Pro‐inflammatory signals induce 20α‐HSD expression in myometrial cells: A key mechanism for local progesterone withdrawal

Abstract Metabolism of progesterone (P4) by the enzyme 20α hydroxysteroid dehydrogenase (20α‐HSD) in myometrial cells is postulated to be a mechanism for P4 withdrawal, which occurs concomitant to uterine inflammation (physiologic or infection‐induced) and associated activation of transcription factors: NF‐кB and AP‐1, common to term and preterm labour. We found that 20α‐HSD protein is significantly increased in human myometrium during term labour, and in mouse uterus during term and preterm labour. Treatment of human myometrial cells with the pro‐inflammatory mediators, lipopolysaccharide (LPS, mimicking infection) and 12‐O‐tetradecanoylphorbol‐13‐acetate (TPA, mimicking inflammation), induced 20α‐HSD gene expression and increased 20α‐HSD protein abundance. LPS treatment decreased P4 release into the culture medium and resulted in up‐regulation of GJA1 in the hTERT‐HM cells. The NF‐кB /AP‐1 transcription factors mediated effects of LPS and TPA on 20α‐HSD gene transcription. Both pro‐inflammatory stimuli induced 20α‐HSD promoter activity in LPS/TPA‐treated cells which was significantly attenuated by inhibition of NF‐кB (JSH: 20 µM) or AP‐1 signalling (T5224: 10 µM). Deletion of NF‐кB consensus sites abrogated LPS‐mediated promoter induction, while removal of AP‐1 sites reversed the TPA‐mediated induction of 20α‐HSD promoter. We conclude that inflammatory stimuli (physiologic or pathologic) that activate NF‐кB or AP‐1 induce 20α‐HSD transcription and subsequent local P4 withdrawal resulting in up‐regulation of GJA1 and activation of myometrium that precedes labour.


