Cardiovascular pharmacology

Induction of autophagy restores the loss of sevoflurane cardiacpreconditioning seen with prolonged ischemic insult

Mayumi Shiomi a, Masami Miyamae b,n, Genzou Takemura c, Kazuhiro Kaneda d,Yoshitaka Inamura d, Anna Onishi d, Shizuka Koshinuma d, Yoshihiro Momota d,Toshiaki Minami a, Vincent M. Figueredo e

a Department of Anesthesiology, Osaka Medical College, Osaka, Japanb Department of Internal Medicine, Osaka Dental University, 8-1 Kuzuha hanazono-cho Hirakata, Osaka 573-1121, Japanc Department of Cardiology, Gifu University Graduate School of Medicine, Gifu, Japand Department of Anesthesiology, Osaka Dental University, Osaka, Japane Institute for Heart and Vascular Health, Einstein Medical Center, and Jefferson Medical College, Philadelphia, USA

a r t i c l e i n f o

Article history:Received 18 July 2013Received in revised form17 December 2013Accepted 18 December 2013Available online 25 December 2013

Keywords:SevofluranePreconditioningAutophagyIschemia-reperfusionChloramphenicol

a b s t r a c t

Sevoflurane preconditioning against myocardial ischemia-reperfusion injury is lost if the ischemic insultis too long. Emerging evidence suggests that induction of autophagy may also confer cardioprotectionagainst ischemia-reperfusion injury. We examined whether induction of autophagy prolongs sevofluranepreconditioning protection during a longer ischemic insult. Isolated guinea pigs hearts were subjected to30 or 45 min ischemia, followed by 120 min reperfusion (control). Anesthetic preconditioning waselicited with 2% sevoflurane for 10 min prior to ischemia (SEVO-30, SEVO-45). Chloramphenicol(autophagy upregulator, 300 mM) was administered starting 20 min before ischemia and throughoutreperfusion in SEVO-45 (SEVO-45þCAP). To inhibit autophagy, 3-methyladenine (10 μM) was adminis-tered during sevoflurane administration in SEVO-45þCAP. Infarct size was determined by triphenylte-trazolium chloride stain. Tissue samples were obtained before ischemia to determine autophagy-relatedprotein (microtubule-associated protein light chain I and II: LC3-I, II), Akt and glycogen synthase kinase3β (GSK3β) expression using Western blot analysis. The effect of autophagy on calcium-inducedmitochondrial permeability transition pore (MPTP) opening in isolated calcein-loaded mitochondriawas assessed. Electron microscopy was used to detect autophagosomes. Infarct size was significantlyreduced in SEVO-30, but not in SEVO-45. Chloramphenicol restored sevoflurane preconditioning lost by45 min ischemia. There were more abundant autophagozomes and LC3-II expression was significantlyincreased in SEVO-45þCAP. Induction of autophagy before ischemia enhanced GSK3β phosphorylationand inhibition of calcium-induced MPTP opening. These effects were abolished by 3-methyladenine. Pre-ischemic induction of autophagy restores sevoflurane preconditioning lost by longer ischemic insult. Thiseffect is associated with enhanced inhibition of MPTP by autophagy.

& 2013 Elsevier B.V. All rights reserved.

1. Introduction

Experimental (Inamura et al., 2009; Kaneda et al., 2008; Kerstenet al., 1997) and clinical studies (Garcia et al., 2005; Julier et al., 2003)have demonstrated that the use of volatile anesthetics constitutes anadditional therapeutic approach in the care of patients at risk ofdeveloping perioperative cardiac complications; known as volatileanesthetic preconditioning (De Hert, 2006). For example, sevofluraneis a popular anesthetic with few clinical side effects and is usedworldwide. The washout time of volatile anesthetics prior to

ischemia (Okusa et al., 2009) and the duration of sustained ischemia(Kevin et al., 2003; Sigaut et al., 2009) have been shown to becrucial to successful volatile anesthetic cardioprotection. For example,cardioprotection by sevoflurane preconditioning is lost if theischemic insult is too long. Kevin et al. demonstrated that cardio-protection by sevoflurane preconditioning is restricted to a range ofischemic durations of 25–40 min in isolated guinea pig hearts (Kevinet al., 2003). Similarly, the neuroprotective effects of clinicallyrelevant concentrations of sevoflurane was observed only for periodsof ischemia of 30 min or less (Sigaut et al., 2009). Finding strategiesthat prolong cardioprotection by sevoflurane preconditioning duringa longer ischemic insults could be clinically useful.

Autophagy is a catabolic process through which damaged orlong-lived protein and organelles are degraded using lysosomal

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/ejphar

European Journal of Pharmacology

0014-2999/$ - see front matter & 2013 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.ejphar.2013.12.027

n Corresponding author.: Tel.: þ81 72 864 3079; fax: þ81 72 864 3179.E-mail address: miyamae0907@gmail.com (M. Miyamae).

European Journal of Pharmacology 724 (2014) 58–66

degradative pathways. This is an evolutionally conserved processcrucial for normal tissue homeostasis. Autophagy has been con-sidered to be one of the modes of cell death, termed “autophagiccell death”. However, no specific autophagy death pathway hasbeen identified (Aviv et al., 2011). Certainly, necrosis and apoptosiscontribute to cardiomyocyte death in ischemia-reperfusion injury.However, accumulating evidence suggests that autophagy is stimu-lated by ischemia and actually contributes to cardiomyocyte survi-val (Kanamori et al., 2011; Yan et al., 2005). Huang et al. in anisolated mice hearts model reported that autophagy induced byischemic preconditioning was essential for cardioprotection (Huanget al., 2010). Sala-Mercado et al. demonstrated that pre-ischemicinduction of autophagy by chloramphenicol succinate limits myo-cardial infarct size in in vivo swine hearts (Sala-Mercado et al.,2010). Yet, excessive self-digestion and degradation of essentialcellular components by autophagy could be detrimental (Scarlattiet al., 2009). Thus, to date, it is unclear whether upregulation ofautophagy is protective or detrimental for ischemic myocardium.

Autophagy has been shown to be triggered by opening ofmitochondrial permeability transition pore (MPTP), which alsoplays an important role in myocardial ischemia-reperfusion injury(Halestrap et al., 2004). Studies have shown that inactivation ofglycogen synthase kinase 3β (GSK3β) by phosphorylation at Ser9

inhibits MPTP opening and protects cardiomyocytes againstischemia-reperfusion injury (Juhaszova et al., 2004). It has notbeen determined whether pre-ischemic induction of autophagyaffects MPTP. We hypothesized that pre-ischemic induction ofautophagy restores loss of sevoflurane preconditioning seenwith prolonged ischemic insult and this effect is associated withMPTP opening. Elucidating the role of autophagy in myocardialischemia-reperfusion injury and finding strategies that prolongcardioprotection by sevoflurane preconditioning during longerischemic insults could prove clinically useful in patients at cardi-ovascular risk during perioperative periods.

