浙江大学 文本 Coronary microembolization does not induce acute preconditioning against infarction in pigs—the role of adenosine_生理学

Curditnuscular Research ELSEVIER Cardiewanrelar Iexearch 61 (J0M)-2 www.ebever.com/looorfcandiores Coronary microembolization does not induce acute preconditioning against infarction in pigs-the role of adenosine Andreas Skyschally",Rainer Schulz".Petra Gres",Ina Konietzka",Claus Martin". Michael Haude,Raimund Erbel,Gerd Heusch "Mstinve Paphyriologk.Zeunvu for Avere Madian LuhernirdrsMwhwu Erson./fafeludrme 55 f5/22 Eesee.Gianssny ha.nAe.wnwf0en包.furnurrblebwe Ecoe Hay点ntng华气f(节人.n四 Reo:ive 19 Janery 2004.novcivud is nevbed fom 26 Mund 2004;serepied 6 April 2004 Awibhle onlite 14 May 200M. Time for primary review 15 dnos Absiract Obfective:After coronary microebolizanon (ME)odenocine i released from ischemic areas ot'the microembolized uyocoedins This odenosine dilotes vessels in adjacent nonembolized myocardium and increases coronzry blood flow.For ichemic precoaditioning (P)lo pmlrel the myneseliim said infnsion,an inenease in the inerslilial arknoone oceoeneration (iAnO)prinr lo the subxgirnt ischemisrperfisires eeessary.We hypothesined tht the adenosine releae aner ME is sufficient to increse iADO d protoel the myocardium against infarction from subsequent ischemiareperfusion.We have theretoce compared myocurdial protection by either coronary microembolization or ichemie precoeditioning peice to ischemtiateperfusine Methods:In aeshetized pigs,the left atterior desceading (1AD)was cannulaed and perfused fom an extracorporeal circuit.In 11 pigs,sumtained ischewtia wis induced by 85%intlo nduution for 90 min (ccetroils).Two other gruups of pigs were sbjoclad cither te IP (e-8:10-min ischemis'l5-min repcrfusion)o aenm=9:ic mienssphenx.42m段.300=厂'.minw）price ln sisained ichemi Commorry venois aderosine concentration (vADO)and iADO (microdalysis)were measared.Infarct size wrs determied afer 2-h reperfusion by triphezyl tetrazolium chloride staining.Results:In pigs subjectod to IP,infaro:size was reduced to 2.6t 1.1%(mean t S.E.M.)vs.17.0t 3.2%in mnls.iAD0w3 sinered fmum24±L.3o13l±58um厂3dr亡gthe reperfusice following1 P.In pegs suljected to M压，aI10 nin aner ME,corcnary blod t3w3.6±3.6o536±43ln-1d￥AD0(025±004o0.4k±0.07mW-)were increas6d wewT,iAD0(20±0.5 at hascline vs2.3土0.6Imo-广I0 min afor ME五时nul incrcase.Infarct six induood by sustainod ischamix fullowing ME (22.5t 52%)wa ahove tha uf control fur amy given suhendocanlial bioud flow.Conclusiun:ME nelasad denosinet the vasculature dicreed coy blood low.The ilure ofiDOonrese with ME possibly explains teck of protection rast infarction after ME. 2004 Furope Society of Canlinlgy Publshed by Fsevier B.V.All righs reerved. Kern:Aderosire:Preoondnonng Microcircularion 1.Introduction identilicd as a potential cause of anhythmias.contractil dysfuneton and intaretlets 1. Atheroselerotie plaque rupture,cecumng cither sponta- In anesthetized dogs,coronary microembolzation with neously or duning coronary interventions,results in the polystyrene microspheres of 42-jm diameer induces an rekase of'ntheromatous andor thromboric material into the mmeline Bchemic eoetractile dystinction whach rocovers coranary circnlation which may then embolize the microvs- over several minutes.Susoquently,myocardial function is cular bed.Such coronary microembolization (ME)was progressively reduced whereas coronary hlond flow remains unchanged or is even increased (perfusicn-con- traction mismatch)2.3 The increase in coronary blood 4 Cuneponinga0x.Td:+4机20l-T234430:→4找-201-723 flow is attributed o vasodilation of adjacent nooembolized 4-481. vessels in response to acenosine relase from the ischemic E4 iY afreas perd heasch-c组d走G.kuhl microembolized myocardium 14.5).The progressive myo- 0008-6363/5 -xr fiuat ialler C 2004 Errupean Soicty uf Crdioligry Publoladl by Ebevier B.V.All righs roservudl. dor10.1016/carones 20M.0H4 001

Coronary microembolization does not induce acute preconditioning against infarction in pigs—the role of adenosine Andreas Skyschallya , Rainer Schulza , Petra Gresa , Ina Konietzkaa , Claus Martina , Michael Haudeb , Raimund Erbelb , Gerd Heuscha,* a Institut fu¨r Pathophysiologie, Zentrum fu¨r Innere Medizin, Universita¨tsklinikum Essen, Hufelandstraße 55, 45122 Essen, Germany bKlinik fu¨r Kardiologie, Zentrum Fu¨r Innere Medizin, Universita¨tsklinikum Essen, Hufelandstraße 55, 45122 Essen, Germany Received 19 January 2004; received in revised form 26 March 2004; accepted 6 April 2004 Available online 14 May 2004 Time for primary review 15 days Abstract Objective: After coronary microembolization (ME) adenosine is released from ischemic areas of the microembolized myocardium. This adenosine dilates vessels in adjacent nonembolized myocardium and increases coronary blood flow. For ischemic preconditioning (IP) to protect the myocardium against infarction, an increase in the interstitial adenosine concentration (iADO) prior to the subsequent ischemia/reperfusion is necessary. We hypothesized that the adenosine release after ME is sufficient to increase iADO and protect the myocardium against infarction from subsequent ischemia/reperfusion. We have therefore compared myocardial protection by either coronary microembolization or ischemic preconditioning prior to ischemia/reperfusion. Methods: In anesthetized pigs, the left anterior descending (LAD) was cannulated and perfused from an extracorporeal circuit. In 11 pigs, sustained ischemia was induced by 85% inflow reduction for 90 min (controls). Two other groups of pigs were subjected either to IP (n = 8; 10-min ischemia/15-min reperfusion) or coronary ME (n = 9; i.c. microspheres; 42 Am Ø; 3000
ml 1
min inflow) prior to sustained ischemia. Coronary venous adenosine concentration (vADO) and iADO (microdialysis) were measured. Infarct size was determined after 2-h reperfusion by triphenyl tetrazolium chloride staining. Results: In pigs subjected to IP, infarct size was reduced to 2.6 F 1.1% (mean F S.E.M.) vs. 17.0 F 3.2% in controls. iADO was increased from 2.4 F 1.3 to 13.1 F 5.8 Amol
l 1 during the reperfusion following IP. In pigs subjected to ME, at 10 min after ME, coronary blood flow (38.6 F 3.6 to 53.6 F 4.3 ml
min 1 ) and vADO (0.25 F 0.04 to 0.48 F 0.07 Amol
l 1 ) were increased. However, iADO (2.0 F 0.5 at baseline vs. 2.3 F 0.6 Amol
l 1 at 10 min after ME) did not increase. Infarct size induced by sustained ischemia following ME (22.5 F 5.2%) was above that of controls for any given subendocardial blood flow. Conclusion: ME released adenosine into the vasculature and increased coronary blood flow. The failure of iADO to increase with ME possibly explains the lack of protection against infarction after ME. D 2004 European Society of Cardiology. Published by Elsevier B.V. All rights reserved. Keywords: Adenosine; Preconditioning; Microcirculation 1. Introduction Atherosclerotic plaque rupture, occurring either sponta￾neously or during coronary interventions, results in the release of atheromatous and/or thrombotic material into the coronary circulation which may then embolize the microvas￾cular bed. Such coronary microembolization (ME) was identified as a potential cause of arrhythmias, contractile dysfunction and infarctlets [1]. In anesthetized dogs, coronary microembolization with polystyrene microspheres of 42-Am diameter induces an immediate ischemic contractile dysfunction which recovers over several minutes. Subsequently, myocardial function is progressively reduced whereas coronary blood flow remains unchanged or is even increased (perfusion –con￾traction mismatch) [2,3]. The increase in coronary blood flow is attributed to vasodilation of adjacent nonembolized vessels in response to adenosine release from the ischemic microembolized myocardium [4,5]. The progressive myo- 0008-6363/$ - see front matter D 2004 European Society of Cardiology. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.cardiores.2004.04.003 * Corresponding author. Tel.: +49-201-723-4480; fax: +49-201-723- 4481. E-mail address: gerd.heusch@uni-essen.de (G. Heusch). www.elsevier.com/locate/cardiores Cardiovascular Research 63 (2004) 313 – 322

