The role of oxytocin in parturition
* Correspondence: Dr S. Thornton, Department of Biological Sciences, University of Warwick, Gibbet Hill Road, Coventry CV4 7AL, UK.
Oxytocin and the oxytocin receptor have two important roles in labour. Evidence in all mammalian species suggests that neurohypophysical oxytocin plays a role in the expulsive phase and, although there are less supporting data, a role for oxytocin in the initiation of labour is likely. The initiation of labour may be mediated in women and rhesus monkeys by paracrine rather than endocrine mechanisms. Although initial characterisation of the oxytocin knockout mouse suggested that oxytocin is not important in this species, subsequent investigations have demonstrated that oxytocin is important for the precise timing of the onset of labour. Studies in knockout mice also confirm important interrelationships between oxytocin and prostaglandins. Oxytocin stimulates prostaglandin release in many species, mainly in the decidua/uterine epithelium. The effects of oxytocin are mediated by tissue-specific oxytocin receptor expression, which leads directly to contraction in the myometrium and prostaglandin formation in the decidua. There is a dramatic increase in oxytocin receptor expression in these tissues in late pregnancy and pharmacological inhibition delays delivery, which suggests that, in contrast to oxytocin, the oxytocin receptor is essential for normal labour.
The precise role of oxytocin in human labour remains a controversial and unresolved issue. Historically oxytocin was assumed to be the initiating factor of parturition, although this thesis has now been substantially modified with more knowledge of the labour process. The current paradigm for parturition, based mainly on the models of the sheep and rhesus monkey, encompasses gestation-wide changes in all intrauterine tissues, which ultimately culminate in the expulsion of the fetus1. The observed gestational changes in the intrauterine tissues are driven by a complex interplay between fetal and maternal signals, many of which occur well before the onset of active labour. The concept of parturition can now be viewed, at least in those species that rely solely on the placenta for steroidal synthesis, as a process that evolves over a number of days or weeks. The evidence for oxytocin playing some role in the process of parturition is overwhelming; it is the nature of the involvement, and to some extent the necessity of oxytocin, which is a matter of debate. Perhaps the most controversial issue is whether or not oxytocin initiates labour, or is solely involved in facilitating the active expulsion of the fetus.
The expulsive Phase of Labour
Historically, oxytocin was assumed to be the initiating factor of parturition because clinical administration initiates labour which is indistinguishable from spontaneous labour. After this ‘proof of concept’ experiment, the role of oxytocin was extensively investigated in many animal species. Initial experiments concentrated on determining the plasma levels of oxytocin during labour. At some stage of the labour process, elevated plasma oxytocin is observed in all placental mammals yet studied, including rabbit2, sheep3, cow4, rat5, goat6, pony7, rhesus monkey8, pig9 and human10,11. Furthermore, evolutionarily related peptides are raised in other species, such as mesotocin in the wallaby12, arginine vasotocin in the chicken at oviposition13, and annetocin in the earthworm at egg laying14.
Detailed analysis of oxytocin in the plasma of many species has demonstrated that, during the labour process at least, the secretion is pulsatile and is maximal with fetal expulsion although there are minor differences in the precise pattern of release. In the rabbit, oxytocin peaks at the time of delivery of the first fetus2, whilst in other species such as the sheep3, cow4, and rhesus monkey8, maximal oxytocin is observed at the birth of the singleton. In the pig, a pulse of oxytocin is detectable at the expulsion of each fetus, whereas in the rat, baseline plasma oxytocin is elevated throughout labour with larger pulses coinciding with pup expulsion5. Thus, in animals, the majority of data support significant increases of pulsatile oxytocin release with fetal expulsion.
Many authors have reported wide variations in oxytocin concentrations during human pregnancy and labour10,11,15–25. The discrepancies can be attributed to a number of technical difficulties in measuring plasma oxytocin in humans. Firstly, as outlined above, studies in animals have clearly demonstrated that oxytocin release is pulsatile and as such, requires a rapid sampling strategy to accurately determine oxytocin levels. Most of the aforementioned studies have failed to employ rapid sampling strategies to identify pulsatile release. Secondly, the metabolic clearance rate of oxytocin increases in pregnancy26 probably due to placental formation of oxytocinase, which increases throughout gestation in human plasma27. Effective oxytocinase inhibition has not been used in all of the studies and is imperative for accurate oxytocin determination15. A third consideration is that some antibodies used to measure oxytocin cross react with oxytocin precursors and may yield spuriously high results28. Only a few of the above studies have addressed all of these issues. Fuchs et al.10 demonstrated discrete pulses of oxytocin, which increased in frequency with the progression of labour. Burd et al.15 also demonstrated pulsatile release in the first stage although this was not confirmed in a later publication by the same laboratory. Despite the absence of pulsatile release in the later report, the same authors did report an increase in oxytocin in some, but not all, women in the second and third stages of labour11. Therefore, the human studies support the animal data in that plasma oxytocin is most likely to be increased in association with the expulsive phase of labour.
