Kv7 and Kv11 channels in myometrial regulation.

Ion channels play a key role in defining myometrial contractility. Modulation of ion channel populations is proposed to underpin gestational changes in uterine contractility associated with the transition from uterine quiescence to active labour. Of the myriad ion channels present in the uterus, this article will focus upon potassium channels encoded by the KCNQ genes and ether-à-go-go-related (ERG) genes. Voltage-gated potassium channels encoded by KCNQ and ERG (termed Kv7 and Kv11, respectively) are accepted as major determinants of neuronal excitability and the duration of the cardiac action potential. However, there is now growing appreciation that these ion channels have a major functional impact in vascular and non-vascular smooth muscle. Moreover, Kv7 channels may be potential therapeutic targets for the treatment of preterm labour.

New Findings

What is the topic of this review?

The topic of this review is how ion channels contribute to the physiology of the uterus, with particular focus on novel potassium channels.

What advances does it highlight?

Two families of potassium channels, encoded by KCNQ and KCNH genes, have been identified as important players in the control of myometrial contraction and may represent interesting novel therapeutic targets.

Introduction

The uterus is an incredibly complex organ that displays considerable physiological plasticity, cellular remodelling and robustness during pregnancy. However, perturbations of the precise orchestrations that regulate the contractile state of the uterus can have negative consequences for the mother and fetus. Early activation of contractility that, for example, results in spontaneous preterm birth can be associated with a high risk of neonatal morbidity and mortality, as well as lifelong ill health and socio-economic consequences. Conversely, delayed delivery or dysfunctional labour due to weak or poorly co-ordinated contractions can lead to fetal hypoxia, clinical intervention and a greater risk of postpartum haemorrhage.

If there are to be improvements in clinical management and development of novel therapeutic strategies for complicated pregnancies then a better understanding of the mechanisms that determine normal and pathophysiological uterine contractility is essential. There are many factors that dictate gestational changes in uterine contractility, such as alterations in the steroid hormone environment, inflammation and uterine stretch that is exerted by the growing feto-placental unit. The impact of these stimuli is a fine tuning of the mechanisms controlling uterine smooth muscle contractility at the cellular level, including gap junctions, G-protein-coupled receptors, calcium regulatory proteins and contractile filament interactions, but ultimately, all converge upon a background electrical rhythm generated by the activity of ion channels, much like a good concerto relies on the precise contributions from individual instruments in an orchestra. Understanding the contribution of these individual instruments to the uterine symphony is very much a work in progress, but recent studies have identified KCNQ and KCNH-encoded K+ channels as new and functionally powerful elements that hold promise as major regulatory mechanisms and potential therapeutic targets for the treatment of intrapartum complications.

The purpose of this article is to provide a brief overview of this field of research, with particular focus on two new pieces of the puzzle rather than a comprehensive summary of the many factors implicated in uterine physiology. The reader is recommended to consult a number of more comprehensive reviews for more depth in specific areas (e.g. Taggart & Tribe, 2001; Wray, 2007).

Inherent excitability

Uterine smooth muscle exhibits spontaneous contractility that can be augmented by receptor agonists, such as oxytocin (Wray, 2007). Spontaneous contractions are intimately related to the generation of slow waves, upon which action potentials are superimposed (Casteels & Kuriyama, 2010; Kuriyama & Suzuki, 2004; Bengtsson et al. 2004; Parkington et al. 2002b). As gestation proceeds towards labour, the resting membrane potential of the uterine smooth muscle becomes progressively more depolarized (Kuriyama & Suzuki, 2004; Bengtsson et al. 2004; Parkington et al. 2002b), and this is associated with an increase in the force and frequency of spontaneous contractions. The initiator of the spontaneous activity, however, remains to be identified unequivocally. In the gastrointestinal tract, peristalsis is driven by multibranched, non-contractile cells that express the c-kit receptor (termed interstitial cells of Cajal or ICC). Similar ICC-like cells have been observed in rodent and human myometrial tissue (Ciontea et al. 1965; Duquette et al. 2002; Allix et al. 2006). Moreover, pharmacological blockade of the c-kit receptor with imantanib or deletion of this gene does affect the frequency of contractions in the myometrium of mice. However, the effects are subtle, and imantanib has negligible effect in human myometrium, suggesting that the impact of ICC-like cells is not as clearly defined in the uterus as it is in the gastrointestinal tract. Irrespective of the genesis of the spontaneous contractility, the operation of specific ion channels maintains contractile activity, and elucidation of the nature of the respective depolarizing (excitatory) and hyperpolarizing (inhibitory) channels remains a key challenge for uterine physiologists.

