immunoreactive bands corresponding to the expected SDS-PAGE position for PR-A and PR-B, showing that the PRs were present in all myometrium specimens. Immunoblotting for the constitutively expressed GAPDH confirmed equal loading of whole-cell lysate (80 μg protein) in each lane.

Specificity of anti-pSer345-PR

The identity of the anti-pSer345-PR immunoreactive protein(s) was verified using the hTERT-HM^A/B immortalized human myometrial cell line in which levels of PR-A and PR-B can be experimentally controlled (Figure 2A). In cells induced to express PR-A and PR-B with DOX and DAH, respectively, progesterone markedly increased anti-pSer345-PR immunoreactivity on protein(s) migrating at the PR-A position and to a lesser extent on protein(s) migrating at the PR-B position. Importantly, the increase in pSer345-PR induced by progesterone occurred only in cells induced to express PR-A and PR-B and was markedly decreased by the knockdown of total PR expression by RNAi. The effectiveness of RNAi for total PR knockdown was confirmed by immunoblotting with the PR-A/B antibody (catalog number M3568; Dako). Anti-pSer345-PR (catalog number 12783; Cell Signaling Technology) IHC immunoreactivity was strong in the nucleus of hTERT-HM^A/B cells expressing PR-A and PR-B and exposed to progesterone (Figure 2B). Only a weak IHC signal was detected in cells under basal conditions, and no immunostaining was detected in cells exposed to nonimmune IgG (data not shown).

Figure 2.

Validation of the pSer345-PR antibody specificity in hTERT-HM^A/B cells. Cells were treated to express PR-A and PR-B (DOX for PR-A and DAH for PR-B), transfected with siRNA targeting PR mRNA or scrambled siRNA, and then exposed overnight to progesterone (P4) or vehicle (ethanol). A, Immunoblot analysis for pSer345-PR (1° antibody; catalog number 12783; Cell Signaling Technology), PR-A/B (1° antibody, catalog number M3568; Dako), and GAPDH (1° antibody, catalog number 32233; Santa Cruz Biotechnology; used as the loading control) in whole-cell protein lysate from treated hTERT-HM^A/B cells. B, IHC using the pSer345-PR antibody in hTERT-HM^A/B cells induced to express PR-A and PR-B ± progesterone. Arrows show examples of strong nuclear immunoreactive staining (brown) in progesterone-treated cells. Bar, 25 μm.

Regulation of pSer345-PR by progesterone in hTERT-HM^A/B cells

Progesterone induced pSer345-PRA in hTERT-HM^A/B cells in a dose- (EC₅₀ ∼1 nM) and time-dependent manner (Figure 3). The abundance of pSer345-PRB was increased by progesterone at early time points (1–6 h) but was undetectable after 24 hours. In contrast, progesterone increased the level of pSer345-PRA and -PRB equally in T47D breast cancer cells after 16–24 hours.

Figure 3.

Regulation of pSer345-PR generation in hTERT-HM^A/B and T47D cells by progesterone (P4). hTERT-HM^A/B cells were induced to express PR-A and PR-B and then exposed to progesterone for various times (A) and at various concentrations (B). A, Immunoblot analysis for pSer345-PR (1° antibody, catalog number 12783; Cell Signaling Technology), PR-A/B (1° antibody, catalog number M3568; Dako), and GAPDH (1° antibody: catalog number 32233; Santa Cruz Biotechnology; used as the loading control) in whole-cell protein lysate from hTERT-HM^A/B cells and T47D cells exposed to progesterone (100 nM) for various times (right panel, typical immunoblot assay; left panel, quantified densitometry data for a typical time course). B, Dose response for progesterone induction of pSer345-PR in hTERT-HM^A/B cells treated with DOX and DAH to express PR-A and PR-B. Left panel, Immunoblot analysis of cells treated for 16 hours with various amounts of progesterone. Right panel, Replicate data (n = 3; mean ± SE) progesterone dose and time (15 min and 3 h) effect on pSer345-PRA induction. Note that both PR-A and PR-B were phosphorylated at serine-345 in response to progesterone in T47D cells, whereas in hTERT-HM^A/B cells progesterone-induced generation of pSer345-PR occurred predominantly on PR-A.