The stimulus for the expulsion-associated increase in oxytocin is well studied in the rat. During pregnancy, the posterior pituitary content of oxytocin increases by 50% and becomes temporally depleted post delivery. During delivery, the baseline plasma oxytocin is increased with large pulses occurring with the delivery of each pup5. The latter are associated with intense synchronised bursts of electrical activity in the hypothalamus29. The stimulus for the burst in action potentials arises from distension by the pup of the highly innervated cervix, from which afferent nerves stimulate the nucleus tractus solitarii in the supraoptic and paraventricular nuclei of the hypothalamus. This phenomenon, known as the Ferguson reflex30, was perhaps most elegantly demonstrated in a cross circulation study in sheep31. The proximal jugular vein of the first animal was connected to the distal jugular of the second. Cervical distension in the first stimulated a contractile response in the mammary gland of the second demonstrating that stimulation of pituitary oxytocin release occurs during cervical distention. Interestingly, and perhaps to protect the uterus from inappropriate oxytocin secretion, this reflex appears to be suppressed during gestation in the rat via an opioid dependant pathway32. Thus, the data suggest that neurohypophysial oxytocin contributes to the expulsive phase of labour via the Ferguson reflex caused by mechanical distension of the cervix.
The Initiation of Labour
The evidence for oxytocin initiating labour is far more controversial and less well documented. The primary evidence comes from studies in the sheep, rhesus monkey and baboon. In the sheep, there is a progressive increase in uterine activity toward the end of pregnancy, which is significantly attenuated by infusion of a specific oxytocin antagonist33. Studies in late gestation primates describe a similar change in uterine activity prior to the onset of labour. There are progressive reversible and repetitive nocturnal changes from contractures (ill defined 5 minute periods of uterine activity) to contractions (defined 1 minute patterns of activity occurring up to 30 times per hour) in the days preceding labour34–37. As with the sheep, nocturnal contractions appear to be caused by maternal oxytocin8 and are prevented by a selective oxytocin antagonist38,39. In the rhesus monkey, the systemic increases in oxytocin can be caused experimentally by infusion of androstenedione into the maternal circulation40. This initiates preterm delivery with the key associated features of normal parturition such as increased fibronectin, normal cervical ripening, and changes in myometrial activity. The androstenedione in these experiments was administered to mimic the diurnal release of fetal dehydroepiandrosterone sulphate (DHEAS), which increases with gestational age and fetal adrenal weight in the rhesus monkey41. Fetal DHEAS is the primary source of substrate for oestrogen production in the placenta of the non-human and human42 primate. It is therefore a good candidate for the origin of the circadian pattern of oestradiol-mediated oxytocin release and delivery. Subsequent experiments demonstrated that the effect of androstenedione was prevented by aromatase inhibitors and not mimicked by oestrogen infusion43. This illustrates that the observed effects are mediated by local metabolism of oestrogen since hypothalamic formation of oxytocin would otherwise be stimulated by systemic oestrogen43. These experiments place oxytocin as integral to a pathway linking fetal maturation with myometrial activation and suggest that the source of oxytocin in this process, in contrast to active labour, is not the hypothalamus.
