Failure of superoxide dismutase and catalase to

Transcription

1 LABORATORY INVESTIGATION MYOCARDIAL ISCHEMIA Failure of superoxide dismutase and catalase to alter size of infarction in conscious dogs after 3 hours of occlusion followed by reperfusion KIM P. GALLAGHER, PH.D., ANDREW J. BUDA, M.D., DIANE PACE, B.S., RICHARD A. GERREN, PH.D., AND MARSHAL SHLAFER, PH.D Downloaded from by on September 29, 218 ABSTRACT Superoxide dismutase (SOD) and catalase (CAT), enzymes that degrade superoxide anion and hydrogen peroxide, respectively, reduce size of infarction in anesthetized, open-chest dogs subjected to coronary occlusion followed by reperfusion. To evaluate potential protective effects of these enzymes in conscious animals, three groups of dogs were instrumented at sterile surgery with a hydraulic occluder on the left circumflex (LCX) coronary artery, sonomicrometers to measure regional wall thickness, and catheters to monitor arterial and left ventricular pressures. Ten to 14 days after surgery, the animals were sedated with morphine sulfate (.5 mg/kg). The LCX artery was occluded for 3 hr by inflation of the hydraulic cuff. Infusions of SOD (n = 7), CAT (n = 6), or saline (control group, n = 7) were begun 15 min before reperfusion and lasted for 45 min of reperfusion. The doses of SOD and CAT were 5 mg/kg, dissolved in 6 ml of saline, and infused at a rate of 1 ml/min. Myocardial blood flow was measured with tracer-labeled microspheres (15 gm diameter) before occlusion, after 5 to 1 min of occlusion, after 15 min of occlusion, and 5 to 1 min after reperfusion. Size of infarction was measured 24 hr later by dual-perfusion staining with Evans blue and triphenyl tetrazolium. Size of infarction (expressed as a percentage of area at risk) did not differ significantly among the three groups: control, % (mean ± SD); SOD, %; CAT, %. Hemodynamic parameters and myocardial blood flows (measured before infusion of any agents) were not significantly different among the three groups. Serum SOD levels in SOD-treated dogs were 19 ± 2 gg/ml at the onset of reperfusion and 29 ± 3 gg/ml at the end of the infusion. Blood assays collected after infusion showed a monoexponential decay of SOD levels with a half-life of 22 ± 6 min. We conclude that myocardial protection by SOD or CAT is model dependent. In conscious dogs subjected to 3 hr of coronary occlusion followed by reperfusion, SOD and CAT failed to alter size of infarction. Circulation 73, No. 5, , CONSIDERABLE INTEREST has focused recently on reduction of myocardial reperfusion injury with enzymes that degrade oxygen radicals, which are highly reactive oxygen species that are toxic to biologic tissues. For example, Jolly et al.1 demonstrated an approximately 5% reduction in infarct size after 9 min of ischemia in anesthetized, open-chest dogs receiving combined superoxide dismutase (2- :2- oxidoreductase; EC ) and catalase (H22:H22 From the Thoracic Surgery Research Laboratory (Departments of Surgery and Physiology), Department of Pharmacology, and Department of Internal Medicine (Division of Cardiology), The University of Michigan Medical School, Ann Arbor. Supported in part by the American Heart Association of Michigan, NIH grants HL29499 and HL3243, NIH Research Career Development Award HL412, and Diagnostic Data, Inc., Mountainview, CA. Address for correspondence: Kim P. Gallagher, Ph.D., Thoracic Surgery Research Laboratory, R3484 Kresge I, Box 56, The University of Michigan, Ann Arbor, MI Received Nov. 8, 1985; revision accepted Jan. 3, Vol. 73, No. 5, May 1986 oxidoreductase EC ) beginning 15 min before reperfusion. Werns et al.,2 using a similar preparation, showed that myocardial protection in this experimental setting was due primarily to superoxide dismutase, there being no significant reduction in size of infarction when catalase alone was administered. Although these studies provided strong support for the view that oxygen radicals contribute importantly to regional myocardial reperfusion injury, the potential therapeutic efficacy of superoxide dismutase or catalase remained to be determined in an experimental preparation that may more closely simulate myocardial infarction in man. Consistent with the recommendation of the National Heart, Lung, and Blood Institute study on animal models for protecting ischemic myocardium, we chose to test the effectiveness of superoxide dismutase and catalase as infarct-reducing agents in conscious dogs.3 A 165

2 Downloaded from by on September 29, 218 GALLAGHER et al. 3 hr occlusion period, followed by 24 hr of reperfusion, was chosen as the setting in which to examine if the enzymes reduced size of infarction. Three hours of ischemia was used because this time corresponds reasonably well with the earliest average times patients are likely to undergo emergency revascularization (either surgically, mechanically, or with thrombolytic therapy) at our institution. In addition, considerable experimental data from conscious or closed-chest dogs exist on the effects of 3 hr of occlusion vs permanent occlusion.f8 Reperfusion was part of the protocol because oxygen radical-scavenging enzymes are thought to provide the most benefit when oxygen is restored to previously ischemic tissue. Methods Animal preparation. The study was performed in conditioned mongrel dogs of both sexes weighing between 18 and 27 kg. After induction with sodium thiamylal, anesthesia was maintained with halothane and a sterile thoracotomy was performed through the left fifth intercostal space of each dog. To measure left ventricular pressure, a Konigsberg (P7) high-fidelity micromanometer was introduced through the apex of the left ventricle. Tygon catheters were placed in the left atrium (for injection of microspheres) and left ventricle (via the apex for calibration of the micromanometer in mm Hg). The proximal circumflex artery was dissected free to allow placement of a hydraulic occluder, which was used to produce coronary occlusion. In some dogs a pulsed Doppler flow probe was placed on the circumflex artery proximal to the occluder to monitor coronary blood flow velocity. The signals from the flow probe were processed with a pulsed Doppler flowmeter constructed at the Dalton Research Institute, University of Missouri (D. Franklin, M. Caldwell; Columbia, MO). Regional myocardial function was evaluated with pairs of sonomicrometers arrayed to measure transmural wall thickness in myocardium supplied by the circumflex artery (ischemic area) to provide a sensitive on-line marker of ischemia. One crystal was inserted tangentially through the myocardium to the endocardium. The other crystal, attached to a Dacron patch, was sewn to the epicardium with shallow sutures after location of the position of least distance between the crystals while monitoring the signals with an oscilloscope.9 1 The dimension gauges were connected to a Triton (Model 12) sonomicrometer for signal processing. The position of the ultrasonic crystals was carefully examined at the time of autopsy to verify correct alignment. The thoracotomy was closed after bringing the wires and catheters subcutaneously to the back of the neck. The dogs were treated with antibiotics and given morphine sulfate (. 