INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

ISCHEMIC CARDIOMYOPATHY is an increasingly prevalent cause of death and disability from cardiovascular disease (20). Unfortunately, major deficiencies exist in our understanding of how chronic coronary artery disease progresses to symptomatic left ventricular (LV) dysfunction due to the lack of experimental animal models where the progression of physiological, cellular, and molecular responses can be evaluated longitudinally. Although some patients develop heart failure due to muscle loss and replacement fibrosis in association with large or multiple myocardial infarcts (1, 6, 29), pathological studies support the view that many patients have global LV dysfunction, disproportionately severe in relationship to the amount of replacement fibrosis identified at postmortem exam (3, 31). Two of the factors that could contribute to this dissociation include a chronic depression in myocyte contractile function due to alterations in sarcoplasmic reticulum (SR) function and LV remodeling with diffuse interstitial connective tissue proliferation and myocyte cell dropout due to apoptosis (28). An additional explanation for the dissociation between the severity of LV dysfunction and fixed structural changes could be reversible contractile dysfunction arising from repetitive episodes of ischemia in regions supplied by severe coronary stenoses with chronically reduced coronary flow reserve (20). Although viable dysfunctional myocardium due to chronic stunning or chronic hibernation has been demonstrated in a subset of patients with reversibly depressed LV function, its importance as a cause of ischemic cardiomyopathy in the absence of infarction has not been clearly established.

In a previous study, we demonstrated that pigs with a chronic left anterior descending coronary artery (LAD) stenosis develop a critical impairment in coronary flow reserve that results in hibernating myocardium but preserved global LV systolic function (18). This was accompanied by cellular features typical of those found in advanced heart failure, including a progressive regional downregulation in SR gene expression (17) and substantial regional myocyte loss due to increased apoptosis. These regional findings suggested that reversible ischemia from chronic reductions in coronary flow reserve may represent an important early and reversible event in the progression of chronic LV dysfunction (22). Importantly, there was minimal fibrosis, no infarction, and no evidence of global LV dysfunction or heart failure in pigs with hibernating myocardium from a single LAD occlusion.

We hypothesized that global LV dysfunction similar to that encountered in patients with chronic ischemic cardiomyopathy could be produced if the extent of myocardium at risk of reversible ischemia was increased. To test this, we produced chronic stenoses on the proximal LAD as well as the proximal circumflex artery of pigs. Echocardiographic assessment of global and regional function was combined with physiological studies to assess systemic and pulmonary hemodynamics to evaluate global LV performance under resting conditions. Our results demonstrate that this instrumentation results in a moderate global impairment in systolic LV function with markedly elevated filling pressures similar to those encountered in compensated congestive heart failure. Moreover, there was viable dysfunctional myocardium with acceleration in the development of phenotypic features typical of hibernating myocardium.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All procedures and protocols conformed to institutional guidelines for the care and use of animals in research. The initial surgical preparation has been previously described (16, 18) but modified in that proximal stenoses were placed on two of the three coronary arteries (15). Briefly, juvenile pigs were premedicated with a Telazol (tiletamine 50 mg/ml and zolazepam 50 mg/ml)/ketamine (100 mg/ml) mixture (0.037 ml/kg im) and given prophylactic antibiotics (cephalothin 50 mg/kg iv and gentamicin 5 mg/kg im). A thoracotomy was performed under isoflurane anesthesia (1-3%). In 40 pigs, the proximal LAD and left circumflex coronary (LC) arteries were dissected free and instrumented with 1.5-mm (ID) Delran constrictors (18). The right coronary artery (RCA) that supplies the posterior descending artery was not manipulated. After the incision was closed and the pneumothorax was evacuated, antibiotics were repeated. Pain was controlled with an analgesic (butorphanol 0.025 mg/kg im) and an intercostal nerve block (2% Marcaine). Sudden death occurred in 25 animals an average of 42 ± 4 days after instrumentation. An additional four animals died after sedation for the terminal physiological study, and one febrile animal was terminated 1 wk after initial instrumentation. One animal had an LC stenosis that was <50%. This animal was excluded from further analysis because functionally significant stenoses were not present in both coronary distributions. Eight animals served as sham-instrumented controls, in which the LAD and LC arteries were dissected free but not instrumented with a stenosis.

