Serotonin (5-hydroxytryptamine) mediates a wide range of physiological effects by activating a large family of receptors that are distributed throughout the body (Hoyer et al., 2002). These cell surface receptors are classified into seven distinct subtypes (5-HT₁ to 5-HT₇), and, with the exception of the ligand-gated ion channel 5-HT₃, they are all members of the G-protein-coupled receptor superfamily (Alexander et al., 2008). In the central nervous system, these receptors are implicated in numerous conditions, including: depression, anxiety, schizophrenia, and migraine (Barnes and Sharp, 1999). Serotonin also plays a role in platelet activation (Kaumann and Levy, 2006) and regulating enteric and vascular smooth muscle contraction (De Ponti, 2004; Kaumann and Levy, 2006).

Serotonin is also known to modulate contractions of the uterus, and the release of serotonin, after the activation of mast cells or platelets, may be an important source of the ligand in this tissue (Rudolph et al., 1993). The exact effect of serotonin, in terms of inhibiting or stimulating contractions, appears to be species specific. Serotonin has been shown to increase contractions in rabbit, rat, guinea pig, and human myometrium (Freyburger et al., 1952; Woolley, 1958; Contractor et al., 1968) but to inhibit contractions of porcine myometrium (Kitazawa et al., 1998). The inhibitory effect of serotonin in porcine myometrium is mediated by the G_s-coupled 5-HT₇ receptor (Kitazawa et al., 1998). The stimulatory effect of serotonin in rat myometrium is mediated by G_q-coupled 5-HT₂ receptors (Cohen et al., 1985), and the increased response to serotonin in late gestation in this species is associated with an increase in the density of 5-HT_2A receptors (Minosyan et al., 2007). A study in a human myometrial cell line demonstrated expression of the 5-HT_2B receptor (Kelly and Sharif, 2006). However, the receptors that mediate the actions of serotonin in intact human myometrium have not been systematically characterized. In the present study, we used selective serotonin receptor agonists and antagonists to characterize the receptor subtypes mediating the effects of serotonin on isolated strips of myometrium obtained at the time of planned caesarean section at term. Candidate receptors were then further studied using RT-PCR, radioligand binding, Western blot, and immunohistochemistry.

Materials and Methods

Tissue Preparation. Human myometrial samples were obtained from nonlaboring patients, undergoing routine elective caesarean section, at 38 to 40 weeks of pregnancy. The study was approved by the Cambridgeshire 2 Research Ethics Committee, and all patients gave their informed, written consent to participate. Indications for caesarean section included breech presentation, twin pregnancy, and prior caesarean section. Specimens were taken from the upper edge of the lower segment uterine incision, after delivery of the baby and placenta, and placed into ice-cold Krebs' solution (119 mM NaCl, 4.7 mM KCl, 1.2 mM MgSO₄, 26 mM NaHCO₃, 1.2 mM KH₂PO₄, 2.5 mM CaCl₂, and 11.1 mM d-glucose). Samples were stored in ice-cold Krebs' solution (with EDTA-free protease inhibitors) at 4°C for up to 24 h before use in isometric tension experiments.

Isometric Tension Measurements. Studies were carried out using an eight-chamber Radnoti tissue organ bath system (ADIn-struments Ltd., Chalgrove, Oxfordshire, UK), and data were recorded using the Powerlab software Chart (version 5.2.2). For each experiment, transducers were calibrated (between 0 and 10 g) before mounting tissue. Uterine biopsies were cleared of serosa (thin fibrous membrane), fibrous or damaged tissue, and any visible blood vessels. The muscle tissue was dissected further, in ice-cold Krebs' solution, into longitudinal strips (following the plane of the muscle fibers) of approximately 1 to 3 × 8 to 12 mm. Tissue strips were mounted in the organ bath, secured with cotton thread and placed under 2-g tension (Bardou et al., 1999), and allowed to equilibrate for a period of 90 min. Organ bath chambers were filled with Krebs' solution, which was warmed to 37°C and gassed with 95% O₂/5% CO₂. The tension was reset at 30 min, and the strips washed with fresh Krebs' solution at 15-min intervals. After 90 min, the tissue was treated with KCl (50 mM, 7 min), then washed with fresh Krebs' solution and left for a further 45 min to allow spontaneous contractions to develop. Stimulatory or inhibitory agonists were then added in a cumulative manner, with subsequent doses added every 20 min. In addition, all potential stimulatory agonists were also studied in the presence of 1 μM forskolin because we have shown previously that this increases the magnitude of responses to serotonin (Cordeaux et al., 2008). Where antagonists were employed, strips were equilibrated with the antagonist for 1 h before agonist addition, with or without forskolin, as appropriate.