13 to.25 mg/kg im) for postoperative analgesia. Studies were performed 1 to 14 days after surgery, when the dogs exhibited normal activity and were afebrile. Experimental protocol. Experiments were carried out with animals lying on their right sides on a table in the laboratory. Morphine sulfate (.5 mg/kg im) was given for analgesia 6 min before coronary occlusion. Additional morphine sulfate (.13 mg/kg) was administered midway through the occlusion period if the dogs exhibited signs of discomfort or pain. A femoral cutdown was performed under local anesthesia to insert an arterial catheter for measurement of arterial blood pressure and for withdrawing a reference arterial sample for myocardial blood flow calculations. 166 The dogs were assigned randomly and blindly to treatment or control groups on the day of the study. The treatment was superoxide dismutase (5 mg/kg; Orgotein, Diagnostic Data, Mountain View, CA) or catalase (5 mg/kg; Calbiochem-Behring, La Jolla, CA). Control recordings were obtained and an injection of microspheres was performed. The left circumflex artery was abruptly occluded by inflation of the hydraulic occluder. Occlusion was verified by cessation of coronary blood flow velocity, prompt elimination of systolic wall thickening, characteristic increases in heart rate, and/or ST segment changes on the electrocardiogram. Five to ten minutes after coronary occlusion, a second injection of microspheres was made to document the extent of reduction of blood flow. The occlusion was maintained for 3 hr. A third injection of microspheres was made at 2.5 hr of occlusion to document any shortterm changes in collateral blood flow to the ischemic area. Fifteen minutes later (15 min before reperfusion), infusion of 6 ml of saline alone or saline containing superoxide dismutase or catalase was begun. The infusion was continued for 45 min after abrupt release of the coronary occluder. Lidocaine (5 mg iv) was administered 3 min before release of the occluder to minimize the likelihood of ventricular fibrillation. Reperfusion was verified by restoration of coronary blood flow velocity or observation of increases in end-diastolic wall thickness measured with the sonomicrometers. A fourth injection of microspheres was made 5 to 1 min after releasing the occluder to provide additional documentation that blood flow had been reestablished. The femoral cutdown was repaired and the animals were returned to their cages 1 to 3 hr after reperfusion. The following day the dogs were returned to the laboratory, deeply anesthetized with sodium pentobarbital, and killed with intravenous KC]. Each heart was rapidly excised and prepared for quantification of infarct size in vitro. Postmortem determination of infarct size. A dual perfusion technique was used to measure size of infarction as a percentage of the region at risk and of left ventricular mass Cannulas were placed in the aorta above the coronary ostia and in the circumflex artery at the site of the occlusion. The circumflex bed was perfused with 1.5% triphenyl tetrazolium hydrochloride (TTC) in potassium phosphate buffer at a pressure of approximately 1 mm Hg. The aorta was perfused simultaneously with.5% Evans blue dye at the same pressure to delineate the nonischemic myocardium. The heart was then cut into eight sections (7 to 1 mm thick) perpendicular to the apical-basal axis. The myocardium perfused by the circumflex artery constituted the region at risk of infarction and it was identified by the presence of a red color (due to the dehydrogenase conversion of the TTC to a red formazan) and absence of blue dye. The infarcted area was apparent as tissue that lacked staining (due to loss of dehydrogenase activity in infarcted tissue) within the region at risk. 14 We recognize that simultaneous dye perfusion of both normal and ischemic myocardium at a pressure that exceeds perfusion pressure of the ischemic zone in vivo may overestimate the area at risk. There was good agreement, however, between the delineation of the ischemic zone by blood flow measurements and that by the dual perfusion dye technique, suggesting that the degree of error was small. Any error that was encountered should have been similar across all three groups in the present study because the same technique was used in each experiment. In addition, this technique provides data that can be compared directly with those of several other investigatorsl who used similar methods. The sections of the left ventricle were traced onto clear plastic overlays and color photographs of each section were obtained to serve as a permanent record. Planimetry of the tracings was CIRCULATION

3 LABORATORY INVESTIGATION-MYOCARDIAL ISCHEMIA Downloaded from by on September 29, 218 performed to measure the size of the anatomic region at risk (as a percentage of the left ventricle) and size of infarction (as a percentage of the region at risk and left ventricular mass). Because previous studies have shown excellent correspondence between this method for quantifying infarct size and results of histologic examination of infarcts," no histopathologic evaluation of the tissue was performed in the present study. Myocardial blood flow measurements. Regional myocardial blood flow was measured with tracer-labeled microspheres (15 gm diameter, New England Nuclear) by the reference withdrawal method. 