INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

IN RESPONSE TO INJURY, myocytes and their surrounding extracellular matrix reorganize resulting in progressive ventricular dilatation, wall thinning, and cardiac dysfunction (37). In both ischemic and nonischemic dilated cardiomyopathy (DCM), cardiac remodeling mediates the transition from compensated to decompensated heart failure (22, 37). Therapies that reduce mortality for patients with heart failure invariably produce favorable effects on the remodeling process and thereby limit disease progression (27). Identifying the molecular regulators of the disruptive remodeling process will facilitate the development of novel therapies for the growing number of patients at risk of congestive heart failure.

The myocardial matrix provides both structural support and a dynamic microenvironment where molecular cues may interact to regulate tissue remodeling (21, 35). In the failing heart, the myocardial matrix is degraded by matrix metalloproteinases (MMPs) resulting in diminished structural support, altered ventricular geometry, and contractile dysfunction (32). The tissue inhibitors of metalloproteinases (TIMPs) are endogenous MMP inhibitors that regulate and maintain matrix homeostasis when present in the dynamic interstitial compartment. TIMPs directly inhibit the disruptive MMPs and have been implicated in the regulation of cell shape, function, and survival independent of their MMP-inhibitory effects (20). In light of these diverse biological roles, the TIMPs may play a central role in cardiac remodeling. Human studies have been limited by end-stage tissue (17, 28, 39) such that a comprehensive analysis of MMP-TIMP interactions and matrix remodeling during the progression of heart failure is not currently available.

The normal human heart expresses all of the four known TIMP species (TIMP-1, -2, -3, and -4). Although the TIMPs are altered during the progression of human heart failure, a consistent and clear pattern has yet to be defined (23). We hypothesized that a specific profile of TIMPs is associated with adverse remodeling of the matrix in both human and experimental heart failure. To investigate this hypothesis, we examined the profile of TIMPs and the associated changes in matrix content, organization, and turnover during the temporal progression of heart failure in the cardiomyopathic hamster. To validate these findings, results were compared with patients with end-stage heart failure.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Experimental animals. The Animal Care Committee of the Toronto General Research Institute approved all procedures performed on animals. In addition, all experiments were performed according to the Canadian Council on Animal Care's "Guide to the Care and Use of Experimental Animals" and the National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85-23, Revised 1985). The Syrian hamster with genetic cardiomyopathy (BIO 53.58 hamster, Bio Breeders; Fitchburg, MA) was used as our heart failure model. Age-matched normal FVG Syrian hamsters served as controls (Charles River; Quebec, Canada). A total of 30 hamsters were used to complete this study (DCM, n = 21; normal, n = 9).

Human samples. The Institutional Review Board approved our protocol for consent to use human tissue for these studies. LV samples (n = 2) were obtained from cardiac donors whose hearts were suitable for transplantation but for which no recipient was available. Patients (n = 7) in end-stage congestive heart failure (LV ejection fraction <20%, New York Heart Association functional class IV), with ischemic cardiomyopathy (n = 3) or DCM (n = 4), were included in the present study. Samples of the anterior LV were collected during cardiac surgery at the time of cardiac transplantation or insertion of a mechanical assist device. Transmural samples were collected avoiding obvious areas of scar or infarction. The samples were immediately snap-frozen in liquid nitrogen and stored at 80°C until use.

Assessment of MMP activity. Snap-frozen samples of the LV anterior wall were pulverized in liquid nitrogen and extracted in ice-cold extraction buffer (20 µl/mg tissue) containing (in mM) 50 Tris · HCl, 150 NaCl, 10 CaCl₂ and 0.2 NaN₃, and 0.01% Triton X-100 (pH 7.6) and stirred for 30 min. After centrifugation, the supernatant was collected and the pellet was reextracted twice more. The supernatant from each extraction step was pooled and stored at 80°C. Gelatinase MMP activity was determined with a commercial bioassay kit (Chemicon International; Temecula, CA). MMP activity in the samples was assessed without preactivation and compared with p-aminophenylmercuric acetate-activated purified human MMP-9 (Chemicon International), which was used as a standard. MMP activity is expressed as the equivalent activity of activated MMP-9. To distinguish MMP activity from nonspecific gelatinase activity, GM6001 (nonspecific MMP inhibitor) was added to a random sample in each group. MMP activity was ~90% of the total gelatinase activity.

