Oxytocin Pretreatment Attenuates Oxytocin-induced Contractions in Human Myometrium In Vitro

OXYTOCIN is a neurohypophyseal hormone that induces uterine contractility and is used for the induction and augmentation of labor as well as to decrease postpartum blood loss. It binds to myometrial oxytocin receptors to induce uterine contractility via activation of phospholipase C and release of inositol 1,4,5-triphosphate, 1,2-diacylglycerol, and intracellular calcium.¹  Oxytocin receptors belong to the family of G-protein–coupled receptors, and like other G-protein–coupled receptors undergo rapid molecular desensitization due to homologous stimulation.²  Continuous and prolonged exposure of cultured human myometrial cells to oxytocin in vitro down-regulates oxytocin receptors, and decreases oxytocin receptor messenger RNA (mRNA) expression and oxytocin binding sites, coincident with a loss of myometrial cellular responsiveness.^3–6  This desensitization phenomenon has been shown to occur after both in vitro and in vivo exposure of myocytes to oxytocin.^5,6 

Desensitization of oxytocin receptors may have important clinical implications because oxytocin is the most commonly used uterotonic drug for prophylaxis and treatment of primary postpartum hemorrhage. The prolonged exposure of the uterus to supraphysiologic concentrations of oxytocin (for induction or augmentation of labor) may reduce the immediate postpartum effect of this first-line drug.^7–9  In a recent study in pregnant rats, we demonstrated that pretreatment with oxytocin leads to a significant decrease in the in vitro oxytocin-induced myometrial contractions in a concentration-dependent manner, likely as a function of the desensitization phenomenon.¹⁰ 

Although the phenomenon of desensitization has been extensively demonstrated in cultured human myometrial cells,^3–6  it has not yet been shown in the isolated human myometrium, and there is no report in the literature using a reliable contractility model for in vitro study of the desensitized human myometrium. We hypothesized that, similar to the observations in human biomolecular and isolated rat myometrial contractility studies, the pretreatment of the human myometrium with oxytocin would cause further decrease in oxytocin-induced myometrial contractility in a concentration and/or time-dependent manner. The objective of this study was to investigate the in vitro contractile response of the human term pregnant myometrium to oxytocin, after pretreatment with different concentrations of oxytocin for variable durations.

After approval by the Research Ethics Board at Mount Sinai Hospital, Toronto, Ontario, Canada, the study was conducted in 62 nonlaboring term pregnant women undergoing elective cesarean deliveries (CD). We obtained written informed consent from all women enrolled in the study. Inclusion criteria were women with gestational age 37–41 weeks requiring an elective primary or first repeat CD under spinal anesthesia for any reason, although, most common indications were breech presentation or repeat CD. Exclusion criteria were: previous uterine surgery other than a single previous CD, a requirement for general anesthesia, and any condition predisposing to uterine atony and postpartum hemorrhage, such as abnormal placentation, multiple gestation, preeclampsia, macrosomia, polyhydramnios, uterine fibroids, bleeding diathesis, chorioamnionitis, or a previous history of postpartum bleeding. The study was registered with the ClinicalTrials.gov registry (NCT-00989027).

Drug solutions were prepared by serial dilutions in sterile double-distilled water. Oxytocin (lyophilized powder), and all salts and reagents used in the preparation of the physiological salt solution (PSS), and 3-N-morpholino propanesulphonic acid solution, were obtained from Sigma-Aldrich Canada Ltd., Oakville, Ontario, Canada. The 3-N-morpholino propanesulphonic acid solution buffer (145 mM NaCl, 4.7 mM KCl, 1.5 mM CaCl₂, 1.17 mM MgSO₄·7H₂O, 1.2 mM NaH₂PO₄·H₂O, 3.0 mM 3-N-morpholino propanesulphonic acid solution, 5.0 mM glucose, and 2.0 mM pyruvate) and PSS (112 mM NaCl, 25 mM NaHCO₃, 1 mM KH₂PO₄, 5 mM KCl, 1.2 mM MgSO₄·7H₂O, 11.5 mM glucose, and 2.5 mM CaCl₂) were prepared in advance and stored at 4–8°C.^10–12 

