Dimitrios Th. Kremastinos a,*, Elias Bofilis a, George K. Karavolias a,Apostolos Papalois a, Loukas Kaklamanis b, Efstathios K. Iliodromitis a
a Second Department of Cardiology, Onassis Cardiac Surgery Center, 356 Sygrou A6enue, 17674 Athens, Greeceb Department of Pathology, Onassis Cardiac Surgery Center, 356 Sygrou A6enue, 17674 Athens, Greece
Received 16 April 1999; received in revised form 25 August 1999; accepted 10 September 1999
Elevations in total cholesterol are a major factorpredisposing to atherosclerosis, with vulnerable plaquescausing acute myocardial infarction [1]. Currently, themost efficient method of reducing mortality in thiscondition is to achieve rapid reperfusion by lysis ormechanical disruption of the occlusive thrombus andplaque [2]. The mortality from acute myocardial infarc-tion under these circumstances is inversely related tothe amount of myocardial salvage achieved by reperfu-sion [2], therefore agents that slow the rate of ischemicnecrosis are likely to save many lives [3].
Ischemic preconditioning describes the resistance toischemic injury that follows brief periods of myocardial
ischemia. The brief ischemic trigger delays the onsetand slows the rate of necrosis occurring in response tosubsequent prolonged ischemia. The resulting level ofprotection is profound, with a typical 4-fold reductionin the volume of infarction after relatively short periodsof coronary occlusion. This striking level of protectionhas resulted in preconditioning being ranked secondonly to reperfusion in ability to reduce infarction [4].The possibility therefore exists that ischemic precondi-tioning, by profoundly slowing the rate of necrosis,increases the amount of myocardial salvage on reperfu-sion. This is certainly consistent with the finding thatpatients with pre-infarction angina tend to have a morebenign course and smaller enzyme derived infarct sizesthan those with unheralded infarction [3].
However, there is also evidence that nitric oxide isinvolved in some animal models in the antiarrhythmiceffect of preconditioning [5]. In addition, it is alsoknown that hypercholesterolemia impairs endothelial
D.T. Kremastinos et al. / Atherosclerosis 150 (2000) 81–8982
relaxation and nitric oxide production [6]. The possibil-ity therefore exists that hypercholesterolemia predis-poses patients to myocardial infarction, whilstdepriving them of the endogenous protection offered byischemic preconditioning.
The aim of this study was to examine the possibilitythat preconditioning is attenuated in a clinically rele-vant animal model where hypercholesterolemia predis-poses to atherosclerosis.
2. Methods
Thirty-six New Zealand White male rabbits weredivided into two series of experiments each composingof two groups. All animals received proper care incompliance with the ‘Principles of Laboratory AnimalCare’, formulated by the National Society for MedicalResearch, and the ‘Guide for the Care and Use ofLaboratory Animals’, prepared by the NationalAcademy of Sciences and published by the NationalInstitutes of Health (NIH publication no. 86-23, revised1985). The first series consisted of 20 animals whichwere fed for 8 weeks on a standard laboratory diet thatwas enriched with 2 g of cholesterol (product USP23,Dolder) mixed with 6% corn-oil in every kilogram offood and tap water. Prior to cholesterol enrichment,blood samples were drawn from an ear vein and cen-trifugated for 10 min at 3000 rpm for baseline choles-terol level assessment. At the end of the feeding periodall the animals were randomly divided into two groups(A and B) control and preconditioned, respectively. Thesecond experimental series consisted of 16 animals fed anormal diet and tap water; groups C and D from thisseries served as normal cholesterol controls having thesame preconditioning trigger and duration of prolongedischemia as groups A and B, respectively.