2. Materials and methods

2.1. Animals

This study was conducted in accordance with the Guidelinesfor Animal Research at Osaka Dental University, and with theapproval of the Animal Experiment Committee of Osaka DentalUniversity, Osaka, Japan. These guidelines conform to those laidout in the Guide for the Care and Use of Laboratory Animals,available from the National Academy of Science. Male Hartleyguinea pigs were fed Lab Diet (RC4, Oriental Yeast, Tokyo, Japan)and given water ad libitum. Chloramphenicol was purchased fromCalbiochem (La Jolla, CA) and was used to induce autophagy.3-methyladenine was purchased from Sigma Aldrich (Ann Arbor,MI) and was used to inhibit autophagy.

2.2. Isolated heart perfusion and measurement of function

Male guinea pigs weighing 550–700 g (12–13 weeks old) weregiven heparin (1000 units intraperitoneally), then anesthetized withpentobarbital (60 mg/kg, intraperitoneally). Hearts were excised andimmediately arrested in cold iso-osmotic saline containing 20 mM KCl.The aorta was cannulated and the isolated hearts were perfused at70 mmHg on a nonrecirculating isovolumic perfused heart apparatus,using a Krebs–Henseleit perfusate and paced at 240 beats/min aspreviously described (Inamura et al., 2009). Left ventricular developedpressure (mmHg) was measured using a 2.5 French, high-fidelitymicromanometer (Nihon-Kohden, Tokyo, Japan) passed into a com-pliant latex balloon, inserted into the left ventricle, and recorded on aPowerLab 2/20 Data Recording System (ADInstruments, Hayward,

Australia). The balloon was connected to a Y-adapter with one endused to advance the micromanometer and the other used to fill theleft ventricular balloon with bubble-free water to an end-diastolicpressure of 10 mmHg. Coronary flow (ml/min) was measured bycollecting effluent. Global ischemia was achieved by clamping theaortic inflow line. During ischemia, hearts were maintained at 37 1C byenclosure in a water-jacketed air chamber. Warmed perfusate kept inthe lower part of the chamber saturated the air with humidity andprevented cooling by evaporation. Heart temperature was continu-ously monitored by a digital thermometer (PTW-100A, Unique Med-ical, Tokyo, Japan). Sevoflurane was insufflated by passing the 95%O2/5%CO2gas mixture through a calibrated vaporizer (ACOMA, Tokyo,Japan). Samples of coronary perfusate were collected anaerobicallyfrom the aortic cannula for measurement of sevoflurane concentrationby an organic vapor sensor (OSP, Saitama, Japan). We measuredsevoflurane concentration at 10 min after administration, when sevo-flurane is known to be equilibrated in the tissue. Sevoflurane was notdetected in the effluent at 50 min after discontinuation of sevoflurane.Thus, at the end of ischemia-reperfusion in the present study, tissuesevoflurane concentrations are negligible.

2.3. Experimental protocol

Animals were assigned to one of 11 groups (n¼8 each; Fig. 1).After a 20 min equilibration, baseline left ventricular developedpressure, left ventricular end-diastolic pressure and coronaryflow were recorded. Hearts were subjected to 30 min (control;CTL-30) or 45 min (control; CTL-45) of ischemia followed by120 min of reperfusion. Anesthetic preconditioning was elicitedby administration of sevoflurane (2%) for 10 min followed by10 min washout before 30 min or 45 min ischemia (SEVO-30,SEVO-45). For induction of autophagy, chloramphenicol (300 mM)was administered starting 20 min before 45 min ischemia andthroughout the reperfusion period in sevoflurane-treated (SEVO-45þCAP) or non-sevoflurane-treated animals (CAP-45) hearts.Additional studies were performed with the autophagy blocker,3-methyladenine (10 μM) was administered during sevofluraneadministration for 20 min and throughout the reperfusion periodin sevoflurane and chloramphenicol treated hearts (SEVO-30þ3MA, SEVO-45þ3MA, CAP-45þ3MA, SEVO-45þCAPþ3MA)and its vehicle (saline) in CTL-45 (CTL-45þ3MA; Fig. 1). Chlor-amphenicol was dissolved in distilled water and 3-methyladenine was dissolved in 0.04% ethanol, and they wereadded to the Krebs–Henseleit perfusate to a final concentration of300 mM and 10 μM, respectively.

2.4. Determination of myocardial infarct size

At the end of experiments, the hearts were quickly frozen at –80 1C for 15 min, then sliced into 2 mm thick transverse sectionsfrom apex to base (6 slices/heart). After removing the rightventricle and defrosting, each slice was weighed and incubatedat 37 1C with 1% triphenyltetrazolium chloride (Sigma Chemicals)in phosphate buffer (pH 7.4) for 10 min and then fixed in 10%formalin for at least 5 h to distinguish red stained viable tissuefrom pale unstained necrotic tissue (Fishbein et al., 1981). Eachslice was photographed and the necrotic area was determinedusing Adobe Photoshops CS (Adobe, San Jose, CA) and multipliedby the weight of the slice, then expressed as a fraction of the leftventricle.

2.5. Western blot analysis

Separate experiments were performed (n¼4 in each group) toexamine expression of autophagy-related protein (microtubule-associated protein light chain I and II: LC3-I, II), Akt, GSK3β and

