Extracellular protease activation and unraveling of the myocardial interstitium: critical steps toward clinical applications

SPECIAL MEDICAL EDITORIALSExtracellular protease activation and unraveling of the myocardial interstitium: critical steps toward clinical applicationsR. English Chapman, and Francis G. SpinaleR. English ChapmanCardiac Surgery, Medical University of South Carolina, Charleston, South Carolina 29403, and Francis G. SpinaleCardiac Surgery, Medical University of South Carolina, Charleston, South Carolina 29403Published Online:01 Jan 2004https://doi.org/10.1152/ajpheart.00609.2003MoreSectionsPDF (275 KB)Download PDF ToolsExport citationAdd to favoritesGet permissionsTrack citations LV FAILURE AND REMODELINGCongestive heart failure is a complex pathophysiological process in which a significant biological underpinning is left ventricular (LV) failure. Part of the natural advancement to failure is myocardial remodeling, most commonly from adaptation to a pathological stimulus. LV remodeling can be defined as molecular, cellular, and interstitial changes within the myocardium that result in alterations in LV geometry and function (2, 4, 11, 21, 63, 71, 78). In patients with intensely progressive LV remodeling, morbidity and mortality are increased (17, 57, 79). This editorial is focused on the relationship between extracellular proteases and myocardial matrix remodeling. It will also explore basic research targets and analyze critical steps necessary in developing basic research into therapeutic and prognostic realizations. There are a number of underlying factors that contribute to the development and progression of myocardial LV remodeling. Specifically, alterations in myocardial tissue structure occur after myocardial infarction (MI), with the development of cardiomyopathies, and in myocardial hypertrophy (2, 4, 11, 21, 63, 71, 78). Although the initiating stimulus and symptoms may appear comparable among the three etiologies, the cellular remodeling processes are distinctly different and the underlying pathologies diverse. One area of research focus is how the myocardial extracellular matrix (ECM) may contribute to this remodeling process. Of particular note is the activation of extracellular proteases, which cause LV remodeling.The purpose of this editorial is to place into context the different patterns of myocardial remodeling to that of extracellular protease activity. It is now recognized that the myocardial remodeling process is not a uniform sequence of events but rather a diverse set of molecular and cellular triggers, which are likely to be distinctive to underlying pathological stimuli (16). This editorial attempts to place into context the clinically encountered etiologies of myocardial remodeling (infarction, cardiomyopathy, and hypertrophy) to a unique pattern of extracellular protease activation. However, it must be recognized that summating the complex process of myocardial remodeling under these categories fails to address the different magnitude and duration of the pathological stimuli. In an attempt to address the complex issue, this editorial will examine both clinical and relevant animal model data in each of the distinct forms of myocardial remodeling with respect to extracellular protease activation.MATRIX METALLOPROTEINASESThe matrix metalloproteinases (MMPs) are a family of zinc-dependent proteases that are essential for normal tissue remodeling in processes such as bone growth, wound healing, and reproduction (48, 52, 76, 83). MMPs are responsible for turnover of the ECM, which in turn facilitates tissue remodeling. MMPs are synthesized as inactive zymogens and are secreted into the extracellular space in a proenzyme form (48, 52, 76, 83). The pro-MMP binds specific ECM proteins and remains enzymatically quiescent until the propeptide domain is cleaved. Cleavage results in exposure of the zinc active site in the catalytic domain and subsequent activation (49). The disruption of the cysteine-zinc interaction in the MMP active site is essential in the activation of MMPs. This "cysteine switch" hypothesis of MMP activation is likely the mechanism by which most MMPs are activated (49). 