Data Analysis. Isometric tension was recorded and analyzed using the Powerlab software Chart (version 5.2.2). Contraction was quantified as the “area under the curve” (AUC) calculated from the “integral over the minimum” using the datapad function. Basal activity was calculated from the AUC in the 20 min before agonist addition, and stimulated activity was the AUC in the 20 min in the presence of the given concentration of agonist. All agonist responses were compared with concurrent time and vehicle controls. The cumulative concentration responses were analyzed by nonlinear regression analysis using GraphPad Prism (version 4.03; GraphPad Software Inc., San Diego, CA), and cumulative EC₅₀ values were obtained. EC₅₀ is the molar concentration of the agonist that produced 50% of the maximal response achieved. pEC₅₀ is the -log₁₀(EC₅₀) value. Antagonist pK_B values were derived from the EC₅₀ values of an agonist in the absence or presence of a given concentration of antagonist, using the Gaddum equation: pK_B = log (concentration ratio - 1) - log (antagonist concentration). All data are expressed as means ± S.E. The n in the text refers to the number of independent experiments performed, using tissue from separate donors. Statistical significance of drug effects, relative to basal (pre-treatment) or time control (untreated strips), were determined using paired Student's t test. Means were compared using Student's unpaired t test or using ANOVA, the latter employed where three or more groups were compared. Multiple comparisons after ANOVA were conducted using the Dunnett's test. Quantitative RT-PCR was analyzed using the appropriate nonparametric methods (Mann-Whitney U test).

RT-PCR. Myometrial tissue samples for RT-PCR were rapidly dissected (under ice-cold Krebs' solution) and snap-frozen in liquid nitrogen. Total RNA was isolated using TRIzol (Invitrogen, Paisley, UK) and treated with DNase (RQ1; Promega, Madison, WI) according to the manufacturer's instructions. cDNA was synthesized using SuperScript II reverse transcriptase (Invitrogen) and random hexamer primers. Quantitative RT-PCR was performed using the Applied Biosystems (Foster City, CA) 7900HT Real-Time PCR System. Dilution series using each of the specific primer/probe sets for each of the RNAs of interest (5-HT_1A, 5-HT_2A, and 18S rRNA) indicated that all amplification rates were similar. Because of the low signal observed when amplifying the 5-HT_1A mRNA, 10 times more cDNA was used for quantitation. Experiments were performed in triplicate, using cDNA obtained from five separate donors.

Preparation of Washed Tissue Homogenates. For Western blotting, fresh biopsies were rapidly dissected (under ice-cold Krebs' solution) into pieces approximately 7 mm³, placed in a bijoux vial, and snap-frozen in liquid nitrogen. Frozen specimens were wrapped in foil and stored at -80°C until use. Samples acquired throughout the study were assigned an arbitrary reference number to ensure patient anonymity. Frozen biopsy tissue was powdered using a metal pestle and mortar, in which the tissue was cooled at all times under liquid nitrogen. Frozen, powdered tissue was transferred into ice-cold RIPA lysis buffer (supplemented with protease inhibitors) and homogenized using a handheld homogenizer. Tissue homogenates were centrifuged at 13,000g for 10 min, and the supernatant was collected and stored in aliquots at -80°C. Protein concentrations were determined using a BCA Protein Assay kit (Pierce Chemical, Rockford, IL), using bovine serum albumin as the standard.

Western Blotting. Protein samples were denatured by heating in Tris-glycine SDS sample buffer supplemented with dithiothreitol (NuPAGE reducing agent; Invitrogen) at 85°C for 2 min. Proteins were separated by SDS-polyacrylamide gel electrophoresis (125 V for 2 h) on 10 or 12% Tris-glycine gels (Invitrogen). Proteins were then transferred to polyvinylidene difluoride membranes (Invitrogen) using an X-Cell II Blot Module (Invitrogen), according to the manufacturer's instructions (25 V, 2 h). After transfer, membranes were incubated in 5% milk/Tris-buffered saline/Tween 20 solution for 2 h at room temperature to reduce nonspecific binding. Membranes were incubated with primary antibodies in 5% milk/TBST solution at 4°C overnight. Membranes were washed in TBST solution (12 × 5-min washes), before incubation with the respective secondary antibody immunoglobulin horseradish peroxidase conjugate (5% milk/TBST solution) for 1 h at room temperature. After a further series of washes with TBST solution (as above), membranes were incubated with Enhanced Chemiluminescence (ECL Plus) detection reagents (5 min, room temperature). Membranes were then wrapped in cling film and exposed to Hyperfilm-ECL X-ray film for up to 5 min.