Local Formation of Oxytocin
The observation in 1998, that there may be an alternative source of oxytocin in labour was not new. Experiments to demonstrate the necessity of oxytocin in parturition by transection of the neurohypophysial stalk in animals have largely demonstrated variable effects but often prolong the labour process44,45. In women with posterior pituitary dysfunction, delivery occurs normally, although quantitative data are not available on the length of labour46. The reported inconsistencies may be explained by a putative role for paracrine oxytocin release. A role for paracrine oxytocin was first postulated in humans following the discovery of oxytocin mRNA in intrauterine tissues at term47. Oxytocin mRNA is increased in labour in the amnion, chorion and, principally, the decidua. Local oxytocin formation is likely to be oestrogen mediated since oxytocin mRNA and peptide synthesis is increased in explant cultures by oestrogen, an effect prevented by tamoxifen48. Local metabolism is likely since both chorion and decidua degrade oxytocin throughout gestation with a similar efficacy to the placenta49. The extent to which paracrine oxytocin affects the adjacent myometrium is therefore dependent on the balance of peptide production and catabolism. Intrauterine oxytocin has also been identified in the uterine epithelium of the rat50. In this species, uterine production is approximately 70 fold greater than the hypothalamus in the three days leading to parturition. Overall, the data support oestrogen mediated formation of oxytocin principally within the decidua.
Gene Deletion Studies
Despite all of the above evidence, the critical need for oxytocin in the process of parturition has recently been questioned due to normal delivery in oxytocin (−/−) null mice51,52. These mice deliver viable pups, which die postpartum due to inability of the mother to eject milk. Thus, in mice, oxytocin is critical for the milk ejection reflex but not parturition. On the face of it, this result is difficult to reconcile with the wealth of data to the contrary. However, subsequent analysis of the oxytocin null mouse in conjunction with other transgenic mice has revealed far more about the role of oxytocin in mouse parturition than was previously described. Perhaps the most important implication of the oxytocin null mouse is the interrelationship of oxytocin with prostaglandins. In mice, a late gestation rise in systemic prostaglandin F2α (PGF2α) is critical for inducing luteolysis and the subsequent progesterone withdrawal necessary for labour. This is demonstrated by failed labour in the PGF2α receptor53, cytosolic phospholipase A2 (cPLA2)54,55, and cyclooxygenase-1 (COX-1) null mice56,57. All of these manipulations prevent PGF2α formation or activity, luteolysis and labour. Surgical or pharmacological luteolysis in these mice restores progesterone withdrawal and normal labour. Taken with the oxytocin null mouse these experiments illustrate that neither oxytocin nor prostaglandins are indispensable for labour once progesterone withdrawal has been initiated.
The prepartum plasma surge of PGF2α leading to luteolysis in the mouse occurs around day 18.5 of gestation57 but the sequence of events leading to PGF2α synthesis occurs from day 15.5. At this time in the uterine epithelium, prostaglandin F-synthase (PGF-S) protein and COX-1 mRNA increase, whilst 15-hydroxyprostaglandin dehydrogenase (PGDH) protein begins to decrease58. The increase in COX-1 mRNA is followed by a gradual increase in COX-1 protein which peaks on the day of delivery, whilst cPLA2 protein and activity remain unchanged throughout gestation. Therefore, arachidonic acid and PGF-S are increased and PGDH decreased by the day of the PGF2α surge, when a maximal increase in COX-1 provides the PGH2 substrate for PGF-S. The extent to which oxytocin interacts with this pathway was revealed when mating COX-1 null mice with oxytocin null mice. Surprisingly the COX-1/oxytocin null mouse initiates labour on the normal day of delivery by luteolysis, but labour is prolonged in some mice over a number of days. Since COX-1 null mice do not labour due to failed luteolysis, this suggests that oxytocin has luteotrophic actions which are only unmasked in the double knock out genotype. Perhaps more importantly, since only the double knock out experiences labour difficulties, significant compensation must occur between PGF2α and oxytocin in generating the uterotonic phenotype of the single knockouts.
The precise regulation of the opposing luteotrophic and uterotonic actions of oxytocin in the mouse were determined in an elegant set of experiments in the oxytocin null mouse59. Infusion of oxytocin on day 15.5 into the oxytocin null and wild type mouse elicits dose dependent inhibitory and stimulatory effects. At lower concentrations, gestation is prolonged by maintenance of the corpus luteum in both mice. However, at higher concentrations both mice initiate premature labour within 24 hours. Interestingly, the responses in both mice are dose dependent but the oxytocin null mouse is more sensitive to the effects of oxytocin. The initiation of premature labour in both mice is not associated with luteolysis and progesterone withdrawal, and is not prevented by co-administration of indomethacin. The induction of premature labour in this model is therefore independent of prostaglandins and mediated by the uterotonic actions of oxytocin. Detailed analysis of the oxytocin receptor mRNA and protein levels in the ovary and uterus indicate a reciprocal relationship whereby ovarian oxytocin receptor levels decrease on day 19 whilst increasing dramatically after progesterone withdrawal in the uterus. Therefore although the absence of oxytocin does not significantly affect the timing of term labour in mice, abnormal modulation of oxytocin action either by peptide or receptor does influence the onset of labour. Taken together these studies suggest that the role of the oxytocin/oxytocin receptor system in mice is to focus the timing of the onset of labour. The oxytocin null mice also illustrate the importance of the tissue specific control of oxytocin receptor expression in modulating the oxytocin receptor system.