Four of the nine dysfunctional animals and all of the sham-instrumented controls had a baseline echocardiogram 1 wk after instrumentation to assess regional and global LV function. After an overnight fast, pigs were sedated with a Telazol/xylazine (100 mg/ml) mixture (0.022 ml/kg im). Transthoracic echocardiography was performed with a 2.25-MHz phased-array transducer (Ultramark 9, ATL) from a right parasternal approach with standard M-mode measurements obtained from a midventricular short-axis view. This permitted simultaneous recording of the anteroseptal (LAD perfusion territory) and posterior walls (normally perfused RCA territory). End diastole was defined as the onset of the QRS complex and end systole was the point of minimum chamber diameter. Myocardial systolic excursion (WT) was defined as end-systolic thickness minus end-diastolic thickness. Percent fractional shortening was defined as [(LV end-diastolic dimension)  (LV end-systolic dimension)]/(LV end-diastolic dimension) · 100.

Experimental protocol. Approximately 2 mo after instrumentation, pigs underwent a terminal physiological study in the closed-chest anesthetized state. After an overnight fast, animals underwent a repeat echocardiogram with Telazol/xylazine sedation. Subsequently, animals were intubated, ventilated with oxygen and anesthetized with isoflurane (1-2%) supplemented with Telazol/xylazine (0.011 ml/kg im as needed). An 8-Fr sheath was placed into the left carotid artery through which a 5-Fr Millar micromanometer was placed into the LV apex for pressure measurement. The side port was used for arterial pressure. An 8-Fr sheath was placed in the right carotid artery and an 8-Fr pigtail catheter was inserted into the LV apex for microsphere flow measurement and ventriculography. A sampling catheter was inserted into the femoral artery for arterial reference sampling during microsphere flow injections. Right heart catheterization was performed through an 8-Fr introducer placed in the jugular vein using a 7-Fr balloon-tipped thermodilution catheter that was advanced into the pulmonary artery and connected to a Gould P23 dB transducer. Animals were heparinized (100 units/kg iv) and hemodynamics were allowed to equilibrate for ~30 min before beginning the protocol.

1

1

1

1

Postmortem, histology, and Western analysis of SR proteins. Hearts were weighed and sectioned into concentric rings. Thin rings were immersed in triphenyl tetrazolium chloride (TTC) to quantify infarction by digital planimetry. Samples were also obtained for quantification of connective tissue and regional SR protein levels. Histological samples were immersed in Z-fix (Anatech), sectioned, and stained with Masson's trichrome, and connective tissue was quantified by point counting (18). Protein was isolated from flash-frozen subendocardial samples, electrophoresed, and separated on SDS polyacrylamide gels, and transferred to nitrocellulose or polyvinylidene difluoride membranes as previously described (17). Linearity of density and protein loading for the candidate SR proteins was demonstrated over a range of 20-60 µg of total protein for the SR Ca²⁺-ATPase, between 25 and 100 µg of total protein for phospholamban, and between 40 and 140 µg total protein for calsequestrin, using commercially available antibodies previously shown to cross react with the swine (17).