| INTRODUC TI ON
The steroid hormone progesterone (P4) is essential for the establishment and maintenance of pregnancy, and in all eutherian species examined so far, P4 withdrawal by decline in P4 levels or disruption of P4 signalling initiates labour. Human parturition that occurs without a systemic decline in maternal P4 levels is speculated to be triggered by local P4 withdrawal in the uterine tissues. 1,2 Local P4 withdrawal can occur by changes in signalling mediated by the nuclear P4 receptor (PR) isoforms, PR-A and PR-B or by increased metabolism of P4 to a PR-inactive form in uterine target cells. Our previous data suggest a combination of these mechanisms whereby local metabolism/inactivation of P4 within the myometrium leads to the loss of P4/PR-B function and dominance of ligand-independent PR-A activity that induces expression of genes encoding contraction-associated proteins (CAPs) that augment myometrial contractility. 2 Local intracrine P4 withdrawal can be mediated by the P4-metabolizing enzyme, 20α-hydroxysteroid dehydrogenase (20α-HSD), 1,2 which belongs to the aldo-keto reductase (AKR) superfamily and acts on a variety of substrates, including steroid hormones and endogenous prostaglandins. 3,4 In particular, human 20α-HSD (encoded by AKR1C1, murine homologue: Akr1c18) catalyses the conversion of P4 into 20α-hydroxyprogesterone (20α-OHP), a relatively weak progestin compared with P4. 5 It has been reported earlier that compared to early gestation (13th week), the P4/20α-OHP ratio is decreased in term (40-42nd week) human myometrium suggesting increased P4 metabolism with advancing gestation. 5 The association of 20α-HSD with labour is evident from studies in rodents. In rats, expression of 20α-HSD in ovaries/corpus luteum negatively correlated with the circulating P4 levels, with luteolysis being associated with increased 20α-HSD expression and the fall in P4 that initiates labour. 6,7 In mice, the knockout of the transcription factor (TF) STAT5b caused early abortion due to increased expression and activity of 20α-HSD in corpus luteum and subsequent P4 withdrawal. 8,9 Moreover, 20α-HSD gene knockout in mice increased P4 levels and delayed parturition, 9,10 suggesting a central role of the enzyme in labour onset. This may be critical in facilitating local P4 withdrawal in the myometrium to trigger labour. In lower mammals despite the decline in peripheral P4, maternal P4 levels at term remain well above the dissociation constant (Kd) for binding to PRs. 11 Similarly in humans, increased expression and abundance of myometrial 20α-HSD 1 has been associated with local withdrawal of P4 and P4/PR function in myometrium. 1,2 It is plausible that PTB is caused by premature induction of 20α-HSD activity in myometrial cells, thus understanding how 20α-HSD is regulated is of paramount importance.
The mechanisms by which 20α-HSD is up-regulated in term myometrium 1,2 in the presence of high P4 levels is unclear since its transcription is suppressed by P4/PR signalling during gestation. 12,13 A possible explanation is the presence of other factors that induce 20α-HSD expression in the term myometrium. Physiologic inflammation 14 as a result of maternal and foetal causes, such as mechanical stretch 15 and foetal lung maturation, [16][17][18] that increase cytokines/chemokine levels and corresponding immune cells infiltration might be one such factor. 19,20 There is ample evidence that in both mice and humans, pro-inflammatory transcription factors (TFs), AP-1 21,22 and NF-кB, 16 are induced in the myometrium during term and preterm labour. In humans, cervical expression of AP-1 and NF-кB has shown to be increased prior to parturition. 23 Since systemic P4 levels remain elevated in humans prior to and during labour, which negatively regulates 20α-HSD expression [reviewed in 24 ], we hypothesized that uterine inflammation causes 20α-HSD induction in myometrium and that the pro-inflammatory TFs, AP-1 and/or NF-кB, play a role in the transcriptional regulation of the gene encoding 20α-HSD.
In this study, we examined the myometrial expression of 20α-HSD in pregnant women and mice, in association with term and preterm labour and determined whether it is affected by proinflammatory stimuli and by the activity of NF-кB or AP-1 TFs. We found that myometrial levels of 20α-HSD increase in association with term and preterm labour and that the activation of NF-кB or AP-1 induces AKR1C1 transcription. Our findings suggest a functional link between uterine inflammation and local P4 withdrawal that triggers labour.

| Ethical approval for human and murine studies
The study involving human tissues was approved by the Research Ethics Board of Mount Sinai Hospital, Toronto, Canada (REB #18-0168-A). All patients provided a written consent to participate in the study. All murine experiments were approved by the institutional Animal Care Committee (AUP # 21-0164H). Hsd:ICR (CD-1) outbred timed pregnant dams were purchased from Harlan Laboratories (http://www.harlan.com/) and housed under specific pathogen-free conditions at the Toronto Centre for Phenogenomics (TCP) on a 12L:12D cycle and were administered food and water ad libitum.

| Murine models of labour; Term labour (TL) model
Female mice were mated overnight with males, and the day of vaginal plug detection was designated gestational day (GD) 0.5 of pregnancy. The average time of delivery was the early morning of GD19.
Mice were killed by carbon dioxide inhalation, and uteri were collected at 10 AM on all days with the exceptions of the labour sample (TL) that was collected once the animals had delivered at least one pup from average number of 14 in two uterine horns. Myometrial tissue was collected from 4 animals on GD8 (early gestation), GD15 (mid-gestation), GD19 (term not in labour, TNIL), and GD19/20 (term labour, TL).

| LPS-induced PTL
The lipopolysaccharide (LPS) used for this study was isolated from Escherichia coli, serotype O55:B5 (Sigma-Aldrich). On GD15, mice underwent mini-laparotomy under general anaesthesia (isoflurane) with intrauterine infusion of 125 μg LPS in 100 μL of sterile saline (LPS group) or 100 μL sterile saline (Sham group) between two amniotic sacs close to the cervix. Animals (n = 3 per group) were killed during LPS-induced PTL (12-24 hours after the infusion) or 24 hours after a saline injection.

