Evaluating aminophylline and progesterone combination treatment to modulate contractility and labor‐related proteins in pregnant human myometrial tissues

Abstract Progesterone (P4) and cyclic adenosine monophosphate (cAMP) are regarded as pro‐quiescent factors that suppress uterine contractions during pregnancy. We previously used human primary cells in vitro and mice in vivo to demonstrate that simultaneously enhancing myometrial P4 and cAMP levels may reduce inflammation‐associated preterm labor. Here, we assessed whether aminophylline (Ami; phosphodiesterase inhibitor) and P4 can reduce myometrial contractility and contraction‐associated proteins (CAPs) better together than individually; both agents are clinically used drugs. Myometrial tissues from pregnant non‐laboring women were treated ex vivo with Ami acutely (while spontaneous contracting) or throughout 24‐h tissue culture (±P4); isometric tension measurements, PKA assays, and Western blotting were used to assess tissue contractility, cAMP action, and inflammation. Acute (1 h) treatment with 250 and 750 μM Ami reduced contractions by 50% and 84%, respectively, which was not associated with a directly proportional increase in whole tissue PKA activity. Sustained myometrial relaxation was observed during 24‐h tissue culture with 750 μM Ami, which did not require P4 nor reduce CAPs. COX‐2 protein can be reduced by 300 nM P4 but this did not equate to myometrial relaxation. Ami (250 μM) and P4 (100 and 300 nM) co‐treatment did not prevent oxytocin‐augmented contractions nor reduce CAPs during interleukin‐1β stimulation. Overall, Ami and P4 co‐treatment did not suppress myometrial contractions more than either agent alone, which may be attributed to low specificity and efficacy of Ami; cAMP and P4 action at in utero neighboring reproductive tissues during pregnancy should also be considered.


| INTRODUC TI ON
Optimal fetal development is achieved after 37-40 weeks of human (term) pregnancy. Globally, preterm birth (before 37 weeks of gestation) is a leading cause of death for children under the age of 5 1 and approximately 70% of all cases occur spontaneously. 2 Molecular mechanisms responsible for the onset of uterine contractions at labor are not fully established, which has hindered development of effective therapies to prevent preterm labor (PTL).
Labor onset is at least partly related to reduced activity of pathways that suppress myometrial contractile apparatus function and cell-to-cell connectivity, which include those regulated by cAMP and progesterone (P4). In myometrium, "P4 functional withdrawal" has been proposed to occur at the end of pregnancy, which involves changes in nuclear P4 receptor (PR) isoform levels and/or function in the absence of a decline in systemic P4 concentrations 3 to reduce P4-suppressed transcription of genes for contraction-associated proteins (CAPs). The ability of cAMP-driven protein kinase A (PKA) activity to prevent formation of actin-myosin complexes is expected to decrease at labor to permit phasic contractions; this is supported by previously observed labor-related changes in expression/activity of cAMP signalling proteins. [4][5][6] P4 prophylaxis is used for PTL prevention but confidence in its effectiveness is currently limited to pregnant women with premature shortening of the cervix; other risk factors, such as multifetal pregnancy, in the absence of a short cervix are less indicative of acceptable P4 efficacy. 7-9 β2-adrenergic receptor (β2-AR) agonists (i.e., betamimetics) that enhance intracellular cAMP concentrations have been used to reduce uterine contractions after PTL has started (i.e., tocolysis) but this has become less common due to maternal side effects. 10 Drawbacks of betamimetics include myometrial β2-AR desensitisation, 11 switch in G-protein coupling 12 and negative feedback by upregulation of cAMP-phosphodiesterases (PDEs). 13,14 An alternative route for enhancing myometrial pro-relaxation effects of cAMP is PDE inhibition. [15][16][17] Both betamimetics and PDE inhibitors are used as bronchodilators, where they have demonstrated dosedependent risks of side effects; however, importantly, their use in combination with steroids has been proposed to improve therapeutic efficacy and reduce side effects in this context. 18 Previously, we demonstrated that treating pregnant mice with a combination of aminophylline (water-soluble theophylline; Ami), a PDE inhibitor, and P4 prior to intrauterine injection of lipopolysaccharide (LPS), a pro-inflammatory inducer of murine PTL, can delay premature pup delivery longer than either agent alone. 19 Ami alone acutely promotes relaxation of ex vivo human myometrium. 20 Muscle relaxant effects of Ami are often attributed to its inhibitory action on cAMP-PDEs that make it a reference drug for bronchodilator design 21 ; although it can also act on cGMP-PDEs, adenosine receptors and histone deacetylases. 22 Ami is used to treat apnoea 23 and acute kidney injury 24 in preterm neonates, which suggests risk of harm related to maternal-fetal transfer 25 is low. Maternal tolerance of Ami was found to be better than ritodrine (betamimetic) in a study that also showed their tocolytic efficacies are similar. 26 Clinical trials for combining betamimetics with other tocolytics to suppress PTL have been conducted 27 but none published yet for PDE inhibitor and P4 combination therapy.
Here, we examined whether Ami and P4 together suppress human myometrial contractions better than either agent alone, by using a novel method for continuous tension recording during tissue culture (TC), and observed for their effect on PKA to assess extent of cAMP involvement. Our hypotheses were (i) Ami at concentrations that promote myometrial relaxation also increase "total" (i.e., whole tissue) PKA activity, (ii) 24-h Ami and P4 treatment suppresses spontaneous contractions and enhances pro-quiescent protein expression/activity more than either agent alone, and (iii) Ami and P4 co-treatment inhibits interleukin-1β (IL-1β)-induced changes in PR and CAPs expression/activity more than either agent alone.