314 A.Shrurkally et a!.Candsnneatir Prreonci di (20H)31-12 cardial contractile dysfunction in the presence of ronal or of Laboratory Animals puhlished by the US National increased blood flow following microemboliztion is asso. Institutes of Health (NIH Publication No.85-23,revised ciatod with an inllammalory response,characterized by 19961 inereased leukocyte mnfiltrion and inereasod tumor necro- s8 Bclor-(TNFo）concentralion韩hin the myocardium 2.J.Experimemul prepurlio 12]TNFa is causally involved in the progressive contrac- tile dysfimetion.sinee intracoranary infision of TNFo Thirty-six Gottinger minipegs (20-40 kg)of either sex memicked contractile dysfunction an the ahsence of micro- were nsed in our experiments.The pigs were fed ad embolization and contrnetile dysfunction following micro- libitum hefore the experiments.Afer initial sedation emholization was prevented by pretreatment with TNFer- ing ketamine hydrochlride (I intmmuseulrly)the antibodies [6].The inflammatory signal cascade also pigs were anestbetized with thiopental (Trapanal,500 mg involves nitric oxide upstream of TNFo and sphingosine intraverously).Thruugh a midline cervical incision,the downstream of TNFo「7l. trachea was intubated for connection to a respirator Afler curonary microembolization.aderosine is re- (Driger.Labock.Germany).Aresthesia was then main- leased liom small ischemic areas in the microemboliaed tained using enflurane (1-1.5%)with an oxygen/nitrous myocardium.This adenosine dilates vessels in the adja- oxide mixture (4%60%).Arlerial blood gases were ont non-embolizcd myocardium and inereases myocardial monitored frxquenly in the initial stas of'the prepara- blood Dow 14.5J.Aderosine is an importan lrigger ol tion until stabl and then periodically throughout the ischemic preconditioning (IP)in mny species including study (Radiometer,Copenhagen,Denmark).Rectal tem. pigs [8]and also humans [9].In the previcus experiments perature was monitored and maintained between 37 and from our laboratory,we have shown that in anesthetized 38 "C by the use of a heated surgical table and drapes. pigs,ane cycle of ischemic preconditioning by 10 min The cammon carotid arteries were carmulated with poly low-flow ischemia and 15-min reperfusion protects the ethylene catheters.oce to measure arterial pressure and myocardium from the conseguences of subsequent 90-min the other to supply blood to the extracorporeal circuit. low-flow ischemia and 120-min reperfusicn and roduces The jugular veins were cannulated for volume replace- infaret size [10.An increase in the inerstitial adenosine ment using warmed 0.9%NaCl and fur the return of concentration (iADX)prior to the sustained ischemia is blood to the animal from the coronary venous line. necessary to establish this protection against infarction A left lateral thorncotomy was performed in the [10,11] fourth intercostal space ad the pericardium opened.A To take ndvantage of our estahlished model of ischemic micromanometer (P7,Konigsherg Instr,Pasadena,CA, preconditioning in pigs [10,11],we have now transferred USA)was placed in the left ventricle through the our microembolization madel from anesthetized dogs to apex together with a saline-filled polyethylne catheter (used aesthetized pigs,and we bypothesized that the adenosine to calibrate the micromanometer in situ).Ultrasanic dimen- release following an episode of coconary microemboliza sion gauges were implanted in the left ventricular (LV) loon s sunlicent to increase the interslitial adenosine myccandium to measure the thickness of'the anenor smd coneentration and to proteet the surroumding myoenrdnm posterior (control)wall.The left anterior descending (LAD) st ischemia/reperfusion-induced infrction.We have coronary artery was dissected over a distance of 1.5 em, therefore compared myocardial protection by either coro- ligated,canmulaed,and perfsod from mm cxtracorporenl nary microemholization or ischemic preeceditioning prior circnit.Prior to comnary cannulation,the pips were anti- to a 901-min low-flow ischemia followed hy 120-min coapulared with 20,0100 IU sodium heparin:additional dases reperfusion.To assess myocardial protection infarct size of 10,000 IL were given m hourly imervals.The system resulting from %-min ischemia/120-min reperfusion was included a roler pump,windkessel,and a side-por for the determined,and the role of adenosine was evaluated by injertion of micmospheres.Coronary arterial pressure was measuring arterial.coronary venous,and interstitial (micro- measured from the sidearm of a polyethylene T-connector dialysis]adenosine concentrations. (Cole-Parmer.Chicago,IL.USA)used as catheter tip with an In our cxperiments,coreary micruembolization did not external transducer (pvb Medizimechnik.Kirchseon.Ger- ptoteet against imlaretion,possibly due to a lack of inerese many).Minimal coronary arterial pressare was held above 7 in interstitial sdenosine prior to the sustained ischemia. mm Hg by adjusting the roller pump of the extrncorporeal which is scen with ischemic preconditioning circuit to avoid hypoperfusion pior to ischemia.The lnage epicardial vein parallel to the LAD coronary artery was dissected and cannulated,Coronary venous hlood was 2.Methods drained to an unpressurized reservoir and then returned to a jugular vein through use of a second roller pump. The experimental protocols were approved by the Bio- Heart rate was controlkd throughout the study by left ethical Commitice of the district of Dusseldorf.and the atrial pacing (Hugo Sachs Elektrunik Type 215/T.Hug- investigation conforms with the Guide for the Care and Use sletten,.Gea）slightly abov地凸e spontancous rate

cardial contractile dysfunction in the presence of normal or increased blood flow following microembolization is asso￾ciated with an inflammatory response, characterized by increased leukocyte infiltration and increased tumor necro￾sis factor-a (TNFa) concentration within the myocardium [2]. TNFa is causally involved in the progressive contrac￾tile dysfunction, since intracoronary infusion of TNFa mimicked contractile dysfunction in the absence of micro￾embolization and contractile dysfunction following micro￾embolization was prevented by pretreatment with TNFa￾antibodies [6]. The inflammatory signal cascade also involves nitric oxide upstream of TNFa and sphingosine downstream of TNFa [7]. After coronary microembolization, adenosine is re￾leased from small ischemic areas in the microembolized myocardium. This adenosine dilates vessels in the adja￾cent non-embolized myocardium and increases myocardial blood flow [4,5]. Adenosine is an important trigger of ischemic preconditioning (IP) in many species including pigs [8] and also humans [9]. In the previous experiments from our laboratory, we have shown that in anesthetized pigs, one cycle of ischemic preconditioning by 10-min low-flow ischemia and 15-min reperfusion protects the myocardium from the consequences of subsequent 90-min low-flow ischemia and 120-min reperfusion and reduces infarct size [10]. An increase in the interstitial adenosine concentration (iADO) prior to the sustained ischemia is necessary to establish this protection against infarction [10,11]. To take advantage of our established model of ischemic preconditioning in pigs [10,11], we have now transferred our microembolization model from anesthetized dogs to anesthetized pigs, and we hypothesized that the adenosine release following an episode of coronary microemboliza￾tion is sufficient to increase the interstitial adenosine concentration and to protect the surrounding myocardium against ischemia/reperfusion-induced infarction. We have therefore compared myocardial protection by either coro￾nary microembolization or ischemic preconditioning prior to a 90-min low-flow ischemia followed by 120-min reperfusion. To assess myocardial protection infarct size resulting from 90-min ischemia/120-min reperfusion was determined, and the role of adenosine was evaluated by measuring arterial, coronary venous, and interstitial (micro￾dialysis) adenosine concentrations. In our experiments, coronary microembolization did not protect against infarction, possibly due to a lack of increase in interstitial adenosine prior to the sustained ischemia, which is seen with ischemic preconditioning. 2. Methods The experimental protocols were approved by the Bio￾ethical Committee of the district of Du¨sseldorf, and the investigation conforms with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996). 2.1. Experimental preparation Thirty-six Go¨ttinger minipigs (20 –40 kg) of either sex were used in our experiments. The pigs were fed ad libitum before the experiments. After initial sedation using ketamine hydrochloride (1 g intramuscularly), the pigs were anesthetized with thiopental (Trapanal, 500 mg intravenously). Through a midline cervical incision, the trachea was intubated for connection to a respirator (Dra¨ger, Lu¨beck, Germany). Anesthesia was then main￾tained using enflurane (1 –1.5%) with an oxygen/nitrous oxide mixture (40%:60%). Arterial blood gases were monitored frequently in the initial stages of the prepara￾tion until stable and then periodically throughout the study (Radiometer, Copenhagen, Denmark). Rectal tem￾perature was monitored and maintained between 37 and 38 jC by the use of a heated surgical table and drapes. The common carotid arteries were cannulated with poly￾ethylene catheters, one to measure arterial pressure and the other to supply blood to the extracorporeal circuit. The jugular veins were cannulated for volume replace￾ment using warmed 0.9% NaCl and for the return of blood to the animal from the coronary venous line. A left lateral thoracotomy was performed in the fourth intercostal space and the pericardium opened. A micromanometer (P7, Konigsberg Instr., Pasadena, CA, USA) was placed in the left ventricle through the apex together with a saline-filled polyethylene catheter (used to calibrate the micromanometer in situ). Ultrasonic dimen￾sion gauges were implanted in the left ventricular (LV) myocardium to measure the thickness of the anterior and posterior (control) wall. The left anterior descending (LAD) coronary artery was dissected over a distance of 1.5 cm, ligated, cannulated, and perfused from an extracorporeal circuit. Prior to coronary cannulation, the pigs were anti￾coagulated with 20,000 IU sodium heparin; additional doses of 10,000 IU were given at hourly intervals. The system included a roller pump, windkessel, and a side-port for the injection of microspheres. Coronary arterial pressure was measured from the sidearm of a polyethylene T-connector (Cole-Parmer, Chicago, IL, USA) used as catheter tip with an external transducer (pvb Medizintechnik, Kirchseon, Ger￾many). Minimal coronary arterial pressure was held above 70 mm Hg by adjusting the roller pump of the extracorporeal circuit to avoid hypoperfusion prior to ischemia. The large epicardial vein parallel to the LAD coronary artery was dissected and cannulated. Coronary venous blood was drained to an unpressurized reservoir and then returned to a jugular vein through use of a second roller pump. Heart rate was controlled throughout the study by left atrial pacing (Hugo Sachs Elektronik Type 215/T, Hug￾stetten, Germany) slightly above the spontaneous rate. 314 A. Skyschally et al. / Cardiovascular Research 63 (2004) 313–322