| Human tissues
Healthy pregnant women with a singleton pregnancy undergoing elective caesarean delivery at term (gestational age ≥37 weeks) were recruited as 'term not in labour' (TNIL, n = 5). Caesarean delivery of 'term in labour' (TL, n = 5) women was performed after the onset of labour

| Murine tissues
Mice were killed by carbon dioxide inhalation at specific gestational days and during TL or PTL. The part of uterine horn close to cervix from the horn from which a foetus was expelled (post-partum tissue) was removed; the remainder was collected for analysis. For biochemical analyses, uterine horns were bisected longitudinally in ice-cold PBS. The decidua basalis and decidua parietalis was removed as described previously. 25 Myometrial samples were snapfrozen in liquid nitrogen and stored at −80℃ for protein analysis. For immunohistochemistry analyses (IHC), one whole uterine horn was fixed in 10% NBF or 4% PFA and processed as described in (A) above.

| Protein extraction and immunoblotting
Myometrial tissues were crushed on dry ice, homogenized in lysis buffer [0.08 M Tris/HCl (pH 6.8), 2% SDS, 10% Glycerol] with freshly added protease and phosphatase inhibitor cocktail (Thermo Fisher Scientific Inc) using Tissuelyser II (Qiagen). hTERT-HM cells were lysed on ice in lysis buffer (same as above), vortexed for 15 seconds twice and incubated on ice for 10 minutes in between. Homogenates from tissues and cell lysates were then sonicated on ice for 10 seconds and centrifuged at 20,000 × g for 25 minutes at 4℃, and supernatant was collected and stored at −20℃ till further processing. Protein concentration was determined by BCA (Thermo Fisher Scientific Inc). Immunoblotting was performed as described earlier. 22 Briefly, equal amount of protein was separated by SDS-PAGE and transferred to a polyvinylidene difluoride membrane (Trans-blot Turbo Midi PVDF, Bio-Rad) using Turbo Trans-Blot system (Bio-Rad). After blocking for an hour with 5% milk prepared in TBS-T, the membranes were incubated with primary antibody at 4℃ for overnight. The membranes were subsequently incubated with horseradish peroxidase (HRP)-conjugated secondary antibody at room temperature for 1 hour. Signals were detected using Luminata HRP substrate (Millipore) or SuperSignal ™ West Femto Maximum Sensitivity Substrate (Thermo Fisher Scientific), and imaging was performed with ChemiDoc imaging system (Bio-Rad). Antibodies used for immunoblotting are listed in Table 1. Immunoblotting for ERK2 was used as loading control. Densitometric analysis was performed using Image Lab system (Bio-Rad).

| Immunofluorescence
Paraffin-embedded human myometrium tissues were sectioned at 5 μm thickness, and slides were baked at 60℃ overnight. The

Antibody
Catalogue # Dilution Source

| Identification of putative NF-кB and AP-1 consensus binding sites in human AKR1C1 promoter
We

| Statistical analysis
The Student t test was used to determine the differences between two groups (non-labouring versus labouring samples or control versus treatment). Differences among several groups were determined by one-way analysis of variance (ANOVA), followed by Dunnet's multiple comparison test using Prism software (GraphPad Prism).
Two-way ANOVA and Tukey's or Sidak's post-tests were used to compare different variables. Differences were considered significant only where P values were less than .05.

| 20α-HSD is up-regulated in myometrium during labour
The onset of labour at term was associated with increased abundance of 20α-HSD protein in human myometrium as assessed by immunoblotting (P = .03, n = 7, Figure 1A) and immunofluorescence ( Figure 1B). In mice, 20α-HSD protein abundance in myometrium assessed by immunoblotting was low during early gestation (GD8) and increased with the advancing gestation (GD15) peaking at GD19 (term not in labour, TNIL) and during TL ( Figure 2A). Myometrial tissue collected during PTB induced by LPS or RU486 also had higher levels of 20α-HSD protein compared with sham/vehicle controls (LPS: P = .04, Figure 2B; RU486: P = .04, Figure 2C).

