INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

ALTHOUGH ISCHEMIC cardiomyopathy is the leading etiological cause of congestive heart failure (15), the mechanisms through which chronic coronary artery disease progresses to decompensated left ventricular dysfunction are complex and poorly understood (33). Severe coronary artery disease and a history of symptomatic ischemia are present in the majority of patients, but the role of reduced coronary flow reserve in the progression of contractile dysfunction has not been resolved. Heart failure in coronary disease can arise from structural alterations consisting of myocyte necrosis and replacement fibrosis (3, 38), but contractile dysfunction due to abnormalities of intracellular calcium and alterations in myocyte function have also been implicated to play an important role (31). Explanted hearts from patients with advanced ischemic cardiomyopathy have reductions in the expression of some, but not all, of the sarcoplasmic reticulum (SR) calcium handling proteins (29), and in vitro studies demonstrate attenuated force-interval relationships in explanted cardiac muscle (34). This suggests that the contractile abnormalities may be due to SR dysfunction. Nevertheless, whether SR dysfunction arises secondary to global systemic hemodynamic changes or neurohormonal activation associated with advanced heart failure, or from repetitive myocardial ischemia due to extensive coronary artery disease has not been studied because it is difficult to dissociate these factors clinically or in experimental animal models of heart failure.

Chronically dysfunctional myocardium that is viable or "hibernating" is frequently found in patients with severe coronary artery disease and collateral-dependent myocardium (2, 45). Although this can occur in the absence of heart failure, it can also contribute to the progression of left ventricular dysfunction to clinical heart failure (15). The mechanisms responsible for chronic contractile dysfunction in hibernating myocardium are unclear, as are their relation to global alterations in advanced ischemic cardiomyopathy. One possibility is that the chronic alterations in SR proteins found in advanced heart failure are actually an early abnormality secondary to an intrinsic myocardial adaptation from intermittent ischemia. This may precede the development of structural fibrosis, neurohormonal activation, and global systemic abnormalities found in end-stage heart failure.

To test this hypothesis, we used a recently developed porcine model of regional, viable, dysfunctional myocardium that has all of the features of hibernating myocardium in humans, including regional reductions in resting flow and function and increased ¹⁸F-2-deoxyglucose uptake (12). The direct effects of a chronic reduction in regional coronary flow reserve could be studied because these animals do not develop systemic changes seen in heart failure. Our first aim was to study whether there was a regional downregulation in selected SR proteins that was similar to heart failure and to determine how this is related to coronary physiology in the intact animal. The second aim was to examine whether a chronic stenosis could be associated with the induction of mechanisms that protected myocytes from ischemia. To do this, we initially chose to examine heat shock protein (HSP)-70 that has been shown to be induced following brief total coronary occlusions and associated with protection against stunning and infarction (27, 43). The results indicate that hibernating myocardium exhibits a molecular phenotype that, while regional, is similar to global changes in ischemic cardiomyopathy. Thus SR dysfunction due to reversible ischemia may occur as an early defect in the progression of left ventricular dysfunction that precedes the development of the structural abnormalities seen in ischemic cardiomyopathy.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All experimental procedures and protocols conformed to Institutional Guidelines for the Care and Use of Animals in Research. Studies were conducted in 24 hibernating animals that had a proximal left anterior descending artery (LAD) occlusion and 13 sham-instrumented controls. Terminal hemodynamic and coronary flow studies were performed in 22 hibernating animals (4 used for protein analysis and 18 used for RNA assessment) and 11 sham controls. The remaining four animals (2 hibernating and 2 sham) were killed without physiological studies to confirm that acute instrumentation had no affect on mRNA levels.