A.Shynroaly e al /Cmndnouralor Rroarh 3 (2 14 2.2.Regioal myocardial bland flow Germany)and 0.1 mM o.B-methylene-adenosing-5-diphos- phmte-solution (Sigma,Taufkirchen,Gemany)stopping the Radiolabeled microspheres (15 um in diameter:Ce. uptake and breskdown of adenosine by crythrocytes [12] Cr.10Ru.Nb.or 4Sc;NEN-DuPonl.Boston,USA) and immediately centrifuged.The plasma was colledlod and were injecled inlo the coronary perliusion circuit lo detet- Utea-70“C. mine the regional myocardial blood now and its distribrtion The measurement of interslitial idenosine concenra- theoughout the LAD perfusion bed (model 5912.Gammas- tion was performed by microdialysis technique.For in zint BF 5300 Packand.Germany). vivo samplig of'dialysate aliguots.the microdialysis probes (CMA/220.Bioanalytienl Systems,West Lafayette 2主.farphoiog USA)were insened into the myceardium within the perfusion bed of the left anterior descending corceary At the end of each stu山y,e heart w部emoved and artery.sach that a 10-mm-long segment containing a sectioced from base to apex into five transverse slices in a dialysis membrane with a molecular cutoff of 20.000 D plane parallel to the atrioventricular groove.The slices spanned the entire transmural wall.Afer insertion.the were immersed in 009 mol-sodium phosphate buffer inflow tube of the probe was conneclod to a gas-tight (pH 7.4)containing 1.0%tripbenyl tetrazolium chkride plastic syringe filled with a modified Krebs-Henseleit (Sigma-Aldrich Chemic.Munich,Giermany)and 8%dex- bufler comsisting of (in mM)118 NaCl.4.7 KCT,1.2 tran (MW 77.800)for 20 min at 37 "C to identify inlardod MgSO4.12 KH:PO4.130 CaCl.25.0 NaHCO:.5.0 tissue.The amount of infarcted tissue is expressed as glucose.The pl was adjusted to 7.4 with NaOH.The percent of the LV area at risk,as determined by the probe was perfused at 2.82 ulmin with a modified microspheres technque [10]. syringe infusion pump (Perfusor ED2.Braun Melsungen, Melsungen,Germany),and the effluent was sampled into 2.4.Adenosine capped microcentrifuge tubes.Each probe was perfused for30mi▣prior to凸e start of1 prococol and these Plasma adenosine concentrations were measured in si- samples were discarded.Then,beginming 10 min before maltaneously drawn coronary venous and arterial blood the first ischemic episode,the efflent was collected in samples.Samples were taken at baseline in all grouns,ar 10-min fmetions thnsghout the whole experiment and 10 min of the reperfision following ischemie precoodition- the samples were immedimely frozen and stored in liquid ing in rroup 2 and at 10 min after coronnry microemboli- nitrogen.The adenosine concentration in the collected zation in group 3.At 5 and 85 min of the sustained dialysate samples was stable,even when stored at room ischemia,samples were again taken in all groups The hlood temperature for several hours. samples were withdrawn with syringes containing ice-cold Adenosine concentrations in arterial and venous plasma dipyridamole solution (0.2 mM,Dr.Karl Thomae,Biberach. and in the dialysate collected from the microdialysis probes Au如1eaw rhdTn 10 11

2.2. Regional myocardial blood flow Radiolabeled microspheres (15 Am in diameter; 141Ce, 51Cr, 103Ru, 95Nb, or 46Sc; NEN-DuPont, Boston, USA) were injected into the coronary perfusion circuit to deter￾mine the regional myocardial blood flow and its distribution throughout the LAD perfusion bed (model 5912, Gammas￾zint BF 5300 Packard, Germany). 2.3. Morphology At the end of each study, the heart was removed and sectioned from base to apex into five transverse slices in a plane parallel to the atrioventricular groove. The slices were immersed in 0.09 mol
l 1 sodium phosphate buffer (pH 7.4) containing 1.0% triphenyl tetrazolium chloride (Sigma-Aldrich Chemie, Munich, Germany) and 8% dex￾tran (MW 77,800) for 20 min at 37 jC to identify infarcted tissue. The amount of infarcted tissue is expressed as percent of the LV area at risk, as determined by the microspheres technique [10]. 2.4. Adenosine Plasma adenosine concentrations were measured in si￾multaneously drawn coronary venous and arterial blood samples. Samples were taken at baseline in all groups, at 10 min of the reperfusion following ischemic precondition￾ing in group 2, and at 10 min after coronary microemboli￾zation in group 3. At 5 and 85 min of the sustained ischemia, samples were again taken in all groups. The blood samples were withdrawn with syringes containing ice-cold dipyridamole solution (0.2 mM, Dr. Karl Thomae, Biberach, Germany) and 0.1 mM a,h-methylene-adenosine-5-diphos￾phate-solution (Sigma, Taufkirchen, Germany) stopping the uptake and breakdown of adenosine by erythrocytes [12] and immediately centrifuged. The plasma was collected and stored at 70 jC. The measurement of interstitial adenosine concentra￾tion was performed by a microdialysis technique. For in vivo sampling of dialysate aliquots, the microdialysis probes (CMA/220, Bioanalytical Systems, West Lafayette, USA) were inserted into the myocardium within the perfusion bed of the left anterior descending coronary artery, such that a 10-mm-long segment containing a dialysis membrane with a molecular cutoff of 20,000 D spanned the entire transmural wall. After insertion, the inflow tube of the probe was connected to a gas-tight plastic syringe filled with a modified Krebs-Henseleit buffer consisting of (in mM) 118 NaCl, 4.7 KCl, 1.2 MgSO4, 1.2 KH2PO4, 1.30 CaCl2, 25.0 NaHCO3, 5.0 glucose. The pH was adjusted to 7.4 with NaOH. The probe was perfused at 2.82 Al
min 1 with a modified syringe infusion pump (Perfusor ED2, Braun Melsungen, Melsungen, Germany), and the effluent was sampled into capped microcentrifuge tubes. Each probe was perfused for 30 min prior to the start of a protocol and these samples were discarded. Then, beginning 30 min before the first ischemic episode, the effluent was collected in 10-min fractions throughout the whole experiment and the samples were immediately frozen and stored in liquid nitrogen. The adenosine concentration in the collected dialysate samples was stable, even when stored at room temperature for several hours. Adenosine concentrations in arterial and venous plasma and in the dialysate collected from the microdialysis probes Fig. 1. Representative chromatogram of adenosine separation by HPLC in a microdialysis sample. The arrow indicates the adenosine peak which is magnified in the inset. The adenosine peak was removed by adding adenosine deaminase (gray tracing) and no interfering peaks were detectable. A. Skyschally et al. / Cardiovascular Research 63 (2004) 313–322 315

316 A.uckoly rt以/Candmaoealr Prreani行20403l-2 were determined by HPLC as described previcusly [11].A TaNe 1 typicul chromatogram of the HPLC separation is given in Myocaal poumeruts of TNFut atd年i由bine tl maber ul Fie.1. infitrating laikncyie fh after erimertal micmemhelizrirn Microereholznd Nar-errhelinad 25.Experimente!mieremilicio in aneefized pigy asterior wall pooterior wall NFa IpaˉWm 1246+266码 658±125 Sphiigmine [pMol-WW] 30±3 136±0 Our pror studies on coronary micrcemholization were Leukocyes Imm 112士3 Tw±5 done in dogs [2,3,7)whereas those on ischemie precondi- vs.posterior wall (doube-skded parired mes) tioning were done in pigs [10.11).We therefore perfommed four control experiments with coronary microembolization in pigs to assure the transfer of the model fiom dogs to pigs. number of loukocytes infiltrating the anterior wall was Coronary microcmbolization was induced by injecting greater than that in the posterior wall (Table 1 Coronary 3000 white-colorod polystyrene microspheres (diameter 42 blood Dow was increased at 10 min alber the microcmbo- um,Dynospheres,Dyno Particks.Lillestrom.Norway)per lization.returned back towards baseline values at 1 h.bul ml-min-coromary inllow into the perliusion system. was inereased again aller 4 h.probably due to hemodilu- During microembolization,coronary inllow was held con- iot(H:107±2g-厂1a1tine8.74土5g-厂'6h) stant.Thereall,coronary perlision pressure was main- (Fig 2 Sham pigs (o-4)atter 6 h have similar inereases tained at baseline levels for 6 h.At the end of the in coroeary blood flow (40.0t3.8 mlmin at baseline expenment,biopsies from the amterior and posterior wall and 62.0+9.3 ml-min'at 6 h)and hemodilution (TIb: were taken and aralyzed for tissue concentrations of TNFo 106+6 g-l at baseline vs.686 gl at 6 h),but [13]and sphingosine [7]In transmural histological sec nchanged wall thickening (46.5 3.4%at baseline vs. tions,the amount of microinfarction was determined [14 40.6+4.3%at61. ad TNFa was visualized by immunostaining [15.The consequenoes of coronary microembolization in pigs were almost identical to thuse previously observed in dogs 3.Experimental protocols Immediately upon comonary micmembolization,fimction of the microembolized myocardium rapidly decreased. Group I (n=11::Following baseline measurements of Afer an initial partial recovery,which occurred within systemic bemodynamics,regional myocardial function and minutes,function progressively decreased over the follow. blocd flow,corocary inflow was redaced to achieve a 85- ing 6 h (Fig.2)Myocardial coexentrations of TNFa 90%reduction in anterior myocardial systolic wall thick. (Tahle I and Fig.3)and sphingosine (Table 1)were higher ening.At 5.and 85-min ischemia,measurements were in the anterior than in the postericr wall at 6 h after repeated.After 90 min ischemia,the myocardium was coronary microemholization.Small necrotic foci were reperfused far 2 h befure infarct size was determined by present in the mnterior wall (4.5+2.4%in total),and the TTC staining. 90 60 70 60 50 40 20 bodline 3-6 min 10 min 1h 2h 3 h 4h 5h 6h po&ME Fit 2.Prugreoive tryocanlal eyoluacliut afler cueotary aioctbuliclion Th:urow isdictles th:injodtion of cibulidt ticttoplns.Cuoutary bloul progressively decreased following soronary mcroembolizatior.Caronary bloed fw (-iseased aner nicembaliaorion