| Materials
Ami from Sigma-Aldrich for acute treatments and Santa Cruz  (Table S1)

| Ethical approval and biopsy collection
This study conformed to standards set by the Declaration of Helsinki; ethics approval was granted by the Brompton and Harefield Research Ethics Committee (10/H0801/45), and written consent was provided by all participants prior to sample collection. One myometrium biopsy per woman (N = 67) was excised from the upper margin of incision made to the lower uterine segment at term gestation (weeks +days of pregnancy: 39 +0 median; 37 +1 to 40 +5 range) while not in labor during caesarean section for singleton birth (Tables S2 and S3). Biopsies were promptly immersed in DPBS for 2-4°C storage until dissected for isometric tension measurements (ITMs) and/or TC within the same day.
Each biopsy was dissected into tissue strips assigned to no more than two different types of experiments (Table S2), the details for which are described in the following subsections and summarized by Figure S1. Dimensions and force measurements of tissue strips are detailed in Tables S4 and S5.

| ITM in organ bath without TC
Each biopsy as one biological replicate was dissected into strips of longitudinal muscle in ice-cold DPBS and tied at both ends with polyester thread to mount onto MLT0201 isometric tension transducers (ADInstruments). After calibration, tissue length was manually adjusted to hold each strip at slack length as defined previously 28 before immersing in physiological saline solution (PSS; composition described previously 15 ) in an 8-channel organ bath (Panlab). Forces were recorded via transducers connected to a ML870 PowerLab unit and computer operating LabChart 7 (ADInstruments). All buffers were maintained at 37°C with 95% O 2 /5% CO 2 aeration. Data recording commenced once the last tissue strip of each experiment was lowered into its chamber. After 5 min, 29.4 mN tension was applied to all tissue strips. PSS was changed once every 60 min for the first 3 h while tissues equilibrated and spontaneous contractions stabilized before treatment. Contractions typically initiated within 1.5-2 h after start of data recording, and tissues that failed to produce stable contractions within 3 h were discarded; at least one tissue strip was treated with vehicle and the remaining strips were each treated with drug.
For acute cumulative concentration experiments, the last of these PSS changes was accompanied with the first concentration of Ami (sterile H 2 O vehicle) or CGS 15943 (DMSO vehicle), and each subsequent concentration of both agents was applied at 25-min intervals without changing PSS. After incubation with the last concentration of Ami or CGS 15943 (0.7% v/v DMSO in total at PSS), PSS at all tissues was exchanged with KPSS (PSS modified to contain 60 mM KCl) to promote maximum tissue depolarisation. After 10 min, KPSS was exchanged with fresh PSS twice and the last of these washes was kept at tissues for at least 10 min, before they were sequentially taken out of their chambers and cut from tied ends to retrieve their sample regions; excess PSS was removed before weighing. Methodology for 1-h exposure to bolus Ami concentrations was as described previously. 15 Data analyzed using LabChart 8 (ADInstruments) to calculate mean integral tension (i.e., area under the curve for force vs. time; MIT) and force per contraction; both normalized to cross-sectional area (CSA) 29 ; contraction frequency per 10 min was attained by manual counting of contraction peaks. For cumulative exposure to Ami or CGS 15943, data for each concentration of these agents were obtained from the last 15 min of treatment, and each normalized to the last 15 min of spontaneous contractions immediately prior to first drug application. For 1-h bolus Ami exposure, data were obtained from the last 30 min of treatment and normalized to the last 30 min of spontaneous activity prior to Ami addition.