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.MMP ClassificationMMPs are classified into subgroups according to substrate specificity and/or structure (48, 52, 76, 83). MMPs that have been identified in the myocardium are listed in Table 1. Interstitial collagenase (MMP-1) and collagenase-3 (MMP-13) possess high substrate specificity for fibrillar collagens, whereas the gelatinases (MMP-2 and MMP-9) demonstrate high substrate affinity for basement membrane proteins (48, 52, 76, 83). The substrate portfolio for stromelysin (MMP-3) includes important myocardial ECM proteins such as aggrecan, fibronectin, and fibrillar collagens (48, 52, 76, 83). Furthermore, past in vitro studies have demonstrated MMP-3 can proteolytically process pro-MMP species (47, 49). For example, Murphy and colleagues (47) reported a 12-fold increase in the conversion of pro-MMP-1 to active MMP-1 in the presence of MMP-3. In addition, other MMPs such as MMP-1, MMP-2, and the membrane-type MMPs (MT-MMPs) can also activate pro-MMPs (24, 47–49, 52, 53, 76, 83). MT-MMPs contain a transmembrane domain and are likely activated intracellularly through a proprotein convertase pathway (24, 47–49, 52, 53, 76, 83). Thus, unlike the secretable MMPs, MT-MMPs are activated once positioned in the cell membrane. It has been demonstrated that MT1-MMP degrades fibrillar collagens and a wide range of ECM components as well as proteolytically processes pro-MMP-2 and pro-MMP-13 (24, 47–49, 52, 53, 76, 83). There is emerging evidence that MMPs also degrade nonmatrix substrates, in turn affecting cell proliferation, migration, and apoptosis (43). Table 1. MMPs identified in human myocardium Class Number Substrate References Collagenases Interstitial collagenase MMP-1 Collagen I, II, III, VII, gelatin, proteoglycan, glycoprotein 16,48,52,67,76,81,83 Collagenase 2* MMP-8* Collagen I, II, III, aggrecan 16,48,52,83 Collagenase 3 MMP-13 Collagen I, II, III, gelatin, proteoglycan, pro-MMP-1 16,48,52,67,76,81,83 Gelatinases Gelatinase A MMP-2 Gelatins, collagen I, IV, V, VII, elastin, fibronectin, laminin, proteoglycan 16,36,48,52,67,76,81,83 Gelatinase B MMP-9 Gelatin, collagen IV, V, XIV, elastin, proteoglycan, glycoprotein 16,36,48,52,67,81,83 Stromelysins Stromelysin-1 MMP-3 Fibronectin, laminin, collagen III, IV, IX, pro-MMP-1,-7,-9 16,48,49,52,67,76,81,83 Matrilysin Matrilysin* MMP-7* Fibronectin, laminin, elastin, gelatin, collagen I, IV 16,48,52,83 Membrane-type MMPs MT1-MMP MMP-14 Collagen I, II, III, fibronectin, laminin-1, glycoprotein, proteoglycan, pro-MMP-2, -13 14,16,24,48,52,76,81,83MMP, matrix metalloproteinase. *May exist in the myocardium in states of inflammation and wound healing MMP Activation/InhibitionThere are presently over 20 members of the MMP family of proteolytic enzymes, and several MMPs are substrate specific for other pro-MMPs. Therefore, activation of a few select enzymes can launch a cascade of proteolytic activity. For example, pro-MMPs are activated not only by MMP-1, MMP-2, MMP-3, and MT-MMP, but also by serine proteases chymotrypsin, trypsin, and plasmin (48, 49, 52). Thus the regulation of these upstream processes and proteolytic cascades may directly affect MMP activity. Because MMPs degrade various components of the ECM, it is important that MMPs be tightly controlled to prevent excessive tissue degradation. Therefore, a group of endogenous proteins, the tissue inhibitors of matrix metalloproteinases (TIMPs), exists to regulate activity of the MMPs. Currently, there are four known TIMP species (7, 19, 22). TIMPs bind to the active site of the MMPs, thereby blocking access to extracellular matrix substrates. The MMP-TIMP complex is formed in a stoichiometric 1:1 molar ratio and forms an important endogenous system for regulating MMP activity in vivo (7, 19, 22). The importance of inhibitory control of MMPs within the myocardium has been demonstrated through the genetic deletion of TIMP-1 expression (62). In TIMP-1 knockout mice, LV remodeling occurred in the myocardial interstitium without a pathological stimulus (62). Although the TIMPs are expressed in an array of cells, TIMP-4 shows a high level of expression in human myocardial tissue (22). The following section assesses MMP and TIMP profiles in relation to LV remodeling following myocardial infarction in cardiomyopathies and in severe myocardial hypertrophy.MMPS AND LV REMODELINGLV Remodeling and MMPs in Patients Postmyocardial InfarctionLV remodeling postmyocardial infarction (MI) is characterized by nonuniform changes in LV myocardial wall geometry (56, 71, 78). After undergoing