Oxytocin-stimulated NFAT transcriptional activation in human myometrial cells.

Oxytocin (OXT) is a peptide hormone that binds the OXT receptor on myometrial cells, initiating an intracellular signaling cascade, resulting in accumulation of intracellular calcium and smooth muscle contraction. In other systems, an elevation of intracellular Ca(2+) stimulates nuclear translocation of the transcription factor, nuclear factor of activated T cells (NFAT), which is transcriptionally active in arterial and ileal smooth muscle. Here we have investigated the role of NFAT in the mechanism of action of OXT. Human myometrial cells expressed all five NFAT isoforms (NFATC1-C4 and -5). Myometrial cells were transduced with a recombinant adenovirus expressing a NFATC1-EFP reporter, and a semi-automated imaging system was used to monitor effects of OXT on reporter localization in live cells. OXT induced a concentration-dependent nuclear translocation of NFATC1-EFP in a reversible manner, which was inhibited by OXT antagonists and calcineurin inhibitors. Pulsatile stimulation with OXT caused intermittent, pulse-frequency-dependent, nuclear translocation of NFATC1-EFP, which was more efficient than sustained stimulation. OXT induced nuclear translocation of endogenous NFAT that was transcriptionally active, because OXT stimulated activity of a NFAT-response element-luciferase reporter and induced calcineurin-NFAT dependent expression of RGS2, RCAN1, and PTGS2 (COX2) mRNA. Furthermore, OXT-dependent transcription was dependent on protein neosynthesis; cycloheximide abolished RGS2 transcription but augmented RCAN1 and COX2 transcriptional readouts. This study identifies a novel signaling mechanism within the myometrium, whereby calcineurin-NFAT signaling mediates OXT-induced transcriptional activity. Furthermore, we show NFATC1-EFP is responsive to pulses of OXT, a mechanism by which myometrial cells could decode OXT pulse frequency.

The uterus primarily functions as a quiescent environment for the developing fetus during pregnancy, before switching from a relaxed to a contractile state at term resulting in parturition. Complications arise when parturition occurs too early or too late, but more often when it is premature. Ten percent of births worldwide are preterm (defined as delivery preceding 37 wk gestation) (1) with 85% of infant morbidity and mortality attributed to prematurity (2). Interestingly, nearly 50% of preterm births are idiopathic, with no obvious etiology (1), and are therefore potentially preventable (3, 4). However, the mechanisms by which labor is initiated have yet to be fully elucidated, making intervention and treatment of preterm labor problematic (5, 6).

Oxytocin (OXT) is a potent uterotonic hormone, released from the posterior pituitary in a pulsatile manner. As parturition progresses, OXT pulse frequency and duration increase (7), thus enhancing the frequency and force of myometrial contractions. Although the clinical use of OXT in the induction and augmentation of labor is well established (8, 9), its importance in initiating parturition physiologically is debatable. Animal experiments have shown that OXT-deficient mice exhibit normal labor and delivery, although the pups do not survive due to loss of the milk ejection reflex (10, 11). This suggests that OXT may play an essential role in the establishment of lactation, but only a permissive role during parturition. In women, there is an increase in OXT receptor (OXTR) density in the myometrium during late pregnancy; however, there is no further increase near the onset of labor (12, 13), indicating that OXTR density is not a defining factor in initiating parturition. Maternal plasma OXT levels are not elevated before the onset of parturition in women (14); however, in addition to being released into the circulation by the posterior pituitary, OXT is also synthesized and secreted by the decidua, with OXT mRNA and protein expression in this tissue increasing prepartum (15–17). This suggests a possible paracrine mechanism by which locally synthesized OXT may be involved in the initiation of parturition at term.

OXT binds to the OXTR, a G protein-coupled receptor (GPCR), on myometrial cells to initiate an intracellular signaling pathway culminating in intracellular calcium ([Ca2+]i) accumulation (18, 19), and thus uterine smooth muscle contraction through activation of myosin light chain kinase (20–22). In conjunction with contraction, Ca2+ signaling also mediates activation of certain transcription factors, i.e. nuclear factor κ-light-chain-enhancer of activated B cells (NFκB), c-JUN N-terminal kinases and nuclear factor of activated T cells (NFAT). NFAT was first identified in T cells, being essential for induction of cytokine gene expression during the immune response (23). Outside the immune system, NFAT has been physiologically implicated in skeletal (24) and cardiac (25, 26) muscle hypertrophy, skeletal muscle growth and development (27, 28), heart valve formation (29), and vascular development (30–32). To date, there have been no studies on the function of NFAT in the myometrium; however, experiments in mice have shown that inhibition of NFAT signaling at term delays delivery by several hours (33), suggesting that the NFAT pathway may be involved in the initiation of parturition.

Five members of the NFAT family of transcription factors have been isolated: NFATC2 (NFAT1/p) (34), NFATC1 (NFAT2) (35), NFATC4 (NFAT3) (36), NFATC3 (NFAT4/x) (36–38), and NFAT5 (TonEBP) (39, 40). NFATC1–4 signaling is mediated by the Ca2+/calmodulin-dependent phosphatase calcineurin (41), whereas NFAT5 activity is activated by osmotic stress (40). In resting cells, the four calcineurin-sensitive NFATC1–4 transcription factors are predominantly localized to the cytosol in a phosphorylated state. On elevation of [Ca2+]i, calcineurin induces NFAT dephosphorylation, unmasking nuclear localization signals (42) and stimulating cotranslocation of the calcineurin-NFAT complex to the nucleus (43). In the nucleus, NFAT binds the NFAT-binding motif in DNA promoter regions and cooperates with cofactors, i.e. activator protein 1, globin transcription factor 1, and myocyte enhancer factor 2 to enhance gene expression (44, 45). As [Ca2+]i declines, NFAT is rephosphorylated by kinases such as c-JUN N-terminal kinase and glycogen synthase kinase 3 (46–48), promoting nuclear export of NFAT via a chromosome region maintenance protein 1/exportin mechanism (49).