| TC with simultaneous ITM
Each biopsy as one biological replicate was dissected into four tissue strips to accommodate all 24-h treatments per experiment; untreated tissues ("t = 0") from each biopsy were dissected at the same time to represent the pre-culture state and used only for comparison to 24-h vehicle controls. Dissection in DPBS and attachment to MLT0420 isometric tension transducers (ADInstruments) using polyester thread (pre-sterilized with ethanol) were undertaken within a biosafety cabinet before transfer into a TC incubator (maintained at 37°C with 95% air/5% CO 2 ).
Transducers were subsequently connected to a 4SP PowerLab unit and computer operating LabChart 7 (ADInstruments), and calibrated prior to setting all tissue strips to slack length. 28 Meanwhile, t = 0 tissues were briefly stored on ice, while their biopsy-matched experiment was being set up in the TC incubator, and flash frozen in liquid nitrogen promptly after Ami ± P4 was added to the last tissue strip. were added to tissues immediately after manual stretch stopped.
Each Ami, P4 and vehicle combination was mixed in its own sterile DMEM aliquot before adding to its tissue strip using a syringe.
Tissue lengths and DMEM were not manually adjusted for the rest of the experiment and the TC incubator was sealed for 24 h of un-

| TC for subsequent organ bath ITM (oxytocin stimulation)
Each biopsy as one biological replicate was dissected into six tissue strips in DPBS to accommodate all 24-h treatments per experiment, and maintained under ~4 mN isotonic tension in TC for 24 h by attaching 0.6-0.7 g mass as described previously. 15 Ami, P4, and their vehicles (sterile H 2 O and ethanol (0.03% v/v in DMEM), respectively) were added to DMEM immediately before immersing tissues to start 24-h TC. Samples of t = 0 tissues were stored as described for TC with simultaneous ITM experiments.
After 24-h TC, tissue strips were mounted onto transducers, which were connected to the same organ bath apparatus used for ITM without TC except LabChart 8 was operated for data acquisition; PSS in chambers was supplemented with P4 or its vehicle (ethanol; 0.3% v/v in PSS) immediately before immersing tissues set at slack length. Data recording for each experiment commenced once the last tissue strip was lowered into its chamber. After 5 min, all tissues were stretched to 1.5× slack length and no further manual adjustments were made. Following 2 h with no PSS change, the first con-

| Total protein extraction & PKA activity assay
Each frozen tissue segment was extracted for total protein content using a Precellys 24 bead-based homogenizer with CK-28R tubes (Stretton Scientific) as described previously. 15 Protein extracts were aliquoted to flash freeze in liquid nitrogen for PKA activity assays or snap frozen on dry ice for protein quantification using Bradford assay reagent (following manufacturer's protocol; BSA standards used) and immunoblotting.
PepTag cAMP-dependent protein kinase assay kit (Promega) was used to measure basal PKA activity as described previously, 15 when reliability of the assay for myometrial tissues in response to five different cAMP/PKA-enhancing agents was demonstrated. Agarose gels were imaged using a G:BOX Chemical XL system (Syngene) and analyzed by histogram-based densitometry using ImageJ v1.5 (National Institutes of Health; Bethesda, MD, USA; Research Resource Identifier (RRID) SCR_003070). Each set of assays included positive (purified PKA catalytic subunit) and background (assay buffer) controls.

