1-1. Heart failure and cardiac remodeling
Heart failure (HF), a condition that impairs the ability of the heart to pump a sufficient amount of blood through the body, is a common cause of morbidity and mortality, and the incidence is increasing because of the aging population (Mathew et al., 2004; Hunt, 2005).
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 (Spinale, 2007). The summation of both cellular and extracellular alterations, a process termed cardiac remodeling, is revealed clinically as changes in the size, shape, and function of the heart (Swynghedauw, 1999; Cohn et al., 2000;
Leri et al., 2005). Histopathologically, it is characterized by a structural rearrangement of components of the normal chamber wall that involved cardiomyocyte hypertrophy, cardiac fibroblast proliferation, fibrosis, and cell death (Swynghedauw, 1999). Irrespective of its cause, this maladaptive remodeling contributes to diminished systolic performance, decreased compliance, and diastolic dysfunction in failing human heart (Brilla and Rupp, 1994).
1-2. ECM Remodeling and fibrosis
The myocardial extracellular matrix (ECM) is made up of fibrillar collagens network, basement membrane, proteoglycans, glycosamioglycans and contains a diverse array of bioactive signaling molecules (Janicki and Brower, 2002; Ito et al., 2005). The fibrillar collagen network ensures the structural integrity of the adjoining myocytes, provides the means by which myocyte shortening is translated into ventricular pump function, and is essential for maintaining alignment of the myofibrils within the myocytes through with a collagen-integrin-cytoskeletal myofibril relation (Janicki and Brower, 2002). 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 (MacKenna et al., 2000;
Borer et al., 2002). In addition to a fibrillar collagen network, the myocardial ECM contains a large reservoir of bioactive molecules that directly influence myocardial ECM synthesis and degradation (Chen et al., 2000; Cucoranu et al., 2005).
Fibrosis, which is a disproportionate accumulation of fibrillar collagen, is an integral feature of the remodeling characteristic of the failing heart (Kostin et al., 2000). Accumulation of type I collagen, the main fibrillar collagen found in cardiac fibrosis, stiffens the ventricles and impedes both contraction and relaxation (Sun and Weber, 2005; Zannad and Radauceanu, 2005). Fibrosis can also impair the electrical coupling of cardiomyocytes by separating myocytes with ECM proteins (Swynghedauw, 1999). Furthermore, fibrosis results in reduced capillary density and an increased oxygen diffusion distance that can lead to hypoxia of myocytes (Sabbah et al., 1995). Thus, it profoundly affectes myocyte metabolism and performance and ultimately alters ventricular function (Schnee and Hsueh, 2000; Manabe et al., 2002).
Fibrosis has been classified into two groups: reparative and reactive fibrosis. Reparative (replacement) fibrosis or scarring accompanies myocyte death. It is a result of a scarring process in which areas of necrosis heal after direct insults such as myocardial infarction (Whittaker, 1995). Reactive fibrosis appears as “interstitial” or “perivascular” fibrosis and does not directly associate with myocyte death. It may be a fibrogenic response of the myocardium to a variety of stimuli. In interstitial fibrosis, fibrillar collagen appears in intermuscular spaces (Kai et al., 2005). Perivascular firosis refers to the accumulation of collagen within the adventitia of intramyocardial coronary arteries and arterioles. Although there are a number of apparent differences between reparative and reactive fibrosis (e.g. cells involved and the time course of fibrotic change), many factors likely work in common to
control fibroblasts function (Kai et al., 2005).
Cardiac fibrosis is not only an increase in the concentration of matrix collagens but also changes in collagen type, organization and cross-links (Whittaker, 1995). Thus, despite the significant increase in collagen production, the replacement collagen is poorly cross-linked (Gunja-Smith et al., 1996). This compromises the supportive scaffolding leading to cell slippage, LV dilation and diminished diastolic compliance (Feldman et al., 2001). Various changes in the composition of collagen types and cross-links have been reported during the development of cardiac fibrosis in different animal models as well as in patients with heart failure (Li et al., 2000).
1-3. ECM remodeling and MMPs
The ECM is a dynamic structure with continuouschanges in the amount and proportions of its structural proteins (Dollery et al., 1995). The integrity of ECM is maintained by a balance in the activity of matrix metalloproteinases (MMPs), a family of enzymes that degrade all the matrix components of the heart, and their tissue inhibitors, TIMPs. Thus, an increase in MMP activity may result in fibrillar collagen degradation, ECM remodeling, and progressive ventricular dilatation (Li et al., 2000). MMPs not only play a role in ECM degradation but also synthesis. The end results is often increased MMPs accompanied with increased fibrosis in HF, and decreased MMPs activity accompanied with decreased fibrosis (Heymans et al., 1999). MMPs may participate in the fibrosis and remodeling process through direct digestion of matrix components, and regulation of the formation of matrikines such as glycyl–histidyl–lysine, derived from several degraded ECM protein can stimulate new connective tissue forming (Maquart et al., 1988), and release of biologically active factors from the ECM (Taipale and Keski-Oja, 1997).
