Exogenous oxytocin modulates human myometrial microRNAs




Objective


MicroRNAs (miRNAs) play a modulatory role in pathways that lead to labor onset, although oxytocin is known to modulate gene expression within the myometrium. We aimed to identify miRNAs whose expression is regulated by oxytocin in pregnant human myometrium.


Study Design


Myometrial miRNA expression profiles were compared between samples collected from women at term before the onset of labor (no labor; n = 8) and after labor onset after early exogenous oxytocin treatment (n = 8). Multivariate modelling was used to assess differences in miRNA profiles. Biologic validation was undertaken on 3 independent patient cohorts (no labor, n = 10; labor induced with oxytocin, n = 8; and spontaneous labor with no oxytocin treatment, n = 10). In vitro studies that used primary myocytes were undertaken to assess target miRNA expression after oxytocin treatment. Target genes of candidate miRNAs were identified in silico and cross-referenced with genes that are known to be associated with labor or expressed in myometrium.


Results


In total, 1309 miRNAs were analyzed by microarray, of which 494 were detected in human myometrium. Multivariate modeling identified 12 target miRNAs the differential expression of which was most responsible for the observed separation of the 2 patient populations in the primary discovery cohorts. Biologic validation in the independent secondary sample cohorts showed that oxytocin independently regulated 5 miRNAs (hsa-miR-146b-3p, hsa-miR-196b-3p, hsa-miR-223-3p, hsa-miR-873-5p, and hsa-miR-876-5p). Additionally, hsa-miR-146b-3p was increased both in labor that was induced with oxytocin and in myometrium from spontaneous labor with no oxytocin treatment compared with no labor samples. Four of the validated miRNAs (hsa-miR-146a-5p, hsa-miR-146b-3p, hsa-miR-196b-3p, and hsa-miR-876-5p) were expressed in primary human myocytes; oxytocin treatment of these cells replicated the directional changes that were observed in vivo.


Conclusion


Oxytocin alters the expression of a unique set of myometrial miRNAs. These results suggest a further role for oxytocin as a signaling molecule that is involved in the regulation of gene expression during parturition.


The mechanisms that control the length of human pregnancy remain poorly understood. Although most women labor spontaneously from 37-42 weeks, approximately 10% deliver preterm and approximately another 10% fail to labor spontaneously at all. Both preterm and postmature births are associated with serious fetal and neonatal morbidity and death. Understanding the pathways of parturition is a critical component of developing effective future therapeutic strategies that will be designed to predict and prevent preterm birth and induce labor more effectively in the case of prolonged pregnancies.


Tightly regulated changes in gene expression in myometrium and fetal membranes are central to the onset of labor. Increasing evidence implicates microRNAs (miRNAs), which are small, noncoding, single-stranded 19-25 nucleotide molecules, as regulators of key gene expression pathways that are involved in parturition. It is thought that approximately one-third of the human genome is regulated by miRNAs, primarily at the posttranscriptional level. MiRNAs bind to the 3’untranslated region of specific target messenger RNAs (mRNAs) to inhibit their translation to protein via imperfect complementary base pairing. A number of recent studies have examined miRNA expression during labor onset in reproductive tissues that have included the placenta, cervix, and fetal membranes. Renthal et al reported that the miR-200 family of miRNAs are up-regulated in laboring murine and human myometrium at term. In mouse myometrium, ZEB1 and ZEB2 were identified as 2 targets that act as transcriptional repressors of the contraction-associated genes, which include the oxytocin receptor and connexin-43, through a mechanism that involves the repression of STAT5b, a transcriptional repressor of the P4 metabolizing enzyme 20α-hydroxysteroid dehydrogenase.


The oxytocin/oxytocin receptor system plays a major role in promoting myometrial contractility in term and preterm labor. Although circulating oxytocin levels do not rise with labor onset, the term pregnant human uterus displays increased sensitivity to oxytocin, mediated via marked up-regulation in oxytocin receptor mRNA and protein expression. Oxytocin is a potent uterotonin that is used for induction of labor and in the augmentation of dysfunctional labor. Conversely, oxytocin receptor antagonists are used as tocolytics in the management of acute preterm labor. In addition to direct effects on contractions via calcium signaling, oxytocin has been shown to function as a signaling molecule that participates in the inflammatory cascade that is associated with labor that can induce uterine activation via prostaglandin and inflammatory cytokine production.


