INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

MANY AUTHORS (3, 17, 25, 35) have hypothesized that the extracellular matrix (ECM) fibrillar collagen network facilitates transduction of cardiomyocyte contraction into myocardial force development and shortening and that changes in collagen may also alter myocardial systolic performance. The basis for this hypothesis is partly because pathological states, which result in the development of systolic dysfunction and congestive heart failure, are associated with significant changes in fibrillar collagen (8, 11, 13, 29, 34, 36, 37). However, these same pathological processes simultaneously change several other cellular and extracellular factors, which can also play a causal role in the development of systolic dysfunction and congestive heart failure (1). In addition, the specificity with which a decrease in collagen can be linked causally to a decrease in systolic function is uncertain because some pathological processes that increase collagen as well as some that decrease collagen can result in systolic dysfunction (8, 11, 13, 29, 34, 36, 37).

Therefore, it is not intuitively obvious whether and to what extent changes in ECM myocardial fibrillar collagen can alter systolic function or contribute to the development of congestive heart failure in pathological states. Determining whether collagen degradation plays a mechanistic role in altering systolic function requires an experiment designed to cause an acute, isolated change in collagen without significantly changing any of the cellular determinants of systolic function. Accordingly, the purpose of this study was to produce an acute, isolated increase in collagen degradation without changing any cellular determinants, measure the resultant effect on myocardial systolic performance, and thereby test the hypothesis that the ECM fibrillar collagen network facilitates transduction of cardiomyocyte contraction into myocardial force development and shortening and helps to maintain normal myocardial systolic performance.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experimental model. Ventricular papillary muscles were isolated from six normal rats, eight normal cats, and eight cats with chronic right ventricular pressure overload hypertrophy (POH) induced by pulmonary artery banding for 4 wk. Initial pilot studies were performed in rats to establish protocol parameters. Protocols were then applied to normal and POH cats. POH was induced by partially occluding the pulmonary artery with a 2.9 mm ID band, with the use of previously described methods (41). Eight cats underwent pulmonary artery banding and then recovered for 4 wk. Four normal adult cats and four sham-operated cats served as controls. Four weeks after pulmonary artery banding or sham operation, catheterization was performed with the use of previously described methods (41). The effects of pulmonary artery banding on in vivo measurements of pressure, oximetry, and mass were similar to those in our previous studies (41), in that pulmonary artery banding caused significant increases in right ventricular systolic pressure and mass. After these hemodynamic studies were completed, left ventricular papillary muscles were isolated from normal rats, and right ventricular papillary muscles were isolated from cats using previously described methods (41). Once isolated, the papillary muscles underwent experimental treatment protocols.

Experimental treatment protocols. First, the isolated muscles were immediately placed in a 50-ml container and superfused with the "experimental treatment buffer," which consisted of the cardioplegia solution, as described previously (41), with either plasmin (plasmin-treated group) or without plasmin (untreated control group) for 240 min. Plasmin is a serine protease that activates endogenous matrix metalloproteinases (MMPs). Activated MMPs have the ability to cleave specific ECM proteins, including fibrillar collagen. Paired muscles were randomly assigned to the untreated control protocol versus the plasmin-treated protocol. The plasmin-treated group was superfused with the experimental treatment buffer containing 0.64 U/ml concentration of plasmin. Initial studies to determine optimal plasmin concentration and length of treatment with plasmin were performed in normal rats. Dose and time titration studies were done from 0.16-0.64 U/ml and from 0 to 240 min. These studies showed that the amount of hydroxyproline liberated into the bath reached a plateau at a plasmin dose of 0.64 U/ml for 240 min. In addition, these dose and time ranges resulted in a significant effect in myocardial systolic performance in normal muscles. After plasmin treatment, the muscle was placed in the isolated muscle chamber and subjected to the mechanical testing protocol described below. Untreated control muscles were treated in a fashion identical to that described above, except that the experimental treatment buffer did not contain plasmin.