severe cellular loss, the nonviable infarct region becomes a thinned, fibrous scar (2, 56, 71, 78). The myocytes in the border zone are susceptible to infarct expansion and undergo hypertrophy, realignment, and slippage (51, 56). Gross LV myocardial geometry changes from an elliptical shape to a spherical shape, and a concomitant decline in LV performance occurs (51, 56, 58). Of importance is that, whereas LV pump function is commonly depressed with post-MI remodeling, contractility in viable isolated myocytes may be preserved (58). These observations are central to the ability of the ECM to translate myocyte shortening into overall ejection. Moreover, these observations emphasize that maladaptive post-MI remodeling is due in part to changes in the ECM rather than in viable myocytes. Accordingly, it is essential to identify the cascade of events that give rise to LV remodeling post-MI.Several studies have demonstrated that increased plasma levels of MMPs occur in patients in the early post-MI period (27, 31). For example, increased peripheral blood levels of MMP-2 and -9 have been reported in patients of acute coronary syndromes, as well as increased MMP-9 and TIMP-1 in the plasma (27, 31). However, in these past observational reports the temporal relationship between MMP and TIMP release into the plasma with respect to the onset of myocardial injury remains unknown. Accordingly, a recent study focused on serial changes in plasma MMP levels following MI induced by alcohol ablation in patients of hypertrophic obstructive cardiomyopathy, and a time-dependent release of MMPs post-MI was observed (6). For example, as seen in Fig. 1, MMP-9 levels increased and remained elevated for up to 50 h following alcohol-induced MI (6). Whereas serial measurements do not reflect direct activity, they may serve as surrogate markers for relative activation of MMPs post-MI. Such studies may be useful in developing prognostic and diagnostic therapies. Fig. 1.Time-dependent release of matrix metalloproteinase (MMPs) postmyocardial infarction (MI) induced by alcohol ablation. A significant increase in plasma MMP-9 levels occurred after alcohol injection and appeared to plateau for up to 50 h after injection. *P < 0.05 vs. time 0; baseline values. Reprinted with permission from Bradham et al. (5).Download figureDownload PowerPoint LV Remodeling and MMPs: Animal Models of MI Demonstrating Cause and EffectA cause-effect relationship between myocardial MMP activation and remodeling has been established through genetic modifications and pharmacological MMP inhibition (18, 55, 61). Altering the expression of certain MMP genes affects tissue remodeling within the MI region as well as influences the degree of post-MI remodeling (18, 55, 61). A recent study by Mukherjee and colleagues (46) utilized MMP inhibition in the setting of post-MI remodeling to evaluate region-specific changes in MMP and TIMP levels. Consistent with previous studies, MMP inhibition attenuated post-MI remodeling. However, as shown in Fig. 2, this study also reports a decrease in the rate of change in regional MI size in the MMP inhibition group, demonstrating that localized MMP activity directly contributes to infarct expansion in later phases of post-MI remodeling (46). Fig. 2.MMP inhibition attenuates post-MI remodeling. Time-dependent change in regional MI size, computed relative to week 2 values, was significantly lower in MI + MMP inhibition (MMPi) group than MI-only group. +P < 0.05 vs. MI only; #P < 0.05 vs. week 2. Reprinted with permission from Mukherjee et al. (46).Download figureDownload PowerPoint A Unique MMP/TIMP Profile After MIA unique temporal profile of MMPs and TIMPs has been reported in the post-MI animal model (46, 55, 82). Wilson et al. described a regional imbalance of MMPs and TIMPs following changes in the remote, transition, and infarct regions post-MI (82). Time-dependent changes in myocardial MMP and TIMP levels after MI induction in rats have also been observed and show that differential profiles of MMPs are released throughout LV remodeling post-MI (55). MMP-13, -2, and -9 were elevated early in the post-MI period, whereas a more gradual elevation of MMP-14 (MT1-MMP) followed 16 wk of post-MI. Furthermore, throughout the time course measured, TIMP-1 and -2 increased while TIMP-4 decreased (55). Taken together, alterations in specific MMP and TIMP levels