15 Four injections were made in the experiments, with the use of one of six available isotopes ('41Ce, 5'Cr, HI3Sn, 13Ru, 95Nb, 46Sc) for each flow determination. The choice of isotopes was determined by which isotopes were available at the time, and the order of their injection was randomized. Approximately 1 to 2 million microspheres were injected into the left atrium for measurement of blood flows. The reference arterial sample was obtained from the catheter inserted through the femoral artery into the aorta at a constant rate (7. ml/min) with a Harvard withdrawal pump; withdrawal of samples was initiated before the injection of microspheres and completed 2 min later. Each bottle of microspheres was thoroughly mixed by vortex agitation before injection, and droplets of the microsphere suspension were periodically examined under a microscope to ensure that adequate dispersal was being achieved. Two of the eight transverse rings of left ventricle (at approximately the midpapillary muscle level) were used for blood flow determinations after overlay tracings and photographs were obtained for quantification of size of infarction. Full-thickness sections were obtained from the central ischemic and control areas (at least 1 cm away from the boundary between areas delineated by TTC and Evans blue dyes). Each section of tissue was divided into three pieces of approximately equal thickness from the endocardial to epicardial surfaces. The location of each piece was recorded and then the tissue samples were weighed and placed in counting vials for assay of radioactivity in a Tracor (Model 1185) gamma scintillation counter. After correcting the counts in each tissue sample for background and overlapping counts with simultaneous equations, blood flow was calculated with the equation'5 Qm = (Cm x Qr)/Cr, where Qm = myocardial blood flow (ml/min); Cm = counts/min in tissue samples; Qr = withdrawal rate of the reference arterial sample (ml/min); Cr = counts/min in the reference arterial sample. Flow per gram of tissue was calculated by dividing flow by the weight of the appropriate sample. Background and overlap corrections and blood flow calculations were performed on an Apple Il Plus microcomputer. Assay for serum levels superoxide dismutase. Venous blood samples (approximately 1 ml each) were withdrawn into heparinized syringes immediately before the infusion of superoxide dismutase was begun and at 15 min intervals thereafter for 18 min. An equivalent volume of.9% NaCl was injected to replace the volume taken by the blood sample. Samples were also taken from one dog that received catalase and from one dog that received plain saline. Each sample was immediately centrifuged and the serum was carefully aspirated and transferred into coded plastic tubes that were capped and frozen at -8 C. The serum samples were sent, on dry ice, to Diagnostic Data, Inc. (Mountain View, CA) for assay. The samples were thawed and well mixed; superoxide dismutase standards were prepared from Orgotein 163/6 (Diagnostic Data, Inc.) diluted in 5% control sheep serum (Pel- Freez ). Dog serum samples, undiluted or diluted 1:5, and superoxide dismutase standards were electrophoresed in Tris glycine (ph 8.5)-buffered agarose gels and stained for superoxide dismutase activity by a modification of the nitroblue tetrazolium/riboflavin method of Beauchamp and Fridovich. 16 Superoxide dismutase levels were estimated independently by four individuals by visual comparison with the superoxide dismutase bands of the standards. There was no detectable superoxide dismutase activity in samples from catalase-treated or control dogs. Data analysis. Recordings were made during each experiment on an eight-channel Hewlett-Packard pressurized-ink recorder and on magnetic tape (Vetter, Model D) for subsequent analysis. Variables analyzed were heart rate, peak left ventricular systolic pressure, left ventricular end-diastolic pressure, peak positive and negative dp/dt, mean arterial blood pressure, pressure-rate product (heart rate x peak systolic blood pressure), regional systolic wall thickening, and myocardial blood flow in control and ischemic areas. Data were analyzed at the four following times: (1) immediately before coronary occlusion (control conditions), (2) 1 min after circumflex occlusion, (3) 2.5 hr after occlusion (15 min before infusion of saline or one of the two enzymes), and (4) 5 to 1 min after release of the circumflex occluder. A minimum of 1 cardiac cycles were averaged for each condition, with use of the beats occurring during the microsphere injections. The data were analyzed by digitizing the analog recordings on tape with the use of a DEC Micro PDP- 1 1 microcomputer system. Analysis of variance was used to analyze differences within groups across the four time periods.'7 When a significant time effect was observed, paired t tests were used to determine which time periods differed from one another. Because multiple comparisons were performed, a Bonferroni correction of the acceptable p level was used.'7 Routinely six comparisons were performed so the conventional p value (.5) was divided by six to give p <.83 as the minimum acceptable alpha level. When "p <.5" is indicated in the tables or text, this will refer to the "corrected value" actually corresponding to p <.83. The Bonferroni adjustment was also applied to comparisons between the control, superoxide dismutase, and catalase groups (during each experimental condition) that were made with unpaired t tests. Data are presented in tables, text, and figures as mean + SD. Results A total of 29 dogs were studied, 2 of which are included in this report: seven in the control group, seven in the superoxide dismutase group, and six in the catalase group. Four dogs of the nine excluded died from ventricular fibrillation between 4 and 2 min after coronary occlusion. Five dogs demonstrated no regional dimensional or hemodynamic changes after coronary occlusion because of extensive native collaterals (confirmed by microsphere measurements of regional blood flow) and were also excluded. None of the 2 dogs included in the analysis had ventricular fibrillation and all survived the full 24 hr after coronary occlusion. Hemodynamic parameters and regional wall thickening. Hemodynamic data are summarized in table 1. There were no significant differences in any parameter values among the three groups of dogs under control conditions or during coronary occlusion, indicating that we had achieved our objective of maintaining uniform hemodynamic conditions across the control, superox- Vol. 73, No. 5, May