Excitatory pathways

In its simplest form, contraction of myometrium, like that of all smooth muscle, is mediated by a rise in [Ca2+] leading to activation of myosin light chain kinase, and the subsequent phosphorylation of myosin light chain at serine 19 allows actin–myosin interaction (see Wray, 2007; Taggart & Tribe, 2001). The rise in [Ca2+]i is mediated by an interplay between increased Ca2+ influx through plasmalemmal channels, Ca2+ release from the sarcoplasmic reticulum and Ca2+ sequestration processes. However, the major precipitatory mechanism is the opening of L-type voltage-dependent Ca2+ channels (VDCCs), as evidenced by the marked effect of dihydropyridines, such as nifedipine, on myometrial contraction (Sperelakis et al. 2007; Wray, 2007). There is evidence that T-type VDCCs may also have some role in maintaining spontaneous contractile activity (Taggart & Tribe, 2001). In addition to VDCCs, voltage-gated sodium channels have been recorded from isolated myometrial smooth muscle (Sperelakis et al. 2007; Seda et al. 2010), and the density of these currents increases in late pregnancy. However, little is known about the molecular nature of the sodium channels and how they contribute to functional activity.

Membrane potential is key

If the influx of Ca2+ through VDCCs is a major determinant of myometrial contractility then logically the influence of membrane potential is central to this mechanism (see Tong et al. 1992 for a computational model). An important question, therefore, is what are the principal mechanisms that propel the membrane potential towards voltages that enhance VDCC open probability and, conversely, which specific ion channels ensure repolarization to more negative membrane potential and closure of VDCCs? In most smooth muscle cells, Ca2+-activated Cl− channels (CACCs) provide the major depolarizing impetus, because smooth muscle cells actively accumulate Cl− ions (Chipperfield & Harper, 2008). As a consequence, the activation of CACCs leads to Cl− ion efflux sufficient to produce membrane depolarization (Leblanc et al. 2000) and, subsequently, to further activation of VDCCs. In relationship to uterine smooth muscle, Cl− currents due to CACC activation have been recorded in rat myometrial cells, and inhibitors of this channel, such as niflumic acid, attenuate myometrial contractility (Jones et al. 2009), although these agents are known to have pluripotent effects (Greenwood & Leblanc, 2005). Preliminary data also show that transcripts for TMEM16A (Caputo et al. 1980; Schroeder et al. 2008; Yang et al. 2007), the putative molecular correlate of CACCs, are present in mouse and human myometrium (AJ Davis, RM Tribe & IA Greenwood, unpublished observations) as well as in vascular smooth muscle cells (Davis et al. 2005). It is worth noting that in the gastrointestinal tract, TMEM16A is expressed by the ICCs, not the smooth muscle cells (Hwang et al. 2008). A second mechanism to produce membrane depolarization is to activate non-selective cation channels, and various members of the ORAI/STIM and TRP gene family that encode for proteins associated with store-operated and receptor-operated calcium entry (see Wang et al. 2011 for overview) are present in rodent and human myometrium (Dalrymple et al. 2000; Yang et al. 2008; Babich et al. 2013). Non-selective cation channels also have a degree of inherent Ca2+ permeability that can potentially contribute to the general rise in [Ca2+] and contraction.

Potassium channels: nature∼s brakes

Co-ordinated contraction of the myometrium relies on hyperpolarizing influences to limit the extent of membrane depolarization (see Fig. 1) and subsequent contraction. Consequently, potassium channels define the magnitude, duration and periodicity of uterine electrical events. Myometrium expresses a number of genes encoding for different potassium channels, including calcium-activated (BKCa; Anwer et al. 2008; Pérez et al. 1999), SKCa (Brown et al. 2005; Pierce et al. 2003), acid-sensitive twin-pore channel TREK-1 (Bai et al. 1993; Buxton et al. 2007), inwardly rectifying ROMK1 (Lundgren et al. 2005) and various voltage-dependent K+ channels, especially members of the Kv4 family (Song et al. 1996; Smith et al. 1999; Greenwood et al. 2003). In terms of functional impact, inhibitors of BKCa, such as paxilline or iberiotoxin, or blockers of SKCa, such as apamin, have negligible effect on rodent or human myometrial contractility (Aaronson et al. 2006; Brown et al. 2005; Smith et al. 1999; Noble et al. 2008). In comparison, the non-selective Kv inhibitor, 4-aminopyridine, enhances contractility (Aaronson et al. 2006; Smith et al. 1999), and the Kv4.2/4.3 blocker, phrixotoxin-2, induces contractions in non-pregnant, but not pregnant, rat myometrium (Smith et al. 1999). Set against this background, two novel types of Kv channel encoded by members of the KCNQ and KCNH gene families have been identified that appear to act as key regulators of uterine contractility and offer new therapeutic targets.