| Immunoblotting
Total protein extracts (same as those prepared for PKA activity assays) were used as previously described for detecting Ser16phosphorylated and total heat shock protein 20 (HSP20), 15 as well as PR, CAPs (cyclooxygenase-2 (COX-2), connexin-43 (Cx43) and oxytocin receptor (OTR)) and housekeeping proteins (glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and β-tubulin) 31 ; 10 and 25 μg total protein per sample was used to detect CAPs and PR, respectively. Specificity validation for primary antibodies shown in Figure S2. Western blots were cut to allow simultaneous detection of targets and housekeeping proteins. Chemiluminescence imaging undertaken using a G:BOX Chemical XL system (Syngene). ImageJ v1.5 used for histogram-based densitometry as described previously. 31

| Data and statistical analysis
Details are provided in Appendix A, figures, and tables.

| Nomenclature of targets and ligands
Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guide topha rmaco logy.  Figure S3). Ami can also inhibit adenosine receptors 35 but CGS 15943, an adenosine receptor antagonist, at ≤10 μM had no effect on contractions ( Figure S4).
Bolus Ami treatment for 1 h reduced force generated per contraction relatively more so than contraction frequency ( Figure 1). An increase in total PKA activity in these tissues was most observable for 750 μM Ami with a mean fold difference of 1.6 (95CI: 1.2-2.0) compared to vehicle ( Figure 2A). HSP20 Ser16-phosphorylation, which is mediated by PKA and promotes actin depolymerisation, 16 was increased in 100 and 750 μM Ami-treated tissues, by mean fold differences of 1.3 (95CI: 1.1-1.6) and 1.7 (95CI: 1.1-2.3), respectively, relative to vehicle ( Figure 2B); total HSP20 abundance was unaffected.

F I G U R E 1
Concentration-dependent potency of aminophylline for acutely suppressing spontaneous contractions in human myometrial tissues. Isometric tension measurements from myometrial tissues (biopsies from term pregnant non-laboring women; N = 10), which were treated with aminophylline (Ami) or H 2 O vehicle for 1 h after establishing stable spontaneous contractions for ≥1 h. Each biopsy (a biological replicate) was dissected into tissue strips, where one strip was assigned vehicle control and their others treated with one concentration of Ami each; 1-3 Ami concentrations tested per biopsy, depending on how many produced stable baseline contractions. Data analyzed for mean integral tension (MIT), force per contraction and contraction frequency per 10 min, all for the periods of 30-min immediately prior to and before the end of Ami treatment for each experiment, which were used to calculate % activity; presented as mean ± SEM, where n = 8 for each Ami concentration paired with biopsy-matched vehicle (left panel). Representative contractility profiles are shown for each Ami concentration, along with vehicle control, all from tissue strips dissected from the same biopsy (right panel). Two-tailed Mann-Whitney test was used for each pair of biopsy-matched Ami concentration versus vehicle control comparison; *p ≤ .05, **p ≤ .01, ***p ≤ .001 [aminophylline] (µM) % mean contraction frequency per 10 min **

