Physiol. Rev. 87: 1285-1342, 2007;
doi:10.1152/physrev.00012.2007
0031-9333/07 $18.00
Myocardial Matrix Remodeling and the Matrix Metalloproteinases: Influence on Cardiac Form and Function
Francis G. Spinale
Division of Cardiothoracic Surgery, Medical University of South Carolina, Charleston, South Carolina
 |
ABSTRACT
|
|---|
It is now becoming apparent that dynamic changes occur within the interstitium that directly contribute to adverse myocardial remodeling following myocardial infarction (MI), with hypertensive heart disease and with intrinsic myocardial disease such as cardiomyopathy. Furthermore, a family of matrix proteases, the matrix metalloproteinases (MMPs) and the tissue inhibitors of MMPs (TIMPs), has been recognized to play an important role in matrix remodeling in these cardiac disease states. The purpose of this review is fivefold: 1) to examine and redefine the myocardial matrix as a critical and dynamic entity with respect to the remodeling process encountered with MI, hypertension, or cardiomyopathic disease; 2) present the remarkable progress that has been made with respect to MMP/TIMP biology and how it relates to myocardial matrix remodeling; 3) to evaluate critical translational/clinical studies that have provided a cause-effect relationship between alterations in MMP/TIMP regulation and myocardial matrix remodeling; 4) to provide a critical review and analysis of current diagnostic, prognostic, and pharmacological approaches that utilized our basic understanding of MMP/TIMPs in the context of cardiac disease; and 5) most importantly, to dispel the historical belief that the myocardial matrix is a passive structure and supplant this belief that the regulation of matrix protease pathways such as the MMPs and TIMPs will likely yield a new avenue of diagnostic and therapeutic strategies for myocardial remodeling and the progression to heart failure.
|
I. INTRODUCTION
|
|---|
Heart failure is a common cause of morbidity and mortality, and the incidence is increasing (115, 169, 177, 190, 260, 359, 465, 493). Following a specific cardiovascular stress, a cascade of compensatory structural events occurs within the myocardium and contributes to eventual left ventricular (LV) dysfunction and the manifestation of the heart failure syndrome. The cascade of structural events that occurs within the myocardium is highly complex, and therefore, for the purposes of this review, the focus will be on structural changes that occur within the myocardium in cardiac disease states such as myocardial infarction (MI), LV hypertrophy (LVH), or cardiomyopathic disease. In these instances, structural changes occur within the myocardial wall that result in changes in LV geometry, a process termed myocardial remodeling (7, 62, 75, 76, 102, 135, 209, 310, 335, 426, 429, 487, 494, 495). Myocardial remodeling is the summation of both cellular and extracellular processes, and a number of these processes have been the subject of past articles in Physiological Reviews (33, 49, 105, 225, 324, 390, 433). With respect to cellular processes, those that have been studied in detail include myocyte growth, apoptosis, and necrosis. With respect to the extracellular compartment and myocardial remodeling, a major clinical and basic research focus has been in the areas of vascular growth and patterning, as well as the conduction system. While historically considered a static structure, it is now becoming recognized that the myocardial extracellular matrix (ECM) is a complex microenvironment containing a large portfolio of matrix proteins, signaling molecules, proteases, and cell types that play a fundamental role in the myocardial remodeling process (28, 43, 52, 77, 78, 113, 134, 178, 183, 275, 303, 392, 486, 491, 535). Accordingly, the purpose of this review is to provide a brief overview of the myocardial ECM as it pertains to remodeling process in relevant cardiac disease states and then to examine in detail how the ECM undergoes remodeling through the induction and activation of a family of proteolytic enzymes, the matrix metalloproteinases (MMPs).
|
II. OVERVIEW OF MYOCARDIAL MATRIX AND REMODELING
|
|---|
A. Structure and Function of the Myocardial Matrix
The architectural complexity of the myocardium and the potential role that the ECM played in maintaining unique myocyte orientations throughout the LV free wall were described by Streeter and Basset (417, 418). Using a structural engineering approach, these authors demonstrated that myocyte orientation and myocardial fiber angles were highly organized and moved in a continuous fashion from the endocardium to the epicardium. It is the structural network of matrix proteins composed of proteins of highly organized structure and architecture, such as type I and type III collagen, that provide structural integrity to adjoining myocytes and contribute to overall LV pump function through the coordination of myocyte shortening. The architectural complexity of the myocardial ECM began to be appreciated through the use of scanning electron microscopy (2, 36, 488). These studies demonstrated the three-dimensional structure of the myocardial ECM and how the fibrillar weave surrounded and supported individual myocytes as well as fascicles of myocytes. Moreover, these initial studies demonstrated the complexity of the ECM and the structural interaction with the vascular compartment. Further research demonstrated that the myocardial ECM maintains alignment of myofibrils within the myocyte through a collagen-integrin-cytoskeleton-myofibril relation (50, 132, 254, 365, 367, 395, 398, 441). In addition to a fibrillar collagen network, a basement membrane, proteoglycans, and glycosaminoglycans, the myocardial ECM contains a large reservoir of bioactive molecules (17, 34, 41, 64, 89, 95, 101, 109, 111, 114, 188, 288, 360, 362, 492). For example, it has been demonstrated that the concentration of bioactive signaling molecules such as angiotensin II (ANG II) and endothelin (ET)-1 are over 100-fold higher within the myocardial interstitium than in plasma (89, 101, 288, 492). Moreover, cytokine activation and signaling such as that for tumor necrosis factor-
(TNF) is highly compartmentalized within the myocardial interstitium (95, 111). Growth factors such as transforming growth factor-
(TGF) are stored in a latent form within the myocardial interstitium and thereby form a reservoir of signaling molecules that directly influence myocardial ECM synthesis and degradation (67, 82, 218, 298, 399, 443). Moreover, mechanical stimuli such as stress or strain are likely transduced through the myocardial ECM to the cardiac myocyte, which in turn would directly affect myocyte growth (35, 37, 69, 217, 249–251, 313, 547). Thus structural changes that would occur within the myocardial ECM would in turn affect myocyte biology and the overall structure and function of the myocardium. As such, a number of studies have examined compositional and structural changes that occur within the myocardial ECM following the induction of a hypertensive stimulus, following myocardial injury such as with MI, and in cardiomyopathic disease. The specific and unique changes to the myocardial ECM as it pertains to these cardiac disease states are examined in the subsequent sections. This examination is not intended to be comprehensive, but rather to provide the foundation for a more in-depth review of a specific proteolytic pathway that contributes to myocardial ECM remodeling in relevant cardiac disease processes.
