Tocolysis: Present and future treatment options

Abstract

In the United States, the generally accepted indication for tocolytic therapy centers on suppression of preterm labor. This may be in the form of preventative therapy with progesterone in women with prior spontaneous preterm birth or as an acute intervention to suppress established uterine contractions associated with cervical change occurring at less than 37 weeks gestation. This article seeks to apply this perspective to tocolytic therapy. Here, we provide a review of current tocolytic options and what the last decade of discovery has revealed about the regulation of myometrial excitability and quiescence. Moving forward, we must incorporate the emerging molecular data that is amassing in order to develop novel and effective tocolytic therapeutic options to prevent preterm labor and spontaneous preterm birth (sPTB).

Introduction

Advancement in medicine is a dynamic process that relies on a critical evaluation of current therapeutic approaches, a commitment to expanding our knowledge base through basic and translational science and a willingness to embark in novel directions based on the marriage of these two concepts. This articles seeks to apply this perspective to tocolytic therapy. Here, we provide a review of current tocolytic options and what the last decade of discovery has revealed about the regulation of myometrial excitability and quiescence. Moving forward, we must incorporate the emerging molecular data that is amassing in order to develop novel and effective tocolytic therapeutic options to prevent preterm labor and spontaneous preterm birth (sPTB).

Indication, therapy, and clinical efficacy

In the United States, the generally accepted indication for tocolytic therapy centers on suppression of preterm labor. This may be in the form of preventative therapy with progesterone in women with prior spontaneous preterm birth or as an acute intervention to suppress established uterine contractions associated with cervical change occurring at less than 37 weeks gestation.

Preventative therapy

With regard to preventative therapy, in 2011 the US Food and Drug Administration (FDA) approved the use of progesterone supplementation (in the formulation of 17-α hydroxyprogesterone caproate, 17 P) during pregnancy to reduce the risk of sPTB in women with a history of prior sPTB. Interestingly, the evidence regarding progesterone-metabolite based supplementation as therapy for sPTB dates back several decades to 1975 when Johnson et al. demonstrated 250 mg of weekly intramuscular 17 P led to fewer preterm births and a lower perinatal mortality rate.¹ Although initial adoption of 17 P was complicated by studies that questioned its efficacy, in the 1980s, both progesterone (vaginal suppositories) and its progestin derivatives (17 P) have gained more widespread use today.^2,3 However, building evidence suggests more stringent patient selection may influence clinical efficacy of progesterone treatment. This notion is reinforced by the results of the OPPTIMUM study, which specifically examined the role of vaginal progesterone. This study represents the largest retrospective double blind study of its kind and although no harmful effects were noted, there was no significant value attributed to vaginal progesterone, particularly in women with a short cervix. The study however did suggest that there was a “general trend” towards benefit and perhaps even significance may have been shown if such a large portion of patients had not been lost to follow up.⁴ We believe this new evidence stresses the need to focus on the possible what chemical, biologic and pharmacologic differences may exist with regard to progesterone and its many derivatives/metabolites that may influence the pathway to sPTB.

Although this recent clinical evidence regarding the efficacy of progesterone and progesterone-based derivatives remains under debate—this therapeutic approach reflects an important shift in our understanding of the potential role this hormone plays in human regulation of labor and how it participates in myometrial quiescence. Progesterone is one the earliest known mediators of embryo implantation and uterine quiescence.⁵ Classically, progesterone production predominates early in the ovary and corpus luteum and is gradually surpassed in gestation by the placenta. Progesterone is known to exert several quiescent-like effects on the myometrium. In particular, it reduces the gap junctions that are formed by the protein connexin 43 (a critical intercellular bridging molecule thought to facilitate tissue-wide myometrial action potentials). Additionally, progesterone has anti-inflammatory properties, which prevent prostaglandin production (a lack of which promotes the state of quiescence).^6,7 In addition, 5-dihydroprogesterone (progesterone metabolite) inhibits human oxytocin receptor (OTR) signaling.⁸ While utilizing progesterone as a treatment is in alignment with the basic science in this field, the conceptual framework that centers on the belief that progesterone supplementation per se can restore physiological homeostasis may be incorrect for some women with a history of PTB. Perhaps we should consider progesterone to be analogous to insulin? We now know not all forms of diabetes result from a “lack” of hormone, but instead reflect a functional disturbance in its signaling. One must consider if the apparent heterogeneity in clinical response to progesterone-based supplementation reflects a heterogeneity in underlying progesterone signaling dysfunction or even progesterone resistance. Indeed, interesting work regarding the potential role of progesterone and progesterone metabolites in regulation of normal human labor may lead to new tocolytic approaches. However, a methodological approach is required in studying women with elevated risk for PTB in order to determine why certain phenotypes have varied responses to specific drug interventions such as vaginal progesterone, 17-OHP, and 17α-hydroxyprogesterone. We believe a precision medicine, individualized approach that screens highly selective phenotype to drug response is required to properly differentiate which genetic/molecular tags are most determinant of positive clinical outcomes. By doing so perhaps we can improve clinical efficacy across all formulations.