Progesterone decreased the rate of migration of both PRs, indicative of ligand-induced phosphorylation. In contrast to breast cancer cells, 24 hours of progesterone treatment in hTERT-HM^A/B cells increased the abundance of PR-A and decreased the abundance of PR-B. However, the extent to which progesterone increased pSer345-PRA was markedly greater than the extent to which it increased total PR-A abundance, especially at the early time points. Serine-345 phosphorylation of PR-A was markedly increased by progesterone treatment within 15 minutes, without a significant change in total PR-A pool.

Effect of labor status on pSer345-PRA abundance in term myometrium

Having established the specificity of the pSer345-PR antibody and that pSer345 occurred predominantly on PR-A in myometrial cells in response to progesterone, we next examined the association of pSer345-PRA abundance with labor status. Our initial immunoblot assays (Figure 1) used to screen multiple pSer-PR forms examined myometrium obtained from women at term (>39 wk) undergoing a scheduled c-section delivery who were in various stages of labor. To examine the association of labor status with pSer345-PRA abundance, we performed immunoblot analyses on a separate cohort of myometrium specimens from women clinically categorized as NIL (n = 6) and in labor (IL; n = 6). Consistent with our original assay results (Figure 1), immunoreactivity with anti-pSer345-PR occurred exclusively on PR-A. Abundance of pSer345-PRA relative to GAPDH and total PR-A increased significantly (P < .05) in association with the onset of active labor at term. As expected, labor was also associated with increased tissue-level inflammation as indicated by increased abundance of the phosphorylated RelA/p65 (pSer536-RelA/p65) subunit of nuclear factor-κB (NF-κB) relative to total RelA/p65 in IL compared with NIL myometrium (Figure 4, A and B). Additionally, pSer345-PR immunoreactivity was detected in the nucleus of myometrial cells and the intensity of IHC staining was qualitatively higher in IL compared with NIL myometrium (Figure 4C).

Figure 4.

Effect of labor status on the abundance of pSer345-PR in human term myometrium. Active labor was associated with increased abundance of pSer345-PRA and activation of NF-κB (indicated by increased abundance of Ser536-RelA/p65) A, Immunoblot analysis for total PR-A/B (1° antibody, catalog number sc-7208; Santa Cruz Biotechnology), pSer345-PR (1° antibody, catalog number 12783; Cell Signaling Technology), pSer536-RelA/p65 (1° antibody, catalog number 3033; Cell Signaling Technology), RelA/p65 (1° antibody, catalog number 6956; Cell Signaling Technology), and GAPDH (1° antibody, catalog number 32233; Santa Cruz Biotechnology; used as loading control) of total cell lysate from term NIL (n = 6) and IL (n = 6) myometrium. B, Quantitative analysis of effects of labor status on the abundance of pSer345-PR (relative to GAPDH and PR-A), the PR-A to PR-B ratio, and abundance of pSer536-RelA/p65 relative to RelA/p65. Data are from digital densitometry of the immunoblot assays shown in panel A. C, Location of pSer345-PR (and IgG negative control in adjacent sections) in representative sections of term NIL and IL myometrium. Positive immunoreactivity (arrows) is indicated by brown stain. Sections were counterstained with hematoxylin to identify cell nuclei (purple). BV, blood vessel. Bar, 100 μm.

Regulation of pSer345-PR in term myometrium

To examine the association between pSer345-PR and inflammation. The effect of progesterone and interleukin-1β (IL-1β), a proinflammatory cytokine that increases in term myometrium in association with the onset of labor (36), on pSer345-PR abundance in explant cultures of term myometrium was examined by immunoblotting using the Li-Cor infrared detection system (Figure 5). In this assay PR-A/B, pSer345-PR, GAPDH, and IL-8 (used a positive control for IL-1β induced gene expression) were simultaneously detected using secondary antibodies conjugated with markers that, in response to infrared excitation, fluoresce red (antimouse IgG) for detection of the PR-A/B; GAPDH, IL-8 antibodies, or green (antirabbit IgG) for detection of the pSer345-PR antibody. We found that in myometrial tissue prior to explant culture pSer345-PRA (green) produced a strong immunoreactive signal compared with total PR-A/B (red). This was likely due to higher affinity of the pSer345-PR antibody for its target than the affinity of the PR-A/B antibody for its target, and a generally higher sensitivity of the system for green fluorescence than red fluorescence. Abundance of PRs detected by both antibodies decreased to barely detectable levels within the first 6 hours of explant culture. This was a consistent finding in multiple explant experiments. Progesterone slightly increased pSer345-PRB and -PRA. IL-1β, alone increased IL-8 abundance but had no effect on pSer345-PRA and -PRB. In contrast, combined exposure of myometrial tissue to progesterone and IL-1β markedly increased the abundance of pSer345-PRA (Figure 5). Interestingly, progesterone and IL-1β markedly increased the abundance of immunoreactive proteins detected by the pSer345-PR antibody (green) without a detectable increase in the amount of protein detected by the PR-A/B antibody. This was expected because the total amount of PR proteins, although near the minimal level for detection with the PR-A/B antibody, was not changed between treatment groups. However, the proportion of pSer345-PRA in the PR pool was increased by progesterone and IL-1β treatment, and this was readily detectable with the pSer345-PR antibody.