2.1. Surgical preparation
All the animals were anesthetized by slowly injecting20 mg/kg sodium thiopentone (Pentothal, Abbott) intoan ear vein. The trachea was opened via midline cervi-cal incision and the animals were intubated and me-chanically ventilated under positive pressure at the rateof 30 respirations per minute at a tidal volume of15–20 ml. A small animal ventilator (MD Industries,AL) was adjusted as necessary to maintain arterialblood gases and pH within the normal range. Twopolyethylene catheters were inserted; one was posi-tioned in the jugular vein for blood sampling, fluidinfusion (1 ml of normal saline containing 1000 IUheparin/100 ml was administered every 30 min) and fortop up anesthesia and the other in the carotid artery forcontinuous blood pressure monitoring. A bipolar chestlead was used for electrocardiographic recording. The
surface electrocardiogram and intraarterial pressurewere monitored using a Nihon-Koden RM 6000recorder. Blood samples were drawn from all the ani-mals for cholesterol level assessment, 8 weeks aftercholesterol enriched diets had commenced (groups Aand B) and for baseline values in groups C and D. Thechest was opened via a left thoracotomy in the fourthintercostal space and the beating heart was exposed.After pericardiotomy a 3.0 silk suture was passedthrough the myocardium around a prominent branchof the left coronary artery. Regional ischemia wasinduced by pulling the ends of the thread through asmall segment of a soft tubing and a clamp thus firmlycompressing the coronary artery. The successful induc-tion of ischemia was verified by visual inspection of acyanotic region and by ST elevation on the electrocar-diogram. Reperfusion was achieved by releasing theclamp and was verified by the ‘blushing’ of the ischemicregion and by refilling of the artery.
2.2. Experimental protocol
Groups A and C, from the first and second series,respectively, were subjected to 30 min regional ischemiafollowed by 180 min reperfusion and served as controls.Groups B and D, again from the first and second series,respectively, were subjected to one cycle of precondi-tioning, comprising 5 min of regional ischemia plus 10min reperfusion, prior to 30 min ischemia followed by180 min reperfusion.
The protocol is presented schematically in Fig. 1.
2.3. Risk area and infarct size
Animals were sacrificed at the end of the experiment,hearts were removed and mounted on a perfusion ap-paratus and perfused (50 mmHg) for 2 min retrogradelyvia the aorta with normal saline (20 ml/min, roomtemperature). When all residual blood had been re-moved from the arteries, the coronary ligature wasretightened and 5 ml of 0.1% Zn–Cd fluorescent parti-cles (10–20 mm, suspended in saline) were slowly in-fused over 3 min for the delineation of the normallyperfused tissue from the risk zone. Hearts were thenfrozen for 24 h at −10°C and sliced in 3 mm thicksections from the apex to the base. The slices wereincubated in 1% triphenyl tetrazolium chloride solutionfor 20 min at 37°C and the infarcted area was definedas the negative stained region. The heart slices werethen immersed in 10% formaldehyde solution for 24 hto delineate the infarcted (tetrazolium chloride nega-tive) areas more clearly. To clarify the borders betweenthe risk zone and the normal area, slices were examinedunder ultraviolet light (wavelength 366 nm). The in-farcted, the risk and the normal areas were traced ontoan acetate sheet, photographically enlarged and were
D.T. Kremastinos et al. / Atherosclerosis 150 (2000) 81–89 83
quantified by planimetry with the aid of a digitizer. Theareas of infarction and of myocardial tissue at risk ofinfarction were automatically transformed to volumesby multiplying by the slice thickness (3 mm). Infarctand risk area volumes were expressed as cm3 and thepercent of infarct to risk area (I/R) calculated.
2.4. Measurement of atherosclerotic plaque
Arterial samples were harvested at the end of theexperiment from the thoracic aorta (one from the rootand the other from the descending), the carotid arteriesand the heart. The ascending thoracic aorta and thecarotids were dissected free of adventitial tissue and anyremaining blood rinsed away. The ostia of the coronaryarteries were also analysed, together with the epicardialportions of the coronary vessels. The tissues were fixedin 10% buffered formalin and then embedded inparafin, and serial tissue sections cut and stained withroutine histochemical stains. Hematoxylin-eosin waspreferred to oil-red stain, because of the better preser-vation of the architecture of the intimal lesions, eventhe early ones. Oil-red stain requires freezing of sectionsby the respective microtome, which usually alters thearchitecture of the early intimal lesions.