were determined by HPLC as described previously [11]. A typical chromatogram of the HPLC separation is given in Fig. 1. 2.5. Experimental microembolization in anesthetized pigs Our prior studies on coronary microembolization were done in dogs [2,3,7], whereas those on ischemic precondi￾tioning were done in pigs [10,11]. We therefore performed four control experiments with coronary microembolization in pigs to assure the transfer of the model from dogs to pigs. Coronary microembolization was induced by injecting 3000 white-colored polystyrene microspheres (diameter 42 Am, Dynospheres, Dyno Particles, Lillestrøm, Norway) per ml
min 1 coronary inflow into the perfusion system. During microembolization, coronary inflow was held con￾stant. Thereafter, coronary perfusion pressure was main￾tained at baseline levels for 6 h. At the end of the experiment, biopsies from the anterior and posterior wall were taken and analyzed for tissue concentrations of TNFa [13] and sphingosine [7]. In transmural histological sec￾tions, the amount of microinfarction was determined [14] and TNFa was visualized by immunostaining [15]. The consequences of coronary microembolization in pigs were almost identical to those previously observed in dogs. Immediately upon coronary microembolization, function of the microembolized myocardium rapidly decreased. After an initial partial recovery, which occurred within minutes, function progressively decreased over the follow￾ing 6 h (Fig. 2). Myocardial concentrations of TNFa (Table 1 and Fig. 3) and sphingosine (Table 1) were higher in the anterior than in the posterior wall at 6 h after coronary microembolization. Small necrotic foci were present in the anterior wall (4.5 F 2.4% in total), and the number of leukocytes infiltrating the anterior wall was greater than that in the posterior wall (Table 1). Coronary blood flow was increased at 10 min after the microembo￾lization, returned back towards baseline values at 1 h, but was increased again after 4 h, probably due to hemodilu￾tion (Hb: 107 F 2 g
l 1 at baseline vs. 74 F 5 g
l 1 at 6 h) (Fig. 2). Sham pigs (n = 4) after 6 h have similar increases in coronary blood flow (40.0 F 3.8 ml
min 1 at baseline and 62.0 F 9.3 ml
min 1 at 6 h) and hemodilution (Hb: 106 F 6 g
l 1 at baseline vs. 68 F 6 g
l 1 at 6 h), but unchanged wall thickening (46.5 F 3.4% at baseline vs. 40.6 F 4.3% at 6 h). 3. Experimental protocols Group 1 (n = 11): Following baseline measurements of systemic hemodynamics, regional myocardial function and blood flow, coronary inflow was reduced to achieve a 85– 90% reduction in anterior myocardial systolic wall thick￾ening. At 5- and 85-min ischemia, measurements were repeated. After 90-min ischemia, the myocardium was reperfused for 2 h before infarct size was determined by TTC staining. Fig. 2. Progressive myocardial dysfunction after coronary microembolization. The arrow indicates the injection of embolizing microspheres. Coronary blood flow was held constant during injection of the embolizing microspheres. Systolic wall thickening of the anterior wall ( ) following an initial rapid recovery progressively decreased following coronary microembolization. Coronary blood flow ( ) increased after microembolization. Table 1 Myocardial concentrations of TNFa and sphingosine and number of infiltrating leukocytes 6 h after experimental microembolization Microembolized anterior wall Non-embolized posterior wall TNFa [pg
g 1 WW] 124.6 F 26.6* 65.8 F 12.6 Sphingosine [pMol
g 1 WW] 330 F 88* 136 F 30 Leukocytes [mm 2 ] 112 F 3* 78 F 3 * p < 0.05 vs. posterior wall (double-sided paired t-test). 316 A. Skyschally et al. / Cardiovascular Research 63 (2004) 313–322

A.Shymtally et al /Canduouraler Reoarei 63 (0 3/-12 1行 antenor wall posterior wall Fig.3 TNFu in myuediem and vaeulalure.Imrchinlegical staining odtiote dipes frum the sntite and he powre wall.In the arlerioe wall distitt 635o时四g3t0nneT的we中门Mo1nr1N:n09g0nr%7力ire is cortned to th0和安r川ang ber 20 pint Group 2 (n-8):Following baseline measurements of tween subendocardial blood flow and infarct size were systemic hemodynamics.regicnal myocardial function and compared by analysis of covariance (ANCOVA).A p blood flow,the myccardium was subjected to one cycle of value less than 0.05 was taken to indcate a.significant 10 min prococditioning ischemia.with a 85-90%roduction difference. in regional myocardial function,followed by 15-min reper fusion.During reperfusion.corooary arterial pressune was maintained at the level measured prior to ischemin by 4.Results continuously ndapting coary inflow with the roller pump Following reperfusion,coronary inflow was again reduced With the onset of the preconditioning ischemia in group to the same level as during the precorditioning ischemia 2 and the sustained ischemia in all groups,regional myo. Thereafier,the protocol of group 2 was identical to that of cardial function in the anterioe wall was significantly re. group I. duced (Table 21. Group 3 (n-9:Following baseline measurements of Anterior systolic wall thickening was slightly decreased systemic hemodynamics,cornary microembolization was at 10 min after injection of the embolizing particles indueed by injecting 3000 white-colored polystyrene (317±5,8%3.28,1±38%n.&1Cw0 ary bleod0w microspheres (dismeter 42 um.)per ml.min-coronary at 10 min ater microcinboliation was inereased from inflow into the perfision system.Durng reactive hyper- 38.6±3.6m53.6±43 ml.min-1(n<05)ni8pk emia,coroary perfuasion pressuare was maintained constant flmf59.5±4.3 ml.min-1nt3.2±0.4 min after m haseline levels.Ten minutes after micmembolization microcmbolizarion. coronary inflow was reduced to n similar extent as in The area at risk was comparable berween the groups group 1.Thereafter,the protocol of group 3 was identical (Fig.4)In group I following 90-min severe myocardial to that of group 1. ischemia and 120-min reperfusion,infirct size averaged 17.0±3.2%【schemic precondit6 oning by10 min ische- 3.1.Dmla annbsis and statirtics mia and 15-min reperfision reduced inlarct size follow- ing 90-min ischemia and 120-min reperfusion to Data are reportod as memn valus+S.E.M.Systemic 2.6 1.1%(p<0.05 vs.groups 1 and 3)and the hemodyname dala and the interstitial knosine concen- teationship belween infaret sine and subendocardial tration were amalyzed by two-way ANOVA.and suberda- blood now was shifted downwards,with a smaller infret cardal blod flow during ischemia,anea a nsk,and infaret size at a given subendocardial blood fow (p<D05 size were anslyzed by one-way ANOVA.Fisher's LSD ANCOVA:Fig 5) tests were used for post-boc comparisons when overall With micruembolizaticn in group 3.infarct size follow significant differences were detected.The time courses of ing 90-min ischemia and 120-min reperfusion was corunary verous plasma adenosine concentration were 22.5+5.2%.Compared to 90-min sustained iscbemia compared to those of the arterial plasma adenosine con- aane,凸e relationship betwe知nfarct s2 e and su止endocar centration by two-way ANOVA.Linear regressions be- dial blood flow was shifted upwards towards higher infarct

