Cardiac Slo2.1 Is Required for Volatile Anesthetic Stimulation of K⁺ Transport and Anesthetic Preconditioning

CLINICALLY relevant doses of halogenated volatile anesthetics (VAs) can protect tissues such as the heart from ischemia–reperfusion (IR) injury,^1–3  a phenomenon known as “anesthetic preconditioning” (APC). APC is evolutionarily conserved from Caenorhabditis elegans⁴  to humans,³  and current American Heart Association/American College of Cardiology (AHA/ACA) guidelines specifically recommend the use of VAs during cardiac surgery for their protective effects.⁵ 

The mechanism of APC is complex and involves many of the same effector signaling pathways as ischemic preconditioning (IPC), including protein kinases C⁶  and A,^7,8  inhibitory G-proteins,⁹  adenosine,^10,11  nitric oxide,¹²  and mitochondrial reactive oxygen species (ROS) generation.^13,14  These signals are thought to converge at the level of mitochondria through distinct K⁺ channels, but the molecular identity of these channels is a subject of debate.

Several K⁺ channel types have been proposed to exist in mitochondria, including adenosine triphosphate (ATP)-sensitive (K_ATP),¹⁵  voltage-sensitive (K_V),¹⁶  and large-conductance (BK) channels of the Slo gene family¹⁷  as well as a K⁺/H⁺ exchanger¹⁸  (for review, see the studies reported by Szabo and Zoratti¹⁹  and Szewczyk et al.²⁰ ). Current evidence favors an involvement of mitochondrial K_ATP channels in IPC,²¹  whereas a mitochondrial Slo channel is thought to underlie APC.^7,22,23 

The mammalian Slo channel family comprises Slo1 (KCNMA1; BK), Slo2.1 (KCNT2, Slick), Slo2.2 (KCNT1, Slack), and Slo3 (KCNU1).²⁴ Slo3 expression is germline restricted,^24,25  but the others are widely expressed. Because pharmacologic Slo1 channel activators can protect the heart against IR injury,^17,26  it was widely assumed that a mitochondrial Slo1 channel (also termed BK, K_Ca, or BK_Ca) was responsible for APC. However, we recently showed that both C. elegans and mice lacking Slo1 channels can still be protected by APC, and mitochondria from these organisms contain a K⁺ channel activated by VA.²² 

We have also previously demonstrated that the K_Ca channel SLO-2 is required for VA-induced protection of C. elegans against IR injury, with slo-2-mutant nematodes also lacking VA- stimulated mitochondrial K⁺ flux.²²  Thus, we hypothesized that one of the mammalian SLO-2 orthologs, Slick or Slack (which are K_Na, not K_Ca channels^24,27 ), may underlie APC in mammals. Herein, we used novel gene-deleted mice (Slo2.1^−/−, Slo2.2^−/−, and Slo2.x double knockout [dKO]) to conclusively demonstrate that Slo2.1 codes the channel required for APC and for VA-stimulated K⁺ flux in cardiomyocytes and mitochondria.

Mice were housed in an Association for Assessment & Accreditation of Laboratory Animal Care-accredited pathogen-free facility with food and water available ad libitum. All procedures were approved by the University of Rochester’s Committee on Animal Resources (protocol no. 2010–030) in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (2011 revision). A Slo2.x dKO strain containing both mutant Slo2.x alleles²⁸  was provided by Dr. Chris Lingle, Ph.D. (Washington University, St. Louis, Missouri). Mice were outcrossed more than six times to a C57BL/6 background, and conventional breeding was used such that experimental wild-type (WT) and knockout mice (males, 8 to 10 weeks old) were littermates. All animals were genotyped by tail-clip polymerase chain reaction (Kapa Biosystems, USA)²⁹  with the following primers: Slo2.1 (Slick), C523-24C 5′-AACTTTATGAGTTCCTCTTCCATG-3′, C320-24F 5′-GAGCATCATACTTTGCTTTTTGGG-3′, KO = 269 base pairs (bp), WT = 579 bp; Slo2.2 (Slack) C311-30 5′-CCCATTCCACACTGCAGCCCTGTCTCTTTC-3′, C315-30 5′-TGTTTACTAGGGTCCAGGGAGAACCT ATGA-3′, KO = 200 bp, WT = 607 bp. Due to personnel limitations (i.e., same person handling mice and doing experiments), it was not feasible to blind the experimenter to animal genotype. However, for all experiments, mice of varying genotypes were randomly assigned to experimental groups, with treatments in randomized order across experimental days.