Histopathological examination and the assessment ofatherogenesis were done by light microscopy. The pres-ence of atherogenetic intimal lesions was evaluated inthe aortas, carotids and coronary arteries. Samplesfrom animals fed normal laboratory chow were simi-larly evaluated (control group).
2.5. Statistical analysis
All results are presented as mean value91 SEM. Alltests were computed with the CSS-Statistical Softwarepackage. Infarct size, risk zone and the infarct to riskratio in percent were compared between groups using
one way analysis of variance (ANOVA). Hemodynamicparameters (heart rate and mean blood pressure) in thecontrol and preconditioning groups were evaluated us-ing 2-factor ANOVA (for group and time) with re-peated measures across the second factor. Post-hocanalyses with the least significant difference test wereused for the statistical comparisons. Comparison of theregression lines (infarct size plotted against risk zone)between control and preconditioning groups, was madeby ANCOVA. A probability of B0.05 was consideredstatistically significant.
3. Results
3.1. Animals and exclusions
One rabbit from the first experimental series diedprior to surgery whilst on the cholesterol diet. Twoanimals (one from group A and one from group B) diedfrom intractable ventricular fibrillation, respectivelyduring ischemia and the final period of reperfusion.Two other animals (one from group B and one fromgroup C) died of progressive hypotension during thefinal period of reperfusion. Thus, 16 animals from thefirst experimental series and 15 from the second seriescompleted the study.
3.2. Hemodynamic 6ariables
Baseline characteristics for all the study groups andtheir trends during the study are presented in Table 1.No significant differences were detected between thegroups.
3.3. Cholesterol le6els
After centrifugation of the blood, cholesterol levels
Fig. 1. Diagrammatic representation of the experimental protocol used to evaluate the effect of preconditioning on infarct size in thehypercholesterolemic and normal rabbit heart in vivo. Periods of ischemia are shown in black and periods of reperfusion in white. Arrowsrepresent the time points of blood sample collection for cholesterol level assessment.
D.T. Kremastinos et al. / Atherosclerosis 150 (2000) 81–8984
Table 1Hemodynamic variables of the groups at the different time pointsa
a PC, preconditioning; BP, mean blood pressure; HR, heart rate.
were evaluated. Mean cholesterol was 58.398.7 mg%in groups A and B at baseline and 14029125 mg%after 8 weeks of cholesterol enriched diet. Baselinecholesterol was 57.595.8 mg% in groups C and D. Thedifference in blood plasma appearance after centrifuga-tion is clearly distinct in Fig. 2.
3.4. Histology
Atherogenesis was detected in the form of subintimalaccumulation of lipids and macrophages. Atherogeneticlesions were present in the ascending aorta, the ostia ofthe coronaries and the epicardial vessels. The lesionswere patchy and interspersed by normal intima. Thecarotids did not show any atheroma formation. Athero-genesis was not detected in any of the vessels removedfrom the animals fed normal laboratory chow (Figs. 3and 4).
3.5. Infarct size
Table 2 details the infarct size and the risk zone size,together with their percent ratios for the various studygroups. Infarcted volume was 0.56 cm3 in group A witha trend towards a smaller volume in group B (0.28cm3). The risk zone volumes were not different betweenthe two groups (1.39 versus 1.50 cm3, respectively). Asshown in Fig. 5 the infarct to risk zone ratio wassignificantly smaller in the preconditioned group(39.396.3 versus 16.793.9%, PB0.01).
In the second series of experiments infarcted volumewas 0.71 cm3 in group C and 0.16 cm3 in group D(PB0.01), while the risk zone volume did not differsignificantly between the two groups (1.65 versus 1.58cm3). Similar to the first series of experiments, theinfarct to risk zone ratio was significantly smaller(41.497.5% versus 10.893.3%, PB0.01) in the pre-conditioned group (Fig. 5). Absolute infarct size plottedagainst absolute risk zone (cm3) for all hypercholes-terolemic hearts in the study can be seen in Fig. 6. Thedifference in the slopes of the two regression lines isreadily apparent (PB0.02).
