Increased Cardiac Expression of Tissue Inhibitor of Metalloproteinase-1 and Tissue Inhibitor of Metalloproteinase-2 Is Related to Cardiac Fibrosis and Dysfunction in the Chronic Pressure-Overloaded Human Heart
Background— Alterations in the balance of matrix metalloproteinases (MMPs) and their specific tissue inhibitors (TIMPs) are involved in left ventricular (LV) remodeling. Whether their expression is related to interstitial fibrosis or LV dysfunction in patients with chronic pressure overload–induced LV hypertrophy, however, is unknown.
Methods and Results— Therefore, cardiac biopsies were taken in 36 patients with isolated aortic stenosis (AS) and in 29 control patients without LV hypertrophy. Microarray analysis revealed significantly increased mRNA expression of collagen types I, III, and IV and transcripts involved in collagen synthesis, including procollagen endopeptidase and lysine and proline hydroxylases, in AS compared with control patients. Collagen deposition was greater in AS than in control patients and was most pronounced in AS patients with severe diastolic dysfunction. Cardiac mRNA expression of TIMP-1 and TIMP-2 was significantly increased in AS compared with control patients (mRNA transcript levels normalized to GAPDH: TIMP-1, 0.67±0.1 in AS versus 0.37±0.08 in control patients; TIMP-2, 9.5±2.6 in AS versus 1.6±0.4 in control patients; P<0.05 for both) but did not differ significantly for MMP-1, -2, or -9. Cardiac TIMP-1 and -2 transcripts were significantly related to the degree of interstitial fibrosis and proportional to diastolic dysfunction in AS patients.
Conclusions— Cardiac expression of TIMP-1 and TIMP-2 is significantly increased in chronic pressure-overloaded human hearts compared with controls and is related to the degree of interstitial fibrosis.
Received October 27, 2004; revision received April 22, 2005; accepted May 16, 2005.
Left ventricular (LV) hypertrophy is an adaptive process that compensates for pressure overload caused by aortic stenosis (AS).1,2 This remodeling process consists of hypertrophic changes of cardiomyocytes and abnormalities of the extracellular network, which are both responsible for changes in systolic and diastolic function.2,3
Altered levels of matrix metalloproteinases (MMPs) and their specific tissue inhibitors (TIMPs) have been involved in remodeling of the extracellular matrix during LV hypertrophy and in the failing heart (reviewed previously4–6). Recently, Hein and colleagues7 showed that the extent of myocyte degeneration in patients with severe AS correlated with the severity of systolic dysfunction, whereas Polyakova et al8 showed enhanced protein expression of TIMPs and MMPs in these patients. These studies included a significant number of patients with moderate to severe systolic dysfunction, and cardiac samples were taken at the subvalvular septum, where myocyte strain and fibrosis are most pronounced. Whether cardiac TIMP or MMP transcript levels in patients with cardiac hypertrophy because of chronic pressure overload correlate with the degree of interstitial fibrosis or with the severity of cardiac dysfunction, however, has not been investigated.
Therefore, in the present study, we first performed a microarray analysis to investigate changes in expression levels of genes involved in collagen synthesis and degradation. Subsequently, we compared cardiac transcript levels of TIMPs/MMPs in patients with AS with those measured in control patients without LV hypertrophy and examined the relation of cardiac TIMP/MMP expression to the degree of cardiac fibrosis and dysfunction.
Thirty-six consecutive patients with isolated AS were included and were submitted to clinical evaluation (Table 1). These patients all underwent surgical aortic valve replacement. Different types of aortic prosthetic valves were implanted (Data Supplement Part I). The duration of aortic cross-clamping was 67±4.3 minutes; that of cardiopulmonary bypass, 97±5.6 minutes; and postoperative length of stay, 18±1.7 days.