| Impact of 24-h Ami ± P4 on PKA activity, PR, and CAPs
Total PKA activity measured from whole tissue extracts after 24-h ITM ( Figure 4 and Figure S6) were all decreased relative to biopsymatched t = 0 (representing untreated (pre-culture) tissues to closely match the in utero state 38 ) except for tissues treated with 750 μM Ami. Less consistency in statistical outcome was observed from comparisons between only 24-h treated tissues, which showed Ami promoted a trend towards increased total PKA activity when compared to H + E; this effect of Ami was enhanced by P4 for all except 250 μM Ami ± 100 nM P4 conditions. HSP20 Ser16-phosphorylation in the same tissues ( Figure 4 and Figure S6) was increased by Ami at 750 μM to a more discernible extent than 250 μM. While "Ami without P4" versus "P4 without Ami", irrespective of Ami concentration, consistently showed that a higher amount of Ser16-phosphorylated HSP20 was present in the former; P4 did not enhance Ami-associated HSP20 Ser16-phosphorylation. Total HSP20 abundance ( Figure S7) was unchanged by all Ami ± P4 combinations in the same tissues.
PR abundance ( Figure 5 and Figure S8A) relative to t = 0 was not altered by H + E treatment during 24-h TC with ITM; PR-A was increased by 750 μM Ami and, to a lesser extent, decreased by 100 nM P4 relative to t = 0. Comparisons between only 24-h treated tissues showed P4 reduced both PR isoforms regardless of whether Ami was F I G U R E 2 Total PKA activity and HSP20 Ser16-phosphorylation in human myometrial tissues after 1-h aminophylline exposure during spontaneous contractions. Total protein extracts were prepared from tissue strips after isometric tension measurements, which were used to assess response to aminophylline (Ami) and its vehicle control during treatment for 1 h (i.e., same tissues represented by Figure 1). These were used for (A) PKA activity assays (purified catalytic PKA subunit (PKAc) used as a positive control and baseline background represented by "buffer control"), and (B) Western blotting for detection of Ser16-phosphorylated ("phospho(S16)") and total heat shock protein 20 (HSP20; a PKA substrate) along with glyceraldehyde 3-phosphate dehydrogenase (GAPDH; loading control). All data presented as mean ± SEM (n = 7 for 250 μM Ami, n = 8 for 100 and 750 μM Ami; each with biopsy-matched vehicle controls). Representative images of agarose gels for PKA assays, where "−P" and "+P" indicate non-phosphorylated and phosphorylated kemptide, respectively, and chemiluminescent Western blots are shown adjacent or above their associated histograms. Two-tailed paired Student's t test (after log transformation for PKA activity data) or Wilcoxon test was used for each biopsy-matched Ami concentration versus vehicle control comparison; *p ≤ .05, **p ≤ .01 also present. PR ratio (PR-A:PR-B; Figure 6 and Figure S8A) was also decreased by P4, which appeared most robust at 300 nM, with and without Ami, when compared to H + E and Ami (without P4); Ami co-treatment did not increase the ability of P4 to reduce PR-A:PR-B.
COX-2 and Cx43 abundance after 24-h TC with ITM were measured to represent myometrium P4-sensitive CAPs 39,40 (Figure 7 and Figure S8B and C). COX-2 was increased in all 24-h treated tissues relative to t = 0 except for those incubated with 250 μM Ami and either 100 or, arguably more so, 300 nM P4. Comparison between only 24-h treated tissues showed 300 nM P4, with and without Ami, robustly reduced COX-2 levels when compared to 250 μM Ami (without P4); the latter was not notably different to H + E. Cx43 was mostly increased during 24-h ITM in TC when compared to t = 0, but there were no differences between only Ami ± P4 combinations (Figure 7 and Figure S8C). OTR was another CAP measured for its abundance in the same tissues ( Figure S9) but showed no differences from the same comparisons tested for COX-2 and Cx43.

| Oxytocin response after 24-h Ami ± P4
Tissues treated with both 250 μM Ami and 100 nM P4 during 24-h TC with isotonic tension were subsequently the most spontaneously contractile (Figure 8). Oxytocin was used to assess whether 24-h Ami ± P4 can alter myometrial response to hormone-augmented contractions; no differences in contractility during cumulative exposure to oxytocin (normalized to pre-oxytocin spontaneous activity) were observed between all Ami ± P4 combinations used for 24-h TC; oxytocin EC 50 values are provided at Table 1.

