Channels of preconditioning: Potassium drain that protects the brain

Perioperative stroke and/or neurological dysfunction, accompanied with significant mortality or incapacitating sequelae, not infrequently complicates cardiac, vascular, or neurological surgery. Indeed a recent study in the Journal showed a remarkably high incidence of perioperative stroke, greater than 1 per 200 such patients.¹ Thus, clinical strategies aimed at either protecting the brain or enhancing its tolerance to ischemic or anoxic insults are worthwhile from the clinical point of view. In an article published in this issue of Anesthesiology, Bantel et al.² provide a fascinating preclinical series of studies examining the mechanisms of brain protection from anesthetics.

Thus, clinical strategies aimed at either protecting the brain or enhancing its tolerance to ischemic or anoxic insults are worthwhile from the clinical point of view.

A particular mechanism of tolerance, by which cells become resistant to subsequent lethal events, is mediated by “preconditioning” and within this context “ischemic preconditioning” develops when brief, sub-lethal episodes of ischemia activate endogenous protective mechanisms rendering organs resistant to subsequent, more severe ischemic events.³ Mimicking ischemic preconditioning, anesthetic preconditioning is a similar protective mechanism whereby exposure to an inhaled anesthetic also renders a tissue tolerant to a subsequent ischemic insult. Anesthetic preconditioning for myocardial protection has been studied extensively and shown to be relevant in clinical situations as well.⁴ Likewise, any benefits conferred by preconditioning strategies, including anesthetics, may be of value for protecting the brain or any neurologic tissue, even though such strategies have not been investigated in the brain as extensively as in the heart. However, several questions have been raised regarding feasibility of neuronal preconditioning, practical aspects, and its possible impact on perioperative outcome, especially after procedures with high incidence of perioperative stroke and adverse neuropsychiatric sequelae.^5–7

Is neuronal preconditioning feasible? Several preclinical and clinical studies positively indicate that neurons can also be preconditioned against ischemia. Mechanisms of neuronal preconditioning have been recently reviewed by Gidday⁸ and Obrenovitch.⁹ It is evident that even without preconditioning, neurons naturally recruit endogenous countermeasures against ischemic injury; adaptive responses mobilized as a response to preconditioning are of a significantly different nature. The preconditioning cascade includes stimuli that, via sensors and transducers, activate downstream transcription factors, ultimately modulating gene expression. Then, novel proteins may act as effectors enhancing the resistance of neurons to ischemia, and/or preexisting proteins may be modulated by post-translational modification acting as effectors, as well. By these cascades a latent neuro-protective phenotype is established prior to ischemia, which, together with other specific cellular responses, eventually results in a reprogrammed brain or other neuronal tissue, capable of tolerating ischemia better.

The classical preconditioning stimuli are brief, sub-lethal episodes of ischemia, but other agents can trigger the preconditioning cascade in neurons, as well. These include molecules (such as glutamate, reactive oxygen species, inflammatory cytokines and caspases) that in high concentrations are deleterious, but at lower concentrations initiate adaptive c-ascades. Of clinical interest is also the fact that drugs, like anesthetics^6,10,11 or the ATP-sensitive potassium (K_ATP) channel openers,¹² may act on and be recognized by molecular sensors in a fashion that initiates the preconditioning signaling cascade. Molecular sensors include a variety of neurotransmitters, neuromodulators, cytokine receptors, ion channels, and redox-sensitive enzymes.⁸ Of particular clinical and pharmacological interest are the K_ATP channels, which in addition to their role as sensors, may also have a parallel role as effectors of the phenotype that provides resistance to ischemia.¹²

Can anesthetics precondition the brain against ischemia? In the heart, preconditioning with volatile anesthetics mimics ischemic preconditioning in a fashion that involves K_ATP channel activation. The same is true in the brain: preconditioning with inhaled anesthetics protects neurons from ischemia-induced death as well.^10,13,14 However, it is not really clear what preconditioning of the central nervous system is and how beneficial it really is. Could it simply be a period of hibernation? This would be beneficial in promoting survival of critical patients, but would not be beneficial for the quality of life. Although inhaled anesthetics may not be practically applicable in all clinical settings wherein neuroprotection is desirable, they nevertheless may be suitably utilized to optimize perioperative outcome after surgical procedures associated with high risk adverse neurological sequelae. Studies of pertinent mechanisms may also lead to identification of novel and clinically more feasible means of neuronal preconditioning in clinical settings.¹⁰