Quantification of MMPs and TIMPs. The relative abundance of MMPs and TIMPs was examined in LV myocardial extracts using standard immunoblotting procedures. In brief, extracts containing equal amounts of protein were fractionated by 10% SDS-PAGE and electrotransferred onto a polyvinylidene difluoride membrane (Bio-Rad). The membranes were blocked with 5% (vol/vol) nonfat dry milk in 20 mM Tris · HCl (pH 7.5), 500 mM NaCl, and 0.1% Tween 20 for 1 h at room temperature. Incubation was performed overnight at 4°C with polyclonal rabbit anti-human antibodies for TIMP-1 (1:5,000), TIMP-2 (1:3,000), TIMP-3 (1:500), TIMP-4 (1:3,000), MMP-2 (1:5,000), and MMP-9 (1:2,000, Chemicon International). Subsequently, the membranes were washed with Tris-NaCl-Tween 20 and incubated with a goat anti-rabbit IgG horseradish peroxidase conjugate (1:2,000, Santa Cruz Biotech) for 1 h at room temperature. Membranes were developed with enhanced chemiluminescence reagent (Amersham Pharmacia) and quantified by image analysis. For TIMP-2 and TIMP-3, only the band that migrated with the positive controls was quantified. The results were presented as the percentage of change compared with nonfailing controls, whose means were arbitrarily set as 100%.

Collagen content, cross linking, and synthesis. Collagen content in the LV was determined by using the method of Strauss and colleagues (36). The minced myocardial sample was divided into two portions. In the first portion, samples were digested overnight by a cyanogen bromide (CNBr) treatment (50 mg/ml in 70% formic acid, 1.0 ml per 50 mg tissue, under N₂), which solubilized all proteins except for cross-linked collagen by cleaving methionine bonds. The supernatant, which contained fragments of collagen and other proteins, was dried and hydrolyzed in 6 N HCl at 110°C for 24 h. The remaining portion of myocardium was dried and hydrolyzed in 6 N HCl at 110°C for 24 h omitting the prior CNBr digestion.

Confocal microscopy of matrix structure. To evaluate the organization of the fibrillar collagen matrix during the transition to decompensated heart failure, LV sections from cardiomyopathic and normal hamsters at 25 wk of age (n = 3 per group) were fixed, processed with graded alcohols, embedded in paraffin, cut into 10- to 15-µm-thick sections, and stained with 0.1% picrosirius red. To enhance the contrast between myocytes and collagen, sections were treated with 0.2% phosphomolybdic acid (pH 1.8-2.2) before being stained. Sections were imaged with the use of a laser scanning confocal system (model MRC 1024, Bio-Rad). Fields were randomly selected from the midmyocardium, excluding large epicardial arteries and veins, cutting or compression artifact, and any obvious focal areas of scar. Only regions that appeared as viable, functional myocardium devoid of cellular necrosis and associated scar were selected for confocal analysis (Fig. 1A). Four random fields were obtained from each of two transverse LV segments. Three hearts for each group were used for confocal analysis to yield a total of 24 images per group. The stained LV sections were then digitized at a final magnification of ×600 and analyzed using an image analysis system (Scion Image, NIH Software). Image analysis was performed to quantify these parameters in blinded fashion. For each image, the total collagen area fraction and the perimysial collagen fiber length and diameter were measured. The average collagen fiber diameter was calculated from 7 to 10 measurements along each fiber.