Data analysis. Hemodynamics were digitized with an analog-to-digital converter at a sampling rate of 500 Hz and analyzed on line with the Notocord Heme software system. LV pressure was differentiated numerically to obtain the first derivative of LV pressure (dP/dt). We quantified regional wall motion from hand tracings of the lateral ventriculograms using the centerline method as previously described in detail (16). Data are presented as means ± SE. Measurements from the LAD and LC regions were compared with the normally perfused RCA region of the same heart using paired t-tests and the Bonferroni correction for multiple comparisons. Differences in hemodynamics and flow between rest and vasodilation were compared using paired t-tests. Experimental groups were compared with shams using unpaired t-tests. P  0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The nine animals instrumented with stenoses were studied 72 ± 3 days after instrumentation when they weighed 49.0 ± 5.0 kg (dysfunctional group). The eight sham-instrumented animals (sham group) weighed 47.1 ± 2.9 kg at 74 ± 1 days after instrumentation (no differences between groups). All animals were in good health at the time of terminal study with no overt clinical signs of congestive heart failure. There were no differences in hematocrit or arterial blood gases between the groups (average values: hematocrit 32 ± 1%, pH 7.42 ± 0.01, PCO₂ 45 ± 1 Torr, and PO₂ 437 ± 12 Torr).

A representative coronary angiogram from a dysfunctional animal is shown in Fig. 1. Significant proximal stenoses were present in both the LAD and LC arteries with an average diameter reduction of 95 ± 2 and 87 ± 3%, respectively. Complete occlusion with collateral-dependent myocardium was present in 4 LAD and 2 LC vessels. Left ventriculography demonstrated severe anteroapical hypokinesis with moderately reduced global function (average ejection fraction 43 ± 2%, P < 0.02 vs. shams, Table 1).

View larger version (124K): [in this window] [in a new window]   Fig. 1.   Coronary angiogram from a dysfunctional animal with proximal stenoses on the left anterior descending (LAD) and left circumflex coronary (LC) arteries. Radiographic contrast injection into the left main coronary artery demonstrated severe proximal stenoses in each animal. The right coronary artery (not shown) was not instrumented. Sham-instrumented animals had smooth luminally normal arteries.

Gross and microscopic pathology. The LV mass-to-body weight ratio was higher in dysfunctional animals but of borderline significance (2.89 ± 0.16 vs. 2.65 ± 0.08 g/kg, P = 0.08). There was TTC evidence of small, healed infarction in five of the dysfunctional animals. The extent of TTC infarction averaged 1.4 ± 0.7% of LV mass. On histological evaluation, point counting of connective tissue averaged 3.6 ± 0.2% in the normally perfused RCA region, 16.5 ± 9.3% in the stenotic LAD region (P < 0.05 vs. RCA), and 8.9 ± 1.7% in the stenotic LC region (P = 0.06 vs. RCA). There were no regional differences in connective tissue staining in sham animals (LAD 3.3 ± 0.4 vs. RCA 4.6 ± 0.9%; P, not significant). Because of the marked variability in connective tissue among individual dysfunctional animals, the difference between dysfunctional and sham groups did not reach statistical significance.

Systemic and pulmonary hemodynamics. Hemodynamic parameters are summarized in Tables 1 and 2 and Fig. 2. Under resting conditions, systemic hemodynamics in dysfunctional animals were similar to sham controls. In contrast, resting LV end-diastolic pressure (LVEDP) was increased in dysfunctional animals (26 ± 3 vs. 17 ± 1 mmHg, P < 0.01), and LV dP/dt was reduced compared with sham controls (1,587 ± 125 vs. 1,894 ± 119 mmHg/s, P = 0.10). Right heart catheterization (Table 1) demonstrated mild pulmonary systolic hypertension (46 ± 3 vs. 37 ± 2 mmHg, P < 0.05) and a moderate increase in pulmonary wedge pressure to 19 ± 2 mmHg compared with 11 ± 1 mmHg in sham controls (P < 0.01). Although LV filling pressures were elevated in dysfunctional animals, there was no alteration in resting cardiac output. Collectively, these hemodynamic findings demonstrate moderate impairment of LV systolic function at rest despite a chronic elevation in LV filling pressure in animals with two-vessel coronary artery stenoses.