Group 2 (n = 8): Following baseline measurements of systemic hemodynamics, regional myocardial function and blood flow, the myocardium was subjected to one cycle of 10 min preconditioning ischemia, with a 85– 90% reduction in regional myocardial function, followed by 15-min reper￾fusion. During reperfusion, coronary arterial pressure was maintained at the level measured prior to ischemia by continuously adapting coronary inflow with the roller pump. Following reperfusion, coronary inflow was again reduced to the same level as during the preconditioning ischemia. Thereafter, the protocol of group 2 was identical to that of group 1. Group 3 (n = 9): Following baseline measurements of systemic hemodynamics, coronary microembolization was induced by injecting 3000 white-colored polystyrene microspheres (diameter 42 Am,) per ml
min 1 coronary inflow into the perfusion system. During reactive hyper￾emia, coronary perfusion pressure was maintained constant at baseline levels. Ten minutes after microembolization, coronary inflow was reduced to a similar extent as in group 1. Thereafter, the protocol of group 3 was identical to that of group 1. 3.1. Data analysis and statistics Data are reported as mean values F S.E.M. Systemic hemodynamic data and the interstitial adenosine concen￾tration were analyzed by two-way ANOVA, and subendo￾cardial blood flow during ischemia, area at risk, and infarct size were analyzed by one-way ANOVA. Fisher’s LSD tests were used for post-hoc comparisons when overall significant differences were detected. The time courses of coronary venous plasma adenosine concentration were compared to those of the arterial plasma adenosine con￾centration by two-way ANOVA. Linear regressions be￾tween subendocardial blood flow and infarct size were compared by analysis of covariance (ANCOVA). A p value less than 0.05 was taken to indicate a significant difference. 4. Results With the onset of the preconditioning ischemia in group 2 and the sustained ischemia in all groups, regional myo￾cardial function in the anterior wall was significantly re￾duced (Table 2). Anterior systolic wall thickening was slightly decreased at 10 min after injection of the embolizing particles (31.7 F 3.8% vs. 28.1 F 3.8%, n.s.). Coronary blood flow at 10 min after microembolization was increased from 38.6 F 3.6 to 53.6 F 4.3 ml
min 1 ( p < 0.05) with a peak flow of 59.5 F 4.3 ml
min 1 at 3.2 F 0.4 min after microembolization. The area at risk was comparable between the groups (Fig. 4). In group 1 following 90-min severe myocardial ischemia and 120-min reperfusion, infarct size averaged 17.0 F 3.2%. Ischemic preconditioning by 10-min ische￾mia and 15-min reperfusion reduced infarct size follow￾ing 90-min ischemia and 120-min reperfusion to 2.6 F 1.1% ( p < 0.05 vs. groups 1 and 3) and the relationship between infarct size and subendocardial blood flow was shifted downwards, with a smaller infarct size at a given subendocardial blood flow ( p < 0.05 ANCOVA; Fig. 5). With microembolization in group 3, infarct size follow￾ing 90-min ischemia and 120-min reperfusion was 22.5 F 5.2%. Compared to 90-min sustained ischemia alone, the relationship between infarct size and subendocar￾dial blood flow was shifted upwards towards higher infarct Fig. 3. TNFa in myocardium and vasculature. Immunohistological staining of tissue slices from the anterior and the posterior wall. In the anterior wall, distinct areas of myocytes adjacent to an embolizing microsphere are stained positive for TNFa; in the posterior wall staining is confined to the vascular wall (scaling bar 20 Am). A. Skyschally et al. / Cardiovascular Research 63 (2004) 313–322 317

118 A或a的TW/Ga山mair2ti200多l-2力 Tabke 2 HR Imn-] LVP [mn Hg dWd:lrim Hes"'] WTane D1 CBF Imi-min Cooup 1 dliae 301±2 93.s±1.5 1347±129 55±3.8 46.5±59 5 7012 2911.9 1017上579 0341c 54⊥03 85 0@+2 81.1+23+ 1069+98* 11+1.1 53+63+ 0甲2 Rwdlie %±4 40土45 1530±151 H6土11 2打1±10 IPRep 明±5 34±43 1232±62 194±324 345±4.1 0±4 7±41* 1079±76+ 68±1.5 63±6.9时 例5 15 1141⊥9" 15⊥1.7+ 43⊥09* 印1 twliae 101土4 101k士16 104± 13士18 样6士16 ME 10见±4 93±3.0 1311±闭 2到.1±3.s 53.6±43+ 10±5 06±3.5+ 919±64 5±1.14 5.)±09 例5 1⊥4 25⊥2中 9%19 90⊥049 57109* 15,I8:d -mn lchemc IPRep:15-mn reperfskn ater he precondinorine ischemi M1 mn aher smary mkmembolation aterier wall:(HP:man LAD comnary inflaw. "p005线beseline. size at a given subendocardinl blood flow (p<005 In grup 3 at 10 min after ooronary microembolization, ANCOVA:Fie 5) the imersritial adenosine eoncentmtion was not diflerer In group 1,the inerstitial adenosine concentration in- from haseline values (2.005 gmol at baseline vs. creased with the onset ot the sustained ischemin from 23±0.6umal-厂1t10 min after microemholization, 4.,0±0.9ual.r'o a maximum of204±4.3mal' but it was incremsed with the onset of the sustained at 20 min (Fig.61. ischemia.The maximal increase during the sustained In group 2.the inkerstitial adenosine concentration in- ich.mia(11.6±2.2me-l厂20min)was less than eased from2.4±1.3umo-l厂at baseline to13.l土5.8 in group I. umol-during reperfision fillowing the preconditioning The coronary venous adenosine concentration was ischemia.Afer a transien drop,the imerstitial adenosine increasod at 5 min aner the onset of the sustained concentration increased agnin during sustained ischemia ischemia in group1.In group工，coronsry venous3 pak68士3.0-广330miLh.&)ut remtained nosine concentration was increised during the precon- lower thimn in group I. ditioning ischemia.During reperfusion following the 50 40 30 20 0048-0,007 rmin 0.0E9±0000mmm9 0c60-0,038wmin9 Group 1 Group 2 Group 3 圆area at risk%tventrick■infaret size[area at ris) Fig 4.Aga at risk.infarct stoe.and regional subendecard al blood flw.There were to significant ditferenoes beoween the free groups in subendecardal Houd flow anl ans i riak.InEre xiee wp nalrvd by ischersic prxundllioting is goug 2.[Gimap I:9-mia isheri 2h rperfinion Ciruap 2:10mis 21 tepersuon)

size at a given subendocardial blood flow ( p < 0.05 ANCOVA; Fig. 5). In group 1, the interstitial adenosine concentration in￾creased with the onset of the sustained ischemia from 4.0 F 0.9 Amol
l 1 to a maximum of 20.4 F 4.3 Amol
l 1 at 20 min (Fig. 6). In group 2, the interstitial adenosine concentration in￾creased from 2.4 F 1.3 Amol
l 1 at baseline to 13.1 F 5.8 Amol
l 1 during reperfusion following the preconditioning ischemia. After a transient drop, the interstitial adenosine concentration increased again during sustained ischemia (peak 6.8 F 3.0 Amol
l 1 at 30 min, n.s.) but remained lower than in group 1. In group 3 at 10 min after coronary microembolization, the interstitial adenosine concentration was not different from baseline values (2.0 F 0.5 Amol
l 1 at baseline vs. 2.3 F 0.6 Amol
l 1 at 10 min after microembolization), but it was increased with the onset of the sustained ischemia. The maximal increase during the sustained ischemia (11.6 F 2.2 Amol
l 1 at 20 min) was less than in group 1. The coronary venous adenosine concentration was increased at 5 min after the onset of the sustained ischemia in group 1. In group 2, coronary venous ade￾nosine concentration was increased during the precon￾ditioning ischemia. During reperfusion following the Fig. 4. Area at risk, infarct size, and regional subendocardial blood flow. There were no significant differences between the three groups in subendocardial blood flow and area at risk. Infarct size was reduced by ischemic preconditioning in group 2. (Group 1: 90-min ischemia/2-h reperfusion; Group 2: 10-min ischemia/15-min reperfusion followed by 90-min ischemia/2-h reperfusion; Group 3: Coronary microembolization followed by 90-min ischemia/ 2-h reperfusion). Table 2 Hemodynamic data HR [min 1 ] LVP [mm Hg] dP/dt [mm Hg
s 1 ] WTant [%] CBF [ml
min 1 ] Group 1 Baseline 101 F 2 93.8 F 1.5 1347 F 129 35.5 F 3.8 46.8 F 7.9 I5 102 F 2 82.9 F 1.9* 1017 F 57* 0.3 F 1.0* 5.4 F 0.3* I85 102 F 2 81.1 F 2.3* 1069 F 98* 1.1 F 1.1* 5.3 F 0.3* Group 2 Baseline 98 F 4 94.0 F 4.5 1530 F 153 34.6 F 3.7 27.7 F 3.0 IPRep 99 F 5 89.4 F 4.3 1232 F 62 19.4 F 3.2* 34.5 F 4.1 I5 99 F 4 80.7 F 4.1* 1079 F 76* 0.8 F 1.5* 6.3 F 0.9* I85 102 F 5 85.0 F 4.8 1141 F 94* 1.5 F 1.7* 6.3 F 0.9* Group 3 Baseline 101 F 4 101.8 F 2.6 1404 F 83 31.7 F 3.8 38.6 F 3.6 ME 102 F 4 99.3 F 3.0 1311 F 60 28.1 F 3.8 53.6 F 4.3* I5 103 F 5 80.6 F 3.5* 919 F 36* 0.5 F 1.1* 5.7 F 0.9* I85 109 F 4 82.5 F 2.8* 994 F 49* 0.0 F 0.6* 5.7 F 0.9* I5, I85: 5- and 85-min ischemia; IPRep: 15-min reperfusion after the preconditioning ischemia; ME: 10 min after coronary microembolization. HR: heart rate; LVP: peak left ventricular pressure; dPdt: maximum of the first derivative of left ventricular pressure; WTant: systolic wall thickening of the anterior wall; CBF: mean LAD coronary inflow. * p < 0.05 vs. baseline. 318 A. Skyschally et al. / Cardiovascular Research 63 (2004) 313–322

