INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

ALTHOUGH REDUCING THE AMOUNT of infarcted myocardium remains a primary goal of treatment for acute myocardial infarction, substantial evidence has accumulated suggesting that the efficiency and quality of tissue repair also play an important part in determining long-term ventricular function and mortality after infarction (32). Among the interventions designed to promote more effective tissue repair, the earliest data were associated with reduction of cardiac afterload and ventricular work. The use of angiotensin-converting enzyme (ACE) inhibitors has now become standard therapy for treatment of large transmural myocardial infarctions. More recently, it has been demonstrated that ACE inhibitors may also favorably influence tissue repair by direct inhibition of localized renin-angiotensin systems in the heart (39).

Over the same historical period, reperfusion of the previously occluded coronary vessel became the standard therapy to reduce tissue injury. Initially, the focus of this effort was to open the coronary artery in an attempt to decrease the degree of infarction. In subsequent studies, however, it also became apparent that reperfusion, even instituted later when tissue infarction could no longer be reduced, effectively promoted tissue repair (15). The combination of an open vessel and the possibility for promoting more effective tissue repair by pharmacological intervention led to a variety of studies assessing factors that play critical roles in tissue repair. It is hoped that methods of augmenting tissue repair will follow.

One of the effects of reperfusion of the infarcted myocardium is an accelerated inflammatory reaction (5) that, under certain circumstances, can actually extend injury (6). However, substantial evidence suggests that leukocyte influx into the previously infarcted area may help promote tissue repair (28) by enhancing phagocytosis (3, 34) as well as production of growth factors and cytokines that may modulate scar formation and angiogenesis (9, 36). Our laboratory and others (1, 21) have described the early influx of monocytes into the infarct area on reperfusion.

Studies in larger animals that describe the cellular and molecular events associated with tissue repair and scar formation are inevitably limited by the ability to deduce the relative importance of the factors described. Treatment with neutralizing monoclonal antibodies over several weeks, even if such antibodies were available, would be of enormous expense and doubtful practicality. For this reason, our laboratory has produced models of chronic myocardial infarction in mice (27). Infarction is created by permanent coronary artery occlusion or by occlusion for a period of 1-2 h, after which the occlusion is removed and coronary flow to the myocardium is restored. These models allow use of genetic modification strategies in the study of the molecular and cellular events associated with the reaction to injury. The models yield mice capable of surviving for long periods after infarction and allow hemodynamic measurements. The purpose of the current study was to characterize repair in mice with myocardial infarction and to define the effects of reperfusion on ventricular remodeling and systolic and diastolic function. The results suggest that, in the absence of reperfusion, infarcts of equivalent size undergo greater expansion that can be directly related to greater ventricular hypertrophy. These anatomic features of remodeling are associated with decreased systolic and diastolic function.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Infarct and reperfusion model. Male C57BL/6 mice 12-16 wk of age (22.5-30.5 g body wt) were obtained from Harlan Sprague Dawley (Houston, TX). All animal protocols were approved by the Institutional Animal Care and Use Committees of the American Association for Accreditation of Laboratory Animal Care-accredited institutions Baylor College of Medicine and Lilly Research Laboratories, in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals [DHHS Publication No. (NIH) 85-23, Revised 1985, Office of Science and Health Reports, Bethesda, MD 20892]. The following brief description is excerpted from a more detailed account (27). Anesthesia was produced by an intraperitoneal injection of pentobarbital sodium (4 mg/ml; 10 µl/g body wt). Mice were placed in a supine position with paws taped to the operating table. With direct visualization of the trachea, an endotracheal tube was inserted and connected to a Harvard rodent volume-cycled ventilator cycling at 100 min¹ with volume sufficient to adequately expand the lungs but not overexpand. The inflow valve was supplied with 100% oxygen.