M. Shiomi et al. / European Journal of Pharmacology 724 (2014) 58–66 59

cyclophilin D. Myocardial tissue samples were collected before45 min ischemia, and homogenized in ice-cold homogenizingbuffer containing 250 mM sucrose, 20 mM HEPES (pH 7.5),10 mM KCl, 2 mM EGTA, 2 mM MgCl2, 25 mM NaF, 50 mM β-glycerophosphate, 1 mM Na3VO4, 1 mM PMSF and protease inhi-bitor leupeptin (10 μg/ml). The homogenate was centrifuged at1000g for 5 min at 4 1C to clean up. The supernatant was re-centrifuged at 10,000g for 15 min at 4 1C to obtain cytosolicfraction. The pellet was designated as crude mitochondrial fractionand was resuspended in homogenizing buffer with 1% TritonX-100, incubated on ice for 1 h, and then re-centrifuged at10,000g for 15 min at 4 1C. The resulting supernatant was usedas mitochondrial fraction. The protein concentration was esti-mated with a Bradford assay. Equivalent amounts (20 μg) ofprotein samples were loaded and separated on a 10% SDS-PAGEgradient gel, then electrically transferred overnight to a polyviny-lidene difluoride membrane (Bio-Rad, Hercules, CA). After blockingwith 5% skim milk in Tris-buffered saline containing 0.1% Tween-20 (TBS-T), the membranes were incubated for 2 h at 4 1C in TBS-Tcontaining 5% skim milk and overnight 1:200–1000 dilution ofrabbit primary antibody for LC3 (Medical & Biological LaboratoriesCo LTD, MA), phospho Akt (Ser47), phospho GSK3β (Ser9) (CellSignaling TECHNOLOGY, Boston, MA) and cyclophilin D (Santa CruzBiotechnology, Santa Cruz, CA). Membranes were incubated with a1:1000 dilution of horseradish peroxidase-labeled anti-rabbitimmunoglobulin G (NA 934V, GE Healthcare, Buckinghamshire,United Kingdom). The same blot was stripped and re-blotted withantibodies to α-tubulin (Santa Cruz Biotechnology) (to confirmequal protein loading), total Akt and GSK3β (Cell Signaling TECH-NOLOGY). Bound antibody signals were detected with enhancedchemiluminescence (Pierce Biotechnology, Rockford, IL) and visua-lized using VersaDoc 5000 Imaging System (Bio-Rad, Hercules,CA). Quantitative analysis of the band densities was performed byQuantity One software (Bio-Rad), and the results are presented asthe ratio of phospho Akt and phospho GSK3β to total (includingnon-phosphorylated and phosphorylated protein) Akt and GSK3β,

respectively. The average light intensity was multiplied by 100 tofacilitate presentation of an x-fold increase.

2.6. Studies in isolated cardiac mitochondria

To investigate the involvement of MPTP in cardioprotection byautophagy, mitochondria were isolated from guinea pig hearts(n¼4 for each group). For this purpose, different hearts from theinfarct size study were used. After perfused and treated with sevo-flurane, chloramphenicol and 3-methyladenine, isolated hearts wereremoved from the Langendorff apparatus, and homogenized in ice-cold MSTEB buffer containing 210mM mannitol, 70 mM sucrose,10 mM Tris/HCl (pH 7.4), 1 mM EGTA 0.5 mg/mL bovine serumalbumin, pH 7.5. After centrifugation (450g, 5 min) two-thirds of thesupernatant was decanted into fresh, pre-chilled tubes. The mitochon-dria were isolated by further centrifugation (5800g, 10 min). Themitochondrial pellet was re-suspended in cold MST buffer containing210mM mannitol, 70 mM sucrose, 10 mM Tris/HCl (pH 7.4), pH7.5 and the previous centrifugation step repeated. The extractedmitochondria were diluted in ice-cold respiratory buffer (5 mg/ml)containing 250mM sucrose, 10 mMHepes, 2 mMK2HPO4, 80 mMKCl,2 mM Mg acetate, pH 7.5 and incubated with 1 μM calcein-AM(Invitrogen Molecular Probes, Carlsbad, CA) for 15 min at roomtemperature. Calcein-AM readily enters the mitochondria and istrapped in the matrix in its free form, which is fluorescent. Aftercalcein was trapped in mitochondria, the mitochondria were washedby KCl buffer containing 120mM KCl, 5 mM TES, 0.1 mM MgCl2,0.2 mM ATP, 10 mM sodium succinate. Calcein-loaded mitochondriawere treated with 50 and 150 μM Caþ þ per milligram of protein, andwere incubated for 10 min at room temperature. Then mitochondrialfluorescences were acquired. Flow cytometric analysis was performedon FACS calibur™ (Becton Dickinson, Franklin lakes, NJ). Mitochondrialabeled with calcein-AM were analyzed by flow cytometry in aninstrument equipped with a 488 nm excitation source.

Fig. 1. Schematic illustration of the experimental protocol used in this study. There are eleven groups. After 20 min baseline period, isolated perfused hearts were subjectedto 30 or 45 min global ischemia and 120 min reperfusion. Anesthetic preconditioning was elicited by administration of 10 min of sevoflurane (2%) with 10 min washoutbefore 30 or 45 min ischemia. chloramphenicol and 3-methyladenine was administered starting 20 min before 30 or 45 min ischemia and throughout the reperfusion periodin hearts from sevoflurane-treated or non-sevoflurane-treated animals. Tissue samples were obtained before ischemia. CTL, control; CAP, Chloramphenicol; 3-MA,3-Methyladenine; SEVO, sevoflurane.

M. Shiomi et al. / European Journal of Pharmacology 724 (2014) 58–6660

2.7. Electron microscopy

To confirm the induction of autophagy, cardiac specimens wereobtained before sustained ischemia. They were fixed withphosphate-buffered 2.5% glutaraldehyde (pH 7.4) and were post-fixed with 1% osmium tetroxide, after which they were conven-tionally prepared for transmission electron microscopy (H-800,Hitachi, Tokyo, Japan).

2.8. Statistical analysis

All data are expressed as mean7S.D.. Statistical power analysisrevealed that a sample size of n¼8 would provide sufficient power(0.8) to detect a difference between mean infarct size indices of 15%(S.D.¼9, α¼0.05). A group size of n¼4 was used for Western blotand calcein studies to provide a power of 0.8 to detect a differencebetween means of 20% (S.D.¼10, α¼0.05). Hemodynamic data weretested for normal distribution and subsequently analyzed by a two-factor repeated-measures analysis of variance for time and treatment.If an overall difference between the variables was observed, compar-isons were performed as one-way ANOVA followed by Tukey0s post-hoc test for inter-group differences and by Dunnett0s for intra-groupdifferences with baseline values as the reference time point. Analysisof infarct size, Western blot and mitochondrial calcein fluorescencewas performed using one-way ANOVA followed by Student0s t-testwith Bonferroni0s correction for multiple comparisons to avoid type Ierror. For changes within and between groups a two-tailed P valueless than 0.05 was considered significant in advance. (SPSS17 forWindows, SPSS Japan, Tokyo, Japan).

3. Results

3.1. General observations and survival rate

Of a total of 205, 7 hearts were not used secondary to intractableventricular fibrillation after reperfusion (3 in CTL-45, 2 in CAP-45, 1 inSEVO-45þ3MA and 1 in SEVO-45þCAPþ3MA) and one heart wasnot used due to aortic rupture. Additional hearts were studied untileach group had n¼8 successful experiments for the infarct sizestudy. There was no significant difference in body weight amonggroups. The concentration of sevoflurane in the coronary perfusateafter 10 min of exposure was 0.2570.02 mM. Sevoflurane was notdetected in the effluent during the baseline, ischemic, and reperfu-sion periods.

