Hypoxia and smooth muscle function: key regulatory events during metabolic stress

Abstract

Hypoxia rapidly reduces force in many smooth muscles and we review recent data that shed light on the mechanisms involved. As many regulated cellular processes are integrated to co-ordinate smooth muscle contractility, the processes responsible for decreased force output with altered metabolism are also likely to be many, acting in concert, rather than the actions of one altered parameter. Nevertheless the aim of this study is to elucidate the hierarchical series of events that contribute to reduced smooth muscle force production during altered metabolism. We conclude that in many phasic smooth muscles the decrease in force can be attributed to impaired electro-mechanical coupling whereby the Ca²⁺ transient is reduced. A direct effect of hypoxia on the Ca²⁺ channel may be of key importance. In tonic vascular smooth muscles K_ATP channels may also play a role in the integrated functional responses to hypoxia. There are also many examples of force being reduced, in tonically activated preparations, without a fall in steady-state Ca²⁺; indeed it usually increases. We examine the roles of altered [ATP], pH, myosin phosphorylation, inorganic phosphate and proteolytic activity on the [Ca²⁺]-force relationship during hypoxia. We find no defining force-inhibitory role for any one factor acting alone, and suggest that force most probably falls as a result of the combination of myriad factors.

Smooth muscle contraction is essential for many homeostatic functions, e.g. the regulation of blood pressure, gastrointestinal motility and the behaviour of the urinogential tract. It is now apparent that many smooth muscles are markedly affected by hypoxia on a rapid time scale. The resulting effect on contraction may be beneficial, e.g. the dilatation of a blood vessel, thereby increasing blood flow to a hypoxic region or preservation of ATP during times of metabolic stress. However, hypoxia can also cause dysfunction, which has been associated with many disease states including atherosclerosis, vasospasm, cerebral haemorrhage and uterine dystocia. The purpose of this review is to draw together the recent data pertaining to hypoxia (or metabolic inhibition with cyanide, which will not generally be distinguished from true hypoxia, apart from when direct effects of low O₂ on ion channels are discussed), and develop mechanistic explanations for its effects. We do not consider hypoxic contraction, the most notable example of which is that which occurs in the lungs, or hypoxic pulmonary vasoconstriction, but confine ourselves to considering how hypoxia reduces force. We begin therefore with a brief overview of smooth muscle force production.

Contraction in smooth muscle

For many of the details and references to this section we recommend the review by Horowitz, Menice, LaPorte & Morgan (1996). The pattern of contractile activity in smooth muscles can be divided into two sorts: phasic, where cycles of contraction and relaxation occur, and tonic, where a maintained level of force exists. Phasic contractions are due to electrical excitation at the surface membrane, leading to action potential generation, and spreading from cell to cell via gap junctions. This depolarization results in Ca²⁺ entry, via L-type Ca²⁺ channels and in the presence of agonists, and IP₃-induced Ca²⁺ release from the sarcoplasmic reticulum (SR) (Somlyo & Somlyo, 1994). Tonic smooth muscles lack spontaneous action potential generation, and force follows a rise in Ca²⁺ following stimulation by neurohormonal agonists. Sustained depolarization can often cause tonic contraction in phasic muscles. The relationship between electrical activity, [Ca²⁺] and force has not been clearly established in intact smooth muscle, as simultaneous measurement of these three parameters has only recently been made (Burdyga & Wray, 1997). When [Ca²⁺]_i rises it combines with calmodulin and activates myosin light chain kinase, which phosphorylates myosin light chains (MLC). There is, then, significant actin-activated myosin ATPase activity for cross-bridge cycling and force production and MLC phosphorylation is thus an important determinant of contractile initiation. During periods of sustained tone MLC phosphorylation levels often decline and force is maintained by slowly cycling cross-bridges (Murphy, 1994). Upon stimulus cessation, [Ca²⁺]_i declines, primarily by plasmalemmal Ca²⁺ extrusion via the Ca²⁺-ATPase (Kosterin, Burdyga, Fomin & Grover, 1994), with possible contributions from the Na⁺-Ca²⁺ exchanger (Suzuki, 1991; McCarron, Walsh & Fay, 1994; Taggart & Wray, 1997b) and sequestration by intracellular organelles such as the SR and mitochondria (see later). As [Ca²⁺]_i declines, MLC₂₀ dephosphorylation is accomplished by a MLC phosphatase and precedes relaxation (Khromov, Somlyo & Somlyo, 1996).