Chronic surgical preparation. The surgical preparation has been previously described (12). Briefly, juvenile pigs were fasted overnight, premedicated with a Telazol (50 mg/ml tiletamine and 50 mg/ml zolazepam)-ketamine (100 mg/ml) mixture (0.037 ml/kg im), and given prophylactic antibiotics (50 mg/kg iv cephalothin and 5 mg/kg im gentamicin). After endotracheal intubation a surgical plane of anesthesia was maintained with a halothane (0.5-2%) and oxygen (balance) mixture. A thoracotomy was performed in the fourth left intercostal space, and the proximal LAD was dissected free and instrumented with a Delran occluder [average ID 1.7 ± 0.1 (SE) mm]. In hibernating animals, the artery was secured within the occluder by a silk ligature or umbilical tape. In sham animals, the groove of the occluder was left open to circumvent the formation of a significant stenosis. After surgery was complete, the chest incision was closed in layers, the intercostal nerves were infiltrated with 2% lidocaine for analgesia, and the pneumothorax was evacuated. A postoperative dose of antibiotics was repeated, and an analgesic (0.025 mg/kg im butorphanol) was given and repeated as required to alleviate pain. Animals were fully recovered within 24 h. They were group housed and fed a diet of Pro-Lean (Agway, Akron, NY) ad libitum.

Hemodynamic studies in hibernating and sham-instrumented animals. Animals were brought back to the laboratory for physiological studies and tissue harvesting after ~3-4 mo. They were fasted overnight, and anesthesia was induced with a Telazol-xylazine (100 mg/ml) mixture (0.022 ml/kg im). After tracheal intubation, a surgical plane of anesthesia was maintained using a halothane (0.5-2%) and oxygen (balance) mixture supplemented with intramuscular doses of Telazol-xylazine (0.011 ml/kg) as required. The left carotid artery was instrumented with a 7-Fr introducer for coronary angiography. The right femoral artery was dissected free and used to pass an 8-Fr pigtail catheter retrogradely into the left ventricle using fluoroscopic guidance. A short catheter was placed into the left femoral artery to withdraw blood for microspheres using the reference sampling technique. Catheters were placed in the femoral veins to administer fluids (normal saline) and pharmacological agents. The animals were heparinized (10,000 U iv), and hemodynamics were allowed to equilibrate for at least 30 min before the protocol was begun.

1

1

1

1

1

1

Microsphere flow and angiographic measurements. Regional perfusion was assessed with colored microspheres (15- µm diameter, Dye-Trak, Triton, San Diego, CA) using previously published techniques (12). Briefly, microspheres suspended in saline with thimerosal (0.01%) and Tween 80 (0.01%) were sonicated and vortex agitated before withdrawal. Approximately 3-5 million microspheres (yellow, red, white, or blue) were administered as a left ventricular bolus after an arterial reference sample was started from the femoral artery. At the end of the study, myocardial samples were weighed, digested in 4 M KOH with 2% Tween 80, and processed using previously published techniques (20). The dyes were eluted from the microspheres using a measured volume of dimethylformamide and aliquots placed in a multiple wavelength spectrophotometer (model U-2000, Hitachi, Tokyo, Japan). Absorbance was measured at the principal absorbance peak of each pure color dye and corrected for overlapping spectra using a matrix inversion technique (18, 20). Regional myocardial perfusion was calculated using the absorbance and flow rate of the arterial reference sample and myocardial absorbance per unit sample weight (20).