Given the known wide range of oxytocin actions, in this study we have set out to determine whether oxytocin may affect the miRNA profile of pregnant human myometrium at term. To do this, we initially compared miRNA expression using microarray analysis in primary discovery cohorts of myometrium in samples taken from women at term who had not been in labor (NL) with samples from women whose labor was induced or augmented with oxytocin (LOxy). This identified a series of candidate miRNAs the expression of which might be effected by oxytocin. We then used quantitative real-time polymerase chain reaction (qRT-PCR) to examine the expression of these candidates in 3 independent, biologic replication sample cohorts that were taken from women at term who had not been in labor, who were in spontaneous labor but had not had oxytocin treatment, or whose labor had been induced or augmented with oxytocin. Finally, we examined the expression of those miRNAs that remained candidates after this secondary biologic validation in isolated human myocytes in vitro and demonstrated that their expression was modulated by exogenous oxytocin.


Materials and Methods


Study design


To examine miRNA expression patterns, myometrial biopsy specimens were obtained initially from women in the following groups: (1) normal term pregnancy, delivered by elective cesarean delivery before labor onset (n = 8 women) and (2) laboring at term after early exogenous oxytocin treatment, delivered by emergency caesarean (n = 8 women). A second, independent population of women was recruited as a biologic validation cohort (NL, 10 women; LOxy, 8 women; and term spontaneous labor without any oxytocin [LSpon], 10 women). Samples in the initial cohort were collected from women at Queen Charlotte’s and Chelsea Hospital, London; the validation samples were taken from women at Chelsea and Westminster Hospital, London.


Normal pregnancy was defined by the absence of medical, surgical, or obstetric complications. Term was defined as delivery at >37 completed weeks gestation. Labor was defined as regular, painful contractions with cervical dilation >3 cm. Early exogenous oxytocin treatment was defined as women who received intravenous oxytocin when cervical dilation was <4 cm and when the duration of treatment was at least for 6 hours. Women with premature rupture of membranes, multiple gestations, stillbirth, or fetal anomalies were excluded. Myometrial biopsy specimens were taken from the upper margin of lower segment incisions during cesarean delivery. Samples were washed in phosphate-buffered saline solution to remove excess blood, immediately snap frozen, and stored at –80°C. All women provided written informed consent for the collection of clinical data and tissue samples. Ethical approval for the collection of myometrial biopsy specimens was granted by the local research ethics committee (REC 3357).


RNA isolation


Total RNA was extracted from myometrial tissue (30 mg; for the discovery and validation groups) or cultured cells (approximately 1 × 10 7 ; for the in vitro model) using the NucleoSpin column kit (Macherey-Nagel, Düren, Germany) according to the manufacturer’s instructions. RNA concentration and integrity were assessed with a nanodrop spectrophotometer and Bioanalyser 2100 (Agilent Technologies, Santa Clara, CA). RNA that was isolated from myometrial tissue was used for miRNA microarray profiling and qRT-PCR validation. RNA from myometrial cells was used for specific miRNA measurement by qRT-PCR.


In vitro modeling: cell culture and cell treatments


Myometrial tissue was washed in phosphate-buffered saline solution and mechanically dissected with 2 sterile blades; the cells were isolated by incubation with 1 mg/mL of collagenase 1A, 1 mg/mL of collagenase X, and 20 mg of bovine serum albumin in 40 mL of Dulbecco’s modified Eagle’s medium for 45 minutes at 37°C. The suspension was filtered through a cell strainer and centrifuged at 400 g for 5 minutes. The cells were resuspended and cultured in Dulbecco’s modified Eagle’s medium that contained 10% fetal calf serum, 2 mmol/L L-glutamine, 100 μg/mL penicillin, and 100 μg/mL of streptomycin and grown in T25 flasks at 5% CO 2 . Myocytes were cultured until confluent and treated with 100 nmol/L oxytocin (Alliance Pharmaceuticals, Chippenham, UK) for 1, 2, and 4 hours or 1 μg/mL interleukin 1β for 2 hours (or alcohol vehicle control).


