Nitric oxide in the human uterine cervix: Endogenous ripening factor

Abstract

The human uterine cervix can produce nitric oxide (NO), a free radical with an ultra‐short half‐life. The release of NO changes during pregnancy and is increased in early nonviable pregnancies compared to normal uncomplicated pregnancies. This review concentrates on the role of NO release in cervical ripening in pregnant women. Also some suggestions on future aspects are discussed.

• cervical ripening
• nitric oxide
• nitric oxide metabolites
• pregnancy

Introduction

Nitric oxide (NO) is a small, uncharged gas molecule that is a highly reactive free radical with an extremely short half‐life of approximately 4 seconds. In the human female reproductive tract NO plays an important role in various physiological processes, including cervical ripening 1.

The uterine cervix has a pivotal role in the physiology of gestation and parturition; it has to be firm enough to retain the conceptus throughout the pregnancy and, on the other hand, to soften before and during labor to enable the birth of the infant. Cervical adaptation during pregnancy is a chronic process, which begins within the first trimester of pregnancy and progressively proceeds until term, and is usually described as softening, ripening, dilation, and finally, remodeling of the cervix. Obviously with this four‐step process, which must occur in sequence, complex systems must be in place to control each forward and irreversible step 2–4. Effacement and dilation are often associated with or contributed to uterine contractions, but it is evident that softening occurs independently of contractions. Further, effacement and dilation are often present without uterine contractions.

The cervix, which is dominated by fibrous connective tissue, is composed of an extracellular matrix consisting predominantly of collagen (70% type I and 30% type III) along with elastin, proteo‐ and glucosaminoglycans, and water, and a cellular portion consisting of smooth muscle, fibroblasts, epithelium, and blood vessels 5. Throughout most of gestation the cervix remains rigid and closed to secure the products of conception. A dramatic functional shift occurs during parturition to secure uncomplicated delivery. The mechanisms involved in cervical ripening are complex and are only known to a certain extent. The cascade responsible for the process of cervical ripening, and which finally enables uterine contractions to efface and dilate the cervix, is complex.

The cervical ripening is actively controlled with features similar to those in inflammation and in rearrangement of the cervical collagen fibers 6, 7. Cervical ripening is thus associated with changes in local cytokines, prostaglandins, and metalloproteases, as well as in other bioregulators which play roles in inflammation and in collagen metabolism 6, 7. These factors take part also in the regulation of NO synthesis and release 1, and the present review is to clarify the role of cervical NO in the ripening of the human uterine cervix.

Synthesis of nitric oxide

Nitric oxide is formed from L‐arginine 8 through nitric oxide synthases (NOS) 9, a group of enzymes that structurally resemble cytochrome P‐450 reductase 10. The biosynthesis of NO takes place from L‐arginine and molecular oxygen utilizing nicotine adenine dinucleotide phosphate as an electron donor, and heme, tetrahydrobiopterin, calmodulin, flavin adenine mono‐ and dinucleotide as cofactors through a reaction that consumes five electrons. The overall reaction consists of a two‐step oxidative conversion of L‐arginine to NO and L‐citrulline via Nw‐hydroxy‐L‐arginine as an intermediate, with monooxygenase I and monooxygenase II, in each step a mixed‐function oxidation taking place.

Three NOS isoenzymes have been characterized: neuronal NOS (type I, nNOS), inducible NOS (type II or iNOS), and endothelial NOS (type III or eNOS) 11–13. Both nNOS and eNOS are expressed constitutively, and their activity is calcium/calmodulin‐dependent, whereas the expression of iNOS is induced by bacterial lipopolysaccharides (LPS) and cytokines, independently of calcium 14 (Table I).

Table I. Randomized controlled trials on nitric oxide donors in cervical ripening in pregnant women.