The expressions of ECM and MMPs change dynamically during the developmental process of heart failure (Moshal et al., 2005). Table 1 summarizes the reported changes in MMPs and TIMPs profiles in dilated cardiomyopathy (DCM) in human. These studies indicate that although maladaptive remodeling in cardiac disease is generally associated with enhanced MMP and reduced TIMP activities, this pattern is not held universally and varies with the etiology, different stages of the disease or the effects of HF treatment (Kassiri and Khokha, 2005). In addition, genetic manipulation of different MMPs or TIMPs in animal models has provided insights into their roles in cardiovascular development and in
progression of cardiac disease. Table 2 summarizes several results from MMPs and TIMPs transgenic studies.
1-4. MMPs : structures and functions
MMPs is a family of extracellular zinc-dependent neutral endopeptidases (Lombard et al., 2005), capable of degrading essentially all ECM components including fibrillar and
non-fibrillar collagens, fibronectin, laminin and basement membrane glycoproteins (Fedarko et al., 2004). MMPs not only play an important role in ECM remodeling in physiologic situations, such as embryonal development, tissue regeneration, and wound repair, also in pathological conditions including rheumatoid arthritis, osteoarthritis, atherosclerotic plaque rupture, tissue ulceration, and in cancer cell invasion and metastasis (Roeb and Matern, 2001;
Jones et al., 2003).
MMPs are generally divided into six groups, interstitial collagenases (MMP-1, -8 and -13), stromelysins (MMP-3, -10, -11 and -12), matrilysins (MMP-7 and MMP-26), gelatinases (MMP-2 and MMP-9), membrane-type MMPs (MMP-14, -15, -16, -17, -24 and -25) and others (Hijova, 2005). Although MMPs are subclassified based on their ability to degrade
various proteins of the ECM, they also play other important roles such as the activation of cell surface receptors and chemokines (Stefanidakis and Koivunen, 2006). In addition, MMP-2 has proteolytic activity to specific targets within the cell to cause acute, reversible contractile dysfunction in cardiac disease (Schulz, 2007). Classification and nomenclature of all the types of MMPs were listed in Table 3. The basic structures of MMPs can be approximately divided into three structurally well-preserved domain motifs, including a catalytic domain, an
N-terminal domain and a C-terminal domain. Zinc-dependent catalytic domain of MMPs is similar with subtle structural differences among the substrate specific groups (Nagase and Woessner, 1999). The N-terminal domain (propeptide domain) contains a unique
PRCG(V/N)PD sequence in which the cysteine residue interacts with the catalytic zinc atom in the active site, prohibiting activity of the MMPs. Thus, the interaction has to be disrupted to “open” the cysteine switch in the process of MMPs activation (van Wart and
Birkedal-Hansen, 1990), which is a critical step that leads to ECM breakdown (Carmeli et al., 2004). The C-terminal hemopexin domain of metalloproteinases has a four-bladed propeller structure and contributes to substrate specificity (Wallon and Overall, 1997). In
membrane-type MMPs, the hemopexin domain contains a transmembrane domain for
anchoring the protein in the membrane; besides, the hemopexin domain in MMP-2 also has a function in the activation of the enzyme (Morgunova et al., 1999; Overall et al., 1999). The regulation of MMPs occurs at many levels, including transcription (the major one),
post-transcriptional modulation of mRNA stability, secretion, localization, zymogen (proenzyme) activation and inhibition of activity by natural inhibitors of MMPs, tissue inhibitor of metalloproteinases (TIMPs).
1-4-1. Collagenase-1 (MMP-1)
MMP-1 was first purified from the tail of tadpole, the collagenolytic activity is required
to digest the collagen of the tail during amphibian metamorphosis (Gross and Lapiere, 1962).
Human MMP-1 cDNA clone and the sequence were obtained from adult skin fibroblasts (Goldberg et al., 1986). Human MMP-1 is produced as two differently glycosylated
proenzymes, a major 52 kDa and a minor 57 kDa form. Activation of these two latent forms generates two active proteinases 42 kDa and 47 kDa in size, respectively (Wilhelm et al., 1986). It was considered that MMP-1 was deficient in rodent until two closely related mouse counterparts to human MMP-1 were cloned: Mcol-A and Mcol-B. The Mcol-A and Mcol-B are expressed during embryo implantation, but only Mcol-A is able to cleave fibrillar collagens (Balbin et al., 2001).