| IL-1β influence on Ami ± P4
Tissues treated with IL-1β or its vehicle, with the same Ami ± P4 combinations used to test oxytocin response (Figure 8 Total PK A activity decreased during 24-h TC in H + E controls, irrespective of whether IL-1β was also present, when compared F I G U R E 5 Protein abundance for PR isoforms A and B in human myometrial tissues after 24-h aminophylline ± progesterone treatment. Total protein extracts were prepared from tissue strips after isometric tension measurements, which were used to assess response at spontaneous contractions to combinations of aminophylline (250 (250A) or 750 (750A) μM; H 2 O vehicle, H) ± progesterone (100 (100P) or 300 (300P) nM; ethanol vehicle, E) during 24-h treatment in tissue culture media (i.e., same tissues represented by Figure 3); "t = 0" (i.e., untreated) biopsy-matched tissues were also extracted. These were used for detection by Western blotting of progesterone receptor (PR), both isoforms A (PR-A) and B (PR-B), and β-tubulin (loading control); representative images of chemiluminescent Western blots shown at Figure 6 and Figure S8A). All data presented as mean ± SEM for 250A ± 300P (n = 8), 750A ± 100P (n = 10) and 250A ± 100P (n = 10); n equates to number of biopsies for each dataset.  Figure S10). HSP20 Ser16-phosphorylation levels were not different between t = 0 and 24-h H + E controls, irrespective of IL-1β co-treatment; whereas comparison between only 24-h treated tissues showed that it was increased by Ami (without P4) relative to H + E irrespective of IL-1β condition ( Figure 10A); total HSP20 abundance was not different for all comparisons ( Figure S10). COX-2 abundance was increased in H + E controls, both with and without IL-1β, when compared to t = 0 ( Figure 10B). COX-2 was also decreased in the absence of IL-1β, but not in its presence, by 250 μM Ami and 300 nM P4 co-treatment when compared to H+E. Cx43 and OTR levels were also measured but no differences were found from the same comparisons described for Ser16-phosphorylated HSP20 and COX-2 ( Figure S11). Myometrial inflammation (associated with labor) can be regulated by

F I G U R E 6 Ratio of PR isoforms A and B at protein level in human myometrial tissues after 24-h aminophylline ± progesterone treatment.
Total protein extracts were prepared from tissue strips after isometric tension measurements, which were used to assess response at spontaneous contractions to combinations of aminophylline (250 (250A) or 750 (750A) μM; H 2 O vehicle, H) ± progesterone (100 (100P) or 300 (300P) nM; ethanol vehicle, E) during 24-h treatment in tissue culture media (i.e., same tissues represented by Figure 3); "t = 0" (i.e., untreated) biopsy-matched tissues were also extracted. These were used for detection by Western blotting of progesterone receptor (PR), both isoforms A (PR-A) and B (PR-B), and β-tubulin (loading control); representative images of chemiluminescent Western blots are adjacent to its associated histogram for 250A ± 300P and Figure S8A for the other two sets of treatment combinations. Data for individual PR isoforms from which PR ratio values were calculated are shown at Figure 5. All data presented as mean ± SEM for 250A ± 300P (n = 8), 750A ± 100P (n = 10) and 250A ± 100P (n = 10); n equates to number of biopsies for whole dataset. PR-A (81 kDa) → cAMP and P4. 50 We previously showed that an adenylate cyclase activator, forskolin, can increase PR-B expression and activity in primary human myometrial cells. 51 Amini et al. has since demonstrated PR-A Ser345-phosphorylation, which is increased by IL-1β to enhance its trans-repressive effect on PR-B, can be reduced by forskolin in a PR-expressing human myometrial smooth muscle cell line. 52 Taken together, these observations suggest that enhancing myometrial P4 and cAMP signalling together prevents labor onset; whether this could be achieved using a clinically approved PDE-inhibiting drug (capable of enhancing cAMP) with P4 on human tissues had yet to be determined.
Here, we used novel 24-h ITM methodology to monitor ex vivo human myometrial contractions for real-time assessment of physiological changes in response to Ami ± P4. From this, we observed no discernible impact except for Ami at a concentration that is not clinically feasible (750 μM) but nevertheless aided proof of concept. Ami at 250 μM neither prevented contractions nor robustly promoted total PKA activity and HSP20 Ser16-phosphorylation;