A.Shrmtally et al Cardnsraler Reoeari 63 (200 313-12 319 40 Gop1 .Droup2 ◇ =巳=Gep3 75 y511+533 20 ◆ 1= 8.26 --300+135 y-0+4月 14.月位 010 01 Q.20的004a500ec.e?1的005.1mǜ.11 subendocsrdis blood llow Iml'min Fig.5.Linear reyrexsion betuoes ahendeardial hloed frw and infaret sise.With micremholiratimn,the regrexin line wan xhidd upuands twards higher r2rtse了agiven suendocardial blood fow.情ischenk preconditioning.nhEl理整是wen subendocardiol同w56ed preconditioning ischemia,coronary venous adencsine con. S.Discusxion centration retumed back to baseline and tended to increase (n.s.)again at 5 min of the sustained ischemia.In gruup 5.了.Reduction of infarer s在eb'chemic precondinoning 3.the coronary venous adenosine concentraticn was but not by coronary microembolization already incresssod st 10 min after mcroembolization and remained inereased during the subsequent sustained ische. In the peresent study,we have transferred our dog model mi (Fie.7). of coronary miemembolization [2,3,7]to pigs.Also we 量G0up1 ★Gmp7 D-Group 3 RB 20 30 50 0 ME nn九m与 the inlereitial acknosne increned sic-fold during eeperfuicn (IPep)foliwing ichemic precmncitioring.In grup 3,he inkrstital adenne cmncestratin remeined uncharged with cororary microembolization (MEI.bit incseased atter me orset of the sustarned ischentia.'ith ischemc precondtionine and potutary miesoctbulalion the masinel inateso i the inpliial adt:couoneriut dariag h:wiml iahana were ulrtuad.(RB:tepeulsd in group 3 pe005 vs.Baseline:pe005 Giroup I vs.Gcup 2 and 3:'pe005 Gooup 1 vs Group 2)

preconditioning ischemia, coronary venous adenosine con￾centration returned back to baseline and tended to increase (n.s.) again at 5 min of the sustained ischemia. In group 3, the coronary venous adenosine concentration was already increased at 10 min after microembolization and remained increased during the subsequent sustained ische￾mia (Fig. 7). 5. Discussion 5.1. Reduction of infarct size by ischemic preconditioning, but not by coronary microembolization In the present study, we have transferred our dog model of coronary microembolization [2,3,7] to pigs. Also we Fig. 6. Interstitial adenosine concentration. In group 1, the interstitial adenosine concentration increased after the onset of the sustained ischemia. In group 2, the interstitial adenosine increased six-fold during reperfusion (IPRep) following ischemic preconditioning. In group 3, the interstitial adenosine concentration remained unchanged with coronary microembolization (ME), but increased after the onset of the sustained ischemia. With ischemic preconditioning and coronary microembolization, the maximal increases in the interstitial adenosine concentration during the sustained ischemia were attenuated. (RB: repeated baseline measurement in group 1; IPRep: 10 min of reperfusion following ischemic preconditioning in group 2; ME: 10 min after coronary microembolization in group 3; *p < 0.05 vs. Baseline; # p < 0.05 Group 1 vs. Group 2 and 3; + p < 0.05 Group 1 vs. Group 2 ). Fig. 5. Linear regression between subendocardial blood flow and infarct size. With microembolization, the regression line was shifted upwards towards higher infarct size at a given subendocardial blood flow. With ischemic preconditioning, infarct size at a given subendocardial blood flow was reduced. A. Skyschally et al. / Cardiovascular Research 63 (2004) 313–322 319

A.Shrurkally et a!.Candnneatir Prreonci di (20H)31-12 .1 三2 ■i词 ■防3 28 1 1 .e 4sdm5好5 Group 3 o四auyw1uset山alin in3 od afler the opet oftla4企al bclams.A simil在eis xee in pu中2 g the pexuslitning icchema Drine reperfnion mollring the preernditioning tocheris.cmorory venois admnesine carcertration dmpped hock m hveline ad tended s iacreae (ns.]wath the ouses of the susuned ischenia Wit oroary microembolizon (group 31.the cveomry veroas odenesise conocmmroa wis inereaad and nemned clvared dharirg the ardained ischema (15.185:5 and 85-mmin iocheri:IP:10 mi uf precendiiuiag ichenic IPRep 15min reperduson atter the precondiconig ischemia:ME:10 min acter comnony microembolizaton:"p0.05 vs amenal) confirmed that ischemic preconditioning by ane cycle of 10. In the present stody,mea infarct size as such was not min ischemia and 15-min reperfusion reduces infart size different hetween groups I and 3;however,the relationships from suhsequent 90 min sustained ischemia and 120 min of between infarct size and suhendocardial blood flow were reperfusion [10.11].In these prioe studies,adenosine was different,indicating that this relationship may be a more identified asa impoctant trigger of ischemic precondition. sensitive endpoint of damage or prutection of ischemic ing.The interstitial adenosine concentration increased dur reperfused myocardium.The small difference in mean ing reperfusion following the preconditioning ischemia infarct size between groups I and 3 is in a range similar [10.11].and procection was lost when adenosine was de to the infarct size induced by corocary microembolization graded by adenosine deaminase [11]. 灯然derived from the four control experiments with A similar finding was reported in a study by Gnind ct al coronary microembolization. [S]where comnary microembolization prioe to 45 min LAD ocelusion inereased infaret size in pigs to 84% $.2.The role of vesewlar and intersirial adenoine in compared to 60%in comtrols Compared to the present Kselewla and candiopmreetimn study,the mch greater increase in infaet size in the study by Grund et al.mmy he attnbuted to different experimental To investigate the role of adenosine with coronary micm- protocols.Grind et al.induced microembolization by emholization,we measured adenosine concentrations in injection of embolizing particles into the lelt atrium which both the vasclar and interstitial space.The high variation resulls in microembolization of the cmtire heart,and sub- in coronary venous adenosine concentration at 5-min iscbe- sequenl sustainod iscbermia was indaced by 45 min LAD mia in group 1.during the preconditioning ischemia,and at oochusion.In our experimentsv the embolizing particles 5 min sustained schemnia in group 2cn be attribued to the were mjeeted direetly into the LAD before a perod of varying severity of the preconditioning or sustained isch- low-flow ischemia which resulted in a much smaller infaret mia,respectively.The observod increse in comrary blood s?eof17.0士32%afee第a5kin6 ntrol.Tns,in flow aner mctocmboliztion could agnin be attbuted to the experiments by Gnind et al.,two factors may acoount the ohscrved releae of adenosine into the vascalature, for the pronourced increase in infarct size with micro- Interestingly,the increuse in cororary verous adenosine embolization.First,the entire heart was microembolized during the early sustained ischemia was attenuated by prior which affected also the border zone of the area at risk.and microembolization or ischemic preconditicning. second,the more severe ischemia protoool results in much An increase in interstitial adeposine prioe to the sustained larger infarction per se. ischemia is obvicusly necessary to induce myocardial pro-

confirmed that ischemic preconditioning by one cycle of 10- min ischemia and 15-min reperfusion reduces infarct size from subsequent 90 min sustained ischemia and 120 min of reperfusion [10,11]. In these prior studies, adenosine was identified as an important trigger of ischemic precondition￾ing. The interstitial adenosine concentration increased dur￾ing reperfusion following the preconditioning ischemia [10,11], and protection was lost when adenosine was de￾graded by adenosine deaminase [11]. A similar finding was reported in a study by Grund et al. [5] where coronary microembolization prior to 45 min LAD occlusion increased infarct size in pigs to 84% compared to 60% in controls. Compared to the present study, the much greater increase in infarct size in the study by Grund et al. may be attributed to different experimental protocols. Grund et al. induced microembolization by injection of embolizing particles into the left atrium which results in microembolization of the entire heart, and sub￾sequent sustained ischemia was induced by 45 min LAD occlusion. In our experimentsv the embolizing particles were injected directly into the LAD before a period of low-flow ischemia which resulted in a much smaller infarct size of 17.0 F 3.2% of the area at risk in controls. Thus, in the experiments by Grund et al., two factors may account for the pronounced increase in infarct size with micro￾embolization. First, the entire heart was microembolized which affected also the border zone of the area at risk, and second, the more severe ischemia protocol results in much larger infarction per se. In the present study, mean infarct size as such was not different between groups 1 and 3; however, the relationships between infarct size and subendocardial blood flow were different, indicating that this relationship may be a more sensitive endpoint of damage or protection of ischemic/ reperfused myocardium. The small difference in mean infarct size between groups 1 and 3 is in a range similar to the infarct size induced by coronary microembolization per se, as derived from the four control experiments with coronary microembolization. 5.2. The role of vascular and interstitial adenosine in ischemia and cardioprotection To investigate the role of adenosine with coronary micro￾embolization, we measured adenosine concentrations in both the vascular and interstitial space. The high variation in coronary venous adenosine concentration at 5-min ische￾mia in group 1, during the preconditioning ischemia, and at 5 min sustained ischemia in group 2 can be attributed to the varying severity of the preconditioning or sustained ische￾mia, respectively. The observed increase in coronary blood flow after microembolization could again be attributed to the observed release of adenosine into the vasculature. Interestingly, the increase in coronary venous adenosine during the early sustained ischemia was attenuated by prior microembolization or ischemic preconditioning. An increase in interstitial adenosine prior to the sustained ischemia is obviously necessary to induce myocardial pro￾Fig. 7. Arterial and coronary venous adenosine concentrations. The arterial adenosine concentrations remained unchanged in all three groups. In group 1, the coronary venous adenosine concentration increased after the onset of the sustained ischemia. A similar increase is seen in group 2 during the preconditioning ischemia. During reperfusion following the preconditioning ischemia, coronary venous adenosine concentration dropped back to baseline and tended to increase (n.s.) with the onset of the sustained ischemia. With coronary microembolization (group 3), the coronary venous adenosine concentration was increased and remained elevated during the sustained ischemia (I5, I85: 5 and 85-min ischemia; IP: 10 min of preconditioning ischemia; IPRep: 15-min reperfusion after the preconditioning ischemia; ME: 10 min after coronary microembolization; *p < 0.05 vs. arterial). 320 A. Skyschally et al. / Cardiovascular Research 63 (2004) 313–322