3.2. Hemodynamics

Hemodynamic data are shown in Table 1. Baseline left ven-tricular developed pressure and coronary flow were similar amongthe eleven groups. Left ventricular developed pressure slightlydecreased during sevoflurane administration, but the decrease didnot reach statistical significance. Administration of chlorampheni-col or treatment with 3-methyladenine did not significantly affectleft ventricular developed pressure or coronary flow before ische-mia. After 120 min reperfusion, only SEVO-30 had significantlyhigher left ventricular developed pressure compared to CTL-30.Left ventricular end-diastolic pressure was significantly increasedin CTL-30 compared with SEVO-30 and SEVO-30þ3MA, whichsuggests impaired diastolic function. This is due to the large infarctsize. Treatment with 3-methyladenine alone(CTL-45þ3MA) didnot affect the recovery of left ventricular developed pressurecompared with CTL-45. There was no significant difference incoronary flow among all the groups throughout the experiment.This suggests that changes in coronary flow could not account forthe improved contractile recovery of the SEVO-30.

3.3. Myocardial infarct size

Myocardial infarct size data are shown in Fig. 2. Myocardial infarctsize in SEVO-30 was significantly reduced by approximately 50%compared with control hearts (SEVO-30:2478% vs. CTL-30:4876%,Po0.05). However, this effect was lost when the ischemic periodwas extended to 45 min (SEVO-45:5376% vs. CTL-45:54711%,P¼NS). Treatment with chloramphenicol restored the protectionlost by prolonged ischemia in SEVO-45 (SEVO-45þCAP: 3478%vs. SEVO-45: 52710%, Po0.05). Restoration of cardioprotectionachieved by chloramphenicol was abolished by 3-methyladenine(SEVO-45þCAPþ3MA: 5178% vs. CTL-45, P¼NS). Importantly,3-methyladenine failed to abolish the reduction in infarct size inSEV-30 (SEVO-30: 2478% vs. SEVO-30þ3MA: 32713%, P¼NS).Treatment with chloramphenicol alone did not limit infarct size(CAP-45: 5179%).

3.4. Electron microscopy

Fig. 3 shows representative electron microphotographs of eachstudy group. Although autophagic vacuoles were rarely seen incardiomyocytes of the untreated control heart (panel A), typicalautophagosomes containing intracellular organelles, such as

Table 1Hemodynamic variables.

Baseline Reperfusion (min)

30 60 120

LVDP (mmHg)CTL-30 109711 3978 32712 2679SEVO-30 103713 6179a 5878a 5078a

CTL-45 108713 2176 25713 2479SEVO-45 101710 23711 24710 2379CAPþ45 109711 2177 2176 2176SEVO-45þCAP 101714 2175 2275 2174CTL-45þ3MA 117712 1876 2077 1776SEVO-30þ3MA 97714 44716b 47717 44717SEVO-45þ3MA 95710 23715 24714 20716CAP-45þ3MA 105715 25711 23710 20710SEVO-45þCAPþ3MA 97711 1979 26716 26716

LVEDP (mmHg)CTL-30 1070 47714 51718 55715SEVO-30 1070 2477a 2475a 2677a

CTL-45 1070 6777 57713 53714SEVO-45 1070 55713 50711 47712CAPþ45 1070 53710 48710 45710SEVO-45þCAP 1070 5376 51710 49711CTL-45þ3MA 1070 6878 66710 66711SEVO-30þ3MA 1070 2777a 2579a 2579a

SEVO-45þ3MA 1070 72719 62717 60719CAP-45þ3MA 1070 63712 57715 52719SEVO-45þCAPþ3MA 1070 59715 51715 44716

CF (mL/min)CTL-30 2876 2175 2175 2078SEVO-30 3177 2374 2275 2276CTL-45 2977 1878 18710 1879SEVO-45 28710 19711 17711 17711CAPþ45 2873 1975 1876 1877SEVO-45þCAP 3474 1875 1979 1979CTL-45þ3MA 3776 2078 1878 1878SEVO-30þ3MA 2978 1877 19711 20711SEVO-45þ3MA 2775 1375 1375 1175CAP-45þ3MA 3179 2079 1979 1979SEVO-45þCAPþ3MA 3179 1977 1977 1876

Data are presented as mean7S.D. LVDP¼ left ventricular developed pressure;LVEDP¼ left ventricular end-diastolic pressure; CF¼coronary flow; CTL¼control;CAP¼chloramphenicol; 3MA¼3-Methyladenine; SEVO¼sevoflurane.

a Po0.05 vs. CTL-30.b Po0.05 vs. SEVO-30 n¼8 for each group.

M. Shiomi et al. / European Journal of Pharmacology 724 (2014) 58–66 61

mitochondria and membrane-like structures (white arrows), wereapparent after sevoflurane administration (panel B: SEVO-45).Additional treatment with chloramphenicol (panel C: SEVO-45þCAP) furthermore increased the number of autophagosomes.Autophagosomes were strikingly diminished by the addition of 3-methyladenine (panel D: SEVO-45þCAPþ3MA).

3.5. Western blot analysis

In hearts pretreated with sevoflurane or chloramphenicol, myo-cardial expression of LC3-II was increased compared with controlhearts. The combination of sevoflurane and chloramphenicol furtherincreased this expression. These increased expressions of LC3-II were

abolished by 3-methyladenine (Fig. 4). The phosphorylation state ofAkt and GSK3β after treatment with sevoflurane, chloramphenicoland 3-methyladenine is illustrated by a representativeWestern blot inFig. 5A. Total Akt and GSK3βwere comparable in all samples. The ratioof phospho Akt to total Akt and phospho GSK3β to total GSK3β weresignificantly increased in SEVO-45 and CAP-45 compared with CTL-45(Fig. 5B). This increase was not caused by unequal loading of thewestern blot, as shown by the detection of α-tubulin. The combina-tion of sevoflurane and chloramphenicol enhanced these increasedexpressions in SEVO-45þCAP. Administration of 3-methyladenineabolished these enhanced expressions in SEVO-45þCAP. There wasno significant difference in expression of mitochondrial cyclophilin Dbetween CTL-45 and SEVO-45þCAP (Fig. 5C).

