Mitochondrial ATP-sensitive K⁺ channels, protectors of the heart

ATP-sensitive K⁺ (K_ATP) channels were first identified in the sarcolemma of cardiac myocytes as inwardly rectifying K⁺ channels that were inhibited by intracellular ATP (Noma, 1983). It was proposed that K_ATP channels would have a cardioprotective effect during ischaemia by shortening action potential duration and thereby decreasing Ca²⁺ influx into myocytes. Then it was found that K⁺ channel opener compounds (KCOs), that are known to activate sarcolemmal K_ATP channels, exert cardioprotective effects under ischaemia–reperfusion (Grover et al. 1989; Grover & Garlid, 2000). However, several groups of investigators found that KCOs exerted cardioprotection at concentrations below those causing action potential shortening (Yao & Gross, 1994; Grover et al. 1995a,b; Garlid et al. 1997), indicating that KCOs may have targets other than sarcolemmal K_ATP channels. Alternative targets for KCOs include mitochondrial K_ATP (mitoK_ATP) channels (Grover & Garlid, 2000).

The mitochondrial inner membrane is polarized by ∼180 mV with the matrix side negative (ΔΨm) due to a H⁺ gradient generated by respiratory enzyme complexes (Fig. 1) (Saraste, 1999). The energy stored in the form of ΔΨm is utilized to make ATP from ADP by ATP synthetase. The inner mitochondrial membrane possesses different ion channels (uniports) through which cations such as K⁺ and Ca²⁺ flow into the matrix under the normal electrochemical gradient and diminish ΔΨm (Bernardi, 1999). Inoue et al. (1991) were the first to identify K⁺-selective channels that were inhibited by ATP and glybenclamide in the inner membrane of liver mitochondria by using patch-clamp methods (Inoue et al. 1991). Dahlem et al. also recently found similar channels in the inner mitochondrial membrane of Jurkat cells by using patch-clamp methods (Dahlem et al. 2004). The existence of mitoK_ATP channels in the heart was confirmed by different techniques such as reconstitution of mitochondrial membranes into bilayer lipid membranes and purified mitochondrial proteins into proteoliposomes (Grover & Garlid, 2000; Ardehali & O’Rourke, 2005). In cardiac myocytes, diazoxide activated mitoK_ATP 1000–2000 times more potently than sarcolemmal K_ATP channels and it exerted cardioprotective effects during ischaemia in this mitoK_ATP-selective concentration range (Garlid et al. 1996, 1997). In addition, an inhibitor of mitoK_ATP channels, 5-hydroxydecanoic acid (5-HD), abolished the cardioprotective effect of diazoxide. Taken together, these results suggested that it was mitoK_ATP channels which play a pivotal role in cardioprotection evoked by KCOs. MitoK_ATP channels were also proposed to be the end-effecter of ischaemic preconditioning (Cohen et al. 2000), a mechanism by which brief periods of ischaemia provide protection against subsequent longer ischaemic periods (Murry et al. 1986).

Figure 1.

Ion transporters on the mitochondrial inner membrane described in the text A horizontal grey bar indicates the mitochondrial inner membrane. This membrane is polarized by ∼180 mV with matrix side negative (ΔΨm) due to a H⁺ gradient generated by respiratory enzyme complexes (REC) (Bernardi, 1999; Saraste, 1999). The energy stored in the form of ΔΨm is utilized to make ATP from ADP by ATP synthetase (AS). Mitochondrial ATP-sensitive channels (mitoK_ATP) mediate K⁺ influx along ΔΨm and cause a decrease in ΔΨm, matrix swelling and regulation of reactive oxygen species (ROS) generation. MitoK_ATP channels are activated by K⁺ channel opener compounds (KCOs) and inhibited by ATP, glybenclamide (Glb) and 5-hydroxydecanoic acid (5-HD). A decrease in ΔΨm inhibits Ca²⁺ influx through Ca²⁺ unipoters (Ca²⁺UP). Ca²⁺UP is inhibited by ruthenium red (RR). Mitochondrial permeability transition pore (PTP) is a specific, voltage-dependent, non-selective high-conductance channel that is activated by an increase in the intra-mitochondrial Ca²⁺ concentration and a decrease in ΔΨm. Ca²⁺ efflux through PTP is also facilitated by a decrease in ΔΨm. PTP is inhibited by cyclosporin A (Cys A).

However, the mechanisms by which mitoK_ATP channels exert their cardioprotective effects were poorly understood. An article from Terzic's laboratory published in The Journal Physiology in 1999 shed light on this issue (Holmuhamedov et al. 1999). By using mitochondria isolated from rat hearts, the authors showed that diazoxide (>1 μm) and another KCO, pinacidil (>10 μm), led to reduction of Ca²⁺ influx through a ruthenium red-sensitive Ca²⁺ uniport and an increase in Ca²⁺ efflux through a cyclosporin A-sensitive mitochondrial permeability transition pore (PTP), a specific, voltage-dependent, non-selective high-conductance channel that is activated by an increase in the intra-mitochondrial Ca²⁺ concentration ([Ca²⁺]_m) and a decrease in ΔΨm (Bernardi, 1999; Halestrap et al. 2004). Holmuhamedov et al. further showed that these effects of KCOs were inhibited by ATP, abolished by removal of extra-mitochondrial KCl and mimicked by the K⁺ ionophore valinomycin. They ascribed these effects to a decrease in ΔΨm induced by the KCOs and thus a decrease in driving force for Ca²⁺ influx, a hypothesis initially proposed by Liu et al. (1998). They also showed that diazoxide exerted a similar effect in a 5-HD-sensitive manner in intact cardiac myocytes. Murata et al. (2001) extended this work and showed that diazoxide reduced [Ca²⁺]_m in isolated cardiac myocytes under simulated ischaemia–reperfusion in a 5-HD-sensitive manner (Murata et al. 2001). Wang et al. (2001) reported that this was also the case in isolated hearts during ischaemia–reperfusion and that the reduction in [Ca²⁺]_m by diazoxide correlated with the recovery of the contractility after reperfusion (Wang et al. 2001).

