INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

THE DEVELOPMENT AND progression of congestive heart failure (CHF) is accompanied by changes in left ventricular (LV) geometry. Past clinical studies in patients with developing CHF have identified that LV dilation is an important prognostic indicator with respect to progressive LV failure and mortality (10, 22, 23). However, the fundamental contributory mechanisms that are responsible for LV remodeling in the setting of CHF remain unclear. The LV myocardium is composed of myocytes and muscle fibers that are connected by a fibrillar collagen matrix (8, 29, 35). The myocardial collagen matrix has been proposed to contribute to maintaining myocyte alignment, coordinating myocardial contraction, and maintaining LV geometry (8, 29, 35). Past clinical and experimental studies have reported that structural changes occur in the LV myocardial collagen matrix in the setting of CHF (16, 22, 31, 36). Therefore, alterations within the LV collagen matrix may be a contributory factor for the initiation and progression of LV dilation in the setting of CHF. An important family of enzymes responsible for the degradation and therefore remodeling of the collagen matrix is the matrix metalloproteinases (MMPs) (5, 37, 39). Previous studies have indicated that the complete digestion of fibrillar collagen occurs through the action of an enzymatic cascade involving interstitial collagenase (MMP-1), gelatinase A (MMP-2), and stromelysin (MMP-3) (18, 39). In addition, several past studies have provided evidence to suggest that increased myocardial MMP activity exists with severe forms of CHF (1, 16, 33). For example, Gunja-Smith et al. (16), using gelatin as an MMP substrate, reported increased activity with the development of CHF. Thus the substrate-specific MMP activity and relative abundance of MMP species that exist within the normal and CHF myocardium remain unclear. Accordingly, the overall goal of this study was to examine potential changes in MMP substrate degradative capacity and MMP species abundance with the development of CHF. This laboratory and others (1, 11, 22, 30, 31, 38) have demonstrated that chronic rapid pacing causes functional and neurohormonal characteristics that are similar to the clinical spectrum of CHF. Specifically, Damiano et al. (11), using radionuclide angiograms, reported that chronic rapid pacing caused progressive LV dilation. Wilson et al. (38) reported that chronic rapid pacing resulted in significant LV dilation and LV pump dysfunction with no overall change in LV mass. Taken together, these studies suggest that with chronic rapid pacing, significant LV remodeling occurs. Therefore, this model of rapid pacing was used in the present study to test the hypothesis that increased MMP substrate-specific activity and abundance occur with LV remodeling in the setting of CHF.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Model of CHF. Twelve Yorkshire pigs (20 kg) were randomly assigned to undergo 3 wk of rapid atrial pacing at 240 beats/min (n = 6) or to be sham controls (n = 6). The pigs were chronically instrumented to measure LV function in the conscious state (30, 32). The pigs were anesthetized with isoflurane (3%, 1.5 l/min) and nitrous oxide (0.5 l/min). A shielded stimulating electrode was sutured onto the left atrium, connected to a modified programmable pacemaker (model 8329, Medtronic, Minneapolis, MN), and buried in a subcutaneous pocket. The pericardium was closed loosely, the thoracotomy was closed, and the pleural space was evacuated of air. After a recovery period of 7-10 days, the pacemaker was activated in the animals selected to undergo chronic rapid pacing. At the completion of this protocol, the animals were sedated with diazepam (20 mg po; Valium, Hoffman-LaRoche, Nutley, NJ) and placed in a custom-designed sling, and the pacemaker was then deactivated (pacing groups only). After a 30-min stabilization period, indexes of LV function and geometry were determined by echocardiography. The pigs were then deeply anesthetized with isoflurane (3%), and the heart was quickly extirpated and placed in cold (4°C) oxygenated Krebs solution. The LV apex and midventricular region were cut into 1-cm³ cubes and subsequently processed for MMP assays.