B. Myocardial Remodeling in Myocardial Infarction
A number of cellular and extracellular factors contribute to the complex process of myocardial remodeling following MI. Specifically, myocyte loss, hypertrophy of remaining myocytes, and increased size and number of nonmyocyte cells all contribute to changes in LV myocardial wall geometry. Significant alterations in the structure and composition of the myocardial ECM occur following MI (75, 76, 166, 182–185, 192, 205, 226, 251, 270, 313, 350, 392, 426, 432, 434, 468, 473, 474, 487, 491, 496, 497, 543, 547). Cardiac wound repair after MI involves temporarily-overlapping phases, which include an inflammatory phase and tissue remodeling phase. The first phase starts after coronary artery occlusion with or without reperfusion and involves degradation of normal ECM, invasion of inflammatory cells at the site of initial injury, and the induction of bioactive peptides and cytokines. Degradation of the ECM during the acute phase is considered to be an essential event that allows for the ingress of inflammatory cells as well as proliferation and maturation of macrophages and fibroblasts, and provides the necessary substructure for scar formation. In 1916, macroscopic/microscopic studies by Karsner and Dwyer (192) reported an early (within 24 h) disappearance of the normal collagen matrix within the infarcted myocardium, which was then followed by significant matrix deposition. However, the mechanisms which caused this early loss of myocardial ECM within the MI region followed by ECM synthesis and deposition remained elusive. In serial studies utilizing biochemical and computer-assisted morphometric techniques, several laboratories reported a loss of normal collagen strut formation, increased release of hydroxyproline (an amino acid primarily found in collagen), and reduced collagen cross-linking within the ischemic region very early post-MI (76, 166, 182, 184, 185, 226, 270, 350, 426, 432, 434, 468, 474, 496, 497, 535, 543). These early ECM events occurred prior to the egress of inflammatory cells into the MI region. For example, in a study by Judd and Wexler (182), an early reduction in hydroxyproline content occurred following myocardial injury in rats, which was then accompanied by increased hydroxyproline during longer time periods (Fig. 1). These reports speculated that the release of endogenous enzymes was responsible for the initial degradation of the myocardial ECM in the early post-MI period, which was then followed by a vigorous inflammatory response that amplified collagen proteolytic activity. Indeed, an early study by the Factor laboratory reported that by 3 h post-MI, the decrease in collagen integrity was accompanied by increased proteinase activity (434). In this early time period, LV myocardial ECM degradation and remodeling were associated with an increased probability of rupture (226). Thus dynamic changes occur within the myocardial ECM in the initial and early phases of the post-MI period, that directly affect the mechanical properties of the LV myocardium. As the MI period progresses over the next several days, an influx of inflammatory cells into the injured myocardium occurs which results in further proteolysis of cellular and ECM proteins. In addition, this inflammatory response causes proliferation and differentiation of fibroblasts and other interstitial cells, and the elaboration of bioactive molecules which contribute to a robust synthesis of ECM for the purposes of scar formation (23, 75, 76, 113, 114, 155, 266, 303, 383, 426, 491). These changes within the MI region yield distinctive cellular and extracellular phenotypic changes. For example, the differentiation and proliferation of fibroblasts within the MI region demonstrate a unique protein signature and function to not only synthesize ECM proteins critical for scar formation, but also contribute to the biophysical properties of the scar itself and have been termed myofibroblasts (34, 55, 75, 76, 114, 195, 331, 491, 499, 522, 547). Early studies utilizing corticosteroids demonstrated the importance of this inflammatory response with respect to scar formation and maturation (205, 473). However, it remains unclear if the exuberant release of proteolytic enzymes from inflammatory cells and myofibroblasts which occur during the initial MI wound healing response are critical for scar maturation.
View larger version (12K):
[in this window]
[in a new window]
|
FIG. 1. Myocardial hydroxyproline content, reflective of fibrillar collagen content, initially fell from reference control values following isoproterenol-induced myocardial infarction (MI) as shown by the open arrow. Myocardial hydroxyproline increased by day 7 (closed arrow) and increased further by  2 wk post MI. This demonstrated that collagen degradation precedes collagen accumulation and scar formation following MI. [From Judd and Wexler (182).]
|
|
The later phase of post-MI remodeling results in ECM changes within all regions of the LV: the MI region, the viable myocardium within the border zone, and the remote region. Within the MI region, the newly formed ECM provides a means to tether viable myocyte fascicles and thereby forms a substrate to resist deformation from the intracavitary stresses generated during the cardiac cycle (217, 249, 251, 313, 546, 547). Failure of ECM support has been associated with LV wall thinning and slippage of myocyte fascicles (7, 310, 494). This adverse remodeling process has been termed "infarct expansion" and occurs in the absence of additional myocyte injury or alterations in LV loading conditions (62, 102, 335, 429, 494, 495). This post-MI remodeling process is a clinically significant problem in that it can lead to LV dilation and dysfunction, the progression to heart failure. Indeed, this post-MI remodeling process, which includes changes in ECM structure and composition, is an independent predictor of morbidity and mortality (62, 495). It has been postulated that an acceleration of ECM degradation occurs within the myocardium within and surrounding the MI (border zone) and facilitates the infarct expansion process in this later phase of post-MI remodeling. The proteolytic systems which likely contribute to continued degradation of the ECM as it relates to infarct expansion will be reviewed in detail in a subsequent section. These synthetic/degradative events within the myocardial ECM occur in a time- and region-dependent manner following MI. With regard to the remote region, increased ECM deposition occurs between hypertrophied, viable cardiac myocytes and fascicles (184, 270, 331, 474, 546). Thus heterogeneous ECM remodeling occurs throughout the LV myocardium post-MI, which in turn will affect overall geometry and function. Therapeutic strategies that are targeted at minimizing the degree of infarct expansion must take into account the absolute requirement to maximize ECM stability within the MI region and identify when adverse ECM proteolytic activity may occur. Elucidating the molecular mechanisms which control ECM degradation in the post-MI period will yield specific therapeutic strategies that will facilitate the wound healing response, but attenuate the adverse myocardial remodeling which gives rise to infarct expansion and LV failure.
