OUP user menu

Mitochondrial plasticity in classical ischemic preconditioning—moving beyond the mitochondrial KATP channel

Jan Minners, Christopher J. McLeod, Michael N. Sack
DOI: http://dx.doi.org/10.1016/S0008-6363(03)00337-7 1-6 First published online: 1 July 2003


Ischemic preconditioning is a powerful biologic phenomenon that activates innate cell survival programs to protect the heart from ischemic injury. The preponderance of research into classical ischemic preconditioning has focused on signaling pathways orchestrating cardioprotection. Conceptually classified into triggers, mediators and end effectors of preconditioning multiple distinct signaling pathways appear to ‘converge’ on the mitochondria possibly via activation of the mitochondrial ATP-sensitive potassium (mKATP) channel. The mechanisms by which mKATP channel activation induces preconditioning are incompletely elucidated but include perturbations of mitochondrial architecture and function. Since evidence invoking the mKATP channel has almost exclusively been based on studies using diazoxide and 5-hydroxydecanote the finding that these two compounds have mitochondrial effects independent of the mKATP channel has initiated a controversy regarding the exclusivity of this particular channel in preconditioning. A concerted effort to characterize the mitochondrial phenotype is important to advance our understanding of the mechanistic events that underlie the robust cardioprotective phenotype unmasked by preconditioning. The purpose of this review is to collate the information available on mitochondrial biology associated with classical preconditioning, to delineate the distinct temporal presentation of these mitochondrial perturbations, to reassess the role of the mitochondrial KATP channel and to propose a working model integrating the mitochondrial adaptations into the biology driving this cyto-protective phenotype.

  • Preconditioning
  • Mitochondria
  • K-ATP channel
  • Energy metabolism
  • Infarction

Time for primary review 28 days.

The preponderance of scientific experiments elucidating the signaling pathways involved in classical or acute first window cardiac preconditioning and suggesting the central role of mitochondrial KATP channel has been performed in in vivo and ex vivo animal studies (reviewed [1]). Due to the technical limitations of studying mitochondrial biology in the in vivo animal or intact heart, the majority of studies to ascertain mitochondrial function in preconditioning have been performed in permeabilized muscle strips, cell culture and isolated mitochondria. In these latter systems, the prototypic mKATP channel activator, diazoxide, has been used as the pharmacologic trigger to induce the preconditioning phenotype [2]. The studies focusing on the mitochondrial response to ischemic and diazoxide preconditioning will be reviewed to illustrate the emerging pattern of mitochondrial phenotypic perturbations in response to preconditioning triggers. The biology associated with ischemic and diazoxide preconditioning has proven to be dynamic and different mitochondrial effects have been described at temporally distinct periods of observation. To analyze these perturbations, this review will discuss the mitochondrial responses at three time points, i.e. in direct response to ischemic preconditioning or diazoxide (trigger phase), immediately prior to the index insult (preconditioned phase) and following the index insult (ischemia–reperfusion phase). These temporally distinct periods are schematized in Fig. 1. In addition, in light of recent findings the exclusivity of the mKATP channel will be critically reviewed and finally a working model integrating the mitochondrial responses to preconditioning triggers will be presented.

Fig. 1

Schematic of the temporally distinct time frames used to dissect out mitochondrial perturbation in response to ischemic preconditioning and in response to diazoxide administration. The hatched boxes represent episodes of ischemia and the solid line represents normal perfusion or maintenance media.

1 Trigger-phase mitochondrial modulation

It has recently been recognized that reactive oxygen species (ROS) signaling orchestrates cell survival programs [3]. A putative role of ROS cytoprotective signaling in ischemic preconditioning was previously demonstrated in the intact rat heart [4] and in isolated cardiomyocytes [5]. ROS generation resulting from preconditioning triggers has now been demonstrated in numerous laboratories (reviewed [1]). The mechanism whereby mKATP channel activation generates mitochondrial ROS production is uncertain, but this ROS production is clearly associated with a cytoprotective phenotype in preconditioning [1]. Hence, a feasible mechanism of mKATP channel activation induced protection could include cytoprotective ROS signaling during the trigger-phase of preconditioning.