F I G U R E 7
Protein abundance for COX-2 and Cx43 in human myometrial tissues after 24-h aminophylline ± progesterone treatment. Total protein extracts were prepared from tissue strips after isometric tension measurements, which were used to assess response at spontaneous contractions to combinations of aminophylline (250 (250A) or 750 (750A) μM; H 2 O vehicle, H) ± progesterone (100 (100P) or 300 (300P) nM; ethanol vehicle, E) during 24-h treatment in tissue culture media (i.e., same tissues represented by Figure 3); "t = 0" (i.e., untreated) biopsy-matched tissues were also extracted. These were used for detection by Western blotting of cyclooxygenase-2 (COX-2) and connexin-43 (Cx43) along with β-tubulin (loading control). previous bioavailability data 53,54 suggests this Ami concentration is also not clinically feasible unless myometrium-targeted drug delivery methods are used. 37 Ami is a weak pan-PDE inhibitor with a relatively poor therapeutic index 55  Although we had previously observed Ami combined with P4 suppresses LPS-induced preterm pup delivery in mice, 19 we had not monitored in vivo intrauterine pressure. We also had not an- P4 functional withdrawal is associated with PR-A dominance and increased inflammation. 62 However, we did not observe increased PR-A:PR-B in 24-h TC vehicle controls despite enhanced COX-2 abundance relative to t = 0 and even in the presence of IL-1β. 63 Furthermore, P4 (300 nM) reduced PR-A:PR-B in IL-1β-treated tissues without affecting COX-2 abundance. Together, these findings suggest that PR-A dominance is not a necessity for COX-2 upregulation and IL-1β may stabilize COX-2 more so than PR-A in ex vivo myometrium.
In agreement with our previous study, 31  tissues that influence labor status, and (iii) a different outcome would be observed in preterm myometrial tissues. These factors will need to be addressed in future investigations to improve evaluation of cAMP and P4 combination therapy for PTL prevention.

ACK N OWLED G EM ENTS
The authors thank all maternity unit patients at Chelsea & Westminster Hospital (London, UK), who kindly consented to provide myometrium biopsies used in this study, and to the NHS staff who assisted their collection. This work was financially F I G U R E 9 Protein abundance and ratio for PR isoforms A and B in human myometrial tissues after 24-h aminophylline ± progesterone ± IL-1β treatment. Total protein extracts were prepared from tissue strips after treatment in serum-free culture media with combinations of aminophylline (250 μM (250A); H 2 O vehicle, H) ± progesterone (100 (100P) or 300 (300P) nM; ethanol vehicle, E) ± interleukin-1β (1 ng/mL in H 2 O vehicle; IL-1β) for 24 h while maintained under isotonic tension (~4 mN); "t = 0" (i.e., untreated) biopsy-matched tissues were also extracted. These were used for detection by Western blotting of progesterone receptor (PR), both isoforms A (PR-A) and B (PR-B), and β-tubulin (loading control); representative images of chemiluminescent Western blots are shown for tissue strips from one biopsy. All data presented as mean ± SEM (n = 7); n equates to number of biopsies for each dataset.

Ethics for this study was approved by the Brompton and Harefield
Research Ethics Committee (London, UK); reference number 10/ H0801/45.

AUTH O R CO NTR I B UTI O N S
Study conceptualisation by PFL and MRJ; study design by PFL and RCY; all experiments, data analysis and writing of original draft manuscript undertaken by PFL; data interpretation, along with review, editing and approval of the manuscript's final version, by PFL, RCY, RMT and MRJ.

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. Some data may not be made available because of privacy or ethical restrictions.