C. Myocardial Remodeling in Overload States
Two distinct patterns of LVH occur in response to a persistent load: pressure overload hypertrophy (POH) and volume overload hypertrophy (VOH). Changes in LV wall stress patterns, which are uniquely different in POH and VOH, likely drive the adaptive hypertrophic response. POH results in a concentric hypertrophy whereby wall thickness is increased while LV diameter remains the same or is decreased. A myocyte undergoes concentric hypertrophy by adding sarcomeres in parallel to achieve an increase in width. In contrast, VOH results in an eccentric hypertrophy whereby wall thickness remains the same or is decreased while LV diameter increases. Eccentric hypertrophy involves the addition of sarcomeres in series to achieve an increase in myocyte length. While in both forms of LVH, alterations in loading conditions form the central mechanical stimulus, local elaboration of growth factors such as TGF, and neurohormonal factors as such as ANG II and ET can modify the growth response, and have particular relevance to the ECM (2, 12, 34, 35, 41, 64, 132, 134, 217, 249, 250, 337, 373, 404, 441, 486, 488, 535). Specifically, since TGF, ANG II, and ET have all been demonstrated to induce increased collagen synthesis in vitro, then it follows that amplification in the formation and/or signaling of these bioactive molecules in vivo would lead to increased ECM accumulation with POH. Indeed, a structural hallmark of prolonged POH is significantly increased collagen accumulation between individual myocytes and myocyte fascicles. The extent and degree of ECM remodeling with POH were clearly exemplified by the landmark studies by Weber and Janicki (2, 488). As shown in Figure 2, these past studies demonstrated thickening of the collagen weave and overall increased relative content between myocytes in a nonhuman primate model of POH. While the accumulation of ECM with POH is not exclusive to collagen, these initial structural studies gave rise to the generic term for this extracellular remodeling process as myocardial fibrosis. In POH, the accumulation of ECM and eventual myocardial fibrosis significantly contributes to LV function. In particular, enhanced synthesis and deposition of myocardial ECM is directly associated with increased LV myocardial stiffness properties, which in turn causes poor filling characteristics during diastole (12, 35, 159, 191, 194, 312, 314, 421, 488). Indeed, recent clinical evidence suggests that progressive ECM accumulation and diastolic dysfunction are important underlying pathophysiological mechanisms for heart failure in patients with POH (196, 544, 545). In contrast to POH, LV end-diastolic volumes increase in VOH due to the persistently elevated preload, and as a result, a much different pattern of ECM remodeling occurs. In large-animal models of VOH due to chronic mitral regurgitation, the LV remodeling process is accompanied by a distinctive loss of collagen fibrils surrounding individual myocytes (88, 330, 409, 450). These changes in ECM support are associated with changes in isolated LV myocyte geometry where the cardiac cells increase in length. Representative scanning electron micrographs taken from a well-established model of canine mitral regurgitation (88, 409) are shown in Figure 3 and demonstrate the profound differences in ECM structure and composition compared with normal myocardium and that of POH. In rodent models of aortocaval fistula, another form of VOH, a similar pattern of LV remodeling and changes in the ECM structure and composition have been reported (96, 124, 127, 244, 274, 484). In both forms of LVH, a loss of normal ECM structure and function occurs which suggests that alterations in ECM degradative processes have occurred. With POH, the highly organized architecture of the ECM is replaced with a thickened, poorly organized ECM. Thus initial degradation of the normal ECM likely occurs with POH, which is then followed by decreased capacity for ECM degradation and turnover. In VOH, increased ECM proteolytic activity likely contributes to the reduced ECM content and support and thereby facilitates the overall LV remodeling process. Proteolytic enzymes that likely contribute to the ECM remodeling with LVH are reviewed in a subsequent section.
View larger version (181K):
[in this window]
[in a new window]
|
FIG. 2. Scanning electron micrographs taken from normal nonhuman primate left ventricular (LV) myocardium and following the induction of pressure overload hypertrophy (POH). These microscopic studies demonstrated thickening of the collagen weave and overall increased relative content between myocytes with POH. [From Abrahams et al. (2).]
|
|
View larger version (133K):
[in this window]
[in a new window]
|
FIG. 3. Representative scanning electron micrographs taken from normal canine LV myocardium and following chronic mitral regurgitation that causes a volume overload hypertrophy (VOH). In this well established model of VOH (88, 330, 409, 450, 484), a loss of normal ECM architecture was demonstrated between individual myocytes (arrows), and the collagen supporting network is poorly organized. (Figure kindly provided by Dr. Louis Dell'Italia, University of Alabama at Birmingham.)
|
|
D. Myocardial Remodeling in Cardiomyopathic Disease
The dilated cardiomyopathies are a classification of primary myocardial disease states which constitute a significant proportion of patients with heart failure (167, 341, 510). While the etiologies for dilated cardiomyopathy (DCM) are diverse, a general pathophysiological classification scheme has been developed and termed ischemic, idiopathic (nonischemic), or infectious. The pathophysiology of DCM is that an increase in LV ventricular chamber radius to wall thickness occurs which results in increased myocardial wall stress. This increased LV myocardial wall stress in turn can promote further dilation. Significant changes in the myocardial ECM occur with cardiomyopathic disease and likely facilitate this remodeling process (104, 146, 206, 326, 391, 489). Within the myocardial ECM, alterations in the myocardial fibrillar collagen architecture have been reported in patients with DCM. For example, morphometric studies have identified a reduction in the number and conformation of collagen struts between adjoining myocytes in samples taken from patients with DCM (104, 489). However, the extent and type of ECM remodeling which occurs with DCM is variable and is likely dependent on the underlying etiology. For example, in idiopathic DCM, the alterations in normal ECM architecture are associated with reduced collagen cross-linking, which will make the ECM more susceptible to degradation (146, 206). Animal models of pacing-induced DCM have demonstrated a time-dependent alteration in normal ECM structure and composition that is accompanied by the progressive changes in LV geometry and function (208, 410, 490, 537). Specifically, in this model of DCM, the progressive loss of normal myocardial ECM is accompanied by LV myocyte lengthening and a loss of myocyte adhesion to critical ECM components (208, 537). These early changes in myocardial ECM structure and composition likely contribute to the progression of LV dilation and dysfunction in this nonischemic model of DCM. In ischemic DCM, large areas of ECM accumulation, that is, focal areas of myocardial fibrosis, are accompanied by other areas of the myocardium in which ECM content and composition are reduced. For example, in an animal model of ischemic DCM created by repeated coronary embolizations, areas of collagen accumulation are accompanied by localized areas of collagen loss (156, 375). Thus a heterogeneous pattern of ECM remodeling occurs in ischemic DCM where enhanced ECM accumulation and degradation occur simultaneously. In cases of infectious DCM, the robust inflammatory response within the myocardium heralds the induction of cytokine and growth factor receptor pathways which can augment ECM synthesis and accumulation (71, 221, 228, 325, 372). Using a selective corrosive technique which reveals the collagen skeletal framework amenable for scanning electron microscopy, Rossi (368) demonstrated that significant ECM remodeling occurred in patients with DCM secondary to Chagastic myocarditis (221). Representative scanning electron micrographs of a normal specimen and one taken from a patient with Chagas' disease are shown in Figure 4. In this form of DCM, the normal myocardial collagen matrix was replaced by a poorly structured, emphysematous matrix between areas of adjacent myocytes and myocardial fascicles. In animal models of infectious/inflammatory DCM commonly induced by viral injections, a similar pattern of ECM degradation paralleled by a abnormal ECM accumulation and structure have been reported (228, 372). While the specific pattern of ECM remodeling is as diverse as the etiologies which give rise to the spectrum of cardiomyopathic disease, some commonalities do exist. First, there is a destruction of the normal ECM framework in terms of structural protein content and architecture. It is likely that this loss in ECM support contributes to the progressive LV dilation and dysfunction that are common pathophysiological underpinnings for DCM. Second, in patients with DCM, there is a heterogeneous and continuous cycle of ECM synthesis and accumulation as well as ECM degradation which occur within myocardium in a regional specific manner. Due to the fact that severe DCM is the most common clinical presentation for cardiac transplantation, a number of studies have been performed on the explanted myocardial specimens and identified ECM proteolytic pathways that likely contribute to the ECM remodeling process. These studies are reviewed in detail in a subsequent section.