Doppler ultrasound measurements. Mice were anesthetized by an intraperitoneal injection of a mixture (acepromazine 1.4 mg/ml, xylazine 8.6 mg/ml, and ketamine 42.8 mg/ml) at a dose of 0.5 µl/g body wt. They were taped to a temperature-controlled laminated plastic board with copper electrodes placed such that the three bipolar limb leads allowed ECG monitoring. Body fur at the left lower sternal border was clipped lightly, and the skin in that area was wetted with warm water to improve sound transmission. A 10-MHz pulsed Doppler probe was positioned at the xiphoid process of the sternum with minimal pressure. The pulsed Doppler range gate depth was set at 4-7 mm to obtain optimal signals from the left ventricular (LV) inflow and outflow tracks substernally as previously described (12, 13). An ECG timing signal was superimposed on the Doppler display using an R wave trigger. Repeated measures were made from each animal to allow for observation at different heart rates and to ascertain the reproducibility of the measurements. For each study, four to six beats were analyzed. The pulsed Doppler instrument and probes were custom made in our laboratories (13), and the Doppler audio signals were processed by a Medasonics (Vasculab SP-25A) spectrum analyzer for subsequent analysis on personal computer using NIH Image. In later studies, the Doppler signals were processed by a high-speed spectrum analyzer and software (Indus Instruments, Houston, TX) specifically optimized for use with mice. In each case the observer was blinded to the experimental conditions. Data were quantified in a Lotus 1-2-3 spreadsheet. Mice were allowed to recover and were followed sequentially, comparing pre- and postprocedure measurements in the same mouse. No deaths were caused by anesthesia in preocclusion or postocclusion Doppler studies.

Ventricular ejection fraction by cardiac nuclear imaging. For murine imaging a multiwire gamma camera (10) optimized for use with ¹⁷⁸Ta was fitted with a 2-mm-diameter pin-hole lens to project an enlarged image of the mouse heart on the image plane at a frame rate of 160 s¹ (Proportional Technologies, Houston, TX). ¹⁷⁸Ta with a half-life (T1/2) of 9.3 min was generated and concentrated on site as needed from ¹⁷⁸W (T1/2 = 21.7 days). The short T1/2 allowed concentration of high radiation doses (20 mCi) in a very small volume (10 µl) to maintain image quality and sequential individual studies timed ~30 min apart.

Assessment of area at risk and infarct size. At 24 h after coronary artery occlusion or occlusion-reperfusion event, hearts were removed and area at risk and infarct size were measured as previously described (27). Briefly, hearts were removed after euthanasia by intravenous injection of 50 mM KCl, 5 mM EGTA, and 10 U/ml heparin. The aorta was cannulated with a 22-gauge Luer stub, and 50 µl of 1% Evans blue were perfused into the aorta and coronary arteries with distribution throughout the ventricular wall proximal to the site of coronary artery ligation. The transverse section cutter used in these studies had blades spaced 1,000 µm apart; therefore, each heart section between the blades had section thicknesses of 1,000 µm, whereas the apical tip and base may vary outside these thicknesses because the first cutter blade was positioned at the site of coronary artery occlusion. In the typical 12- to 16-wk-old mouse, this cutting gave three equal 1,000-µm-thick pieces, an ~2,000-µm-thick basal section, and an apical section generally <1,000 µm thick.

Perfusion fixation and histology. For assessment of changes in cardiac structure, the intravenous cardioplegic solution contained (in g/l) 4.0 NaCl, 3.7 KCl, 1.0 NaHCO₃, 2.0 glucose, 3.0 2,3-butanedione monoxime, 3.8 EGTA, and 0.0002 nifedipine. This solution was perfused through the jugular vein to promote quiescence and relaxation. After being excised and rinsed in cold cardioplegic solution, the aorta was cannulated with a 22-gauge blunt needle and the left atrial appendage with a PE-50 catheter pushed across the mitral valve into the LV and secured in place. Hearts were fixed for 10 min by aortic perfusion of 10% zinc-buffered Formalin with the LV drain held vertically 16 cm above the base of the heart to maintain constant LV pressure at ~16 cmH₂O. After the cannulas were removed, fixation by immersion continued for 2 h, after which the hearts were dehydrated, cleared, and embedded in paraffin. The entire heart was cross-sectioned into 5-µm sections from base to apex (see Fig. 1). Sections were stained with hematoxylin and eosin or Masson trichrome stain.