LV myocardial MMP extraction. After stringent washing in ice-cold saline, the LV myocardial samples were homogenized (3- to 30-s bursts) in 5 ml of an ice-cold extraction buffer (1:3, wt/vol) containing cacodylic acid (10 mM), NaCl (0.15 M), ZnCl (20 mM), NaN₃ (1.5 mM), and 0.01% Triton X-100 (pH 5.0). The homogenate was then centrifuged (4°C, 10 min, 800 g), and the supernatant was decanted and saved on ice. The zymography and immunoblotting samples were then concentrated using a Centriplus concentrator at 4°C. Briefly, the homogenate was centrifuged (6,000 revolutions/min for 4.5 h), and the protein concentration of the myocardial extracts was then determined using a standardized colorimetric assay (Bio-Rad protein assay, Bio-Rad, Richmond, CA). The extracted samples were then aliquoted and stored at 20°C before use.

MMP zymography. Extracts were thawed on ice, diluted to a final protein concentration of 400 µg/ml, and then incubated in activation buffer containing 0.005% Brij-35 and 1 mM CaCl₂ for 5 min at 37°C. The extracts were incubated in the presence and absence of 2.5 µM trypsin (13, 18, 34). After 5 min of proteolytic activation by trypsin, the reaction was stopped by placing the mixtures in an ice bath. This concentration of trypsin and the incubation time were selected based on preliminary studies that demonstrated that this procedure provided maximal activation of LV myocardial extracts. The myocardial extracts were normalized by protein content (4 µg total protein) and then directly loaded onto electrophoretic gels (SDS-PAGE) containing 1 mg/ml of either gelatin, type III denatured collagen, or -casein (Sigma, St. Louis, MO) under nonreducing conditions (11, 13, 21, 33, 34). The gels were run at 15 mA/gel through the stacking phase (4%) and at 20 mA/gel for the separating phase (10%), maintaining a running buffer temperature of 4°C. After SDS-PAGE, the gels were washed twice in 2.5% Triton X-100 for 30 min each, rinsed in water, and incubated for 12 h (72 h for -casein) in a substrate buffer at 37°C (50 mM Tris · HCl, 5 mM CaCl₂, and 0.02% NaN₃, pH 7.5). In an additional set of experiments, gelatin zymogram lanes were also incubated for 12 h at 37°C in the same substrate buffer containing concentrations of an MMP inhibitor, gelardin, ranging from 0 to 50 nM (2, 14). These concentrations of gelardin were selected on the basis of past reports (2, 14) as well as preliminary studies performed by this laboratory. After incubation, the gels were stained using 0.1% Amido black, destained in water, analyzed, and dried for permanent record. As a positive control for SDS-PAGE zymography, conditioned media (2 µg total protein) from HT-1080 cells incubated for 24 h in the absence and presence of 100 ng/ml phorbol 12-myristate 13-acetate (PMA) were included in all zymograms (25, 28).

MMP immunoblotting. Before immunoblotting, the LV myocardial extracts were diluted to the appropriate loading concentration using sample buffer (0.1 M Tris · HCl, 0.2 M dithiothreitol, pH 6.8, containing 4% SDS and 0.01% bromphenol blue). LV myocardial extracts (4.0 µg total protein) were loaded onto an 8% SDS-polyacrylamide gel and separated at 40 mA in 0.02 M Tris base, 0.2 M glycine, pH 6.8, containing 0.1% SDS. The separated proteins were transferred at 100 V to a nitrocellulose membrane (Trans-blot transfer medium, 0.45 µm, Bio-Rad Laboratories, Hercules, CA) in 0.025 M Tris base, 0.2 M glycine, pH 8.2, containing 20% methanol (vol/vol) (6, 7). Membranes were blocked with 0.2 M Tris base, 1.4 M NaCl, pH 7.6, containing 5% powdered goat milk, 0.1% Tween-20, and 0.02% NaN₃. After being washed with 0.2 M Tris base, 1.4 M NaCl, pH 7.6, containing 0.1% Tween-20, membranes were incubated overnight at 4°C in specific monoclonal antibodies corresponding to MMP-1, -2, or -3 (1.0 µg/ml, Oncogene Research Products, Cambridge, MA). The primary antisera were diluted in 0.2 M Tris base, 1.4 M NaCl, pH 7.6, containing 1% powdered goat milk, 0.1% Tween-20, 0.08% BSA, 13% Dulbecco's modified Eagle's medium-Ham's F-12 cell culture medium (GIBCO, Grand Island, NY), and 0.02% NaN₃. After stringent washing, the membranes were incubated for 1 h in horseradish peroxidase-conjugated goat anti-mouse antibody (1:5,000 dilution, Bio-Rad Laboratories). The membranes were washed again, and the horseradish peroxidase-conjugated secondary antibody was activated with peracid and luminol (enhanced chemiluminescence Western blotting detection reagents, Amersham Life Science, Arlington Heights, IL). The luminescent signal was detected by exposure to X-ray film (Eastman Kodak, Rochester, NY) for exactly 5 min. Positive controls for MMP-2 and -3 were included in all immunoblots and were obtained from human epithelial and fibroblast cell lines (AG771 and AG770, respectively, Chemicon International, Tenecada, CA). Cell culture medium from PMA-stimulated HT-1080 fibrosarcoma cell line (25, 28) was also used as a positive control for immunoblotting. Prestained molecular weight markers (Bio-Rad Laboratories) were used to ensure adequate protein separation and transfer.