Pulsatile release of hormones is important for regulating physiological processes and coordinating Ca2+ signaling and thus transcription within a cell; i.e. pulses of OXT determine luteolysis and maintenance of pregnancy in the sheep (50), and pulses of GnRH regulate gene expression in the gonadotrophs (51, 52). A pulsatile system is physiologically advantageous in comparison with sustained release of a hormone for several reasons: intermittent stimulation of a receptor does not result in desensitization and internalization of the receptor, as with sustained stimulation (53), and allows for recycling of the receptor to the surface between stimulations for a maximal response. Moreover, the energy cost to the cell of intermittently elevating [Ca2+]i is far less than the energy required to maintain elevated [Ca2+]i levels for a sustained period of time. Furthermore, pulse frequency is an important mechanism for encoding transcriptional responses; i.e. GnRH pulse frequency is decoded within gonadotrophs to stimulate differential transcriptional readouts (51), and a similar system may be involved in GH-regulated gene transcription (54). NFAT nuclear shuttling has been shown to be dependent on Ca2+ oscillation and hormonal pulse frequency, with the greater the frequency, the greater the amplitude of NFAT nuclear localization (51, 55, 56). This suggests that NFAT is an important mechanism in decoding pulse frequency to differentially regulate gene transcription.

In this report, we demonstrate for the first time that NFAT is involved in the mechanism of action of OXT in human myometrial cells. OXT induces translocation of NFAT to the nucleus via a calcineurin-mediated mechanism, with NFAT nuclear localization mirroring pulses of OXT in a frequency-dependent manner. We show that pulsatile OXT can provide a more efficient signal for NFAT activation than a sustained stimulus and that OXT-induced gene expression is dependent on calcineurin-NFAT activity.

Materials and Methods

Reagents

DMEM, fetal calf serum (FCS), collagenase type II, dispase, penicillin, streptomycin Hoechst nuclear stain, fluo4 Ca2+ indicator, β-mercaptoethanol, and primers were ordered from Invitrogen (Paisley, UK). Deoxyribonuclease, elastase, atosiban, coenzyme A, and ATP was purchased from Sigma (Dorset, UK). L368899, cyclosporin A (CyclA), FK506, and cycloheximide (CX) were purchased from Tocris (Bristol, UK). OXT was from Calbiochem (Darmstadt, Germany), ionomycin from Ascent Scientific (Abcam, Cambridge, UK), and luciferin from Promega (Southampton, UK). All basic chemicals were from Sigma unless otherwise stated.

Antibodies

Primary antibodies were mouse monoclonal anti-NFATC1 (7A6) antibody (SC7294) from Santa Cruz Biotechnology (Santa Cruz, CA) and rabbit polyclonal antidesmin (ab15200), mouse monoclonal anti-actin (ab18147), and rabbit monoclonal anticalponin (ab46794) from Abcam (Cambridge, UK). Secondary antibodies were Alexa fluor 488 goat antirabbit IgG (heavy and light chains) from Invitrogen, and DyLight 488 AffiniPure donkey antimouse IgG (heavy and light chains) from Jackson ImmunoResearch Laboratories (Suffolk, UK).

Myometrial cell dispersion and culture

This study was approved by the South West Research Ethics Committee. Human myometrial biopsies were obtained, with informed consent from premenopausal women at hysterectomy. Myometrial cell isolation and culture was adapted from the protocol described by Phaneuf et al. (57). Briefly, dissected tissue was incubated at 37 C, with gentle shaking, in DMEM (4.5 g/liter glucose, glutamate), with 300 U/ml collagenase type II, 0.3 U/ml dispase, 30 U/ml deoxyribonuclease, and 0.09 U/ml elastase for 1 h. The supernatant was spun for 10 min at 378 × g, and the cell pellet was resuspended in DMEM supplemented with 10% FCS and 1% penicillin-streptomycin. After two additional incubations, collected cells were pooled into T75 flasks for culture. Cultured cells were used experimentally from passages 3–10.

IN Cell Analyzer 1000 image acquisition

Imaging experiments were performed using the IN Cell Analyzer 1000 (GE Healthcare, Buckinghamshire, UK) semiautomated imaging platform as previously described (51). Images were acquired in a single field of view for live cell imaging and three fields for fixed cell imaging. Experiments were performed in triplicate wells with 50–200 cells per field.

Measurement of [Ca2+]i

Myometrial cells were seeded at 5000 cells per well in black-walled 96-well sample plates (Appleton Woods, Birmingham, UK). Next day, cells were washed before loading with 2.5 μm fluo4 Ca2+ indicator and 0.5 μm Hoechst nuclear stain, along with other compounds if required (as described in figure legends), in phenol red-free DMEM (10% FCS, 25 mm HEPES) and incubated at 37 C for 30 min. After incubation, sample plates were loaded into the IN Cell Analyzer 1000 for imaging. Live cell populations were imaged before stimulation with OXT (as described in figure legends), using a robotic needle, with subsequent images acquired every 12 sec for 60 sec.

NFATC1-EFP nuclear translocation

Myometrial cells were seeded at 5000 cells per well in black-walled 96-well plates. Next day, cells were transduced with a NFATC1-EFP adenovirus (GE Healthcare, Buckinghamshire, UK), at 2 × 104 plaque-forming units (pfu)/μl in DMEM (2% FCS) for 16–24 h. On the day of experiment, cells were loaded with 0.5 μm Hoechst nuclear stain in physiological salt solution (PSS) [127 mm NaCl, 0.5 mm NaH2PO4, 1.8 mm CaCl2, 2 mm MgCl2, 5 mm KCl2, 5 mm NaHCO3, 10 mm HEPES (pH 7.5), 0.1% BSA, 10 mm glucose] and incubated at 37 C for 30 min. For live cell experiments, cell populations were imaged before stimulation with OXT, with subsequent images being acquired at the relevant time points after stimulation (see figure legends). For repeated stimulation experiments, cells were washed five times with PSS between 5-min stimulations with OXT (the frequency and amplitude of stimulation is described in the figure legends). For fixed cell imaging, cells were pretreated with the relevant compound (as described in figure legends) and stimulated with OXT or ionomycin before terminating the reaction with ice-cold PBS, fixation in 4% paraformaldehyde (PFA) for 20 min, and imaging.