Immunohistochemistry. Tissue samples for immunohistochemistry were immediately dissected (under ice-cold Krebs' solution) into pieces approximately 4 mm³ and fixed in formalin (10% solution) at room temperature for 2 to 4 h. Samples were embedded in paraffin wax, and 5-μm sections were cut. After deparaffination and rehydration, background peroxidase activity was quenched with 10% hydrogen peroxide solution, and nonspecific binding was blocked with 20% rabbit serum (in 1% bovine serum albumin, 30 min at room temperature). Slides were then incubated with primary antibody (goat anti-human 5-HT_2A; Santa Cruz Biotechnology, Inc., Santa Cruz; CA, 0.2 mg/ml; 4°C, overnight) or negative control (goat IgG). After a series of washes (phosphate-buffered saline), slides were incubated with secondary antibody (rabbit anti-goat IgG biotin; diluted 1:600) for 1 h at room temperature. Immunostaining was visualized by incubating slides with streptavidin HRP (30 min, room temperature) and diaminobenzidine reagent (45 s) before counterstaining with hematoxylin. Slides were dehydrated with ethanol solutions and xylene before mounting with Depex.

Radioligand Binding. Tissue samples for radioligand binding were immediately dissected (under ice-cold Krebs' solution), placed in a bijoux vial, and snap-frozen in liquid nitrogen. Ten-micron sections were cut on a cryostat, thaw-mounted onto gelatin-coated slides, and stored at -80°C until use. Adjacent sections were cut and solubilized in RIPA lysis buffer. Protein concentrations were determined using a BCA Protein Assay kit (Pierce Chemical), using bovine serum albumin as the standard. Slide-mounted sections of tissue from three separate donors were thawed and preincubated at room temperature in 50 mM Tris-HCl, pH 7.4, for 15 min. For 5-HT_1A and 5-HT_2A receptor binding, sections were incubated with a single dose of either [³H]WAY100635 (3 nM) or [³H]ketanserin (10 nM), respectively, at room temperature for 2 h. Nonspecific binding was defined in the presence of 10 μM WAY100635 or 10 μM ketanserin, respectively. Binding assays were terminated by washing sections in ice-cold Tris-HCl buffer (3 × 5 min). Sections were wiped from each slide (using a GF/C glass filter article) into a scintillation vial, and scintillation fluid was added and counted on a beta counter.

Materials. DAB tablets, formalin solution (10%), forskolin, 5-hydroxytryptamine hydrochloride, 8-hydroxy-2-dipropylaminotetralin [(R)-(+)-8-OH-DPAT], rabbit serum, Tris-buffered saline with Tween 20, pH 8.0, (S)-WAY100635 hydrochloride, and Whatman GF/C filter papers were purchased from Sigma Chemical (Poole, Dorset, UK). AS19, BW723C86 hydrochloride, cisapride, ketanserin tartrate, α-methyl-5-hydroxytryptamine maleate, Ro 60-0175 fumarate, RS102221 hydrochloride, SB269970 hydrochloride, and SB204741 were purchased from Tocris Bioscience (Bristol, UK). Imigran (sumatriptan succinate) was purchased from GlaxoSmithKline (Uxbridge, Middlesex, UK). Syntocinon (oxytocin) was purchased from Alliance Pharma plc (Chippenham, Wiltshire, UK). Ultima Gold scintillation fluid was purchased from PerkinElmer Life and Analytical Sciences (Waltham, MA).