The patients were administered spinal anesthesia by the attending anesthesiologist and underwent CD via a Pfannenstiel incision, as per the routine in this hospital. The obstetrician excised a small sliver of myometrium from the upper border of the incision on the lower uterine segment after the delivery of the neonate, just before oxytocin administration. The collected myometrial tissue was immediately immersed in the 3-N-morpholino propanesulphonic acid solution buffer, which was placed on ice for tissue dissection. Four longitudinal myometrial strips (2 × 2 × 10 mm³ each) were isolated and suspended vertically in individual, temperature-controlled organ bath chambers, containing 10 ml of PSS at pH 7.4. Each myometrial strip was clipped between the base of the organ bath chamber and an isometric force transducer (Radnoti four unit tissue-organ bath system, model 159920; Harvard Apparatus Canada, Saint-Laurent, Quebec, Canada). The organ bath solution was continuously aerated with 95% oxygen and 5% carbon dioxide at 37°C.

As described in previous studies,^10–12  during each experiment, each isolated myometrial strip was allowed to equilibrate in PSS at 1 g tension for at least 20 min of spontaneous regular contractions. If the spontaneous contractions did not start immediately, the PSS was changed by flushing every 10–15 min until spontaneous contractions were observed. After equilibration, each myometrial strip was exposed to an isosmotic 120 mM of potassium chloride solution to induce a contraction that reflected the maximum contractile capacity of the tissue. The potassium chloride solution was then drained from the organ baths and any residual solution was removed by washing it three times with PSS.

Individual strips were then pretreated with oxytocin at 10⁻¹⁰, 10⁻⁸, or 10⁻⁵M concentration (experimental group) or bathed in PSS without oxytocin (control group) for a 2-, 4-, 6-, or 12-h period. After pretreatment, the strips in the control group were given fresh PSS, and all oxytocin-pretreated strips in the experimental groups were washed twice with PSS. All strips were then subjected to a dose–response testing with oxytocin 10⁻¹⁰M to 10⁻⁵M (fig. 1), increased cumulatively in a pattern of one log molar concentration every 10 min. At the end of the dose–response period, the myometrial strips were washed with PSS and exposed to a final isosmotic 120 mM potassium chloride solution to assess the muscle viability. The strips were then removed from the apparatus, manually dried and weighed.

The amplitude and frequency of myometrial contractions were continuously recorded by force displacement transducers connected to a data acquisition system (MP 100 with Acknowledge 3.9.0 software; Biopac System Inc., Goleta, CA). The data sets from samples that failed to contract or those with technical errors were discarded.

In each group, the amplitude (g) and frequency (number of contractions in 10 min) of contractions were recorded, and the motility index (amplitude × frequency/10 min; Montevideo units) and the area under the curve (AUC; integral force; gram × minute) were derived to reflect the uterine activity and the strength of contractions, respectively.^12–14  These parameters measured during the dose–response period were compared among various study groups. We used linear regression models (maximum likelihood estimates for parameter determination) adjusted for repeated measures through a compound symmetry covariance structure. Generalized estimating equations (GEE) were generated, including duration of oxytocin preexposure, oxytocin concentration during preexposure, and oxytocin concentrations during the dose–response period, as independent variables. Outcomes (amplitude, frequency, motility index, and AUC) were square root (SR) transformed (selected on the basis of Akaike information criteria) to adjust for their skewed distribution of both motility index and AUC. The model was adjusted for baseline tone, weight of sample, maximum amplitude after potassium (administered before the pretreatment period), and measures of amplitude, frequency, motility index, and AUC of baseline contractions, during the equilibration period. All these values were added as variables in the multivariable regression model. The statistical analyses were performed using SAS statistical software v9.2 (The SAS Institute, Cary, NC). All hypotheses were tested using two-tailed testing, and a value of P less than 0.05 was considered statistically significant.