Papillary muscle servo control system. A computer-controlled dual-mode (length and force) servo system was used to mechanically test the papillary muscles. Studies examining the effects of treatment with plasmin on myocardial diastolic viscoelastic properties from these same papillary muscles are the subject of a separate report. This study examined the effects of treatment with plasmin in normal and POH on myocardial systolic performance. The characteristics of the servo control system, the preconditioning protocol, and the methods used to measure muscle length (ML) and cross-sectional area, have been described in detail (41). Myocardial systolic performance was examined in ventricular papillary muscles by measuring the following: extent and velocity of isotonic shortening, extent and rate of isometric force development, and the developed tension versus resting ML relationship. Length was expressed as a percentage of the rest length at maximum length (L_max) ML and force was normalized by muscle cross-sectional area (mN/mm²). Maximum isotonic shortening extent and velocity (dL/dt) were measured as ML and ML/s, respectively. Maximum isometric developed tension and the rate of tension development (+dT/dt) were measured as mN/mm² and mN · mm² · s¹, respectively. The developed tension versus resting ML relationship was derived from the L_max determinations in each muscle.

MMP activation. In a subset of studies (n = 3), the experimental treatment buffer from the untreated control and the plasmin-treated muscles were rapidly decanted and flash frozen. These samples were subjected to substrate-specific MMP zymographic analysis, as described previously (21, 27, 32). The buffer was initially lyophilized and then reconstituted in zymographic buffer solution. Purified MMP-9 and MMP-2 standards (CalBiochem; La Jolla, CA) were included in all zymograms. After electrophoresis and incubation, the zymograms were stained and digitally quantitated by image analysis (Gel Pro Analyzer, Media Cybernetics) (21, 27, 32).

Hydroxyproline assay. Collagen degradation was quantified by measuring the amount of hydroxyproline present in the experimental treatment buffer of both untreated control and plasmin-treated normal rat papillary muscles. Hydroxyproline concentration was determined by using a modified method of Stegemann and Stalder (30). Hydroxyproline standard solutions were made, and a standard curve was obtained. Two milliliters of the experimental treatment buffer were lyophilized, resuspended with the assay buffer, and mixed with 500 ml of chloramine-T reagent. An aldehyde-perchloric acid solution was added, the samples were immersed in a 60°C bath for 15 min, and then cooled in tap water. The absorbency of the samples was obtained, and the hydroxyproline concentration was assessed using the curve obtained from the standard solutions.

ECM characterization scanning and transmission electron microscopy. Fibrillar collagen distribution and geometry were characterized using scanning electron microscopy (EM), as previously described (14, 28, 29). After measurements of systolic performance were completed, muscles were held at a constant length equal to the length at L_max. The mechanical testing buffer was then removed and replaced by 2.5% gluteraldehyde fixative for 3 h. The myocardial samples were critical point dried and then freeze fractured. The freeze-fractured samples were dehydrated and gold sputter coated (Hummer II; Technics). The sections were examined with a JEOL JSM-25S scanning electron microscope at an accelerating voltage of 15 kV. Myocardial samples were prepared in triplicate and at least 10 photomicrographs of the extracellular space were obtained from each sample. Additional muscle sections were prepared for EM as described previously (29). Briefly, thin myocardial sections were prepared from each papillary muscle, stained with uranyl acetate and lead citrate, and examined with a JEOL JEM-1210 electron microscope. The central portion of each section was digitized at a precalibrated magnification.