The effect of hypoxia on smooth muscle force

Figure 1 shows typical examples of the effects of hypoxia (simulated with cyanide) in a phasically contracting (Fig. 1A) and tonically activated preparation (Fig. 1B). The effects of hypoxia on force are both rapid and marked, and in this example from the ureter, such effects would reduce the flow of urine from the kidneys to the bladder (Bullock & Wray, 1997). A similar rapid attenuation of force with hypoxia or cyanide occurs in many phasic smooth muscles including uterus (Wray, Duggins, Iles, Nyman & Osman, 1992; Earley & Wray, 1993), stomach (Huang, Chowdhury, Kobayashi & Tomita, 1993) and portal vein (Sward, Josefsson, Lydrup & Hellstrand, 1993) and decline in force such as seen in Fig. 1B can be seen in many tonically activated preparations including uterus (Taggart, Menice, Morgan & Wray, 1997), taenia caeci (Ishida & Paul, 1990), portal vein (Sward et al. 1993) and cerebral arterioles (Taguchi, Heistad, Kitazono & Faraci, 1997). Given the dependence of force on Ca²⁺ it is natural to ask if the contractile inhibition seen in Fig. 1 is due to a reduction in [Ca²⁺].

Figure 1. Simultaneous force (top) and intracellular [Ca²⁺] (bottom) measurements before, during and after cyanide application to guinea-pig ureter.

A, phasic contractions produced by action potentials following electrical stimulation. B, tonic contraction produced by high K⁺ depolarization. The Ca²⁺ measurements were made from the ratio of the fluorescence signals of indo-1 emitted at 400 and 500 nm. Adapted from Bullock & Wray, 1998.

What are the effects of hypoxia on [Ca²⁺] in smooth muscle?

Several studies have shown that changes in the Ca²⁺ transient amplitude or duration cause similar changes in the phasic contractions (Himpens, Lydrup, Hellstrand & Casteels, 1990; Huang et al. 1993; Taggart, Burdyga, Heaton & Wray, 1996). Figure 1 also shows the simultaneously recorded intracellular Ca²⁺ changes associated with the contractile activity during control and hypoxic conditions. Cyanide concomitantly reduces the Ca²⁺ transients along with the phasic contractions; similar finding have been reported for several different smooth muscles (Sward et al. 1993; Taggart et al. 1997; Bullock & Wray, 1997). Thus we can explain the fall of force in Fig. 1A by a reduction in the Ca²⁺ transient. In Fig. 1B, however, contraction is also reduced in hypoxic conditions but under conditions of unchanged steady-state [Ca²⁺]. Similar findings of unchanged or even increased [Ca²⁺] (see Fig. 3) with hypoxia have been reported in many smooth muscles (see Table 1). It is therefore clear that we must discuss two main questions: (1) what are the mechanisms underlying the decreased Ca²⁺ and force transients in phasic tissue? and (2) what are the mechanisms disrupting the Ca²⁺-force relationship in tonically activated smooth muscle? To answer the first question we require a detailed appreciation of the control of smooth muscle membrane excitability, as this underlies the Ca²⁺ transient.

Figure 3. Changes in uterine force and myosin phosphorylation with cyanide.

A, simultaneous force (top) and Ca²⁺ (bottom) records showing the effect of hypoxia on a depolarized rat uterine preparation. Note the increased Ca²⁺ record in hypoxia, but the fall of force. B, MLC phosphorylation measurements during contraction for 10 min produced by high K⁺, in the presence and absence of hypoxia. There was no significant difference in the amount of MLC phosphorylation at any time point; mean data from densiotometric scans of silver stained 2-D gels (see inset). U, unphosphorylated; P, phosphorylated. Figure adapted from Taggart et al. 1997.

Table 1.