Figure 5.

Regulation of pSer345-PR by progesterone and IL-1β in human term myometrium explant tissue. Exposure to progesterone and IL-1β induced pSer345-PRA generation in term myometrium and increased the proportion of pSer345-PRA in the PR pool. A, Immunoblot assay using the Li-Cor infrared system for PR-A/B (1° antibody, catalog number M3568; Dako; red), pSer345-PR (1° antibody, catalog number 12783; Cell Signaling Technology; green), IL-8 (1° antibody, catalog number MAB208; R&D Systems; red), and GAPDH (catalog number 32233; Santa Cruz Biotechnology; red) in whole-cell lysate from myometrium before explant culture (T0) and tissue explants exposed to media only (−), progesterone (P4; 100 nM), IL-1β (1 ng/mL), and IL-1β + P4 for 6 hours. Merged image (ie, red + green) is shown. Data are from two time-separated experiments using term myometrium from two patients and are representative of three experiments. B, Mean (±SEM; n = 3) pSer345-PRA relative to GAPDH assessed by digital densitometry in tissue exposed to P4 (100 nM) alone or P4 + IL-1β (1 ng/mL). C, Immunoblot analysis using the Li-Cor infrared system for PR-A/B (red) and pSer345-PR (green) in whole-cell protein lysate from T47D cells exposed to media alone (−) or media containing the progesterone agonist R5020 (10 nM) for 16 hours. Merger (red + green) image is shown.

Effect of pSer345 on PR-A function

Wild-type and ^S344/345APR-A were expressed in hTERT-HM cells (express only low levels of PR-A and no PR-B) in conjunction with wild-type PR-B (Figure 6A). Effects on the capacity for PR-A, in response to progesterone, to repress the transcriptional activity of PR-B was assayed by the following: 1) progesterone/PR-B-induced PRE-LUC activity and 2) progesterone/PR-B repression of IL-1β-induced IL-8 expression. In response to progesterone, PRE-LUC activity increased in cells expressing PR-B, and this was markedly decreased by coexpression of native PR-A. The transrepressive effect of PR-A was maintained in cells expressing ^S344/345APR-A (Figure 6B). Previous studies have shown that progesterone via PR-B exerts antiinflammatory effects in myometrial cells (11, 23). Because parturition involves tissue-level inflammation, it was proposed that progesterone promotes myometrial quiescence for most of pregnancy at least in part by PR-B-mediated suppression of inflammation in the myometrium. In this context the PR-A/PR-B hypothesis posits that at parturition PR-A inhibits the antiinflammatory activity of PR-B. We therefore examined whether pSer344/345 affects the capacity for PR-A to suppress progesterone/PR-B-mediated inhibition of myometrial cell responsiveness to an inflammatory stimulus. We found that in the absence of IL-1β, abundance of IL-8 mRNA was low and barely detectable in hTERT-HM cells and not affected by progesterone or PR status. As expected, IL-1β increased expression of IL-8. In cells expressing PR-B, progesterone inhibited IL-1β-induced IL-8 expression. PR-A partially inhibited the repressive effect of progesterone/PR-B. In contrast, ^S344/345APR-A did not affect the antiinflammatory activity of PR-B (Figure 6C).

Figure 6.