View larger version (65K): [in this window] [in a new window]   Fig. 1.   A: representative light micrograph of the left ventricle in a dilated cardiomyopathic (DCM) hamster (BIO 53.58) indicating patchy areas of focal scar and cellular necrosis (solid arrows) adjacent to areas of viable nonscarred myocardium (open arrows). Collagen matrix appears in blue and is highly localized to patchy areas of cell necrosis (Masson's trichrome staining, ×25; inset, ×100). B: gelatinase matrix metalloproteinase (MMP) activity, assessed in the presence of native tissue inhibitor of MMP (TIMP) species, was increased in cardiomyopathic hearts (*P = 0.03) and peaked at 25 wk during the transition to overt failure. Normal (n = 5; n = 1 at 15 wk, n = 2 at 25 wk, n = 2 at 35 wk) vs. DCM (n = 6 at 15 wk, n = 6 at 25 wk, and n = 6 at 35 wk) values are shown.

Statistical analysis. Results are presented as means ± SE. Comparison between groups was performed by one-way ANOVA. If the F ratio was significant, pairwise tests of individual group means were compared using the Student-Newman-Keuls test. All statistical procedures were performed with SAS software (SAS Institute; Cary, NC). The critical -level for these analyses was set at P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

MMP gelatinase activity. Gelatinase (MMP-2 and MMP-9) activity was assessed by a commercial plate assay, which allows determination of MMP activity in the presence of native TIMP species. MMP activity was elevated above normal levels in all ages of DCM and peaked significantly at 25 wk during the transition from compensated dysfunction to overt failure (Fig. 1B). MMP activity was three times greater than the activity in normal myocardium at 25 wk of age.

Protein levels of MMPs and TIMPs. Protein levels of myocardial MMP-2, MMP-9, and TIMP-1 to TIMP-4 were determined by Western blotting with appropriate commercially available antibodies. The mean levels of myocardial MMP-2 and MMP-9 were not significantly different from normal age-matched controls. Although MMP activity was elevated in DCM hearts, the protein levels of these MMPs were not different between normal and failing hearts at any age.

View larger version (46K): [in this window] [in a new window]   Fig. 2.   A: representative immunoblots indicating differential TIMP-1 and TIMP-3 protein expression during progressive left ventricular (LV) remodeling and the transition to cardiac failure in cardiomyopathic hamsters compared with normal nonfailing controls. B and C: by immunoblotting, TIMP-3 was decreased during the transition to decompensation at 25 wk (74 ± 5% of normal) and during end-stage disease at 35 wk (69 ± 10% of normal; *P = 0.001), whereas TIMP-1 increased progressively (P = 0.001). Normal (n = 5; n = 1 at 15 wk, n = 2 at 25 wk, n = 2 at 35 wk) and DCM (n = 6 at 15 wk, n = 6 at 25 wk, and n = 6 at 35 wk) values are shown. D and E: reductions in TIMP-3 were associated with increased MMP activity suggesting that TIMP-3 regulates cardiac matrix turnover in the failing heart. In contrast, the endogenous MMP inhibitor TIMP-1 increased in association with elevated MMP activity perhaps indicating an inadequate compensatory response to elevated matrix degradation.

Myocardial collagen content, synthesis, and cross-linking. Collagen content was significantly increased in the DCM group compared with age-matched controls (Fig. 3A). In DCM hearts, collagen content increased in parallel to the progression of heart failure.

View larger version (18K): [in this window] [in a new window]   Fig. 3.   A: myocardial hydroxyproline (HPRO), a measure of collagen content, increased with the progression to overt failure in cardiomyopathic hamsters (*P < 0.0001). Normal (n = 5; n = 1 at 15 wk, n = 2 at 25 wk, n = 2 at 35 wk) and DCM (n = 6 at 15 wk, n = 6 at 25 wk, and n = 6 at 35 wk) values are shown. B: cross-linked collagen (ratio of insoluble to soluble collagen) was reduced at 15 and 25 wk (P = 0.003) in cardiomyopathic hamsters in parallel to LV dilatation. Normal (n = 5; n = 1 at 15 wk, n = 2 at 25 wk, n = 2 at 35 wk) and DCM (n = 6 at 15 wk, n = 6 at 25 wk, and n = 6 at 35 wk) values are shown. C: collagen synthesis was determined by the rate of [14C]proline incorporation and was elevated in cardiomyopathy before the onset of overt failure (*P < 0.05). Normal (n = 3 at 15 wk, n = 3 at 25 wk, n = 3 at 35 wk) and DCM (n = 6 at 15 wk, n = 6 at 25 wk, and n = 6 at 35 wk) values are shown. CPM, counts per minute.