Twenty-nine patients undergoing off-pump coronary artery bypass graft surgery (CABG) with normal ejection fraction, diastolic function, and without unstable angina, previous history of myocardial infarction, or LV hypertrophy at echocardiography or ECG measurements served as control subjects. Patients with concomitant renal, hepatic, rheumatic, cancerous, or other severe diseases were excluded. In addition, 6 LV biopsies of nonused donor hearts with normal cardiac function were included.
The institutional ethics committee of the University Hospital of Leuven approved the study, and all patients gave informed consent.
Each patient underwent an M-mode, 2D, and Doppler echocardiographic study (Hewlett-Packard Sonos 5500). All echocardiographic data represent the mean of 3 measurements on different cardiac cycles.
M-mode LV end-diastolic diameter was measured and normalized to body surface area. Diastolic patterns of LV filling were evaluated by Doppler echocardiography at the level of the mitral valve and pulmonary vein and were considered normal, prolonged relaxation, pseudonormal, or restrictive as previously described in detail.9
During open-heart surgery, before patients were placed on the extracorporeal circulation, 2 or 3 transmural true-cut needle biopsies, each weighing 1 to 2 mg, were taken from the anterior LV at close proximity between the left descending coronary artery and the circumflex coronary artery. One biopsy was immediately frozen in liquid nitrogen and stored at −80°C, whereas the second one was mildly fixed in 1% paraformaldehyde and embedded in paraffin. For electron microscopy, samples were embedded in Epon, sectioned, and stained according to a standard protocol.10
RNA Isolation, Real-Time Polymerase Chain Reaction, and Microarray Analysis
RNA was isolated from biopsies of 24 AS patients and 20 control patients with the RNeasy Mini kit (Qiagen).11 RNA quality was measured with a Bioanalyser Nanochip (Agilent Technologies), and RNA quantity with the NanoDrop ND-1000 UV-Vis Spectrophotometer (Nanodrop Technologies).
Cardiac mRNA levels of MMP-1, MMP-9, TIMP-1 and TIMP-2, collagen α1 type I (COLIa1), and collagen α1 type III (COLIIIa1) were determined by use of real-time fluorescence detection. The genes of interest and a housekeeping gene (GAPDH) were amplified with the ABI Prism 7700 Sequence Detection System (Perkin-Elmer). Transcript levels were determined in duplicate, and the results were expressed relative to GAPDH. The primer and probe sequences are listed in the Data Supplement Part II.
For microarray analysis, 50 ng total RNA from 19 AS patients and 7 control patients were amplified in 2 rounds with the 2-cycle cDNA synthesis kit from Affymetrix. First-round cDNA (600 ng) was used in a second round of amplification with biotin labeling. Biotin-labeled cDNA (11 μg) was then hybridized to Affymetrix human U133A GeneChips containing 18 400 genes. Gene transcript levels were determined with Microarray Analysis Suite Software, version 5.0 (MAS 5.0) (Affymetrix).
Primary antibodies against desmin, ubiquitin, MMP-1, MMP-2, MMP-9, TIMP-1 (all R&D Systems) and TIMP-2 (Calbiochem) were used for immunostaining. Relative surface area of MMP/TIMP immunoreactivity was expressed as percentage immunoreactive area divided by total area. The collagen fraction in the endocardial to epicardial regions was determined after collagen-specific picrosirius red staining with a Zeiss Axioplan2 microscope, a 3CCD video camera (DXC-93OP, Sony), and KS300 software as previously described.10
MMP Zymography and Immunoblotting Analysis
For in situ zymography, cryosections of AS and control biopsies were covered with fluorescently labeled pigskin gelatin (Oregon Green 488, Molecular Probes) and incubated for 72 hours at 37°C.12 The gelatinase activity, indicated by the gray discoloration caused by gelatin degradation, was scored on a scale of 0 (no activity); 1, focal activity; 2, moderate; and 3, diffuse activity by a blinded observer. To confirm specificity of gelatinase activity in the degradation assay, consecutive sections were coincubated with EDTA, an inhibitor of MMP activity.