View larger version (104K):
[in this window]
[in a new window]
|
FIG. 4. Representative scanning electron micrographs of a normal human LV myocardial specimen and one taken from a patient with Chagas' disease. In these micrographs, the cardiac cells have been digested using a corrosive technique (368). With the use of this approach, the complexity and three-dimensional aspects of the human myocardial ECM can be readily appreciated. In dilated cardiomyopathy secondary to Chagas' disease, a dense poorly organized matrix that has an emphysematous appearance (arrows) can be seen. [From Rossi (368).]
|
|
|
III. MYOCARDIAL MATRIX METALLOPROTEINASES
|
|---|
As a historical perspective, the recognition that a specific enzyme system existed for degradation of the ECM has been attributed to the seminal observations by Gross and Lampier (140). In these studies, it was demonstrated that incubating sterile sections of tadpole tail on an opaque collagen gel caused an area of lysis. In later studies, using the recently developed polyacrylamide gel electrophoretic method, Gross and Nagai (141) reported a collagenase reaction occurring at 70 kDa. These landmark studies provided the backbone for future work which utilizes the combination of a collagen substrate and electrophoresis to examine the presence of collagenolytic enzymes, termed zymography. Indeed, as detailed in a subsequent section, the recognition that collagenolytic enzymes existed within the myocardium were first described utilizing this zymographic technique (11, 43, 75, 128, 141, 455, 457, 458). While the initial proteolytic enzyme was termed collagenase, it became clear from subsequent studies that this was a heterogeneous family of enzymes that has now been termed the MMPs. The MMPs are a family of more than 25 species of zinc-dependent proteases that are essential for normal tissue remodeling in processes such as bone growth, wound healing, and reproduction (131, 252, 300, 315, 316, 323, 349, 394, 412, 414, 447, 456). Moreover, increased induction and elaboration of MMPs have been identified to hold biological significance in a number of pathological conditions that include cancer, inflammatory disease, and cardiovascular disease. As a consequence, the initial observations made by Gross and colleagues have resulted in an explosion of research regarding MMPs. In a review by Woessner (504), another pioneer in the field of ECM and proteolytic enzymes, it was estimated that by 1998 over 8,500 publications had appeared regarding MMPs, and the publication rate exceeded 1,000/year. As of the time of this review, over 9,000 publications in the past 5 years have been archived by the National Library of Medicine in the subject area of MMPs. Accordingly, an exhaustive treatise on the biochemistry, basic biology, and regulation of MMPs would be not be possible, nor entirely relevant to this review. Rather, a generalized description of the structure of MMPs and the current classification scheme will be briefly provided.
A. Taxonomy and Structure of the MMPs
A great deal of confusion can surround the nomenclature of MMPs, due to the fact that initially MMPs were named based on recognized substrates. However, as detailed in a subsequent section, a large and diverse portfolio of substrates exists for MMPs, and a great deal of crossover exists between substrates and MMP types. Table 1 provides the taxonomy number, generic and commonly used group names, and approximate molecular weights for MMP types that have thus far been identified to exist within the myocardium. However, substrate portfolios for specific MMPs are not provided, since active research is occurring in this area, and therefore, any attempt to compile such a list would need to be considered incomplete or obsolete. The generic or common names can be a source of confusion, and therefore, the specific MMP type under question is often referred to by number. For example, MMP-13 has been referred to as rodent collagenase, since this is the primary collagenase found in rodent myocardial samples (96, 227, 228, 334, 504). However, robust levels of MMP-13 have been reported in large animal and human myocardial specimens (39, 348, 503). MMP-8 has been commonly called neutrophil collagenase; however, this MMP type has localized to a number of cell types including fibroblasts and smooth muscle cells (1, 118, 355, 403, 405). Thus, while relatively high levels of specific MMPs may be found in certain tissue types or cells, all of the MMPs shown in Table 1 can be expressed by the major cell types found within the myocardium: cardiac myocytes, fibroblasts, smooth muscle cells, endothelial cells, and macrophages.
In general terms, MMPs are synthesized as inactive zymogens and are secreted into the extracellular space as proenzymes (pro-MMPs). The pro-MMP binds specific ECM proteins and remains enzymatically quiescent until the propeptide domain is cleaved. Although each member of the MMP family is the product of a different gene, analysis of their protein structure has revealed four well-conserved modular structures (157, 293, 505). The signal peptide and propeptide sequences constitute the NH2-terminal domain. The catalytic domain contains the zinc (Zn2+) binding region and is responsible for proteolytic activity. The hemopexin/vitronectin domain is found in all MMPs except MMP-7 and confers substrate specificity (472). For MMP activation to occur, a sequence of proteolytic events must take place. In the latent state, the MMP catalytic domain is concealed by the propeptide mediated by a cysteine-Zn2+ interaction (66, 281, 322, 501). 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. This biochemical interaction of concealing this catalytic site with subsequent activation is known as the cysteine switch (66, 322). Additional autolytic or exogenous cleavages often occur following the initial cleavage of the propeptide domain, resulting in lower molecular weight active forms of MMP. Instead of a sporadic distribution of pro-MMPs throughout the ECM, there is a specific allotment of these proteolytic enzymes within the extracellular space. Moreover, a large reservoir of recruitable MMPs exists, which upon activation can result in a rapid surge of ECM proteolytic activity.