WAY100635 [methoxy-³H] (specific activity, 73 Ci/mmol) and ketanserin hydrochloride [ethylene-³H] (specific activity, 67 Ci/mmol) were purchased from American Radiolabeled Chemicals (St. Louis, MO). Benchmark prestained protein ladder, Invitrolon polyvinylidene difluoride filter article, Novex Tris-Glycine gel, Novex Tris-Glycine Transfer buffer, Novex Tris-Glycine SDS running buffer, Novex Tris-Glycine SDS sample buffer, NuPAGE reducing agent, SuperScript II reverse transcriptase, and TRIzol reagent were purchased from Invitrogen. RQ1 RNase-free DNase was purchased from Promega. ECL Plus Western blotting detection agent and Hyperfilm ECL X-ray film were purchased from GE Healthcare (Chalfont St. Giles, UK). The BCA Protein Assay kit was purchased from Perbio Science UK Ltd. (Chester, Cheshire, UK), protease inhibitor cocktail tablets were purchased from Roche Diagnostics (Basel, Switzerland), and RIPA lysis buffer was purchased from Millipore Corporation (Billerica, MA). Goat IgG, anti-goat IgG Biotin, anti-goat IgG HRP, anti-mouse IgG HRP, and anti-rabbit IgG HRP were purchased from Dako UK Ltd. (Ely, Cambridgeshire, UK). 5-HT_1A antibodies were purchased from Aviva Systems Biology, LLC (San Diego, CA), Santa Cruz Biotechnology, Inc., and Everest Biotech Ltd (Upper Heyford, Oxfordshire, UK). 5-HT_2A antibody was purchased from Santa Cruz Biotechnology, Inc. Streptavidin HRP was purchased from Vector Laboratories (Peterborough, UK).

5-HT_1A, 5-HT_2A, and 18S TaqMan Gene Expression Assays were purchased from Applied Biosystems. All other reagents were analytical grade.

Results

Isometric Tension Measurements. Development of spontaneous contractile activity before exposure to potassium was highly variable. However, incubation in 50 mM potassium resulted in an immediate contractile response in all strips, and, after washout of potassium, virtually all strips (>99%) developed regular spontaneous contractions. Forskolin (1 μM) reduced the resultant spontaneous activity (data not shown), and all potentially stimulatory agonists were also studied in the presence of forskolin because we have shown previously that it increases the magnitude of response to serotonin (Cordeaux et al., 2008).

5-HT₁ Receptors. In the absence of forskolin, the maximal response to the 5-HT_1A-selective agonist 8-OH-DPAT (10^-5 M) was 123 ± 12% basal activity, which was slightly greater than that of time-control strips (79 ± 14% basal activity) (p = 0.047, n = 9, Student's t test), although in a single experiment, 8-OH-DPAT seemed to produce a robust response (maximum 5223% basal activity, with a pEC₅₀ value of 6.25). In the presence of forskolin, the maximal response to 8-OH-DPAT (10^-5 M) was 280 ± 77% basal activity, which was significantly greater than that of time-control strips (85 ± 16% basal activity) (p = 0.03, n = 13, Student's t test). However, the responses were highly variable between samples from different women. In three of the 13 samples, there seemed to be a concentration-dependent increase in contractions (pEC₅₀ = 5.75 ± 0.15, n = 3).

The effect of antagonists was studied on strips that exhibited a clear response to 8-OH-DPAT. All these experiments were conducted in the presence of 1 μM forskolin. The pEC₅₀ to 8-OH-DPAT was similar in vehicle (5.80 ± 0.33, n = 3) (Figs. 1A and 2A) and in the presence of the 5-HT_1A-selective antagonist WAY100635, both at 10^-7 M (5.43 ± 0.22, n = 3) (Figs. 1B and 2A) and 10^-6 M (5.06 ± 0.28, n = 3) (p = 0.25, one-way ANOVA) (Fig. 2A). However, the maximal response to 8-OH-DPAT in vehicle (700 ± 63% basal activity) was greater than in the presence of WAY100635 at 10^-7 M (434 ± 64%) or 10^-6 M (456 ± 43%) (p = 0.03, one-way ANOVA), and the difference was statistically significant at both concentrations (p < 0.05 for each treatment, n = 3, Dunnett's multiple comparison test).

Because the pEC₅₀ of 8-OH-DPAT was not significantly affected in the presence of the 5-HT_1A-selective antagonist and because this ligand has been shown previously to activate α adrenoceptors (Castillo et al., 1993), we determined the effect of the selective α adrenoceptor antagonists, prazosin (α₁, 100 nM) and yohimbine (α₂, 100 nM), on the response to 8-OH-DPAT. The pEC₅₀ of 8-OH-DPAT was similar in the presence of vehicle (0.001% DMSO, final concentration) (5.15 ± 0.33), prazosin (5.01 ± 0.46), and yohimbine (5.01 ± 0.37) (p = 0.96, one-way ANOVA, n = 3) (Fig. 2B), as was the maximal response (1209 ± 471, 1229 ± 280, and 870 ± 320% basal activity, respectively) (p = 0.75, one-way ANOVA, n = 3). The possibility that 8-OH-DPAT was acting at either 5-HT₇ or 5-HT_2A receptors was addressed by determining the effects of the selective antagonists SB269970 and ketanserin, respectively. The pEC₅₀ of 8-OH-DPAT, in the presence of SB269970 (100 nM) (5.26 ± 0.07) or ketanserin (100 nM) (4.91 ± 0.09), was similar to that obtained in control conditions (5.17 ± 0.25) (p = 0.31, one-way ANOVA, n = 5), as was the maximal response (813 ± 256, 743 ± 149, and 636 ± 62% basal activity, respectively (p = 0.78, one-way ANOVA, n = 5).