RNA isolation and Northern blot analysis. Total myocardial RNA was isolated from subendocardial samples (inner one-third) of the central LAD and normal remote regions by guanidinium thiocyanate-phenol chloroform extraction (9). Because large variations in HSP-70 expression were found, we also examined its distribution in mid- and subepicardial samples. We simultaneously electrophoresed 20-µg aliquots of RNA on 0.8% agarose formaldehyde gels. These were vacuum transferred onto nylon membranes and baked in a vacuum oven at 80°C. We hybridized Northern blots with a 2.3-kb probe for human HSP-70 (47) [American Type Culture Collection (ATCC), Rockville, MD]. Control for RNA loading was confirmed and quantified using a 1.2-kb probe to glyceraldehyde-3-phosphate dehydrogenase (GAPDH; ATCC) (44). The remaining probes for Ca²⁺-ATPase, calsequestrin, calmodulin, phospholamban, -myosin heavy chain, and the ryanodine receptor were synthesized by PCR. First-strand cDNA was synthesized by reverse transcribing myocardial mRNA with M-MLV Reverse Transcriptase at 37°C for 1 h. This was subsequently amplified by PCR using primers for the specific nucleotide sequence coding for each protein. After an initial denaturation for 5 min at 94°C, samples were cycled 25 times at 94°C for 30 s, at 45°C for 1 min, and at 72°C for 1 min. This was followed by a final 7-min extension at 72°C. The following pairs of oligonucleotide primers were used to amplify probes from porcine myocardial cDNA with the exception of SR Ca²⁺-ATPase, for which rat myocardial cDNA was used: 1) rat SR Ca²⁺-ATPase (24), sense primer 5'-TTGGCTTGGTTCGAAGAAGG-3' (+223 to +242 nt) and antisense primer 5'-CCAAGAGCCACCATGAACTG-3' (+864 to +845 nt) to produce a 642-bp fragment; 2) calsequestrin (40), sense primer 5'-AAGCTTGCCAAGAAGCTGGG-3' (+301 to +320 nt) and antisense primer 5'-GCAAAGGCCACAATGTGGAT-3' (+821 to +802 nt) to yield a 521-bp product; 3) calmodulin (13), sense primer 5'-AGGAGTTGGGGACAGTGATG-3' (+192 to +211 nt) and antisense primer 5'-ATGTCAGCCTCCCTGATCAT-3' (+492 to +473 nt) to produce a 301-bp segment; 4) phospholamban (46), sense primer 5'-TCAGCTTTCTCTTGACGGCT-3' (52 to 33 nt) and antisense primer 5'-ACCCCTAGTTCATCCTCAGA-3' (+474 to +455 nt) to produce a 526-bp fragment; 5) -myosin heavy chain (21), sense primer 5'-TGAAGGAGAACATCGCCATC-3' (+788 to +807 nt) and antisense primer 5'-GTAGGTGAGCTCCTTGATGC-3' (+1344 to +1325 nt) to yield a 557-bp product; and 6) the cardiac ryanodine receptor (RYR2) (22), sense primer 5'-GGAAATCCATTCTGAATTCT-3' (+21 to +40 nt) and antisense primer 5'-GCAGTCACAAACGGCTCGGT-3' (+488 to +507 nt) to produce a 487-bp segment. The amplified cDNA products were inserted into a vector [PCR-TRAP kit from GenHunter (Boston, MA) or TA Cloning kit from Invitrogen (San Diego, CA)] and transformed into Escherichia coli. HSP-70 and GAPDH probes were labeled by random priming. All other probes were labeled by PCR in the presence of [-³²P]dCTP (30 cycles at 94°C, 30 s; 50°C, 1 min; 72°C, 1 min).