While the MMPs were initially considered to be secreted in a soluble, proenzyme form, a unique subfamily of MMPs initially described by Sato and colleagues in 1994 were discovered to be membrane bound (378). These membrane-type MMPs (MT-MMPs) of which MT1-MMP is the most well characterized, are a fully active enzyme once inserted into the cell membrane and contains an extracellular catalytic domain, a transmembrane domain, and an intracellular domain, all of which are critical for full functioning of the protease (224, 317, 353, 378). It has been demonstrated that trans-Golgi processing and intracellular activational steps, such as those by the intracellular protease furin, play a critical role in processing the mature, full-length MT1-MMP (148, 327, 378). Thus, unlike the secretable MMPs, MT-MMPs are inserted into the cell membrane already activated and therefore can serve as a localized, pericellular site of ECM proteolytic activity.
B. Myocardial MMP Substrates
Several comprehensive reviews using artificial substrates and computer-based modeling have been completed and emphasize the diversity of MMP substrates (170, 321, 413, 452). For the purposes of this review, those substrates that hold potential relevance to the myocardium will be identified. The interstitial collagenase (MMP-1), neutrophil collagenase (MMP-8), and collagenase-3 (MMP-13) possess high substrate specificity for fibrillar collagens, as well as other ECM proteins such as aggrecan, perlican, versican, and proteoglycans. The gelatinases (MMP-2 and MMP-9) demonstrate substrate affinity for denatured fibrillar collagen, basement membrane proteins such as collagen type IV, fibronectin, and laminin. MMP-2 and MMP-9 also exhibit proteolytic activity against elastin and proteoglycans. The substrate portfolio for stromelysin (MMP-3) includes all basement membrane proteins, elastin, and proteoglycans. MMP-7 appears to have a wide substrate portfolio that may be due to the fact that this MMP type lacks the hemopexin domain and therefore may lack critical substrate recognition/docking sites (472). MMP-7 possesses proteolytic activity against the fibrillar collagens I and III, basement membrane proteins (collagen IV, fibronectin), and proteoglycans. One of the more proteolytically diverse MMP types is the MT-MMPs, where for example MT1-MMP can degrade all fibrillar collagens, all basement membrane components, ECM proteins such as perlican and versican, and the chondroitin sulfates.
Another important proteolytic feature of MMPs is that other pro-MMPs also serve as substrates. This results in the ability of one active MMP type to induce proteolytic activation of other MMPs (507). One of the first to be recognized for this particular function was stromelysin, MMP-3 (431, 506). MMP-3 participates in the activational cascade by forming intermediary complexes with other pro-MMPs, ultimately yielding a fully active MMP. For example, Murphy and colleagues reported a 12-fold increase in the conversion of pro-MMP-1 to active MMP-1 in the presence of MMP-3 (290). An emerging body of evidence suggests that a fundamental mechanism for MMP-2 activation is through proteolytic cleavage by MT1-MMP (207, 290, 297, 419). In this activational pathway, pro-MMP-2 is "presented" to MT1-MMP, and the prodomain is cleaved yielding a fully active MMP-2. In addition, it is likely that MT1-MMP processes other pro-MMPs, such as pro-MMP-13 (431). Thus, while further research is required to identify the extent and conditions under which these localized MMP activational cascades would occur, this process would provide a means for a significant and localized amplification of ECM proteolytic activity.
There is emerging evidence to suggest that MMPs can also degrade nonmatrix substrates such as cytokines, bioactivate peptides, and growth factors, which in turn would affect a number of biological processes within the myocardium (125, 170, 210, 213, 222, 280, 306, 309, 318, 413, 435, 463, 500, 523). For example, MT1-MMP and MMP-7 have both been identified to process membrane-bound TNF- to a soluble form (125, 213). Angiostatin, a small bioactive peptide which can modulate the effects of vascular endothelial growth factor (VEGF), can be formed by MMP processing (500). In more recent studies, it has been demonstrated that MMP-3, MMP-7, MMP-9, and the MT-MMPs can proteolytically process VEGF and therefore modify the local biological activity of this growth factor (210, 302, 500). Other growth factors, such as the release of active insulin-like growth factor, can be mediated by several MMP types (256, 276, 299, 508, 541). Thus it is likely that a number of important nonmatrix MMP substrates exist which hold relevance to the LV remodeling process.
C. Transcriptional Regulation of MMPs
Transcriptional activity is an important event in the ultimate synthesis and release of soluble MMPs as well as the MT-MMPs within the myocardium. Following either biological and/or mechanical stimuli, a cascade of intracellular events culminates in the formation of a number of transcription factors. These transcription factors can bind to the promoter region of MMP genes and induce transcription. A generalized schematic for transcription factor binding sites within representative MMP promoter regions are shown in Figure 5, which was adapted from several generalized reviews on this subject (20, 59, 110). Two major cis-acting elements are found in a majority of the MMP promoters: activator protein-1 (AP-1) and polyoma enhancer A binding protein-3 (PEA-3) which interact with the Fos and Jun family and Ets family of transcription factors, respectively. MMP-2 is interesting in that it lacks both the AP-1 and PEA-3 elements, and also lacks a TATA box (20). Due to the fact that MMP-2 lacks these canonical transcription factor binding sites, it was originally assumed that MMP-2 was constitutively expressed and under limited transcriptional regulation. While indeed MMP-2 has been demonstrated to be constitutively expressed in tissues at substantial levels, there is evidence that external stimuli can influence an additional increase in MMP-2 production (215, 261). While AP-1 and PEA-3 are found in most MMP types, there are other elements found in individual MMP promoter regions. For example, MMP-1, MMP-7, MMP-13, and MT1-MMP all have one or more TGF inhibitory elements (TIEs) which bind the family of SMAD transcription factors (84, 110, 150, 234, 245, 462). Other transcription factors such as SP1 likely bind to the GC box contained within the promoter regions of MMP-2, MMP-9, and MT1-MMP (20, 245). In addition, MMP-9 has a nuclear factor-
B (NF-B) binding site (110, 308). MMP-1 and MMP-13 both have been described to have an NF-B-like binding site which responds to the NF-B transcription factor (470). Thus the transcription factor binding sites, the number of these binding sites, and the type of upstream stimuli that are operative under certain cardiac disease states will ultimately determine the transcriptional activity of select MMP types. It is this level of transcriptional regulation that likely adds regional specificity to MMP induction within the myocardium. Specific examples of how specific signaling cascades, known to be operative within the myocardium, can influence MMP transcriptional activity are discussed in the following paragraphs.