4. Discussion
Our study demonstrates that preconditioning is ableto confer protection against myocardial infarction inboth normally fed and hypercholesterolemic rabbitswith atherosclerosis.
Progress in research and new therapeutic develop-ments have revealed that apart from the restoration ofblood flow there are also ‘endogenous mechanisms ofprotection’ which render the heart more tolerant toprolonged ischemia [7,8]. Preconditioning is one ofthese protective mechanisms and a great number of invivo and in vitro studies have reconfirmed that itprotects the heart of various species [7,9], independentlyof collaterals [9,10]. Since there is no experimentalinformation, an open question remains as to whetherpreconditioning is able to protect the heart againstinfarction in a hypercholesterolemic and atheroscleroticanimal model.
Patients with coronary artery disease have atheroscle-rosis and are in many cases hypercholesterolemic. How-ever it is not known whether preconditioning protectsin such patients with chronic CAD, as it does in thenormal heart under experimental conditions. Clinicalstudies so far have generated indirect evidence of theexistence of preconditioning, i.e. tissue biopsies fromthe human heart [11], improved exercise tolerance [12],reduced ST elevation after repeated balloon inflation indiseased coronary arteries [13] and various pharmaco-logical interventions [14]. Because of the remaininguncertainty, we have investigated whether precondition-ing protects hypercholesterolemic rabbits with athero-sclerosis.
The attenuation of ST elevation represents an expres-sion of preconditioning [15] and this can be used as anindirect end point of protection in animals and inhumans. Using ST elevation as an index of protection,Szilvassy et al. found that pacing-induced precondition-ing is lost in hypercholesterolemic and atheroscleroticconscious rabbits [16]. Although we agree that cyclic-GMP may be involved in preconditioning [17] it is
D.T. Kremastinos et al. / Atherosclerosis 150 (2000) 81–89 85
Fig. 2. Serum samples from normally fed (first two from the left) and hypercholesterolemic rabbits (remaining samples) indicating the majordifference in macroscopic appearance of serum between the two types of diet.Fig. 3. Cross sections of the aortic wall from normal aorta (a), a case with early subintimal accumulation of lipids and macrophages (b), and acase (c) with increased deposition of lipids in the subintimal area (HE, ×400)
D.T. Kremastinos et al. / Atherosclerosis 150 (2000) 81–8986
Fig. 4. Myocardial branch of coronary vessel (a) with subintimal deposition of lipids and narrowing of the lumen (HE, ×400). The ostia of acoronary artery (b) with early atherogenesis (HE, ×400).
D.T. Kremastinos et al. / Atherosclerosis 150 (2000) 81–89 87
Table 2Infarct size and risk zone in the various study groupsa
Risk zone volume (cm3)Groups % I/RInfarct size (cm3)
A (hypercholesterolemic control) 1.3990.130.5690.11 39.396.31.5090.180.2890.09 16.793.9*B (hypercholesterolemic PC)
a PC, preconditioning; % I/R, infarct to risk ratio (%).* PB0.01 versus control.
Fig. 5. The effect of preconditioning on infarct size (expressed as a percent of the risk zone size) in the hypercholesterolemic (1st series) and normal(2nd series) rabbits. Groups A and C represent control while groups B and D represent preconditioned hearts.
questionable whether nitric oxide is responsible for thisincrease, as atrial natriuretic peptide has been alsoshown to increase cyclic-GMP in cultured vascular cellsand in balloon injured arteries [18]. There are a numberof differences between the Szilvassy’s study and ourown protocol. They used a 5 min rapid pacing period asthe global ischemic preconditioning stimulus, with a 5min reperfusion interval before a sustained 10 minrapid pacing period, as a surrogate for global sustainedischemia. In the present study, we used a more tradi-tional model of 5 min regional ischemia, with 10 minreperfusion interval as the preconditioning trigger be-fore sustained (30 min) regional ischemia. Furthermore,we used infarct size limitation as the final expression ofprotection; this is regarded as the gold standard end-point in preconditioning studies. Finally, we did not useST elevation and our rabbits were anesthetized and notconscious.