Measurements of cardiomyocyte systolic performance. In a subset of studies (n = 4 normal cats), cardiomyocyte systolic performance was examined by measuring the extent and velocity of isotonic cell and sarcomere shortening. Cardiomyocytes were isolated from feline myocardium with the use of previously described methods (18). Cardiomyocytes were divided into two aliquots; one was treated with plasmin (Krebs-Henseleit buffer with plasmin, 0.64 U/ml for 240 min) and the second was the untreated control (Krebs-Henseleit buffer without plasmin for 240 min). A minimum of 15 cells from each of the two aliquots from each of four isolations was studied (n = 60 cells/group). The cells were placed in a custom-designed study chamber, superfused with Krebs-Henseleit buffer, including 2.5 mM calcium, maintained at pH 7.4, 37°C, and electrically stimulated to contract at 1 Hz. Cardiomyocyte and sarcomere length and velocity transients were determined using computer assisted videomicroscopy and computer controlled video edge-detection protocols (IonOptix; Milton, MA). The experimental apparatus used in this study did not allow examinations of isometric contractions in isolated cardiomyocytes and did not allow isolated cardiomyocytes to undergo isotonic contractions at L_max preloaded length. Therefore, the mechanical studies performed in isolated cardiomyocytes do not precisely correspond to the studies performed in isolated papillary muscles. However, they were sufficient to determine the effects of plasmin on cardiocyte contractile state.

Statistics. Data are presented as means ± SE for each data group. Differences between normal and POH groups at baseline and after plasmin treatment were determined with multiple analysis of variance and a Newman-Keuls multiple-sample comparison test and were considered significant at P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Effects of plasmin on MMP activation and collagen degradation. Representative zymograms from the experimental treatment buffer from untreated control and plasmin-treated muscles were analyzed (Fig. 1). In the untreated controls, there was no detectable zymographic activity in the buffer. This absence of MMP zymographic activity suggested that MMPs were not spontaneously activated and released into the buffer during this treatment period. In marked contrast, however, MMP zymographic activity contained within the 50-90 kDa region, consistent with proteolytic activity for MMP-9 and MMP-2, respectively (38), was significantly increased with plasmin treatment. These results indicated that treatment with the serine protease plasmin caused MMP activation and release into the experimental treatment buffer superfusing the muscle. Moreover, there appeared to be a greater degree of MMP zymographic activity contained within the experimental treatment buffer from plasmin-treated muscles from the POH muscle samples compared with the normal muscles treated with plasmin. Densitometric analysis of the proteolytic region spanning from 50-90 kDa revealed a greater lytic region contained within the POH samples compared with normal samples (1,049 ± 33 × 10³ vs. 922 ± 2 × 10³ pixels, respectively, P = 0.033, Wilcoxon's rank sum test).

View larger version (58K): [in this window] [in a new window]   Fig. 1.   Representative gelatin zymograms obtained from the experimental treatment buffer. Lane 1: proteolytic activity for purified pro-matrix metalloproteinase (MMP)-9 (~92 kDa) and pro-MMP-2 (~72 kDa). Lane 2: no observable gelatinolytic activity was present in the experimental treatment buffer in the absence of plasmin pretreatment (Plasmin). However, robust gelatinolytic activity was observed in the experimental treatment buffer of normal + pressure overload hypertrophied (POH) myocardium (lane 3) after plasmin treatment (+Plasmin) (lane 4).

View larger version (184K): [in this window] [in a new window]   Fig. 2.   Scanning electron microscopy (EM) was performed on papillary muscle samples, which had been fixed at maximum length (Lmax) preload. In normal myocardium (top left), a fine fibrillar collagen weave was observed within the interstitium. In normal myocardium after treatment with plasmin (top right), significant loss of the fibrillar collagen weave could be readily detected in all samples examined. In POH muscles (bottom left), a significantly thickened collagen weave surrounded individual myocytes and a robust fascicular collagen matrix was observed throughout the sections examined. In POH muscles after plasmin treatment (bottom right), noticeable defects in the thick collagen fascicular weave were readily apparent.