3.6. Mitochondrial calcein fluorescence

The mitochondrial calcein fluorescence values after treatmentwith Ca2þ were taken as the values from which any reduction influorescence was measured. Exposure of mitochondria to 50 μMCa2þ did not induce MPTP opening (Fig. 6A). However, exposure ofmitochondria to 150 μM Ca2þ induced MPTP opening, representedby a reduction in calcein fluorescence of �66.9713.0% in CTL-45.This Ca2þ-induced reduction in calcein fluorescence was attenuatedin the SEVO-45 and CAP-45 groups (�40.078.0%, �44.071.8% vs.CTL-45, Po0.05, respectively; Fig. 6B). This effect was enhanced bythe SEVO-45þCAP (�15.570.3% vs. SEVO-45 and CAP-45, Po0.05),which was then abrogated by treatment with 3-methyladenine(SEVO-45þCAP þ3MA: �74.677.8%) (Fig. 6B).

4. Discussion

This study demonstrated that induction of autophagy beforeischemia restores the loss of sevoflurane preconditioning cardio-protection with prolonged ischemic insult. This effect is associated

Fig. 2. Infarct size as a percentage of left ventricle in eleven groups. After 120 minreperfusion, infarct size was significantly reduced in SEVO-30 but not in SEVO-45compared to corresponding CTL. This cardioprotection of SEVO-30 was notabolished by 3-methyladenine. Treatment with chloramphenicol in SEVO-45restored the protection lost by prolonged ischemia in SEVO-45. Chloramphenicolalone did not limit infarct size after 45 min of ischemia. Restoration of cardiopro-tection achieved by chloramphenicol was abolished by 3-methyladenine. 3-methy-ladenine alone did not affect infarct size. Data are presented as mean7S.D.;nPo0.05 vs. CTL-30; †Po0.05 vs. SEVO-45. CTL, control; CAP, Chloramphenicol;3-MA, 3-Methyladenine; SEVO, sevoflurane (n¼8 for each group).

Fig. 3. Electron microphotographs of cardiomyocytes, showing pre-ischemic induction of autophagosomes by reagents. Autophagosomes were rarely seen in the untreatedcontrol heart (panel A), were apparently observed (white arrows) after treatment with SEVO-45 (panel B). Autophagosomes appear to be more abundant after additionaltreatment with chloramphenicol (SEVO-45þCAP) compared with SEVO-45 alone (panel C). Appearance of autophagosomes strikingly diminished after treatment with3-methyladenine (panel D). Nucl, nucleus. Bars, 1 μm. (A) CTL -45, (B) SEVO -45, (C) SEVO -45 +CAP, (D) SEVO -45 +CAP +3MA.

M. Shiomi et al. / European Journal of Pharmacology 724 (2014) 58–6662

with enhanced phosphorylation of GSK3β before prolonged ische-mia which results in elevation of the threshold of MPTP openingby Ca2þ overload, but not expression of mitochondrial cyclophilinD. There is a consensus that the use of volatile anesthetics duringperioperative period could be a promising strategy to reducemyocardial ischemia-reperfusion injury for patients at risk ofdeveloping cardiac complication. However, the beneficial effectsof volatile anesthetics are limited to a certain range of sustainedischemia. Thus, strategies that prolong the protective effect ofvolatile anesthetics during longer ischemic insults should be

explored. This study demonstrated for the first time that chlor-amphenicol treatment could be a novel strategy for this purpose.

Chloramphenicol is an antibiotic that has been clinicallyused for over 30 years. It is a cytochrome P450 monooxygenaseinhibitor which has been shown to inhibit mitochondrial proteinsynthesis, causing mitochondrial stress. Chloramphenicol has beenshown to induce autophagy and preconditioning-like cardiopro-tection, but the mechanisms underlying this protective effectremain to be established. Thus, we examined whether treatmentwith chloramphenicol affects expression of mitochondrial cyclo-philin D, which is a component of MPTP that controls mitochon-drial pore-dependent Ca2þ exchange (Elrod et al., 2010; Hausenloyet al., 2010), and is a critical mediator of cardioprotection byischemic preconditioning. There were no significant differences inexpression of mitochondrial cyclophilin D between CTL-45 andSEVO-45þCAP. This suggests that cyclophilin D is not involved incardioprotection by chloramphenicol. Treatment of this drug hasbeen shown to stimulate the autophagic transcript Atg 12 byinhibiting mitochondrial protein synthesis, which results in induc-tion of autophagy (Prigione and Cortopassi, 2007). This studyfound that pretreatment with chloramphenicol before ischemiaincreased expression of LC3-II which leads to formation of autop-hagosomes. This was confirmed by electron microscopy.

McCormick et al. demonstrated that enhancing autophagyconfers protection against ischemia-reperfusion injury in cardio-myocytes (McCormick et al., 2012). Using in vivo swine models,Sala-Mercado et al. showed that induction of autophagy bychloramphenicol succinate before ischemia limited infarct size(Sala-Mercado et al., 2010). However, the precise mechanismsby which pre-ischemic induction of autophagy confers cardiopro-tection remain to be elucidated. In the present study, the autophagyantagonist 3-methyladenine abolished the chloramphenicol-inducerestoration of sevoflurane preconditioning after prolonged ischemic

Fig. 4. Representative western blot analysis of LC3-I, II from left ventricularsamples obtained before ischemia. The expression of LC3-II was slightly increasedin SEVO-45 and CAP-45 compared to CTL-45. The ratio of LC3-II to LC3-I (LC3-II/LC3-I) was further increased in SEVO-45þCAP compared to SEVO-45 and CAP-45.Administration of 3-methyladenine abolished this increased LC3-II/I in the SEVO-45þCAP group. nPo0.05 vs. CTL-45, †Po0.05 vs. SEVO-45 and CAP-45; CAP,Chloramphenicol; LC3-I, II, microtubule-associated protein light chain I and II;3-MA, 3-Methyladenine; SEVO, sevoflurane.

Fig. 5. (A) Representative western blot of phosphorylated GSK3β from left ventricular samples acquired before ischemia. Expression of phosphorylated GSK3β wassignificantly increased in the SEVO-45 and CAP-45 group. The combination of sevoflurane and chloramphenicol further increased this expression of phosphorylated GSK3βand Akt in the SEVO-45þCAP group. Administration of 3-methyladenine abolished this enhanced expression of phosphorylated GSK3β in the SEVO-45þCAP group.(B) Densitometric evaluation of four experiments as the x-fold increase in average light density vs. CTL-45. The results are presented as the ratio of the phosphorylation stateto total protein. The average light intensity was multiplied by 100 to facilitate the presentation of an x-fold increase. Data are presented as mean7S.D. nPo0.05 vs. CTL-45,†Po0.05 vs. SEVO-45, CAP-45 and SEVO-45þCAPþ3 MA. (C) Cytosolic and mitochondrial expression of cyclophilin D in CTL-45, CAP-45 and SEVO-45þCAP. There was nosignificant difference in expression between the three groups. CTL, control; CAP, Chloramphenicol; CyF D, cyclophilin D; GSK3β, glycogen synthase kinase-3; 3-MA,3-methyladenine; PHB, prohibitin; SEVO, sevoflurane.