4 GALLAGHER et al. Downloaded from by on September 29, 218 TABLE 1 Hemodynamic data under control conditions, early (1 min) and late after occlusion (2.5 hr), and early (1 min) after reperfusion Early Late Early Control occlusion p valuca occlusion p valuea reperfusion Heart rate (bpm) C (n = 7) ±25C < NS SOD (n = 7) ±26C NS NS B CAT (n = 6) B NS NS Left ventricular systolic pressure (mm Hg) C (n = 7) NS 126±21 NS SOD (n = 7) NS 141 ± 12 NS CAT (n = 6) NS 142±12 NS 14+2 Left ventricular end-diastolic pressure (mm Hg) C (n = 7) 11.1 ± B NS 17.6±5.2B NS 18.8±7.4B SOD (n = 7) NS NS CAT (n = 6) NS 19.2±5.2 NS Mean arterial blood pressure (mm Hg) C (n = 6) 17±15 11±12 NS 115±12 <.1 16±13 SOD (n = 7) NS 19± 11 NS 18±13 CAT (n = 6) 111±14 116±9 NS 11+9 NS 11±16 Peak positive dp/dt (mm Hg/sec) C (n = 7) NS NS 2627±37c SOD (n = 6) 3534± ±224 NS 3395± 158 NS 2965±549 CAT (n = 5) ±536 NS NS 2826±47 Peak negative dp/dt (mm Hg/sec) C (n = 7) ±428 NS 233+±537 NS 233 ±537 SOD (n = 6) ±22 NS 2769±468 NS CAT (n = 5) ±682 NS 2764±483 NS Pressure-rate product (LVSP x HR; x 13) C (n = 7) 11.91± ±2.86B NS 14.9±3.67 NS 14.5±4.42 SOD (n = 7) ± ±2.92B NS ±6.43 NS CAT (n = 6) 1.81± ±6.6B NS 13.95±2.89 NS 13.5± 1.9 t Data are mean ± SD. C = control group; SOD = superoxide dismutase group; CAT = catalase group; HR = heart rate. AFor difference between early and late occlusion and between occlusion and early reperfusion data (Bonferroni-adjusted paired test). Bp <.5; Cp <.1, for difference from control condition (Bonferroni-adjusted paired t test). There were no significant differences between C, SOD, and CAT groups during any condition. ide dismutase, and catalase groups. The only significant hemodynamic changes after coronary occlusion were increases in heart rate and pressure-rate product (table 1). There was a tendency for left ventricular enddiastolic pressure to increase as well, but this was statistically significant only in the control group. Early after reperfusion hemodynamic data were changed little compared with those recorded in the late occlusion period, but 17 of the 2 dogs exhibited ventricular arrhythmias at that time, making interpretation of the data difficult. The only significant differences among groups during the early reperfusion phase were in heart rate and the pressure-rate product between the superoxide dismutase and catalase groups. 168 Systolic wall thickening was successfully measured in the central ischemic area of four dogs in the control group, six dogs of the superoxide dismutase group, and five dogs of the catalase group. Before coronary occlusion systolic wall thickening was % in the control group; early after coronary occlusion, wall thickening was replaced by wall thinning ( %) that was not significantly different when measured at 2.5 hr of occlusion ( %). Similar patterns were observed in the superoxide dismutase and catalase groups (figure 1), indicating that the intensity of ischemia (measured as regional wall motion abnormalities) was similar in all three groups of dogs. Early after reperfusion no significant improvement in CIRCULATION

5 LABORATORY INVESTIGATION-MYOCARDIAL ISCHEMIA Downloaded from by on September 29, d WT (IS AREA) Ei -2OCULUN CLSIO CONTROL EARLY LATE -2 OCCLUSION OCCLUSION FIGURE 1. Systolic wall thickening under control conditions and early and late after coronary occlusion in the ischemic area (IS AREA). Wall thickness was measured with sonomicrometers. No significant differences were observed among the three experimental groups. dwt = end-systolic minus end-diastolic wall thickness; SOD = superoxide dismutase. TABLE 2 Myocardial blood flow in the central ischemic area wall thickening was evident. We present this finding cautiously, however, given the high incidence of postreperfusion arrhythmias encountered in the experiments. Myocardial blood flow. Blood flow data from the central ischemic and control areas are summarized in tables 2 and 3, respectively. Because of technical problems, blood flow data from one dog in the control group and one dog in the catalase group could not be used. Before coronary occlusion there were no significant differences between control and ischemic zone blood flow values within groups. There were also no significant differences among the control, superoxide dismutase, and catalase groups except with respect to midmyocardial blood flow (control vs superoxide dismutase group, p <.5). Five to ten minutes after coronary occlusion, there were substantial reductions in central ischemic blood flow that did not differ significantly among the three groups (table 2). The usual transmural perfusion gradient favoring subendocardial blood flow was reversed, with the subendocardium demonstrating the most dramatic flow reductions (-94% to -98%). Although reduced significantly, subepicardial perfusion was affected less severely (-57% to - 84%; table 2) by coronary occlusion. Simultaneously measured blood flow in the control zone was significantly elevated in all three layers early after coronary occlusion (table 3), consistent with the increases observed in heart rates and pressure-rate products (table 1). Late (2.5 hr) in the coronary occlusion period central ischemic zone blood flow remained depressed compared with that under control conditions, but small increases, possibly attributable to acute opening of collateral vessels, were evident. In the catalase group, subendocardial blood flow increased significantly (from.5.4too.1 +.6ml/min/g;p<.5) Early Late Early Control occlusion p valuea occlusion p valuea reperfusion Subendocardial MBF (ml/min/g) C.77 ±.17.3 ±.2c NS.4 ±.4c < ±.8 SOD 1.1 ±.31.2 ± O.Olc NS.4 ±.3c < ±.82 CAT ±.4c <.5.1±.6c < ±1.11 Midmyocardial MBF C ±.4c <.5.1±.7c < ±.81B SOD.97±.22.4±.4C NS.12±.12c < CAT ±.8c <.5.24±.12c < ± 1.43 Subepicardial MBF C ±.7 <.1.31 ±.8C < ±.8 SOD c <.1.27±.16B <.5.98±.17 CAT <.1.5±.22 < ±.73 Mean transmural MBF C ±.4 <.1.l15 ±.6c < ±.69 SOD.92±.21.6±.4c <.1.14A±.1c < ±.35B CAT ±.1B <.1.29±.12B <.1 2.± 1.3 Data are mean ± SD. MBF = myocardial blood flow; C = control group (n = 6); SOD = superoxide dismutase group (n = 7); CAT = catalase group (n - 5). AFor difference between early and late occlusion and between late occlusion and early reperfusion values (Bonferroni-adjusted t test). Bp <.5; cp <.1, for difference from control condition (Bonferroni-adjusted paired t test). Vol. 73, No. 5, May 1986 ff SAL INE 169