Mouse hearts were perfused as previously described.^22,26,29  In brief, after tribromoethanol anesthesia (100 mg/kg ip), the aorta was rapidly cannulated and perfused without pacing at a constant flow of 4 ml/min per 100 mg with Krebs–Henseleit buffer (118 mM NaCl, 4.7 mM KCl, 25 mM NaHCO₃, 10 mM d-glucose, 1.2 mM MgSO₄, 1.2 mM KH₂PO₄, and 2.5 mM CaCl₂, gassed with 95% O₂ and 5% CO₂ at 37°C). Left ventricular (LV) pressure was measured via a water-filled transducer-linked low density polyethylene balloon. LV and coronary root pressures were monitored and digitally recorded at 1 kHz (DATAQ, USA). Parameters calculated from the LV balloon throughout perfusion included heart rate, systolic and end-diastolic pressures, LV-developed pressure (systolic − diastolic), rate × pressure product, dP/dT_MAX (contraction), and dP/dT_MIN (relaxation).

Hearts were equilibrated for 20 min before initiating data acquisition. A total of 163 hearts were analyzed, and none were excluded from the study. A 30-min normoxic perfusion was followed by 30 min of no flow global ischemia and then 60 min of reperfusion (fig. 1). Treatment groups were as follows: (1) control IR injury; (2) IPC comprising three cycles of 5-min ischemia plus 5-min reperfusion to replace the 30-min perfusion; (3) 10-min isoflurane (Henry Schein Animal Health, USA) infusion (100 µM or approximately 0.35 minimum alveolar concentration [MAC] for a C57BL/6 mouse^30,31 ); (4) 20-min 7-chloro-3-methyl-4H-1,2,4-benzothiadiazine 1,1-dioxide (diazoxide, a K⁺ channel activator³² ; Sigma, USA) infusion (30 µM); and (5) 10-min 2,2’-sulfanediylbis(4,6-dichlorophenol) (bithionol a Slo channel activator³³ ; TCI America, USA) infusion (2.5 nM). Pharmacologic agents were delivered via syringe pump just above the perfusion cannula for the indicated times, followed by 30-s washout before index ischemia. At the end of IR protocols, hearts were sliced, stained in 1% (wt/vol) 2,3,5-triphenyltetrazolium chloride for 20 min and fixed in 10% neutral buffered formalin for 24 h. Heart slices were imaged and infarct size was analyzed by planimetry as previously described.^22,26,29 

Mouse adult ventricular cardiomyocytes were isolated by collagenase perfusion as previously described.²⁶  In brief, after anesthesia (tribromoethanol, see Ex Vivo Perfused Heart, above), hearts were cannulated and perfused with isolation buffer (IB) at 37°C (IB: 120 mM NaCl, 15 mM KCl, 0.6 mM Na₂HPO₄, 0.6 mM KH₂PO₄, 1.2 mM MgSO₄, 10 mM HEPES, 4.6 mM NaHCO₃, 30 mM taurine, 5.5 mM d-glucose, and 10 mM 2,3-butanedione monoxime, pH = 7.4) for less than 2 min before switching to digestion buffer (35 ml IB plus CaCl₂ 12.5 μM final, trypsin 400 μl at 2.5% [w/v], collagenase A 6.525 units, and collagenase D 15.375 units), and 8 min later placed in 2 to 3 ml of stop buffer (IB plus CaCl₂ 12.5 μM, 10% heat-inactivated Fetal Bovine Serum). Cardiomyocytes were dissociated and filtered through 75 μm mesh, and then, gravity sedimentation and resuspension were used to bring [Ca²⁺] stepwise to 1.8 mM, and the final pellet was placed into 1 ml of Minimal Essential Media (GIBCO, USA). Cell viability and yield were determined using Trypan blue and a hemocytometer, and preparations with greater than 85% viable rod-shaped cells were used experimentally. Cells were seeded (2500 cells per well) in a Falcon 96-well plate and incubated at 37°C for 1 h and then loaded with Thallos reagent (2 µM; Teflabs, USA) in 65-μl Hanks buffered salt solution (GIBCO) and monitored for thallium (Tl⁺) uptake by addition of 4 mM final Tl₂SO₄ in stimulus buffer (SB: 276 mM Na⁺ gluconate, 2.6 mM CaSO₄, 1.6 mM MgSO₄, 11.2 mM d-glucose, and 40 mM HEPES, pH = 7.3) in a BioTek Synergy plate reader in kinetic mode measuring every 7 s. (λex 488 nm; λem 525 nm) in the presence or absence of channel modulators. N-benzyl-N-(3-isobutoxy-2-pyrrolidin-1-yl-propyl)aniline (Bepridil; Sigma) is a calcium channel blocker that also inhibits Slo2 channels.³³  Eight measurements were taken before additions. Paired control wells contained Tl⁺-free SB, enabling determination of Tl⁺-specific fluorescence changes.