Reference	No. of women	Drug (dose)	Compared with	Trim ^a	Exposure time (hrs)	Comparison of efficacy ^b
Thomson et al. 1997 32	48	IMN (40 mg); GTN (0.5 mg)	Placebo; Placebo	I	3; 3	IMN>Placebo; GTN = Placebo
Thomson et al. 1998 85	66	IMN (40 mg); IMN (80 mg)	Gemeprost (1 mg); Gemeprost (1 mg)	I	3; 3	IMN<Gemeprost; IMN<Gemeprost
Facchinetti et al. 2000 94	36	SNP (5 mg); SNP (10 mg)	Placebo; Placebo	I	6; 3	SNP>Placebo; SNP>Placebo
Ledingham et al. 2001 107	65	IMN (40 mg)	Misoprostol (0.4 mg)	I	2–3	IMN<Misoprostol
Li et al. 2003 88	126	IMN (40 mg)	Placebo or Misoprostol (0.4 mg)	I	4–6	IMN = Placebo; IMN<Misoprostol
Chan et al. 2005 95	200	SNP (10 mg)+Placebo	Misoprostol (0.4 mg)+Placebo	I	3	SNP<Misoprostol
Arteaga‐Troncoso et al. 2005 99	60	IMN (80 mg)	Misoprostol (0.4 mg)	I	12 (max 4 doses)	IMN>Misoprostol
David et al. 2005 91	30	IMN (40 mg)	Misoprostol (0.2 mg)	I	3	IMN = Misoprostol
Radulovic et al. 2007 106	120	IMN (40 mg)	Misoprostol (0.2 mg)	I	10	IMN<Misoprostol
Li et al. 2003 89	100	IMN (40 mg)	Placebo	II	12 (after 1 dose Miso.)	IMN = Placebo
Eppel et al. 2005 90	72	IMN (40 mg)+Gemeprost (1 mg)	Placebo+Gemeprost (1 mg)	II	48 (max 3 doses/d)	IMN>Placebo
Chanrachakul et al. 2000 96	110	GTN (0.5 mg)	Dinoprost (3 mg)	III	6	GTN<Dinoprost (CS rate 35% versus 35%)
Nicoll et al. 2001 86	36	IMN (20 mg); IMN (40 mg)	Vaginal exam.; Vaginal exam.	III	6	IMN = vag. exam. (CS rate 46% versus 33%); IMN = vag. exam. (CS rate 18% versus 33%)
Ekerhovd et al. 2003 87	60	IMN (40 mg)	Placebo	III	4	IMN>Placebo (Elective CS all)
Sharma et al. 2005 97	65	GTN (0.5 mg); GTN (0.5 mg)	Misoprostol (0.05 mg); Dinoprost (3 mg)	III	6	GTN<Misoprostol (CS rate 43% versus 48%); GTN<Dinoprost (CS rate 43% versus 52%)
Wolfler et al. 2006 92	110	ISM (40 mg)+Dinoprost (3 mg)	Placebo+Dinoprost (3 mg)	III	48 (max 2 doses/d)	IMN = Placebo (CS rate 33% versus 25%)
Osman et al. 2006 93	200	ISM (40 mg×2), then Dinoprost (1 mg×2)	Dinoprost (2 mg×2), then Dinoprost (1 mg×2)	III	40	IMN<Dinoprost (CS rate 33% versus 31%)
Nunes et al. 2006 98	196	GTN (0.5 mg)+Dinoprost (2 mg)	Placebo+Dinoprost (2 mg)	III	6	GTN>Dinoprost (CS rate 34% versus 32%)

^a Trim: pregnancy trimester.

^b Comparison of efficacy: which drug caused more cervical ripening.

IMN = isosorbide mononitrate; GTN = glyceryl trinitrate; SNP = sodium nitroprusside; CS = cesarean section.

Endothelial NOS is mainly expressed in endothelial cells 11 and platelets 15, while nNOS is expressed in the cerebellum and skeletal muscle 13. Inducible NOS was first cloned from activated mouse macrophages 12, and thereafter, it has been demonstrated in various human cells including macrophages 16.

Nitric oxide as a mediator

Nitric oxide serves as a highly diffusible molecule that can affect cells both directly and indirectly. The direct effects are mediated by the NO molecule itself, while the indirect ones are mediated by reactive nitrogen produced by the interaction of NO with oxygen or superoxide radicals (O₂^‐). At the low concentrations of NO produced through eNOS and nNOS, the direct effects prevail, while at higher concentrations of NO, produced through iNOS, the indirect effects dominate 17.

The formation of cyclic guanosine 3',5'‐monophosphate (cGMP) accounts for many of the direct physiological effects of NO 17. Nitric oxide may also interact with nonheme iron‐containing and zinc‐containing proteins, or form S‐nitrosothiols by nitrosylation 18.

The indirect effects of NO include oxidation, nitrosation, and nitration. Cytokine‐induced NO production mediates cytotoxicity in the target cells of macrophages 19. In a reaction with O₂ (autooxidation) NO forms dinitrogen trioxide (N₂O₃), which can mediate DNA deamination and nitrosylation. By reacting with superoxide NO produces peroxynitrite (ONOO‐), which is a toxic nitrating agent and a powerful oxidant, modifying proteins, lipids, tyrosine, and nucleic acids.