The 5-HT_1B/1D-selective agonist, sumatriptan, did not stimulate contractions in either the presence or absence of forskolin. In control conditions, the AUC for contractions in the presence of sumatriptan (10 μM) was 82 ± 11% that of basal, which was not significantly different from the vehicle time control (69 ± 9% basal, p = 0.12, n = 5). In the presence of forskolin (1 μM), the AUC for contractions in the presence of sumatriptan (10 μM) was 49.5 ± 10% basal and in vehicle was 51 ± 12% basal (p = 0.79, Student's t test, n = 6).

5-HT₂ Receptors. The nonselective 5-HT₂ receptor agonist, α-Me-5-HT (10^-10 to 10^-5 M), produced an increase in contractions in the absence of forskolin. The maximal response (with 100 nM to 1 μM agonist) was 166 ± 24% basal activity, which was significantly greater than vehicle time control (65 ± 11% basal activity, p = 0.01, n = 5). In four of five experiments, the response was concentration-dependent, and the pEC₅₀ value for α-Me-5-HT was 7.35 ± 0.17 (n = 4). In the presence of forskolin, α-Me-5-HT (300 nM to 3 μM) stimulated contractions to 730 ± 223% basal activity, which was significantly greater than vehicle time control (39 ± 10% basal activity, p = 0.04, n = 5). The pEC₅₀ value for this response was 7.21 ± 0.20 (n = 5).

The 5-HT_2B-selective agonist, BW723C86, failed to produce a significant increase in contractions in the absence of forskolin. The AUC value, in the presence of the maximally effective dose of BW723C86, was 117 ± 13% basal activity and was 70 ± 10% basal activity in the vehicle time control strips (p = 0.12, n = 3, Student's t test). In the presence of forskolin, this ligand-stimulated contractions in one experiment (439% basal, pEC₅₀ = 6.10). In five other tissue samples, BW723C86 failed to produce a significant increase in contractions, compared with time control (115 ± 25% basal and 74 ± 24% basal activity, respectively) (p = 0.15, n = 5, Student's t test).

In control conditions, the 5-HT_2C-selective agonist Ro-60-01-75 (1-10 μM) stimulated contractions to 131 ± 17% basal, which was significantly greater than in time control strips (68 ± 9% basal activity, p = 0.01, n = 6, Student's t test), although the pEC₅₀ (>micromolar) could not be defined. In the presence of forskolin, Ro-60-01-75 failed to stimulate contractions in two experiments. In five other experiments, the ligand produced a significant increase in contractions, compared with time control (673 ± 201% basal activity and 87 ± 23% basal, respectively, p = 0.03, n = 5, Student's t test). In four of these experiments, a concentration-response curve could be defined (pEC₅₀ = 6.45 ± 0.15, n = 4).

The 5-HT_2A-selective antagonist ketanserin (100 nM) produced a rightward shift in the concentration-response curve to α-Me-5-HT (Figs. 3B and 4A). In the absence and presence of antagonist, the pEC₅₀ values for α-Me-5-HT were 7.60 ± 0.10 (n = 5) and 6.15 ± 0.13 (n = 5), respectively (p = 0.001, Student's t test, n = 5) (Fig. 4A). The corresponding pK_B value of ketanserin (as determined from Gaddum equation) was 8.47 ± 0.16 (n = 5). Ketanserin did not inhibit oxytocinmediated contractions; the pEC₅₀ values for oxytocin in the absence and presence of ketanserin (1 μm) were 7.54 ± 0.03 and 7.45 ± 0.03, respectively (p = 0.22, n = 3, Student's t test) (Fig. 4B).