To semiquantitatively investigate protein expression of MMP-2 and MMP-9 in cardiac extracts, immunoblotting of MMP-2 and MMP-9 (antibodies by R&D Systems) was performed on additional pooled biopsies of AS and control patients.11
Data are expressed as mean±SEM. Normal distribution of all continuous variables was tested by use of the method of Kolmogorov and Smirnov. The characteristics data all passed the normality test, and therefore, unpaired t tests were used for comparing patient characteristics in AS with those of control patients. Histological data and cardiac mRNA expression levels were compared by the unpaired t test or the Mann-Whitney test when the standard deviations of 2 groups were significantly different. A standard Pearson correlation test was used when 2 groups followed Gaussian distribution, whereas the nonparametric (Spearman) correlation was selected when groups did not follow Gaussian distribution.
Patients with AS showed typical characteristics of concentric cardiac hypertrophy, including significantly increased LV mass index (LV mass/m2), septal and posterior wall thickness, relative wall thickness, ECG criteria (Table 1), and cross-sectional area of cardiomyocytes (Table 2). Systolic function was similar in AS and control patients. Only 2 patients with AS had marked systolic dysfunction (ejection fraction [EF] <50%). Overt LV dilatation (end-diastolic diameter >58 mm) was present in only 3 AS patients, including the 2 patients with an EF of <50%. All patients with AS except one presented with diastolic dysfunction at echocardiography. Twenty-one patients with AS presented with an abnormal relaxation without signs of increased LV filling pressures, whereas the 16 remaining patients showed a pseudonormal8 or restrictive filling,8 indicating increased LV filling pressures. Systolic or diastolic dysfunction was absent in the control group (exclusion criteria). No significant difference in drug intake, except for β-blockers, was observed between the 2 groups.
Collagen Synthesis, Fibrosis, and Systolic Function
Microarray analysis showed that collagens and transcripts involved in collagen synthesis were significantly upregulated in AS compared with control patients (Data Supplement Table I and Figure 1). Expression of COLIa1 and COLIIIa1 was 1.3- and 2.3-fold upregulated, respectively. Upregulation of COLIa1 and COLIIIa1 mRNA was confirmed by Taqman real-time polymerase chain reaction (PCR) and expressed as fold increase in AS compared with control patients (COLIa1, 2.6-fold, P=0.05; COLIIIa1, 2.8-fold, P=0.02).
Prolyl 4-hydroxylase and lysyl hydroxylase, critical for procollagen synthesis in the endoplasmic reticulum, were upregulated 1.6- and 1.4-fold. Procollagen type III N-endopeptidase and procollagen type I and type II C-protease enhancer, which catalyze the cleavage of procollagens to ready-to-secrete collagens, were upregulated 1.3-, 1.2-, and 2.0-fold, respectively, in AS patients. Furthermore, lysyl oxidase like-2, which cross-links collagens to form triple helices in the extracellular space, was upregulated 1.4-fold. In addition, TIMPs were slightly but significantly upregulated by 1.4-fold (TIMP-1) and 1.2-fold (TIMP-2), whereas there was no differential expression of MMP-1 and -9 and a 1.2-fold increase in MMP-2 expression.
Transmural interstitial fibrosis as visualized by Sirius red staining was pronounced in all biopsies of AS compared with control patients (Table 2; Figure 2, A and B). Collagen deposition in AS patients was most prominent in the subendocardial area (Table 2), where necrotic myocytes were present, suggesting replacement fibrosis. Fibrosis was also pronounced in perivascular areas of AS patients, indicating reactive fibrosis (Figure 2B). To investigate the impact of fibrosis on diastolic dysfunction, we divided the AS patients into 2 groups according to diastolic filling (normal plus abnormal relaxation versus pseudonormal plus restrictive filling, assuming that the second group identifies patients with diastolic dysfunction and increased LV end-diastolic pressure). Patients with pseudonormal or restrictive patterns had a significantly higher degree of interstitial fibrosis compared with patients with normal or abnormal relaxation patterns (% collagen deposition, 17.8±0.3 versus 14.6±0.2, respectively, P<0.01).