HEK293 cells were seeded on glass 12-mm coverslips and transfected using Lipofectamine 3000 (Invitrogen, USA) with the expression vector pKT122 expressing an mCherry:rat Slick complementary DNA fusion in a pcDNA5/TO vector backbone (Invitrogen). After 24 h, cells were incubated in 1 ml of Hanks buffered saline solution containing 2 µM Thallos reagent (TefLabs) for 30 min and then transferred to an open-flow perfusion rig attached to a Nikon TE2000 microscope (Nikon, USA) equipped with a monochromator (TILL Photonics, Germany), charge coupled device camera (PCO-TECH, USA), and appropriate filter sets. Transfected cells were identified via mCherry fluorescence and perfused with SB, with or without Slo2 channel modulators, then SB with 4 mM Tl₂SO₄, and images acquired every 5 s (λex 488 nm; λem 535 nm). Changes in fluorescent emission were quantified for all cells in the field of view, both transfected and untransfected, using TILLvisION software (TILL Photonics).

After anesthesia (tribromoethanol as above), mitochondria were isolated from three pooled mouse hearts, loaded with BTC-AM (Life Technologies, USA) and monitored for Tl⁺ uptake (a surrogate for K⁺ channel flux) as previously described.^22,29,34 

In perfused hearts, isoflurane infusion via syringe pump afforded no possibility for volatilization before entering the heart. For C57BL/6 mice, 100 µM isoflurane equates to approximately 0.35 MAC.^30,31  For cell and mitochondrial experiments, incubations generally lasted less than 1 min such that isoflurane volatilization was considered negligible. As such, isoflurane concentrations are denoted as “initial.”

Significance was determined by using a two-way ANOVA and when warranted post hoc unpaired t testing (two tailed) analysis using Bonferroni multiple comparisons correction. P values reported in the results are from t tests, with P value less than 0.05 (following multiple comparisons correction) considered statistically significant. Sample size was determined based on experience and previous studies.

In C. elegans, SLO-2 is encoded by a single gene and contributes to VA-stimulated mitochondrial K⁺ flux and APC.²²  However, in mammals, slo-2 has diverged into two paralogs (Slo2.1 and Slo2.2) with differing ion sensitivity to the C. elegans channel. By using a recently developed mouse strain with the genes coding for Slick (Kcnt2 or Slo2.1) and Slack (Kcnt1 or Slo2.2) both deleted,²⁸  our goal herein was to determine whether either or both of these mammalian gene products are orthologous to the worm SLO-2 channel and facilitate K⁺ flux across the mitochondrial inner membrane in response to VA. First, we verified using genomic polymerase chain reaction that the Slo2.x dKO contained the expected lesions in each of the two genes and that each allele was unambiguously detectable (fig. 2, A and B). Next, we used Western blot analysis to query whether the channels themselves were similarly ablated. Slick and Slack are abundantly expressed in neural tissue, and our results demonstrate that we can readily detect both paralogs in brain lysates as well as their absence in the dKO background (fig. 2C). Finally, we report that Slo2.x dKO mice were viable and had normal cardiac electrical function (table S1, Supplemental Digital Content 1, http://links.lww.com/ALN/B256).