MMP data analysis. The control and pacing CHF LV myocardial extracts were loaded in a random fashion onto SDS-PAGE gels that were used for zymography and immunoblotting. The zymograms and the immunoblots were digitized (CCD 100, Dage-MTI, Michigan City, IN) using a constant light intensity. For the zymograms, the size-fractionated banding pattern, which indicated MMP activity, was determined by quantitated image analysis (Gel Pro, Media Cybernetics, Silver Spring, MD). A fixed area of interest (0.5 × 0.5 mm) was then placed over each of the lysis areas, and two-dimensional integrated optical density (IOD) was computed as follows: OD(x,y) = 1/{log [intensity(x,y)  black reference]/incident light  black reference}, where OD is optical density. With respect to the MMP inhibitor studies, the values of proteolytic activity obtained from the gelatin zymograms incubated in each of the concentrations of gelardin were expressed as a percentage of basal activity (0 nM). For the immunoblots, a single linear array (5-pixel width) was placed over the center of each lane, and the IOD was computed for each molecular weight species. The IOD was normalized to control samples and assumed to be 100%. The LV myocardial extracts used for zymography and immunoblotting were coded before densitometric analysis, and after this analysis, the code was broken and the data were grouped accordingly.

Statistical analysis. Indexes of LV function were compared using the Student's t-test. MMP zymographic activity was compared using analysis of variance (ANOVA). The ANOVA was used to compare two main effects, pacing CHF and trypsin activation. If the ANOVA revealed significant differences, pairwise tests of individual group means were compared using Bonferroni probabilities. With respect to the MMP inhibitor studies, linear regression analysis was performed to examine the relationship between gelardin concentration and percent zymographic activity. From this linear regression analysis, an effective concentration inhibiting 50% of proteolytic activity, or the IC₅₀, was computed. With the use of linear regression analysis, the relationship between indexes of LV geometry and MMP zymographic activity were examined. For the immunoblotting studies, it was not assumed that the data followed the normal distribution. Accordingly, the relative abundance of each MMP species was analyzed using the Wilcoxon signed-rank test in which control values were normalized to 100%. All statistical procedures were performed using the BMDP statistical software package (BMDP Statistical Software). Results are presented as means ± SE. Values of P < 0.05 were considered to be statistically significant.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

In the present study, all of the pigs that underwent chronic rapid pacing successfully completed the study protocol and exhibited symptoms of CHF that included dyspnea, tachypnea, and ascites.

LV function with pacing-induced CHF. Indexes of LV function are summarized in Table 1. With the development of pacing-induced CHF, ambient resting heart rate was increased. In the pacing CHF group, LV end-diastolic dimension was increased by 67%, and fractional shortening decreased by 68%. LV wall thickness was decreased by 40% with pacing CHF. Thus pacing-induced CHF was accompanied by LV pump dysfunction. Moreover, pacing-induced CHF was associated with LV dilation and wall thinning indicative of LV myocardial remodeling.