IN Cell Analyzer image analysis

Image analysis and fluorescence quantification was performed using the IN Cell Analyzer Work station version 3.5 software (IN Cell Investigator; GE Healthcare). Cytoplasmic and nuclear regions were defined by green channel (fluo4 and NFATC1-EFP) and blue channel (Hoechst nuclear stain) images, respectively. The fluorescence readouts for each region were used to calculate cell population and individual cell fluorescence. For graphical representation and statistical analysis, fluo4 nuclear intensity (background subtracted) was standardized to resting cell nuclear intensity, and NFATC1-EFP cytoplasmic and nuclear intensity were used to calculate the nuclear to cytoplasmic ratio (N:C) as previously described (51).

Immunocytochemistry

Myometrial cells were seeded at 5000 cells per well in black-walled 96-well plates. Next day, cells were washed with PBS, fixed in 4% PFA for 20 min, and permeabilized with 0.1% Triton X-100 for 10 min. Nonspecific primary antibody binding was blocked with normal goat serum (Invitrogen) (desmin and calponin) or normal donkey serum (Sigma) (NFATC1 and actin) for 2 h. Cells were treated with primary antibody for 24 h at 4 C. Primary antibodies used were NFATC1 (1:100), desmin (1:150), actin (1:400), and calponin (1:150). Next day, cells were washed in PBS six times before incubation with goat antirabbit (1:200) (desmin and calponin) and donkey antimouse (1:200) (NFATC1 and actin) secondary antibody. Cells were again washed with PBS six times before staining with Hoechst nuclear stain for 30 min and image acquisition by the IN Cell Analyzer 1000 as previously described. Image analysis was by IN Cell Analyzer work station version 3.5 software as previously described. Primary antibody cytoplasmic fluorescence was defined by green channel images. A filter was set for cytoplasmic intensity at 260, 250, and 300 arbitrary fluorescent units for desmin, actin, and calponin images, respectively, to calculate the percentage of positive cells for each primary antibody. For graphical representation and statistical analyzes, a NFATC1 N:C was calculated as previously described and normalized to the untreated control.

Luciferase assay

Myometrial cells were seeded at 5000 cells per well in black-walled 96-well plates. Next day, cells were transduced with a NFAT response element luciferase reporter (NFAT-RE-Luc) (Addgene, Cambridge, UK; plasmid 10959) and β-galactosidase (β-Gal) adenovirus as previously described (51, 58) at 1 × 104 pfu/μl and 5 × 103 pfu/μl, respectively, in DMEM (2% FCS) for 16–24 h. Cells were treated with inhibitors where necessary and stimulated with OXT in PSS at the desired concentration and time (see figure legends) before the reaction was terminated with ice-cold PBS, treated with lysis buffer, and frozen then thawed to aid lysis. Luciferase reagent [33.3 mm dithiothreitol, 270 μm coenzyme A, 530 μm ATP, 470 μm luciferin in assay buffer containing 20 mm Tricene, 0.1 mm EDTA, 1.07 mm (MgCO3)4Mg(OH)2, and 2.67 mm MgSO4] was added to cell lysates to assess luminescence with a Lumat LB 9507 luminometer (Berthold Technologies, Hertfordshire, UK). β-Gal activity was used to correct for adenovirus transduction efficiency, being measured by incubation of the remaining lysate with chlorophenol-red-galactopyranoside (Roche Applied Sciences, West Sussex, UK) mix (878 μm chlorophenol-red-galactopyranoside, 175 μm β-mercaptoethanol, 60 mm Na2HPO4, 40 mm NaH2PO4, 10 mm KCl, and 1 mm MgSO4) at 37 C for 48 h, with β-Gal activity measured using a microplate reader (Revelation version 4.22; Dynex Technologies, West Sussex, UK). Light units were standardized to β-Gal activity and normalized to control for graphical representation and statistical analysis.

Quantitative real-time PCR

Myometrial cells were seeded at 100,000 cells per well in six-well plates and grown to 60–80% confluency. Cells were treated with OXT, plus or minus inhibitors, for the appropriate times (as described in figure legends). The stimulations were terminated with ice-cold PBS (ribonuclease free) before addition of lysis buffer (50 μm mercaptoethanol). RNA was isolated using an RNeasy mini kit (QIAGEN, West Sussex, UK), according to the manufacturer's protocol. Quantity and quality of RNA was assessed by the NanoDrop 1000 full spectrum 220/750-nm spectrophotometer (Labtech International, East Sussex, UK). cDNA was synthesized using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA) following the manufacturer's instructions and diluted 4-fold on reaction completion. Gene expression was quantified by real-time PCR (qPCR) with Power SYBR Green Master Mix (Applied Biosystems, Foster City, CA), using a 7500 real-time PCR system (Applied Biosystems, Foster City, CA). A template of 2 μl cDNA was run per reaction in a final reaction volume of 20 μl (75 nm forward and reverse primers, Table 1) for 45 cycles of 95 C for 15 sec and 60 C for 60 sec. A dissociation stage followed to ensure the amplification of a single product. Relative mRNA expression levels were quantified by the Pfaffl method of relative quantification (59), as used by Peeters et al. (60), and standardized to the housekeeping genes 18S and RNA polymerase II. Expression levels were normalized to control values for graphical representation and statistical analysis.

Table 1.

Primer pairs for qPCR

Gene nomenclature	Forward primer (5′–3′)	Reverse primer (5′–3′)
RNA18S1 (18s)	(1332) ATGGCCGTTCTTAGTTGGTG	(1548) CGCTGAGCCAGTCAGTGTAG
POLR2A (RNAPOLII)	(4453) GCACCACGTCCAATGACATTG	(4719) GTGCGGCTGCTTCCATAAGC
NFATC2 (NFAT1)	(2114) AAGAGCCAGCCCAACATGC	(2220) CGTTTTCTCTTCCCATTGATGAC
NFATC1 (NFAT2)	(1302) CTGTGCAAGCCGAATTCTCTGG	(1379) ACTGACGTGAACGGGGCTGG
NFATC4 (NFAT3)	(2189) GTCCTGATGGGAAGCTGCAATGG	(2258) AGCGTCACCTCGTTGCTCTGC
NFATC3 (NFAT4)	(303) GCGGCCTGCAGATCTTGAGC	(404) TGATGTGGTAAGCAAAGTGGTGTGGT
NFAT5	(2458) GACACTGGCGGTGGACTGCG	(2517) CTGGCTTCGACATCAGCATTCCT
RGS2	(236) CTTTCATCAAGCCTTCTCCTGA	(299) GCTAGCAGCTCGTCAAATGC
RCAN1 (DSCR1)	(493) GCTCAGACCTTACACATAGGAAGC	(598) CCACTTGTTTCCATCCCACTG
PTGS2 (COX2)	(590) CAAAGGTAAAAAGCAGCTTCCTG	(671) CTGGGGATCAGGGATGAACT

The gene sequence position of each primer's 5′-terminal nucleotide is indicated in parentheses preceding each primer sequence.