Effect of pSer345 on PR-A function in hTERT-HM cells. A, Validation of ^S344/345APR-A expression in hTERT-HM cells. Cells were transfected with empty vector, wild type PR-A (PR-A), or ^S344/345APR-A and exposed to progesterone (P4) or vehicle for 16 hours. In cells transfected with the native PR-A and ^S344/345APR-A expression plasmids, PR-A was detected by the PR-A/B (catalog number M3568; Dako). In contrast, the pSer345 antibody (catalog number 12783; Cell Signaling Technology) detected immunoreactive PR-A only in cells transfected with the wild-type PR-A. The PR-A signal in the ^S344/345APR-A lane is attributable to the low levels of PR-A expressed from the endogenous PR gene in hTERT-HM cells. B, Effect of pSer344/345 on PR-A transrepression of PR-B at the PRE-LUC reporter. hTERT-HM cells were transfected with wild-type PR-B and/or wild-type PR-A or ^S344/345APR-A. ^S344/345APR-A caused the same level of transrepression of PR-B as wild-type PR-A. C, Effect of pSer344/345 on PR-A transrepression of PR-B at the IL-8 promoter. hTERT-HM cells were transfected with wild-type PR-B and wild-type PR-A and/or ^S344/345APR-A and exposed to progesterone (100 nM) and IL-1β (10 ng/mL) overnight. PR-A inhibited PR-B repression of IL-β-induced IL-8 expression (assayed by qRT-PCR relative to GAPDH mRNA), and this effect was lost in cells expressing ^S344/345APR-A (n = 3; mean ± SE; groups with the same letters are not different; P > .05; P value is shown for other comparisons.)

Discussion

Progesterone withdrawal is considered to be the primary trigger for parturition, yet our understanding of how this occurs in women is unclear. A conundrum exists because, unlike most viviparous species, parturition in humans is not associated with a systemic fall in maternal progesterone levels. Although this suggests that human parturition is not triggered by progesterone withdrawal, labor in women is induced by disrupting progesterone/PR signaling with PR antagonist such as mifepristone (14). Thus, it is hypothesized that human parturition is triggered by a physiologically controlled change in PR function such that the capacity for progesterone to promote uterine quiescence and the block to labor is lost. This process is referred to as functional progesterone withdrawal, and various mechanisms have been proposed to explain how it occurs (37, 38). The present study tested the hypothesis that site-specific serine phosphorylation of the PRs is involved in altering PR function to cause functional progesterone withdrawal as part of the human parturition process.

Phosphorylation occurs at multiple serine residues on both PR isoforms and affects PR function by modulating sensitivity to progesterone, nuclear localization, transcriptional activity, and gene promoter targeting (27, 28, 39). This mode of PR regulation is thought to mediate pleiotropic actions of progesterone in different cell types. Using immunoblot assays with site-specific pSer-PR antibodies, we found that pSer345-PRA is a predominant immunoreactive phospho-PR form in term myometrium (Figure 1). Immunoreactivity signals for pSer81-PR, pSer190-PR, pSer294-PR, pSer400-PR, and pSer554-PR were weak, suggesting low levels of these PR phosphoforms in term myometrium. We therefore focused on pSer345-PR.

The specificity of the pSer345-PR antibody was confirmed using the hTERT-HM^A/B immortalized human myometrial cell line in which levels of PR-A and PR-B could be experimentally controlled. Those studies (Figure 2) showed that in myometrial cells pSer345-PR was induced by progesterone and was preferential for PR-A, consistent with immunoblot assays of term myometrium lysate. Even in cells induced to express high levels of both PR isoforms, progesterone increased the abundance of pSer345-PRA within 15 minutes with only a slight effect on pSer345-PRB abundance (Figures 2 and 3). This result suggests that in human myometrial cells, exposure to progesterone favors the generation of pSer345-PRA. The preferential induction of pSer345-PRA occurred in myometrial cells but not in T47D breast cancer cells in which progesterone increased abundance of pSer345-PRA and pSer345-PRB (Figures 3 and 5). This likely reflects cell type-specific differences in the protein kinases (and/or phosphatase) and protein degradation pathways that affect steady-state levels of pSer-PR isoforms and highlights the concept of differential site-specific PR serine phosphorylation as a mechanism for cell-specific modulation of PR activity. Importantly, the T47D cell data showed that the pSer345-PR antibody detects pSer345-PRB, confirming that the abundance of pSer345-PRB in pregnancy myometrium and hTERT-HM cells was low and not due to a failure of the immunoassay to detect pSer345-PRB.

Our data also confirm previous studies showing that phosphorylation at serine-345 on the PRs is accompanied by phosphorylation at the adjacent serine-344. We found that the pSer345-PR antibody (catalog number 12783; Cell Signaling Technology) exhibits affinity for pSer344-PR and pSer345-PR. Based on those findings, it is uncertain whether the immunoreactive signal detected by the pSer345-PR antibody in term myometrium and in hTERT-HM cells was due to pSer344, -345, or both. Nonetheless, it is clear that phosphorylation of PR at the serine-344/345 residues occurs in the human pregnancy myometrium and is ligand dependent in hTERT-HM cells. Therefore, pSer345 is hitherto referred to as pSer344/345.