A.Syerbally e al Cendnouralor Recrrb 3 (3!-12 2知 tection,supparting its role ns trigger [&]The time course of lower than that for myocardialcardioprotective effects [22]. changes in the inerstitial adenosine concentrtion with This could he due to the fact that the vasculnr wall contains ischemic procooditioning obeervod in the present study is a lrge receplor reserve for adnosine A:receptors [231. consistent with a previous study [11].The inerstitial aden- wherens the potential proloctive clfocts of inlerslitial aden- osine coecentration remained ucanged afer 10 mimn of osine on cardiomyocytes are cverted through activation of microembolization.This lack of increase of interstitial adenosine A and A:receptors [24]. adenosine shortly after microembolization is possibly attrib- uted to an augmented washou of adenosine from the 天3 wleraclio of cuesuvine amd TNF microembolized myocardium by the ancreased coronary blood fow in the adjacent microvasculatire.With coronary While the lack of increase in the interstitial adenosine microembolization,the interstitial adenosine increased anly concentration with caronary microembolizabion is a plausi after the onset of the sustained ischemia d this increase ble explanation for the lack of protecticn from infarction,an during the sustained ischemia was attenuated.An attenua alemative explanation relates to TNFa.TNFo is causally tion of increases in the interstitial adenosine concentrations involved in the development of infarct size.since pretreat during the prolonged ischemia alter ischemic procoodition- ment with TNFo antibodies roduces infarct site in anesthe ing was reported for rabbits [16.17L pigs [18]and dogs 19] tizod rabbits 25]Possibly therefore the incressod TNFa and was also observed in the present study.However.this levels counderacted the beneficial ellects that might other- alteruation of an increase of the interstittal adenosine wise hinve resulted from coronary microembolization.On the concentration appears to he urelated to a protective effect other hand,TNFa is also involved in the mechanism of In rabbits,either ischemic preconditioning or pharmacologic preconditioning.Pretreatment with TNFa mimics precondi preconditioning by intravenous infusion of ademosine was tioning in isolated rat hearts [26]and mice [271,and in TNFa able to redce infarct size after prolonged isbemia hut anly knockout mice ishemic preconditianing is ahrogated [27].It ischemic precoditioning attenuated the increse in ine is therefore possible that increasing levels of TNTa may tial adenosine cencentraticn during the prolonged iscbemia protect the myocardium from infarctian by sustained ische 16].An attenuaticn of the increase in the interstitial mia in a later phase following corenary microembelization. adenosine concentration during prolonged ischemia was and this will have to be studied in fiture experiments. also ohserved in rabhits nfter peretrearment with HOF 140 We conclude that comnary microemholimntion dees not c 8-phenyl-thenphylline which both hlocked the infaret nrotect the myocardm acutely from the consequences of size-limiting effect of ischemic preconditioning [17].Our subsequent sustained ischemin.The Inck of protection current results further confirm that protection and nttenua- against infarction is nttributed to n failure of interstitinl tion of increase in interstitial adenosine during sustained adencosine to increase with coronary microembolization. ischemia are unrelated,since we observed an atenuated increase with either coronary microemholizatian,which was associated with a lack of prutection,ar ischemic precendi. References toning.which protected the myocardium by decresing nlaret siac. 1]Erbel B.Heh G.Bnetrev:oy mroemJAr The return to bascline in the intersttial adenosine con- C道C者d2003长22-4. centration in all groups during the late phase of the snstained 21 D3eg:H Ncuman T.Berends M ct al.Perfisicn-ro ischemis may refleet the ongoing development of neerosis. 程iech with conmary micn止由n:n:ole of inflamra tin.Am J Prysid Hean Cro Phorsol 200279-H2587-92. Severely ischemie cell contnhute most to the nmount af 3]Skywhally A,Sdrale R.Erbel R.Hetoch G Roiaul cenaary und ineerstitial adenosine,hut these cells are also the first to underpo necnsis. Booet Cin Ptysiol 2002 292H611-4. The different patem of coronary venous and interstitial [4]Huri M,oue M,Kiskre:M,et :l.Ruke uf adetaoine in hyperemric adenosine concentrations likely relstes to their different 程of comn可d fow in r不circelation.Am J Physin Bot Cir:PLysid 1956250.HS0)-18 sources:the interstitial aderosine largely reflects myucardial [5]Ginnd F.Sommerechild HT.lyberg T.Kirkehren K.lehekk A production and rekase,whereas coronary venous plasma Miocembolition in pigs:effects on coronary blood fow and myo adenosine is at kast transicntly and to a certain exbent crdul iachemic lukrce.An J Paysiel Heart Cirr Phycil rekased fiom th cndothelimm 120.Even at bascline,the 19927T:11513-12 间Dinge H.Sch业R,B:oeow S,ct al.Coeotrry sictocrbo2alut nerstilial adenosine concentrition is about one magnitude hgher thaan the cororary venous concentration,and with 22:1451-2 sustainod eemis,it is inereased to an even preater extent 7Thielaen M.TXge IL Mortn C.er al.Myocadial dystincnon with This difference prohably not only reflects a diffusion gra dient but also active uptake ofadenosine by the endothelium 22:017=13. 21].In any case,if the corunary vencus aderosine lvels 8Sdiale R.Cuer MV.Bearenah M.Dowaey JM,Heack G.Sigal are assumed to reflect the levels in the vascular wall,the threshold for coronary vasomotor effects is cansiderably 181-8

tection, supporting its role as trigger [8]. The time course of changes in the interstitial adenosine concentration with ischemic preconditioning observed in the present study is consistent with a previous study [11]. The interstitial aden￾osine concentration remained unchanged after 10 min of microembolization. This lack of increase of interstitial adenosine shortly after microembolization is possibly attrib￾uted to an augmented washout of adenosine from the microembolized myocardium by the increased coronary blood flow in the adjacent microvasculature. With coronary microembolization, the interstitial adenosine increased only after the onset of the sustained ischemia and this increase during the sustained ischemia was attenuated. An attenua￾tion of increases in the interstitial adenosine concentrations during the prolonged ischemia after ischemic precondition￾ing was reported for rabbits [16,17], pigs [18] and dogs [19] and was also observed in the present study. However, this attenuation of an increase of the interstitial adenosine concentration appears to be unrelated to a protective effect. In rabbits, either ischemic preconditioning or pharmacologic preconditioning by intravenous infusion of adenosine was able to reduce infarct size after prolonged ischemia but only ischemic preconditioning attenuated the increase in intersti￾tial adenosine concentration during the prolonged ischemia [16]. An attenuation of the increase in the interstitial adenosine concentration during prolonged ischemia was also observed in rabbits after pretreatment with HOE 140 or 8-phenyl-theophylline which both blocked the infarct size-limiting effect of ischemic preconditioning [17]. Our current results further confirm that protection and attenua￾tion of increase in interstitial adenosine during sustained ischemia are unrelated, since we observed an attenuated increase with either coronary microembolization, which was associated with a lack of protection, or ischemic precondi￾tioning, which protected the myocardium by decreasing infarct size. The return to baseline in the interstitial adenosine con￾centration in all groups during the late phase of the sustained ischemia may reflect the ongoing development of necrosis. Severely ischemic cells contribute most to the amount of interstitial adenosine, but these cells are also the first to undergo necrosis. The different pattern of coronary venous and interstitial adenosine concentrations likely relates to their different sources; the interstitial adenosine largely reflects myocardial production and release, whereas coronary venous plasma adenosine is at least transiently and to a certain extent released from the endothelium [20]. Even at baseline, the interstitial adenosine concentration is about one magnitude higher than the coronary venous concentration, and with sustained ischemia, it is increased to an even greater extent. This difference probably not only reflects a diffusion gra￾dient but also active uptake of adenosine by the endothelium [21]. In any case, if the coronary venous adenosine levels are assumed to reflect the levels in the vascular wall, the threshold for coronary vasomotor effects is considerably lower than that for myocardial/cardioprotective effects [22]. This could be due to the fact that the vascular wall contains a large receptor reserve for adenosine A2 receptors [23], whereas the potential protective effects of interstitial aden￾osine on cardiomyocytes are exerted through activation of adenosine A1 and A3 receptors [24]. 5.3. Interaction of adenosine and TNFa While the lack of increase in the interstitial adenosine concentration with coronary microembolization is a plausi￾ble explanation for the lack of protection from infarction, an alternative explanation relates to TNFa. TNFa is causally involved in the development of infarct size, since pretreat￾ment with TNFa antibodies reduces infarct size in anesthe￾tized rabbits [25]. Possibly therefore the increased TNFa levels counteracted the beneficial effects that might other￾wise have resulted from coronary microembolization. On the other hand, TNFa is also involved in the mechanism of preconditioning. Pretreatment with TNFa mimics precondi￾tioning in isolated rat hearts [26] and mice [27], and in TNFa knockout mice ischemic preconditioning is abrogated [27]. It is therefore possible that increasing levels of TNFa may protect the myocardium from infarction by sustained ische￾mia in a later phase following coronary microembolization, and this will have to be studied in future experiments. We conclude that coronary microembolization does not protect the myocardium acutely from the consequences of subsequent sustained ischemia. The lack of protection against infarction is attributed to a failure of interstitial adenosine to increase with coronary microembolization. References [1] Erbel R, Heusch G. Brief review: coronary microembolization. J Am Coll Cardiol 2000;36:22 – 4. [2] Do¨ rge H, Neumann T, Behrends M, et al. Perfusion – contraction mismatch with coronary microvascular obstruction: role of inflamma￾tion. Am J Physiol Heart Circ Physiol 2000;279:H2587 – 92. [3] Skyschally A, Schulz R, Erbel R, Heusch G. Reduced coronary and inotropic reserves with coronary microembolization. Am J Physiol Heart Circ Physiol 2002;282:H611 – 4. [4] Hori M, Inoue M, Kitakaze M, et al. Role of adenosine in hyperemic response of coronary blood flow in microcirculation. Am J Physiol Heart Circ Physiol 1986;250:H509 – 18. [5] Grund F, Sommerschild HT, Lyberg T, Kirkeboen K, Ilebekk A. Microembolization in pigs: effects on coronary blood flow and myo￾cardial ischemic tolerance. Am J Physiol Heart Circ Physiol 1999;277:H533 – 42. [6] Do¨ rge H, Schulz R, Belosjorow S, et al. Coronary microembolization: the role of TNFa in contractile dysfunction. J Mol Cell Cardiol 2002;34:51 – 62. [7] Thielmann M, Do¨ rge H, Martin C, et al. Myocardial dysfunction with coronary microembolization: signal transduction through a sequence of nitric oxide, tumor necrosis factor-a and sphingosine. Circ Res 2002;90:807 – 13. [8] Schulz R, Cohen MV, Behrends M, Downey JM, Heusch G. Signal transduction of ischemic preconditioning. Cardiovasc Res 2001;52: 181 – 98. A. Skyschally et al. / Cardiovascular Research 63 (2004) 313–322 321