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Baseline hemodynamics in dysfunctional and sham animals. Dysfunctional animals developed an increase in left ventricular (LV) end-diastolic pressure (LVEDP) and pulmonary wedge pressure (PCW). Although the first derivative of LV pressure (dPLV/dt) and systolic pressure (LVSYS) were lower in dysfunctional animals, these differences did not reach statistical significance.

Echocardiographic measurements of function. Transthoracic echocardiography performed at the time of terminal study confirmed impaired global and regional LV function in animals instrumented with chronic stenoses (Table 3, Figs. 3 and 4). Initial echocardiograms performed 7 ± 1 days after instrumentation (1 wk) demonstrated no differences in regional or global function between dysfunctional and sham-instrumented animals (Table 3, Fig. 3). At ~10 wk after instrumentation, WT had increased commensurately with end-diastolic wall thickness in sham-instrumented animals and in the normally perfused posterior region of the dysfunctional group (Table 3). However, anteroseptal WT in dysfunctional animals failed to increase as the animals grew and regional function was significantly reduced compared with sham controls (systolic excursion 2.1 ± 0.3 vs. 4.6 ± 0.4 mm, P < 0.01; wall thickening 26 ± 4 vs. 65 ± 7%, P < 0.05). Function of the normally perfused posterior wall was similar in each group.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Echocardiographic measurements of function in dysfunctional and sham animals. Individual data for animals with chronically dysfunctional myocardium are compared with sham controls. Echocardiograms performed 1 wk after initial instrumentation demonstrated similar levels of systolic excursion (A) and fractional shortening (B) in each group. However, ~10 wk after instrumentation, anteroseptal systolic excursion and fractional shortening in dysfunctional animals was reduced compared with shams. Systolic excursion in the normally perfused posterior wall was similar between the two groups of animals at each time point.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Echocardiographic measurements of global LV function. Dysfunctional animals exhibited a reduction in ejection fraction (LVEF) and an increase in the end-systolic volume index (LVESVI). There was also an increase in the end-diastolic volume index (LVEDVI) that was of borderline significance. An increase in LV mass-to-body weight ratio in dysfunctional animals was also of borderline significance.

Regional myocardial perfusion. Systemic hemodynamics corresponding to the microsphere flow measurements at rest and during adenosine vasodilation are shown in Table 2. There were no differences in heart rate, aortic pressure or double product between dysfunctional and sham animals either under resting conditions or during vasodilation. There were no regional differences in flow in sham-instrumented animals. Full-thickness perfusion in sham animals averaged 1.32 ± 0.09 ml · min¹ · g¹ and was higher in the subendocardium with an endo/epi flow ratio of 1.22 ± 0.03. During adenosine, there was more than a sixfold increase in flow in sham animals to 8.66 ± 1.20 ml · min¹ · g¹ (P < 0.01 vs. rest) with a slight reduction in the endo-to-epi ratio to 0.88 ± 0.04, (P < 0.01 vs. rest).

1

1

1

1

1

1

1

1

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Regional perfusion in dysfunctional animals. Compared with the normally perfused RCA region (C), resting flow was significantly reduced in the subendocardial (Endo) and midmyocardial (Mid) layers of the LAD region (A). During adenosine vasodilation the normally perfused RCA region demonstrated a 5.4-fold increase in flow over baseline values. In contrast, the LAD and LC perfusion (B) territories had a severe reduction in maximal flow with subendocardial samples unable to increase flow over baseline values. Thus there were significant reductions in flow during vasodilation limiting perfusion to 2 of the 3 major perfusion territories of the left ventricle. Epi, epicardial.