We next examined the abundance and localization of pSer344/345-PR in term myometrium and whether these change in association with the onset of labor (Figure 4). In term myometrium abundance of pSer344/345-PRA increased in association with the onset of labor and localized exclusively to the nucleus of myometrial cells (Figure 4). Interestingly, pSer344/345-PRA abundance in myometrium correlated with increased tissue-level inflammation, as indicated by the abundance of the pSer536-RelA/p65 subunit of NF-κB. It has been shown that NF-κB is activated in myometrium as part of the parturition process and that expression of IL-1β, a major proinflammatory cytokine, in myometrium also increases in association with the onset of labor (40). These correlative data suggesting that the generation of pSer344/345-PR-A is linked to inflammation, led us to examine the effect of IL-1β on of pSer344/345-PRA generation in term myometrium explant cultures (Figure 5).

We found that in term myometrium tissue explants, IL-1β alone induced an inflammatory response, indicated by increased abundance of IL-8, but did not affect pSer344/345-PR abundance. In contrast to its effects in hTERT-HM^A/B cells and in T47D cells, progesterone alone induced only a slight increase in pSer344/345-PRA and -PRB in myometrium explants. However, exposure of myometrial tissue to progesterone and IL-1β markedly increased the abundance of pSer344/345-PRA while decreasing the abundance of pSer344/345-PR-B. These data suggest that the protein kinase activity that generates pSer344/345-PRs is constitutively active in hTERT-HM^A/B and T47D cells, and as such production of the pSer344/345-PRs is dependent only on ligand binding. In contrast, the protein kinase in term myometrium is inactive under basal conditions and activated by inflammatory stimuli. Our data also suggest that the IL-1β-activated kinase preferentially targets PR-A. In serum-starved T47D cells, the generation of pSer344/345-PR was found to be dependent on progesterone and induced by MAPKs, especially ERK1/2 (29). We have found that ERK1/2 is constitutively active in hTERT-HM^A/B cells under standard culture condition used in the present studies (data not shown). In this context where ERK1/2 phosphorylates PR at serine-344/345, progesterone alone is required for pSer344/345-PR generation in hTERT-HM cells. Taken together, these results suggest that generation of pSer344/345-PRA in pregnancy myometrium is ligand dependent and is induced by inflammatory cytokines associated with tissue-level inflammation at parturition. This would explain the low level of pSer344/345-PR in quiescent NIL myometrium even though the tissue was subjected to long-term (∼40 wk) exposure to relatively high levels of progesterone (41) and suggests that the protein kinase activity responsible for generating pSer344/345-PRA in response to ligand binding is inactive prior to the onset of labor and is activated by inflammatory stimuli as part of the parturition process.

The preferential effect of progesterone on the induction of pSer344/345 on PRA in myometrial cells was unexpected, and the mechanism for this effect is not clearly understood. A possible explanation is that in myometrial cells progesterone increases the stability of PR-A while decreasing the stability of PR-B, which would explain the increased abundance of PR-A and the decreased abundance of PR-B after long-term exposure to progesterone (see Figure 3: 24 h time point ± progesterone), leading to lower levels of pSer344/345-PRB. However, this effect cannot explain the rapid (within 15 min) increase in pSer345-PRA, but not pSer345-PRB, detected in hTERT-HM^A/B cells in response to progesterone. Clearly, in contrast to breast cancer cells, the process by which PR isoforms are phosphorylated in myometrial cells in response to progesterone favors the generation of pSer345-PRA and an increase in the proportion or pSer345-PRA in the PR-A pool (see Figure 4B: pSer345-PR-A/PR-A).