Promoter sequence analysis for NFAT-binding motifs

Regulator of G protein signaling 2 (RGS2), regulator of calcineurin 1 (RCAN1), and cyclooxygenase 2 (COX2) promoter sequences were derived using the UCSC genome browser (UCSC Genome Bioinformatics, University of California, Santa Cruz, CA). The promoter sequences were entered into the Meme Suite program (61) and using the Fimo Algorithm (62) scanned for NFAT-binding motifs. Statistical significance was taken as P < 0.001.

Statistical analysis

Statistical analysis was performed with GraphPad Prism version 5 (GraphPad Software, La Jolla, CA) using one-way or two-way ANOVA with Dunnett's or Bonferroni post hoc test as indicated. Results shown are mean ± SEM, with P < 0.05 considered statistically significant. All experiments were repeated at least three times using cells from three or more separate donors.

Results

Characterization of myometrial cell phenotype

The phenotype of myometrial cells in our culture system were validated using primary antibodies for the smooth muscle cell markers desmin, actin, and calponin (Supplemental Fig. 1, published on The Endocrine Society's Journals Online web site at http://mend.endojournals.org). From a sample size of 1200–2400 cells, 95.4 ± 1.3% of cells were positive for desmin, 94 ± 0.28% positive for actin, and 99.1 ± 0.33% positive for calponin. Within the no-primary-antibody control group, 96.3 ± 0.54% of cells were negative for desmin, 99 ± 0.003% negative for actin, and 98.4 ± 0.9% negative for calponin.

Expression of NFAT isoforms in human myometrium

Because there has been no previous description of NFAT expression in myometrial smooth muscle, we designed probes to identify NFAT isoforms by real-time PCR and found that NFATC1, NFATC2, NFATC3, NFATC4, and NFAT5 were expressed in cultured myometrial cells. The calcineurin-insensitive NFAT5 was expressed in greatest abundance in comparison with the calcineurin-dependent NFATC1–4 (P < 0.001). The levels of expression of the calcineurin-sensitive isoforms were NFATC4 more than NFATC1, and NFATC3 (P < 0.01) and NFATC4 more than NFATC2 (P < 0.001) (Fig. 1).

Fig. 1.

NFAT mRNA expression in human myometrial cells. Real-time PCR was performed on cDNA collected from cultured human myometrial cells. Data shown are relative mRNA copy number, mean ± se, from five independent donors, calculated by the Pfaffl method of analysis and normalized to the housekeeping genes 18S and RNA polymerase II. Statistical analysis was performed by one-way ANOVA, with Bonferroni multiple-comparison post hoc test. Statistics shown are in comparison with relative NFATC4 expression. *, P < 0.01; **, P < 0.001.

OXT-induced NFATC1-EFP nuclear translocation in human myometrial cells

The NFAT signaling pathway is activated by an elevation of [Ca2+]i in arterial and ileal smooth muscle cells (63, 64). To confirm OXT induction of Ca2+ signaling in myometrial cells, a fluo4-acetoxymethyl ester (AM) Ca2+ indicator was used. Treatment of myometrial cells with OXT caused a concentration-dependent (EC50 = 1.18 ± 0.18 nm) and time-dependent increase in [Ca2+]i (Fig. 2A). The response was maximal 12 sec after stimulus at the higher concentrations (10−6 and 10−7 m) and 24 sec for the lower concentrations (10−8-10−10 m), with fluo4-AM intensity returning to near baseline within 1 min (Fig. 2A).

Fig. 2.

OXT induces accumulation of [Ca2+]i and NFATC1-EFP nuclear translocation. Human myometrial cells were loaded with fluo4-AM (A) or transduced with a NFATC1-EFP adenovirus (B) and stained with Hoechst nuclear stain before imaging. Cells were stimulated with OXT, in a concentration-dependent manner at the concentration (molar) and time indicated. The concentration of OXT used (molar) is indicated by the symbols, which apply to both A and B. The control (Ctrl) group was not stimulated with OXT. Live cell image acquisition and analysis was performed as described in Materials and Methods. Representative images shown are before and after treatment with OXT (10−7 m). Scale bar, 60 μm. Data shown are fluo4 intensity (background subtracted) normalized to time zero (A) and NFATC1-EFP intensity (background subtracted) N:C ratio (B), mean ± se, from three separate experiments and independent donors. Statistical analyses were performed by one-way ANOVA, with Dunnett's post hoc test, which show statistical difference compared with the untreated control in terms of area under the curve in A and B, as follows: for a concentration of OXT of 10−9 m, P < 0.01, and for 10−8-10−6 m, P < 0.001.

To assess the effect of OXT on NFAT activity, myometrial cells were transduced with a recombinant adenovirus expressing a NFATC1-EFP imaging reporter. In unstimulated cells, NFATC1-EFP was largely cytoplasmic; however, on treatment with OXT, NFATC1-EFP translocated to the nucleus (Fig. 2 and Supplemental Fig. 1). This effect was quantified by an automated imaging system, which defined the NFATC1-EFP fluorescence intensity in the cytoplasm and nucleus to calculate the NFATC1-EFP (N:C) intensity ratio in cell populations and individual cells (Supplemental Fig. 2). Sustained treatment with OXT induced a concentration-dependent increase in NFATC1-EFP N:C ratio (EC50 = 1.14 ± 0.15 nm) (Fig. 2B) within the cell population, which paralleled the effect of OXT on fluo4-AM intensity. The effect was slower in onset than [Ca2+]i accumulation, with NFATC1-EFP translocating to the nucleus within 5 min. The NFATC1-EFP N:C ratio was maximal from 30–40 min and remained sustained for at least 70 min (Fig. 2B).