1-4-2. Gelatinase A (MMP-2)
In 1978, Sellers et al. were first to separate a gelatinase activity from collagenase and stromelysin in the culture medium from rabbit bone (Sellers et al., 1978). A similar enzyme, acting on basement membrane type IV collagen was reported by Liotta et al. (1979) in the following year. Gelatinase was purified from human skin, mouse tumor cells, rabbit bone, and human gingival. The completed sequence of the human MMP-2 except for the signal peptide was reported by Collier et al. (2001). Gelainase A has a triple repeat of fibronectin type I domains inserted in the catalytic domain; this domain participates in binding to the gelatin substrates of the enzyme (Libson et al., 1995; Lee et al., 1997). MMP-2 is ubiquitously expressed in the cells which comprise the heart and is found in normal cardiomyocytes, as well as in endothelium, vascular smooth muscle cells and fibroblasts (Coker et al., 1999).
1-4-3. Gelatinase B (MMP-9)
In 1972, Harrwas and Krane detected a gelatinase activity in rheumatoid synovial fluid.
Sopata et al. described a gelatinase from human polymorphonuclear leukocytes (Sopata and Wize, 1979). Rabbit macrophages produce a very similar enzyme which is able to digest type V collagen (Horwitz et al., 1977). The neutrophil collagenase and gelatinase were resolved in 1980 (Murphy et al., 1980). Purification of MMP-9 protein was achieved in 1983 and
sequencing of the cDNA was completed in 1989. An interesting phenomenon, still not fully understood, is the binding of TIMP-1 to proMMP-9 to form a complex (Sakyo et al., 1983;
Stetler-Stevenson et al., 1989). Human neutrophil MMP-9 commonly occurs as a complex with lipocalin (Fernandez et al., 2005). A series of papers concerned a 95 kDa protein in plasma that binds to gelatin culminated in the identification of this protein as MMP-9 (Makowski and Ramsby, 1998).
1-5. TIMPs: structures and functions
The family of TIMPs presently numbers four distinct gene products that are specific inhibitors of the MMPs through binding in a 1:1 stoichiometry reversibly (Cook et al., 1994;
Okada et al., 1994; Silbiger et al., 1994; Greene et al., 1996). These secreted proteins are thought to regulate MMPs activity during tisssue remodeling (Baker et al., 2002). All four mammalian TIMPs have many basic similarities, but they exhibit distinctively structural features, biochemical properties and expression patterns (Table 4). This suggests that each TIMP has specific roles in vivo. The local balance between MMPs and TIMPs is believed to play a major role in ECM remodeling during process of diseases such as cancer and arthritis (Anand-Apte et al., 1996). The TIMPs have molecular weights of ~20 to 30 kDa and are variably glycosylated (Baker et al., 2002). They have six disulphide bonds and comprise a three-loop N-terminal domain and an interacting three-loop C-subdomain. Most of the biological functions of these proteins discovere thus far are attributable sequences within the N-terminal domain, although the C-subdomain mediated interaction with the catalytic
domains of some MMPs (Li et al., 1999) and with the hemopexin domains of MMP-2 and MMP-9 (Brew et al., 2000). The TIMPs are secreted proteins, but may be found at the cell surface in association with membrane-bound proteins; for example, TIMP-2, TIMP-3 and TIMP-4 can bind MMP-14, a membrane-type (MT) MMP. All four TIMPs inhibit active forms of all MMPs studies to date, their binding constants being in the low picomolar range, although TIMP-1 is a poor inhibitor of MMP-19 and a number of the MT-MMPs (Baker et al., 2002).