We examined the effect of pSer344/345 on PR-A function in myometrial cells and whether it may contribute to functional progesterone withdrawal. The PR-A/PR-B hypothesis for functional progesterone withdrawal posits that the PR-mediated progesterone block to parturition is functionally eliminated by PR-A transrepression of PR-B. We therefore examined whether pSer344/345 affects the capacity for PR-A to transrepress PR-B in hTERT-HM myometrial cells. This was achieved by comparing native PR-A, which is phosphorylated at serine-344/345 in response to progesterone, with a mutated PR-A that cannot be phosphorylated at serine-344/345 (^S344/345APR-A). Two readout parameters were used to assess effects on progesterone/PR-B activity: 1) PRE-LUC activity (ie, the direct transcriptional activity of PR-B at a canonical PRE), and 2) PR-B-mediated inhibition of IL-1β-induced IL-8 expression (ie, the antiinflammatory activity of PR-B at a non-PRE promoter). We found that pSer344/345 did not affect the capacity for PR-A to repress PR-B at the canonical PRE (Figure 6B). However, pSer344/345 was necessary for PR-A inhibition of antiinflammatory activity mediated by PR-B at the IL-8 promoter (Figure 6C). These data suggest that the capacity for PR-A to inhibit the antiinflammatory activity of PR-B at the IL-8 promoter is dependent on its phosphorylation at serine-344 and/or 345. It is also possible that ^S344/345APR-A, like PR-B, inhibits IL-1β-induced IL-8 expression and that phosphorylation at serine-344 and/or 345 inhibits this effect. Further studies are needed to resolve this issue. Nonetheless, it is clear that serine-344 and/or 345 of PR-A affects PR-mediated antiinflammatory activity of progesterone in human myometrial cells. Our data also suggest that progesterone/PR-A/B transcriptional activity at a canonical PRE promoter is distinct from that at a noncanonical promoter (eg, the IL-8 promoter), with repression of the latter by PR-A dependent on pSer344/345. Importantly, effects at the IL-8 promoter are relevant to parturition because the generation of pSer344/345-PRA may inhibit the capacity for progesterone to block tissue-level inflammation within the myometrium. Although the mechanism for this effect of pSer344/345 on PR-A activity at the IL-8 promoter is not known, studies in breast cancer cells suggest an indirect mechanism whereby pSer344/345 promotes functional interaction of PR to other transcription factors (29), leading to modulation of gene promoters that lack a PRE. Whether this is the basis for pSer344/345-PRA repression of PR-B at the IL-8 promoter is an intriguing possibility that requires further study.

Our data suggest that pSer344/345-PRA is generated in pregnancy myometrium in response to tissue-level inflammation. Based on the recognized causal link between inflammation and parturition (30, 31, 40, 42), we propose the following model (Figure 7): 1) for most of human pregnancy, progesterone, via the PRs, promotes uterine quiescence via various genomic mechanism; 2) proinflammatory/prolabor influences accumulate with advancing gestation leading to an increased inflammatory load (ie, the aggregate of multiple proinflammatory/prolabor signals) on the myometrium; 3) an inflammatory load threshold exists above which protein kinase activity in myometrial cells is activated to induce the ligand-dependent phosphorylation of PR-A at serine-344/345, leading to an increased abundance of pSer344/345-PRA, and 4) pSer344/345 induces the capacity for PR-A to inhibit antiinflammatory/progestational actions of progesterone mediated by PR-B. This model predicts that the timing of birth is determined by the set point for the inflammatory load threshold and/or the rate at which the inflammatory load increases. The increased incidence of preterm birth associated with infection/inflammation and fetal/maternal stress supports this concept (33). Therefore, therapeutics designed to alter the inflammatory load trajectory and/or uncouple inflammation and phosphorylation of PR-A at serine-344/345 in myometrial cells may be effective strategies to prevent preterm birth.

Figure 7.

Working model. We propose that for most of a pregnancy, progesterone acting via the PR-B promotes uterine quiescence mainly via antiinflammatory mechanisms. With advancing gestation, prolabor signals from multiple sources (some listed) increase the inflammatory load on the pregnancy uterus until a threshold is reached above which protein kinase pathways are activated in myometrial cells that induce pSer344/345-PRA. This changes the transcriptional activity of PR-A, using it to inhibit the antiinflammatory activity of PR-B, leading to tissue level inflammation and transition of the myometrium to the laboring phenotype.

We acknowledge that pSer344/345 may be only one of many posttranslational modifications that affect PR function in the pregnancy myometrium and suggest that unbiased assays, such as mass spectrometry, are needed to determine the full spectrum of pSer-PRs, and other posttranslationally modified PRs, in the pregnancy myometrium. However, this study provides an initial insight and suggests that ligand- and inflammation-induced phosphorylation of PR-A at serine-344/345 in myometrial cells plays a role in the physiology of human parturition.