The 5-HT_2B-selective antagonist SB204741 (100 nM) did not inhibit the response to α-Me-5-HT. In the presence of antagonist, the pEC₅₀ value of α-Me-5-HT was 7.26 ± 0.13 (n = 4), which was not significantly different from the vehicle control (7.36 ± 0.15, p = 0.54, n = 4, Student's t test) (Fig. 4C). The 5-HT_2C-selective antagonist, RS102221 (100 nM), was also ineffective in shifting the α-Me-5-HT response. In the presence of the antagonist, the pEC₅₀ value of α-Me-5-HT was 7.41 ± 0.15 (n = 4), which was not significantly different from the vehicle control (7.23 ± 0.10, p = 0.31, n = 4) (Fig. 4D).

5-HT₄ and 5-HT₇ Receptors. The 5-HT₄ receptor agonist cisapride (10^-5 M) did not inhibit spontaneous activity, compared with the vehicle time control (0.5% DMSO) (49 ± 9 and 57 ± 8% basal, respectively, p = 0.41, n = 7, Student's t test). The 5-HT₇ receptor agonist AS19 (10^-5 M) also failed to reduce spontaneous contractions, compared with vehicle time control (0.5% DMSO) (50 ± 9 and 53 ± 8% basal, respectively, p = 0.83, n = 7 Student's t test).

RT-PCR. RT-PCR analysis indicated that the mRNAs encoding both 5-HT_1A and 5-HT_2A were present in myometrial total RNA. After correcting for possible differences in RNA extraction and cDNA synthesis efficiencies using the 18S rRNA data, it was noticed that there was considerable patient-patient variability in the levels of mRNA detected (5-HT_1A: median in arbitrary units, 0.046; range, 0.013-0.656; 5-HT_2A: 11.6, 3.96-23.53). Nonetheless, there was a significant difference between the two groups (p < 0.008, Mann-Whitney U test, n = 5). The ratio of 5-HT_2A/5-HT_1A mRNA within individual patient samples showed considerable variation (median, 139.5; range, 17.6-514.8).

Western Blotting. Three antibodies raised against the human 5-HT_1A receptor were used to detect the protein in human myometrium using Western blotting. A major band corresponding to a molecular mass of 35 kDa was observed with all three antibodies (Santa Cruz Biotechnology, Inc., Aviva Systems Biology, and Everest Biotech) in six samples of myometrium, from separate donors. A fainter band, 48 kDa, was observed with two of the antibodies (Everest Biotech and Aviva), which was close to the predicted molecular mass of 46 kDa, as determined from the amino acid sequence. In addition, lower bands were seen at 14 to 18 kDa with Everest Biotech and Aviva antibodies, and a higher band of 80 kDa was observed to varying extents between donors with Everest and Santa Cruz Biotechnology, Inc. antibodies (Fig. 5A). A single antibody (Santa Cruz Biotechnology, Inc.), raised against the human 5-HT_2A receptor, was used to detect the protein in the human myometrium using Western blotting. A major band consistent with the predicted molecular mass of 55 kDa was observed in tissue obtained from six separate donors. A slightly lower band of 50 kDa was also seen, to varying extents, between donors. In addition to this protein doublet, a higher band corresponding to a molecular mass of approximately 100 kDa was also seen in three of the six samples (Fig. 5B).

Immunohistochemistry. Formalin-fixed sections, treated with anti-human 5-HT_2A antibody, showed positive staining in a fraction of the smooth muscle cells (Fig. 6B). No staining was observed in negative controls, where the primary antibody was replaced with the same concentration of nonimmune goat IgG (n = 3, Fig. 6A).

Radioligand Binding. The radiolabeled 5-HT_2A receptor antagonist, [³H]ketanserin, demonstrated specific binding to sections of myometrium obtained from three separate donors (Fig. 7) (p = 0.002, paired Student's t test, n = 3). Specific binding in the presence of 10 nM radioligand represented 73 ± 2% (mean ± S.E.M.) of total binding (n = 3) and corresponded to an expression level of 34 ± 4 fmol/mg protein (mean ± S.E.M., n = 3). The radiolabeled 5-HT_1A receptor antagonist, [³H]WAY100635, demonstrated a low but significant level of specific binding to sections of myometrium (Fig. 7) (p = 0.01, paired Student's t test, n = 3). However, specific binding in the presence of 3 nM radioligand represented 25 ± 2% (mean ± S.E.M.) of total binding (n = 3) and corresponded to an expression level of 1.5 ± 0.1 fmol/mg protein (mean ± S.E.M., n = 3).