The uterine oxytocin receptor increases at term in all mammalian species tested to date, such as rats60, guinea-pigs61, rabbits62,63, sheep64, mice59, cows65 and humans66. The effect of the dramatic increase is dependent on the site of expression, which occurs in both the myometrium and decidua/uterine epithelium. The direct uterotonic effect of oxytocin in myometrium is well established but the oxytocin receptor effects in the decidua/uterine epithelium are more complex and less completely described. There is little doubt however, that oxytocin causes prostaglandin formation in many tissue types. In explant culture of human chorio–decidua, oxytocin markedly increases the production of PGF2α, PGE2 and leukotrienes in contrast to amnion where PGE2 is the primary prostaglandin product67–69. In the cow and sheep endometrium during pregnancy, oxytocin stimulates PGF2α formation, which increases with gestation and correlates with oxytocin receptor binding65,70. The mechanism of this oxytocin-stimulated release of prostaglandins has been studied at term in the rabbit amnion71. Oxytocin increases cPLA2, COX-1 and COX-2 activity leading to PGE2 formation. Interestingly, in this system cortisol upregulates oxytocin-mediated PGE2 release 100-fold, whilst combined cortisol and forskolin treatment increases the effect 5600-fold, suggesting interaction with other activating mechanisms72. Thus the oxytocin receptor interacts both directly with the myometrium in stimulating uterine contractions, but also indirectly, via prostaglandin formation in other tissues. Taken together, the data suggest that although oxytocin may be dispensable for labour, the oxytocin receptor may not. Although definitive evidence will only come when an oxytocin receptor null mouse is fully characterised, further evidence for the involvement of the oxytocin receptor in labour is provided by the relative success of oxytocin receptor antagonists in arresting contractions in many species.
The oxytocin receptor antagonists are effective in inducing uterine quiescence in rhesus monkeys73, baboons74, wallabys75, guinea-pigs76, rats77, and humans78,79 at term. Atosiban has recently been reported to be effective in the treatment of preterm labour compared to placebo80 or beta-agonists81 in two multi-centre, randomised trials. Although the trial design may have influenced the outcome variables82, atosiban was as effective as ritodrine, but with a reduction in maternal side effects.
Overall, the data support a marked increase in oxytocin receptor expression in uterine tissues at the end of pregnancy. The effect is tissue dependent; in the myometrium, the oxytocin receptor directly causes contraction whereas in other tissues the effect may be mediated by prostaglandins. Experimental laboratory and clinical data strongly support a major role for the oxytocin receptor in labour.
In summary, oxytocin and the oxytocin receptor have two important roles in labour. Evidence in all mammalian species suggests that neurohypophysial oxytocin plays a role in the expulsive phase. Data supporting a role for oxytocin in the initiation of labour is less established but remains likely. This process may be mediated in women and rhesus monkeys by paracrine rather than endocrine mechanisms. Although initial characterisation of the oxytocin knockout mouse suggested that oxytocin is not important in this species, subsequent investigations have demonstrated that oxytocin is important for focussing the timing of labour. Studies in knockout mice also reveal important interrelationships between oxytocin and prostaglandins. Oxytocin stimulates prostaglandin release in many species, mainly in the decidua/uterine epithelium. The effects of oxytocin are therefore mediated by tissue-specific oxytocin receptor expression, which leads to direct contractile effects in myometrium and prostaglandin formation in the decidua. There is a dramatic increase in oxytocin receptor expression in these tissues in late pregnancy and pharmacological inhibition delays delivery, which suggests that, in contrast to oxytocin, the oxytocin receptor may be essential for normal labour.
This work was supported by Wellbeing (ref 434) and the University Hospitals of Coventry and Warwickshire Trust.