Figure 1

Schematic representation of the functional role of potassium channels in uterine smooth muscle contraction

Left-hand panel shows that open K+ channels result in membrane hyperpolarization that indirectly limits the opening of voltage-dependent calcium channels shown in blue. This results in a less contracted smooth muscle. In the right-hand panel, the potassium channels are non-functional due to blockade, loss-of-function mutations or trafficking defects. This leads to membrane depolariziation, and the open probability of the calcium channels increases. The concomitant influx of calcium contributes to smooth muscle contraction.

KCNQ- and ERG-encoded potassium channels

Ether-à-go-go-related genes or ERGs (ERG1, 2 and 3) are members of the KCNH gene family. All genes encode for voltage-dependent K+ channels (Kv11.1–11.3) that assemble as a tetramer to generate a Kv channel with unique voltage-dependent properties due to an over-riding c-type inactivation (Smith et al. 2002). ERG1 (KCNH2) exists mainly as two splice variants (ERG1a and 1b; London et al. 1976) and is expressed predominantly in cardiac myocytes, where it contributes to the late repolarizing phase of the cardiac action potentials; mutations to the underlying gene underpin a major component of hereditary arrhythmias. ERG2 and ERG3 are located in neurones and contribute to the suppression of membrane excitability (Selyanko et al. 2007). The KCNQ gene family contains five members (KCNQ1–5), and each gene encodes a Kv channel (Kv7.1–7.5, respectively) with low activation threshold (V0.5 ≈ −35 mV) and minimal inactivation (Haitin & Attali, 2007). Kv7 channels also exist as tetramers, with Kv7.1 assembling homomerically. Kv7 activity is modulated by local phosphoinositide levels (Hernandez et al. 2009; Haitin & Attali, 2007), calmodulin and association with auxiliary proteins encoded by the KCNE gene family (McCrossan & Abbott, 2009). KCNQ genes have a well-defined pattern of expression, with KCNQ1 located predominantly in the heart as well as the inner ear; KCNQ2, 3 and 5 are mainly neuronal where they comprise the so-called M-channel in neurones (Brown & Adams, 1984; Selyanko et al. 2008); and KCNQ4 is restricted to the inner ear and auditory nerves (Kharkovets et al. 2013). Mutations to KCNQ genes underlie hereditary arrhythmias (KCNQ1), epilepsy (KCNQ2/3) and deafness (KCNQ4).

KCNQ- and ERG-encoded potassium channels and smooth muscle

The impact of ERG- and KCNQ-encoded K+ channels on cardiac and neuronal physiology was established over 10 years ago. However, both gene families have been ascribed new roles of late through their identification as key players in the regulation of smooth muscle activity.

Expression of KCNQ in smooth muscle was first identified in rat stomach by Ohya et al. (2010a). Since then, KCNQ transcripts have been identified in mouse, rat and human blood vessels (e.g. Ohya et al. 2002a; Yeung et al. 2008; Makie et al. 2011; Ng et al. 2008), as well as in the gastrointestinal tract, urinary tract and airways (see Jepps et al. 2011 for comprehensive overview). KCNQ channel blockers, such as linopirdine or XE991, evoke contractions in the quiescent smooth muscles, such as arteries, or enhance spontaneous contractility (e.g. Yeung & Greenwood, 2002, Jepps et al. 2009, Rode et al. 1993; Ipavec et al. 2008; Anderson et al. 1999). Serendipitously, there are also activators of KCNQ-encoded channels, such as the novel anticonvulsant retigabine, that relax smooth muscles (see Jepps et al. 2011).

Expression of ERG has been determined in the gastrointestinal tract (Akbarali et al. 2006; Ohya et al. 2010a; Farrelley et al. 2010; Parr et al. 2003), mouse portal vein (Ohya et al. 2011b) and bovine epididymis (Mewe et al. 2004), where the smooth muscles exhibit phasic contractions. In these tissues, ERG channel blockers, such as dofetilide or E4031, augment spontaneous contractions tremendously and often cause individual events to fuse into a tonic contraction.