Acute intervention

With regard to acute tocolytic therapy and maintenance, there has been a wide variety of drugs that have been employed to suppress uterine contractions (Table). Interestingly, while several distinct drug classes remain in clinical use for this indication—there are currently no Food and Drug Administration (FDA) approved tocolytic drugs available on the market. Ironically, the only FDA approved drug for acute treatment of PTL has been ritodrine (β₂-adrenoceptor agonist) which is currently unavailable in the United States. In addition to beta-sympathomimetics, other known tocolytic drugs include calcium channel blockers (e.g., nifedipine, magnesium sulfate), nitrates (e.g., nitroglycerin), oxytocin receptor antagonists (e.g., atosiban), and prostaglandin inhibitors (e.g., indomethacin). The absence of FDA approval for any of the above drugs reflects the general lack of clinical efficacy. It should be noted, however, that the methodology of many randomized clinical trials (RCT) focusing on these drugs has been limited by lack of sufficient number of patients involved and lack of comparison with placebo.⁹ Although such limitations make it difficult to evaluate the efficacy of classes of tocolytics using meta-analysis, the current evidence suggests that the administration of tocolytic drugs is more effective than placebo/control for delaying delivery for 48 h thus allowing administration of corticosteroids. Interestingly, tocolytic therapy does not result in statistically significant reduction in important clinical outcomes such as neonatal respiratory distress and survival.¹⁰ While an in-depth evaluation of the clinical efficacy of each of these drugs remains outside the scope of this chapter, we have chosen to highlight some of the specific information pertaining to efficacy for each drug class and their representative members.

Table –

Current tocolytic classes.

Tocolytic	Mechanism	Issues
Selective beta 2 (β2) agonists: Ritodrine Terbulatine	– Impairs intracellular cyclic AMP concentration and facilitate myometrial relaxation.	– Lack of long-term benefit – Maternal side effects include tachysystole, chest pain, cardiac arrhythmias, electrolyte disturbances, pulmonary edema, lower blood pressure, tremor.
Nitric oxide (NO): Nitroglycerine (NTG)	– Powerful vasodilator synthesized during an amino acid oxidation process catalyzed by NO synthase. It increases cGMP content by interaction with guanylyl cyclase.	– Intravenous use of NTG required to obtain the desired degree of uterine relaxation is extremely variable in an acute setting and may be accompanied by significant hypotension.
Prostaglandin-synthase or cyclooxygenase (COX) isoforms: Indomethacin	– Nonspecific COX inhibitor – COX-1 and −2 are essential enzymes for converting arachidonic acid to prostaglandins. – Prostaglandins are thought to promote uterine contraction by enhancing myometrial gap junction and increasing intracellular calcium concentration.	– Use should be restricted in duration and limited to pregnancies below 32 weeks because of fetal ductus arteriosus closure risk and decreased urine production responsible for oligohydramnios. – These treatments also have maternal side effects including gastric ulcer or asthma recurrence.
Oxytocin receptor (OTR) antagonists: Atosiban	– Oxytocin mediates its myometrial effects primarily through activation of oxytocin receptors (OTR) which classically Gq couple to activate phospholipase-C beta (PLCb) thereby triggering IP3 and DAG pro-contractile second messenger pathways, the main effect of which seems to involve enhancing intracellular calcium levels. – OTR antagonists block this pathway.	– Mixed affinity between both the OTR and vasopressin (V1a receptors), limited parenteral route of administration, and variable bioavailability. – Possible fetal concerns if used prior to 28 weeks. – Lack of long-term efficacy.
Voltage-Gated Calcium channel (VGCC) blockers: Nifedipine	– Interferes with calcium ion transfer through the myometrial cell membrane to decrease intracellular free calcium concentration and promote myometrial relaxation.	– As a peripheral vasodilator it may cause symptoms such as nausea, flushing, headache, dizziness, and palpitation, hypotension. – Lack of long-term efficacy.
Magnesium sulfate:	– Exact mechanism has not been completely delineated. – It likely competes with calcium at the level of the plasma membrane voltage-gated channels, hyperpolarizes the plasma membrane, inhibits myosin light-chain kinase activity by competing with intracellular calcium.	– Maternal side effects include diaphoresis, flushing, headache, magnesium toxicity. – Lack of long-term efficacy.

Selective beta 2 (β2) agonists.

These drugs have been used in clinical practice for preterm labor since the 1980s. These drugs impair intracellular cyclic AMP concentration and facilitate myometrial relaxation.^11,12 Randomized controlled studies reviewing placebo vs. β agonists for inhibiting preterm labor reported that these agents were more efficient than placebo for delaying preterm birth for 2 days.¹³ In addition, there is no evidence that maintenance therapy with these tocolytic preparations (given after the acute mitigation of preterm contractions) prolong gestational age.¹⁴ Unfortunately, the lack of benefit for long-term tocolysis (effect restricted to 7 days) and perinatal mortality is compounded by reports of an enhanced maternal morbidity rate.^12,15,16 This latter issue has been a consistent concern even with selective β₂ adrenergic receptor agonists. For example, although ritodrine is the only drug approved by the US FDA for the treatment of acute preterm labor, it was discontinued in US markets by the manufacturer in 1992 due to rare, potentially fatal maternal side effects such as chest pain, cardiac arrhythmias, and pulmonary edema.¹⁷ This led to many practitioners using terbutaline, another β agonist, in its place, and has led the FDA to issue warnings in 2011 and 2016 regarding the use of prolonged terbutaline treatment for preterm labor because of continued reports of serious maternal side effects.¹⁸

Beyond maternal risk are also reports that prolonged in utero exposure to β-adrenergic receptor agonists can result in deleterious effects in offspring.¹⁹ These data suggest that the use of terbutaline should be limited to short-term use (<48 h) as a tocolytic or for the acute management of uterine tachysystole to avoid significant maternal side effects (such as tachycardia, dyspnea, hypokalemia, hyperglycemia, and chest pain).^12,15,16,20 In summary, these complications resemble other deleterious fatal reports of chronic beta-agonists use in asthma.²¹ Furthermore, it is not too surprising to find that this particular drug class loses its effectiveness after 48 h of continual/chronic treatment, since receptor desensitization (and internalization) is a well-described phenomenon.²² While not examined specifically in pregnancy—there is a strong possibility that since these drugs are systemically administered the process of β agonist-induced desensitization is likely occurring ubiquitously after 48 h. Given the well-established relationship between a markedly reduced responsiveness of the beta-adrenergic receptor system and the development of heart failure [where levels of β-adrenergic receptors are reduced and activity of desensitizing pathways (e.g., GRK2) are upregulated] it is biologically plausible that the loss of beta-adrenergic receptor responsiveness following maintenance B-agonist therapy may be the culprit leading to impairment of maternal cardiac function.²²

Nitric oxide (NO).