Protein isolation and Western blot analysis. To confirm that mRNA alterations reflected steady-state changes in protein levels, subendocardial samples were obtained from four animals with regional reductions in resting perfusion that underwent identical instrumentation. Protein was isolated from flash-frozen samples using an extraction buffer containing 20 mM Tris (pH 7.4), 50 mM NaCl, 1 mM EDTA (pH 8), 1 µg/ml pepstatin, 0.5 µg/ml leupeptin, 5 mM -mercaptoethanol, 0.2 mM sodium vanadate, 10% SDS, and 0.2 mM phenylmethylsulfonyl fluoride. Protein concentration for each sample was measured in triplicate using the Lowry method (25). Aliquots of protein (50 µg/lane for SR Ca²⁺-ATPase, 75 µg/lane calsequestrin and phospholamban, and 80 µg/lane for HSP-70) were electrophoresed and separated on 10 or 14% SDS-polyacrylamide gels. They were transferred to a nitrocellulose membrane (HSP-70, Protran, Schleicher and Schuell, Keene, NH) or a polyvinylidene difluoride membrane (other proteins, Immobilon-P, Millipore, Bedford, MA) and blocked in 3% nonfat dry milk in PBS. After the membrane was rinsed, it was incubated overnight with the primary antibody [SR Ca²⁺-ATPase anti-canine mouse monoclonal antibody (32), 1:1,500 dilution (MA3-919, Affinity Bioreagents, Golden, CO)]; phospholamban anti-canine mouse monoclonal antibody, 1:1,000 dilution (05-205, Upstate Biotechnology, Lake Placid, NY); calsequestrin anti-canine rabbit polyclonal antibody, 1:800 dilution (06-382, Upstate Biotechnology); and HSP-70 anti-human mouse monoclonal antibody, which recognizes both the constitutive (73 kDa) and inducible (72 kDa) forms of HSP-70, 1:5,000 dilution (SPA-820, Stress Gen Biotechnology, Victoria, BC). After being rinsed in PBS, membranes were incubated with peroxidase-labeled goat antibodies to IgG. Bands were visualized with 3,3',5,5'-tetramethylbenzidine (TMB) membrane peroxidase substrate (Kirkegaard and Perry, Gaithersburg, MD) and quantified on a Laser Densitometer (Bio-Rad, Melville, NY). Linearity of density and protein loading for the SR Ca²⁺-ATPase was demonstrated over a range of 20-60 mg of total protein, between 25 and 100 mg of total protein for phospholamban, and between 40 and 140 mg total protein for calsequestrin and HSP-70.

Histology. Myocardial samples adjacent to those taken for RNA and protein analyses were immersed in Z-fix (Anatech, Battle Creek, MI). Thin sections were stained with Masson's trichrome stain to delineate connective tissue and fibroblast staining from normal myocytes. Percent connective tissue staining was quantified with standard point counting techniques using a 121-point grid at a magnification of ×200 as we have previously described (12). Two full-thickness regions were analyzed in each sample and selected to represent the area of greatest and least connective tissue staining in the sample. At least 30 fields were quantified and averaged for each sample. Connective tissue staining was expressed as a percentage of the total. Perivascular connective tissue staining and endocardial staining were included in the reported results.

Data analysis. All data are presented as means ± SE. Differences between sham and hibernating animals were determined using group t-tests for similar parameters and interventions. Measurements in the LAD region were compared with corresponding measurements in the normal region using paired t-tests. Differences between rest and pharmacological interventions were assessed using an ANOVA followed by paired t-tests and the Bonferroni correction for multiple comparisons. P < 0.05 level was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Myocardial perfusion and hemodynamics in hibernating versus sham controls. All animals were in good health at the time of study. Data for the entire hibernating group (n = 22) are summarized below. Body weight increased from 9.2 ± 0.4 kg at the time of instrumentation to 83.0 ± 3.4 kg at the time of study (average growth interval 102 ± 3 days). Left ventricular mass averaged 217 ± 9 g, and mass-to-body weight ratio averaged 2.63 ± 0.07 g/kg. Arterial blood gases immediately before resting measurements of myocardial perfusion averaged pH 7.40 ± 0.01, PCO₂ 38 ± 1 mmHg, and PO₂ 477 ± 28 mmHg. Hematocrit averaged 33 ± 1%. Under resting conditions, heart rate averaged 83 ± 3 beats/min, and systolic blood pressure averaged 121 ± 4 mmHg. Resting subendocardial perfusion averaged 0.73 ± 0.05 ml · min¹ · g¹ in LAD regions and 1.01 ± 0.06 ml · min¹ · g¹ in normal remote regions (P < 0.001). Corresponding full-thickness values were 0.80 ± 0.05 and 0.92 ± 0.06 ml · min¹ · g¹ (P < 0.001). During adenosine vasodilation, subendocardial flow in the LAD region did not increase significantly over rest values [1.0 ± 0.17 ml · min¹ · g¹, P = not significant (ns) vs. rest], but in normal remote regions it increased to 3.84 ± 0.26 ml · min¹ · g¹ (P < 0.001 vs. rest). Corresponding full thickness values were 1.42 ± 0.16 ml · min¹ · g¹ in the LAD region (P < 0.01 vs. rest) and 4.33 ± 0.24 ml · min¹ · g¹ in the remote normal zone (P < 0.001 vs. rest).