post-MI must be considered, because these certain species warrant further study for target intervention.A loss of TIMP-mediated control has been reported in LV remodeling following MI (46, 55, 82). In the rat MI model, MMP mRNA levels increased early post-MI but were not associated with a concomitant increase in TIMP mRNA levels (55). In an in vitro system of ischemia and reperfusion, TIMP-1 expression was reduced in the early reperfusion period (4). These findings suggest a loss of endogenous MMP inhibitory control occurs early in the post-MI period (4).Identifying proteolytic systems, which promulgate regional geometry changes, offers greater potential for direct therapeutic targeting. However, developing specific targets will be predicated on recognition of the phases of post-MI remodeling. The early phase of post-MI remodeling occurs immediately after an acute myocardial insult. Localized MMP activity contributes to infarct scar formation, and healing is accompanied by neutrophil infiltration into the infarct area (2, 71, 78). With longer post-MI periods, however, rampant MMP activity in the border region of the infarct scar contributes to adverse remodeling and progressive infarct scar expansion (46). Thus targeting the MMP system post-MI must account for both the critical healing phase and the adverse post-MI remodeling phase. To determine the ideal time for MMP inhibitor deployment in the setting of LV remodeling post-MI, it is necessary to couple upstream cues that instigate adverse remodeling with improved imaging modalities.LV Remodeling and MMPs in Dilated CardiomyopathyRemodeling in dilated cardiomyopathy (DCM) is characterized by an increase in the ratio of LV chamber radius to wall thickness, which increases myocardial wall stress and eventually results in progressive LV chamber dilation (84). The ECM undergoes structural realignment, loss of myocyte connections, degradation of the normal collagen fibrillar weave, and abnormalities in collagen cross-linking (20, 78). Markers of collagen degradation have been shown to be increased in DCM patients compared with age-matched control subjects (64). Thus it is likely that changes in ECM structure and function are largely responsible for LV remodeling in DCM.Whereas the etiologies of DCM are diverse, the general classifications are ischemic, idiopathic (nonischemic), and infectious (84). Myocarditis is an inflammatory cardiomyopathy, most commonly viral or autoimmune in origin (41, 60). The fact that myocarditis often progresses to DCM underscores the potential significance of the inflammatory pathways in ECM remodeling. Furthermore, the inflammatory pathway induces cytokines such as tumor necrosis factor-α (TNF-α), and in DCM there appears to be a unique portfolio of MMPs that emerge and are likely under the control of upstream signals such as TNF-α (81). For example, TNF-α receptor activation occurs in patients with cardiomyopathic disease as evidenced by increased plasma levels of soluble TNF-α receptors (81). Moreover, a relationship has been established between plasma levels of TNF receptors 1 and 2 and plasma MMP-9 levels in patients of cardiomyopathy (Fig. 3) (81). Fig. 3.Plasma levels of MMP-9 and soluble tumor necrosis factor (TNF) receptors 1 and 2 (TNFR1 and TNFR2) were serially examined by ELISA in congestive heart failure (CHF) patients at study entry and up to 6 mo of follow-up (n = 24 patients: age 61 ± 3 yr). Increased MMP-9 levels accompanied an absolute increase in plasma TNFR1 from baseline. An increase in TNFR2 plasma levels from baseline was associated with a relative increase in MMP-9 plasma levels in CHF patients. Reprinted with permission from Wilson et al. (81).Download figureDownload PowerPoint Clinical Studies of MMPs and TIMPs in Dilated CardiomyopathyA number of studies have examined relative MMP and TIMP expression in end-stage human DCM (36, 67, 73, 74). Tyagi and colleagues (74) demonstrated increased in vitro myocardial MMP zymographic activity in DCM samples. Also, it has been observed that myocardial MMP-9 was increased due to either ischemic or nonischemic origins, and MMP-3 was increased in nonischemic DCM myocardial extracts (67, 73). A selective upregulation of certain MMP species has been observed in DCM. For example, the collagenase MMP-13 is expressed at low levels in normal myocardium but is significantly increased in