Holmuhmedov et al. have investigated isolated cardiac mitochondria in response to pharmacologic mKATP channel activation [6]. Here numerous perturbations in function including depolarization of the inner mitochondrial membrane, accelerated respiration, slowed ATP production, release of accumulated mitochondrial Ca2+, mitochondrial swelling and the efflux of intermembrane proteins such as adenylate kinase and cytochrome c, have been associated with mKATP channel activation [6]. These data are conflicting in that the collective phenotype mechanistically supports both cell survival and cell death promoting cellular events. On the other hand, Kowaltowski et al. were unable to demonstrate a significant decrease in mitochondrial membrane potential in response to diazoxide [7] arguing that the predicted change in potential resulting from mKATP channel activation would be in the range of 1–2 mV, too low to be detected with potentiometric dyes. Similarly, diazoxide had no uncoupling effect on in situ mitochondria in permeabilized cardiac muscle fibers [8]. However, a substrate dependent inhibition of mitochondrial respiration was noted [8] resulting in an emerging concept postulating that moderate mitochondrial ‘stress’, albeit through unexplained mechanisms, primes the entry of the mitochondria into a ‘stress-resistant state’ [9,10]. The phenotypic response to mKATP channel activation could induce this moderate mitochondrial ‘stress’. This concept is further supported where diazoxide, albeit at higher dose than is routine, has been shown to acutely induce transient low conductance opening of the mitochondrial permeability transition pore (MPTP) [11]—a putative alternate ‘mitochondrial-stress’ that programs the subsequent cell-survival phenotype.

Mitochondrial flavoprotein fluorescence has also been demonstrated to be increased in diazoxide treated cardiomyocytes [12,13]. The initial interpretation was that diazoxide opening the mKATP channel resulted in potassium influx, thereby decreasing inner mitochondrial membrane potential. A subsequent acceleration of electron transfer, combined with a shortage of electron donors would result in oxidation and hence increased fluorescence of flavoproteins. As pointed out by Ovide-Bordeaux and Ventura-Clapier [8] diazoxide has been shown to inhibit succinate dehydrogenase (complex II) in a dose dependent manner [14]. Therefore, an alternative interpretation of the increased flavoprotein fluorescence in diazoxide treated cells is based on inhibited substrate oxidation within the electron chain. Here diazoxide could induce a transient ‘mitochondrial-stress’ which then, possibly via ROS production, activates the ischemia-tolerant phenotype.

2 Mitochondrial phenotype during the preconditioned-phase

The time period following the postischemic ‘trigger’ or after the washout of diazoxide represents the cell and mitochondrial ‘stress-resistant phase’. The mitochondrial perturbations identified here also reflect mitochondrial function immediately prior to the index ischemia–reperfusion insult.

Mitochondrial calcium [Ca2+]m overload in ischemia is detrimental to the cell [15]. A reduction in mitochondrial calcium uptake is an attractive hypothesis whereby mKATP channel activation mediates cytoprotection [6]. In intact adult cardiomyocytes diazoxide does not affect basal [Ca2+]m [16]. However, in response to ouabain-induced Ca2+ overload diazoxide significantly attenuated [Ca2+]m in these cardiomyocytes over a 30-min period [16]. Interestingly, this was associated with a concomitant depolarization of the inner mitochondrial membrane potential (ΔΨm). Mitochondrial depolarization during the preconditioned-phase has been demonstrated in both isolated mitochondria as well as in whole cells [17,18] and as mitochondrial Ca2+ uptake via the uniporter is driven, in part by ΔΨm [17,19] a reduction in ΔΨm may limit [Ca2+]m.

Our group questioned whether mitochondrial depolarization during the preconditioned-phase was exclusively linked to [Ca2+]m homeostasis or whether this electrochemical modulation was associated with broader energetic perturbations. In two distinct cell lines we demonstrated that diazoxide and ischemic preconditioning induced uncoupled oxidative phosphorylation during the preconditioned phase [18]. This was evident in that the mitochondria were depolarized with augmented O2 consumption and reduced cellular ATP levels. This observation was recently confirmed in the intact rat heart in response to the KATP channel opener, P-1075 [20]. The exact effect of uncoupling of oxidative phosphorylation in preconditioning has not been established, however, plausible effects include the reduction in ROS generation with a concomitant limitation in oxidative damage [10] and/or an uncoupling-induced augmentation of glucose uptake [21].

Kowaltowski et al. are skeptical of a direct mKATP channel activation effect on mitochondrial coupling and propose that the primary effect of mKATP channel activation is to augment mitochondrial matrix volume [7]. These investigators propose that mKATP channel activation-induced increases in mitochondrial matrix volume counteract matrix contraction that would otherwise occur during increased rates of ATP synthesis. This biochemical scenario could be quite compatible with the preconditioned-phase where an ATP deficit exists [18,22].