View larger version (140K): [in this window] [in a new window]   Fig. 1.   Coronary angiogram and tracings of a resting left ventriculogram from a hibernating pig. Angiogram and tracings of left ventriculogram are in the left lateral projection. Coronary angiography in animals in the hibernating group (A) usually showed total occlusion of proximal left anterior descending artery (LAD; arrow), which was associated with brisk opacification of distal LAD by angiographically visible collaterals. Ventriculography (B) showed severe hypokinesis or akinesis of anteroapical wall of left ventricle. In contrast, sham-instrumented animals had mild stenosis of the proximal LAD and preserved anteroapical wall motion. Solid line, end-diastolic (ED) tracing; dotted line, end-systolic (ES) tracing; LC, left circumflex artery.

1

1

Alterations in SR Ca²⁺-ATPase and phospholamban mRNA in hibernating myocardium. Figure 2 is a representative Northern blot for paired samples from four hearts comparing mRNA levels in dysfunctional hibernating regions with their corresponding expression in normal remote regions from the same heart. Figure 3 summarizes the results of representative SR mRNAs from paired samples from all of the hibernating animals that were studied. Data were normalized to the GAPDH expression in each sample, which did not vary between regions (subendocardial GAPDH 2.96 ± 0.23 densitometric units in hibernating LAD regions versus 3.03 ± 0.24 densitometric units in normal remote regions, P = ns). Although there was variability among animals, we found highly significant reductions in the level of mRNA for phospholamban and the SR Ca²⁺-ATPase in hibernating LAD regions compared with normal regions, whereas mRNA for the SR calcium binding protein calsequestrin was unchanged. The relative changes in mRNA (LAD/normal) were also reduced in comparison to sham-instrumented controls as summarized in Fig. 4 (phospholamban 0.76 ± 0.08 in hibernating vs. 1.07 ± 0.06 in shams, P < 0.02; SR Ca²⁺-ATPase 0.83 ± 0.06 in hibernating vs. 1.02 ± 0.06 in shams, P < 0.05; calsequestrin 1.08 ± 0.06 in hibernating vs. 0.96 ± 0.05 in shams, P = ns). Analysis of the ryanodine receptor mRNA was performed only in hibernating animals. The expression of RYR2 relative to GAPDH was reduced in hibernating compared with normal remote regions (LAD RYR2/GAPDH 0.12 ± 0.01 vs. 0.15 ± 0.01 in normal regions, n = 17, P < 0.05; LAD/normal 0.85 ± 0.07). These data indicate that there were significant reductions in the mRNAs for the SR calcium uptake and release proteins, whereas the constancy of mRNA for calsequestrin suggests that this did not reflect a nonspecific alteration in the expression of all SR proteins.

View larger version (66K): [in this window] [in a new window]   Fig. 2.   Representative Northern blot of sarcoplasmic reticulum (SR) calcium handling protein mRNA from subendocardium in normal (Nl) and hibernating (H) regions. Each pair of lanes depicts subendocardial mRNA from individual animal. Bottom panel shows that there was relatively uniform loading and no change in expression of control probe for glyceraldehyde-3-phosphate dehydrogenase (GAPDH). In contrast, there was reduction in expression of phospholamban (both upper and lower bands) and SR Ca2+-ATPase in hibernating LAD regions. Although variable, the degree to which phospholamban was reduced generally paralleled that for the SR Ca2+-ATPase in individual animals.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Paired comparison of subendocardial SR gene expression in hibernating LAD vs. normal remote regions (n = 20). A: phospholamban. B: SR Ca2+-ATPase. C: calsequestrin. Although there was variation among individual animals, there were significant reductions in the mRNA levels for phospholamban and the SR Ca2+-ATPase in hibernating vs. normally perfused regions. In contrast, expression of calsequestrin remained unchanged. Solid symbols represent group averages (means ± SE).