This molecule is a powerful vasodilator synthesized during an amino acid oxidation process catalyzed by NO synthase. It is present in myometrium and increases cGMP content by interaction with guanylyl cyclase. There is a direct relationship between NO production and magnitude of uterine relaxation.^11,23 This pathway can be exogenously induced by systemic administration of intravenous nitroglycerine (NTG) to relax smooth muscle by releasing nitric oxide, increasing intracellular cGMP levels, and causing prompt cervico-uterine relaxation. Indeed, NTG has been used in obstetrics for multiple purposes.²⁴ Several recent case reports and series have described the effective use of NTG for acute uterine relaxation in the antenatal, intrapartum, and postpartum periods. Therapeutic indications for NTG range from facilitating external cephalic version, the Ex Utero Intrapartum Treatment Procedure (EXIT), difficult vaginal or cesarean section delivery, and manual exploration of the uterus due to its potency as a tocolytic. However, intravenous use of NTG required to obtain the desired degree of uterine relaxation is extremely variable in an acute setting and may be accompanied by significant hypotension.²⁵ Other NTG preparations (transdermal) have been used in preterm labor but only in small series. Its use was associated with a better tocolytic effect (than placebo) significantly delaying delivery for 2 days. Its effect was similar to ritodrine.^26–28 A meta-analysis of randomized trials that compared glyceryl trinitrate by any route to placebo (three trials), beta adrenergic receptor agonists (nine trials), and nifedipine (one trial) found the use of glyceryl trinitrate did not significantly prolong pregnancy by >48 h, reduce preterm birth, or result in improved neonatal outcomes compared with any other tested tocolytic.²⁹

Prostaglandin-synthase or cyclooxygenase (COX) isoforms.

COX-1 and −2 are essential enzymes for converting arachidonic acid to prostaglandins. Prostaglandins are thought to promote uterine contraction by enhancing myometrial gap junction and increasing intracellular calcium concentration.^11,12,16,28 Indomethacin, a nonspecific COX inhibitor, has been reported in studies and in a recent meta-analysis to be an efficient tocolytic drug compared to placebo, significantly delaying preterm delivery.^15,30 It can be administrated rectally or orally. Its use should be restricted in duration and limited to pregnancies below 32 weeks because of fetal ductus arteriosus closure risk and decreased urine production responsible for oligohydramnios.^16,20,27,31 These treatments also have maternal side effects including gastric ulcer or asthma recurrence.^16,20,27 COX-2 inhibitors such as nimesulide or rofecoxib have been studied in animal but not in humans and are currently not recommended for preventing preterm labor in clinical practice.³² In conclusion, indomethacin is an efficient tocolytic drug that is indicated for short-term effects limited to the second trimester of pregnancy.

Oxytocin receptor antagonists.

Oxytocin is an endogenous nona-peptide secreted by the posterior pituitary gland in a pulsatile fashion. Although oxytocin has neuro-physiological effects found to be important in psycho-emotional and social contexts, its uterine-contracting properties were first discovered in cats over a century ago by the British pharmacologist Sir Henry Dale in 1906. This hormone mediates its myometrial effects primarily through activation of oxytocin receptors (OTR) which classically Gq couple to activate phospholipase-C beta (PLCb) thereby triggering IP3 and DAG pro-contractile second messenger pathways, the main effect of which seems to involve enhancing intracellular calcium levels.³³ Consistent with a role in preparing the uterine smooth muscle bed to become more excitable at term, myometrial OXT receptor concentration is found to be relatively low at 13–17 weeks gestation but increases twelvefold by 37–41 weeks.³⁴ Given that OXT receptor expression is maximally expressed after the initiation of labor irrespective of gestational age (i.e., preterm labor also induces maximal OTR expression) antagonizing this receptor makes logical sense as a semi-selective blocker of myometrial excitability. The only drug in this class that currently has been used in the United States is atosiban. Atosiban is a competitive, reversible antagonist of the OTR that has been shown to reduce intracytoplasmic calcium release and downregulate prostaglandin synthesis.^11,26 A first multi-center randomized trial comparing atosiban and ritodrine demonstrated a similar tocolytic effects but fewer adverse effects with atosiban.^20,28 A meta-analysis published in 2005 reported no benefit in terms of preterm delivery rate and neonatal outcome in 1695 patients treated either by atosiban or placebo.³⁵ The same systematic review and meta-analysis found that atosiban was as effective as beta agonists for preventing preterm birth within 48 h of initiating treatment.³⁶ However, atosiban was associated with a significantly lower risk of maternal side effects than beta agonists. However, a subsequent trial (APOSTEL III) randomly assigned 510 women with preterm labor to receive oral nifedipine or intravenous atosiban for 48 h and found the drugs resulted in similar composite perinatal outcomes and a similar proportion of pregnancies that did not deliver within 48 h.³⁷ The US FDA declined to approve the use of atosiban for tocolysis because of concerns about the drug’s safety when used in pregnancies less than 28 weeks of gestation. However, European studies have failed to confirm this concern, and atosiban remains widely used in clinical practice there because of its low side effects profile.^16,20 A German meta-analysis based on 6 randomized trials, among them 3 double blinded studies, confirmed a similar tocolytic action between atosiban and β2 adrenergic receptor agonists with significantly lower incidence of adverse effects. Moreover, they found lower overall cost associated with atosiban when considering length of hospital stay and enhanced morbidity-associated costs when compared to continuous fenoterol administration.³⁸ There are some disadvantages associated with atosiban such as mixed affinity between both the OTR and vasopressin (V1a receptors), limited parenteral route of administration, and variable bio-availability.