Hoshida et al. have shown in hypercholesterolemicrabbits that a nitric oxide donor ameliorates the sever-ity of myocardial injury after a period of sustainedischemia, with no prior preconditioning, and they con-
clude that less nitric oxide production is responsible forlarger myocardial infarcts [19]. They did not investigate
Fig. 6. Absolute infarct volume plotted against absolute risk zone size(cm3) in hypercholesterolemic rabbits. Open symbols depict individ-ual hearts from control group while closed symbols depict precondi-tioned hearts. Lines are from linear least square fits to the two sets ofdata.
D.T. Kremastinos et al. / Atherosclerosis 150 (2000) 81–8988
the effect of the nitric oxide donor in hypercholes-terolemic rabbits after a preconditioning stimulus butother studies have shown that inhibition of nitric oxidenot only does not abolish preconditioning [20] but infact reduces infarct size [21]. We have previously shownthat cyclic-GMP, a product of nitric oxide and otherprecursors [22], increases into the atheromatic humancoronary arteries during PTCA [23,24].
More than 10 years since the seminal study by Murryet al. [7], many factors have been proposed as triggersof preconditioning. Agonists binding to adenosine A1[25] and adrenergic a1 [26] receptors are known tocouple to protein kinase C and other kinases in themyocyte; the translocation of protein kinase C from aninactive cytosolic form to an active membrane formplays a key role in the mechanism of protection [27,28].Although production of mediators by the endotheliumis reduced in hypercholesterolemic conditions [6],protein kinase C is increased in the aortas of choles-terolemic rabbits [29]. Therefore, it would be reasonableto assume that PKC may also play a key role inischemic preconditioning in our hypercholesterolemicanimals. Hypercholesterolemia and atherosclerosis con-tribute to the induction of various autoantigens, such asheat shock proteins, which also protect the ischemicheart. Whether aortic and coronary artery rings arenitric oxide deficient or heat shock proteins enriched inhypercholesterolemic and atheromatic rabbits is thesubject of an other study.
In conclusion, our study reports for the first time thathypercholesterolemia and atherosclerosis do not abolishprotection against myocardial infarction in rabbits afterone cycle of 5 min regional ischemia. Our findingssuggest that the positive effect of preconditioning islikely to be present in humans with coronary arterydisease.
Acknowledgements
This work was made possible in part by a grant fromEEC (Biomed II Concerted Action BMH4-CT 95-0838).
References
[1] Libby P. Molecular bases of the acute coronary syndromes.Circulation 1995;91:2844–50.
[2] Mc Murray J, Rankin A. Treatment of myocardial infarction,unstable angina, and angina pectoris. Br Med J 1994;309:1343–50.
[4] Kloner RA, Bolli R, Marban E, Reionli L, Braunwald E. Medicaland cellular implications of stunning, hibernation and precon-ditioning. An NHLBI Workshop. Circulation 1988;97:1848–67.
[5] Vegh A, Szekeres L, Parratt JR. Preconditioning of the ischemicmyocardium; involvement of the L-arginine nitric oxide pathway.Br J Pharmacol 1992;107:648–52.
[6] Ross R. The pathogenesis of atherosclerosis; a perspective for the1990s. Nature 1993;362:801.
[7] Murry CE, Jennings RB, Reimer KA. Preconditioning withischemia: a delay of lethal cell injury in ischemic myocardium.Circulation 1986;74:1124–36.
[8] Marber MS, Latchman DS, Walker JM, Yellon DM. Cardiacstress protein elevation 24 hours after brief ischemia or heat stressis associated with resistance to myocardial infarction. Circulation1993;88:1264–72.
[9] Hale SL, Kloner RA. Effect of ischemic preconditioning onregional myocradial blood flow in the rabbit heart. CoronaryArtery Dis 1992;3:133–40.
[10] Ovize M, Kloner RA, Hale SI, Przyklenk K. Coronary cyclic flowvariations preconditioning ischemic myocardium. Circulation1992;85:779–89.
[12] Tomai F, Crea F, Danesi A, Perino M, Gaspardone A, Shini AS,Cascarano MT, Chiariello L, Gioffre PA. Mechanisms of thewarm-up phenomenon. Eur Heart J 1996;17:1022–7.