View larger version (79K):
[in this window]
[in a new window]
|
FIG. 5. Transcription factor binding elements located within the proximal promoters of representative matrix metalloproteinase (MMP) types. C/EBP-, CCAAT enhancer-binding protein; TIE, transforming growth factor- inhibitory element; AP-1, activator protein-1; PEA-3, polyoma enhancer A binding protein-3; OSE-2, osteoblastic cis-acting element; TATA, TATA box; AP-2, activator protein-2; Sil, silencer; GC, Sp-1 binding site; NF-B, NF-B binding site; CCAAT, CCAAT box; SPRE, stromelysin-1 platelet-derived growth factor responsive element.
|
|
1. Modification of MMP transcription by prototypical stimuli
A) BIOACTIVE MOLECULES.
ANG II is an important mediator in the myocardial remodeling process associated with MI, LVH, or DCM. ANG II receptor activation ultimately causes the induction of the Janus kinase-signal transducers and activators of transcription (JAK-STAT) pathways. Transcription factors activated by this pathway include the AP-1 family of transcription factors, STATs and NF-B. Thus it would be anticipated that ANG II receptor activation would potentially cause activation of MMPs. Indeed, exposure of ANG II in cell culture systems causes a concordant induction of transcription factors that would bind to MMP promoter regions and increased MMP levels (22, 45, 180, 471, 480). For example, Wang et al. (480) reported that stretch-induced induction of MMP-2 and MT1-MMP was abrogated by coincubation with an ANG II receptor antagonist. In an early study by our laboratory, it was reported that exposure of isolated cardiac myocytes to a fixed concentration of ANG II caused a robust increase in MMP-2 levels as assessed by gel zymography (79). It should be recognized that zymography entails the use of detergents during the extraction phase and during electrophoresis. This results in unfolding of the MMP structure and thereby exposes the catalytic domain of the proenzyme. As a consequence, both the proform and the active form of MMPs can be visualized using this substrate zymographic approach. In the next set of studies, ANG II stimulation of neonatal rat ventricular myocytes triggered the mobilization of cytoplasmic NF-B to the nucleus, which in turn increased MMP-9 transcription (371). STAT proteins have been demonstrated to participate in the ANG II-mediated transduction pathway as well. However, there is a considerable amount of cross-talk and feedback between receptor pathways, and the initial increase in MMP levels that can be induced by ANG II may not be sustained or, in fact, may actually be suppressed by prolonged exposure. For example, stimulation of rat cardiac fibroblasts with ANG II induced the production of both NF-B and AP-1 transcription factors, and this was associated with an increase in collagen type-1 production as well as a decrease in MMP-1 expression (63). A study by Chen et al. (68) demonstrated that STAT1- and STAT3-mediated ANG II induced factors that would inhibit MMP activity in human proximal tubular epithelial cells, suggesting a role in the fibrotic process. Indeed, it is likely that ANG II receptor activation induces TGF, which in turn causes multiple effects on MMP transcription, which is discussed in greater detail in a subsequent paragraph. Very often the production of ANG II is paralleled by increased levels of aldosterone. While it has been identified that myocardial overexpression of aldosterone induces production of fibrotic proteins such as collagen (87, 342), whether and to what degree aldosterone influences MMP transcription directly remains unknown. There is no direct evidence to date to suggest that a MMP promoter region contains a hormone response element (HRE) (389). Although it has been demonstrated that other steroids can modulate MMP expression, the effects of these steroids on the transcriptional regulation are likely due to indirect mechanisms of action (94, 223). Nevertheless, since decades of research have clearly demonstrated that ANG II and aldosterone contribute to adverse myocardial remodeling, most notably myocardial fibrosis, to what degree and by what mechanisms these signaling molecules affect MMP transcription warrant further study.
ET is another bioactive molecule that likely contributes to the adverse myocardial remodeling process. One intracellular event following ET receptor activation is to activate members of the protein kinase C (PKC) family. These activated PKC isotypes ultimately lead to the activation of transcription factors such as c-Jun, GATA-4, a member of the Ets family, and NF-B, which in turn could lead to MMP transcriptional activation (295, 296, 424). For example, ET increased MMP-1 levels in vascular endothelial cell cultures and MT1-MMP levels in cardiac myocyte preparations (79, 296). In a study by Tsurdua et al. (449), it was demonstrated that ET exposure could amplify the release of MMPs in a myocardial fibroblast preparation. Furthermore, Podesser et al. (338) demonstrated that ET receptor inhibition attenuated the degree of MMP levels in an intact mouse model of MI.
Increased synthesis and release of the catecholamine norepinephrine (NE) has been considered a biological hallmark of heart failure progression (112, 198). Thus increased levels of a number of signaling molecules such as ANG II and ET in the context of myocardial remodeling and heart failure are likely due, at least in part, to systemic neurohormonal activation and the elaboration of NE. NE infusion in rats resulted in an increased MMP-2 mRNA 3 days post infusion (42). In addition to the indirect effects of NE on MMP transcription, it is likely that NE can directly influence MMP transcription through activation of the - and -receptor pathways (79, 393). Thus whether the effects of catecholamines on MMP transcription are directly related to the actions of NE, or whether other bioactive molecules are involved remains to be determined.
B) CYTOKINES.
While a number of cytokines/chemokines have been identified in the myocardial remodeling process, for the purpose of this review TNF and interleukin (IL)-1 will be presented as prototypical examples. While the TNF signaling cascade is divergent and can evoke multiple pathways, activation of the mitogen-activated protein kinases (JNK and p38) and formation of the transcription complexes of the AP-1 family and NF-B are common events (16). Thus it follows that TNF would modify MMP transcription and ultimately myocardial MMP levels. Cardiac-specific overexpression of TNF in transgenic mice caused progressive myocardial remodeling and was temporally associated with MMP induction (399). In another transgenic construct, cardiac restricted overexpression of TNF and disruption of NF-B signaling directly altered MMP mRNA levels (197). Specifically, with TNF overexpression only, MMP-9 increased, whereas with the TNF overexpression and disrupted NF-B, MMP-9 levels were significantly attenuated (197). Other in vitro studies demonstrated that TNF induced MMP-13 and that this induction was reduced by both AP-1 and NF-B inhibitors (232). The proinflammatory cytokine IL-1, through its cognate receptor, can cause activation of the NF-B pathway as well as the AP-1 family of transcription factors (246, 257). For example, IL-1 has been demonstrated to increase both NF-B and AP-1 DNA binding in neonatal rat myocytes (246). Using myocardial fibroblast cultures, Siwik and colleagues (246, 400) provided conclusive measurements that TNF as well as IL-1 could stimulate the de novo production of certain MMP types. In these studies, relative levels of MMP-2 and putatively MMP-9 were increased in the conditioned fibroblast media with exposure to TNF or IL-1 when examined by gelatin zymography (Fig. 6). In other studies, stimulation of adult rat cardiac fibroblasts with IL-1 induced an upregulation of both MMP-2 and MMP-9 protein levels, which was reduced with NF-B inhibition (511). Additionally, MMP-1 levels increased following IL-1 stimulation and were effectively attenuated with mitogen-activated protein kinase inhibitors (200). These studies provide a clear cause-effect relationship between cytokine stimulation and MMP transcriptional activity. Moreover, it is likely that IL-1 and TNF cause induction of specific MMP types through both NF-B-dependent and -independent pathways.