Recent efforts to improve selective antagonism for the OTR resulted in discovering another peptide-based antagonist (barusiban) that displays higher affinity for human OTR and lower affinity for the V1a receptor. In in vitro studies employing isolated human myometrium, barusiban exhibited higher potency and a longer duration of action than atosiban. Investigators have also shown that barusiban inhibits oxytocin-induced myometrial contractions of both preterm and term human myometrium.³⁹ Despite encouraging results indicating superiority of barusiban (compared to atosiban) in in vivo pre-term labor monkey and non-human primate models,^40–42 in a subsequent multicenter trial in Europe barusiban proved no more effective than placebo in preventing preterm labor.⁴³ While new efforts are now directed toward discovery of non-peptide OTR antagonists for the management of preterm labor, most candidate therapies remain at an exploratory phase with very few clinical studies successfully completed.

One concern with chronic antagonism of the OTR (as for maintenance tocolysis) requires reconciling the lack of long-term efficacy (>48 h) with this drug class as a consequence of GPCR-mediated sensitization. Indeed, the opposite is known to be true since under chronic agonist activation the myometrial OTR has been shown to desensitize and even down-regulate expression.^44–46 These findings corroborate the clinical experience where prolonged IV induction utilizing oxytocin infusions have been shown to increase the risk of uterine atony.⁴⁷ While no study specifically has addressed the possibility that chronic competitive OTR antagonism can lead to over-sensitization this phenomenon is known to occur with other GPCRs. The other possibility why OTR antagonists have not proven as effective as investigators have hoped is that while this system certainly can magnify myometrial contractility—it may not be a critical mediator of human labor and other pro-inflammatory triggers of labor may bypass the effect of OTR blockade through alternative (complementary) signaling pathways.

Voltage-gated calcium channel (VGCC) blockers.

This class of antagonists interferes with calcium ion transfer through the myometrial cell membrane to decrease intracellular free calcium concentration and promote myometrial relaxation.^26–28 Among the drug classes that target this receptor family, nifedipine is the most commonly used drug for preterm labor inhibition at a daily dose of 30 mg daily. Randomized controlled trials report a similar tocolytic effect for nifedipine compared with β-agonists, however calcium channel blockers showed statistical benefit over β-agonists with respect to prolongation of pregnancy, minimizing serious neonatal morbidities, and maternal adverse outcome.⁴⁸ A meta-analysis of randomized trials of calcium channel blockers compared with placebo/no treatment for pre-term labor, use of calcium channel blocker reduced the risk of delivery within 48 h, but there was no statistical reduction in this outcome compared with other classes of tocolytics. A Cochrane Database review meta-analysis published in 2003 reported a decreased number of deliveries within 7 days following treatment and also, a reduced incidence of neonatal respiratory distress syndrome.⁴⁹ Given the established importance of calcium in contractility and the central role this particular class of channels play in generating myometrial action potentials, the strategy of targeting VGCCs makes rational sense. However, these channels are also critical in vascular smooth muscle (VSM) with several in vitro studies showing nifedipine potency on VSM exceeds that demonstrated on myometrial cells. Indeed, since these drugs were developed primarily as anti-hypertensives (and not as tocolytics) it is not surprising that clinically their use becomes limited by maternal cardiovascular effects.

Although each of the above drug classes have been shown to be effective at inhibiting uterine smooth muscle contractility (hours to days)—their ability to maintain tocolysis over longer periods of time (weeks) has not been clinically demonstrable. This draws on an interesting disconnect between how these drugs are often assessed for efficacy in experimental studies and how their effectiveness is ultimately judged in clinical practice. At present, the most adopted method for assessing a drugs capacity as a tocolytic remains either ex vivo studies that utilize strips of human myometrium in contractility experiments or employing in vivo studies in animal models that are not fully representative of human biology. With regard to the former, while this screening process is important method for gauging acute drug potency—it offers very little information regarding the impact effective concentrations play in vivo or whether their effectiveness is sustainable. Part of the way forward should reconcile this disparity in models that closely resemble human biology. It is time to develop and encourage experimental platforms that assess human tocolytic capacity and move away from models of maintenance tocolysis that are mechanistically divergent (i.e., rodent).

Mechanisms contributing to normal and premature labor

Successful parturition represents one of the most dynamic processes a human body can undertake. While only a component of the myriad of physiological adaptations that is part of this process, labor is certainly a critical player. Clearly a major contributor to this “normal” physiological adaptive process is the uterus, yet it is important to remember that there are several other players that all contribute to the successful coordination of uterine contractions and subsequent delivery. Although tocolytics have traditionally focused on the uterine smooth muscle cell, perhaps better tocolytic efficacy in the future should employ multiple synergistic strategies that overlap with and between all the systems known to be contributory to labor.