Cervical nitric oxide

All three NOSs (nNOS, iNOS, and eNOS) are present in the human uterine cervix 20–22: neuronal NOS is localized in stroma and epithelial cells 22, iNOS in the epithelial cells and stromal spindle cells 20, and eNOS in vascular endothelium 20.

In animal studies, NO induced cervical ripening 23, and cervical NO release was elevated during labor 24. In a recent bovine study, relative iNOS mRNA levels were found the highest in the beginning of the third trimester and decreased significantly from towards parturition 25. It could be argued that the decrease in iNOS protein expression was due to changes in the composition of the extracellular matrix (ECM) in the cervix, which occur during late pregnancy and parturition 26. In addition, NO induces relaxation of smooth muscle in cervix 27.

Nitric oxide shares with tumor necrosis factor α (TNF‐α) the unique ability to initiate and to block apoptosis, depending on multiple variables that are being elucidated 28. Therefore, NO is both an antiapoptotic and apoptotic substance, which may arrest the cellular turnover and allow the reorganization of the collagen 23, 26. Nitric oxide may act in concert with prostaglandin E2 (PG E2) by inducing local vasodilatation and by increasing vascular permeability and leukocyte infiltration 29. In addition, NO may directly regulate the activity or the production of matrix metalloproteases (MMPs) 1, although Ledingham et al. 30 demonstrated that the secretion of MMP‐2 and MMP‐9 in cervical fibroblasts was not regulated by exogenous NO. If NO modulates MMPs, the action of NO both in uterus and cervix may be mediated partly by MMPs.

During normal pregnancy

In human studies, iNOS has been reported to become stimulated in the cervix during cervical ripening (Figure 1) 20, 21, 31, although not in all studies 32. In addition, data are not uniform as regards the changes in expression of cervical nNOS and eNOS during term vaginal delivery: in some studies no change was seen 20, 21, but in the others, the cervical nNOS expression became stimulated 22.

Figure 1 Immunoreactivity of iNOS and eNOS in cervixes of women with early pregnancy. Biopsy samples from mifepristone‐treated (A–D) and control women (E, F). Strong iNOS staining is seen in the endothelium in A (×200) and in the cervical glands in B (×400), but not in the control woman (E, ×200). Endothelial NOS staining is restricted to vascular endothelial cells in mifepristone‐treated (C, ×200; D, ×400) and control women (F, ×200).

Cervical NO release increases during human pregnancy obtained by various methods 20–22, 33 (Figure 2). Pregnancy‐related increase in cervical NO release has been detected also in animal studies 1, 24, 34–36. Cervical NO release relates to Bishop scores 33 and, moreover, Chiossi et al. found a significant positive correlation between changes in cervical NOx (nitric oxide metabolite) levels and Bishop score modification (P<0.01; r = 0.494), as well as between the modification of NO metabolites concentration and cervical shortening (P<0.01; r = 0.307) 37. Parous women have higher cervical fluid NOx concentrations than nulliparous women 33. A recent study showed that the iNOS gene was downregulated in the cervixes of women after term vaginal labor 38. This may imply that during active labor iNOS‐derived cervical NO release is no longer needed, because the primary task of cervical iNOS is in cervical ripening.

Figure 2 Cervical fluid NOx(nitric oxide metabolite) concentrations and group‐specific medians in nonpregnant and pregnant (pregn.) women (µmol/L) in logarithmic scale.

Figure 3 Correlation between changes in cervical length and changes in cervical NOx(nitric oxide metabolite), 6 hours after dinoprostone application. From reference 37 with permission.

In women the concentration of NOx in vaginal secretions has been reported to be elevated before preterm birth 39, 40. Although the source of this NOx is not known, both infiltrating inflammatory cells and cells in the uterine cervix may be responsible. Because NO can activate MMPs 41, 42 and induce apoptotic cell death 28, 43, the overproduction of NO may be involved in cervical ripening, fragility of membranes, and subsequent premature delivery.