Patients were recruited for this study between August 2008 and December 2010. Of 129 women approached for participation in the study, 65 consented, and myometrial samples were obtained from 62 subjects. Each sample was divided into four pieces for four separate experiments, and a total of 163 experiments were conducted successfully (fig. 2). Among the excluded experiments, there was a failure to obtain oxytocin-induced contractions during the dose–response in 0, 6, 4, and 15% of samples in the 2-, 4-, 6-, and 12-h groups, respectively. The clinical characteristics of the patients were similar among the study groups (table 1).

After adjustment for all covariates and baseline contractility parameters, the duration of oxytocin preexposure had significant effect on all the contractility parameters, except frequency. The amplitude, motility index and AUC during the dose–response period decreased with the increased duration of oxytocin preexposure from 2 to 12 h; however, the effect was mainly driven by the 6- and 12-h preexposure groups. Longer preexposure was associated with significantly lower amplitude (type 3 GEE analysis: chi-square = 14.0; df = 3; P = 0.003), motility index (type 3 GEE analysis: chi-square = 9.3; df = 3; P = 0.03; fig. 3), and AUC (type 3 GEE analysis: chi-square = 10.5; df = 3; P = 0.02; fig. 4), but had no statistically significant effect on frequency of contractions (P = 0.10). These effects were seen when all groups, including oxytocin preexposed and control, were analyzed together.

After adjustment for all covariates and baseline contractility parameters, we found that the amplitude of contractions was significantly lower during the dose–response period with oxytocin pretreatment 10⁻⁵M (P = 0.02; type 3 GEE for all four concentrations: P = 0.06), however, there was no statistically significant difference between 10⁻⁸M (P = 0.07) or 10⁻¹⁰M (P = 0.93) pretreatment groups and the control groups (table 2). The frequency of contractions was significantly lower with oxytocin pretreatment 10⁻⁵ and 10⁻⁸M (P = 0.03 and P = 0.002, respectively; type 3 GEE for all four concentrations: P = 0.02) as compared with the control groups, however, there was no difference between oxytocin 10⁻¹⁰M pretreatment and the control group (table 2). The mean (standard error) values for the amplitude and frequency of contractions in various oxytocin pretreatment groups are shown in table 3.

The mean (standard error) values for the motility index and AUC in various oxytocin pretreatment groups are shown in table 4. Overall, there was a statistically significant effect of the pretreatment concentration on the motility index during the dose–response period (P = 0.02). The pretreatment concentrations of 10⁻⁵M (P = 0.005) and 10⁻⁸M (P = 0.02) were associated with lower motility index than the control group, whereas the 10⁻¹⁰M concentration did not differ (P = 0.35) from control (table 2; fig. 3). For the pretreatment concentration of 10⁻⁵M, the effect of decreased myometrial contractility versus control was preserved at 2 h (estimate [EST; standard error] = −1.131 [0.483] SR unit, P = 0.02) and 4 h (EST = −0.957 [0.375] SR unit; P = 0.01). The same trend of decreased myometrial contractility versus control was not seen at 6 h (EST = −0.684 [0.445] SR unit; P = 0.32) or 12 h (EST = −0.442 [0.864] SR unit; P = 0.61) of preexposure. For the pretreatment concentration of 10⁻⁸M, there was no effect at 2 h (EST = −0.600 [0.393] SR unit; P = 0.13), a statistically significant effect at 4 h (EST = −1.088 [0.511] SR unit; P = 0.03), and no effect at 6 h (EST = −0.656 [0.507] SR unit; P = 0.20) or 12 h (EST = +0.150 [0.436] SR unit; P = 0.58) of preexposure. The pretreatment with 10⁻¹⁰M oxytocin did not have any significant effect on the motility index with any pretreatment duration.

Overall, the pretreatment concentration had no statistically significant effect on the AUC during the dose–response period (P = 0.10), however, pretreatment concentration of 10⁻⁵M (P = 0.04) was associated with significantly lower AUC than control, whereas 10⁻⁸M (P = 0.05) and 10⁻¹⁰M (P = 0.65) were not statistically significant when compared with control (table 2). The decrease in AUC compared with the control in the oxytocin 10⁻⁵M pretreatment group was preserved at 2 h (EST = −8.750 [2.790] SR unit; P = 0.002) and 4 h (EST = −10.164 [2.269] SR unit; P < 0.001), but not at 6 h (EST = −2.530 [2.667] SR unit; P = 0.34) or 12 h of preexposure (EST = −5.883 [5.081] SR unit; P = 0.25). The pretreatment with 10⁻⁸ and 10⁻¹⁰M oxytocin did not have any statistically significant effect on the AUC for any pretreatment duration (fig. 4).