Microarray


Microarray expression analysis of 1309 miRNAs was performed with the miRCURY LNA miRNA microarray (version 8.0; Exiqon, Vedbaek, Denmark). Samples were labeled with miRCURY LNA miRNA Hi-Power Labeling Kit, Hy3/Hy5, and were hybridized on the array. To minimize nonbiologic experimental variation, samples were randomized for the array and spike-in oligo nucleotide controls were included. The quantified signals (background corrected) were normalized with the global locally weighted scatterplot smoothing regression algorithm. The positive effect of normalization was confirmed by assessment of log ratio/log mean-intensity plots.


Real-time qRT-PCR for miRNA


QRT-PCR was used to quantify the expression of candidate miRNAs with the miRCURY LNA kit (Exiqon) with 0.05 ng complementary DNA. LNA assays were custom designed (Exiqon) for miRNAs of interest: hsa-miR-21-5p (204230), hsa-miR-29a-3p (204698), hsa-miR-29a-5p (204430), hsa-miR-29b-3p (204679), hsa-miR-29c-3p (204729), hsa-miR-29c-5p (204132), hsa-miR-141-3p (204504), hsa-miR-146a-5p (204688), hsa-miR-146b-3p (204374), hsa-miR-196b-3p (204619), hsa-miR-223-3p (204256), hsa-miR-873-5p (204175), and hsa-miR-876-5p (204527). MiRNA expression data were normalized to 5S ribosomal RNA (203906).


Statistical analysis


Background signal was defined as expression 1.2 times the 25th percentile of overall signal intensity and subtracted from the array dataset. MiRNAs that were detected above background in ≥6 samples in any phenotypic group, and therefore expressed in most of those samples, were included in the analysis. Array data were subjected to unsupervised principal components analysis (PCA) and supervised partial least squares–discriminatory analysis (PLS-DA) with the use of SIMCA-P (Soft independent modelling of class analogies-P; version 13.0.2; Umetrics, Umeå, Sweden). Data were scaled to unit variance by the division of each variable by 1/(S k ), where S k represents the standard deviation value of the variable to ensure that all variables retained equal importance during the generation of the model. Goodness-of-fit and predictive ability parameters were calculated (the latter by a 7-round internal cross-validation of the goodness-of-fit parameter) and used to assess model fit. Although PCA permits an overall examination of covariance in the datasets, PLS-DA was used to examine specific differences in the miRNA profiles that were associated with patient groups. PLS-DA analysis differs from PCA in that the algorithm is provided with class membership of clinical sample groups and components are projected and rotated around the axis to obtain maximum separation between these classes, thus permitting the best description of class-related variability. MiRNAs that differed most between classes were identified by examination of their calculated variable importance for the projection (VIP) score, which was calculated in SIMCA-P with a proprietary algorithm. The VIP score describes the relative contribution of each miRNA (X variables) to patient class variation (Y variables) accounted for by each component in the PLS-DA model. A VIP score >1 is considered to indicate an important contributor to the separation of patient classes that are observed in the PLS-DA model. In this study, miRNAs that exhibit VIP scores >1.5 were defined as significant and selected for further investigation.


For qRT-PCR, relative expression of individual miRNAs in tissue and cell samples was calculated by the ΔΔCT method. Statistical significance of the data was inferred with a 2-tailed Mann-Whitney test for tissue samples and a 1-way analysis of variance for treated cells. A probability value of < .05 was considered statistically significant.


Identification of potential target genes


Target genes of the candidate differentially expressed myometrial miRNAs were sought with the use of the miRTarBase platform, which describes experimentally validated miRNA targets in the current literature ( http://mirtarbase.mbc.nctu.edu.tw ). These candidate genes were cross-referenced with genes that previously were reported in gene microarray studies of pregnant human myometrium and those known to be implicated in parturition. This method produced a focused short-list of validated potential gene targets.