OXT-induced translocation of NFATC1-EFP to the nucleus was inhibited by the OXTR antagonists atosiban and L368899 (Fig. 3A) and by the calcineurin inhibitors CyclA and FK506 (Fig. 3B). Pretreatment with the intracellular Ca2+ chelator BAPTA [1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid]-AM or the use of Ca2+-free PSS blocked NFATC1-EFP nuclear translocation in response to OXT (Fig. 3C). The Ca2+ ionophore ionomycin also induced NFATC1-EFP nuclear localization, the effect being suppressed by CyclA, FK506, BAPTA-AM, and Ca2+-free media (Fig. 3, B and C) but not inhibited by L36899 or atosiban.

Fig. 3.

OXT-induced NFATC1-EFP nuclear translocation is inhibited by OXT antagonists, calcineurin inhibitors, and inhibition of Ca2+ signaling. Human myometrial cells were transduced with a NFATC1-EFP adenovirus before pretreatment with OXT antagonists atosiban (10−5 m) and L368899 (10−5 m) (A), calcineurin inhibitors CyclA (10−6 m) and FK506 (10−6 m) (B), and the [Ca2+]i chelator BAPTA-AM (10−6 m) and replacement of PSS with Ca2+-free solution (C) for 30 min. Cells were stimulated with OXT (10−7 m) or ionomycin (2 μm) for 60 min before terminating the reaction with ice-cold PBS, fixing in 4% PFA, and staining with Hoechst. Fixed cell image acquisition and analysis were performed as described in Materials and Methods. Data shown are NFATC1-EFP intensity (background subtracted) N:C ratio, mean ± se, from three separate experiments and independent donors. Statistical analyses were performed by one-way ANOVA with Dunnett's post hoc test compared with the untreated control. *, P < 0.01; **, P < 0.001.

Effect of pulsatile OXT on NFATC1-EFP nuclear translocation

To investigate whether NFATC1-EFP nuclear translocation is dependent on sustained OXTR occupancy or triggered by the initial binding of OXT to its receptor, we measured responses in the presence of brief or sustained OXT exposure. Myometrial cells were treated with OXT for 5 min followed by a repeated wash period to remove the agonist (Fig. 4). Within the initial 5-min stimulation, NFATC1-EFP translocated to the nucleus (35% of maximal response seen with sustained receptor occupancy). After the wash, there was no additional increase in the NFATC1-EFP N:C ratio, with the signal gradually declining toward baseline within 40 min. This shows that OXT-induced nuclear translocation of NFATC1-EFP is reversible and requires constant receptor occupancy for a sustained signal. This reversibility raised the possibility that pulsatile OXT treatment would, in turn, cause pulsatile NFATC1-EFP translocation responses, so we examined the effect of repeated 5-min pulses of OXT on NFATC1-EFP nuclear localization. Myometrial cells were subjected to 5-min stimulations with OXT, followed by five washes, over a time course of 3 h, at varying concentrations and pulse frequency. Each exposure to OXT caused a concentration-dependent translocation of NFATC1-EFP to the nucleus, with similar kinetics, i.e. maximal response at 5 min, followed by a gradual decline of NFATC1-EFP N:C after washing, despite the variance in amplitude (Fig. 5A) and pulse frequency (Fig. 5, B and C). Treatment of cells with a pulse of OXT every 30 min resulted in the NFATC1-EFP signal not returning to baseline between pulses, resulting in a cumulative effect (Fig. 5B). Furthermore, the areas under the curve of the NFATC1-EFP N:C traces in response to sustained treatment with OXT and pulsatile stimulation (30-min wash periods) over a 3-h time course were of similar magnitude (Fig. 5D).

Fig. 4.

OXT induction of NFATC1-EFP nuclear translocation is reversible. Human myometrial cells were transduced with a NFATC1-EFP adenovirus and stained with Hoechst nuclear stain before imaging. Cells were stimulated with either sustained OXT (10−7 m) for 70 min, or briefly, 5 min, followed by five washes (gray rectangle). The control (Ctrl) group was not stimulated with OXT. Images were acquired at the time points indicated. Data shown are NFATC1-EFP intensity (background subtracted) N:C ratio, mean ± S.E. from three separate experiments and independent donors. Statistical analysis was by two way ANOVA which demonstrated that brief vs. sustained stimulation with OXT was a significant source of variation, P < 0.01.

Fig. 5.

NFATC1-EFP nuclear translocation during pulsatile OXT treatment. Human myometrial cells were transduced with a NFATC1-EFP adenovirus and stained with Hoechst nuclear stain before imaging. A–C, Cells were stimulated with either sustained or 5-min pulses of OXT at the indicated concentration (molar) (A) or at 10−7 m (B and C) and indicated time points, followed by washes (five times) (gray rectangle). D, Histogram of area under the curves of NFATC1-EFP traces in B and C. Ctrl, Control; PF, pulse frequency. Live cell image acquisition and analysis were performed as described in Materials and Methods. Data shown are NFATC1-EFP intensity (background subtracted) N:C ratio, mean ± se, from at least three separate experiments and independent donors. In A, the y-axis is offset for clarity so the traces are clearly distinguishable, the amplitude of the offset is as follows: 10−9, +0.5 N:C; 10−8, +1 N:C; 10−7, +1.5 N:C. Statistical analysis was by two-way ANOVA with Bonferroni post hoc test, which demonstrated that OXT concentration was a significant source of variation at P < 0.05 (A), and one-way ANOVA with Dunnett's post hoc test compared with the untreated control: *, P < 0.05; ** P < 0.01 (D).

OXT stimulates endogenous NFAT nuclear translocation and transcriptional activity

Because OXT stimulated translocation of a NFATC1-EFP reporter construct to the nucleus of myometrial cells, we investigated the effect of OXT on endogenous NFAT nuclear localization using a NFATC1 primary antibody. A 30-min treatment with OXT (10−6 m) stimulated translocation of NFATC1 to the nucleus with CyclA inhibiting the response (Fig. 6). NFATC1 nuclear translocation was also induced by ionomycin [NFATC1 N:C ratio (fold change): control, 1 ± 0; ionomycin, 1.62 ± 0.19; data not shown].

Fig. 6.