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Relative induction of subendocardial SR Ca2+ regulatory protein mRNA in hibernating (n = 20) vs. sham-instrumented animals (n = 10). Relative mRNA induction is expressed as ratio of LAD to normal zone mRNA normalized to respective expression of GAPDH in each region. Sham-instrumented animals demonstrated no regional alteration in expression of phospholamban, SR Ca2+-ATPase, or calsequestrin. Hibernating animals demonstrated significant regional reductions in SR Ca2+-ATPase and phospholamban mRNA in comparison to shams, whereas expression of calsequestrin mRNA was unaltered. Pattern of gene expression demonstrated on regional basis is similar to global changes found in advanced congestive heart failure.

Transmural induction of HSP-70 mRNA in hibernating animals. We found a graded induction of mRNA for HSP-70 across the wall of the LAD region of hibernating animals. Figure 5 depicts the results of a representative gel from an individual animal with hibernating myocardium. There was a low uniform HSP-70 mRNA expression in all layers of the normally perfused region that was similar to that seen in sham-instrumented animals. In contrast, there was an induction of HSP-70 mRNA in hibernating LAD regions that was greatest in the subendocardium and least in the subepicardium. Figure 6 compares the induction of HSP-70 mRNA relative to GAPDH in the LAD region of hibernating animals versus sham-instrumented controls. HSP-70 levels were uniform in all transmural layers of the normally perfused regions as well as in the LAD region in shams. In contrast, there was a greater than twofold induction of HSP-70 mRNA in the subendocardium of hibernating hearts, which decreased toward the subepicardium (P < 0.05 vs. each corresponding layer in the normal region). Figure 7 summarizes the results of subendocardial HSP-70 mRNA in comparison to mRNA levels of the cytosolic proteins calmodulin and -myosin heavy chain. The levels of HSP-70 mRNA were increased in hibernating compared with sham animals. There was no regional induction of calmodulin and -myosin heavy chain mRNA in the LAD region of hibernating animals relative to normal remote regions. Whereas sham-instrumented animals showed no regional difference in mRNA for HSP-70 or calmodulin, there was a slight but significant increase in the expression of -myosin heavy chain that was secondary to increased levels in the LAD zones of sham-instrumented animals.

View larger version (79K): [in this window] [in a new window]   Fig. 5.   Expression of heat shock protein (HSP)-70 mRNA in representative Northern blot from hibernating animal. Total RNA from subendocardium (Endo), mid-myocardium, and subepicardium (Epi) of normal and hibernating regions demonstrated uniform intensity of the GAPDH signals in all layers confirming equal RNA loading. There was weak and uniform hybridization for HSP-70 in normally perfused regions that did not vary across the wall of the heart. In contrast, HSP-70 was induced in the hibernating LAD regions. There was a distinct transmural pattern with induction of HSP-70 mRNA most pronounced in subendocardial layers and least marked in subepicardial layers.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Summary of induction of HSP-70 mRNA in hibernating (A, n = 20) vs. sham-instrumented (B, n = 11) animals. Hibernating animals had a more than twofold induction of HSP-70 mRNA in subendocardium. This progressively decreased in outer layers of the myocardial wall but remained significantly higher than corresponding normal remote region. In contrast, HSP-70 expression in LAD and normal regions from sham-instrumented animals was uniform across myocardial wall.