A plausible additional effect of the preconditioning ‘trigger’ could be the sustained activation of the mKATP channel. This scenario has been suggested by data that initiated the debate as to whether mKATP channel activation acts as a ‘trigger’ or a mediator of the preconditioning program [23]. In support of this concept, Akao et al. demonstrated that continuous exposure of rat neonatal cardiomyocytes to diazoxide resulted in cytoprotection against oxidative stress [24]. However, the temporal and mechanistic complexity of invoking the mKATP channel in this effect of diazoxide is illustrated in another study where reconstituted mKATP channels are shown to be activated by superoxide signaling [25].

Taken together, the diversity of mitochondrial effects described during the preconditioning-phase may not be mutually exclusive and could implicate combinatorial adaptive biology in response to mKATP channel activation.

3 Adaptive mitochondrial perturbations during the ischemia–reperfusion phase

The sine-quo-non of preconditioning is the significant diminution of cell death following myocardial ischemia and reperfusion event, above that achieved in the control heart exposed to the same index insult [26]. Mitochondrial function is significantly impaired during ischemia and reperfusion and has been identified as a central rheostat in orchestrating cell death. Hence, a logical extension of preconditioning should include a greater restoration of mitochondrial function in the preconditioned versus control heart following an ischemia–reperfusion insult. In reviewing the literature, the restoration of mitochondrial energetic function does appear to be a common feature in preconditioning across all species examined [27].

The role of mKATP channel activation in the restoration of mitochondrial function at reperfusion has been demonstrated in numerous laboratories in a variety of different models. Akao et al. demonstrated that the mKATP channel activators diazoxide and pinacidil maintained rat neonatal cardiomyocyte integrity in response to oxidative stress by preventing mitochondrial depolarization with a concomitant reduction in apoptotic inducing events such as cytochrome c translocation and caspase-3 activation [24]. Complementary data from Dos Santos et al. demonstrate that diazoxide pretreatment preserves the low outer mitochondrial membrane permeability to nucleotides and cytochrome c in response to ischemia and reperfusion in the rat heart [28]. Here it is postulated that mKATP channel activation results in a small decrease in intermembrane volume that in turn, increases the half-saturation constant for ADP stimulation of respiration with a resultant reduction in ATP hydrolysis. These effects are proposed to lead to preservation of adenine nucleotides during ischemia with augmented energy transfer upon reperfusion [28]. As an extension to this mechanistic insight, these investigators have also demonstrated that ischemic preconditioning preserves the functional coupling between mitochondrial creatine kinase and adenine nucleotide translocase [29]. In the direct evaluation of the mitochondrial response to anoxia and reoxygenation following diazoxide administration Ozcan et al. demonstrate a greater capacity to restore ADP-dependent respiration compared to vehicle treated control mitochondria [30]. In addition, in this same isolated mitochondrial protocol these investigators show that preadministration of diazoxide or an alternate mKATP channel opener (nicorandil) attenuates ROS generation and cytochrome c release at reoxygenation [31]. Finally, diazoxide has also been shown to suppress [Ca2+]m accumulation during ischemia and reperfusion in rabbit ventricular cardiomyocytes [32]. Whether the reduced uptake of [Ca2+] can be explained by the depolarizing effect of diazoxide remains to be established. Alternative channels situated in the inner mitochondrial membrane including a recently described calcium dependent potassium channel [33] and the mitochondrial permeability transition pore (MPTP) may well play a fundamental role in preconditioning. The effect of ischemic preconditioning and diazoxide inhibition on the MPTP in the ischemia–reperfusion phase of preconditioning has recently been explored [34,35]. However, as the preconditioning phase is associated with normalization of two of the adverse cellular components required to activate MPTP [36], i.e. [Ca2+]m overload and diminished mitochondrial ATP synthesis with a concordant low ΔΨm, the primary role of MPTP regulation in preconditioning needs to be further delineated. Collectively, however, the consequence of all of these restorative effects of ischemic preconditioning and diazoxide, putatively via mKATP channel activation, on postischemic mitochondrial function is probably central to the cytoprotective program induced by preconditioning.