Cardiac MMP/TIMP Expression and Fibrosis
To quantitatively assess the differential expression of MMP/TIMPs in AS (n=21) compared with control patients (n=19), transcript levels of MMP-1, MMP-2, MMP-9, TIMP-1, and TIMP-2 were investigated by real-time PCR. Transcript levels of TIMP-1 and TIMP-2 and TIMP-1/MMP-2, TIMP-2/MMP-2, and TIMP-2/MMP-9 ratios were significantly increased in AS compared with control patients (Figure 3, A–C). Importantly, transcript levels of both TIMP-1 and TIMP-2 were significantly related to the degree of fibrosis in AS patients (Figure 4, A and B). In contrast, mRNA expression of MMP-1, -2, or -9 did not differ significantly between AS and control patients (Figure 3A), nor was it related to collagen deposition (MMP-1, r=−0.11, P=NS; MMP-2, r=−0.18, P=NS; MMP-9, r=−0.14, P=NS) in AS patients. Significantly higher transcript levels of TIMP-1 and TIMP-2 and TIMP-2/MMP-2 ratio were observed in AS patients with pseudonormal or restrictive patterns compared with normal or abnormal relaxation (mRNA expression relative to GAPDH: TIMP-1, 1.89±0.4 versus 0.41±0.04, respectively, P<0.001; TIMP-2, 18.7±5.3 versus 3.2±1.2 respectively, P<0.01; TIMP-2/MMP-2, 538±213 versus 139±71, respectively, P=0.02).
Semiquantitative analysis revealed significantly increased TIMP-1 and -2 immunoreactivity in the left ventricle of AS (Table 2; Figure 2, D and F) compared with control patients (Table 2; Figure 2, C–E), whereas MMP-1, -2, or -9 immunoreactivity did not differ significantly between the 2 groups (Table 2; Figure 5, A–F). In AS hearts, immunoreactivity of TIMP-1 and -2 was superimposable and was most prominent in areas of pronounced hypertrophy and in the extracellular matrix (Figure 2, D and F) but was weak in normal myocardium (Figure 2, C–E). Cardiomyocytes surrounding fibrotic areas also strongly expressed TIMP-1 and -2, suggesting that not only fibroblasts8 but also cardiomyocytes13 produce TIMPs.
MMP Zymography and Immunoblot Analysis
Immunoblotting of extracts from additional pooled biopsies (n=4 per group) showed similar levels of protein expression of MMP-2 and MMP-9 in AS compared with control patients (% expression in AS patients relative to control patients: MMP-9, 89±8% in AS versus 100±15% in control patients; for MMP-2, 122±13% in AS versus 100±24% in control patients) (Figure 6G), concordant with similar MMP-2 and MMP-9 immunoreactivity.
In situ gelatin (MMP-2 and MMP-9) zymography of cryosections (n=6 per group) revealed consistently decreased MMP gelatinolytic activity in AS compared with control patients (Figure 6, A–F). All of the 6 randomly studied AS patients showed grade 0 (minimal) gelatinolytic activity, whereas 4 of 6 control patients showed grade 2 (moderate), 1 grade 3 (diffuse), and 1 grade 1 (focal) gelatinolytic activity (gray coloration), which was inhibited in the presence of EDTA.
Validation of CABG Patients as a Control Group
To ensure that heart tissue obtained during CABG indeed reflects relatively “healthy” myocardium, we performed additional structural analysis in CABG compared with AS tissues and also determined fibrosis, MMP and TIMP immunoreactivity, and MMP immunoblotting in CABG compared with nonused donor hearts.