View larger version (210K): [in this window] [in a new window]   Fig. 3.   Representative transmission electron micrographs from normal muscle treated with (B) and without (A) plasmin. In untreated normal myocardium, the sarcolemma was intact and clear sarcomeric registration within individual myofibrils could be appreciated (A, top and bottom). The mitochondria revealed mild swelling, which separated the christae; there were no differences between control vs. plasmin-treated muscles. This effect was likely due to the 3-h storage of the muscle segments in a crystalloid buffering solution. In sections taken from muscle segments treated with plasmin, there were no gross defects in the sarcolemmal membrane or other architectural evidence of significant myocyte injury. However, in certain sections examined, there appeared to be a higher prevalence of adjacent myofibrils, in which poor registration of Z-bands occurred (B, top and bottom).

Effects of plasmin on myocardial systolic performance in normal muscles. The effects of plasmin treatment on the systolic performance in normal rat papillary muscles are shown in Fig. 4. When normal rat papillary muscles were treated with plasmin, there was a significant decrease in dL/dt and a significant decrease in +dT/dt. The effects of plasmin treatment on the systolic performance in normal cat papillary muscles are shown in Fig. 5. When normal feline papillary muscles were treated with plasmin, there was a significant decrease in dL/dt from 0.17 ± 0.01 ML and 0.91 ± 0.06 ML/s in the untreated muscles to 0.09 ± 0.01 ML and 0.62 ± 0.16 ML/s in the plasmin-treated muscles (P < 0.05). Plasmin caused a significant decrease in +dT/dt from 45 ± 4 mN/mm² and 201 ± 11 mN · mm² · s¹ in the untreated muscles to 36 ± 3 mN/mm² and 173 ± 15 mN · mm² · s¹ in the plasmin-treated muscles (P < 0.05).

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Effects of plasmin treatment on normal rat myocardial systolic performance. A: example of myocardial shortening of an untreated control (solid trace) and of a plasmin-treated (dashed trace) papillary muscle from a normal rat. B and C: group data from treated and untreated normal rat papillary muscles. Plasmin treatment significantly decreased shortening extent and velocity. D: example of myocardial tension development of an untreated control and of a plasmin-treated papillary muscle from a normal rat. E and F: group data from treated and untreated normal rat papillary muscles. Plasmin treatment significantly decreased developed tension and the rate of tension development (+dT/dt). ML, muscle length.

View larger version (24K): [in this window] [in a new window]   Fig. 5.   Effects of plasmin treatment on normal cat myocardial systolic performance. A: example of myocardial shortening of an untreated control (solid trace) and of a plasmin-treated (dashed trace) papillary muscle from a normal cat. B and C: group data from treated and untreated normal cat papillary muscles. Plasmin treatment significantly decreased shortening extent and velocity. D: example of myocardial tension development of an untreated control and of a plasmin-treated papillary muscle from a normal cat. E and F: group data from treated and untreated normal cat papillary muscles. Plasmin treatment significantly decreased +dT/dt.

Effects of plasmin on myocardial systolic performance in POH cat. The effects of plasmin treatment on systolic performance in the POH cat papillary muscles are shown in Fig. 6. When POH muscles were treated with plasmin, there was a significant decrease in dL/dt from 0.08 ± 0.01 ML and 0.34 ± 0.03 ML/s to 0.04 ± 0.2 ML and 0.21 ± 0.04 ML/s (P < 0.05), and in +dT/dt from 20 ± 2 mN/mm² and 94 ± 11 mN · mm² · s¹ to 13 ± 4 mN/mm² and 65 ± 14 mN · mm² · s¹ (P < 0.05).

View larger version (27K): [in this window] [in a new window]   Fig. 6.   Effects of plasmin treatment on POH cat myocardial systolic performance. A: example of myocardial shortening of an untreated control (solid trace) and of a plasmin-treated (dashed trace) papillary muscle from a POH cat. B and C: group data from treated and untreated POH cat papillary muscles. Plasmin treatment significantly decreased shortening extent and velocity. D: example of myocardial tension development of an untreated control and of a plasmin-treated papillary muscle from a POH cat. E and F: group data from treated and untreated POH cat papillary muscles. Plasmin treatment significantly decreased +dT/dt.