end-stage DCM (Fig. 4) (67, 73). Interstitial collagenase (MMP-1), in contrast, is decreased (67, 73). In addition, because TIMP-1 binds with less affinity to MMP-13 than to MMP-1, there is a loss of endogenous control as MMP-13 levels increase (32). Furthermore, an over threefold increase in MT1-MMP has been shown in cardiomyopathic samples (Fig. 4) (67). Because of its biologically diverse actions, MT1-MMP is not only able to degrade surrounding matrix but also is capable of initiating cascades of proteolytic activity (24, 48, 52, 53, 76, 83). Therefore, the emergence of certain MMP species within the DCM myocardium may contribute to increased susceptibility of the myocardial fibrillar collagen network to degradation and subsequent maladaptive myocardial remodeling. Fig. 4.Myocardial levels of MMP-13 and MT1-MMP were determined by using immunoblotting techniques in normal and dilated cardiomyopathy (DCM) myocardium. MMP-13 was expressed at low levels in normal human myocardium (N) but increased significantly in nonischemic DCM (D) and ischemic cardiomyopathy (I). MT1-MMP was present in normal myocardium (N) and increased by over threefold in cardiomyopathic samples (D, I). MMP-13 and MT1-MMP were localized to positive controls (+). Reprinted with permission from Spinale et al. (67).Download figureDownload PowerPoint Myocardial TIMP levels appear to be variably expressed in end-stage DCM (36, 73). Thomas and colleagues (73) found an increased abundance of TIMP-1 and TIMP-2 in DCM myocardium samples. In a study by Li and colleagues (36), TIMP-1 and TIMP-3 levels were reduced in cardiomyopathic samples, whereas TIMP-2 levels were unchanged compared with control myocardium. Previous studies of cardiomyopathic disease have suggested that changes occur in the stoichiometric ratio of MMPs to TIMPs (36, 67). For example, in both ischemic and nonischemic DCM, an absolute reduction in MMP-1/TIMP-1 complex formation was observed (67). Similar to MMPs, TIMPs are encoded by unique genes and differ in promoter regions (13, 80). Thus whereas further research is necessary, extracellular stimuli may induce differential expression of TIMPs in DCM. Nevertheless, the quantitative reduction of MMP-TIMP complexes within the myocardium in end-stage DCM has significant relevance to the LV remodeling process. As observed post-MI, a change in the MMP-TIMP stoichiometric ratio in DCM could favor persistent MMP activation, resulting in unbridled ECM turnover and progressive LV remodeling.MMPs and TIMPs in an Animal Model of DCMA model of heart failure that has reported a relationship between LV remodeling in DCM and MMP activity is the pacing-induced cardiomyopathy model (12, 68, 69). After the pacing protocol in pigs was instituted, MMP activity measured by an in vitro assay was increased (12, 68, 69). Not only was an increased abundance of myocardial MMP-2 and MMP-3 observed, but MMP activity was also shown to coincide with the onset of LV remodeling and dilation (69). Because LV remodeling preceded significant defects in myocardial contractility, activation of MMPs may be an early in the progression to In the myocarditis model, an upregulation of MMP-3 and MMP-9 mRNA and has been reported with a of TIMP-1 and TIMP-4 mRNA and MMP levels in this model also to increased expression of several cytokines Remodeling and MMPs in Patients with Myocardial cause for symptoms associated with heart failure is to LV hypertrophy In to LV remodeling post-MI and with cardiomyopathy, heart disease is often characterized by an in normal LV Whereas the cause of this is changes in the ECM are in collagen as well as alterations in collagen have been observed in with There is evidence to such observations from of the ECM For example, in in collagen directly affects There are distinct patterns of that occur in to persistent hypertrophy and hypertrophy The patterns of myocardial remodeling differ significantly between these states and are and MMPs in of myocardial fibrillar collagen influences of the and results in 78). studies have reduced MMP plasma levels in patients with and For example, decreased plasma levels of MMP-1 were reported in patients with hypertrophy, and reduced plasma levels of MMP-9 were observed in patients The ratio of TIMP-1 to MMP-1 was also increased in patients of severe Taken together, these studies suggest an decrease in MMPs and increase in TIMPs