M. Shiomi et al. / European Journal of Pharmacology 724 (2014) 58–66 63

insult, whereas it did not block the infarct size sparing effect ofSEVO-30. This suggests that induction of autophagy plays a role incardioprotection against ischemia-reperfusion injury after pro-longed ischemia. It has been demonstrated that induction ofautophagy during ischemia is protective whereas further enhance-ment of autophagy is detrimental during the reperfusion period(Matsui et al., 2008). Thus, it remains controversial whether autop-hagy is beneficial or harmful in myocardial ischemia-reperfusioninjury.

MPTP opening has been identified as a crucial determinantof ischemia-reperfusion injury (Hausenloy et al., 2009). MPTPopening has been also shown to induce autophagy (Gustafssonand Gottlieb, 2009). Elmore et al. demonstrated that autophagicstimulation by starvation caused an increase of spontaneouslydepolarizing mitochondria (presumably transient MPTP opening)in rat hepatocytes (Elmore et al., 2001). It is well documented thattransient MPTP opening before prolonged ischemia is a crucial

step to make the heart more tolerant to subsequent ischemia-reperfusion injury (Hausenloy et al., 2004; Saotome et al., 2009).We have previously demonstrated that exposure of sevoflurane tothe heart increases the threshold of MPTP opening by Ca2þ

loading (Onishi et al., 2012). In the present study, using isolatedcardiac mitochondria, we investigated whether administration ofchloramphenicol to sevoflurane-treated mitochondria enhancedinhibition of MPTP opening. Exposure of mitochondria to 150 μMCa2þ , not 50 μM Ca2þ , induced MPTP opening, represented bya reduction in calcein fluorescence in CTL-45. This reductionwas attenuated in CAP-45 and SEVO-45, and attenuated to aneven greater degree in SEVO-45þCAP. These data suggests thattransient MPTP opening by induction of autophagy with chlor-amphenicol increases the elevation of the threshold of MPTPopening by sevoflurane alone.

The threshold for MPTP opening has been shown to be elevated byphosphorylation of GSK3β at Ser9 through phosphatidylinositol 3-

Fig. 6. (A) Representative flow cytometric profile of isolated cardiac mitochondria loaded with calcein showing the effects of calcium (50 μM) on MPTP opening asdemonstrated by reductions in mitochondrial calcein fluorescence. Exposure of mitochondria to 50 μM Ca2þ did not induce MPTP opening in any group. (B) Exposure ofmitochondria to 150 μM Ca2þ induced MPTP opening in CTL. This Ca2þ-induced reduction in calcein fluorescence was attenuated in SEVO-45 and CAP-45. This effect wasenhanced by the combination of sevoflurane and chloramphenicol, which was then abrogated by treatment with 3-methyladenine. (C) Effect of calcium (150 μM) on MPTPopening as demonstrated by reductions in mitochondrial calcein fluorescence. Data are presented as mean7S.D. Percent change from control in the presence or absence of3-MA (10 μM). nPo0.05 vs. CTL-45; †Po0.05 vs. SEVO-45 and CAP-45 groups. CTL, control; CAP, Chloramphenicol; 3 MA, 3-Methyladenine; SEVO, sevoflurane. n¼4 foreach group.

M. Shiomi et al. / European Journal of Pharmacology 724 (2014) 58–6664

kinase (PI3K)/Akt signaling, which inactivates this kinase (Juhaszovaet al., 2004). Whether there is an association between autophagy andGSK3β phosphorylation in reducing ischemia-reperfusion injury is notknown. Recently, it has been reported that ultraviolet B-inducedautophagy activates AMP-activated protein kinase, an importantregulator of autophagy through the inhibition/phosphorylation ofGSK3β (Yang et al., 2012). Our data demonstrated that induction ofautophagy increased expression of phosphorylated GSK3β at Ser9,accompanied by increased Akt phosphorylation, which could lead toenhanced inhibition of MPTP opening. Both sevoflurane and chlor-amphenicol have been shown to activate PI3K/Akt signaling which isupstream of GSK3β (Li et al., 2010; Onishi et al., 2012). Class III PI3K isinvolved in formation of autophagosomes whereas class I PI3K inhibitsthe induction of autophagy via phosphorylation of Akt and themammalian target of rapamycin (mTOR) (Petiot et al., 2000). Further-more, there is evidence of self-regulation of autophagy by autophagy-induced inhibition of mTOR (He and Klionsky, 2009). Thus, theinteraction of Akt with mTOR is multifaceted and bidirectional.Nevertheless, our data suggest that autophagy and PI3K-Akt-mTORsignaling might yield additive benefit against ischemia-reperfusioninjury (Fig. 7). However, excessive autophagy is detrimental. Atpresent, it is hard to control autophagy. Further study is needed todetermine the optimal dose of chloramphenicol to induce clinicallybeneficial autophagy.

In contrast to the previous in vivo porcine study by Sala-Mercado et al. (2010), treatment with chloramphenicol alonedid not limit infarct size in the present study, despite the increa-sed expression of the autophagy-related protein LC3-II. However,chloramphenicol-induced autophagy did prolong the sevofluranepreconditioning cardioprotection into a longer ischemic insult.Although the reason for the discrepancy remains to be elucidated,differences in ischemia-reperfusion model and species could playa role.

The following study limitations should be acknowledged. First,GSK3β is a substrate of multiple pro-survival protein kinases,including ε-protein kinase C, extracellular signal-regulated kinase1/2 and PKG (Cohen and Frame, 2001). We cannot exclude thepossibility that these kinases other than PI3K/Akt signaling maycontribute to enhanced phosphorylation of GSK3β by

chloramphenicol. Second, the pharmacokinetics of 3-methyladenine are not well known, especially in high doses. Thedose (10 μM) of 3-methyladenine used in the present study was lowcompared to previous studies using 5 mM or 5 nM (McCormicket al., 2012; Yang et al., 2012). However, this dose of 3-methyladenine has been shown to effectively block LC3-II expres-sion in H9c2 cells as was seen in the present study (Cetrullo et al.,2012). Third, endothelial function was reported to be one of themediators for cardioprotection (Heusch et al., 2008). Novalija et al.demonstrated that sevoflurane mimics ischemic preconditioningeffects on nitric oxide release (Novalija et al., 1999). Chlorampheni-col was also reported to increase oxygen species such as nitric oxide(Holt and Bajoria, 1999). However, in a perfused heart model,coronary arteries are maximally dilated. Hence effects of sevoflur-ane and chloramphenicol on endothelial function are controlled for.Fourth, stimulation of autophagy with chloramphenicol reducedinfarct size but did not improve functional recovery. One possibleexplanation for this finding is myocardial stunning caused byprolonged global ischemia in a much greater area than the infarctedarea of LV.