OXT-induced nuclear translocation of endogenous NFATC1. Human myometrial cells were stimulated with OXT (10−6 m) for 30 min or pretreated with CyclA (10−6 m) for 30 min before stimulation with OXT. After stimulation, the reaction was terminated with ice-cold PBS and fixed in 4% PFA. Immunocytochemistry was performed as described in Materials and Methods. Images show myometrial cells stained with NFATC1 primary antibody. The control (Ctrl) group was not stimulated with OXT. Images were acquired and analyzed as described in Materials and Methods. Images shown are representative of 2500–3000 cells. Scale bar, 60 μm. Data shown are NFATC1 N:C fold change over control group, mean ± se, from at least three separate experiments and independent donors. Statistical analysis was by one-way ANOVA with Dunnett's post hoc test compared with the untreated control. *, P < 0.01

To assess endogenous NFAT transcriptional activity, myometrial cells were transduced with a NFAT response element luciferase reporter (NFAT-RE-Luc). Cells were then stimulated with a sustained concentration of OXT for 6 h, which elicited a concentration- and time-dependent increase in NFAT-RE-Luc activity (EC50 = 84.6 ± 0.24 nm) (Fig. 7, A and B). OXT-stimulated (10−7 m) NFAT-RE luciferase activity was abolished after pretreatment with the OXTR antagonist L368899 (10−5 m) and with the calcineurin inhibitor CyclA (10−6 m), [luciferase activity (fold induction): control, 1 ± 0; OXT, 2 ± 0.11; L368889, 0.65 ± 0.15; CyclA, 0.86 ± 0.14; data not shown].

Fig. 7.

OXT induces transcriptional activity of endogenous NFAT. Human myometrial cells were transduced with Ad-NFAT-RE-Luc and Ad-β-Gal before sustained stimulation with OXT (10−6 m) for the indicated time points (A) or in a concentration-dependent manner for 6 h (B). The control (Ctrl) group was not stimulated with OXT. After stimulation, the reaction was terminated with ice-cold PBS before cell lysis. Data shown are luciferase activity in luminescence units, standardized to β-Gal activity, and normalized to control. Mean ± se is displayed for at least three separate experiments and independent donors. Statistical analyses were performed by one-way ANOVA with Dunnett's post hoc test compared with time zero (A) or the untreated control (B). *, P < 0.05; **, P < 0.01; ***, P < 0.001.

The increase in NFAT-RE-Luc activity in response to OXT led us to investigate the effect of OXT on transcriptional activity. For this, we chose genes whose expression was known to be either OXT dependent in the myometrium or NFAT dependent in other smooth muscle cells. The genes selected for analysis were RGS2 (65) and prostaglandin-endoperoxide synthase 2 (PTGS2/COX2) (66) (OXT dependent) and RCAN1 (67), coagulation factor II (thrombin) receptor (F2R/PAR1) (68), and transient receptor potential cation channel, subfamily C, member 1 (TRPC1) (69) (NFAT dependent). COX2 expression has also been shown to be NFAT dependent (67). Gene activation was measured by real-time PCR.

Exposure of myometrial cells to OXT increased mRNA levels of RGS2, RCAN1, and COX2 (Fig. 8A). The response was significant after 4–6 h of stimulation with OXT, with a 2.5- to 4-fold increase in mRNA. Pretreatment with L368899 or CyclA prevented the increase in mRNA (Fig. 8B), indicating that OXT-induced gene expression is mediated through the OXTR and dependent on the calcineurin-NFAT signaling pathway. OXT did not influence expression of PAR1 or TRPC1 (data not shown).

Fig. 8.

OXT induces gene expression via a calcineurin-NFAT-dependent pathway. Human myometrial cells were stimulated with sustained OXT (10−7 m) for the indicated time points (A) or pretreated with L368899 (10−5 m) or CyclA (10−6 m) for 30 min before treatment with sustained OXT (10−7 m) for 4 h (B). The control (Ctrl) group was not stimulated with OXT. After stimulation, the reaction was terminated with ice-cold PBS before RNA extraction and cDNA conversion. Real-time PCR was performed on cDNA to quantify mRNA of RGS2, RCAN1, and COX2. Data shown are relative mRNA expression calculated by the Pfaffl method of analysis, standardized to the housekeeping genes 18S and RNA polymerase II and normalized to time zero (A) or the untreated control (B). Mean ± se is displayed for at least three separate experiments from independent donors. Statistical analyses were performed by one-way ANOVA with Dunnett's post hoc test compared with time zero or untreated control. *, P < 0.05; **, P < 0.01.

The effect of OXT on gene expression requires both transcription and translation

Despite induction of RGS2 gene expression being mediated by the calcineurin-NFAT pathway, it has previously been reported that the RGS2 promoter does not contain the NFAT consensus-binding motif (67), raising the question whether OXT-induced activation of gene expression may involve the synthesis of intermediary proteins. To search for NFAT-binding motif expression in the promoter regions of RGS2, RCAN1, and COX2, we used the Meme Suite program as described in Materials and Methods. The search identified NFAT-binding motifs within the promoter regions of RCAN1 and COX2 but did not identify the sequence within ±2.5 kb of the RGS2 promoter. Therefore, we investigated the role protein neosynthesis may play in OXT-induced gene expression. Human myometrial cells were pretreated with the protein synthesis inhibitor CX before stimulation with OXT. CX alone did not influence RGS2 transcription but had a tendency to increase transcription of RCAN1 and COX2. Joint stimulation with OXT and CX inhibited OXT-induced RGS2 transcription but amplified the OXT-induced transcriptional readouts for RCAN1 and COX2 (Fig. 9). These data indicate initiation of RGS2 transcription is dependent on protein neosynthesis, whereas induction of RCAN1 and COX2 transcription is protein neosynthesis independent.

Fig. 9.