Electron microscopy revealed pronounced interstitial fibrosis and slight degeneration of cardiomyocytes in AS but a normal appearance in CABG patients (Figure 7, A and B). Immunoreactivity of ubiquitin, a ubiquitous polypeptide with a pivotal role in intracellular protein degradation,14 was significantly increased in AS compared with control patients (Table 2; Figure 7, C and D). Finally, distribution of desmin was disturbed in AS but more homogeneous in CABG patients (Figure 7, E and F). These data thus confirmed that the selected control CABG patients did not show signs of ongoing myocyte ischemia, suffering, or degeneration, in contrast to pronounced myocyte degeneration, cell death, and fibrosis in AS patients.
When comparing nonused donor samples with CABG samples, collagen deposition, immunoreactivity of MMPs and TIMPs, and MMP immunoblotting did not differ significantly (Table 3; Figure 8, A–L) (% immunoblotting in nonused donor relative to CABG LV: MMP-9, 96±11% in donor versus 100±12% in CABG; for MMP-2, 88±9% in donor versus 100±27% in CABG; P=NS), indicating that LV tissue as obtained during CABG does not differ from myocardial tissue of nonused donor hearts with respect to the levels of TIMPs, MMPs, and fibrosis.
The present study investigates the expression of MMP/TIMP in cardiac transmural biopsies of patients with pronounced LV hypertrophy and fibrosis caused by AS. Fibrosis is a crucial determinant of cardiac dysfunction during pressure overload caused by AS,1,2,15,16 which highlights the need for a better understanding of the mechanisms contributing to fibrosis during chronic pressure overload by AS.
Microarray analysis reveals significant upregulation of transcript levels of collagens and enzymes involved in collagen synthesis in AS patients compared with control patients, congruent with increased collagen deposition in AS patients. In addition, cardiac transcript levels (reverse transcription–PCR) of TIMP-1 and -2 and TIMP-1/MMP-2, TIMP-2/MMP-2, and TIMP-2/MMP-9 ratios are significantly increased in AS compared with control patients. Importantly, both TIMP-1 and -2 transcript levels in AS patients correlate with the degree of LV fibrosis. The excess of myocardial collagen seen in AS patients may therefore be the result of the net balance between a higher degree of collagen synthesis relative to collagen degradation, as indicated by increased transcript levels of collagens and enzymes involved in collagen synthesis and decreased MMP activity in LV biopsies of AS compared with control patients.
MMPs and their physiological inhibitors, TIMPs, play a key role in collagen remodeling in the heart. Importantly, a time- and space-dependent window of TIMP/MMP balance may exist during the development of LV hypertrophy and its progression to heart failure.4,5 Acute pressure overload in animals is associated with increased myocardial MMP expression17 and activity.18 Recently, we demonstrated that MMP-9-gene inactivation or TIMP-1-gene overexpression in mice significantly reduced hypertrophic growth of cardiomyocytes and prevented dilatation during acute LV pressure overload.10 During chronic pressure overload of the LV, the relation between TIMP and MMP levels, LV fibrosis, and remodeling may be more complex. Evidence exists supporting the concept that diminished myocardial MMP activity could facilitate collagen deposition in developing hypertrophy during prolonged pressure overload.19–21 In the spontaneously hypertensive rat, the development of compensated hypertrophy is associated with increased myocardial TIMP levels, which would imply reduced MMP activity.20 When LV hypertrophy progresses to decompensation in the spontaneously hypertensive rat, myocardial TIMP levels fall below normal levels, which would favor increased MMP activity.19,20 Our AS patients presented with pronounced LV interstitial fibrosis but without overt LV dilatation or systolic dysfunction. In concordance, the shifted balance of TIMPs versus MMPs favored increased TIMP levels and decreased MMP activity.