In conclusion, induction of autophagy before ischemia restoresthe loss of sevoflurane preconditioning cardioprotection withprolonged ischemic insult. This effect is associated with enhancedphosphorylation of GSK3β before prolonged ischemia which isassociated with enhanced inhibition of MPTP opening. Givenrecent clinical evidence that the efficacy of sevoflurane precondi-tioning is lost after prolonged ischemia, induction of autophagy bychloramphenicol could be an additional strategy for reducingperioperative myocardial ischemia-reperfusion injury in at riskpatients.

Acknowledgments

We thank Akiko Tsujimoto (Gifu University Graduate School ofMedicine) for technical assistance. This study was supported byGrant-in-Aid for Scientific Research (C) 23593008 from the Min-istry of Education, Culture, Sports, Science and Technology of Japan(Masami Miyamae) (Tokyo, Japan).

References

Aviv, Y., Shaw, J., Gang, H., Kirshenbaum, L.A., 2011. Regulation of autophagy in theheart: you only live twice. Antioxid. Redox Signal. 14, 2245–2250.

Cetrullo, S., Tantini, B., Flamigni, F., Pazzini, C., Facchini, A., Stefanelli, C., Caldarera, C.M.,Pignatti, C., 2012. Antiapoptotic and antiautophagic effects of eicosapentaenoicacid in cardiac myoblasts exposed to palmitic acid. Nutrients 4, 78–90.

Cohen, P., Frame, S., 2001. The renaissance of GSK3. Nat. Rev. 2, 769–776.De Hert, S.G., 2006. Anesthetic preconditioning: how important is it in today0s

cardiac anesthesia? J. Cardiothorac. Vasc. Anesth. 20, 473–476.Elmore, S.P., Qian, T., Grissom, S.F., Lemasters, J.J., 2001. The mitochondrial

permeability transition initiates autophagy in rat hepatocytes. FASEB J. 15,2286–2287.

Elrod, J.W., Wong, R., Mishra, S., Vagnozzi, R.J., Sakthievel, B., Goonasekera, S.A.,Karch, J., Gabel, S., Farber, J., Force, T., Brown, J.H., Murphy, E., Molkentin, J.D.,2010. Cyclophilin D controls mitochondrial pore-dependent Ca(2þ) exchange,metabolic flexibility, and propensity for heart failure in mice. J. Clin. Invest. 120,3680–3687.

Fishbein, M.C., Meerbaum, S., Rit, J., Lando, U., Kanmatsuse, K., Mercier, J.C., Corday, E.,Ganz, W., 1981. Early phase acute myocardial infarct size quantification: valida-tion of the triphenyl tetrazolium chloride tissue enzyme staining technique. Am.Heart J. 101, 593–600.

Garcia, C., Julier, K., Bestmann, L., Zollinger, A., von Segesser, L.K., Pasch, T., Spahn, D.R.,Zaugg, M., 2005. Preconditioning with sevoflurane decreases PECAM-1 expres-sion and improves one-year cardiovascular outcome in coronary artery bypassgraft surgery. Br. J. Anaesth. 94, 159–165.

Gustafsson, Å.B., Gottlieb, R.A., 2009. Autophagy in ischemic heart disease. Circ. Res.104, 150–158.

Halestrap, A.P., Clarke, S.J., Javadov, S.A., 2004. Mitochondrial permeability transi-tion pore opening during myocardial reperfusion-a target for cardioprotection.Cardiovasc. Res. 61, 372–385.

Fig. 7. Schematic diagram of potential mechanisms of autophagy-induced cardio-protection conferred by chloramphenicol and sevoflurane. Phosphorylation ofGSK3β by PI3-Akt pathway inhibits MPTP opening. Class III PI3K is involved information of autophagosomes whereas class I PI3K inhibits the induction ofautophagy via phosphorylation of Akt and mTOR. Treatment with chloramphenicolor sevoflurane alone activates class III PI3K to induce autophagy. Simultaneoustreatment with chloramphenicol and sevoflurane enhances this pathway whichresults in enhanced phosphorylation of GSK3β and elevation of the threshold ofMPTP opening. PI3K¼phosphatidylinositol 3-kinase; Akt¼protein kinase B;GSK3β¼glycogen synthase kinase 3β; mTOR¼mammalian target of rapamycin;MPTP¼mitochondrial permeability transition pore.

M. Shiomi et al. / European Journal of Pharmacology 724 (2014) 58–66 65

Hausenloy, D., Wynne, A., Duchen, M., Yellon, D., 2004. Transient mitochondrialpermeability transition pore opening mediates preconditioning-induced pro-tection. Circulation 109, 1714–1717.

Hausenloy, D.J., Lim, S.Y., Ong, S.G., Davidson, S.M., Yellon, D.M., 2010. Mitochon-drial cyclophilin-D as a critical mediator of ischaemic preconditioning. Cardi-ovasc. Res. 88, 67–74.

Hausenloy, D.J., Ong, S.B., Yellon, D.M., 2009. The mitochondrial permeabilitytransition pore as a target for preconditioning and postconditioning. BasicRes. Cardiol. 104, 189–202.

He, C., Klionsky, D.J., 2009. Regulation mechanisms and signaling pathways ofautophagy. Annu. Rev. Genet. 43, 67–93.

Heusch, G., Boengler, K., Schulz, R., 2008. Cardioprotection: nitric oxide, proteinkinases, and mitochondria. Circulation 118, 1915–1919.

Holt, D.E., Bajoria, R., 1999. The role of nitro-reduction and nitric oxide in thetoxicity of chloramphenicol. Hum. Exp. Toxicol. 18, 111–118.

Huang, C., Yitzhaki, S., Perry, C.N., Liu, W., Giricz, Z., Mentzer Jr., R.M., Gottlieb, R.A.,2010. Autophagy induced by ischemic preconditioning is essential for cardio-protection. J. Cardiovasc Transl. Res. 3, 365–373.

Inamura, Y., Miyamae, M., Sugioka, S., Kaneda, K., Okusa, C., Onishi, A., Domae, N.,Kotani, J., Figueredo, V.M., 2009. Aprotinin abolishes sevoflurane postcondi-tioning by inhibiting nitric oxide production and phosphorylation of proteinkinase C-δ and glycogen synthase kinase 3β. Anesthesiology 111, 1036–1043.