In nonviable early pregnancy

Cervical NO release is increased in women with failed early intrauterine pregnancy, but not in women with tubal pregnancy 44. Increased cervical NO, although a very reactive molecule, is scarcely a primary cause of miscarriage. This is supported by the reduced placental expression of iNOS at the fetomaternal interface in women with spontaneous abortion when compared with that in women with early viable pregnancy 45. The increased cervical NO release in nonviable intrauterine pregnancy could have the following explanations. First, elevated release of NO in the dying fetus, decidua, and fetal membranes has been found in LPS‐induced abortions in animal experiments 46, 47, but no such human data exist so far. However, it appears plausible that remnants of the conceptus could have released NO excessively, which could have leaked, as either NO or NOx, into the cervical canal. This hypothesis is supported by our findings: cervical fluid NOx levels were elevated only in intrauterine miscarriages, not in tubal ones, and missed abortion, with potentially more abundant remnants of conception, was characterized by higher cervical fluid NOx levels than cases of blighted ovum 44. Second, miscarriage is often associated with a local inflammatory reaction in the cervix, and this may result in the stimulation of NO release through PGs or cytokines 1, 7. In addition, many other hormones, such as inhibins 48–50, may be involved in spontaneous abortion and may have secondarily stimulated cervical NO release. Third, increased cervical NO release may be a specific phenomenon in miscarriage, perhaps triggered by a fall in the level of progesterone either locally or in the serum (Figure 5).

In postterm pregnancy

Cervical NO release is deficient in postterm pregnancy (Figure 4) 51. It remains unclear if cervical NO deficiency is a primary phenomenon, and thus a true contributing factor to postterm pregnancy, or whether it is a reflection of relative insufficiency of PGs, cytokines, MMPs, or some other agents which may be primarily involved in cervical ripening 7, 41, 52–58 and which may stimulate NO release. A high progesterone level seems to downregulate the synthesis and release of cervical NO. Recent data indicate that among women with an initial postterm labor there is a significant risk for subsequent postterm deliveries 59, 60. This characteristic seems to be genetically determined 61, and therefore we speculate that ‘postterm genes’ are functionally linked to the genes regulating NO synthases. Such women might benefit from the administration of a vaginal NO donor when induction of labor is needed 62.

Figure 4 Cervical fluid nitric oxide metabolite(NOx) concentrations in women going postterm and in women delivering spontaneously at term (µmol/L). Medians are shown by lines. The detection limit of the assay was 3.8 µmol/L.

Figure 5 A simplified schematic model of the possible role of cervical nitric oxide in human cervical ripening(NOSs = nitric oxide synthases; PG = prostaglandin; MMP = matrix metalloprotease).

Relationships to prostaglandins

Nitric oxide and PGs operate jointly in many cells 63–67. Nitric oxide may either stimulate or inhibit the release of COX‐2 (cyclooxygenase), and likewise PGs may have a stimulatory or inhibitory effect on iNOS, depending on the cell type and/or the presence of cofactors 1, 67–69. Misoprostol as a PG analogue induces NO release in the uterine cervix of pregnant women, and furthermore the response of cervical NO release to PG becomes enhanced from early to late pregnancy 70. Thus, PG analogues such as misoprostol can perhaps initiate a chain reaction in the cervix of pregnant women; the initial NO stimulation caused by misoprostol is followed by endogenous release of PGs triggered by NO. As a result, NO, PG, and COX pathways may have a joint action in human cervical ripening, as schematically shown in Figure 5.

Reducing effect of progesterone

Numerous animal experiments support the view of progesterone having opposing effects on NO release in the endomyometrium and cervix; it upregulates NO release in the former, but downregulates it in the latter 1, 24, 34–36. The data imply a link between progesterone and cervical NO (Figure 5). First, the lower the progesterone level the higher the detection rate of cervical fluid NOx in nonpregnant women 33, 70. In fact, 93% of women in the follicular phase, versus 46% of women in the luteal phase, showed detectable cervical fluid NOx. Circulating NOx levels are also higher during the follicular phase and at the time of ovulation than in the luteal phase 71. Second, cervical NO release was inversely related to circulating progesterone concentrations in early nonviable pregnancy 44. Women with threatened abortion or preterm birth have considerably lower levels of circulating progesterone than women with normal pregnancy 72. Progesterone insufficiency could well have stimulated cervical NO release in nonviable intrauterine pregnancy, which may be needed for ripening of the cervix during the course of miscarriage. Third, cervical NO responded to the antiprogestin mifepristone with a 17‐fold increase in cervical fluid NOx and with increased expression of iNOS in early viable pregnancy 31. Furthermore, mifepristone induced the appearance of iNOS in cervical glands 31 (Figure 1).