To interrogate the individual contribution of each Slo2.x paralog to APC, the mutant alleles were genetically separated through extensive backcrossing to a C57BL/6 background. Ex vivo perfused hearts from WT and Slo2-ablated mice (both the Slo2.x dKO and single mutant alleles) were then subjected to IR injury with optional APC (isoflurane in perfusion media). Injury was assessed by measuring post-IR functional recovery (fig. 3A) and infarct size (fig. 3B). APC was protective in WT hearts, with a 95% improvement in the post-IR recovery of rate pressure product (RPP; heart rate × left ventricular–developed pressure; P = 0.0005) and a 44% reduction in infarct size (P = 0.0043) versus control hearts. Slo2.2^−/− mice exhibited a similar degree of protection by APC (132% improvement in recovery of RPP, P = 0.00016, 50% reduction in infarct size, P = 0.000005) versus control hearts. However, APC in Slo2.1^−/− mice resulted in no protection (12% worse recovery of RPP, P = 0.057, 2% reduction in infarct size, P = 0.817, compared with IR alone). As expected, APC in dKO mice was similarly ineffective (8% worse recovery of RPP, P = 0.725, 2% increase in infarct size, P = 0.854, compared with IR alone). None of the genotypes exhibited significant differences in baseline susceptibility to ischemia.

To discount the possibility that lack of protection by APC in the Slo2.1^−/− or dKO hearts was due to an inability of these hearts to be protected at all, we examined two additional cardioprotective treatments that are unrelated to VA: namely IPC (fig. 4) and the mitochondrial K_ATP channel opener diazoxide (fig. 5).

The improvement in functional recovery (RPP) induced by IPC for each genotype was as follows: WT 148%, dKO 184%, Slo2.1^−/− 108%, Slo2.2^−/− 193%, compared with IR alone, P < 0.025 for all genotypes (fig. 4A). The reduction in infarct size induced by IPC was as follows: WT 53%, dKO 34%, Slo2.1^−/− 45%, Slo2.2^−/− 49%, compared with IR alone, P < 0.0025 for all genotypes (fig. 4B). Similar to IPC, diazoxide delivered consistent cardioprotection across all genotypes. Namely, the improvement in functional recovery (RPP) induced by diazoxide: WT 98%, dKO 171%, Slo2.1^−/− 125%, Slo2.2^−/− 143%, compared with IR alone, P < 0.01 for all genotypes (fig. 5A). The reduction in infarct size induced by IPC was as follows: WT 37%, dKO 45%, Slo2.1^−/− 55%, Slo2.2^−/− 48%, compared with IR alone, P < 0.001 for all genotypes (fig. 5B). Thus, cardioprotection by either IPC or diazoxide was consistent regardless of genotype, suggesting Slo2.1^−/− hearts are still capable of being protected by these interventions and that cardioprotection, even via mechanisms thought to involve mitochondria, does not invariably require Slick.

In addition to functional recovery measured as RPP, data for dP/dT_MAX and dP/dT_MIN, LV-developed pressure, and LV end-diastolic pressure, for all genotypes and treatment regimens, are provided in Supplemental Digital Content 1, http://links.lww.com/ALN/B256, tables 2–4, respectively.

Given the proposed role of a mitochondrial large-conductance K⁺ channel in APC,^7,22,23  we examined VA stimulation of K⁺ flux in both primary cardiomyocytes and isolated cardiac mitochondria by using a thallium (Tl⁺)-based fluorescence assay.^22,34  All four genotypes were studied at the isolated mitochondrial level, but due to our initial results (vide supra) indicating that Slick but not Slack is involved in APC, studies in cardiomyocytes were limited to WT and Slo2.1^−/−.