The effect of inhibiting protein neosynthesis on OXT-induced gene expression. Human myometrial cells were stimulated with sustained OXT (10−7 m) for 4 h, with or without pretreatment with CX (50 μm) for 30 min. After stimulation, the reaction was terminated with ice-cold PBS before RNA extraction and cDNA conversion. Real-time PCR was performed on cDNA to quantify mRNA of RGS2, RCAN1, and COX2. Data shown are relative mRNA expression calculated by the Pfaffl method of analysis, standardized to the housekeeping genes 18S and RNA polymerase II and normalized to the untreated control. Mean ± se is displayed for at least three separate experiments from independent donors. Statistical analyses by two-way ANOVA revealed OXT and CX to be significant sources of variation: RGS2, P < 0.05 (OXT and CX); RCAN1, P < 0.001 (OXT) and P < 0.01 (CX); COX2, P < 0.001 (OXT and CX). The OXT-CX interaction was also found to be significant: RGS2, P < 0.05; RCAN1, P < 0.05; COX2, P < 0.01. Post hoc Bonferroni tests revealed statistical differences between treatment and control groups. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Discussion

OXT is a neuropeptide hormone that stimulates contraction of the pregnant uterus; it is involved in the augmentation and possibly the initiation of parturition and uterotonic activity after delivery. OXT acts via the OXTR, a phospholipase C-coupled GPCR, to increase [Ca2+]i and activate Ca2+ effector proteins, such as myosin light chain kinase, to initiate myometrial contractions. In this study, we are interested in the potential for other Ca2+ effector proteins (i.e. transcription factors such as NFAT) to mediate longer-time effects of OXT on gene expression. We focused on NFAT for two reasons: first, NFAT is a Ca2+-dependent transcription factor that regulates gene expression in other smooth muscle cells (67–70); second, NFAT activation is known to be dependent on stimulus pulse frequency in other systems (51, 55, 56), and therefore, NFAT may decode OXT pulse frequency. In this report, we have defined NFAT isoform expression in myometrial cells by qPCR, monitored NFAT activation by the OXTR with a NFATC1-EFP reporter, assessed endogenous NFAT activity with an NFATC1 primary antibody and NFAT-RE-Luc construct, and finally, used qPCR to investigate putative NFAT and OXT target genes within the myometrium.

Myometrial cells in our cell culture system were characterized using primary antibodies raised toward the smooth muscle cell markers desmin, actin, and calponin. Over 90% of all myometrial cells were positive for desmin, actin, and calponin, indicating that cell dispersion from myometrial biopsies yielded a 90% pure myocyte population. In addition, myometrial cells in our cell culture system have previously demonstrated robust phospholipase C and Ca2+ signals in response to OXT (19, 57), thus indicating they retain their functional responsiveness in culture and are a good model for investigating OXT-stimulated NFAT activation.

This is the first description of NFAT isoforms in the myometrium, defined by qPCR. mRNA for all known NFAT isoforms (NFATC1–C4 and -5) were expressed, in contrast with arterial and ileal smooth muscle, which express only NFATC1, -C3, and -C4 (70) and NFATC3 and -C4 (64), respectively, indicating NFAT isoform expression is tissue specific.

Expression of NFAT isoforms in the myometrium led us to study the effect of Ca2+ on NFAT activation by using NFATC1-EFP, because translocation of this reporter from the cytoplasm to the nucleus can provide a live cell readout for NFAT activation (51), as opposed to immunocytochemistry for endogenous NFAT, which requires cell fixation. First, we demonstrated that OXT stimulates accumulation of [Ca2+]i using a fluo4 Ca2+ indicator with an EC50 in the low nanomolar range as previously described (19). The fluo4 response peaked faster with high concentrations of OXT than at lower concentrations; this delay could simply reflect the fact that augmenting the concentration of OXT increases the rate at which OXT-sensitive intracellular stores are depleted and Ca2+ mobilized.

To assess whether mobilization of Ca2+ by activation of the OXTR induces an NFAT response myometrial cells were transduced with a NFATC1-EFP and stimulated with OXT. We found that OXT stimulated NFATC1-EFP translocation to the nucleus with an EC50, similar to OXT-induced accumulation of [Ca2+]i. NFATC1-EFP translocation is far slower in onset than Ca2+ mobilization (within minutes rather than seconds), suggesting there is a rate-limiting step between Ca2+ release and nuclear translocation of NFATC1-EFP. Ionomycin also stimulated NFATC1-EFP nuclear localization with similar response kinetics (data not shown), demonstrating that translocation of NFATC1-EFP is dependent on Ca2+ entry as opposed to other aspects of OXTR signaling. GPCR activation and ionomycin have previously been shown to stimulate nuclear translocation of NFATC1-EFP in gonadotrophs and HeLa cells with similar kinetics (51). However, in arteries stimulated with endothelin NFATC3, nuclear translocation is slower in onset than in the myometrium, and nuclear localization is not sustained (70), whereas in the ileum, ionomycin failed to induce NFATC3 nuclear translocation (64), indicating that NFAT activation is likely to be isoform, tissue, and stimulus dependent.

OXT-induced nuclear translocation of NFATC1-EFP is mediated through the OXTR and calcineurin, as demonstrated by inhibition of NFAT activity by the OXT antagonists atosiban and L368899 and the calcineurin inhibitors CyclA and FK506. Furthermore, removal of extracellular Ca2+ and the use of the [Ca2+]i chelator BAPTA-AM both abolished the NFATC1-EFP response, indicating that Ca2+ influx into the cell and elevation of [Ca2+]i is necessary for the initiation of NFATC1-EFP nuclear translocation.

Because prolonged exposure of OXT to myometrial cells induced sustained nuclear localization of NFATC1-EFP, we investigated whether NFATC1-EFP nuclear translocation is dependent on constant receptor occupancy or whether the initial binding of OXT to its receptor is sufficient to generate a sustained NFATC1-EFP response. We found that washing off OXT after stimulation resulted in nuclear export of NFATC1-EFP, demonstrating that nuclear localization of NFATC1-EFP is reversible. This is in concordance with previous work in other systems that found sustained GPCR occupancy, and [Ca2+]i elevation is required to maintain NFATC1-EFP localization in the nucleus (51, 55, 56).