Zymographic activity with pacing-induced CHF. A representative zymogram using gelatin as a proteolytic substrate for LV myocardial extracts taken from control and pacing CHF groups is shown in Fig. 1. The conditioned media taken from HT-1080 cells, which were incubated for 24 h in the absence and presence of PMA, yielded zymographic activity consistent with past reports (13, 25, 28). Zymographic activity in the LV myocardial extracts was examined in the basal, nonactivated state as well as after activation with the serine protease trypsin. In both the nonactivated and activated states, zymographic activity was increased in the pacing CHF LV myocardial extracts. Activation with trypsin increased zymographic activity in both the control and pacing CHF LV myocardial extracts by at least 59%. The zymographic activity in the nonactivated state and after activation was quantitated using densitometric methods. An illustration of densitometric profiles in the control and pacing CHF LV myocardial extracts for the proteolytic substrate gelatin is shown in Fig. 2. Lytic bands could be identified between 84 and 30 kDa in both the control and pacing CHF groups, which are consistent with the active forms of several MMP species (1, 5, 12, 15, 37, 39). In the activated LV myocardial extracts, proteolytic activity was observed at lower molecular weights, which may reflect additional breakdown products or distinct species of MMPs (4, 17, 24-28). Total zymographic activity for the substrate gelatin is summarized in Table 2. In the pacing CHF group, LV myocardial zymographic activity was increased by over 100% in the nonactivated state and after activation when compared with controls.

View larger version (65K): [in this window] [in a new window]   Fig. 1.   Zymographic activity in left ventricular (LV) myocardial extracts was examined in control (Con) and pacing congestive heart failure (CHF) groups using gelatin as a proteolytic substrate. Proteolytic activity was examined in LV myocardial extracts with and without activation using the serine protease trypsin. Conditioned media from an HT-1080 cell line incubated in absence () and presence (+) of phorbol 12-myristate 13-acetate (PMA) were included in all zymograms. In nonactivated LV myocardial extracts, proteolytic activity was observed at 60- to 90-kDa region. In activated LV myocardial extracts, a number of proteolytic bands were observed at lower molecular weight regions. In both nonactivated and activated states, zymographic activity appeared to increase with pacing CHF. Total zymographic activity and molecular weight fractionation for this gelatin substrate were quantitated using densitometric methods (Fig. 2) and are summarized in Tables 1 and 2, respectively.

View larger version (62K): [in this window] [in a new window]   Fig. 2.   LV zymographic activity for specific substrates was quantitated by densitometry. After background subtraction, a 5-pixel-wide scan line was placed over each lane, and an intensity profile was generated with respect to molecular weight. Integrated optical density for specific proteolytic bands was quantitated from densitometric profile. Left: representative gelatin zymograms from nonactivated and activated control and pacing CHF LV myocardial extracts. Right: specific proteolytic bands were displayed with respect to optical density. In both nonactivated and activated states, increased zymographic activity was observed in pacing CHF LV myocardial extracts. After activation of LV myocardial extracts with trypsin, total zymographic activity increased in both control and pacing CHF groups. Quantitated results obtained from these densitometric profiles are shown in Tables 1 and 2.

View larger version (129K): [in this window] [in a new window]   Fig. 3.   Caseinolytic activity was determined in LV myocardial extracts taken from control and pacing CHF groups to more carefully examine recruitable matrix metalloproteinase (MMP)-3 activity (13, 15). A distinct proteolytic banding pattern consistent with MMP-3 activity was observed (4, 5, 13, 17, 37, 39). In nonactivated state, proteolytic activity was observed in LV myocardial extracts in 50-kDa region (arrow). This proteolytic activity likely reflects 57-kDa stromelysin, or MMP-3, in which the active form can be localized to 48 kDa. In activated LV myocardial extracts, an additional lower molecular weight proteolytic band was observed (arrows). Appearance of proteolytic activity at a lower molecular weight likely reflects activation of MMP-3. With pacing CHF, total caseinolytic activity was increased from control values (Table 1).