6 GALLAGHER et al. TABLE 3 Myocardial blood flow in the control (nonischemic) area Early Late Early Control occlusion p valuca occlusion p valuea reperfusion Subendocardial MBF C B NS 1.24±.42 NS SOD B NS 1.25±.37 NS CAT ±.51B NS 1.28±.47 NS Midmyocardial MBF C ±.35B NS NS SOD B NS NS CAT c NS NS Subepicardial MBF C B NS NS SOD ±.22B NS NS.9±.11 CAT ±.46c NS NS Mean transmural MBF C ±.34B NS NS 1.7±.63 SOD ±.24B NS NS 1.14±.11 CAT.83± ±.5c NS 1.14±.44 NS Data are mean ± SD. MBF = myocardial blood flow; C = control group (n = 6); SOD superoxide dismutase group (n = 7); CAT = catalase group (n = 5). AFor difference between early and late occlusion and between late occlusion and early reperfusion values (Bonferroni-adjusted t test). Bp <.5; Cp <.1, difference from control condition (Bonferroni-adjusted paired t test). Downloaded from by on September 29, 218 compared with the early occlusion value, but this level of subendocardial perfusion remained nonsignificantly different from that in the superoxide dismutase or control group. Midmyocardial blood flow increased in the control and catalase groups (table 2). Blood flow was augmented most appreciably in the subepicardium, where it approximately doubled (compared with early occlusion values) in all three groups. Control zone blood flow was not changed significantly compared with the early occlusion value (table 3). Five to ten minutes after reperfusion, blood flow in the central ischemic area was restored to control levels and there were no significant differences among the three groups. Although reperfusion flows tended to exceed control condition averages, no statistically significant differences from control were detected (with the exception of midmyocardial blood flow in the superoxide dismutase group) given the high level of vari- 3 ~~~ Lu ability in these data. 2 Part of the variability was likely related to the arrhythmias observed in most of the 1 experiments. Because the reperfusion data were obtained during nonsteady-state conditions, we present it only to provide additional documentation that reperfusion occurred after release of the coronary occluder. Control zone blood flow was not significantly different compared with that at late occlusion (table 3). Size of myocardial infarction. Left ventricular weights 17 were not significantly different among the three groups (control group, 186 ± 36 g; superoxide dismutase group, 176 ± 23 g; catalase group, 156 ± 2 g). As depicted in figure 2, the regions at risk in vitro delineated by the dual perfusion dye technique were quite similar across groups and did not differ significantly (control group, % of left ventricle; superoxide dismutase group, 4.7 ± 2.9%; catalase group, 41.9 ± 7.2%). Size of infarction, expressed as a percentage of the NS 6 U NS SALINE 6 ~~~~~NS NS NS NS SOD 5o CATALASE (D 4 4 ~~~~~~~~~~~~~NS ~~~~~~NS R.A.R. I.S. I.S. (% of LV) (% of R.A.R.) (% of LV) FIGURE 2. Data on region at risk and infarct size. The region at risk (R. A.R.) represents the percentage of the left ventricle (LV) not stained with Evans blue dye. Infarct size (I.S.) is expressed as a percentage of the region at risk and as a percentage of the left ventricle. No significant differences were observed among the three experimental groups. CIRCULATION

7 LABORATORY INVESTIGATION-MYOCARDIAL ISCHEMIA 1-1 E -J ` 3 A, 1 n o:3 2 _., region at risk, was 32.4 ± 16.5% in the control group, 38.4 ± 17.4% in the superoxide dismutase group, and 27.1 ± 16.6% in the catalase group. No significant differences were demonstrable, indicating that no myocardial protection (in the form of reduction of infarct size) was provided by superoxide dismutase or catalase in this study. Likewise, no significant differences were evident when size of infarction was expressed as a percentage of the left ventricle. The size of infarction measured in this manner in the control group averaged 14.6 ± 7.1% and that in the superoxide dismutase and catalase groups averaged 15.5 ± 6.9% and 13.3 ± 8.6%, respectively (figure 2). Blood levels of superoxide dismutase. In figure 3 are shown superoxide dismutase levels at different times during and after its administration. There was relatively little variability between animals and blood levels were near their peak when the coronary occlusion was released, allowing reperfusion to occur. That peak levels were attained near the onset of the reperfusion phase is important because it is early during reperfusion that oxygen radical-mediated damage is conventionally thought to be maximal. The plasma half-life of superoxide dismutase in this conscious animal preparation was 22 ± 6 min (n = 7). Analysis of variability of size of infarction. Because there were sizable standard deviations for the data on infarct size in our study, we evaluated the influence of size of the region at risk, pressure-rate product, and collateral blood flow on size of infarction.3 As shown in figure 4, there was no relationship evident between size of the area at risk and infarct size (r =.32, p = NS), largely because the size of the risk regions were quite similar within and across groups. The pressurerate product also failed to correlate closely with size of 3 a Downloaded from by on September 29, 218 E cm1 FIURE LU 3. Su REPERFUSE min FIGURE 3. Superoxide dismutase (SOD) levels at different times during and after its administration. Infusion of SOD (in 6 ml of saline) was initiated 15 min before reperfusion and continued for 45 min after reperfusion. The dosage was 5 mglkg. Peak levels were attained at or near the time of reperfusion. The blood half-life of the enzyme in conscious dogs was approximately 22 min. %/ l S INF LV WTJ * C o SOD * CAT 2 1 6) %IS 4 [ N F LRA R J2- - II a 31. U 3 4 RA R *i. U o a i m EQ U (r =_.32) 5 U (r-.23) P-R PRODUCT FIGURE 4. Infarct size (percentage of left ventricle) as a function of the size of the region at risk and infarct size (percentage of region at risk) as a function of the pressure-rate product. The correlations were not significant, in part due to the relatively narrow range of risk region and pressure-rate values encountered in the experiments. RAR = region at risk; P-R product = pressure-rate product; %IS = percentage infarct size; INF = infarct weight; LV WT = left ventricular weight; C = control group; SOD = superoxide dismutase group; CAT = catalase group. Vol. 73, No. 5, May