[13] Tomai F, Crea F, Gaspardone A, Versaci F, De paulis R, Pentade Peppo A, Chiariello L, Gioffre PA. Ischemic preconditioningduring coronary angioplasty is prevented by glibenclamide, aselective ATP-sensitive K+ channel blocker. Circulation1994;90:700–5.
[14] Leesar MA, Stoddard M, Ahmed M, Broadbent J, Bolli R.Preconditioning of human myocardium with adenosine duringcoronary angioplasty. Circulation 1997;95:2500–7.
[16] Szilvassy Z, Ferdinandy P, Szilvassy J, Nagy I, Karscu S, LonovicsJ, Dux L, Koltai M. The loss of pacing-induced preconditioningin atherosclerotic rabbits: role of hypercholesterolemia. J Mol CellCardiol 1995;27:2559–69.
[17] Iliodromitis EK, Papadopoulos CC, Markianos M, ParaskevaidisIA, Kyriakides ZS, Kremastinos DT. Alterations in circulatingcyclic guanosine monophosphate (c-GMP) during short and longischemia in preconditioning. Basic Res Cardiol 1996;91:234–9.
[18] Levin ER, Gardner DG, Samson WK. Natriuretic peptides. NewEngl J Med 1998;339:321–8.
[19] Hoshida S, Hishida M, Yamashita N, Igarashi J, Hori M,Kamada T, Kuzuya T, Tada M. Amelioration of severity ofmyocardial injury by a nitric oxide donor in rabbits fed acholesterol-rich diet. J Am Coll Cardiol 1996;27(4):902–9.
[20] Weselcouch EO, Baird AJ, Sleph P, Grover GJ. Inhibition of nitricoxide synthesis does not affect ischemic preconditioning isolatedperfused rat hearts. Am J Physiol 1995;268:H242–9.
[22] Schulz S, Chinkers M, Garbers DL. The guanlylate cyclase/recep-tor family of proteins. FASEB J 1988;3:2026–35.
[23] Iliodromitis EK, Karavolias GK, Markianos M, Kyriakides ZS,Vlahakos DV, Kremastinos DTh. Intracoronary and systemicrelease of the atrial natriuretic factor and cyclic-guanosinemonophosphate during coronary angioplasty. Lett Peptide Sci1996;3:221–6.
[24] Kremastinos DTh, Iliodromitis EK, Markianos M, Apostolou TS,Kyriakides ZS, Karavolias GK. Intracoronary cyclic-GMP andcyclic-AMP during percutaneous transluminal coronary angio-plasty. Int J Cardiol 1996;53:227–32.
[25] Liu GS, Thornton JD, VanWinkle DM, Stanley AWH, OlssonRA, Downey JM. Protection against infarction afforded bypreconditioning is mediated by A1 adenosine receptors in therabbit heart. Circulation 1991;84:350–6.
D.T. Kremastinos et al. / Atherosclerosis 150 (2000) 81–89 89
[26] Banerjee A, Locke-Winter C, Rogers KB, Mitchell MB, BrewEC, Cairns CB, Bensard DD, Harken AH. Preconditioningagainst myocardial dysfunction after ischemia and reperfusionby an alpha 1-adrenergic mechanism. Circ Res 1993;73:656–70.
[27] Liu Y, Tsuchida A, Cohen MV, Downey JM. Pretreatment withangiotensin II activates protein kinase C and limits myocardialinfarction in isolated rabbit hearts. J Mol Cell Cardiol1995;27:883–92.
[28] Iliodromitis EK, Miki T, Liu GS, Downey JM, Cohen MV,Kremastinos DTh. The PKC activator PMA preconditions rab-bit heart in the presence of adenosine receptor blockade:is 5%-nucleotidase important? J Mol Cell Cardiol 1998;30:2201–11.
[29] Sirikci O, Ozer NK, Azzi A. Dietary cholesterol-induced changesof protein kinase C and the effect of vitamin E in rabbit aorticsmooth muscle cells. Atherosclerosis 1996;126:253–63.