1

1

1

1

1

1

1

1

1

1

Regional reductions in SR proteins. Subendocardial samples from the LAD and normally perfused region were assayed for candidate proteins to determine whether there were regional alterations as we have previously found in pigs with hibernating myocardium from a single LAD stenosis (17). Protein levels for the SR Ca²⁺-ATPase (LAD 6.5 ± 1.3 vs. 9.3 ± 1.4 densitometric units in normal hearts, P < 0.05) and phospholamban (LAD 7.7 ± 2.1 vs. 12.1 ± 2.4 densitometric units in normal hearts, P = 0.05) were reduced by ~30-35% in the dysfunctional LAD region compared with normally perfused remote RCA regions. There were no regional differences in calsequestrin nor were there regional differences in sham-instrumented control animals.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

There are two important new findings from this investigation. First, pigs instrumented with two proximal coronary artery stenoses develop regional dysfunction accompanied by a moderate reduction in LV performance. This was characterized by reductions in LV ejection fraction in the setting of pronounced increases in LV filling pressures at rest. Although this did not produce reductions in resting cardiac output nor clinical signs characteristic of advanced right heart failure, it was similar to the hemodynamic profile of patients with moderate (NYHA Class II and III) compensated congestive heart failure in the setting of chronic coronary artery disease. Our results also clearly show that hibernating myocardium can be responsible for chronic global LV dysfunction in the absence of infarction when the extent of myocardium at risk of reversible ischemia involves at least two of the three major coronary artery distributions. Development of hibernating myocardium was accelerated compared with the 3-4 mo time frame required in pigs instrumented with a single LAD stenosis, where global LV function and filling pressures were normal (17, 18). Thus our results support the hypothesis that chronic ischemic heart disease can lead to heart failure with viable, chronically dysfunctional myocardium.

Ischemic cardiomyopathy and viable dysfunctional myocardium. Ischemic cardiomyopathy is the most common cause of congestive heart failure and accounts for nearly 70% of the patients enrolled in recent large clinical trials (20). Patients with cardiomyopathy due to coronary artery disease have a survival rate poorer than those with idiopathic dilated cardiomyopathy (2, 20, 29). Postmortem studies demonstrate extensive arteriosclerosis with total occlusion of at least one of the three major epicardial arteries, LV dilatation, and an increase in LV mass (1, 31). History of prior myocardial infarction may be decades remote from the development of heart failure, and the percentage of the LV involved with healed scar is generally less than that found in patients dying of pump dysfunction after an acute myocardial infarction. This supports the hypothesis that viable, chronically dysfunctional myocardium contributes importantly to the development and progression of heart failure. Nevertheless, the extent to which reversible ischemia can be responsible for contractile abnormalities and/or structural changes associated with ischemic cardiomyopathy remains unclear (5).

Relationship to previous models of heart failure from ischemic heart disease. Although there are several experimental models to study heart failure, most work in large animals has focused on pacing-induced heart failure or the decompensated state of hypertrophy arising from hypertension or chronic aortic outflow tract obstruction. Large animal models of heart failure due to ischemic heart disease have been limited. These have been primarily focused on heart failure in association with large myocardial infarcts or myocardial microembolization (30). Because the extent of acute infarction needed to produce heart failure is associated with an extremely poor short-term survival in large animals, the majority of studies have focused on the rat because it can survive myocardial infarcts approaching 60% of LV mass. The one exception to this has been the circumflex occlusion model in growing swine developed by Zhang et al. (37). This model leads to biventricular hypertrophy with variable degrees of heart failure and LV remodeling. Nevertheless, LVEDPs were only increased to 16 mmHg and were lower than what we found in the present study.

Comparison with the one-vessel model of hibernating myocardium in pigs. Results of the present study extend previous work from our laboratory as well as others, in which pigs instrumented with a stenosis on the proximal LAD developed viable, chronically dysfunctional myocardium (17, 18, 26, 27). We have previously shown there was a progression from chronically stunned to hibernating myocardium as stenosis severity increases (9, 16). Transition to hibernating myocardium occurred after 3-4 mo (average 107 days) and was accompanied by regional reductions in oxygen consumption (27) and flow (17, 18). Although the LAD supplies ~40% of the pig left ventricle (35), LV ejection fraction in animals with hibernating myocardium and one-vessel occlusion was the same as sham-instrumented controls (17). In the present study, both the LAD and LC arteries were instrumented, thereby increasing the total LV mass at risk. Based on previous measurements in pigs, we would estimate the risk area subjected to reversible ischemia to exceed 60% of the left ventricle (35).