4 Beyond the mitochondrial KATP channel

The exclusivity of mKATP channel activation in the mitochondrial perturbations described above has begun to be questioned. Terzic's laboratory has demonstrated that the ROS attenuating effects of mKATP channel openers in the ischemia–reperfusion phase were maintained in a nominally K+-free medium (Ozcan et al. [31]). These investigators went on to show that oxidant stress attenuation was present in the presence or absence of the K+-ionophore valinomycin and was mimicked by malonate, a modulator of the mitochondrial redox state [31]. These investigators postulated that a K+ conductance-independent pathway may be activated by the mKATP openers to confer mitochondrial protection.

Several laboratories have supported the concept of K+ conductance-independent effects of diazoxide [8,37,38]. These groups all demonstrate that a cardioprotective dose of diazoxide selectively attenuates succinate oxidation. Interestingly, Hanley et al. also go on to demonstrate that pinacidil did not alter succinate oxidation, but selectively inhibited NADH oxidation [37]. These data support a mechanistic link between partial inhibition of electron transport and preconditioning. Interestingly, partial inhibition of the mitochondrial electron transfer chain complex II with 3-nitropropionic acid confers cardioprotection in rabbits [39] and we and others have demonstrated cardioprotection with the transient administration of the mitochondrial uncoupler dinitrophenol [9,40]. Finally, the pretreatment of rats with cyclosporin A resulted in a reduction in ATP levels and a subsequent reduction in infarct size [41]. Collectively, these data support a role, albeit poorly understood, whereby partial respiratory inhibition evokes the preconditioning program. This hypothesis is also consistent with the uncoupled phenotype described during the preconditioning phase [18].

Hanley et al. also demonstrate that the classical mKATP channel blocker, 5-hydroxydecanoate (5-HD), is converted to 5-HD-CoA [37]. Here they postulate that β-oxidative metabolites of 5-HD-CoA could enter the electron-transfer chain at the level of ubiquinone, thereby compensating for the partial inhibition of the respiratory chain by diazoxide and pinacidil at complexes II and I, respectively. This mitochondrial respiration augmenting effect of the fatty-acid 5-HD-CoA may explain a mechanism whereby 5-HD inhibits preconditioning. If 5-HD is indeed metabolized the oxidation of fatty acids at reperfusion will adversely affect cellular recovery via inhibition of glucose metabolism [42] providing an alternative hypothesis on the mechanism of 5-HD.

5 Working model of the temporal events shaping the preconditioned mitochondria

Collectively the investigations described in this mini-review can be incorporated into a dynamic model of mitochondrial function that may facilitate the development of the cardioprotective phenotype evident in preconditioning (schematized in Fig. 2). In brief, the trigger-phase is associated with mitochondrial stress as evident by the multitude of reversible perturbations in mitochondrial function. A probable component of this mitochondrial stress is the blunting of mitochondrial respiration, which in turn, may generate reactive oxygen species [43]. These, in turn, are postulated to act as cytoprotective signaling intermediates, although downstream targets have not been extensively investigated in classical preconditioning. The preconditioned phase is typified by mitochondrial uncoupling, optimized mitochondrial architecture to facilitate respiration and diminished mitochondrial Ca2+ stores. The preconditioning-induced prosurvival phenotype during postischemic reperfusion is characterized by an enhanced capacity to restore ATP production, blunted generation of ROS, maintenance of mitochondrial outer membrane integrity, diminished mitochondrial calcium overload and the possible induction of yet unidentified cytoprotective peptides. In addition, this prosurvival phenotype probably augments the threshold below which MPTP opening facilitates cell death.

Fig. 2

Schematic of the mechanistic paradigm of preconditioning orchestrated mitochondrial perturbations that promote an ischemia-tolerant phenotype. This working model is described in the final section of the text.

The role of the mKATP channel in preconditioning is still under intense investigation and its further molecular characterization should ultimately provide insight into its mechanistic function. However, the data presented in this short review caution against an exclusive role of the mKATP channel in preconditioning. We have generated a working model of mitochondrial perturbations which, we believe, should facilitate a greater understanding of the role of the mitochondria in conferring the cellular protection seen in preconditioning.


The research in the Hatter Institute for Cardiology Research is funded in part by the Wellcome Trust (MNS), The Aaron Beare Foundation (JM) and the South African Medical Research Council (CJM).


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
  33. [33]
  34. [34]
  35. [35]
  36. [36]
  37. [37]
  38. [38]
  39. [39]
  40. [40]
  41. [41]
  42. [42]
  43. [43]
View Abstract