Confocal microscopy and quantitative image analysis. Confocal microscopy was performed to assess the architecture of the myocardial fibrillar collagen network in hamsters at the time of TIMP-3 deficiency (25 wk) before end-stage disease and was compared with age-matched normal controls. Images were selected exclusively from areas of viable myocardium avoiding focal areas of scar and large blood vessels. In the DCM group, all hamsters showed marked disruption of the structure of the fibrillar collagen matrix (Fig. 4). By quantitative image analysis (Table 1), the local collagen matrix content was significantly reduced in the DCM group indicating a loss of matrix components due to excessive degradation. The length and diameter of the perimysial collagen fibrils surrounding the myofibrils was reduced in the DCM group indicating fragmentation of the fibrillar collagen network.

View larger version (62K): [in this window] [in a new window]   Fig. 4.   Representative confocal micrographs of the left ventricle showing cardiomyocytes (green due to autofluorescence) and the surrounding fibrillar collagen network (stained red by picrosirius red) in normal (A) and DCM hamsters at 25 wk (B). In areas of myocardium devoid of scar, fragmentation and disruption of perimysial collagen fibers was evident in the cardiomyopathic hamsters before the onset of overt failure. Bar equals 50 µm.

Comparative analysis of human samples. Results of the comparative analysis of end-stage human failing myocardium are presented in Fig. 5. MMP activity was increased in human patients with congestive heart failure (n = 7) compared with nonfailing controls (n = 2). TIMP-3 was decreased in patients with congestive heart failure compared with nonfailing controls, whereas the content of the remaining TIMP species, including TIMP-1, were not significantly different.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Consistent with the progression of hamster cardiomyopathy, human myocardium from patients with end-stage congestive heart failure (CHF; n = 7) had elevated gelatinase activity (*P = 0.02) (A) associated with isolated reductions in TIMP-3 [55 ± 5% of nonfailing (NF) hearts; P = 0.003] (B).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The extracellular matrix may play a central role in the complex process of myocardial remodeling. Reorganization of the myocardial matrix alters the structural support for cardiomyocytes as well as the environmental cues necessary for their function and survival (21, 35). In the present study, we provide a comprehensive analysis of matrix remodeling during the temporal progression of remodeling in the cardiomyopathic hamster model of heart failure and compared these results with those obtained in patients with end-stage heart failure. Our data suggests that reductions in myocardial TIMP-3 parallel significant changes in matrix turnover, content, and structure and may contribute to decompensation in the failing heart.

Experimental model. The Syrian cardiomyopathic hamster strain used in this study has a genomic deletion of the -sarcoglycan gene (26) that disrupts the vascular smooth muscle cell sarcoglycan-sarcospan complex (29), thus inducing microvascular spasm (10). An accumulation of focal myocardial infarcts results in a reproducible, irreversible, and gradually progressive congestive heart failure (14). In our experience, LV function and structure is grossly preserved at 15 wk of age, whereas ventricular dysfunction with significant LV dilatation is evident by 25 wk, but without overt failure (compensated stage) (25). However, by 35 wk, the LV is severely dilated with compromised function and overt failure (decompensated stage). Remodeling of the myocardial matrix has been studied extensively in this hamster strain by using similar time points for comparison (9, 11, 24). The BIO 53.58 strain lacks a hypertrophic response similar to patients with DCM. In fact, -sarcoglycan gene mutations occur in human patients and can result in DCM (38). Thus the -sarcoglycan-deficient hamster is similar to human cardiomyopathy (13), both acquired (ischemic) and idiopathic (inherited) dilated forms, and is thus an appropriate model to study the process of LV remodeling.