View larger version (38K): [in this window] [in a new window]   Fig. 4.   To confirm that observed zymographic activity was due primarily to MMPs, a previously characterized MMP inhibitor, gelardin (2), was employed. Inset: representative gelatin zymogram lanes from control (n = 3) and pacing CHF (n = 3) groups were exposed to each concentration of gelardin ranging from 0 to 50 nM. A concentration-dependent reduction in zymographic activity was observed in both control and pacing CHF groups, indicating that zymographic assays in present study primarily reflect MMP activity. This procedure was repeated in triplicate, and with the use of the inhibition curve generated from these experiments, an effective concentration inhibiting 50% of proteolytic activity or an IC50 of 1.6 ± 0.6 nM was computed.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   To examine the relationship between LV geometry and gelatin zymographic activity, linear regression analysis was performed. LV end-diastolic dimension and wall thickness were regressed to total gelatin zymographic activity in nonactivated state. A positive linear relationship was observed with respect to gelatin zymographic activity and LV end-diastolic dimension, whereas a negative linear relationship was observed for LV wall thickness. Similar results were obtained for zymographic substrates collagen III and casein, and these are summarized in RESULTS.

MMP abundance with pacing-induced CHF. Immunoblotting was performed on LV myocardial extracts taken from control and pacing CHF groups for interstitial collagenase (MMP-1), gelatinase A (MMP-2), and stromelysin (MMP-3), and representative immunoblots are shown in Fig. 6. For MMP-1, a positive immunoreactive band was observed at the 52/57-kDa region in LV myocardial extracts, which corresponds to the proenzyme form of MMP-1 (24, 37, 39). Conditioned medium from PMA-stimulated HT-1080 human fibrosarcoma cells, which was used as a positive control for MMP-1, revealed an immunoreactive band corresponding to MMP-1 that is consistent with past reports (5, 24, 37, 39). With pacing CHF, the abundance of MMP-1 was increased. MMP-2 was localized to the 72/66-kDa region, and MMP-3 was localized to the 54-kDa region, corresponding to molecular weights consistent with species of MMPs reported previously (5, 12, 17, 18, 37, 39). Positive controls for MMP-2 and MMP-3 that were obtained from human epithelial and fibroblast cell lines (AG771 and AG770, respectively) revealed immunoreactive bands corresponding to MMP-2 and MMP-3. The relative abundance of each MMP species was quantitated using desitometric analysis. With pacing CHF, the relative abundance of MMP-1 was increased by 319 ± 94%, the relative abundance of MMP-2 was increased by 194 ± 31%, and the relative abundance of MMP-3 was increased by 493 ± 159%. With pacing CHF, the abundance of all these MMP species was significantly increased when compared with controls (P < 0.05).

View larger version (66K): [in this window] [in a new window]   Fig. 6.   Immunoblotting was performed on LV myocardial extracts taken from control (C) and pacing CHF groups for interstitial collagenase (MMP-1), gelatinase A (MMP-2), and stromelysin (MMP-3). Top: a positive immunoreactive band was observed at 52/57-kDa region in LV myocardial extracts that corresponded to proenzyme form of MMP-1. In addition to substitution of nonimmune antisera, which abolished all staining, cell culture media from PMA-stimulated HT-1080 fibrosarcoma cell line was used as a positive control (+) for immunoblotting and revealed an immunoreactive band corresponding to MMP-1. Middle: MMP-2 was localized to 72/66-kDa region. Bottom: MMP-3 was localized to 54-kDa region, corresponding to molecular weights consistent with species of MMPs reported previously. Positive controls (+) for MMP-2 and MMP-3 were obtained from human epithelial and fibroblast cell lines (AG771 and AG770, respectively). Relative abundance of each MMP species was quantitated using densitometric methods and is summarized in RESULTS.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