In the present study, we observed increased transcript levels of TIMP-1 and -2 in AS compared with control patients, which was most pronounced for TIMP-2. TIMP-2 plays a regulatory role in the proteolytic activation of proMMP-2.22,23 At low concentration, TIMP-2 serves as a receptor for proMMP-2, resulting in increased activation of proMMP-2 by MT1-MMP. However, at high concentration, TIMP-2 neutralizes MT1-MMP and prevents MMP-2 activation, indicating that preferential inhibition of MMP-2 by high levels of TIMP-2 might facilitate interstitial fibrosis.22,23
The primary action of TIMPs is to inhibit matrix metalloproteinases, but numerous studies have reported cell growth–promoting, antiapoptotic, steroidogenic, and antiangiogenic activities (reviewed previously22,24). A part of these functions is attributed to MMP inhibition, but TIMPs also exhibit cellular activities that seem to be independent of MMP inhibition. Importantly, both TIMP-1 and TIMP-2 may stimulate the growth of fibroblasts in vitro,22,24,25 apart from their MMP inhibition. Lovelock et al25 clearly demonstrated a predominant role of TIMP-2 over TIMP-1 in stimulation of collagen production by cardiac fibroblasts, independent of its ability to inhibit MMPs. Thus, it remains possible that the upregulation of TIMPs increases cardiac fibrosis not just by inhibiting MMPs but also by separate direct profibrotic mechanisms.
Fibrosis is an important structural substrate for cardiac failure and sudden death.26,27 A major result of this study was to reveal that increased expression of cardiac TIMP-1 and TIMP-2 in AS compared with control patients was related to the degree of interstitial fibrosis. Therefore, TIMP-1 and -2 deserve further investigation as potential targets for therapeutic interventions in AS patients.
CABG patients without cardiac dysfunction, unstable angina, ischemic heart disease, or LV hypertrophy were used as control subjects. Coronary artery disease has been associated with increased plasma levels of MMPs/TIMPs.28 However, altered cardiac MMP/TIMP expression levels in patients with coronary artery disease but without LV dysfunction, ischemia, or hypertrophy have, to the best of our knowledge, never been demonstrated. Nonused donor hearts, as an alternative for control hearts, are also not likely to be completely normal. Confounding factors linked to brain death,29 cause of death (trauma), age, cardiac ischemia, and drug administration (inotropes)30 have been described to affect myocardial function and structure. In our study, we showed that myocardial tissue as obtained during CABG did not differ from myocardial tissue of nonused explanted donor hearts with respect to the levels of TIMPs, MMPs, and fibrosis.
The duration of AS-induced chronic pressure overload before aortic valve surgery could not be evaluated in the present study, because most of the patients presented only when severe AS resulted in dyspnea, angina, or syncope. Because of ethical considerations, follow-up biopsies could not be taken.
Use of β-blockers was different in control and AS patients. Whereas β-blockade reduces MMP-9 activity after acute myocardial infarction in rats,31 an effect of β-blockade on mRNA expression of cardiac TIMPs/MMPs has never been demonstrated. In the present study, TIMP/MMP transcript levels did not differ in AS or CABG patients with or without β-blockade (Data Supplement Table II).
This study was supported by a Dr Dekkers grant from the Netherlands Heart Foundation (NHS, 2003T036) to Dr Heymans, a research grant from the Leuven University, Belgium (OT-0346) to Dr Heymans and Dr Van de Werf, and a VIDI grant (016.036.346) from the Netherlands Organization for Scientific Research (NWO) to Dr Pinto. Dr Janssens is a clinical investigator for the Fund for Scientific Research-Flanders and holder of a chair financed by Astra-Zeneca.
The online-only Data Supplement can be found at http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.104.516963/DC1.
Spinale FG. Matrix metalloproteinases: regulation and dysregulation in the failing heart. Circ Res. 2002; 90: 520–530.
Creemers EE, Cleutjens JP, Smits JF, Daemen MJ. Matrix metalloproteinase inhibition after myocardial infarction: a new approach to prevent heart failure? Circ Res. 2001; 89: 201–210.