To date, there is a great deal of exciting evidence that is emerging from basic science studies that suggests there are at least 5 different “players” involved in normal human labor. These include the myometrium, the cervix, the placenta, the fetal adrenals, and the maternal neuroendocrine axis. Perhaps the one molecule that overlaps the most with respect to all of these distinct contributors is corticotropin releasing hormone (CRH) and its related peptides. There are two sources of CRH-related peptides—maternal and placental. Under normal conditions, maternal CRH-related peptides levels are thought to remain relatively static but circulating levels of placental-derived CRH have been shown to increase exponentially as human pregnancy progresses towards term.⁵⁰ Indeed, peak levels of CRH have been tightly associated with the onset of labor with placental CRH activation being driven by the fetal HPA axis in a forward-feedback loop that coincides with fetal maturation.⁵¹ In both the hypothalamus and placenta, CRH is a product of the same gene located on the long arm of chromosome eight.⁵² However, while CRH is negatively regulated by glucocorticoids in the hypothalamus, Robinson et al.⁵³ found that glucocorticoid positively regulates CRH in the term placenta. The positive regulation of placental CRH by glucocorticoid establishes a feed-forward loop between the fetus and placenta that appears to drive CRH production during pregnancy. These two phenomenons have led some to postulate that placental-derived CRH serves as a ‘placental clock’ to determine the length of pregnancy. Interestingly, studies have also shown that prostaglandin administration for labor induction increases CRH levels in maternal plasma via both a reduction in the levels of circulating CRH binding protein and an enhanced secretion of the peptide thereby initiating a positive feedback effect of prostaglandin on maternal CRH bioactivation that seems to facilitate the development of active labor.⁵⁴ While these relationships are clearly important, the mechanism by which CRH is linked to the initiation of labor and rhythmic contractions is still incompletely understood.

What is known is that human myometrial smooth muscle cells express functional CRH receptors whose activity and expression appear to be dynamically related to pregnancy and labor. However, even though expression patterns in the myometrium appear coupled to the pattern of placental CRH secretion⁵⁵ subsequent studies primarily based on in vitro cellular and molecular models suggest the primary role of this system is to maintain uterine quiescence and prevent premature myometrial contractions.⁵⁶ Throughout most of pregnancy the myometrium expresses CRH type 1 receptors. Activation leads to Gsα stimulation of adenylate cyclase and cAMP, which classically promotes myometrial relaxation.⁵⁷ However, at the end of pregnancy an alternative splice variant of the CRH receptor is expressed and the level of expression of Gsα subunits declines which theoretically would shift the myometrium out of a quiescent phenotype.⁵⁸

Another distinctive feature that implicates CRH in the pathogenesis of pre-term labor is its relationship as an adaptive response to environmental stress. Indeed, stress eliciting changes in CRH as a neuroendocrine response is well established and is increasingly being recognized as an important risk factor for preterm birth.⁵¹ The response to stress also occurs at the placental level, as indicated by in vitro studies of human placental cells that show CRH is released in a dose-dependent manner in response to major biological effectors of stress (including cortisol, catecholamines, oxytocin, angiotension II, and IL-1). In vivo studies have also found significant correlations between maternal psychosocial stress and the levels of CRH, ACTH, and cortisol in maternal plasma. Several studies have related early increases in maternal plasma CRH levels to pre-term birth. Hobel et al. conducted serial assessments of CRH levels over the course of gestation and found that women delivering preterm had significantly elevated CRH levels compared with those in women delivering at term, as well as a significantly accelerated rate of increase in CRH levels over the course of their gestation. In addition, they found that maternal psychosocial stress levels at mid-gestation significantly predicted the magnitude of increase in maternal CRH levels between mid-gestation and later time points of gestation.⁵⁹ These studies implicate that a relationship between maternal psychological stress and prematurity may be mediated by prematurely increased levels of placental CRH. However, instead of placental CRH activation being driven by the fetal HPA axis—it is the maternal HPA axis (along with the maternal sympathetic-adreno-medullary system) that may hijack placental CRH expression.⁵¹ Under this scenario, maternal stress (occurring in advance of fetal maturity) can lead to elevated cortisol and epinephrine which can prematurely accelerate placental CRH expression to induce preterm labor.⁵⁹

Recent work by Wang and Rosen has further deepened our understanding of CRH revealing unique relationships to glucocorticoid, NF-κB, and progesterone signaling. This group found that in term human placental cells, glucocorticoid drives the non-canonical NF-κB pathway to bind to a newly identified κB response element in the CRH promoter.⁶⁰ This group has subsequently shown the mechanism occurs in an epigenetic fashion that involves CREB binding protein (CBP) and the histone deacetylase HDAC1 at the CRH promoter.⁶¹ Given the prior observation that progesterone-mediated activation of progesterone receptor-A (PR-A) inhibits NF-κB (by preventing p100 processing to p52)⁶² this group has revealed the intriguing possibility that the non-canonical NF-κB pathway may be a mechanistic linchpin between CRH, progesterone, inflammation and infection in the pathogenesis of pre-term labor/birth.

What remains missing are the subsequent signals and effector targets that get triggered in the uterus and cervix. In other words, while the placenta may serve as a “brain” that initiates the trigger for labor—the process also requires functional recruitment of the myometrium and cervix. How that is done exactly and what the effector molecules are in the myometrium and cervix remains unknown. However, studies from the last decade have advanced our understanding about many of the molecular changes that accompany myometrial excitability and have revealed the importance of myometrial electro-mechanical coupling as a necessary component of labor.

Since the myometrium is devoid of specialized neural input or classical pacemaking cells, electro-mechanical coupling (EMC) in the uterus is fundamentally related to the various myometrium ion channels and their differential influence on developing rhythmic waves of contractility. Here, myometrial cells take on a myogenic phenotype capable of generating action potentials that trigger widespread changes in intracellular calcium to enhance cycling of actin-myosin mediated development of smooth muscle tension. It appears the uterus accomplishes this in part through a complicated interplay between potassium, calcium, sodium, and chloride ion channels to regulate membrane excitability and action potential generation. A novel description of how myometrium recruits “excitability” is offered by Smith et al. who paints an apt analogy between the spreading of excitability that occurs in myometrial cell membrane potential at term with the emergent behavior of a crowd that takes on a chant at a sports game.⁶³ Under this theory, the uterus relies on building up an “excitability” of sufficient magnitude that it not only reaches a threshold but also then spreads through the tissue (optimized by intercellular connections called gap junctions). However, in totality the wide-spread development of coordinated contractions relies on several factors including removal of the quiescent “brakes” that allows underlying pro-contractile mechanisms to emerge.