Discussion

In the present study, we sought to characterize the receptor subtypes by which serotonin mediates an increase in contractions in human myometrium. Although there are 14 distinct serotonin receptors (grouped into seven receptor subtypes), the most likely candidates to mediate effects on human myometrium are identifiable from their known signal transduction pathways in other tissues (Alexander et al., 2008). 5-HT₁ receptors are typically coupled to G-proteins of the G_i family. If present in the myometrium, their activation could lead to increased contractile activity, both by inhibiting adenylyl cyclase and cAMP accumulation (via G_iα subunits) and by stimulating phospholipase C to increase intracellular calcium (via G_iβγ subunits) (Katz et al., 1992). 5-HT₂ receptors couple to G-proteins of the G_q family to activate the phospholipase C/intracellular calcium pathway (Roth et al., 1984) (via G_qα), in a manner analogous to that of oxytocin receptors. Their role in mediating contractions in rat myometrium is well described (Cohen et al., 1985; Minosyan et al., 2007). The 5-HT₃ receptor is an ion channel that was not characterized in this study and, therefore, will not be discussed further. 5-HT₄ receptors couple to G-proteins of the Gs family and, hence, to cAMP accumulation to mediate smooth muscle relaxation, as is seen in the gut (McLean and Coupar, 1996). The signaling pathways of 5-HT₅ and 5-HT₆ receptors are not yet well characterized, and their study is hampered by a lack of selective agonists. 5-HT₇ receptors are coupled to Gs and cAMP accumulation, and their activation has been shown to cause relaxation of porcine myometrium (Kitazawa et al., 1998).

The key finding of this article is that there is strong evidence to suggest that the 5-HT_2A receptor mediates contractile effects of serotonin in pregnant human myometrium. The 5-HT₂ receptor agonist α-Me-5-HT exhibited a clear, dose-dependent stimulation of contraction. This was observed as either an increase in the amplitude of contractions or as an increase in frequency and amplitude of contractions (dependent on the donor). Consistent with previous findings with serotonin (Cordeaux et al., 2008), the response to α-Me-5-HT was potentiated by forskolin. α-Me-5-HT stimulates all three 5-HT₂ receptor subtypes, and we employed a series of more selective agonists and antagonists to identify the receptor subtype. The selective 5-HT_2B and 5-HT_2C receptor agonists, BW723C86 and Ro-60-01-75, produced inconsistent responses, with potencies that were lower than would be expected for activation of their cognate receptors. Moreover, the selective antagonists SB204741 (5-HT_2B) and RS102221 (5-HT_2C) failed to inhibit the response mediated by α-Me-5-HT. The lack of effect of the 5-HT_2C receptor antagonist was not surprising because the distribution of this receptor is thought to be limited to the central nervous system (Hoyer et al., 2002).

In contrast, the 5-HT_2A-selective antagonist, ketanserin, caused a rightward shift in the dose-response curve to α-Me-5-HT, with a pK_B value of 8.47 ± 0.16, which is consistent with previous reports for this receptor (pK_i = 8.1-9.7; Alexander et al., 2008). Moreover, the effect of ketanserin was specific because it had no effect on oxytocinmediated responses. We confirmed the presence of mRNA for the 5-HT_2A receptor by RT-PCR. Radioligand binding demonstrated specific binding of [³H]ketanserin, and this was estimated as approximately 34 fmol/mg protein. Antibodies specific for the 5-HT_2A receptor detected a protein doublet at around 55 kDa, which was consistent with the predicted molecular mass of the 5-HT_2A receptor (53 kDa) and with previous reports (Wu et al., 1998). Immunohistochemistry localized this receptor to the smooth muscle cells of the myometrium, although not all muscle cells were labeled. This uneven labeling of the myometrium, where positively stained cells are interspersed with weak or unstained cells, is similar to that previously described for the oxytocin receptor (Kimura et al., 1996). Hence, we conclude that this study provides strong evidence that the 5-HT_2A receptor mediates contraction of human myometrium.