Reported changes in smooth muscle steady-state or basal [Ca²⁺]_i during conditions of altered metabolism

Intervention	Smooth muscle	Change steady-state in calcium	Reference
Hypoxia	Coronary	⇄	Paul et al. 1991
	Small pulmonary	↑	Salvaterra & Golfman, 1993; Post et al. 1995; Urena et al. 1996
	Conduit pulmonary	↓	Urena et al. 1996
	Cerebral and mesenteric	⇄	Aalkjaer & Lombard, 1995
	Cerebral	⇄	Gebremedhin et al. 1994
	Taenia coli	⇄	Obara et al. 1997
Metabolic inhibition	Ileum	↑	Hori et al. 1989
	Aorta	↑	Ishida & Honda, 1991
	Stomach	↑/⇄	Huang et al. 1993; Drummond & Fay, 1996
	Portal vein	↑	Sward et al. 1993; Lydrup et al. 1994
	Pulmonary	↑	Yuan, 1995
	Uterus	↑	Taggart et al. 1997
	Ureter	↑/⇄	Bullock & Wray, 1997

Membrane excitability

Many topics (and references) that can only be touched on here can be found in the recent book on smooth muscle excitation, edited by Bolton & Tomita (1996). Resting membrane potential in smooth muscles (approximately -75 to -45 mV) is dependent upon the balance of K⁺, Ca²⁺ and Cl⁻ channel activity, which may change during hypoxia. Many different K⁺ channels have been described in smooth muscle cells including those modulated by ATP, O₂, Ca²⁺ and voltage. The activity and distribution of ion channels depend not just upon the particular smooth muscle preparation, but on developmental and gestational state (Khan, Smith, Morrison & Ashford, 1993; Weir & Archer, 1995; Urena, Franco-Obregon & Lopez-Barneo, 1996). Such differences may well account for much of the diversity found between smooth muscles. Any effect of hypoxia on these channels will affect resting membrane potential and is therefore likely to affect force. A general point to consider is that many ion channels require a basal level of phosphorylation for activation, and hence a fall in [ATP] may reduce current flow (Tewari & Simard, 1994; Hilgemann, 1997), although some Ca²⁺ channels do not appear to be activated by phosphorylation (Klockner & Isenberg, 1985; Ohya, Kitamura & Kuriyama, 1988; Sperelakis, Xiong, Haddad & Musuda, 1994). The changes in metabolites and ions that occur with hypoxia and which may influence ion channel activity are summarized in Table 1. As will become apparent, in many smooth muscles increased outward K⁺ current and decreased inward Ca²⁺ current can be demonstrated with hypoxia or cyanide, along with increased K⁺ efflux and decreased Ca²⁺ entry.

What are the effects of hypoxia on K⁺ channels?

ATP-gated K⁺ channels (K_ATP) are present in smooth muscle and have been proposed as a critical link between cell metabolism and electrical activity, and in particular as forming part of the mechanism underlying the dilatation of coronary and cerebral arterioles during hypoxia (Standen, Quayle, Davies, Brayden, Huang & Nelson, 1989; Daut, Maier-Rudolph, Von Beckerath, Mehrke, Gunther & Goedel-Meinen, 1990; Dart & Standen, 1995; Taguchi et al. 1997). Teleologically this makes sense; this channel is inhibited by intracellular ATP (inhibition constant, k_i, ∼10-100 μM). In hypoxia, the fall in [ATP] and increase in [ADP], along with a decrease in intracellular pH (pH_i) (see Table 2), will be expected to lead to the opening of the channel and hyperpolarization of the smooth muscle cell. This will result in relaxation of the vessel and increased blood flow to the hypoxic region. Recent work has drawn attention to the gating of similar channels by other nucleotide diphosphates (Zhang & Bolton, 1996), but their role in hypoxia remains to be investigated. So, do K_ATP channels open in hypoxia? There is good evidence that they do in the vascular studies quoted above. However, the reliance in several studies on glibenclamide to block these channels and show their involvement in the hypoxic process, has been questioned (Zhang & Bolton, 1996). It is also difficult to control for a direct effect of low oxygen on channels. It is also clear that a small membrane depolarization, rather than the predicted hyperpolarization, occurs with hypoxia or cyanide in several phasic smooth muscles (Huang et al. 1993; Nakayama, Chihara, Clark, Huang, Horiuchi & Tomita, 1997). In these tissues and others (Heaton, Wray & Eisner, 1993) glibenclamide is also unable to reverse the functional effects of hypoxia. Finally as can be seen in Fig. 1A, cyanide reduced the amplitude of the contractions but not their frequency. As K_ATP channel opening should hyperpolarize the membrane, an effect on frequency would be expected. Thus for phasic muscles there is little compelling evidence for K_ATP channel opening, and hyperpolarization being responsible for the fall in Ca²⁺ and force. This may reflect a difference between phasic and tonic smooth muscle responsiveness to hypoxia, but it would also be useful to have more measurements of membrane potential, Ca²⁺, force and K_ATP activity in these microvessels. There is also a discrepancy between the k_i for ATP and the [ATP] found in smooth muscle during hypoxia, i.e. does it fall low enough to cause significant channel opening? (see Table 2).