The mechanisms behind mifepristone‐induced NO release are not known, but a progesterone receptor‐mediated pathway may be involved 73. Local progesterone withdrawal in the cervix brought about by mifepristone may lead specifically to the stimulation of iNOS. Alternatively, the antiglucocorticoid properties of mifepristone, blocking glucocorticoid receptors, may stimulate iNOS 74. This would fit well with data showing that the levels of glucocorticoid receptor decrease in the human cervix during labor 55. Furthermore, it is possible that mifepristone triggers an influx of inflammatory cells, such as macrophages, neutrophils, and monocytes, which are rich in iNOS. Moreover, mifepristone upregulates various MMPs and/or the secretion of cytokines in cervical cells 1, 55, 75. These mediators may either induce or inhibit iNOS, depending on the availability of various cofactors 74, 75. Furthermore, NO may act in concert with the COX pathway, especially with COX‐2 1, 28, 65, 76 (Figure 5). Nitric oxide in turn may soften the cervix by remodeling the ECM 1 (Figure 5), where cytokine‐induced NOS is centrally involved 1, 20, or by inducing apoptosis of cervical cells 1, 28. Taken as a whole, there is strong evidence 1, 31, 34, 44, 77, 78 that NO in the cervix, and progesterone, are functionally related in pregnancy (Figure 5). In fact, administration of progesterone is used in treatment of threatened abortion or preterm birth 79–83.

Possible use of nitric oxide in Obstetrics and gynecology

In animals, NO donors were found to ripen the cervix 77, 84. In women, NO donors, such as isosorbide mononitrate (IMN) 85–93, sodium nitroprusside (SNP) 94, 95, and glyceryl trinitrate (GTN) 85, 96–98, administered intravaginally or intracervically, ripen cervix during pregnancy (Table I). In general, NO donors appear less effective than PGs, at least in viable pregnancies, but in nonviable early pregnancy there is discrepancy concerning the effectiveness of IMN compared to misoprostol 91, 99, 100 (Table I). Interestingly, although the combination of NO donor and PG, simultaneously applied, did not ripen the cervix more effectively in 6 hours, it shortened the time interval between intervention and active labor 98. Sodium nitroprusside ripens cervix even in nonpregnant women 101.

Glyceryl trinitrate administrated intracervically induced cervical NO release at term 33, and intravaginal administration of GTN together with PG E2 (Dinoprost®) resulted in a shorter induction‐to‐vaginal‐delivery time 98. Isosorbide mononitrate induced COX‐2, but not COX‐1 in human cervix at term 102.

Nitric oxide donors are safe and have no major side effects on fetus or mother 62, 86, 87, 93, 100, 103–106. When compared with misoprostol NO donors were less effective 106, 107 but did not cause the uterine hyperstimulation 86, 108. Thus, NO donors hold promise in cervical ripening in women, although additional data are needed before they can be routinely used in clinics.

Cervical fluid NOx may perhaps be used as a marker of cervical ripeness in clinics. Clinicians could benefit from its assessment both in early and late pregnancy because this test may indicate the readiness of the cervical canal for misoprostol or mifepristone priming 31, 70. In early pregnancy, women with nonviable early pregnancy and ‘high’ cervical fluid NOx could perhaps be treated expectantly 44, because these women might belong to the 50% of such women who will abort spontaneously within 2 weeks 109. If the test result is ‘low’, priming of the cervix with PGs or mifepristone can be considered. These questions should be studied in clinical trials. Likewise, as deficient cervical NO release was related to failed progression of labor 51, the cervical fluid NOx level may perhaps be used to predict the likelihood of successful induction of labor. In early as well as in postterm pregnancy, NO donors may hold a promise, but this question has been studied only in postterm pregnancy so far 62.

No reliable biochemical marker for the identification of women at risk of preterm birth exists. Fetal fibronectin and insulin‐like growth factor‐binding protein‐1 (IGFBP‐1) have been extensively studied as marker candidates, but they have poor predictive values and there are major confounders in sample collection 110–112. Transvaginal ultrasonography is also insufficient in predicting preterm birth, mostly because of the wide variance in predictive value 113, 114. Cervical fluid NOx might prove to be a feasible marker for this condition. An epidemiological study in this connection would need a huge number of women, the risk of spontaneous preterm birth being 2%–3% in Finland. Therefore, we should first study women who report preterm contractions in order to see if this test can differentiate between those women who carry to term and those who deliver preterm. Some preliminary data suggest that this test could be useful in such women 40.

Acknowledgements

Work was financially supported by grants from the Research Funds of Helsinki University Central Hospital, the Emil Aaltonen Foundation, and the Research Foundation of Orion Corporation.