The VA isoflurane was found to stimulate K⁺ flux in cardiomyocytes (fig. 6A) as well as mitochondria (fig. 6B) from WT mice, and this effect was blocked by the inhibitor bepridil. Bepridil is thought to target cardiac K_Na³⁵  and other channels³⁶  and has recently been shown to function as an Ebola antiviral.³⁷  Its use here was motivated by its ability to inhibit Slo2 channels,³³  which we validated using recombinant Slick expressed in HEK293 cells (fig. S1, Supplemental Digital Content 1, http://links.lww.com/ALN/B256). Notably, VA stimulation of K⁺ flux was absent in both cardiomyocytes and mitochondria from Slo2.1^−/− hearts (fig. 6, A and B). At the mitochondrial level, the Tl⁺ flux pattern in Slo2.2^−/− was similar to WT, and the pattern in dKO was similar to Slo2.1^−/−. These data indicate that isoflurane induces a Slo2.1-dependent K⁺ flux in both cardiomyocytes and mitochondria.

To substantiate the hypothesis that K⁺ flux and cardioprotection occur through Slick activation, we turned to a second distinct reagent. Bithionol has been shown to block Slo-type channels³³  and inhibits recombinant Slick expressed in HEK293 cells (fig. S1, Supplemental Digital Content 1, http://links.lww.com/ALN/B256). At the cardiomyocyte level, bithionol induced a bepridil-sensitive K⁺ flux in WT cells that was absent in Slo2.1^−/− cells (fig. 6C). In mitochondria, bithionol induced a bepridil-sensitive K⁺ flux in WT that was absent in the dKO mitochondria and reduced in both the Slo2.1^−/− and Slo2.2^−/− mitochondria (fig. 6D). The cell data fully support our hypothesis. However, the mitochondria data suggest that both Slick and Slack partially contribute to bithionol-stimulated Tl⁺ uptake, which is potentially interesting (see Discussion). However, it is likely that bithionol stimulates other channel activities or even mitochondrial function through unknown mechanisms of action, and very low levels of Slo2.2 are expressed in the heart.²⁸ 

Finally, additional support for our hypothesis came from the observation that bithionol was cardioprotective in WT hearts (fig. 7, A and B: 129% improvement in RPP recovery, P = 0.0098, 57% reduction in infarct size, P = 0.00008, compared with IR alone). This protection was partially lost in Slo2.1^−/− hearts (bithionol induced 97% improvement in RPP recovery, P = 0.028, 14% reduction in infarct size, P = 0.15). Overall, these data support a K⁺ transport pathway with the genetic and pharmacologic characteristics of Slick, being required for the cardioprotective effects of APC.

Herein, we demonstrated that the K_Na channel Slick, encoded by the Slo2.1 gene, is required for cardioprotection by APC in mice. We also showed that Slick is responsible for VA-stimulated K⁺ flux at both the cardiomyocyte plasma and mitochondrial membranes and that pharmacologic Slick openers are cardioprotective.

Although SLO-2 is a K_Ca channel in C. elegans, in mammals it has diverged into two paralogs, Slick and Slack, which are K_Na channels.^24,27  These channels may contribute to various physiologic functions attributed to K_Na channels, including metabolic sensing and neuronal plasticity,^24,27,38  and although the major focus of K_Na channel studies to date has been neuronal, K_Na channels are found in cardiac myocytes.³⁹  Notably, they have not been reported in mitochondria. Although these results cement a cardioprotective role for Slick, the relative contribution of plasma membrane versus mitochondrial Slick channels to this effect remains unclear and is a topic of ongoing research.

Many signaling pathways such as APC and IPC are evolutionarily conserved,²²  and there are many redundancies in their signaling cascades. For example, the mechanism of APC and IPC share commonalities such as protein kinase C activation,⁴⁰  ROS generation,^14,41–43  and mitochondrial K⁺ channels.^44–46  Although the precise mechanism of APC is not well understood, evidence suggests that APC involves both altered mitochondrial function and K⁺ channel activity (reviewed in the study by Agarwal et al.⁴⁷ ).