8 GALLAGHER et al. infarction (figure 4). Collateral blood flow, however, correlated with infarct size reasonably well (figure 5), especially the blood flow data obtained 2.5 hr after coronary occlusion. Although the collateral blood flow data could not account for all of the variability in size of infarction, we think collateral blood flow was the major contributing factor. Discussion Univalent reduction products of oxygen, some of which are free radicals, may contribute importantly to myocardial damage occurring with ischemia and reperfusion. These metabolites include superoxide anion (2-), hydrogen peroxide (H22), and hydroxyl radical (OH -). It is exceedingly difficult to assay them Downloaded from by on September 29, %I S 3 %I S 7 5. %I S 3o ~ ' co me', \ i i 'je \ jo EARLY 11 mini (r=.68).1 ENDO MBF 1 * \' * \% em1 \ C) \**' s \ * \'o O\ '.2.4 EPI MBF *C o SOD a CAT -7.2 %I S (r =. 7) %I S vw.2 MEAN MBF (r =.75) %I S * ' ' a *.3 \ \, 5 \ --. a (r-.66) EPI MBF \ (r (31\\ LATE 12.5 hrl.1 ENDO MBF 5 1WW1 1 i,.-.w-.2 MEAN MBF (r =.86) A ) FIGURE 5. Infarct size (percentage of region at risk) as a function of collateral blood flow (ml/min/g) in the central ischemic area. The relationship between infarct size (IS) and collateral blood flow is presented for blood flow at 1 min after occlusion (EARLY) and that at 2.5 hr after occlusion (LATE). Data on infarct size are plotted vs blood flow to the subendocardium (ENDO) and subepicardium (EPI), as well as average (MEAN) blood flow across the wall. Regression lines (dashed) and correlation coefficients are superimposed on the graphs. Linear regression was adequate to describe the relationships for EARLY after occlusion but cubic (ENDO) and quadratic (EPI and MEAN) equations were necessary for the LATE data. Because significant correlations between collateral blood flow and infarct size were evident, but 11o significant correlations were demonstrable between infarct size and region at risk or pressure-rate product, we conclude that the primary reason for differences in infarct size was variability in collateral blood flow. Regression equations: EARLY ENDO, y x ; EARLY EPI, y = x ; EARLY MEAN, y = x ; LATE ENDO, y = x x x ; LATE EPI, y = 121.6x x ; LATE MEAN, y = 598.9x x MBF myocardial blood flow; C = control group; SOD = superoxide dismutase group; CAT = catalase group CIRCULATION

9 LABORATORY INVESTIGATION-MYOCARDIAL ISCHEMIA Downloaded from by on September 29, 218 directly or to identify their sites of generation. Consequently, there is little definitive proof that they are involved in ischemic tissue damage. Nevertheless, indirect support for their involvement has been provided by studies showing reduction of global or regional ischemic or hypoxic damage" 2, 13, by enzymes such as superoxide dismutase or catalase, which selectively degrade superoxide anion and hydrogen peroxide, respectively. Other chemicals, such as antioxidants, iron chelators, or inhibitors of enzymes thought to generate oxygen metabolites (e.g., allopurinol inhibition of xanthine oxidase), also reduce ischemic or hypoxic damage in ways that are less selective biochemically but are nevertheless consistent with inhibition of known pathways of oxygen metabolite formation or reactivity. The effects of superoxide dismutase or catalase in the setting of experimental myocardial infarction were recently evaluated by Werns et al.2 Reduction of infarct size of approximately 5% was achieved in anesthetized, open-chest dogs that were subjected to 9 min of left circumflex coronary occlusion followed by reperfusion. Superoxide dismutase was infused at the same dose (5 mg/kg) we used, beginning 15 min before release of coronary occlusion. The potential benefit of superoxide dismutase was clearly demonstrated in the study by Werns et al.,2 and insights were provided on the mechanism of oxygen radical-mediated damage during reperfusion. Considered with the reports of Romson et al. and Jolly et al.,, the findings of Werns et al.2 suggest that neutrophil-derived cytotoxic oxygen metabolites, and superoxide in particular, act as extracellular contributors to ischemic myocardial damage. It was our objective, however, to evaluate the potential therapeutic efficacy of superoxide dismutase and catalase in a more physiologic and (potentially) clinically relevant experimental preparation of myocardial infarction. Although any experimental results from animals must be extrapolated to the clinical setting with extreme caution, we thought a conscious dog preparation exposed to a longer period of ischemia would constitute a more realistic and demanding test of the enzymes. This view also appears consistent with the report of the Animal Models for Protecting Ischemic Myocardium (AMPIM) group,3 which suggested that conscious preparations should be used for more rigorous tests of therapeutic efficacy. No significant differences among control, superoxide dismutase, or catalase groups were detected in terms of size of region at risk or, most importantly, size of infarction (figure 2). Therefore, the data dem- Vol. 73, No. 5, May 1986 onstrate no protective effect of either superoxide dismutase or catalase after 3 hr of coronary occlusion followed by reperfusion in conscious dogs. The failure to show a reduction in infarct size with either of the two enzymes could not be attributed to differences among the groups with respect to hemodynamic parameters, severity of wall thickening dysfunction, level of collateral blood flow early or late in the occlusion period, or blood flow 5 to 1 min after reperfusion. There are several major differences between our study and that by Werns et al.2 that could account for our failure to show significant reductions in size of infarction. Because our animals were conscious during the study, we used morphine for ethical reasons and lidocaine just before reperfusion to reduce the incidence of lethal arrhythmias. Although either drug might confer intrinsic cardioprotection and mask potential beneficial effects of superoxide dismutase or catalase, both are used routinely in humans for their cardiac and extracardiac actions. Therefore, their use in this study does not constitute a major deviation from a human protocol. It is not known with certainty whether in dogs 9 min of regional ischemia,2 rather than 3 hr, produces ischemic damage that is more comparable to that occurring during clinically relevant ischemic periods in man. Related to this is the possibility that damage produced by 3 hr of ischemia is not amenable to reduction by drug intervention. The 3 hr ischemic period that we used, however, corresponds well to the shortest average times that patients are likely to have suffered ischemia before undergoing emergency revascularization, either surgically or with thrombolytic therapy, at our institution. Earlier interventions, such as in the experimental study of superoxide dismutase by Werns et al.2 or the clinical study of streptokinase-induced thrombolysis by Koren et al.,25 may well increase the likelihood of myocardial protection. The methodologic difference that we think is most likely to explain the disparity between our results and those obtained by Wems et al.2 is the use of closedchest dogs. It is conceivable that trauma produced by thoracotomy could activate the complement system or neutrophil chemotaxis and production of oxygen radicals'3 to a degree that exceeds considerably the activation produced only by occlusion of a coronary artery. In the open-chest preparation, neutrophil-derived oxygen metabolites might be intercepted effectively by pharmacologic doses of drugs such as superoxide dismutase. In the absence of thoracotomy-induced trauma, generation of oxygen radicals by myocardial parenchymal or endothelial cells, or by neutrophils, may 173