F I G U R E 2 Myometrial 20α-HSD increases with advancing gestation and during preterm labour in mice. A, Representative
Western blots and densitometric analysis of 20α-HSD protein in mouse myometrium from A) early gestation (gestational day 8, GD8), mid-gestation (GD15), term not in labour (GD19, TNIL) and term labour (TL). Graph represents mean relative optical density of 20α-HSD protein with respect to ERK2 ± SEM, n = 4 per GD. '**' denotes statistical significance at P < .01 and '***' at P < .001 for respective groups vs GD8, determined through ordinary oneway ANOVA followed by Dunnett's multiple comparisons test. B, LPS-induced preterm labour model where GD15 mice underwent mini-laparotomy under general anaesthesia with intrauterine infusion of 125 µg LPS in 100 µL of sterile saline (LPS group) or intrauterine infusion of 100 µL sterile saline (Sham group). Graph represents samples from GD16: sham (non-labouring) and LPS (labouring, n = 3). C, RU486-induced preterm labour model where GD15 mice received intraperitoneal injection of progesterone antagonist RU486 (150 µg). Graph represents samples from GD16: vehicle (non-labouring) and RU486 (labouring n = 4). Densitometry data show relative optical density of 20α-HSD protein with respect to total ERK2, represented as mean ± SEM, '*' denotes statistical significance at P < .05. Unpaired t test was used to determine the differences between the treatment vs control group   Figure 3C). Furthermore, LPS increased GJA1 mRNA (encodes CX43) ( Figure 3D) and GJA1 protein in hTERT-HM cells ( Figure 3E).

| Pro-inflammatory stimuli induce human AKR1C1 promoter activity in vitro, NF-кB mediates the effect of LPS while AP-1 transcription factors mediate the effect of TPA on human AKR1C1 promoter activity
To assess the transcriptional regulation of human 20α-HSD by the pro-inflammatory stimuli, the human 20α-HSD promoter (−886 to +43) with a luciferase reporter (pWT) was transfected into human HEK293T cells and treated with LPS or TPA. We found that LPS (0.25-10 µg/mL) and TPA at ≥20 ng/mL were able to induce 20α-HSD promoter activity above the basal transcription levels (P < .05; Figure 4A).
We determined that LPS treatment of hTERT-HM cells induced phosphorylation of NF-kappa-B inhibitor alpha (IκBα) protein ( Figure 4B) and subsequent nuclear translocation of NF-кB, which was blocked by the JSH (NF-кB inhibitor) ( Figure 4C). TPA treatment induced phosphorylation and activation of AP-1 factor cFOS protein, which was blocked by the AP-1 inhibitor T5224 ( Figure 4D). The luciferase activity of LPS-and TPA-induced AKR1C1 promoter was decreased by NF-kB and AP-1 inhibitors, respectively ( Figure 4E). LPS-induced AKR1C1 promoter-luciferase activity was significantly reduced by the deletion of the NF-κB response elements but not by deletion of the AP-1 response elements ( Figure 5B, left). TPA-induced AKR1C1 promoter-luciferase activity was inhibited by deletion of AP-1 response elements ( Figure 5B, right).  The AP-1 family of TFs were previously implicated in regulation of 20α-HSD gene promoter activity in the monkey. 42 In this study, we determined that the 5′ flanking region of AKR1C1 (−860 to +43) is responsive F I G U R E 6 Schematic representation of putative molecular mechanism by which different inflammatory stimuli regulate P4 withdrawal and myometrial activation. 20α-HSD expression is up-regulated in myometrium in response to 1) infection (LPS), which activates the NF-кB transcription factor and mediates its nuclear translocation through phosphorylation and subsequent degradation of IκBα in TLR4/ MyD88/TRAF-dependent pathway, or 2) inflammatory stimuli (TPA), which mediate phosphorylation, activation and translocation of AP-1 factors via PKC/PI3K pathways. Up-regulation of 20α-HSD expression causes decrease of P4 levels and induces transcription of CAPs (such as GJA1/Cx43), resulting in enhanced myometrial cell connectivity through the formation of gap junction channels, activation of myometrium and labour onset. The inflammatory signals (LPS and TPA) generate a secondary response via secretion of multiple cytokines and chemokines by myometrial cells to elicit autocrine and paracrine positive feedback loop, thereby amplifying the inflammatory response. Lipopolysaccharide (LPS), 12-O-tetradecanoylphorbol-13-acetate (TPA), 20alpha hydroxysteroid dehydrogenase (20α-HSD), nuclear factor-κB (NF-кB), activator protein 1 (AP-1), inhibitor of nuclear factor kappa B (IкBα), Toll-like receptor 4 (TLR4), myeloid differentiation primary response protein (MyD88), tumour necrosis factor receptor (TNFR)-associated factor 6 (TRAF6), protein kinase C (PKC), phosphoinositide 3-kinase (PI3K), progesterone (P4), contraction-associated proteins (CAPs). Note that this figure was created with Biorender.com