In terms of the myometrium, all KCNQ isoforms are expressed in non-pregnant mice, with KCNQ1 being dominant, and the transcript level for all isoforms remains stable throughout the oestrus cycle (McCallum et al. 1997). In pregnant mice, the expression of all KCNQ genes drops dramatically at early stages of gestation but recovers to robust levels by late stages (McCallum et al. 1997), suggesting that their main role is to regulate contractility at the end of pregnancy rather than to induce quiescence in early pregnancy. Transcripts for all KCNQ genes except for KCNQ5 have also been detected in myometrium from women undergoing Caesarean section at term (McCallum et al. 1997). Of the three ERG genes, only ERG1 is expressed in mouse (Greenwood et al. 2003) and human myometrium (R. M. Tribe & I. A. Greenwood, unpublished observations). In the BALB/c mouse myometrium, both splice variants of ERG1 were detected, with the longer C-terminal ‘a’ isoform dominant (Greenwood et al. 2003), and the expression of this gene did not vary throughout mouse gestation or following parturition (Greenwood et al. 2003). All members of the KCNE gene family whose expression products alter the membrane insertion capabilities and biophysical properties of KCNQ- and ERG-encoded channels (McCrossan & Abbott, 2009) are also expressed in virgin and pregnant mouse myometrium (Greenwood et al. 2003; McCallum et al. 1997). Moreover, transcripts for KCNE2 and KCNE4 increased markedly in mouse myometrium throughout pregnancy (Greenwood et al. 2003; McCallum et al. 1997), an observation that was mirrored at the protein level (Greenwood et al. 2003).

A functional role for both KCNQ- and ERG-encoded K+ channels has been determined in isometric tension and single-cell electrophysiological studies. Linopirdine and XE991 are specific inhibitors of all KCNQ channel isoforms that increase contractile activity in either non-pregnant or pregnant mouse myometrium, mainly through an increase in the frequency of contractions (McCallum et al. 1997, 1997). These agents have similar effects on term non-labouring samples of human myometrium (McCallum et al. 1997). In line with a working hypothesis that increased K+ channel activity limits membrane depolarization and suppresses voltage-dependent Ca2+ influx, the KCNQ-encoded K+ channel activators, flupirtine and retigabine, produce rapid inhibition of spontaneous and oxytocin-driven contractility in mouse and human myometrium (McCallum et al. 1997, 1997). This tocolytic activity is more marked in myometrium from late pregnant mice compared with early pregnant mice (McCallum et al. 1997).

Specific blockers of ERG-encoded channels, such as dofetilide or E4031, have a more striking effect on spontaneous contractility of mouse myometrium than KCNQ channel blockers (mean integral of tension increases by ∼300%, in comparison to ∼50% seen with XE991) that is usually manifest as an increase in the amplitude and duration of individual contractions (Greenwood et al. 2003). Inhibitors of ERG-encoded channels also have a dramatic effect on oxytocin-mediated contractions in mouse myometrium, with tissues often generating sustained contractions of considerable magnitude (Greenwood et al. 2003). Activators of ERG-encoded K+ channels (NS1643 or PD118057) also attenuate contractions in mouse uterus. However, in contrast to KCNQ channel modulators, the effects of channel blockers and activators is lost in the final stages of mouse pregnancy (Greenwood et al. 2003). This is associated with an inability to record dofetilide-sensitive K+ currents in isolated myometrial smooth muscle cells that are present in cells from non-pregnant animals (Greenwood et al. 2003). Modulators of ERG channels become effective again in tissues harvested only 3 h after delivery (Greenwood et al. 2003). Currently, the effects of ERG inhibitors in human myometrial tissues have only been studied in samples obtained from non-labouring woman at term (end of pregnancy), so it is not yet confirmed whether a similar molecular mechanism exists in humans. However, this redundancy in the functional impact of ERG-encoded channels in late mouse pregnancy represents a potential pivot point in the switch from a quiescent system to an excitable system able to generate considerable rhythmic contraction in order to facilitate fetal delivery.

Conclusion

The uterus remains an enigma. Despite much research, there is still much to ascertain with regard to the mechanisms that drive the switch from quiescence to contractile activity preceding labour, and little is known about the stimulus for induction of preterm labour. Furthermore, existing therapies are far from being the ideal tocolytics. The recent findings that KCNQ- and (ERG) KCNH-encoded K+ channels have a major impact on myometrial contractility and that the functional impact of KCNH-encoded channels diminishes in an animal model of term pregnancy represent progression towards answering some of these questions.