View larger version (46K): [in this window] [in a new window]   Fig. 1.   Representative mouse heart imaging schematic illustrating that a 5.5-mm heart is cross-sectioned into 1,100 5-µm-thick sections, from which 23 are chosen for analysis of ventricular mass and dimension. LAD, left anterior descending coronary artery.

Analysis of ventricular mass and scar formation. The mass of the LV and septum was obtained by integration of a curve created by plotting the LV area measured in each cross section against the distance of the cross section from the base of the heart (see Fig. 1). The basal cross section used as the starting point in each heart was identified as the last cross section containing any part of the aortic valve and every 50th cross section (intervals of 5 µm × 50 = 250 µm) was analyzed. The sum of the areas in each of the partitions was obtained by integration of this curve that was carried out by dividing the interval of integration (x-axis) into 250-µm partitions representing each of the cross sections, beginning at the first basal cross section and ending at the last apical cross section. The sum was then used as an estimate of total ventricular space occupied by the myocardium in each heart. LV myocardial mass was derived from this space estimate by multiplying by the commonly assumed muscle tissue density of 1.065 mg/µl. The mass of the interventricular septum was separately analyzed using identical methods. In the latter case, the LV septum was described in the first basal section and continued through the last section to show evidence of the RV chamber. The edges of the septum were defined by the approximate sites of the anterior and posterior interventricular (longitudinal) sulci on each cross section. The actual line of demarcation was arbitrarily drawn on each section with consideration of the sites of these sulci and with the understanding of a slight rightward (toward RV) convexity of the interventricular septum.

Computation of ventricular expansion ratio. In each cardiac cross section, a thin segment was identified in the LV free wall. The thickness of these segments was characteristically <40% of the remaining ventricular free wall in the cross section. The thickness of thin segments varied <10% when compared in all of the cross sections from a single heart. The length of the endocardial surface underlying each thin segment was measured as well as the length of the total endocardial surface in the cross section. Thin segment length and total endocardial surface length were plotted against the distance of the cross section from the base of the heart. (see RESULTS). Each plot was integrated by dividing the interval of integration (x-axis) into 250-µm partitions representing the distance between the cross sections taken for analysis, beginning at the first basal cross section and ending at the last apical cross section, and obtaining the sum of the areas in each of the partitions. The sums were used as estimates for the thin segment area and for the total endocardial surface area in the LV of each heart. The ventricular expansion ratio was calculated by dividing thin segment endocardial area by total endocardial area and multiplying by 100.

Statistics. Differences between groups were determined using analysis of variance. Specific group differences were determined using an unpaired t-test with Dunnett's correction for multiple comparisons. The data not representing normal distribution were tested by nonparametric rank-sum tests. Correlation between variables was tested by linear regression analysis. Values listed are means ± SE. For all testing, P < 0.05 was used to determine significance.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

A total of 35 mice successfully completed the 24-h infarct size protocol, which represented 95% survival for sham, 75% for permanent occlusion, and 60% for occlusion-reperfusion mice. The hemodynamic assessment with ultrasound and nuclear imaging and the ventricular dimension experiments necessitated an additional 22 sham, 32 permanent occlusion, and 24 occlusion-reperfusion mice; this represented survival averaging 92% for sham, 60% for permanent occlusion, and 45% for occlusion-reperfusion mice.

Infarct size. Our laboratories have used a coronary occlusion site on the mouse that results in an infarction of ~30% of the LV (27). Mice reperfused after 30 and 60 min of occlusion demonstrated significantly reduced myocardial infarct size compared with permanent occlusion mice. In the current study, we examined and compared tissue repair under conditions of identical infarct size. In reperfused mice, we measured infarct size as a function of occlusion time with the intent to determine the time of reperfusion that would not result in a significant reduction in myocardial infarct size. When the coronary artery was occluded for 120 min, myocardial infarct size was identical to that seen with a permanent occlusion, resulting in an infarct of 30% of the LV, representing ~65% of the area at risk (Fig. 2). The remainder of the study, therefore, compared hemodynamic and anatomic consequences of a permanent occlusion with those of an occlusion of 2 h followed by reperfusion.