In addition to hormonal suppression of inter-myocyte communication through these gap junctions,⁶⁴ uterine quiescence is also maintained by several families of potassium (K⁺) channels. While several different types of K⁺ channels are expressed in the uterus, most contribute outward potassium flux to hyperpolarize the myometrial cell in order to maintain a relatively lower resting membrane potential. Throughout most of gestation, K⁺ channel efflux is the primary ionic current responsible for maintaining the resting membrane potential—however, as gestation progresses the myometrium becomes increasingly depolarized (and closer to action potential thresholds).⁶⁵ The prevailing belief is that a loss of K⁺ channel currents underlies this observation. Indeed, several studies have revealed the feedback between some of these K⁺ channels and the hormonal changes that may lead up to and accompany labor. In particular; the large conductance BKCa channel (responsive to changes in adenylyl cyclase, NO, and relaxin) the SK3 channel (responsive to estrogen signaling mediators in the mouse), the voltage-gated potassium channel Kv4.3 (undergo 17β-estradiol mediated inhibition of current) may play a role in shaping human myometrial membrane potential.^66–68 For an excellent in-depth review on this topic please refer to Brainard et al.⁶⁹

More recently, elegant work by McCloskey et al.⁷⁰ have implicated another K channel—the inwardly rectifying potassium channel KIR7.1—which when antagonized or knocked-down results in this shift toward a more depolarized state and enhanced excitability. Similarly, Parkington et al.⁷¹ demonstrated a role for the voltage-gated K channel hERG proteins in human myometrium in late pregnancy. Consistent with a quiescent role on successful action potential generation, this group found levels of the hERG b-inhibitory subunit are elevated in laboring myometrium while a decrease in hERG activity resulted in more “robust” action potential duration, and enhanced contractility.⁷¹ This group also demonstrated that obese women have reduced expression of the hERG inhibitory b-subunit suggesting that this channel may be responsible for reports that link high BMI to ineffective labor and increased risk for cesarean delivery.⁷² Whether it is driving resting membrane potential channels, or voltage-dependent modulation of the action potential,⁶⁸ or initiating changes in intracellular calcium release through BKca and SK3,^74,75 or influencing intracellular ATP concentration through K_ATP channels⁷³ potassium channels are clearly critical to the suppression or development of excito-mechanical coupling.^73–75

One consequence of a functional withdrawal of K⁺ channel mediated quiescence is that this phenomenon allows for other previously dormant ion channels to take on a larger role to more positively influence myometrial excito-mechanical coupling. This is consistent with the landmark observation by Parkington et al.,⁶⁵ which utilized micro-electrode recordings to reveal membrane potential of human myometrium shifts from baseline values of approximately −70 mV (at 29 weeks of gestation) to more positive (depolarized) resting potentials of approximately −55 mV (at term/during labor). The shift toward a more depolarized membrane potential is thought to be important because it closely approximates the voltage thresholds required to activate other voltage-gated channels. Indeed, there several channel families that seem to facilitate excitability and AP generation in the pregnant human myometrium and include (at a minimum) the voltage-gated calcium channels (VGCC), Ca-activated chloride channels (CaCC), and TRP channels. How these ion channels individually affect the action potential is an important consideration—since several studies have suggested the characteristics of the action potentials can change based the hormonal/gestational context whereby transitory single AP spikes become more robust, last longer, and feature plateaus.⁷⁶ This finding correlates with other studies that have linked greater contractility to bursts of multiple spikes (plateau type).⁷⁷ Understanding which ion channels contribute to the stability of plateau formation is an area of active investigation and its development may be due to interplay between several factors. To date, the greatest amount of work among these ion channels has focused on the VGCC, of which two sub-types (L-type and T-type) have been shown to be present in human myometrium.⁷ Classically, these two sub-types are distinguishable based on differential thresholds of activation where the T-types are voltage activated at lower (classically at about −55 mV) thresholds than their L-type counterparts (classically activate at approximately −40 mV). In humans,⁷⁸ L-Type VGCCs clearly play a critical role in generating an action potential by allowing for robust calcium influx (depolarizing current). Parkington et al.⁶⁵ definitively demonstrate the in vitro ability of nifedipine (10–8 M) to abolish VGCC-mediated increases in intracellular calcium and contraction in pregnant human uterine smooth muscle at term. Blanks et al.⁷⁹ showed that Ca_v3.1 is the major T-type VGCC-mediated current found in term human myometrium, but only 55% of human myocytes at term actually express this channel. This was in contrast to 100% of term myocytes that displayed a functional L-type current. This group also demonstrated that functional antagonism of the T-type channels resulted in significant reductions in contractile frequency but no significant effect on maximum contractile force. However, their findings are similar to the observed relationship VGCC T-type channels contribute to the rhythmic beating of the heart.⁸⁰ This observation also mirrors the earlier work of Young et al. who reported differential drug response among currents recorded from human (pregnant) dissociated myocytes. They found that nifedipine (10⁻⁶ M) blocked L-type currents (no effect on T-type currents) whereas magnesium (8 × 10⁻³ M) effectively reduced T-type currents but did not reduce L-type currents.⁷ Interestingly, while studies clearly support the L-type VGCC is the major carrier of the calcium current responsible for action potentials, there is some suggestion L-type channel drug sensitivity may become diminished during human labor. Longo et al.⁸¹ in a comparison of drug response between laboring and non-laboring myometrium found reduced responsiveness to nifedipine in samples taken from “laboring” parturients—independent of gestational age. While other investigators that have shown enhanced potency to nifedipine in non-pregnant myometrial strips obtained following hysterectomy.⁸² The underlying mechanism for this observation has not been explored, but may be another reason why the effectiveness of nifedipine differs so greatly between clinical experience and experimental paradigms.