Effects of POH and plasmin on developed tension versus length relationship. POH caused a shift downward of the developed tension versus resting ML relationship such that for any given rest length, developed tension was less in the POH muscles (Fig. 7). In normal rat, normal cat, and POH cat muscles, plasmin treatment caused a downward shift in this relationship, indicating that plasmin decreased myocardial systolic performance.

View larger version (19K): [in this window] [in a new window]   Fig. 7.   A: relationship between developed force and ML in normal rat papillary muscles without and with plasmin treatment. B: relationship between developed force and ML in normal cat papillary muscles without and with plasmin treatment and in POH papillary muscles without and with plasmin treatment. *P < 0.05 vs. normal untreated; #P < 0.05 vs. POH untreated. Plasmin treatment moved the developed force vs. ML relationship downward.

Effects of plasmin on systolic performance of normal cat cardiomyocytes. The effects of plasmin treatment on the systolic performance in normal cat cardiomyocytes are shown in Fig. 8. Plasmin treatment did not alter the cardiomyocyte or sarcomere dL/dt.

View larger version (20K): [in this window] [in a new window]   Fig. 8.   Effects of plasmin treatment on normal cat cardiomyocyte systolic performance. A-D: group data for normal (NML) cat cardiomyocytes. Plasmin treatment did not affect cardiomyocyte shortening extent (A), cardiomyocyte shortening velocity (B), sarcomere shortening extent (C), or sarcomere shortening velocity (D).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The major findings of this study include the following: 1) treatment of isolated papillary muscles from normal and POH cats with serine protease plasmin caused activation of MMPs and collagen degradation; 2) myocardial force development and shortening were significantly decreased after treatment with plasmin; however, 3) plasmin treatment did not alter cardiomyocyte shortening. Thus acute disruption of the fibrillar collagen network caused a decrease in myocardial systolic performance without changing cardiomyocyte contractility. Therefore, these data support the hypothesis that fibrillar collagen facilitates transduction of cardiomyocyte contraction into myocardial force development and helps to maintain normal myocardial systolic performance.

There are at least two possible mechanisms by which fibrillar collagen might facilitate transduction of cardiomyocyte contraction into normal myocardial contractility. First, myocardial fibrillar collagen forms a complex three-dimensional matrix, which surrounds, invests, and connects individual cardiomyocytes and groups of cardiomyocytes (fascicles). Fibrillar collagen may help to maintain cardiac muscle structural organization, which consists of geometrically positioned cardiomyocyte fascicles. Within the muscle, cardiomyocyte fascicles act in a coordinated and homogeneous fashion to produce efficient contraction. Degradation of the fibrillar collagen matrix, such as occurs with MMP activation, may result in discontinuities in the collagen matrix and disruption of the matrix-mediated support, geometric alignment, and coordination of adjacent cardiomyocyte fascicle contraction. Therefore, one mechanism that might cause the reduction of force development and shortening observed in the present study after treatment with the serine protease plasmin was a loss of optimal geometric alignment and coordination of cardiomyocyte fascicles within the papillary muscle.

A second mechanism might involve the loss of the normal collagen matrix-basement membrane-integrin relationship. Collagen attaches to basement membrane proteins, which in turn form connections to integrin-mediated focal adhesion complexes at the surface of cardiomyocytes. These connections occur predominantly at the sarcomeric Z-line and the cardiomyocyte intercalated disks. Endomysial collagen struts form lateral connections between cardiomyocytes. These collagen connections are thought to contribute to the processes holding sarcomeres in registry at equal end diastolic lengths across myofibrils at the start of contraction. This creates equally preloaded sarcomeres across myofibrils. In addition, these collagen connections are thought to contribute to the processes keeping sarcomeres in registry throughout contraction, creating a homogeneous contraction. Furthermore, disruption of this sarcomeric registration might lead to a nonhomogeneous contraction creating both asynchrony and asynergy, which, in turn, result in a decrease in contractile state. Data from the present study support this notion. In plasmin-treated muscles, the reduction in contractility was associated with some alterations in the normal sarcomeric Z-band registry.