View larger version (33K): [in this window] [in a new window]   Fig. 2.   Infarct size and area at risk after LAD occlusions of 30 (n = 5), 60 (n = 7), and 120 (n = 10) min followed by reperfusion or permanent LAD occlusion (n = 13). Hearts were removed at autopsy 24 h after occlusion. Data for permanent occlusion and 30- and 60-min occlusion followed by reperfusion are from Ref. 27. Values are means ± SE. * P < 0.05 vs. permanent occlusion. LV, left ventricle.

Hemodynamic function. Using a nuclear imaging camera developed in our laboratory (10), we were able to measure ejection fractions in mice at various times after permanent occlusion or reperfusion. As previously described (11), acute occlusion of the coronary vessel reduces ejection fractions to values commonly below 30%. The ejection fractions after permanent occlusion or occlusion and reperfusion appeared to be stable from 2 wk to several months after infarct and therefore were combined. The ejection fraction was 26.4 ± 3.1% (n = 23) for mice with permanent occlusions and was almost identical (24.1 ± 4.0%; n = 10) for the reperfused group; sham-operated mice had significantly higher ejection fractions of 47 ± 2.8% (n = 10; P < 0.05). Representative end-diastolic and end-systolic images from a normal, nonoperated mouse and a mouse with a permanent coronary artery occlusion are shown in Fig. 3. In this example, the ejection fractions are 60% in the normal mouse and 17% in the mouse with the permanent occlusion. The resolution of the images was sufficient to appreciate the dilated and poorly contracting infarcted cardiac apex compared with the motion of the same region in the normal mouse and to recognize the heterogeneity in wall motion within a given ventricle.

View larger version (22K): [in this window] [in a new window]   Fig. 3.   First-pass nuclear images of LV blood pool taken at end diastole and at end systole from a normal mouse and from a mouse with a permanent LAD occlusion. Images were taken with multiwire gamma camera fitted with a 2-mm pin-hole lens at a rate of 160 frames/s. 178Ta was concentrated to 20 mCi in 10 µl of fluid, injected as a bolus into jugular vein, and flushed with 10 µl of 0.9% NaCl. Occluded mouse has a larger ventricular cavity with reduced ejection fraction (EF) and with dilatation and paradoxical wall motion at apex. HR, heart rate; bpm, beats per minute.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   A: peak aortic flow velocity followed for 5-6 mo in mice subjected to sham operation (), 2-h occlusion followed by reperfusion (), and permanent occlusion (). Data are % of preoperative values and are expressed as means ± SE. Preoperative values: sham (n = 15), 104 ± 20 cm/s; permanent occlusion (n = 24), 111 ± 17 cm/s; reperfusion (n = 13), 102 ± 10 cm/s. * P < 0.05, permanent occlusion vs. sham; # P < 0.05, reperfusion vs. sham. B: average aortic flow velocity followed for 5-6 mo in mice subjected to sham operation, 2 h of ischemia followed by reperfusion, and permanent occlusion. Data are % of preoperative values and are expressed as means ± SE. Preoperative values: sham (n = 15), 74 ± 16.9 cm/s; permanent occlusion (n = 24), 81 ± 15.0 cm/s; reperfusion (n = 13), 69 ± 5.9 cm/s. * P < 0.05, permanent occlusion vs. sham; # P < 0.05, reperfusion vs. sham.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Peak early filling velocity followed for 5-6 mo in mice subjected to sham operation (), 2-h ischemia followed by reperfusion (), and permanent occlusion (). Data are % of preoperative values and are expressed as means ± SE. Preoperative values: sham (n = 15), 69 ± 3.1 cm/s; permanent occlusion (n = 24), 71 ± 1.9 cm/s; reperfusion (n = 13) 66 ± 2.1 cm/s. * P < 0.05 vs. sham.

Ventricular dimensions. We characterized the anatomic changes associated with tissue repair and adaptation to the injury. LV from sham-operated and infarcted mice were examined over the period from 2 to 14 mo after infarct. This was the period during which the postinfarct functional changes in both systolic and diastolic parameters remained stable and clearly distinguished mice reperfused after coronary occlusion from those in which the occlusion was permanent. Our intent was to characterize the anatomic counterparts of these differences.