10 Downloaded from by on September 29, 218 GALLAGHER et al. be inadequate to cause appreciable damage. It is also possible that our use of intravenous drugs rather than identical intra-atrial doses2 attenuated the delivery of effective drug concentrations to the myocardium. However, our data show that blood flow to ischemic myocardium was restored during reperfusion, and it is unlikely that superoxide dismutase or catalase concentrations in coronary artery blood would be appreciably lower than those measured in the peripheral circulation. Another possibility is that oxygen radical-mediated damage caused by prolonged ischemia may be so great, regardless of the cellular sources of the metabolites, that the drug doses used are inadequate to prevent it. Considerable variability was evident in the data on infarct size even though the regions at risk in vitro were very similar within and across groups (figure 2). Because hemodynamics were also quite similar, we think much of the variability was due to differences in collateral blood flow. Weak but statistically significant relationships existed between subendocardial, subepicardial, and mean transmural blood flow (in the central ischemic area) and size of infarction expressed as a percentage of the region at risk (figure 5). Although the catalase group was characterized by somewhat higher collateral flow than the other two groups, the data points from each group were distributed relatively uniformly, suggesting that variability in collateral flow was common to each experimental group. Unlike the AMPIM study group,3 we evaluated collateral blood flow in the subendocardium and transmurally (as well as in the subepicardium) early and late after coronary occlusion. Significant relationships were evident at both times, but there were substantial differences between the two. Consistent with the increases in collateral blood flow observed 2.5 hr after coronary occlusion (table 2), the regression lines were shifted rightward and they required nonlinear solutions. The data support the importance of early collateral blood flow as a determinant of ultimate size of infarction but also emphasize that later changes in collateral perfusion may occur in the acute infarction phase, contributing another potential source of variability. Previous studies in conscious dogs have also documented high levels of variability in infarct size. For example, Reimer et al.3 reported that infarct size after permanent occlusion averaged % (mean + SD) of the region at risk in dogs without any pharmacologic intervention. Their coefficient of variation was approximately 47%, similar to that which we observed. Baughman et al.7 reported that dogs with 3 hr of occlusion had infarcts that included approximately % of the left ventricle, with a coefficient of variation over 65% (table 2 in Baughman et al.7). Although they could demonstrate improvement in survival after 3 hr of occlusion followed by reperfusion compared with after permanent occlusion, the large degree of variability in the data made it impossible to demonstrate a significant difference between the groups in terms of size of infarction. Anesthetized dogs, on the other hand, have been reported to have much larger infarcts, on the average, after 3 hr of occlusion (26 ± 14% of the left ventricle, see Reimer et al. 3). Expressed as a percentage of the region at risk, anesthetized dogs had infarcts of 6 ± 27%,3 nearly twice that which we observed in our control group. Consequently, in contrast to the anesthetized, open-chest dog, the conscious dog preparation makes it difficult to demonstrate a significant drug effect on infarct size. The relatively small infarcts in the control group and substantial variability constitute limitations of the present study, and they may have contributed to our failure to demonstrate a significant reduction in size of infarction with superoxide dismutase or catalase. An important issue raised by our findings and those of other investigators' relates to the sources of oxygen radicals and the ability of the interventions to reach them. If neutrophils were the sole source of damaging metabolites, and they were activated to a lesser extent than occurs after thoracotomy, then superoxide dismutase should have protected the myocardium, as predicted by the data from Wems et al. That this drug did not reduce infarct size suggests that neutrophils did not act during the time of drug administration as a major source of oxygen radicals in our experimental preparation. Another possibility is that superoxide anion is not the major cytotoxic oxygen species in cardiac damage due to ischemia. Previous studies from our laboratories22 24 have failed to demonstrate protection by superoxide dismutase in experimental preparations in which other radical-related interventions, including catalase, have reduced damage. Since superoxide dismutase plus catalase reduces ischemic damage to blood-free isolated hearts, '9 endothelial or myocardial parenchymal cells may also act as significant sources of damaging oxygen metabolites via the action of xanthine oxidase,26 the mitochondrial electron transport chain, arachidonate metabolism, or other pathways. Both enzymes that we tested were accessible to the vascular face of the endothelium, and may have also gained access to the parenchymal side, so involvement of the endothelium as a source of oxygen metabolites is questionable in this preparation. An alternative explanation. therefore, is that the large mo- CIRCULATION