View larger version (20K): [in this window] [in a new window]   Fig. 7.   Comparison of subendocardial HSP-70 induction to the cytosolic proteins -myosin heavy chain and calmodulin in hibernating (n = 20) vs. sham-instrumented (n = 10) animals. There was twofold relative induction (LAD/normal) of HSP-70 in hibernating animals, whereas mRNA levels of calmodulin and -myosin heavy chain remained unchanged. There was no alteration in HSP-70 or calmodulin mRNA in sham animals, but a small increase in -myosin heavy chain expression was statistically significant.

Alterations of candidate proteins in hibernating myocardium. Figure 8 and Table 3 summarize the Western blot analysis for SR Ca²⁺-ATPase, calsequestrin, phospholamban, and HSP-70 protein from four hibernating animals. Subgroup analysis showed no differences in resting or vasodilated flow compared with the group used for mRNA measurements. In comparison to normally perfused remote myocardium, hibernating regions had significantly lower protein levels of SR Ca²⁺-ATPase and phospholamban but no significant change in calsequestrin. Immunoblots for HSP-70 in hibernating regions revealed two bands consistent with the constitutive (73 kDa) and inducible (72 kDa) forms. Although densitometric analysis revealed significant increases in both forms, induction was most evident in the 72-kDa band with a twofold increase. Thus the quantitation of candidate proteins confirmed directionally similar changes as seen for the mRNA.

View larger version (63K): [in this window] [in a new window]   Fig. 8.   Western blot analysis of SR Ca2+-ATPase, calsequestrin, phospholamban, and HSP-70 in hibernating (H) vs. normal remote regions (Nl). All 4 animals demonstrated reduction in SR Ca2+-ATPase protein (105-110 kDa) and phospholamban (monomeric form 6 kDa), whereas calsequestrin remained unchanged. HSP immunoblots demonstrated both 72- and 73-kDa bands that were increased in hibernating LAD regions in comparison to normally perfused remote regions.

Relation of SR gene expression to regional adenosine flow reserve. Figure 9 highlights the relation between relative adenosine flow (LAD/normal) and mRNA changes for selected SR proteins. A novel observation was that the magnitude of mRNA changes in dysfunctional LAD regions correlated with the physiological significance of the stenosis or collateral vasodilator reserve. Sham-instrumented animals showed no relation between mRNA and relative flow during adenosine. In contrast, the expression of phospholamban and the SR Ca²⁺-ATPase mRNA were significantly correlated with the severity of flow reserve reduction. At modest levels of flow impairment, there appeared to be little systematic change in mRNA despite the fact that myocardial function was chronically depressed. When flow reserve was more severely reduced, there was a reduction in SR Ca²⁺-ATPase and phospholamban mRNA. Calsequestrin expression was unchanged over a wide range of stenosis severity. These results suggest that the molecular adaptations seen in hibernating myocardium are likely related to repetitive ischemia that will increase in frequency and severity as adenosine flow reserve is reduced.

View larger version (16K): [in this window] [in a new window]   Fig. 9.   Relation between SR protein mRNA levels and relative subendocardial flow during adenosine. A: phospholamban. B: SR Ca2+-ATPase. C: calsequestrin. Variability in mRNA expression for phospholamban and SR Ca2+-ATPase was to a large extent related to variations in physiological significance of flow limitation to hibernating regions. Animals in which adenosine vasodilation was most severely limited had the largest reductions in the expression of SR uptake proteins. There was a poorer correlation between mRNA levels and resting subendocardial flow (data not shown) [relative phospholamban expression = 0.39 ln(relative resting flow) + 0.93, r = 0.45, P < 0.01; relative SR Ca2+-ATPase expression = 0.26 ln(relative resting flow) + 0.96, r = 0.39, P < 0.01; relative calsequestrin mRNA = 0.051 ln(relative resting flow) + 1.06, r = 0.09, P = ns].