Table 2.

A summary of the changes in some important metabolites and ions occurring as a consequence of hypoxia

Substance	Control	Hypoxia
ATP	3 mm	1 mm
ADP	100 μm	200 μm
Pi	2 mm	7 mm
PCr	4 mm	< 1 mm
pH	7.2	6.9
Mg2+	100 μm	1 mm
Ca2+ (basal)	50 nm	200 nm

These values are representative of those reported in many studies. P_i, inorganic phosphate. PCr, phosphocreatine.

There are K⁺ (and Cl⁻) channels that are activated by a rise in Ca²⁺ (Carl, Lee & Sanders, 1996; Mironneau, Arnaudeau, Macrez-Lepretre & Boittin, 1996; Large & Wang, 1996). Opening of both these channels can occur spontaneously, due to brief focal releases of Ca²⁺ from the SR, referred to as Ca²⁺ sparks, and can be enhanced by Ca²⁺ entry. The resulting currents are referred to, respectively, as spontaneous transient outward currents (STOCS) associated with hyperpolarization and relaxation (Benham & Bolton, 1986; Nelson et al. 1995) and spontaneous transient inward currents (STICS) associated with depolarization and contraction (Large & Wang, 1996). Clearly if there is a significant alteration in STOC activity due to hypoxia affecting SR Ca²⁺ release (see later) or calcium-activated K⁺ current (I_K(Ca)), due, e.g. to the rise in basal Ca²⁺ commonly observed (see Fig. 3A), then this could explain the fall of force in hypoxia. However, Large & Wang (1996) suggested that there is a heterogeneous distribution of K⁺ and Cl⁻ channels and that when Ca²⁺ rises and activates both I_K(Ca) and calcium-activated Cl⁻ current (I_Cl(Ca)), depolarization is the net result. Little is known about the effect of hypoxia on these currents but in pulmonary vessels the Cl⁻ current is activated and the K⁺ current is inhibited, and depolarization is seen (Greenwood, Helliwell & Large, 1997). The effects of hypoxia on K⁺ channels and hence membrane potential, will influence the gating of L-type Ca²⁺ channels. It is also likely that hypoxia will have direct effects on Ca²⁺ channels, and these effects will be discussed next.

Calcium channels

Before intracellular [Ca²⁺] or Ca²⁺ currents could be measured during hypoxia, there was good evidence that it reduced Ca²⁺ entry into smooth muscle cells (van Breemen, Wuytack & Casteels, 1975; Pearce, Ashwal, Long & Cuevas, 1989). Recent more direct approaches have shown reductions in Ca²⁺ currents and relaxation by cyanide, hypoxia or oxidation (Ohya & Sperelakis, 1989; Okashiro, Tokuno, Fukumitsu, Hayashi & Tomita, 1992; Chiamvimonvat et al. 1995; Urena et al. 1996; Rekalov, Juranek, Malekova & Bauer, 1997). The first two groups of authors consider the effect of hypoxia in reducing Ca²⁺ current to be a direct one on the channel, i.e. low O₂ directly inhibits smooth muscle Ca²⁺ channels. Interestingly the evidence for such effects has followed the demonstration that low oxygen inhibits K⁺ currents in pulmonary myocytes, causing depolarization and Ca²⁺ entry, which could explain hypoxic pulmonary contraction (Yuan, Goldman, Tod, Rubin & Blaustein, 1993; Franco-Obregon & Lopez-Barneo, 1996; Haddad & Jiang, 1997). An oxygen-sensitive Ca²⁺ channel as a mechanism for vasodilatation is appealing, especially as it is rapidly effective, and sensitive to low oxygen in a dose-dependent manner and over the physiological range.