Hein S, Arnon E, Kostin S, Schonburg M, Elsasser A, Polyakova V, Bauer EP, Klovekorn WP, Schaper J. Progression from compensated hypertrophy to failure in the pressure-overloaded human heart: structural deterioration and compensatory mechanisms. Circulation. 2003; 107: 984–991.
Lombardi R, Betocchi S, Losi MA, Tocchetti CG, Aversa M, Miranda M, D’Alessandro G, Cacace A, Ciampi Q, Chiariello M. Myocardial collagen turnover in hypertrophic cardiomyopathy. Circulation. 2003; 108: 1455–1460.
Schroen B, Heymans S, Sharma U, Blankesteijn WM, Pokharel S, Cleutjens JPM, Porter JG, Evelo CTA, Duisters R, van Leeuwen REW, Janssen BJA, Debets JJM, Smits JFM, Daemen MJAP, Crijns HJGM, Bornstein P, Pinto YM. Thrombospondin-2 is essential for myocardial matrix integrity: increased expression identifies failure-prone cardiac hypertrophy. Circ Res. 2004; 95: 515–522.
Lehoux S, Lemarie CA, Esposito B, Lijnen HR, Tedgui A. Pressure-induced matrix metalloproteinase-9 contributes to early hypertensive remodeling. Circulation. 2004; 109: 1041–1047.
Coker ML, Jolly JR, Joffs C, Etoh T, Holder JR, Bond BR, Spinale FG. Matrix metalloproteinase expression and activity in isolated myocytes after neurohormonal stimulation. Am J Physiol. 2001; 281: H543–H551.
Kostin S, Pool L, Elsasser A, Hein S, Drexler HC, Arnon E, Hayakawa Y, Zimmermann R, Bauer E, Klovekorn WP, Schaper J. Myocytes die by multiple mechanisms in failing human hearts. Circ Res. 2003; 92: 715–724.
Krayenbuehl HP, Hess OM, Monrad ES, Schneider J, Mall G, Turina M. Left ventricular myocardial structure in aortic valve disease before, intermediate, and late after aortic valve replacement. Circulation. 1989; 79: 744–755.
Villari B, Vassalli G, Monrad ES, Chiariello M, Turina M, Hess OM. Normalization of diastolic dysfunction in aortic stenosis late after valve replacement. Circulation. 1995; 91: 2353–2358.
Schubert A, Walther T, Falk V, Binner C, Loscher N, Kanev A, Bleiziffer S, Rauch T, Autschbach R, Mohr FW. Extracellular matrix gene expression correlates to left ventricular mass index after surgical induction of left ventricular hypertrophy. Basic Res Cardiol. 2001; 96: 381–387.
Nagatomo Y, Carabello BA, Coker ML, McDermott PJ, Nemoto S, Hamawaki M, Spinale FG. Differential effects of pressure or volume overload on myocardial MMP levels and inhibitory control. Am J Physiol. 2000; 278: H151–H161.
Li H, Simon H, Bocan TM, Peterson JT. MMP/TIMP expression in spontaneously hypertensive heart failure rats: the effect of ACE- and MMP-inhibition. Cardiovasc Res. 2000; 46: 298–306.
Visse R, Nagase H. Matrix metalloproteinases and tissue inhibitors of metalloproteinases: structure, function, and biochemistry. Circ Res. 2003; 92: 827–839.
Li YY, McTiernan CF, Feldman AM. Interplay of matrix metalloproteinases, tissue inhibitors of metalloproteinases and their regulators in cardiac matrix remodeling. Cardiovasc Res. 2000; 46: 214–224.
Lovelock JD, Baker AH, Gao F, Dong JF, Bergeron AL, McPheat W, Sivasubramanian N, Mann DL. Heterogeneous effects of tissue inhibitors of matrix metalloproteinases on cardiac fibroblasts. Am J Physiol. 2005; 288: H461–H468.