Another family of ion channels that have a functional role in pregnant myometrium are calcium activated chloride channels (CaCC). These channels are voltage-gated (classically activate at −50 mV to −30 mV) but this voltage sensitivity is enhanced by elevations in intracellular calcium. For example, in response to an increase in intracellular calcium, CaCC activation leads to an outward current that results in greater depolarization of the cell membrane and increased excitability of the myometrium. CaCC currents have been demonstrated in several smooth muscle tissues and have been shown to play an important role in myogenicity leading to spontaneous contractile activity in the gut and urethra.^83,84 In the uterus, CaCC effects have been described for some time⁸⁵ although most published studies have examined their role in rodent models.^86–88 With regard to exito-mechanical coupling, Jones et al. demonstrated that CaCC can be activated by calcium entry through VGCC channels and contribute to both spontaneous and oxytocin-stimulated contractions in the rat myometrium. Furthermore, others have shown enhanced CaCC flux following oxytocin stimulation of the rat uterus⁸⁹ illustrating the potential for their effects to be associated with labor-induced hormonal fluctuations. Under this context, CaCC channel mediated depolarization may assist in further VGCC’s (L-type) activation to enhance AP generation. In addition, some evidence suggests CaCC channel activation may play a role in stabilizing plateau potentials. Although studies on laboring human uterine smooth muscle cells reveal plateau potentials of −30 mV over a time period of approximately 45–100 seconds⁶⁵ the mechanisms that underpin the stability of the plateau potential are not well understood. In rat myometrium Young et al. discovered that chloride channel blockade reduced impulse, duration of spiking activity, and number of AP spikes generated in each contraction.⁹⁰ Given the importance of plateau potentials (mentioned earlier) for successful uterine contraction if this finding is shared in human myometrium—CaCC inhibition may represent a means to disrupt robust AP plateau formation and serve as a novel tocolytic approach. Our lab has identified specific members of a sub family of CaCC (anoctamin 1 and 2) in late gestation human uterine smooth muscle⁹¹ and have illustrated blockade of these channels can greatly diminish murine uterine contractility. While CaCC functional expression was found to be in 33% of the rat myometrium, a recent publication examining the transcriptomes of late gestation human myometrium suggested potentially higher expression of the human CaCC anoctamin 1. However promising, more studies are required to determine if this ion channel family holds any merit as a potential tocolytic and whether a suitably selective antagonist can be found (a problem that has plagued CaCC studies in the past).

Given their capacity to generate fast inward currents to promote AP in other tissues, voltage-gated sodium (Na⁺) channels may also play a role in human uterine excito-mechanical coupling. Indeed, expression of two voltage-gated Na⁺ channels (Na_v2.1 and Na_v2.3) have been described in human and mouse myometrium.⁹² While studies in rats have shown an increase in Na⁺ current as gestation progresses (suggesting that a high Na⁺ current close to labor may contribute to excito-mechanical coupling) the distinct roles of Na_v2.1 and Na_v2.3 channels in human labor remains unclear since these channels have not been functionally characterized. Of some interest however are older clinical studies that showed a significant association between epidural use (higher dose local anesthetic regimens then are typically in use today) and an increased rate of arrest of labor/c-section rates.^93,94 While it seems plausible that local anesthetic-mediated voltage-gated Na⁺ channel blockade may have been related to this clinical phenomenon, in vitro studies required several fold higher concentrations of drug than clinically relevant to show contractile inhibition. However, the validity of Na⁺ channel blockade as a means of in vivo tocolysis in labor has not been adequately studied and may be a potential adjunct in the future.

Like VGCCs, nonspecific cation channels have the capacity to allow for calcium flux into the myocyte to enhance excitability. Given that certain ion channels in this category are activated by factors known to accompany labor (i.e., swelling and stretch) selective antagonism of these channels may represent novel targets for laboring myometrium. Indeed, a recent publication by Ying et al. examined the role of the nonspecific cation transient receptor potential vanilloid 4 (TRPV4) channel in rodent models of preterm labor and in human myometrial cells in vitro. This group demonstrated that TRPV4 channel gene and protein expression increases with gestation, and becomes proportionally more functional at term by exhibiting higher membrane expression relative to the whole cell.⁹⁵ This transition seems to be tied to b-arrestin signaling (primary mechanism that mediates down-regulation and sensitivity to b-agonists), which may translate poorly as a chronic tocolytic for humans. However, while the relevance of TRPV4 to human labor has yet to be determined, it may prove to be an important tocolytic target.

Beyond the contributions of ion channels that generate membrane excitability, there are other important pathways in myocytes that enhance mechanical coupling. Chief among these relates to the importance of internal sensitization to calcium within a myocyte.⁹⁶ and the mechanisms in place that enhance transmission of excitability through a three dimensional tissue. The cellular mechanisms that govern calcium sensitivity are numerous, tend to be ubiquitous among many smooth muscle beds (not just the myometrium), and culminate their effects by modulation of actin-myosin dynamics. Ultimately, uterine contraction must lead to coupling of actin and myosin, which largely depends on the phosphorylation state of myosin. This either occurs by activation of myosin light chain kinase (MLCK) primarily activated by calcium-calmodulin after an increase in intracellular calcium levels or by inhibition of MLC phosphatase.