To assess the physiological importance of stimulus pulse frequency on NFATC1-EFP nuclear translocation, we studied the effect of pulses of OXT on the NFATC1-EFP N:C. We found that NFATC1-EFP N:C mirrored OXT pulse frequency in a concentration-dependent manner, demonstrating a cumulative, sawtooth effect. This is in concordance with previous work that shows NFATC1-EFP nuclear translocation in response to pulsatile GnRH (51). OXT is released in pulses in vivo; therefore, it is important that NFATC1-EFP is responsive to a pulsatile OXT system, which highlights the physiological relevance of NFATC1-EFP N:C being dependent on OXT pulse frequency. Pulsatile systems are important for preventing desensitization, signaling efficiency, and encoding information. Within our system, there was no reduction in the amplitude of the NFATC1-EFP N:C ratio with each pulse, indicating that NFATC1-EFP nuclear translocation is not desensitized within the conditions studied. Furthermore, the amplitude of the NFATC1-EFP N:C ratio with pulsatile OXT stimulation over 4 h was similar to that with sustained treatment, suggesting a pulsatile system is a more efficient mechanism of OXT-stimulated NFATC1-EFP nuclear translocation than a sustained stimulus (although this observation cannot be extended and has not yet been tested in regard to OXT-induced gene expression). Moreover it costs the cell less energy to maintain high [Ca2+]i levels for repeated short intervals than for a constant time period. Interestingly, pulsatile administration of OXT is more efficient than constant infusion for the induction/augmentation of labor in women (8, 71).

OXT-stimulated nuclear translocation of the NFATC1-EFP reporter construct led us to investigate the role of OXT in inducing endogenous NFAT nuclear translocation and transcriptional activity using a NFATC1 primary antibody and NFAT-RE-Luc reporter, respectively. OXT stimulated endogenous NFATC1 nuclear translocation with the effect being mediated by calcineurin, indicating NFATC1-EFP nuclear translocation can be extended to an endogenous mechanism. OXT also induced a time- and concentration-dependent increase in NFAT-RE-Luc activity, demonstrating that activation of NFAT by OXT in myometrial cells results in a transcriptional output.

We have also investigated whether OXT induces gene expression through the calcineurin-NFAT pathway. We have shown for the first time that OXT stimulates the expression of the calcineurin inhibitory factor RCAN1 in the myometrium. Moreover, we have confirmed that OXT induces expression of RGS2 (a regulator of G protein signaling) and COX2 (an enzyme involved in prostaglandin synthesis) as previously shown in myometrial cells (65, 66). The calcineurin inhibitor CyclA attenuated OXT-induced expression of all three genes, demonstrating the importance of the calcineurin-NFAT signaling pathway in mediating OXT-induced gene expression in the myometrium. The physiological significance of OXT-induced up-regulation of RGS2 and RCAN1 in myometrial cells is not established, but it's likely to be important for the negative regulation of OXT signaling; RGS2 uncouples G protein-coupled receptor subunits (72), therefore controlling OXT signaling at the level of the OXTR. RCAN1 inhibits calcineurin activity (73), thus inhibiting NFAT-induced gene transcription in response to OXT. On the other hand, the induction of COX2 by OXT would result in production of prostaglandins and amplification of the parturition cascade (66). Prostaglandins are uterotonic agonists that could work in synergy with OXT to induce myometrial contractions at term; moreover, they participate in cervical ripening, which would facilitate a successful vaginal delivery (74).

Although the increase in RGS2 was found to be mediated by calcineurin-NFAT, the absence of the NFAT consensus-binding motif in the RGS2 promoter suggests that NFAT itself is not directly responsible for enhancing RGS2 gene expression, and there must be another intermediary factor. To test this, we pretreated cells with CX (an inhibitor of protein synthesis) before stimulation with OXT. CX was found to completely inhibit OXT-induced RGS2 gene transcription, suggesting that protein neosynthesis is necessary for this mechanism of OXT action. Because the calcineurin inhibitor CyclA inhibited Oxt induced RGS2 gene expression independently, it is possible that in OXT-stimulated cells NFAT participates in the transcription of another transcription factor that, when translated into active protein, in turn induces RGS2 expression.

In contrast, we have shown that CX does not inhibit OXT-induced transcription of RCAN1 and COX2, indicating this mechanism of OXT action is protein synthesis independent. Instead, CX amplifies the effect of OXT stimulation. A similar increase in transcription has also been reported in fibroblasts pretreated with CX before stimulation with TNFα (75). Protein neosynthesis-dependent negative feedback loops are often found in signaling cascades; for example, ERK-mediated expression of dual-specificity phosphatases (52, 76) and NFκB-mediated expression of nuclear factor of κ light polypeptide gene enhancer in B-cells inhibitor α (75), where the phosphatases and IκBα negatively regulate ERK and NFκB activity, respectively. The same could be true of NFAT-mediated expression of RCAN1 and COX2, where CX blocks protein synthesis, removing putative negative regulators of NFAT signaling, thus increasing the transcriptional response of NFAT target genes. These hypotheses need to be tested in additional experiments.

We have identified a novel signaling pathway within the myometrium, whereby OXT initiates NFAT signaling, inducing transcriptional activity. These findings are likely to be of physiological significance in the context of parturition. Pregnant mice administered with the calcineurin inhibitor FK506 experienced delayed labor (33); thus, it is possible that OXT, through induction of the calcineurin-NFAT pathway, is responsible for the regulation of the initiation of labor. The transcriptional activity stimulated by OXT in the myometrium could produce factors involved in the initiation and maintenance of uterine contraction, i.e. COX2 through the synthesis of prostaglandins, or regulatory factors such as RCAN1 and RGS2. Identification of a wider array of OXT- and calcineurin-NFAT-responsive genes in the myometrium will provide insight into how parturition is initiated, enabling development of therapies to prevent spontaneous premature labor.

In summary, we have shown that human myometrial cells express the four Ca2+/calcineurin-sensitive NFAT isoforms (NFATC1–C4) as well as the Ca2+/calcineurin-insensitive NFAT5. We demonstrate for the first time that the OXTR mediates translocation of a NFATC1-EFP reporter from the cytoplasm to the nucleus and that this mechanism is mediated by calcineurin. We also show how NFATC1-EFP translocation kinetics can make signaling more efficient with a pulsatile OXT signal in contrast to sustained stimulation. Furthermore, we show that OXT induces nuclear translocation of endogenous NFATC1 and stimulates endogenous NFAT-mediated activation of a NFAT-Luc reporter. Moreover, we identify OXT transcriptional induction of three calcineurin-NFAT-sensitive genes (RGS2, RCAN1, and COX2) as well as providing preliminary evidence for translation-dependent negative feedback regulation of OXTR-mediated NFAT signaling. The data provide novel insights into the role of OXT in uterine function.