Reductions in Ca²⁺ currents with secondary effects of hypoxia (e.g. changes in ATP, pH, Ca²⁺ and Mg²⁺) have also been reported; for example, in portal vein the Ca²⁺ current was reduced when ATP was decreased (Lorenz & Paul, 1997) and acidification also reduces Ca²⁺ current in many smooth muscle cells (Shmigol, Smith, Taggart, Wray & Eisner, 1995). Basal Ca²⁺ is elevated during hypoxia in many smooth muscles (see Fig. 1B) and elevated [Ca²⁺] can also inhibit L-type Ca²⁺ channels (Ohya et al. 1988). It is, of course, as Mg-ATP that ATP is found within the cell and a decrease in [ATP] will increase [Mg²⁺]. McHugh & Beech (1996) suggested that, during blockade of oxidative metabolism, in basilar artery Mg²⁺ block on the channels is responsible for the reductions in current seen. Interestingly they found no effect of inhibiting glycolysis on the inward current. Thus it differs from another report (Lorenz & Paul, 1997) which indicates a role for glycolysis in supporting Ca²⁺ channel activity and directly altering Ca²⁺ entry. Both studies provide a link between Ca²⁺ channel function and energy production, although which mechanism predominates may well depend upon the particular smooth muscle.

Thus it appears that hypoxia, by both its direct effect on the Ca²⁺ current and via secondary changes in other parameters, e.g. low pH inhibits gap junction conductance (Turin & Warner, 1980), will act to reduce the Ca²⁺ current (see Fig. 2). These effects will also be reinforced by any increases in K⁺ conductance produced by hypoxia, which will favour a decrease in L-type channel activity. The net result of these processes in phasic smooth muscles will be a decrease or abolition of the Ca²⁺ transient in the cell and a decrease in force (Fig. 2). Consistent with this mechanism is the fact that if the decrease in membrane excitability is overcome, e.g. by application of high-K⁺ depolarizing solution or agonist, then contraction, albeit of reduced magnitude, can be initiated in the continued presence of cyanide or hypoxia (Paul, 1989; Heaton et al. 1993). This makes it unlikely that the rapid effects of hypoxia on phasic force are due to critical changes in ATP, pH_i or inorganic phosphate (P_i) around the myofilaments. These effects on the initiation of Ca²⁺ transients, which are dependent upon electrical activity, may be distinct from those changes occurring in basal or steady-state [Ca²⁺]_i which are discussed next.

Figure 2. A scheme to show the possible mechanism underlying the fall of phasic contractions in hypoxia.

Hypoxia reduces O₂, pH and ATP, and increases Ca²⁺, Mg²⁺ and ADP. These changes affect the opening of Ca²⁺ channels, which underly the Ca²⁺ transient.

What causes the rise in calcium?

As shown in Table 1, basal or steady-state Ca²⁺ increases with hypoxia or cyanide in many smooth muscles. There is as yet no clear indication of the source of this. Several intracellular organelles are capable of storing and releasing Ca²⁺ i.e. the SR, nucleus and mitochondria, and may therefore be considered likely candidates. All three organelles are often found close to one another within the cytosol (Broderick & Broderick, 1990; Nixon, Migneri & Somlyo, 1994), offering the possibility that Ca²⁺ movements from each organelle may be important in co-ordinating [Ca²⁺]_i signalling, energy provision and gene expression with contractile activity.