We found some evidence that pregnant human myometrium may be stimulated by activation of the 5-HT_1A receptor. The 5-HT_1A-selective agonist 8-OH-DPAT produced a small increase in contractions in control conditions. The response was significantly increased in tissue that had been prerelaxed with forskolin, consistent with previous studies with serotonin (Cordeaux et al., 2008). The increase in contractions could also be defined as an increase in either amplitude or frequency or both and was donor-dependent. The extent of the response to 8-OH-DPAT and its potency were highly variable. This might reflect variability between donors in the expression level of the receptor(s) activated by this ligand and/or the components of the downstream signaling pathway(s). The effects of the 5-HT_1A-selective antagonist WAY100635 were also inconclusive. The antagonist depressed the maximal response to 8-OH-DPAT. However, it had no significant effect on the pEC₅₀ value of the agonist at concentrations where it is known to block the 5-HT_1A receptor (pK_i, 7.9-9.2; Alexander et al., 2008). WAY100635 has also been shown to act as an agonist at dopamine receptors (Chemel et al., 2006), which, although not yet characterized in human myometrium, could confound the interpretation of data obtained with this ligand. 8-OH-DPAT did not seem to be acting through α adrenoceptors, as has been demonstrated in rat studies (Mihályi et al., 2003), because there was no effect of blockade of these receptors. 8-OH-DPAT is also known to stimulate 5-HT₇ receptors (Bard et al., 1993; Krobert et al., 2001). However, these receptors are positively coupled to adenylyl cyclase and, therefore, unlikely to stimulate contractions. As expected, the 5-HT₇ antagonist SB269970 had no effect on the agonist-mediated response. The possibility of 8-OH-DPAT stimulating the 5-HT_2A receptor (which we have identified as being present in human myometrium) was also dismissed because ketanserin failed to inhibit the agonist-mediated response. Molecular analysis of 5-HT_1A expression was also ambiguous. The median level of 5-HT_1A mRNA was more than 100-fold lower than that of the transcript encoding the 5-HT_2A receptor. Radioligand binding studies with [³H]WAY100635 indicated a very low level of specific binding. This was 20-fold lower than that of [³H]ketanserin and was calculated at being approximately 1.5 fmol/mg protein. Antibodies used to detect the 5-HT_1A receptor by Western blotting detected multiple protein bands, and none of the three antibodies employed demonstrated a band at the predicted molecular mass of 46 kDa. Hence, our combined functional and molecular analysis does not provide strong evidence for the presence of functional 5-HT_1A receptors in human myometrium.

The 5-HT_1B/1D agonist sumatriptan, which is used clinically to treat migraine, did not stimulate contractions in human myometrium in control conditions or after relaxation with forskolin. This would suggest that the 1B and 1D receptor subtypes are not involved in modulating contractions in this tissue. Characterization of the 1E and 1F receptor subtypes is hampered by the lack of commercially available selective agonists. However, sumatriptan has been shown to display high affinity for the human 5-HT_1F receptor (Adham et al., 1993; Agosti, 2007); hence, if the receptor were abundant, a response from this ligand might have been expected.

The present study produced no evidence for inhibitory 5-HT receptors in human myometrium. We studied the effects of the 5-HT₄ receptor agonist, cisapride, and the 5-HT₇ receptor agonist, AS19, on the contractility of human myometrial strips. Neither drug elicited effects that were greater than their respective time and vehicle controls. However, there are only limited reports of the use of AS19 in the literature (Perez-García and Meneses, 2005; Perez-García et al., 2006), and its activity at 5-HT₇ receptors is not well characterized. The potent agonist 5-CT was avoided in studying the 5-HT₇ subtype because it also has been shown to activate the 5-HT_1A receptor with nanomolar potency (Newman-Tancredi et al., 1992) and bind to the 5-HT_2A receptor with submicromolar potency (Knight et al., 2004). Hence, a full characterization of these serotonin receptors, associated with smooth muscle relaxation, will depend on development of more specific agonists, but the present study did not provide evidence for such inhibitory effects using currently available methods.

These data are of potential clinical relevance. First, sumatriptan is quite widely used in pregnancy because migraine is relatively common among women in their reproductive years. Some studies have suggested an association between sumatriptan and preterm birth (Olesen et al., 2000), although other studies could not confirm this (for recent review, see Soldin et al., 2008; Källén and Lygner, 2001). The present study suggests that this drug would not be expected to stimulate myometrial contractility in pregnant women. The present data do suggest that the 5-HT_2A receptor may be a potential therapeutic target. Antagonists of this drug may have activity as uterine tocolytics, particularly in situations where endogenous levels of serotonin may be increased. Spontaneous preterm labor is known to have many of the characteristics of an inflammatory response (Lindström and Bennett, 2005). Given that the uterus has many mast cells in close proximity to the myometrium (Rudolph et al., 1993), it is possible that a drug that inhibited the response of the myometrium to serotonin may have a tocolytic effect that could be used to treat preterm labor. Conversely, 5-HT_2A agonists may have a role in contracting human myometrium, for example, in the induction of labor or in the treatment of atonic postpartum hemorrhage. In conclusion, we present strong evidence for the expression of the 5-HT_2A receptor subtype in human myometrium and that this receptor is coupled to pathways that stimulate uterine contraction.