Irrespective of the etiology, the development and progression of CHF is accompanied by changes in LV geometry and myocardial remodeling. Past clinical trials, such as with angiotensin-converting enzyme inhibition, have emphasized the importance of LV dilation with respect to morbidity and mortality (10, 22, 23). The LV myocardial collagen matrix has been proposed to play an important role in the maintenance of LV geometry and myocyte alignment (8, 29, 35). Alterations in the composition and structure of the LV myocardial collagen matrix have been reported to occur in several cardiac pathologies (16, 31, 36). For example, Weber et al. (36) reported a discontinuity of the fibrillar collagen weave in myocardial sections taken from patients with end-stage dilated cardiomyopathy. The overall content and composition of the LV myocardial collagen matrix is primarily determined by the balance between matrix synthesis and degradative processes (5, 31, 35-37, 39). The family of enzymes that has been implicated to play an important role in extracellular matrix degradation is the MMPs (5, 37, 39). Recently published clinical and experimental studies have shown increased MMP activity within the LV myocardium following the development of end-stage CHF (1, 16, 33). For example, Gunja-Smith et al. (16) reported increased MMP zymographic activity for a gelatin substrate with dilated cardiomyopathy. However, the relative abundance of specific MMP species within the normal and CHF myocardium and the relation to MMP activity remains unclear. Chronic rapid pacing in animals has been demonstrated previously to cause LV dilation and pump dysfunction, which subsequently leads to CHF (1, 11, 22, 30, 31). Accordingly, the present study examined MMP activity and species abundance in LV porcine myocardium in the control state and with the development of pacing CHF. The significant and unique finding of this study was that the development of pacing CHF was accompanied by an increase in the abundance of specific MMP species as well as increased MMP zymographic activity. Thus these findings suggest that changes in MMP abundance and activity potentially contribute to the LV myocardial remodeling that occurs with the development of CHF.

The MMP family is currently divided into 4 groups, or subclasses, which include 14 distinct MMP species (5, 37, 39, 40). For example, the interstitial collagenases, such as MMP-1, have been characterized as having a high affinity for the fibrillar collagens, types I, II, and III (5, 37, 39). Another subclass, the gelatinases, which include MMP-2, cleaves denatured collagen products as well as basement membrane components such as collagen type IV (5, 37, 39). Finally, an important subclass, the stromelysins, which includes MMP-3, cleaves a wide variety of extracellular matrix components (4, 5, 39), and more importantly, other MMP species can be activated by MMP-3 (17-19, 26, 40). Thus the digestion of extracellular matrix components is achieved through the activity of several MMPs. For example, interstitial collagen degradation is mediated by a MMP enzymatic cascade (5, 18). Accordingly, we examined the MMPs required for the complete degradation of interstitial collagen, MMP-1, -2, -3, and -9. In the present study, gelatin, collagen III, and casein were used as zymographic substrates to measure LV myocardial MMP activity. Zymographic activity against the substrate gelatin or collagen III was observed in control and pacing CHF LV myocardial extracts at the region between 84 and 40 kDa. This activity may reflect MMP-1, MMP-2, or MMP-3 activity. The lytic zone observed at 84 kDa was particularly pronounced in the pacing CHF group when compared with controls. This indicates that MMP-9 induction may occur with pacing CHF. Past studies have suggested that casein may be a specific substrate for MMP-3 (12, 15). In both control and pacing CHF LV myocardial extracts, caseinolytic activity was observed near the 50-kDa region, which is consistent with the molecular weight of MMP-3 (5, 12, 17, 37, 39). Thus, in both control and pacing-CHF LV myocardial extracts, zymographic activity was observed against the substrates gelatin, collagen III, and casein, which likely reflects the activity of several species of MMPs. More importantly, in pacing CHF LV myocardial extracts, zymographic activity for all of these substrates was increased with the development of pacing CHF.

The MMPs are synthesized intracellularly and after posttranslational modification are secreted into the extracellular space in a proenzyme (zymogen) form. Within the extracellular space, it has been proposed that MMPs may be associated with various components of the extracellular matrix while remaining in the latent, nonactivated state (37, 39). In this latent state, the MMP catalytic site is concealed by a propeptide domain mediated by a Cys-Zn²⁺ bond. For MMP activation to occur, the NH₂-terminal sequence of the propeptide domain is cleaved, resulting in the exposure of the Zn²⁺ binding site of the catalytic domain (5, 37, 39). Past in vitro studies have demonstrated that MMP activation can be achieved by chaotrophic agents and organomercurials (5, 37, 39). However, it has been proposed that in vivo MMP activation occurs through the cleavage of the NH₂-terminal sequence by serine proteases (5, 26, 37, 39). For example, Nagase et al. (27) demonstrated that the serine proteases, plasmin and trypsin, generated an active form of MMP-3. Thus, in the present study, we also incubated LV myocardial extracts with trypsin to determine total recruitable MMP activity. After trypsin activation, LV myocardial zymographic activity against the substrates gelatin and collagen III was increased by approximately twofold. Importantly, total recruitable LV myocardial zymographic activity was increased with the development of pacing CHF.