Increasing the concentration of oxytocin by 10⁻¹M increments from 10⁻¹⁰ to 10⁻⁵M during the dose–response period led to a statistically significant change in the motility index (P < 0.001) as well as the AUC (P < 0.001) in a consistent manner, with an initial increase in the slope until 10⁻⁶M concentration, which was followed by a plateau. The response was similar for all durations of preexposure (i.e., at 2, 4, 6, and 12 h). There was no difference in the slopes of the dose–response curves for motility index and AUC between oxytocin-pretreated and control groups, or across the various durations of exposure (type 3 GEE analysis for preexposure time interaction with concentration: P = 0.09 for motility index and P = 0.20 for AUC). The lack of interaction between the effect of preexposure concentration and duration suggests that these variables are independent of each other (i.e., additive), but are not multiplicative of each other. The representative recordings of myometrial contractions after pretreatment with various concentrations of oxytocin for 2 h are shown in figure 5.

Our study demonstrates that oxytocin-induced contractile response of the isolated term human myometrium is attenuated by previous oxytocin exposure, and suggests that this response is dependent on both concentration and duration of oxytocin preexposure. Our results reproduce our previous findings in rat myometrium and are, therefore, of practical importance because this effect may be seen with clinically used doses of oxytocin.¹⁰  The similarity of the response in both rat and human myometrium is an important validation step in utilizing the rat myometrium model for in vitro myometrial contractility studies.

The most likely explanation for our findings, in keeping with the existing evidence in the literature, is the oxytocin receptor desensitization phenomenon.^3–6  Desensitization of receptors is a complex physiologic process, one that has evolved to prevent endogenous hyperstimulation by impairing signal transduction during prolonged receptor activation. The possible mechanism of oxytocin receptor desensitization involves phosphorylation, uncoupling from the G protein and internalization, leading to recycling of the receptor to the cell membrane or degradation by lysosomes.¹⁵  This phenomenon has been shown in a variety of ways at the biomolecular level.

Robinson et al. worked with an in vitro human myocyte culture model and showed that cells preexposed to oxytocin 10⁻⁸M undergo desensitization in the form of inhibition of oxytocin-induced calcium efflux. This effect is seen within 3 h of oxytocin stimulation, resulting in 50% desensitization at 4.2 h and complete desensitization within 6 h.³ 

Phaneuf et al.⁵  demonstrated the same phenomenon by reporting a 50-fold down-regulation of oxytocin receptor mRNA in myometrium obtained at emergency CD, when labor had been stimulated with intravenous oxytocin for more than 12 h. The same investigators subsequently demonstrated that in women with spontaneous and induced labors, resulting in CD after more than 10 h of labor, oxytocin receptor mRNA levels decreased by 60 and 300 times, respectively, when compared with nonlaboring women.⁶  They also found a decrease in oxytocin binding related to the increasing duration and dose of oxytocin infusion administered for labor induction.⁶ 

Previous in vitro studies in rat myometrium demonstrated inhibition of oxytocin-induced myometrial contractions after exposure to oxytocin 10⁻⁸M for 1 h,^10,11  without any obvious dependence on the duration of preexposure. In the human myometrial preparation used in our current study, however, pretreatment with either oxytocin 10⁻⁵M for at least 2 h, or 10⁻⁸M for at least 4 h was required for producing such effect. This suggests that with a low concentration of oxytocin preexposure, a longer duration of exposure might produce similar inhibitory effects on myometrial contractility, whereas, with a higher concentration, this effect may be seen earlier. We observed that with increasing concentration of oxytocin pretreatment, both amplitude and frequency of contractions during the dose–response are attenuated, thereby, decreasing the motility index as well as the AUC. However, the decrease in the motility index and AUC seen with increasing the duration of preexposure to oxytocin are likely to be predominantly due to the change in amplitude, rather than in frequency of contractions. Our results also suggest that the rat myometrium may be prone to the desensitization effect at lower oxytocin concentrations than in human myometrium, but further studies are warranted to confirm these findings.