However, Holmuhamedov's view was challenged by Kowaltowski et al. (2001) (Garlid, 2000; Kowaltowski et al. 2001). They argued that the bioenergetic effects observed by Holmuhamedov et al. with high concentrations of KCOs (i.e. >100 μm diazoxide or >50 μm pinacidil) resulted not from activation of mitoK_ATP channels but from the drugs’ protonophore activity and inhibitory effect on respiration. Furthermore, they found that the decrease in ΔΨm induced in isolated mitochondria by diazoxide and pinacidil (<50 μm) was too small (1–2 mV) to account for their cardioprotective effect. Instead, they found that the KCOs significantly increased the mitochondrial volume by causing a K⁺ influx and they suggested that this protected mitochondria during ischaemia–reperfusion by preserving the architecture of the intermembrane space with consequent slowing of ATP hydrolysis and preservation of the ability to use creatine efficiently as substrate on reperfusion. On the other hand, Korge et al. (2002) found that although diazoxide hardly decreased ΔΨm in energized mitochondria, it did so clearly in de-energized mitochondria (Korge et al. 2002). They showed that diazoxide thereby decreased Ca²⁺ influx and prevented Ca²⁺-induced opening of PTP, consistent with Holmuhamedov's view. Since widespread irreversible opening of PTP inevitably results in the necrosis of cardiac myocytes (Halestrap et al. 2004), they ascribed cardioprotection by KCOs to this effect. They also found that diazoxide prevented the release of cytochrome c from the intermembrane space perhaps by causing mitochondrial swelling. This would prevent cardiac myocytes from undergoing apoptosis (Akao et al. 2001). It should be noted that KCO-induced opening of mitoK_ATP channels may also cause cardioprotection by regulating the synthesis of reactive oxygen species during ischaemia–reperfusion (Ardehali & O’Rourke, 2005).

Thus, following Holmuhamedov's work (Holmuhamedov et al. 1999), a number of investigators have proposed different mechanisms by which mitoK_ATP channels can cause cardioprotection. Probably, these mechanisms are not mutually exclusive but coordinately cause cardioprotection during ischaemia–perfusion (Ardehali & O’Rourke, 2005). In spite of this remarkable progress, there still remain a number of questions regarding mitoK_ATP channels. For instance, diazoxide and 5-HD are reputed to specifically target mitoK_ATP channels but in fact both have other non-channel targets in mitochondria (Schafer et al. 1969; Hanley, 2002; Lim et al. 2002; Ozcan et al. 2002; Drose et al. 2006). Furthermore, it has been shown that diazoxide can activate sarcolemmal K_ATP channels especially in the presence of intracellular ADP (D’Hahan et al. 1999), and mouse atrial sarcolemmal K_ATP channels are highly sensitive to diazoxide (Zhang et al. 2009). Thus, one must be cautious in interpreting the effects of these agents. In addition, the molecular identity of mitoK_ATP channels remains unclear (O’Rourke, 2000, 2004). The pharmacological similarities between mitoK_ATP and sarcolemmal K_ATP channels might suggest that mitoK_ATP channels are composed of sulfonylurea receptors (SUR1, SUR2A or SUR2B) (receptors for KCOs and sulfonylureas) and pore-forming subunits (Kir6.1 or Kir6.2) as sarcolemmal K_ATP channels (Seino, 1999). Indeed, Grover & Garlid (2000) tentatively identified a 63 kDa sulfonylurea-binding protein and a putative pore-forming subunit of 55 kDa from mitochondria. Although some immunological analyses indicated the presence of these subunits in mitochondria (Suzuki et al. 1997; Lacza et al. 2003a,b; Singh et al. 2003; Cuong et al. 2005; Jiang et al. 2006), these observations were not confirmed by other investigators (Seharaseyon et al. 2000; Kuniyasu et al. 2003; O’Rourke, 2004; Foster et al. 2008). Liu et al. indicated that SUR1/Kir6.1 channels closely resembled mitoK_ATP channels in their pharmacological properties (Liu et al. 2001). However, diazoxide-induced protection of the brain from ischaemia was observed equally in SUR1 knockout and wild-type mice in a 5-HD-sensitive manner, indicating that SUR1 is not a required component of mitoK_ATP channels (Munoz et al. 2003). Seharaseyon et al. (2000) showed that transfection of a dominant negative construct of Kir6.1 did not affect mitoK_ATP channel activity in isolated rabbit ventricular myocytes (Seharaseyon et al. 2000), indicating that Kir6.1 is also not included in mitoK_ATP channels. Recently, Ardehali et al. proposed an alternative hypothesis that mitoK_ATP channels may be formed as a macromolecular complex containing mitochondrial ATP-binding cassette protein 1, phosphate carrier, adenine nucleotide translocator, ATP synthetase and succinate dehydrogenase (Ardehali et al. 2004). Thus, mitoK_ATP and sarcolemmal K_ATP channels may be completely different molecules. Identification of the molecular structure of mitoK_ATP channels will lead to more precise delineation of the mechanism underlying regulation of the channels and the development of drugs selectively acting on the channels. Therefore, further investigations are clearly needed in order to deepen our understanding of this important field of cardiovascular pathophysiology and pharmacology.