Matrix turnover. In the cardiomyopathic hamster, MMP gelatinase activity was elevated during the temporal progression of heart failure. These data support the concept of excessive MMP activity in failing hearts (32, 33). However, the majority of previous studies employed techniques that were unable to account for the influence of TIMPs on overall MMP activity. In the present study, gelatinase activity was assessed using a bioassay that, unlike traditional gelatin zymography, measures MMP activity in the presence of and relative to their natural inhibitors and, hence, is more likely to reflect in vivo MMP and TIMP activity. The presence of TIMPs in the samples was confirmed by immunoblotting (Fig. 2A). Our data suggests that when the influence of TIMP inhibition is considered, MMP activity is still in excess of normal and is particularly elevated during the transition to overt failure.

TIMP regulation of matrix remodeling. MMP activity can be elevated from increased MMP content and/or increased MMP activation (35). Interestingly, MMP-2 and MMP-9 protein abundance was not significantly different from normal controls, suggesting that increased MMP activation in hamsters with cardiomyopathy elevates MMP activity. These data suggest a loss of endogenous inhibitory control, perhaps from a reduction in TIMPs. The four TIMP species all bind and inactivate the various MMPs, including MMP-2 and MMP-9, but with different affinities. TIMP-3 is a more potent inhibitor of MMP-9 than the other TIMPs (35). In our study, whereas TIMP-2 and TIMP-4 levels were unchanged, TIMP-3 was reduced in parallel to significant alterations in the matrix turnover, content, and structure. The reorganization of matrix components was temporally associated with LV dilatation and the loss of function. Whereas no clear pattern of overall TIMP expression has been defined in heart failure (23), reductions of TIMP-3 have been consistently reported in end-stage human heart failure (17, 39) and in ventricular hypertrophy after experimental aortic banding (31). In our study, TIMP-3 was significantly reduced during the progression of hamster cardiomyopathy and in patients with end-stage congestive heart failure. In contrast to TIMP-1, TIMP-3 abundance was negatively correlated with gelatinase activity, which supports its role in regulating matrix degradation.

Matrix content and organization. In the failing heart, both the content and organization of fibrillar collagen are believed to influence myocardial structure and function (23, 37). MMP-TIMP stoichiometery may regulate the balance of matrix elements and maintain chamber geometry and contractile function. Although an accumulation of myocardial collagen (fibrosis) is frequently associated with LV dilatation (37), so too is a degradation of the fibrillar collagen network (33). This apparent contradiction attests to the complexity of the remodeling process. In cardiomyopathic hamsters, total myocardial collagen content increased with the progression of LV dilatation to failure. In addition, myocardial collagen synthesis was elevated in disease, supporting an increased myocardial collagen content. In contrast, by a laser confocal microscopic analysis confined to viable myocardial segments avoiding areas of focal scar, we demonstrate that regional collagen content is indeed reduced, with significant collagen fibril thinning and fragmentation indicating degradation of the fibrillar collagen network (Table 1). The appearance of the degraded fibrillar collagen network was similar to reports of human DCM (41). Light microscopic analysis indicated that most of the new collagen was directed to scar formation in areas of cell necrosis as replacement fibrosis (Fig. 1A). In contrast, matrix degradation appeared to occur in the viable myocardial regions remote from areas of focal necrosis and scar formation (Fig. 4). These regional differences in matrix turnover may explain how collagen breakdown and increased collagen content can occur simultaneously.

TIMP-3 and cardiac remodeling: a novel role. Unlike its counterparts, TIMP-3 is unique in that it binds to and is localized within the extracellular matrix (2, 16), where it acts to neutralize MMPs, inhibit TNF--converting enzyme (TACE) (1, 4), and influence cell survival (5, 12, 43). In failing myocardium, reductions of TIMP-3 may contribute to cardiac remodeling by elevating MMP activity, increasing TNF- activation, and promoting cell death (Fig. 6). These unique properties may afford TIMP-3 a key regulatory role in the process of myocardial remodeling.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Loss of myocardial TIMP-3 may induce MMP-mediated matrix disruption, tumor necrosis factor- (TNF-)-converting enzyme (TACE) activation, and cardiomyocyte apoptosis resulting in LV remodeling, cardiac dilatation, and contractile dysfunction.