11 LABORATORY INVESTIGATION-MYOCARDIAL ISCHEMIA Downloaded from by on September 29, 218 lecular weights of superoxide dismutase (32,) and catalase (24,) limited access to the sarcolemma, where they might degrade superoxide or hydrogen peroxide that diffuses from parenchymal cells, or access through the sarcolemma to the sarcoplasm, where they might degrade their substrates closer to intracellular sites of generation. The nature of our experimental preparation precludes determination of which, if any, of these possibilities is correct, and further investigation will be required to address these issues. One interpretation of our data is that cytotoxic oxygen metabolites did not participate in the ischemic damage observed in our closed-chest model. Although we cannot conclusively reject this possibility, substantial evidence exists in the literature to implicate oxygen radicals in diverse preparations of cardiac ischemic or hypoxic damage. 2,1824 In addition, coronary occlusion in closed-chest dogs rarely produces the total ischemia and complete anoxia that would be necessary to completely abolish oxygen metabolism. The data from our study show that ischemia was not complete since some collateral blood flow to the ischemic zone persisted (table 2). This indicates that the tissue was not anoxic, but rather hypoxic. Therefore, an environment may have existed that was favorable for univalent oxygen reduction. There was substantial time (2.75 hr before drug intervention) for the development of damage by a free radical-mediated process that would not have been influenced by administration of superoxide dismutase or catalase much later. If this is correct, it incriminates oxygen radicals as contributors to myocardial damage in our preparation and emphasizes the need for early drug intervention. We propose that significant myocardial damage mediated by oxygen radicals can occur during a prolonged ischemic period, such as in our preparation, and is not necessarily restricted to the reperfusion phase (the focus of most experimental studies). In deciding whether superoxide dismutase or catalase can effectively limit size of infarction, a key question to answer is which experimental preparation best simulates myocardial infarction in man. Unfortunately, an unequivocal answer to that question is not currently available. It is our conclusion that the effects (or lack of effects) of these enzymes on infarct size are strongly model dependent. An implication of this conclusion, in our view, is that the enzymes superoxide dismutase or catalase may prove useful only in selected circumstances and patients, which are yet to be adequately defined. We appreciate the technical assistance of Thomas B. McClanahan, Russell A. Grinage. and Mila Frurla, and thank Tarry Goble for word processing of the manuscript. We also thank Dr. J. Hurley Myers (Department of Physiology and Pharmacology, Southern Illinois University School of Medicine) for advice and counsel in planning the experiments. We thank Diagnostic Data, Inc. for the supply of superoxide dismutase and analysis of blood samples for superoxide dismutase. References 1. Jolly SR, Kane WJ, Bailie MB, Abrams GD, Lucchesi BR: Canine myocardial reperfusion injury. Its reduction by the combined administration of superoxide dismutase and catalase. Circ Res 54: 277, Wems SW, Shea MJ, Driscoll EM, Cohen C, Abrams GD, Pitt B, Lucchesi BR: The independent effects of oxygen radical scavengers on canine infarct size. Reduction by superoxide dismutase but not catalase. Circ Res 56: 895, Reimer KA, Jennings RB, Cobb FR, Murdock RH, Greenfield JC Jr, Becker LC, Bulkley BH, Hutchins GM, Schwartz RP, Bailey KR, Passamani ER: Animal models for protecting ischemic myocardium: results of the NHLBI cooperative study. Comparison of unconscious and conscious dog models. Circ Res 56: 651, Ginks WR, Sybers HD, Maroko PR, Covell JW, Sobel BE, Ross J Jr: Coronary artery reperfusion. II. Reduction of myocardial infarct size at 1 week after the coronary occlusion. J Clin Invest 51: 2717, Costantini C, Corday E, Lang T-W, Meerbaum S, Brasch J, Kaplan L, Rubins S, Gold H, Osher J: Revascularization after 3 hours of coronary arterial occlusion: effects on regional cardiac metabolic function and infarct size. Am J Cardiol 36: 368, Schaper W: Residual perfusion of acutely ischemic heart muscle. In Schaper W, editor: The pathophysiology of myocardial perfusion. New York, 1979, Elsevier/North Holland Biomedical Press, pp Baughman KL, Maroko PR, Vatner SF: Effects of coronary artery reperfusion on myocardial infarct size and survival in conscious dogs. Circulation 63: 317, Lavallee M, Cox D, Patrick TA, Vatner SF: Salvage of myocardial function by coronary artery reperfusion 1, 2, and 3 hours after occlusion in conscious dogs. Circ Res 53: 235, Sasayama S, Franklin D, Ross J Jr, Kemper WS, McKown D: Dynamic changes in left ventricular wall thickness and their use in analyzing cardiac function in the conscious dog. Am J Cardiol 38: 87, Gallagher KP, Matsuzaki M, Koziol JA, Kemper WS, Ross J Jr: Regional myocardial perfusion and wall thickening during ischemia in conscious dogs. Am J Physiol 247(Heart Circ Physiol 16):H727, Romson JL, Bush LR, Jolly SR, Lucchesi BR: Cardioprotective effects of ibuprofen in experimental regional and global myocardial ischemia. J Cardiovasc Pharmacol 4: 187, Kirlin PC, Romson JL, Pitt B, Abrams GD, Schork MA, Lucchesi BR: Ibuprofen-mediated infarct size reduction: effects on regional myocardial function in canine myocardial infarction. Am J Cardiol 5: 849, Romson JL, Hook BG, Kunkel SL, Abrams GD, Schork MA, Lucchesi BR: Reduction of the extent of ischemic myocardial injury by neutrophil depletion in the dog. Circulation 67: 116, Fishbein MC, Meerbaum S, Rit J, Lando U, Kanmatsuse K, Mercier JC, Corday E, Ganz W: Early phase acute myocardial infarct size quantification: validation of the triphenyl tetrazolium chloride tissue enzyme staining technique. Am Heart J 11: 593, Heymann MA, Payne BD, Hoffman JIE, Rudolph AM: Blood flow measurements with radionuclide-labeled particles. Prog Cardiovasc Dis 2: 55, Beauchamp C, Fridovich I: Superoxide dismutase: improved assays and an assay applicable to acrylamide gel. Anal Biochem 44: 276, Morrison DF: Multivariate statistical analysis. New York, 1976, McGraw-Hill, pp Shlafer M, Kane PF, Kirsh MM: Superoxide dismutase plus catalase enhances the efficacy of hypothermic cardioplegia to protect the globally ischemic reperfused heart. J 83: Thorac Cardiovasc Surg Vol. 73, No. 5, May