1-6. Doxorubicin and cardiotocixity
Doxorubicin (Dox; adriamycin) is one of the original anthracyclines isolated in the early 1960s from the pigment-producing bacterium Streptomyces peucetius (Takemura and
Fujiwara, 2007). Dox is one of the most widely used antitumor drugs. It is effective against a wide spectrum of cancers including acute leukemia, Hodgkin’s and non-Hodgkin’s lymphoma, and breast cancer (Weiss, 1992; Singal and Iliskovic, 1998). The mechanisms of antitumor that have been suggested include: (1) intercalation into DNA, leading to inhibition of synthesis of macromolecules; (2) generation of reactive oxygen species (ROS), leading to DNA damage or lipid peroxidation; (3) DNA binding and alkylation; (4) DNA cross-linking;
(5) interference with DNA unwinding or DNA strand separation and helicase activity; (6) direct membrane effects; (7) initiation of DNA damage via inhibition of topoisomerase II; and (8) induction of apoptosis in response to topoisomerase II inhibition (Gewirtz, 1999; Minotti et al., 2004). Despite the usefulness, the side effects of Dox can cause dilated cardiomyopathy in a dose-dependent manner. The sharp increase in the incidence of cardiomyopathy at
cumulative doses above 550 to 600 mg of Dox per square meter of body surface area has formed the basis to set an empirical dose limit of 500 mg/m2 (Minotti et al., 2004).
Progressive ventricular dysfunction and congestive heart failure may occur after the patients
even have stopped Dox treatments (Singal et al., 1987). The mechanism of the cardiotoxicity remains unclear, but it is likely to be distinct from the mechanism of antitumor. Most studies support the view that an increase in oxidative stress, evidenced increases in the levels of ROS (Kalyanaraman et al., 1980; Doroshow, 1983), and lipid peroxidation (Singal et al., 1985;
Singal et al., 1987), along with reductions in the levels of antioxidants and sulfhydryl groups (Odom et al., 1992), play a key role in the pathogenesis of Dox-induced cardiomyopathy.
1-7. Monitor of cardiotoxicity 1-7-1. Electrocardiography
Electrocardiography (ECG) is a widely used and inexpensive technique. It is also considered useful in the identification of cardiotoxic effects induced by Dox. The changes of ECG on patients treated with Dox may occur including reduction in the voltage of the
QRS-wave, T-wave flattening (Huang et al., 2004) and Q-T interval extension (Nousiainen et al., 1999). Although ECG is not a functional parameter, several studies have described the ECG changes in the animals upon administration of anthracyclines. The increased S-T
interval was monitored by telemetric ECG system in non-anesthetic mice with administration of Dox (van Acker et al., 1996; Fisher et al., 2005). The widening of the S-T interval, which stands for the prolongation of the repolarization phase, may be explained by the prolongation of the action potential (van Acker et al., 1996). The action potential has been prolonged in Purkinje fibers after incubation with Dox (le Marec et al., 1986). On the other hand, Dox induced a lengthening of the P-R (Puri et al., 2005), Q-T interval (Sacco et al., 2001) in rats and a decreased amplitude of R-wave in canines were reported (de Souza and Camacho, 2006).
1-7-2. Natriuretic peptides
There are three major natriuretic peptides, atrial natriuretic peptide (ANP), brain natriuretic peptide (B-type NP, BNP), and C-type natriuretic peptide. Natriuretic peptides belong to a family of structurally related peptides and share the similar features of bioactivity.
ANP, which was isolated from right atrium extracts in 1981 (de Bold et al., 1981), is preferentially synthesized and secreted from atria under physiological conditions (de Bold, 1985). BNP is synthesized in both atria and ventricles, but is predominantly released from the latter (Mukoyama et al., 1991). However, ANP and BNP can be synthesized in either chamber under pathologic conditions (Yasue et al., 1994), which play roles as hormones to act in various tissues in body and induce vasodilation, natriuresis, and diuresis to protect the cardiovascular system from overload (Nakao et al., 1992). The plasma levels of natriuretic peptides have been shown to be increased in patients with CHF (Wei et al., 1993), besides, the amount of natriuretic peptides highly correlating to cardiovascular diseases as a potential marker have been reported (Daniels and Maisel, 2007).
In Dox-induced cardiomyopathy, the measurement of plasma ANP (Bauch et al., 1992;
Hayakawa et al., 2001), N-terminal proANP (Tikanoja et al., 1998), BNP (Pinarli et al., 2005) and N-terminal proBNP (Soker and Kervancioglu, 2005) during anthracycline treatment in clinical studies were reported, suggest that ANP and BNP are also useful makers of LV dysfunction in patients undergoing anthracycline therapy (Bryant et al., 2007). In the
experiment of animal model, the increase of ANP mRNA in the hearts of Dox-treated dog, rat (Rahman et al., 2001) and rabbit (Boucek et al., 1999) were demonstrated. In addition, the Dox-treated neonatal piglets revealed increased BNP mRNA levels in LV and RV (Torrado et al., 2003), and the plasma level of BNP but not ANP was significantly increased in
Dox-treated rats (Koh et al., 2004). Nevertheless, the plasma BNP was not augmented in a canine model (Alves de Souza and Camacho, 2006).