1

1

1

1

Perfusion-contraction matching in viable chronically dysfunctional myocardium. Controversy exists regarding the mechanisms responsible for viable dysfunctional myocardium and whether resting flow is actually normal or depressed (8-11, 21, 36). Some investigators have suggested that regional dysfunction is disproportionately depressed relative to the modest reductions in flow in regions of hibernating myocardium (8). Others have suggested heterogeneity of mechanisms when myocardial flow and function are analyzed on a segmental basis, with chronic stunning and hibernating myocardium being present in the same heart (19, 33). Our results with direct microsphere measurements indicate that flow is chronically reduced in hibernating LAD myocardium compared with normal remote regions as well as sham controls. Data to assess whether flow and function are matched or mismatched in clinical studies are complicated by the wide variations in regional perfusion among patients (7, 11, 21) and the paucity of information regarding the relationship between flow and radial wall motion versus more sensitive indices of regional function such as wall thickening.

View larger version (12K): [in this window] [in a new window]   Fig. 6.   Flow-function relations in pigs with hibernating myocardium. Relative function in hibernating myocardium is shown in relationship to relative subendocardial (A) and full-thickness (B) perfusion. Data from the hibernating region are shown relative to values from the normally perfused posterior region. Shaded symbols are individual data; solid symbols depict means ± SE. As emphasized by the identity relationship (dashed line), the reduction in regional function was greater than the reduction in resting flow in the majority of animals. These results suggest that either additional stunning was been superimposed on hibernating myocardium [as initially proposed by Bolli (4)] or that there may be an imperfect matching between flow and function in chronic hibernating myocardium.

Methodological limitations. Although we have previously demonstrated that pigs with a single LAD stenosis have contractile reserve to -adrenergic stimulation, we did not assess this in pigs with two stenoses because of a high rate of sudden death during spontaneous excitement in this model. We have no reason to believe that this would be absent because there was minimal evidence of irreversible injury and a similar degree of increased connective tissue, but confirming this will require additional studies. Delineation of the LC region was done by subjectively evaluating the coronary anatomy in conjunction with the distribution of flow during vasodilation. Because flow was analyzed with colored microspheres and we needed to obtain samples for protein isolation and histology, we were unable to identify the perfusion territory of individual vessels with vital dyes. Although we were careful to restrict our analyses to core regions and avoided potential border areas between perfusion territories, it is possible that the intermediate reductions in vasodilated flow in the LC region could, to some extent, be a result of an admixture of normal and severely stenotic tissue. Because of this and an inability to assess regional wall thickening in the lateral region by M-mode echocardiography or ventriculography, we have confined our primary flow and function analyses to the LAD and normally perfused RCA regions. Furthermore, because we could not analyze wall thickening and flow in small myocardial regions, we cannot determine whether heterogeneity in mechanisms exists with chronically stunned and hibernating myocardium coexisting in the same perfusion territory. Additional studies will be required to evaluate flow-function relations on a three-dimensional basis in smaller myocardial segments.

Clinical implications. Our study shows that viable, chronically dysfunctional myocardium can lead to reductions in global function and elevations in LV filling pressure that are characteristic of compensated heart failure. Thus global LV dysfunction from repetitive episodes of ischemia can develop in the absence of myocardial infarction. Because large dysfunctional regions after myocardial infarction can stimulate global LV remodeling, it is possible that ischemic cardiomyopathy could develop from remodeling in response to viable dysynergic regions. Because apoptosis has been demonstrated in hibernating myocardium, myocyte loss may be accentuated as the extent of dysfunctional myocardium increases. Whether extending the time frame over which the observations are conducted or increasing the area at risk of reversible ischemia leads to remote zone remodeling and/or progresses to more advanced phases of heart failure will require further study.