One related area of investigation that is of particular interest is the role of the family of phospholipase C (PLC) as a bridge between membrane receptors and subsequent downstream signaling pathways that sensitize the myocyte to calcium and enhance contractility through modulation of actin-myosin bridging. Dr. Sanborn provides an eloquent in-depth review on this topic.⁹⁷ The PLC family of proteins is of unique interest because several isoforms are expressed in human myometrium and as a target they are not susceptible to classical desensitization pathways (seen with b-agonist and OXTR antagonists). In particular, phospholipase CB3 (PLCB3) has been shown to be a substrate for multiple pro-contractile and relaxant signaling molecules/kinases and serves as a point of convergence between several signaling pathways in the myometrium.⁹⁸ Other investigators have demonstrated a role for PLC activation on calcium oscillations—a process shown to be important for contraction frequency in human myometrium.⁹⁹ In addition to stimulating calcium release through sarcoplasmic stores, various studies have explored phospholipase-coupled pathways as mechanisms of regulation of gap junctions. Activation of both the phosphatidylinositol (PI)-specific phospholipase C (PLC) pathway and the phospholipase Delta pathways are known to activate Protein kinase C leading to inhibition of gap junctions in mammalian myometrium.^100,101 Given these three levels of influence related to uterine contractility, PLC inhibition represents an important target for future tocolytic development.

The mechanisms that enhance transmission of excitability and mechanical coupling at the level of the 3 dimensional myometrium are clearly of fundamental importance to the development of coordinated, tissue-wide effectual contractions. Beyond the importance of plateau type APs and gap junction proteins already mentioned—the enhancement of contractility afforded by stretch is another area of active investigation. Stretch has been shown to independently enhance expression of oxytocin receptors, gap junctions, and pro-contractile mechano-receptors and has been proposed as the primary means by which synchronization and coordination of uterine myocytes contractile activity occurs.¹⁰² The combination of these effects result in tissue wide electrical activity that can be recorded as an electromyogram. Indeed, while the capacity to directly record both electrical and mechanical activity of the uterus in humans dates back over 30 years,^103,104 only more recently has it been successfully applied as a diagnostic modality with greater predictive capability than the presently utilized and clinically accepted tests. In particular, Garfield et al. showed superiority of uterine EMG when compared to other “predictive” technologies such as intrauterine pressure catheters and external uterine monitors.^105,106 More recently, Luckovnik et al. were the first to demonstrate EMG propagation velocity is a highly sensitive diagnostic tool for true pre-term labor,¹⁰⁷ suggesting this technology platform can provide direct objective measurements to assist early detection of PTL. It has also been postulated that electrical inhibition of contracting myometrium can be achieved by applying relatively long electrical pulses with a plurality of wave patterns, and constant current output.¹⁰⁸ Such device invention could suggest an entirely novel approach to inhibiting of labor using non-excitatory electrical stimulation.

Future considerations for improving tocolysis

Clearly, the volume and complexity of mechanisms underlying labor is daunting. While our understanding of the molecular underpinnings involved in initiating and promoting labor have grown in leaps and bounds, the armamentarium for tocolysis remains sadly antiquated. We have presented throughout the chapter several conceptual failings of current tocolytic strategies—many of which are tied to issues of drug desensitization or off-target side effects. Fortunately, we believe there are several ways to advance treatment for tocolysis.

To begin, we must recognize that many of the current drugs used for tocolysis were not developed primarily for that use. In particular, nifedipine and terbutaline represent drugs with higher potency in other tissues—so it should be of no surprise that monotherapy employing these drugs is prone to failure due to the drugs’ side effects. Perhaps one of the primary steps we need to make is to borrow a page from cardiovascular pharmacology and begin testing drug combinations with proven mechanistic synergy in studies employing human myometrium. In this way, individual drug concentrations may be reduced with the hope that it will reduce the tachyphylaxis that occurs during monotherapy with higher drug concentration requirements. Some papers have examined combinational therapy—but none were based on regimens shown to possess synergy. Given the persistent issues many systemic pharmacologic drugs possess with regard to off-site effects, we need to advance our concept of therapy to include other regional delivery modalities. In particular, can uterine EMG platforms evolve to serve an analogous capacity as modern day Automated Implantable Cardioverter-Defibrillators (AICD)? These devices not only monitor electrical activity as a diagnostic tool but also deliver therapeutic current. Given the great complexity of ion channel splice variants, receptor polymorphisms, enzymatic isoforms, post-translational protein modifications, additional insights into combinational approaches may best be afforded by in silico mathematical models of labor. This field has blossomed over the past decade with considerable progress being made by several investigators in the field of myometrial modeling.^109–111 We feel this approach eventually will liberate us from reductionist methods of mechanistic assessment that often does not allow for biological assessment of relevance. Along these lines, we should acquiesce that most animal models do not recapitulate human reproductive biology and in many cases have led the scientific community into assumptions that have proven false. While abandoning animal models is irrational, at the very least emphasis should be paid to cross-validate mechanistic findings in animal models in some meaningful way with human tissues or models that reflect human biological systems. New advances in CRISPR genetic editing platforms and human engineered tissue chips may help in this effort.¹¹² Finally, for the field of informatics to yield important insights into plausible variations in PTL pathogenesis greater scrutiny needs be paid to phenotype stratification. In particular, a precision medicine approach that stratifies patient groups by objective measures (including drug response) is required to distill (from the current heterogeneity of drivers that can lead to PTL) markers that can guide patient-centered therapeutic approaches.