View larger version (28K):
[in this window]
[in a new window]
|
FIG. 6. Effects of inflammatory cytokines on MMP release in conditioned media of cultured myocardial fibroblasts. MMP levels were assessed by zymography, and therefore, the identity of MMP types was estimated from molecular weights shown on the left. Zymography entails treatment of the myocardial extracts with detergent that will unfold the MMP structure causing exposure of the catalytic domain. Thus a proteolytic band corresponding to both the proform and active form of MMPs will be visualized by this technique. Annotations of likely MMP types revealed by zymography are indicated on the right. The high molecular weights likely reflect dimerization and binding of MMPs, whereas molecular weights around the 90-kDa region likely reflect MMP-9, those around the 70-kDa region reflect MMP-2, and those at lower molecular weights likely reflect MMP-13. Fibroblasts were exposed to interleukin (IL)-1 (4 ng/ml), tumor necrosis factor (TNF) (100 ng/ml), IL-6 (10 ng/ml), or interferon- (500 U/ml). The cytokines IL-1, TNF, and IL-6 caused a significant increase in MMP levels that was likely due to increased formation of transcription factors which bind to MMP promoters and in turn increased MMP transcription. [From Siwik et al. (400).]
|
|
C) MATRICELLULAR FACTORS.
Certain signaling molecules that have a significant effect on the ECM have been termed matricellular factors and include osteopontin (OPN) and thrombospondin (TSP). These factors can function as both a cell attachment molecule as well as a cytokine. While OPN is highly expressed during embryonic development, OPN expression in the adult myocardium is quite low. However, following a pathophysiological stimulus such as MI or POH, OPN levels increase rapidly and robustly (80, 383). Past studies have demonstrated that in OPN-deficient mice, the hypertrophic response and matrix accumulation were attenuated in models of POH or ANG II infusion (236, 262, 512). Binding of OPN to transmembrane receptor complexes results in the production of the NF-B transcription factor (85, 346).
Past studies have demonstrated that OPN can induce MMP-2 and MMP-9 activation through an NF-B-dependent mechanism (336, 345, 513). Conversely, it has been shown that OPN can reduce IL-1-mediated induction of MMP-1 in adult rat cardiac fibroblasts (513). TSPs are a family of secreted glycoproteins that appear to modulate cell-matrix interactions through the coalescence of membrane proteins and signaling molecules at specific contact points on the cell surface. TSP-2 and TSP-4 have been shown to be upregulated in the transition from myocardial hypertrophy to failure in renin overexpressing rats and spontaneously hypertensive rats, respectively (374, 388). In addition, TSP-2 null mice show abnormalities in matrix structure and composition (521). TSP signaling likely involves mitogen-activated protein kinase and ultimately the formation of AP-1 transcriptional complexes (48, 97, 137). Although it is not yet known what the direct effects of TSP are on MMP transcription, fibroblasts isolated from TSP-2 null mice demonstrated abnormal ECM adhesion and increased levels of MMP-2 (521). In addition, TSP fragments stimulated differential production of MMP-2 and MMP-9 in bovine endothelial cells (97). In light of the fact that these matricellular signaling proteins are produced within the adult myocardium following a period of cardiovascular stress, it is likely that these play a role in MMP transcription as well as contribute to the overall ECM remodeling process.
D) GROWTH FACTORS.
The TGF superfamily consists of a large number of growth-signaling proteins, but for the purpose of this review, TGF will refer to the most abundant and well-studied member TGF-1. Heightened TGF signaling has been implicated to contribute to myocardial and ECM growth following MI and with pressure overload hypertrophy (POH) (364). TGF can bind to a number of receptor complexes, and through a series of translocations, ultimately results in phosphorylation of a unique set of intracellular signaling proteins, the SMADs. Through a series of SMAD interactions, eventual binding to the TIE promoter region occurs, and given that MMPs-1 and -7 as well as MT1-MMP contain TIE binding domains, it is suggestive that these MMP types would be altered with increased levels of TGF. Since TGF is classically considered a "profibrotic" molecule, it was assumed that TGF would inhibit MMP transcription, and thereby reduce ECM turnover. Indeed, a number of studies demonstrated that TGF could reduce MMP transcription (150, 152, 233, 462, 532). A study by Hall et al. (150) demonstrated that the proximal AP-1 site is essential for the TGF-mediated repression of MMP-1 in cultured fibroblasts. In another study using dermal fibroblasts, it was demonstrated that MMP-1 transcriptional suppression was mediated through the SMAD3 and SMAD4 activation (532). It has also been demonstrated that SMADs can directly interact with members of the AP-1 family of transcription factors (233, 516). Therefore, while TGF may inhibit MMP transcription through the SMAD pathway, simultaneous activation of transcription factors that will induce certain MMP types can occur. For example, TGF can stimulate MMP-13 transcriptional activity through the activation of Fos and Jun transcription factors (462). Moreover, other studies have demonstrated that TGF can activate other transcriptional complexes, such as the Ets family of transcription factors (84), which would also induce certain MMP types. Thus it is likely that TGF causes a duality of responses with respect to MMP transcriptional activity which are likely to be concentration, cell, and time dependent.
2. Additional biological pathways to MMP transcription
A novel transmembrane protein has been identified that induces the expression of specific MMPs in vitro (29, 56, 149, 289, 339, 446, 518). The nomenclature of this protein has varied somewhat (Basigen, collagenase stimulatory factor, CD147) but has most commonly been identified as the extracellular matrix metalloproteinase inducer protein (EMMPRIN) (29, 56). The most active area of EMMPRIN investigation has been with respect to tumor growth, where increased EMMPRIN expression has been localized to the area with intense remodeling activity or highly invasive tumors (29, 56, 289, 339, 446, 518). In this pathological tissue remodeling process, it is likely that EMMPRIN is a contributory factor in the stimulation of MMPs required for tumor invasion and metastasis. Our laboratory identified that robust levels of EMMPRIN exist within the human myocardium and increase significantly with end-stage DCM (406). Moreover, increased EMMPRIN levels have been associated with ECM remodeling and atherosclerotic plaque progression and in monocytes in patients post-MI (386, 530). Furthermore, EMMPRIN caused an induction of MMP types in vascular smooth muscle cell cultures (386). Therefore, EMMPRIN may be an important factor in the transcriptional regulation of MMPs in the context of myocardial remodeling.