Juhaszova, M., Zorov, D.B., Kim, S.H., Pepe, S., Fu, Q., Fishbein, K.W., Ziman, B.D.,Wang, S., Ytrehus, K., Antos, C.L., Olson, E.N., Sollott, S.J., 2004. Glycogensynthase kinase-3β mediates convergence of protection signaling to inhibitthe mitochondrial permeability transition pore. J. Clin. Invest. 113, 1535–1549.

Julier, K., da Silva, R., Garcia, C., Bestmann, L., Frascarolo, P., Zollinger, A., Chassot, P.G.,Schmid, E.R., Turina, M.I., von Segesser, L.K., Pasch, T., Spahn, D.R., Zaugg, M.,2003. Preconditioning by sevoflurane decreases biochemical markers for myo-cardial and renal dysfunction in coronary artery bypass graft surgery: a double-blinded, placebo-controlled, multicenter study. Anesthesiology 98, 1315–1327.

Kanamori, H., Takemura, G., Goto, K., Maruyama, R., Ono, K., Nagao, K., Tsujimoto, A.,Ogino, A., Takeyama, T., Kawaguchi, T., Watanabe, T., Kawasaki, M., Fujiwara, T.,Fujiwara, H., Seishima, M., Minatoguchi, S., 2011. Autophagy limits acutemyocardial infarction induced by permanent coronary artery occlusion. Am.J. Physiol. Heart Circ. Physiol. 300, H2261–2271.

Kaneda, K., Miyamae, M., Sugioka, S., Okusa, C., Inamura, Y., Domae, N., Kotani, J.,Figueredo, V.M., 2008. Sevoflurane enhances ethanol-induced cardiac precon-ditioning through modulation of protein kinase C, mitochondrial KATP chan-nels, and nitric oxide synthase, in guinea pig hearts. Anesth. Analg. 106, 9–16.

Kersten, J.R., Schmeling, T.J., Pagel, P.S., Gross, G.J., Warltier, D.C., 1997. Isofluranemimics ischemic preconditioning via activation of K(ATP) channels: reductionof myocardial infarct size with an acute memory phase. Anesthesiology 87,361–370.

Kevin, L.G., Katz, P., Camara, A.K., Novalija, E., Riess, M.L., Stowe, D.F., 2003.Anesthetic preconditioning: effects on latency to ischemic injury in isolatedhearts. Anesthesiology 99, 385–391.

Li, C.H., Cheng, Y.W., Liao, P.L., Yang, Y.T., Kang, J.J., 2010. Chloramphenicol causesmitochondrial stress, decreases ATP biosynthesis, induces matrix metalloproteinase-13 expression, and solid-tumor cell invasion. Toxicol. Sci. 116, 140–150.

Matsui, Y., Kyoi, S., Takagi, H., Hsu, C.P., Hariharan, N., Ago, T., Vatner, S.F.,Sadoshima, J., 2008. Molecular mechanisms and physiological significance ofautophagy during myocardial ischemia and reperfusion. Autophagy 4, 409–415.

McCormick, J., Suleman, N., Scarabelli, T.M., Knight, R.A., Latchman, D.S., Stephanou,A., 2012. STAT1 deficiency in the heart protects against myocardial infarction byenhancing autophagy. J. Cell. Mol. Med. 16, 386–393.

Novalija, E., Fujita, S., Kampine, J.P., Stowe, D.F., 1999. Sevoflurane mimics ischemicpreconditioning effects on coronary flow and nitric oxide release in isolatedhearts. Anesthesiology 91, 701–712.

Okusa, C., Miyamae, M., Sugioka, S., Kaneda, K., Inamura, Y., Onishi, A., Domae, N.,Kotani, J., Figueredo, V.M., 2009. Acute memory phase of sevoflurane precon-ditioning is associated with sustained translocation of protein kinase C-α and ε,but not δ, in isolated guinea pig hearts. Eur. J. Anaesthesiol. 26, 582–588.

Onishi, A., Miyamae, M., Kaneda, K., Kotani, J., Figueredo, V.M., 2012. Directevidence for inhibition of mitochondrial permeability transition pore openingby sevoflurane preconditioning in cardiomyocytes: comparison with cyclos-porine A. Eur. J. Pharmacol. 675, 40–46.

Petiot, A., Ogier-Denis, E., Blommaart, E.F., Meijer, A.J., Codogno, P., 2000. Distinctclasses of phosphatidylinositol 30-kinases are involved in signaling pathwaysthat control macroautophagy in HT-29 cells. J. Biol. Chem. 275, 992–998.

Prigione, A., Cortopassi, G., 2007. Mitochondrial DNA deletions and chlorampheni-col treatment stimulate the autophagic transcript ATG12. Autophagy 3,377–380.

Sala-Mercado, J.A., Wider, J., Undyala, V.V., Jahania, S., Yoo, W., Mentzer Jr., R.M.,Gottlieb, R.A., Przyklenk, K., 2010. Profound cardioprotection with chloramphe-nicol succinate in the swine model of myocardial ischemia-reperfusion injury.Circulation 122, S179–184.

Saotome, M., Katoh, H., Yaguchi, Y., Tanaka, T., Urushida, T., Satoh, H., Hayashi, H.,2009. Transient opening of mitochondrial permeability transition pore byreactive oxygen species protects myocardium from ischemia-reperfusioninjury. Am. J. Physiol. Heart Circ. Physiol. 296, H1125–1132.

Scarlatti, F., Granata, R., Meijer, A.J., Codogno, P., 2009. Does autophagy have alicense to kill mammalian cells? Cell Death Differ. 16, 12–20.

Sigaut, S., Jannier, V., Rouelle, D., Gressens, P., Mantz, J., Dahmani, S., 2009. Thepreconditioning effect of sevoflurane on the oxygen glucose-deprived hippo-campal slice: the role of tyrosine kinases and duration of ischemia. Anesth.Analg. 108, 601–608.

Yan, L., Vatner, D.E., Kim, S.J., Ge, H., Masurekar, M., Massover, W.H., Yang, G.,Matsui, Y., Sadoshima, J., Vatner, S.F., 2005. Autophagy in chronically ischemicmyocardium. Proc. Natl. Acad. Sci. USA 102, 13807–13812.

Yang, Y., Wang, H., Wang, S., Xu, M., Liu, M., Liao, M., Frank, J.A., Adhikari, S., Bower,K.A., Shi, X., Ma, C., Luo, J., 2012. GSK3β signaling is involved in ultravioletB-induced activation of autophagy in epidermal cells. Int. J. Oncol. 41,1782–1788.

M. Shiomi et al. / European Journal of Pharmacology 724 (2014) 58–6666