32壮 A.Srachally et al/Cohorarevar是awh43G32】 例Leesar MA.Sooddand M.Ahinod M.Bioadbest人，oll R.Preoend [18]Wiksom BG.Ronguis G.Waldesim A.Dyramies of myocardial tioning in homn myecantiam wih adenosine during coronary ango netabelins in the procond riosod porcine hoar sdied using oontin. 4ay复culaton I9T96400-7 uos microdalesis Bar lount J 19963-9 明St山R.os H.Vhhous C,fuch G.Ischerig precaedtioning场 [l9习nknD区、pt国scheric precendtioning o国oda pigs:a gradod pheaomen hsatkn adro ard beadyiinn e mnabolie国ctte aocunultion during9 rdal ische Cale19989k02-9 aCaa4国19322w1-. I月st白民.ae人os 1.laudh G.Iolveno of eadogo4s 可Nsssaa A1atG,Sn细人Contrbutian of corona可m如 adenosine in ischaomic preconditianing in owind,Pfligors Arch aal cels to cand ac aderosing prodacion Pfligers Anch 196a06 19950273-1. 615-14 lD人cnd年人aw4 onts of admasir6,nda,d句 21]Desssea A.Srappen M.Schafer $Kelm M.Qunrifieation of estra a自ana.Jg9430价.404-9 cellal and inreceluar adeneoine pooduction.Understandig he [13]Bclesjorw S Schule R.Dirge H.Schade FU,Heasch G.Esdotroin raasrierbeatous0 enceniratio0dieL.Cireulatien5对.纲 and ischemie preoeedtimng:TNF.n cercealmrien and myocantial 21=7. infaret devekpment n tabits.Am J Plosiel Heant Cire Plrysiol [22]Belardinel L,Shryoc IC Dees aderonise fanction a a retdlinory 1999277H240-5. mctabolie it the heat News Plysiol Sci 1992.7:52-6. 4月cumnimn T.Konictda I.van de Send A,AegS,S业民，etd Sk,saw小ySBr适G.Adis roe国 G.HestiScation of socnotic tissee by plre-postsnl migooscepy al an 3aef6 ronery wsedilien.Curculation199网9线7II-g. caly sage of aoue mpocandil infarcton.Lob Invest 2000.80981-2 B可Hca山kk JP,Hack B.Ashen K Aone adenosinorgic cardieprotee. Il分Aber 5.Bek味erow S.Konictzka 1.ct d.The sorun b e6争sehanc-reperfased h6 ns.Am J Ploysiol Hean Cin探Plysial myocardial TNFa cenceatsrion is incrensel is pacing-indaced heart 203,15H1797=18. failune is rabbis.Am J Plysiol Regul Inkr Comp Pysiol 2003:285 间B:wes,Belle L Daschin A.Hetsch G.S业R.TNFo an R463=-9. bodies are as efRective as ischemic preoonditioning in rodecing infarct IL利Lank与RD.Keeyn队e JO.Mestoer Jr RM.EITees of ische siae is nstis.An J Posiol Hean Cire Ploysiol 2003:284H927-30. sie and adcnosine precond riosing6角样la以月al fuid adesesine and [26]Leceer S.Smith RM.Woodword B.Opie LH.Recheme L Sack MK. myocandial infact siae.Am J Physial Hoart Cin Plysiol 199538 kericarien of a rovel role for sphingolipid sinaling in TNFen and H1460=6 schemic preoosditioning medated cardiopronection.J Mol Cell Cae l门MunT,sau由Ksh如k,rn0.Suapprexuion o起 diol 200;145-1X degradation of adeaise nackotides deng ichena ray aot be a E7 Smih k3以Sukrrat N M.Cathy山sak.Cic inchemi bat wufficicrt mechasim ar infrct aine lrrittion by prooaaditonng nat pharmaolgic procroditonng i abeegalod folkwing penctic Baic Hes Cantiol 199691-425-32. shlatin of the TNFn geme.Cardirvae Kex 2002:55:551-60

[9] Leesar MA, Stoddard M, Ahmed M, Broadbent J, Bolli R. Precondi￾tioning in human myocardium with adenosine during coronary angio￾plasty. Circulation 1997;95:2500 – 7. [10] Schulz R, Post H, Vahlhaus C, Heusch G. Ischemic preconditioning in pigs: a graded phenomenon. Its relation to adenosine and bradykinin. Circulation 1998;98:1022 – 9. [11] Schulz R, Rose J, Post H, Heusch G. Involvement of endogenous adenosine in ischaemic preconditioning in swine. Pflu¨gers Arch 1995;430:273 – 82. [12] Ontyd J, Schrader J. Measurements of adenosine, inosine, and hypo￾xanthine in human plasma. J Chromatogr 1984;307:404 – 9. [13] Belosjorow S, Schulz R, Do¨ rge H, Schade FU, Heusch G. Endotoxin and ischemic preconditioning: TNF-a concentration and myocardial infarct development in rabbits. Am J Physiol Heart Circ Physiol 1999;277:H2470 – 5. [14] Neumann T, Konietzka I, van de Sand A, Aker S, Schulz R, Heusch G. Identification of necrotic tissue by phase-contrast microscopy at an early stage of acute myocardial infarction. Lab Invest 2000;80:981 – 2. [15] Aker S, Belosjorow S, Konietzka I, et al. The serum but not the myocardial TNFa concentration is increased in pacing-induced heart failure in rabbits. Am J Physiol Regul Integr Comp Physiol 2003;285: R463 – 9. [16] Lasley RD, Konyn PJ, Hegge JO, Mentzer Jr RM. Effects of ische￾mic and adenosine preconditioning on interstitial fluid adenosine and myocardial infarct size. Am J Physiol Heart Circ Physiol 1995;38: H1460 – 6. [17] Miura T, Suzuki K, Shimamoto K, Iimura O. Suppression of the degradation of adenine nucleotides during ischemia may not be a sufficient mechanism for infarct size limitation by preconditioning. Basic Res Cardiol 1996;91:425 – 32. [18] Wikstro¨m BG, Ronquist G, Waldenstro¨m A. Dynamics of myocardial metabolism in the preconditioned porcine heart studied using contin￾uos microdialysis. Eur Heart J 1995;16:563 – 9. [19] Van Wylen DGL. Effect of ischemic preconditioning on interstitial purine metabolite and lactate accumulation during myocardial ische￾mia. Circulation 1994;89:2283 – 9. [20] Deussen A, Mo¨ser G, Schrader J. Contribution of coronary endothe￾lial cells to cardiac adenosine production. Pflu¨gers Arch 1986;406: 608 – 14. [21] Deussen A, Stappert M, Scha¨fer S, Kelm M. Quantification of extra￾cellular and intracellular adenosine production. Understanding the transmembranous concentration gradient. Circulation 1999;99: 2041 – 7. [22] Belardinelli L, Shryock JC. Does adenosine function as a retaliatory metabolite in the heart. News Physiol Sci 1992;7:52 – 6. [23] Shryock JC, Snowdy S, Baraldi PG, et al. A2A-adenosine receptor reserve for coronary vasodilation. Circulation 1998;98:711 – 8. [24] Headrick JP, Hack B, Ashton KJ. Acute adenosinergic cardioprotec￾tion in ischemic-reperfused hearts. Am J Physiol Heart Circ Physiol 2003;285:H1797 – 818. [25] Belosjorow S, Bolle I, Duschin A, Heusch G, Schulz R. TNFa anti￾bodies are as effective as ischemic preconditioning in reducing infarct size in rabbits. Am J Physiol Heart Circ Physiol 2003;284:H927 – 30. [26] Lecour S, Smith RM, Woodward B, Opie LH, Rochette L, Sack MN. Identification of a novel role for sphingolipid signaling in TNFa and ischemic preconditioning mediated cardioprotection. J Mol Cell Car￾diol 2002;34:509 – 18. [27] Smith RM, Suleman N, McCarthy J, Sack MN. Classic ischemic but not pharmacologic preconditioning is abrogated following genetic ablation of the TNFa gene. Cardiovasc Res 2002;55:553 – 60. 322 A. Skyschally et al. / Cardiovascular Research 63 (2004) 313–322