Results


Exogenous oxytocin modulates the expression of a subset of miRNAs in human myometrium


Analysis of patient demographic characteristics revealed no significant differences in age, ethnicity, parity, or gestation between those women who provided myometrial biopsy specimens that were taken before labor onset (n = 8) and those who received exogenous oxytocin (n = 8; Table 1 ). A total of 494 miRNAs were detected above background expression levels and were used for multivariate modeling of LOxy and NL myometrium profiles ( Figure 1 ). Unsupervised PCA was first used to assess overall variation of the dataset. As seen in Figure 1 , A, no obvious clustering of the 2 patient populations was detected, which indicates that most miRNA variation in the sample population was independent of oxytocin-induced/augmented labor. To focus on oxytocin-specific–induced variation in the miRNA profiles, supervised PLS-DA was used ( Figure 1 , B). Analysis of the calculated VIP values ( Figure 1 , C) permitted selection of a panel of 12 miRNAs that were most responsible for the observed separation of the 2 patient populations. Those miRNAs with a VIP score ≥1.5 were defined as significantly contributing to the separation in the model. These comprised hsa-miR-21-5p, hsa-miR-29a-3p, hsa-miR-29b-3p, hsa-miR-29c-3p, hsa-miR-29c-5p, hsa-miR-141-3p, hsa-miR-146a-5p, hsa-miR-146b-3p, hsa-miR-196b-3p, hsa-miR-223-3p, hsa-miR-873-5p, and hsa-miR-876-5p.



Table 1

Demographics of original sample cohorts































Characteristic Term P value
Not in labor (n = 8) In labor receiving oxytocin (n = 8)
Median maternal age, y (range) 37 (24–44) 31 (27–36) .08 a
Median parity, n (range) 1 (0–2) 0 (0–1) .06 a
Black ethnicity, n (%) 1 (11) 1 (13) 1.0 b
Median gestational age at delivery, wk (range) 39 (38–39) 40 (38–41) .06 a

The maternal age, parity, ethnicity, and gestation of the original cohort participants are described. There were no significant differences in age, ethnicity, parity, or race between the 2 clinical groups.

Cook. Oxytocin modulates human myometrial miRNAs. Am J Obstet Gynecol 2015 .

a Continuous variables compared by Mann-Whitney tests


b Discrete variables compared by χ 2 test.




Figure 1


Multivariate modeling of microRNA expression in myometrium of women laboring with oxytocin and those women not in labor

A, Unsupervised principal components analysis of myometrial microRNA expression did not reveal any clear clustering of patients not in labor ( blue ) or patients laboring with oxytocin ( red ), which indicates that most variance in the data was independent of oxytocin treatment. B, Supervised partial least squares–discriminatory analysis of microRNA expression profiles facilitated separation of the patient classes. C, Variable importance plot scores show which individual microRNA variance contributes to the separation of phenotype (ie, VIP, ≥1.5).

LOxy , labor with oxytocin; NL , not in labor; PC1 , first principal component; PC2 , second principal component; VIP , variable importance plot.

Cook. Oxytocin modulates human myometrial miRNAs. Am J Obstet Gynecol 2015 .


The laboring myometrial miRNA profile of women who were treated with oxytocin differs from those in spontaneous labor