Studies have implicated the mitochondrial ATP-sensitive K⁺ channel (mK_ATP) in APC largely through the use of nonspecific measures of channel activity (e.g., flavoprotein fluorescence) and drugs that are known to have off-target mitochondrial effects (e.g., diazoxide or 5-hydroxydecanoate).^{44–46,48,49} Although these pharmacologic approaches suggest the involvement of the mK_ATP, the effects could have also been due to another K⁺ channel such as Slick. The current lack of a molecular identity for the mK_ATP channel precludes a definitive answer to this dilemma.

Moreover, the data presented herein do not rule out a potential cross talk between the mK_ATP and Slick. For example, because both diazoxide- and IPC-mediated cardioprotection were preserved in Slo2.1^−/− mice, Slick activation may be upstream of mK_ATP channel activation. Overall, although there are many similarities between APC and IPC, there is also evidence for distinct signaling pathways,⁵⁰  and herein we show that Slick appears to be involved only in APC, not in IPC.

In the field of APC, the canonical BK channel Slo1, which may be present in cardiac mitochondria,⁵¹  emerged as an early candidate effector of APC protection.^7,23  However, the importance of this channel was recently questioned by the evidence that both IPC and APC protection are robust in Slo1-ablated organisms.²²  Although the role of Slo1 in the heart remains to be fully elucidated, it is important to emphasize that our data do not preclude coexistence of Slo1 and Slick in cardiomyocytes or cardiac mitochondria. Rather, they conclude that only Slick is required for APC in the heart.

Our results are consistent with the findings in C. elegans, which showed that SLO-2 was required for APC.²²  Despite the large sequence similarity (approximately 74% identical)³⁸  between Slick and Slack, we found no role for Slack in cardiac APC or IPC. This is in agreement with the low expression levels of Slo2.2 in the heart.²⁷  However, this does not rule out a role for Slack in protecting other tissues from ischemia. Both Slo2.1 and Slo2.2 are highly expressed in the nervous system,²⁴  and there is evidence that Slick and Slack may coassemble into channels.⁵²  In this respect, it is intriguing that both Slick and Slack appeared to contribute to bithionol-induced K⁺ flux in purified mitochondria (fig. 6D). However, this was the only condition where both Slick and Slack were additive in our hands. Moreover, the mechanism of action of bithionol is unknown, but its activity as an antihelminthic may rely on suppressing succinate oxidation.⁵³  Hence, it is premature to conclude that Slack contributes to mitochondrial function. Nevertheless, it is possible that in other tissues such as the brain, Slo2.2 may play a role protecting against stress.

It is also interesting to speculate that Slick did not evolve to be opened by VAs, and given our results, we speculate that it may have a physiologic role in the mitochondrion. Although some mitochondrial K⁺ channels are demonstrated to play roles in pathophysiology (e.g., protection from IR injury),¹⁹  most are only reported at the phenomenological level as mitochondrial swelling or K⁺ flux sensitive to activators/inhibitors, with very little known about the normal function of these channels. The activation of mitochondrial K⁺ channels is hypothesized to modulate mitochondrial matrix volume, Ca²⁺ uptake capacity, and ROS production,⁵⁴  and although these events are involved in the pathophysiology of IR injury, they also have effects on regular mitochondrial function and metabolism. For example, Ca²⁺ stimulates oxidative phosphorylation, and matrix volume can determine the efficiency of high-energy phosphate transport between the mitochondrion and cytosol. Ongoing research in our laboratory is aimed at determining the role of Slick in regulating physiologic mitochondrial function.

A potential limitation of this study is the low dose of isoflurane (approximately 0.35 MAC) used to elicit cardioprotection. It is unclear if higher doses, or other VAs (e.g., desflurane, sevoflurane), will exhibit similar Slo2.1-dependent cardioprotective effects. Nevertheless, we note that the current widespread clinical use of VAs, and the ACC/AHA recommendations for their use during cardiac surgery,⁵  suggests that the identification of Slick as a mediator of APC may drive development of novel cardioprotective drugs targeting this channel.

The authors thank Chris Lingle, Ph.D. (Department of Anesthesiology, Washington University School of Medicine, St. Louis, Missouri), for providing Slo2.x knockout mice.

Supported by the National Institutes of Health (Bethesda, Maryland; grant no. R01-GM087483 to Drs. Brookes and Nehrke).

The authors declare no competing interests.