PRETERM LABOUR TERM LABOUR
to LPS and TPA. Moreover, blocking NF-кB or AP-1 activity, using specific inhibitors, attenuated AKR1C1 promoter activity suggesting that both TFs regulate human AKR1C1 transcription. The targeted deletion of NF-кB and AP-1 consensus sites revealed that NF-кB mediates infection (LPS)-induced AKR1C1 transcription, while the AP-1 pathway is critical for inflammation (TPA)-induced AKR1C1 transcription. Given that both AP-1 and NF-кB are activated in myometrium prior to labour, 21,22,36,43,44 we suggest that these TFs are pivotal in the up-regulation of 20α-HSD and consequent localized P4 withdrawal in myometrial cells. We have previously found that the AP-1 specific dimer composition differentially regulates gene expression and that the activity of nuclear AP-1 heterodimers (FRA1/JUNB and FRA2/JUND) is associated with human labour. 22 Elucidation of the specific role of different AP-1 dimers in the regulation of 20α-HSD transcription will be important to delineate the molecular regulation of AKR1C1 expression.
Taken together, our findings provide a molecular mechanism linking uterine inflammation and myometrial P4 withdrawal, which precedes labour onset (summarized in Figure 6). We propose that up-regulation of myometrial AKR1C1/20α-HSD during labour is conserved in term (mice and human) and preterm parturition (mice). We found that inflammatory signals, driven by mimics that are widely accepted as representing infection and inflammation, induce AKR1C1 expression through the activation of AP-1 or NF-кB TFs. The resultant increase in 20α-HSD may cause local P4 withdrawal that remove inhibition of CAP gene expression, especially GJA1, that cause the contraction of labour by increasing myometrial cell contractility and excitability. We suggest that present data on local metabolism of P4 in myometrium explain the conflicting results on the effectiveness of clinical P4 treatment aimed at preventing PTB. [45][46][47] It may be possible, therefore, to clinically inhibit labour and prevent PTB using anti-inflammatory therapeutics that inhibit the activation of AKR1C1 expression and 20α-HSD activity in myometrial cells, thus preserving the relaxatory and pro-gestational actions of P4.

ACK N OWLED G EM ENTS
We thank Dr Caroline Dunk for her technical advice on immunofluorescence and Mr Adam Boros-Rausch for the illustration in Figure 6 created using BioRender's tool for scientific illustrations (biorender. com).

CO N FLI C T O F I NTE R E S T S
The authors have no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.