To validate our findings, qRT-PCR was used to examine the expression of the 12 selected target miRNAs in myometrial samples that were isolated from secondary, independent validation cohorts of women at term not in labor (n = 10), at term in labor after early oxytocin treatment (n = 8) and additionally women in spontaneous labor not receiving oxytocin (n = 10). Demographic characteristics of the validation cohort are provided in Table 2 . No significant differences in age, ethnicity, or gestation between the patient groups were observed, although all women who received oxytocin were nulliparous, which is consistent with the common clinical use of exogenous oxytocin. To examine whether parity impacts on the expression of our target miRNAs, we undertook logistic regression on available PCR data from 6 miRNAs that were analyzed in our discovery set. When adjusted for parity, relative expression of miRNAs was not affected ( Table 3 ). Similar analyses were not performed in the validation cohort because all women in the LOxy group were primiparous and most women in the NL group were multiparous. Differential expression of 6 of the 12 candidate miRNAs (hsa-miR-146a-5p, hsa-miR-146b-3p, hsa-miR-196b-3p, hsa-miR-873-5p, hsa-miR-876-5p, and hsa-miR-223-3p) was confirmed in the myometrium of women who received oxytocin in labor (LOxy vs NL; Figure 2 ). Expression of hsa-146a-5p, hsa-miR-196b-3p, hsa-miR-873-5p, and hsa-miR-876-5p was decreased in LOxy myometrium (–4 fold change, P = .02; –3.3 fold change, P = .04; –5 fold change, P = .01, and –10 fold change, P = .02, respectively), whereas hsa-miR-146b-3p and hsa-miR-223-3p levels were increased in the myometrium of oxytocin-treated women (2.7 fold change, P = .0055; 9 fold change, P = .0025, respectively). Only hsa-miR-146b-3p was found to be up-regulated in both LSpon (3.3 fold change, P = .04) and LOxy groups ( Figure 2 ).



Table 2

Demographic data from the biologic validation cohorts









































Characteristics Term P value Term, in labor, receiving oxytocin (n = 9) P value
Not in labor (n = 10) In labor, no exogenous oxytocin (n = 8)
Median maternal age, y (range) 33 (19–41) 30 (19–37) .7 a 36 (26–41) .1 a
Median parity, n (range) 1 (0–2) 0.5 (0–2) .56 a 0 .01 a
Black ethnicity, n (%) 1 (10) 0 1.0 b 1 (10) 1.0 b
Median gestational age at delivery, wk (range) 39 (37–41) 39 (37–41) .32 a 40 (38–42) .08 a

The maternal age, parity, ethnicity, and gestation of the validation cohort participants are described. There were no significant differences in age, ethnicity, or gestation between the nonlaboring and laboring with exogenous oxytocin groups, however, women in the laboring with exogenous oxytocin groups were more likely to be primiparous ( P = .01). There were no significant differences in age, parity, ethnicity, or gestation between the nonlaboring and spontaneous laboring groups.

Cook. Oxytocin modulates human myometrial miRNAs. Am J Obstet Gynecol 2015 .

a Continuous variables compared by Mann-Whitney tests


b Discrete variables compared by χ2 test.



Table 3

Expression of microRNAs in original cohort using polymerase chain reaction


































MicroRNA Women in labor who received exogenous oxytocin compared with women not in labor, P value
Unadjusted Adjusted for parity
hsa-miR-200c-3p .25 .15
hsa-miR-29c-3p .023 .012
hsa-miR-29b-3p .022 .034
hsa-miR-223-3p .012 .0038
hsa-miR-873-5p .0007 .0040
hsa-miR-876-5p .0009 .0042

Univariate logistic regression modeling was used to compare expression of a selection of individual myometrial microRNAs that were measured with real-time polymerase chain reaction in the same samples as those analyzed by microarray. Samples were taken from women in labor who received oxytocin and women not in labor. The findings reproduce those of the microarray. Multivariate logistic regression was also used to adjust for parity, which although not significantly different in the 2 groups, was slightly dissimilar ( P = .06). The probability values remained consistent, and we conclude that small variations in parity did not introduce bias to our results. Analysis was carried out using STATA software (StataCorp LP, College Station, TX).

Cook. Oxytocin modulates human myometrial miRNAs. Am J Obstet Gynecol 2015 .

Only gold members can continue reading. Log In or Register to continue

Stay updated, free articles. Join our Telegram channel

May 6, 2017 | Posted by in GYNECOLOGY | Comments Off on Exogenous oxytocin modulates human myometrial microRNAs

Full access? Get Clinical Tree

Get Clinical Tree app for offline access