Sarcoplasmic reticulum

The SR is the major intracellular storage site of releasable Ca²⁺ for contractile activation (Somlyo & Somlyo, 1994). The SR is found throughout the cell and forms a continuous tubular network from regions closely apposed to the plasma membrane (peripheral SR) to deeper in the cell (central SR) and the perinuclear space (Nixon et al. 1994; Golivina & Blaustein, 1997; Arnaudeau, Boittin, Macrez, Lavie, Mironneau & Mironneau, 1997). Ca²⁺ is accumulated in the SR by the action of a Ca²⁺-ATPase. Agonist- and IP₃-dependent release of SR Ca²⁺ can directly activate the contractile apparatus in the absence of trans-sarcolemmal Ca²⁺ influx (Somlyo & Somlyo, 1990). Additionally, there is evidence for [Ca²⁺]_i amplification by ryanodine-sensitive Ca²⁺-induced Ca²⁺ release from some (Ganitkevich & Isenberg, 1992; Kamishima & McCarron, 1996; Arnaudeau et al. 1997; Kamishima & McCarron, 1997) but not all investigations (Ganitkevich & Isenberg, 1995; Kamishima & McCarron, 1996; Taggart & Wray, 1997a). Thus Ca²⁺ released from the SR may play a role in smooth muscle contractility by directly activating the myofilaments. The direct influence of hypoxia or metabolic inhibition on SR Ca²⁺ handling in smooth muscle has not been extensively studied. An early study (van Breemen et al. 1975) suggested that iodoacetate or dinitrophenol increased the fractional loss of ⁴⁵Ca²⁺ from intracellular stores. Ishida & Honda (1991) reported that [Ca²⁺]_i and tension responses to either noradrenaline or caffeine were inhibited by cyanide application. Indeed, conditions expected to occur with hypoxia such as decreased [ATP] and pH, and increased Mg²⁺ and [P_i] and altered cellular redox state, may lead to a reduction in SR-releasable Ca²⁺ (Iino, 1991; Tsukioka, Iino & Endo, 1994; Kuemmerle & Makhlouf, 1995; Bootman, Missiaen, Parys, De Smedt & Casteels, 1995; Kosterin, Babich, Shlykov & Rovenets, 1996). Presently, however, there are few data to support a major role for the SR in either the rise in Ca²⁺ or fall in force, during hypoxia.

Mitochondria

The mitochondrion is a high capacity, low affinity Ca²⁺ storage site which several recent studies have suggested may play a role in removing Ca²⁺ from the cytosol upon cell stimulation. In non-muscle cells, oxidizable substrates that increase mitochondrial membrane potential and respiration, and presumably Ca²⁺ uptake, were found to alter the IP₃-induced Ca²⁺ wave profile (Jouaville, Ichas, Holmuhamedov, Camacho & Lechlelter, 1995). In stomach smooth muscle cells, mitochondrial Ca²⁺ increased upon SR Ca²⁺ release by caffeine (Drummond & Fay, 1996). In addition, it has recently been reported that the decay of smooth muscle [Ca²⁺]_i transients, evoked by membrane depolarization, is impaired by a variety of agents interfering with mitochondrial function (Drummond & Fay, 1996; McGeown, Drummond, McCarron & Fay, 1996). Furthermore, inhibition of oxidative phosphorylation slowed the time course of decay of Ca²⁺-activated chloride currents elicited by membrane depolarization, again suggesting that Ca²⁺ signals may be regulated by mitochondrial uptake mechanisms (Greenwood et al. 1997).

Ca²⁺ accumulation by the mitochondria may be a means of limiting cytosolic Ca²⁺ increases and thereby protecting against deleterious protease activity. It may also be a mechanism linking energy supply to demand as Ca²⁺ has been implicated in the regulation of several key mitochondrial enzymes (Broderick & Broderick, 1990). Organellar swelling, and increased Ca²⁺ leakiness, with mitochondrial inhibition has been reported (Somlyo & Somlyo, 1990; Piper, Null & Siegmund, 1994) and may therefore result in several important consequences for the smooth muscle cell including contributing to the rise in steady-state Ca²⁺ both directly, and indirectly by activation of depolarizing currents (Greenwood et al. 1997).

Nucleus

Evidence from a large number of cell types, including smooth muscle, indicates that the nucleus can also act as a Ca²⁺ store. Changes in smooth muscle nuclear Ca²⁺ concentration ([Ca²⁺]_n) following depolarization or agonist stimulation have been reported (Williams, Becker & Fay, 1987; Burnier, Centeno, Burki & Brunner, 1994; Bkaily et al. 1997). It has been suggested that IP₃ receptors are located on the inner surface of the nuclear envelope membrane (Gerasimenko, Gerasimenko, Tepikin & Petersen, 1995) such that any agonist-induced release of stored Ca²⁺ will be directed into the nucleoplasm (Fay, Shlevin, Granger & Taylor, 1979; Himpens, De Smedt & Casteels, 1994; Gerasimenko et al. 1995). Although the SR forms a continuous tubular network with the perinuclear space, Ca²⁺ may not be stored homogeneously throughout this region (Golivina & Blaustein, 1997). Given a dual role in Ca²⁺ homeostasis and gene expression, the nucleus may perform crucial functions in the cellular responses to hypoxia, e.g. regulation of transcription of genes encoding carbohydrate metabolism (Bunn & Poyton, 1996). We can, however, find no study examining the contribution of nuclear Ca²⁺ to changes in cytosolic Ca²⁺ found in hypoxia. We will now discuss the possible causes of the reduction of force in hypoxia when [Ca²⁺] is not reduced.