View larger version (58K): [in this window] [in a new window]   Fig. 6.   Cross sections of representative mouse hearts at 5 sites (basal to distal) illustrating appearance of sham, permanent occlusion, and occlusion-reperfusion mice.

View larger version (14K): [in this window] [in a new window]   Fig. 7.   Septal mass is plotted as a function of expansion ratio (length of thinned endocardial surface/total endocardial surface × 100). For permanent occlusion ( and dotted line, n = 8), regression line was y = ax + b, r = 0.85, P < 0.007; for reperfusion (, n = 7), there was inadequate distribution for regression analysis; for combined permanent occlusion and reperfusion (solid line) regression, r = 0.675, P = 0.006;  = sham ± SD (n = 9).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The ability to treat the reperfused infarct directly suggests that there might be an opportunity to favorably influence tissue repair in the reperfused myocardial infarct, but a greater understanding of the variables associated with this process is necessary. Taking advantage of the ability to genetically manipulate the mouse, it may be possible to further clarify cellular and molecular mechanisms of tissue responses in the tissue repair after myocardial infarction. Mice with genetic deletions of adhesion molecules (24, 40), cytokines (20), and mast cells (46) have been developed and examined with respect to their effects on inflammatory fibrogenic influences.

We examined the physiological and related anatomic effects of late-reperfusion myocardial tissue remodeling in two types of myocardial infarction in the mouse. We considered release of occlusion at 2 h to be late because our studies concluded that this was the minimal coronary artery occlusion time necessary to produce infarcts equal to those seen with a permanent occlusion.

Recently, functional and anatomic consequences of ventricular remodeling after permanent coronary artery occlusion in mice (30) and partial occlusion (23) were described. In the current study, we have elected to pursue both the physiological and the anatomic consequences of tissue repair in the presence and absence of reperfusion. To our knowledge, this represents the first attempt to characterize the effects of reperfusion on both the functional and anatomic consequences of cardiac remodeling in mice.

The protocol for permanent coronary occlusion and late reperfusion used in this study resulted in identical myocardial infarct size at 24 h. The functional studies demonstrated that basal ejection fraction was not different between the two groups. Despite the absence of significant differences in basal ejection fractions, both systolic and diastolic function of the permanently occluded ventricle were significantly impaired compared with either sham-operated or reperfused ventricles. In fact, after 2 wk following occlusion-reperfusion, abnormalities were not apparent in either systolic or diastolic function in the reperfused ventricles despite the fact that their mean ejection fraction was reduced by ~50% (see Fig. 4, A and B, and Fig. 5). These data suggest that the mice with permanent coronary occlusion acquired a functional handicap, in addition to the reduction in ejection fraction, that decreased their ability to adjust to unrelieved ischemic myocardial injury.

The anatomic data (Fig. 7) confirmed this mechanical disadvantage. The permanently occluded mice demonstrated both higher expansion ratios and greater compensatory ventricular hypertrophy than mice with reperfused infarcts. A highly significant correlation between the ventricular expansion ratio and the degree of septal hypertrophy was obtained, suggesting a positive relationship between the degree of ventricular hypertrophy and the area of dyskinetic myocardium. The relationship between these parameters did not appear to be a function of reperfusion; late reperfusion appeared to exert its effect by reducing ventricular expansion.

From a functional point of view, animals with greater hypertrophy and larger dyskinetic segments would be expected to have reduced ventricular compliance and reduction in ventricular filling velocity as demonstrated in Fig. 3. Similarly, reduction in aortic ejection velocity might result from mechanical disadvantages imposed by the thin segment or by the necessity for generating a higher preload to fill the less compliant ventricle.

In patients, survival of acute myocardial infarction after the initial hospital phase is largely dependent on residual ventricular function. Although reduction of myocardial infarct size after rapid reperfusion has become the hallmark of successful intervention, infarct size is not the only variable that affects postinfarct function. Ventricular unloading and inhibition of angiotensin II formation or signal transduction (which also unloads the ventricle) both modify resultant ventricular anatomy associated with infarction and ventricular repair (33, 39). Pfeffer and Braunwald (32) emphasized the prognostic importance of the geometric disadvantage associated with infarct expansion in patients.