The findings of our study may have important clinical implications. Previous studies have demonstrated that women with oxytocin-augmented labors require nine-fold greater doses of oxytocin (ED₉₀ [95% CI] = 2.99 [2.32–3.67] IU), to produce effective uterine contractions during CD as compared with nonlaboring women (ED₉₀ [95% CI] = 0.35 [0.18–0.52] IU). Despite the use of this higher dose of oxytocin, these women bled twice as much as the nonlaboring women (1,178 [716] vs. 693 [487] ml).^16,17  A meta-analysis of 11 randomized clinical trials on the use of oxytocin for labor induction demonstrated that more aggressive or high-dose oxytocin protocols resulted in more episodes of uterine hyperstimulation, lower rate of spontaneous vaginal delivery, and higher incidence of postpartum maternal infection and hemorrhage, compared with low-dose regimens, in which doses were not increased more than every 30 min.⁸  Similarly, a recent study demonstrated a higher incidence of severe postpartum hemorrhage secondary to uterine atony, when oxytocin was used for labor augmentation in larger doses and for longer durations.⁹  These clinical findings are likely to be due to oxytocin receptor desensitization and signal attenuation^3,5,6  that have been demonstrated by molecular mechanism and are now confirmed to affect myometrial contractility in vitro.

The concentrations of oxytocin in our study are likely to be on the upper limit of physiological levels during labor (10⁻¹⁰M) and/or circulating levels observed with high doses of oxytocin induction protocols (10⁻⁸M) in humans. Although there is a huge variability of these data in the literature, the plasma levels of oxytocin in laboring women, and those with oxytocin augmentation, have been shown to vary from 10⁻⁹ to 10⁻¹²M,^18–21  and the concentrations reached after delivery are likely to be much higher due to the practice of administration of high doses of oxytocin within a short timeframe. In fact, plasma levels of oxytocin have been shown to correlate with the rate of oxytocin infusion rather than with the uterine activity during labor.²²  The rationale for choosing different preexposure durations was based on the time-dependent oxytocin receptor desensitization seen in human myometrium.^3,5,6 

The most important limitation of our study is that it was an in vitro preparation. The effect of the same magnitude may be difficult to achieve in in vivo settings. Our experimental model produced consistent results up to 6-h duration, showing increased myometrial inhibitory effect with increasing concentration of oxytocin pretreatment. However, at 12 h, such effect was no longer detectable, which could be related to the limitations of the model itself. We observed a higher failure rate and an overall reduction in myometrial contractility during the dose–response period in the 12-h group, perhaps due to tissue exhaustion and/or deprivation of metabolites in the in vitro settings. Hence we suggest that this experimental model should not be used for pretreatment duration longer than 6 h.

Our findings suggest the need for a reappraisal of current regimens used for prevention and treatment of primary postpartum hemorrhage in women who have received oxytocin augmentation of labor. Further, our study attests to the validity of an in vitro model, demonstrating reduced response to oxytocin after oxytocin pretreatment of the human myometrium. Our model can be used for future in vitro studies on the oxytocin-pretreated human myometrium, to define new strategies to optimize uterine contractility, and reduce complications associated with postpartum hemorrhage.

The authors thank Cedric Manlhiot, B.Sc., Clinical Research Program Manager, Labatt Family Heart Centre, The Hospital for Sick Children, Toronto, Ontario, Canada, for his help in statistical analysis of this study. The authors also thank Samah Hassan, M.D., Research Assistant, Mount Sinai Hospital, Toronto, Ontario, Canada, for assisting with experimentation, and Stephen J. Lye, Ph.D. (Associate Director, Research), and S. Lee Adamson, Ph.D. (Senior Investigator), from the Samuel Lunenfeld Research Institute, Mount Sinai Hospital, Toronto, Ontario, Canada, for their continued guidance and support for our research.