To identify the relative abundance of MMP species, we used immunoblotting procedures. With the development of pacing CHF, the relative abundance of all studied MMPs, interstitial collagenase (MMP-1), gelatinase A (MMP-2), and stromelysin (MMP-3) was increased. MMP-1 has a high affinity for the fibrillar collagens (5, 37, 39), and it has been demonstrated previously that alterations in LV myocardial fibrillar collagen occurred with the development of pacing CHF (31). The results of the present study suggest that a contributory mechanism for changes in LV myocardial fibrillar collagen with the development of pacing CHF may be increased MMP-1 abundance. The present study showed that, with pacing CHF, increased myocardial gelatinase activity occurred as observed in human hearts with the development of CHF (1, 16, 33). The results of the present study suggest increased LV myocardial gelatinase activity observed with CHF may be because of an absolute increase in MMP-2 abundance. MMP-3 is an important constituent of the MMP family that degrades a wide range of extracellular matrix components and activates other latent MMPs (5, 37, 39). In addition to increased MMP-3 abundance, LV myocardial zymographic activity against the substrate casein, which has been suggested to be specific for MMP-3 (12, 15), was increased with the development of pacing CHF. Therefore, the increased MMP-3 abundance and activity observed with the development of pacing CHF may play a contributory role in LV myocardial collagen remodeling directly through the degradation of collagen matrix components as well as the activation of endogenous myocardial MMPs.

Past clinical and experimental studies have shown that cardiac pathologies such as cardiomyopathy are accompanied by both alterations in the myocardial fibrillar collagen matrix and increased MMP activity (1, 16, 31, 33, 36). For example, in a previous study, this laboratory examined changes in the composition of the fibrillar collagen matrix with the development of pacing CHF (31). Specifically, myocardial collagen content, as measured by hydroxyproline, was decreased by 23% with pacing CHF. Furthermore, morphometric analysis revealed that the continuity of the fibrillar collagen weave was decreased with pacing CHF. Thus the present study builds on these past studies by demonstrating that with pacing CHF, LV myocardial MMP substrate specific activity and abundance were increased. Furthermore, the results of the present study indicate that there is a correlative relationship between LV geometry and myocardial MMP activity. Taken together, these studies suggest that a possible mechanism for the collagen remodeling that occurs with pacing CHF may be increased MMP activity and abundance.

There are limitations to the present study that must be recognized. First, in the present study, in vitro zymographic assays were performed that may not reflect endogenous in vivo MMP activity and activational state. Furthermore, in vivo MMP activity has been demonstrated to be regulated by the tissue inhibitors of metalloproteinases or TIMPs (5, 37, 39). Thus the in vitro assays presently used possibly disrupted TIMP/MMP complexes that may have been present in vivo. Future studies that examine endogenous in vivo MMP activity and activational state are warranted. Second, the present project employed a model of chronic rapid pacing that produced changes in LV functional and neurohormonal characteristics similar to that of the clinical spectrum of CHF (1, 3, 9, 20, 22, 31, 34). With the use of this animal model of CHF, changes in MMP activity and abundance were examined in the absence of confounding influences that may be encountered in clinical studies. However, it must be recognized that any animal model will not fully represent the complex clinical spectrum of CHF. Specifically, the changes in LV myocardial structure that occur with pacing CHF are not similar to clinical forms of CHF because of chronic ischemia or hypertensive disease (20). Furthermore, this model of chronic rapid pacing for 3 wk produces a rapidly progressive model of CHF that is in contrast to clinical forms of CHF that occur over a period of months to years. Nevertheless, the results of the present study provide direct evidence that a likely contributory mechanism for the LV remodeling process that occurs with CHF may be changes in MMP activity and abundance.