Although the permanently occluded and the reperfused animals showed a similar reduction in ejection fraction, peak aortic flow velocity was decreased after 2 wk only in the permanently occluded group. LV ejection fraction is clearly dependent on infarct size (32), but the relationship between ejection fraction after infarction and maximum cardiac output or exercise capacity is not as clear. Franciosa et al. (7) described individuals with very low LV ejection fractions who were capable of very vigorous exertion. Similarly, peak aortic flow velocity and ejection fraction may vary independently. There are reports of overlap between peak aortic flow velocity values at rest for normal subjects and those with considerable heart disease, reduced ejection fractions, or experimental ischemia (16, 17, 26, 38). Peak aortic flow velocities for normal patients, and those with coronary artery disease, diverged significantly in response to exercise or dipyridamole stress (22, 26). At the present time, we have not developed a satisfactory, standardized stress test for the highly labile mice used in the present study, but we would anticipate that stress might uncover ventricular dysfunction in the reperfused mice in a manner similar to that found by others (22, 26). Nevertheless, the data clearly demonstrate a more severe defect in both systolic and diastolic function in the permanently occluded animals evident even in the basal state.

In recent years, it has been recognized that reperfusion may have an effect on myocardial healing and ventricular anatomy beyond its ability to limit infarct size. Several clinical trials of thrombolytic therapy showed that late reperfusion reduces mortality (2, 15, 25, 29, 43, 45). Other clinical studies indicated that late reperfusion may reduce LV dilatation and remodeling (4, 18, 19, 42). Recent work (43) also suggested that the absence of a patent infarct vessel is associated with greater ventricular scar formation and ventricular remodeling observed by detecting collagen precursors circulating in the serum.

Several animal studies offer more detail on the role of reperfusion in ventricular remodeling. In the dog, late reperfusion resulted in accelerated inflammatory response and increased rate of effective infarct repair (35). Although the dog model did not manifest cardiac rupture or chronic infarct expansion as seen in human models, it was postulated that the acceleration of the early healing response would limit pathological remodeling (35). Similar studies done in rat suggested that late reperfusion limited infarct expansion (3). The extensive study by Boyle and Weisman (3) demonstrated that late reperfusion resulted in accelerated clearance of dead myofibrils and improved myocardial healing. Infarct expansion measured by the ratio of endocardial segment lengthening to infarct wall thinning was likewise reduced with reperfusion at both 7 and 21 days after myocardial infarction.

The factors associated with reperfusion that influence ventricular healing have not been carefully defined. Reperfusion in the first day may result in accelerated inflammatory responses that could favorably influence healing of the infarcted myocardium (5). Morita and co-workers (28) suggested that accelerated influx of monocytes results in increased monocyte-macrophage function that might favorably influence healing. Frangogiannis et al. (8) demonstrated that the reperfused myocardium manifests increased influx of mast cell precursors with resultant 5- to 10-fold increases of mast cell number in the healing infarct. Mast cells secrete histamine, basic fibroblast growth factor, and tryptase, all of which are potent fibrogenic agents (14, 31, 37). Frangogiannis and co-workers (8) demonstrated a high correlation between the presence of tissue mast cell and mitotically active myofibroblasts in the ventricle scar. In the final analysis, effective tissue repair requires effective removal of necrotic tissue, transition and ultimate suppression of the inflammatory response, myofibroblast influx, effective collagen formation, and ultimately, marked reduction in myofibroblast number. Optimal scar formation and ventricular topography are dependent on all of the above processes, and each is controlled in a complex manner.

In summary, the ability to manipulate postinfarct coronary blood flow in the mouse affords both functional and anatomic end points that can be followed in evaluating healing response to myocardial infarction in the presence and absence of reperfusion. Using mice with specific genetic manipulations may allow definition of the cellular and molecular factors associated with reperfusion-dependent improvement of ventricular healing and function.