Isoflurane, the most studied inhaled anesthetic, confers ischemic preconditioning in brain neurons of various species^10,13,14 in an age-specific and gender-specific fashion.¹⁵ Preconditioning effects have also been shown for halothane^10,13,14 and sevoflurane.¹⁶ In this issue of anesthesiology, Bantel et al.² provide further evidence supporting the role of plasmalemmal K_ATP channels as mediators and possibly effectors of neuronal preconditioning induced by the novel anesthetic gas xenon. Their study is also indicative of the multiple, diverse mechanisms by which neuronal preconditioning is attained. They show that xenon and sevoflurane at clinical concentrations precondition neurons, but via different pathways. While only xenon induced preconditioning via opening of K_ATP channels; sevoflurane did not. Because K_ATP channels are proven ‘metabolic sensors’ the agents activating these channels might initiate the preconditioning response in a way that most closely resembles that of the ‘physiological’ ischemic preconditioning. The next task is to elucidate which signal it is that the K_ATP channels generate in order to bring about longer lasting cellular changes.

Many studies do not distinguish between mitochondrial K_ATP channels, neuronal K_ATP channels, and cardiac muscle K_ATP channels. These are clearly separate molecular entities with their own specific, but overlapping, pharmacological profile. Bantel et al.² attempted to unambiguously distinguish between these, and one would hope that this will become a recurrent theme in future studies of neuronal preconditioning. K_ATP channels are octamers, consisting of 4 pore-forming (Kir6.x) coupled to 4 regulatory sulfonylurea receptor subunits (SURx). Three tissue-specific subtypes have been identified: Kir6.2/SUR1 (β-cell/neuroendocrine/neuronal), Kir6.2/SUR2A (cardiac and skeletal muscle type), and Kir6.2 (or 6.1)/SUR2B in smooth muscles.^17,18 In contrast to other types, neuronal K_ATP channels remain relatively under-investigated and their roles less well understood. Plasmalemmal Kir6.2/SUR1 channels are present in the brain, in areas such as hypothalamus, forebrain, and striatum.¹² These channels are activated by ADP, in the presence of Mg²⁺, and inhibited by high or physiological intracellular ATP concentrations.^12,20 Oxygen depletion leads to anoxic depolarization, whereby opening of voltage-gated channels results in Ca²⁺ overload activating lethal cascades and promoting cell death.¹² However, parallel activation of K_ATP channels (from reduction of intracellular ATP) produces outward K⁺ flux, membrane hyperpolarization and reduced excitability. Concomitant limitation of transmembrane ionic fluxes results in energy conservation mainly by decreased activity of ion pumps that consume about half of the total energy utilized by the brain.¹² Activation of K_ATP channels also reduces excitatory neurotransmission, including glutamate release, thus limiting the extent of excitotoxic injury. The protective effects of K_ATP channels against neuronal death from ischemic insults have been confirmed by a number of studies,^18,19 including observations in mutant mice lacking Kir6.2.²⁰

Mitochondrial K_ATP channels are also inhibited by ATP and mediate cytoprotective actions and preconditioning against cerebral and cardiac ischemia.¹² Both types of channels may act in a coordinated fashion,¹² but Bantel et al. showed that this was not the case with xenon.² Nevertheless, xenon, in contrast to halogenated anesthetics, at clinical concentrations independently enhanced plasmalemmal K_ATP channel activity by 50%, which is sufficient to trigger preconditioning. These findings should be verified in more complex systems: e.g., is xenon preconditioning, but not sevoflurane preconditioning, lost in K_ATP knock-out animals? The authors also implied that effective preconditioning may involve more than one diverse pathway in which K_ATP channels may or may not participate. These pathways may produce neuroprotection by converging to the same downstream effectors,⁸ will require further investigation.

While there is no doubt that xenon may limit ischemic neuronal injury via a preconditioning effect,^{21, 22} controversy arises from the notion that it may exacerbate ischemic damage to brain neurons and aggravate neurological sequelae by combined cardiopulmonary bypass and cerebral air emboli.²³ Therefore, further work is needed to ultimately document its efficacy and safety prior to indisputable, widespread acceptance for clinical use as a preconditioning and neuroprotective agent. The findings by Bantel et al.² are important and might stimulate the future investigations to combine systems-, cellular-, and molecular-level approaches. Clearly, in this challenging task, there is a need for more crosstalk between the different approaches taken by the large number of independent findings in order to make lasting contributions.