Chronic changes in fibrillar collagen. Several studies have attempted to define the relationship between changes in ECM fibrillar collagen and myocardial systolic performance. Disease states such as POH, dilated cardiomyopathy, and myocardial infarction cause decreases in systolic function, which have been associated with a decrease, an increase, or no change in fibrillar collagen. This variability makes it difficult to assign a causal relationship between changes in collagen and the development of systolic dysfunction.

Acute changes in fibrillar collagen. Previous experimental approaches (6, 16, 31, 33, 40) attempted to accomplish the goal of an acute increase in collagen degradation using methods such as acute ischemia, treatment with disulfide reagents, or treatment with exogenous collagenase. The results of these studies have been difficult to interpret because the results were not consistent and because many determinants of systolic performance, including intracellular mechanisms affecting energetics and calcium metabolism and extracellular mechanisms affecting myocardial edema, were altered simultaneously.

POH. The present study used normal and POH muscle preparations to examine specific determinants of cardiac muscle function after plasmin treatment and putative MMP activation. A complete quantification of the POH muscle preparations was beyond the scope of this report. The characteristics of this POH preparation and the alterations in cellular and ECM structure have been described previously. However, two observations were made about changes in the ECM and MMPs in these POH muscles, which warrant further comment. First, the scanning electron micrographs revealed a thickened fibrillar collagen weave in the POH muscles, which is consistent with this form of hypertrophy. Second, a more robust gelatinolytic activity was realized in POH muscles after plasmin treatment. This likely reflects a greater quantity of recruitable MMPs in the POH myocardium. Whereas the increased collagen accumulation and MMP abundance in the POH myocardium may appear paradoxical, there are several regulatory processes that contribute to MMP activational states. First, MMPs are synthesized in an inactive state and require an exogenous biochemical trigger or physical stimulus to yield a fully active MMP. Second, a set of regulatory proteins, tissue inhibitors of MMPs (TIMPs), exists within the myocardium that rapidly and tightly binds to active MMPs and interrupt proteolytic activity. The induction of a pressure overload has been demonstrated to cause a shift in the stoichiometry between MMPs and TIMPs, which would favor reduced MMP proteolytic activity and collagen accumulation. The results of the present study demonstrated that a significant amount of recruitable MMPs exist within the ECM of the POH myocardium. Future studies that more fully determine the relative abundance and MMP species, which exist within the POH myocardium, are warranted.

Components of ECM. There are other components of ECM that may also contribute to the maintenance of systolic function. In addition to fibrillar collagen, the myocardial ECM is composed of the following: 1) other fibrillar protein such as elastin; 2) proteoglycans; and 3) basement membrane proteins, such as collagen type IV, laminin, and fibronectin. There is no clear evidence that proteins other than fibrillar collagen change during clinical pathological processes or what role these changes would have. Nonetheless, this may have relevance to the present study because plasmin does not act directly on fibrillar collagen; rather, plasmin degrades collagen indirectly by activating a cascade of MMPs, which in turn can degrade collagen. This cascade may also activate MMPs, whose specific protein targets include other proteins within the ECM in addition to fibrillar collagen. To prove what role each MMP species and each MMP target protein contributes to change in systolic performance will require additional studies. This kind of specificity may be possible either by treatment with selective recombinant MMPs or by the addition of selective MMP inhibitors to the plasmin treatment. However, the first step in such experimental designs was taken in the current study, providing proof of concept activation of endogenous MMPs and degradation of collagen resulted in a decrease in myocardial systolic performance.