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

There are two important new findings from our study. First, pigs with hibernating myocardium develop regional alterations in the expression of the SR calcium uptake proteins phospholamban and Ca²⁺-ATPase, which may be one of the mechanisms responsible for producing resting regional dysfunction. Western analysis confirmed that the regional alterations in mRNA were associated with regional changes in protein content. Whereas the pattern of expression we found was similar to that reported in cardiomyopathy and decompensated left ventricular hypertrophy, the changes were regional and occurred without global alterations in hemodynamics compared with sham control animals, indicating that they arise from chronic intermittent myocardial ischemia. Second, we found a regional induction of stress protein HSP-70 with increases in both the 72- and 73-kDa proteins by Western blot analysis. To our knowledge, this is the first demonstration that chronic reversible subendocardial ischemia can alter myocyte phenotype on a regional basis.

Downregulation in expression of selected SR proteins in hibernating myocardium. The results of our study indicate that the SR proteins involved in the uptake and release of calcium, the SR Ca²⁺-ATPase, phospholamban, and the ryanodine receptor are coordinately downregulated in our porcine model of hibernating myocardium. The changes were restricted to the dysfunctional LAD region and did not appear to be related to a generalized reduction in the SR proteins or a loss of SR within myocytes because mRNA and protein levels for the SR calcium binding protein calsequestrin remained unchanged. They also appear unrelated to myocyte loss because mRNA levels for cytosolic proteins such as GAPDH, calmodulin, and -myosin heavy chain were not significantly reduced in hibernating myocardium. Histological samples showed no myocardial necrosis, and point counting showed only a small increase in connective tissue staining in the hibernating regions, similar to that reported in patients with collateral-dependent myocardium distal to chronic LAD occlusions and clinically documented hibernating myocardium (5). These changes were much less extensive than those reported by Elsässer et al. (10) where up to 58% of the wall was fibrotic in reversibly dyssynergic myocardium, indicating that our findings may be representative of an earlier stage of disease. Thus the selective reduction in the expression of SR calcium uptake and release proteins supports the hypothesis that the chronic contractile abnormalities found in hibernating myocardium are secondary to SR dysfunction and arise from chronic intermittent ischemia. This precedes the advanced structural changes that could affect contractile function in an independent fashion.

Transmural induction of HSP-70 distal to chronic stenosis. We also found that the stress protein HSP-70 was variably increased in hibernating myocardium. The magnitude of HSP-70 mRNA induction was most pronounced in the subendocardial layers. Previous studies have demonstrated that HSP-70 can be induced after repetitive brief total coronary occlusions in pigs and that this is associated with accelerated recovery of myocardial function when a series of brief occlusions is performed 24 h later (preconditioning against myocardial stunning) (43). Other studies have demonstrated a small, statistically insignificant reduction in infarct size in pigs following prolonged occlusions when a repetitive ischemic stimulus is performed 24 h earlier (second window of preconditioning)(35). Despite these experimental findings, the relevance of this endogenous protective mechanism in chronic coronary artery disease is uncertain since ischemia produced by brief repetitive total coronary occlusions is somewhat artificial and, aside from unstable coronary syndromes, unlikely to mimic clinical circumstances in which the myocardium will be subjected to demand-induced ischemia.

Methodological limitations. We assessed regional wall motion with left ventriculography to circumvent potential nonspecific effects of instrumentation on inducing mRNA (6). Because the relation between regional perfusion and radial wall motion has not been established, we could not ascertain the extent to which chronic recurrent myocardial stunning (which would be associated with a dissociation between subendocardial flow and the degree of dysfunction) could have contributed to the chronic dysfunction. Nevertheless, stunning alone cannot explain our observations since it is associated with normal rather than reduced values of resting perfusion (4, 8, 36, 41). In pigs studied at 1 and 2 mo after instrumentation we have found viable dysfunctional myocardium without a reduction in resting flow (11). These data along with the present findings are consonant with the notion that there is a temporal transition from chronic stunning with normal flow to hibernation with depressed resting flow during the progression of a chronic stenosis, as we previously reported in dogs (8).