Altered [Ca²⁺]-force relationship during hypoxia

In a variety of different smooth muscles hypoxia, or metabolic inhibition, has resulted in a significant reduction in force, but no change or increases in [Ca²⁺]_i either at rest or in activated tissues (see Table 1 for references) as illustrated in Figs 1B and 3A. These studies thus led to the conclusion that hypoxia can produce a dissociation of the [Ca²⁺]_i-force relationship in smooth muscle (Taggart et al. 1997) and indirectly supported earlier observations that raising extracellular Ca²⁺ under such conditions did not prevent hypoxic-induced reductions in force (Ishida & Paul, 1990). Therefore we will now review which other factors may be important in causing the reduction of force.

Myosin light chain (MLC) phosphorylation

As previously described the phosphorylation of MLCs is a critical determinant of smooth muscle force production (Somlyo & Somlyo, 1994). As hypoxia or metabolic inhibition result in a reduction of cytosolic [ATP], it is possible that the phosphorylation potential of the tissue declines, especially as MLC₂₀ phosphorylation has been suggested to account for up to 50 % of the ATP cost associated with contraction (Wingard, Paul & Murphy, 1994). However, the findings of recent studies suggest that neither mitochondrial inhibition, nor hypoxia, result in significant changes in levels of MLC₂₀ phosphorylation (Taggart et al. 1997; Obara, Bowman, Ishida & Paul, 1997; and see Fig. 3B). Thus, altered smooth muscle metabolism results in a dissociation of both the [Ca²⁺]_i-force and MLC₂₀-force relationships, as shown in Fig. 3.

Myosin ATPase activity

The above experiments and Fig. 3 illustrate that the fall in force with altered metabolism is unrelated to the phosphorylation of MLC₂₀. As cytosolic [ATP] during hypoxia or as a result of cyanide addition remains > 1 mM (Wray, 1990), in the absence of microdomains of extremely low ATP, it is unlikely that [ATP] surrounding the myofilaments is limiting the myosin ATPase (Arner & Hellstrand, 1985) and hence contraction. Furthermore as creatine kinase is localized to smooth muscle myofilaments it will also help maintain ATP : ADP (Clark, Khuchua, Kuznetsov, Saks & Ventura-Clapier, 1993; Clark & Dillon, 1995). Several studies are thus in agreement that energy supply to the myofilaments is not inhibiting force during hypoxia (Hardin, Wiseman & Kushmerick, 1992; Harrison, Larcombe-McDouall, Earley & Wray, 1994; Clark & Dillon, 1995; Nakayama et al. 1997; Taggart et al. 1997).

Intracellular pH

Significant decreases in intracellular pH occur with hypoxia and metabolic inhibition and often can be correlated with an increase in lactate production and efflux (Vogel, Lilja & Hellstrand, 1983; Wray, 1990; Ishida & Paul, 1990; Harrison et al. 1994; Taggart & Wray, 1995). For example, both in vivo and in vitro uterine hypoxia cause pH_i to fall by around 0.3 pH units and lactate efflux to increase fivefold (Wray, 1990). Hypoxia and ATP depletion both inhibit Na⁺-H⁺ exchange, which would also contribute to a reduction in pH_i (Burns, Homma & Harris, 1991; Noel & Pouyssegur, 1995). Decreases in pH_i with metabolic inhibition were suggested to contribute to impaired mechanical output especially in spontaneously active tissues (Wray, 1990). However, altered pH_i appears not to be the prime factor responsible for force inhibition for several reasons: (a) when cyanide-induced pH_i changes are removed, force is still attenuated (Taggart & Wray, 1995); (b) the tonic force of many smooth muscles is often promoted by intracellular acidification (Taggart et al. 1996); (c) pH-induced alterations in Ca²⁺-activated force of permeabilized smooth muscles cannot explain hypoxic-induced force inhibition (Crichton, Taggart, Wray & Smith, 1993; Wu & Vaughan-Jones, 1994); and (d) hypoxia can result in force inhibition without concomitant reductions in pH_i (Aalkjaer & Lombard, 1995; Obara et al. 1997).