will favor matrix to patterns of collagen in LV remodeling with is characterized by excessive ECM degradation and LV collagen in states disruption of normal fibrillar collagens, which may be due to proteolytic activity of myocardial MMPs It is that early of MMPs facilitates LV profiles and time-dependent changes of MMPs and TIMPs in patients with Remodeling and MMPs in Animal Models of hypertrophy progresses to and LV time-dependent changes in the activation of myocardial MMPs are likely to In several studies have demonstrated such changes in myocardial MMP levels throughout the development of In the rat (Fig. the development of hypertrophy is associated with an increase in TIMP-4 which a reduction in MMP activity However, over MMP-2 and MMP-9 levels increase and TIMP levels myocardial remodeling and changes in the ECM to and failure et al. reported MMP and TIMP levels in the of compared with the to CHF Whereas MMP-2 and TIMP activity remained unchanged at the MMP-2 activity increased significantly and that of the TIMPs as to LV failure Fig. ventricular (LV) from the level of the and with a shown are from a a heart failure and a rat with the MMP inhibitor from et al. figureDownload PowerPoint It is that both and stimuli LV remodeling in and the are likely time For example, an animal model of acute in increased myocardial MMP-9 expression and zymographic activity With MMP-9 zymographic activity to and was accompanied by changes in TIMP-1 levels Furthermore, it has been demonstrated in an animal model of that the of the stimulus directly MMP and TIMP levels Therefore, the stimulus to LV remodeling in states is likely wall such as or increased myocardial MMP levels and zymographic activity have been observed of have shown not only an increase in MMP-2 and MMP-9 zymographic activity, which was associated with changes in LV function and but also an in and adverse LV remodeling following MMP inhibition with a concomitant of normal LV function Taken together, these studies suggest that changes in myocardial MMP and TIMP levels likely contribute to the LV remodeling process a or different patterns of MMP and TIMP expression occur on the and duration of and 3) increased levels of MMPs and TIMPs may be decreased with pharmacological inhibition or of the AND MMPs studies have demonstrated the of progressive LV dilation and post-MI through the of MMP inhibition (46, 55, The clinical of MMP however, may be Specifically, MMP inhibition has been associated with which that inhibition may with normal tissue turnover Therefore, the specific MMP in the adverse LV remodeling process must be A study by et al. selective MMP which MMP-1, in a model of post-MI ventricular remodeling and reported that progressive LV dilation was Thus a has been established in which selective MMP inhibition is a pathway to LV remodeling post-MI. with must also be for in post-MI remodeling. The early phase of remodeling immediately after MI is essential for infarct scar formation and cell and it is important to the for the post-MI. For example, past clinical studies that with the early acute inflammatory were associated with adverse of this acute critical healing phase from that of the adverse remodeling phase is Thus, in targeting the MMP proteolytic cascades as well as initiating events and temporal profiles must be in MMP MMPs in development and progression of cardiomyopathic disease is a but there are likely underlying processes that may MMPs and have on myocardial remodeling. Specifically, signals such as wall stress may be an important in patients of from a study by Li et al. MMP-2 and MMP-9 levels and after with the LV Although MMP-2 remained MMP-9 was reduced in of patients after (Fig. Thus, with as a stimuli are in MMP levels in patients with The of inflammatory pathways is therapeutic target in cardiomyopathy because elevated levels of inflammatory cytokines have been reported in cardiomyopathy patients Therefore, in cardiomyopathy, both and biological signals are which likely MMP and TIMP Fig. in cardiomyopathy patients a reduction in certain MMPs and is most in patients with significant adverse remodeling. MMP-2 zymographic activity not change but MMP-9 levels in general decreased after LV MMP-9 zymographic activity, expressed as a of set at decreased in of patients duration of LV LV [Reprinted with permission from Li et al. figureDownload PowerPoint MMPs in