Cardiovascular disease states such as ischemia-reperfusion and MI are often associated with oxidative stress and the formation of reactive oxygen species (ROS), such as superoxide and hydrogen peroxide. Increased ROS production can have multiple effects, which include direct modification of proteins-posttranslational modification, directly inducing intracellular factors that can modulate MMP transcription, as well as the induction of signaling molecules such as cytokines which in turn will alter MMP transcription. Thus oxidative stress and the formation of ROS cause the activation of MMPs through two primary mechanisms: posttranslational modification of pro-MMPs to an active form which will be discussed in a subsequent section, and the formation of transcriptional complexes which will yield newly formed MMP species (121, 303, 387, 401, 402, 478, 483). Studies in several laboratories have demonstrated a direct relationship between the formation of ROS and the induction of MMPs in the context of ischemia/reperfusion and cardiac surgery (181, 235, 268, 269). Moreover, in vitro studies have demonstrated that direct exposure of ROS will induce MMP release (19, 72, 536). For example, a brief incubation of hydrogen peroxide in cardiac fibroblasts increased MMP levels (402). In other studies, exposure of endothelial cells to hydrogen peroxide caused the induction of MMP-2 and MT1-MMP (19). Increased nitric oxide levels, which can also induce reactive species, have been shown to increase MMP-13 mRNA levels in a vascular endothelial cell preparation (536). Since oxidative stress does not occur in isolation, but rather in the context of a number of signaling events, it would follow that heightened ROS or reactive nitrogen levels could potentially synergize MMP transcription. However, future studies will be necessary to dissect out and differentiate the effects of oxidative stress in regard to MMP transcriptional and posttranslational effects.
3. Mechanical stimuli and MMP transcription
Perhaps the most underappreciated and understudied mechanism for the regulation of MMP transcription is that of mechanical stimuli. Since the ECM is integrated to the intracellular compartment through a series of transmembrane proteins, the integrins, it is likely that changes in mechanical stress placed on the myocardium would alter intracellular signaling pathways involved in MMP transcriptional activity (51, 249, 356, 366, 373, 396, 448). Increased transmural pressure increased relative MMP-2 and MMP-9 levels using an ex vivo arterial preparation (70). In a cell culture system of dynamic tensile strain, increased mRNA levels for several classes of MMPs were documented, which appeared to be time dependent (92). In endothelial cell cultures, cyclic strain has been shown to induce transcription factors that would bind to the MT1-MMP promoter and in turn cause increased MT1-MMP mRNA and protein levels (515, 533). In other studies, it has been demonstrated that cyclic strain can cause an induction of MMP-2 in both endothelial and vascular smooth muscle preparations (305, 475). However, the relative degree of MMP induction is likely to be time and stimulus-type dependent. For example, a small uniform biaxial strain reduced MMP-1 expression in human vascular smooth muscle cells (519). Whereas in another study, MMP-1 levels have been reported to be increased with cyclic stretching in human vascular smooth muscle cells (107). In other studies, relative MMP-2 levels were demonstrated to be altered as a function of the magnitude of cyclic strain and the duration of the strain (475, 519). In a study by Garcia and colleagues (332), it was demonstrated that local alterations in myocardial strain in vivo could alter local MMP levels. Specifically, in an intact canine preparation, local alterations in lengthening patterns by high rate pacing (i.e., changing the strain patterns) was associated with a robust increase in relative MMP-9 levels (Fig. 7). However, in this study, local oxidative stress and inflammatory markers were also increased within the region with abnormal strain patterns, and therefore, whether and to what degree the increased MMP-9 is from local synthesis or margination of inflammatory cells with a preformed reservoir of MMP-9 remains to be determined. Nevertheless, this in vivo study demonstrated an important relationship between changes in local myocardial deformation patterns to MMP induction.
View larger version (35K):
[in this window]
[in a new window]
|
FIG. 7. In a canine model of rapid pacing of the LV (LVP) in which the stimulation was confined to a focal region of the myocardium, increased total amounts of MMP-9 were observed as evidenced by proteolytic bands appearing by zymography (92/86 kDa). A: the proteolytic bands for MMP-9 were only increased in the LVP site, whereas there was no change in the LV remote (LVR) or right ventricle (RV). B: quantification of the MMP-9 proteolytic bands revealed that focal LVP increased both the latent and active forms. Thus a stimulus which caused a heterogeneous myocardial deformation caused an induction of MMPs, likely due to local changes in stress/strain patterns. [From Garcia et al. (123).]
|
|
4. Integration of signals and MMP transcription
The intent of this section was to provide a brief review on prototypical signals that would potentially cause transcriptional induction of MMPs. However, it must be recognized that all of these signals are not operative in a consistent fashion and are likely differentially induced depending on the cardiac disease state. Thus significant generalizations have been taken into how these biological stimuli can alter MMP transcription. Furthermore, it must be recognized that these pathways do not work in isolation but rather in a dynamic state within the myocardial interstitium. For example, mechanical stimuli can give rise to bioactive molecules and cause oxidative stress, which in turn cause MMP induction. For example, mechanical stretch increased MMP-2 mRNA levels, which was dependent, at least in part, on a ROS signaling cascade. Thus it is likely that simultaneous signals give rise to a multitude of transduction pathways and transcription factors that are integrated and eventually drive MMP transcription (Fig. 8) (142). Whether and to what degree one, several, or all of these pathways play a dominant role in MMP induction is likely dependent on the specific disease state such as MI, POH, or DCM. Indeed, as discussed in the sections which address each of these myocardial remodeling processes, a unique portfolio of MMPs is expressed within the myocardium that is likely due in large part to a specific set of operative biological/mechanical signals and ultimately differential MMP transcriptional processing.
View larger version (87K):
[in this window]
[in a new window]
|
FIG. 8. An important regulatory point of myocardial MMP abundance is transcriptional activity. A number of transcriptional binding elements such as AP-1, Ets, Sp1, and others exist on the promoter region of MMPs (Fig. 5). The formation of transcription factors in myocardial cells includes extracellular stimuli by bioactive molecules and cytokines and by mechanical signals. However, there are also factors on the promoter region that can potentially repress MMP transcriptional activity, and these include growth factors and subsequent activation and binding of SMAD proteins. [Adapted from Thomas et al. (444).]
|
|
D. Translational/Posttranslational Modification of MMPs
The most important step regarding posttranslational modification is through the proteolytic activation of the proform of MMP to the active form,
Top
Next