Inorganic phosphate (P_i)

Hypoxic-induced elevations of [P_i] have been reported in both in vivo and in vitro situations (Ishida & Paul, 1990; Harisson et al. 1994). P_i inhibits Ca²⁺-activated force of permeabilized smooth muscle fibres (Itoh, Kanmura & Kuriyama, 1986; Crichton et al. 1993) independently of the level of MLC₂₀ phosphorylation (Gagelman & Guth, 1987; Osterman & Arner, 1995). P_i also accelerates the relaxation from rigor upon photolysis of caged ATP, even in the presence of increased [MgADP], an effect attributable to P_i driving the actomyosin equilibrium to a condition favouring weakly attached or detached cross-bridges (Somlyo & Somlyo, 1990; Nishiye, Somlyo, Torok & Somlyo, 1993). Thus, as suggested by Taggart et al. (1997), the hypoxia-induced increase in [P_i] appears to be a suitable candidate for modulating force in the face of elevated [MgADP] (which would promote the lifespan of attached cross-bridges; Nishiye et al. 1993) and unaltered MLC phosphorylation. Indeed, when the increases in [P_i] with altered metabolism were reduced, the resultant inhibition of force (Ishida & Paul, 1990), and rates of relaxation (Hardin et al. 1992; Taggart & Wray, 1997c), were also less. The absolute (Hardin et al. 1992) reduction in Ca²⁺-activated force of permeabilized smooth muscle with elevated [P_i], however, is less than the inhibition of force in intact arteries during metabolic alteration (e.g. compare uterine smooth muscle studies of Crichton et al. 1993 and Taggart et al. 1997). It is unlikely, therefore, that [P_i] changes alone can account for tonic force attenuation with metabolic inhibition.

Cytoskeleton

An additional possible force-modulatory consequence of hypoxia/metabolic inhibition of smooth muscle, especially as [Ca²⁺]_i remains elevated, is the activation of Ca²⁺-dependent actin-capping proteins and/or proteases (Gailly, Lejuene, Capony & Gillis, 1990) which have been suggested to contribute to attenuated cardiac contractility following ischaemic reperfusion (Gao, Liu, Mellgren & Marban, 1995). Consistent with this notion are the findings that actin-capping agents inhibit tonic smooth muscle contraction independently of [Ca²⁺]_i and MLC₂₀ phosphorylation (Hori, Saito, Shin, Ozaki, Fusetani & Karaki, 1993; Obara & Yabu, 1994). Additionally, the extent of force inhibition is similar to that observed during mitochondrial inhibition (Taggart et al. 1997). Changes in the cytoskeleton will also affect ion channel activity (Hilgemann, 1997) and therefore also influence the response to hypoxia.

Conclusions

There are many potential regulatory sites of the excitation- contraction coupling pathway in smooth muscle during conditions of altered metabolism. Hypoxia and cyanide can inhibit force production either with or without altering the profile of the activating [Ca²⁺]. Hypoxic changes of plasma membrane channel activity contribute to reduced contractility. We propose that in phasic smooth muscles decreased electromechanical coupling and reduction of the Ca²⁺ transient are responsible for the decline in spontaneous force seen in hypoxic conditions. We further speculate that this is predominantly a direct effect on the Ca²⁺ channels, rather than via an increase in K⁺ conductance. In tonically active smooth muscle, a dissociation between Ca²⁺ and force is commonly encountered with hypoxia and cyanide. The decrease in force appears not to be due to limitation of ATP around the myofilaments. We suggest rather, that a combination of the factors, including increased [P_i], acting directly on the myofilaments is the most likely cause. The SR, nucleus and mitochondria may play a wider role in smooth muscle Ca²⁺ homeostasis than previously appreciated, and their involvement in the acute and chronic changes occurring with hypoxia pose many intriguing questions for smooth muscle homeostasis.