|Home | About | Journals | Submit | Contact Us | Français|
Myocardial fibrosis is an integral component of most cardiac pathologic conditions and contributes to the development of both systolic and diastolic dysfunction. Because of the availability of genetically manipulated animals, mouse models are essential for understanding the mechanisms involved in the pathogenesis of cardiac fibrosis. Accordingly, we characterized the inflammatory and fibrotic response in a mouse model of cardiac pressure overload due to transverse aortic constriction (TAC). Following TAC, mouse hearts exhibited induction of chemokines and proinflammatory cytokines, associated with macrophage, but not neutrophil, infiltration. Induction of inflammatory cytokines was followed by a late upregulation of Transforming Growth Factor (TGF)-β isoforms, activation of the Smad2/3 and Smad1/5 pathways, induction of matricellular proteins, and deposition of collagen. Inflammatory activity decreased after 28 days of TAC; at this timepoint established fibrosis was noted, accompanied by ventricular dilation and systolic dysfunction. Late induction of inhibitory mediators, such as TGF-β may play an essential role in the transition from inflammation to fibrosis by suppressing inflammatory gene synthesis while inducing matrix deposition. Our findings identify molecular mediators and pathways with a potential role in cardiac fibrosis laying the foundations for studies exploring the pathogenesis of fibrotic cardiac remodeling using genetically targeted mice.
Myocardial fibrosis is an integral component of most cardiac pathologic conditions (Brown et al. 2005). Accumulation of extracellular matrix in the cardiac interstitium disrupts the coordination of myocardial excitation-contraction coupling in both systole and diastole and may result in profound functional impairment (Janicki and Brower 2002). Beyond its effects on function, myocardial fibrosis also promotes arrhythmogenesis through impaired anisotropic conduction and subsequent generation of reentry circuits (Khan and Sheppard 2006).
Because the heart has negligible regenerative capacity, most forms of cardiac injury ultimately result in the development of fibrosis. Myocardial infarction is associated with death of a large number of cardiomyocytes and sets into motion a reparative response leading to formation of a collagen-based scar (Cleutjens et al. 1995). Extensive evidence suggests the involvement of inflammatory mechanisms in post-infarction cardiac repair (Frangogiannis 2006b). Cardiomyocyte death triggers an acute inflammatory reaction resulting in infiltration of the infarcted myocardium with leukocytes that clear the wound from dead cells and matrix debris. Subsequent activation of inhibitory mediators, such as TGF-β, suppresses proinflammatory gene synthesis while promoting fibrosis (Bujak and Frangogiannis 2007). Although the inflammatory reaction is critically involved in post-infarction reparative fibrosis (Frangogiannis 2006a), (Frangogiannis 2008) the mechanistic basis of fibrotic cardiac remodeling in response to injurious stimuli that do not result in cardiomyocyte death remains poorly understood.
Pressure overload induced by hypertension or aortic stenosis results in extensive fibrotic remodeling of the heart (Heymans et al. 2005), (Cingolani et al. 2004), initially associated with diastolic dysfunction that frequently progresses to ventricular dilation and combined diastolic and systolic heart failure (Berk et al. 2007). The availability of a murine model of transverse aortic constriction (TAC) provides us with a valuable tool to explore the mechanisms involved in cardiac fibrosis following pressure overload using genetically targeted animals. However, exploration of the pathogenesis of fibrotic cardiac remodeling requires detailed characterization of the inflammatory and fibrotic response in pressure-overloaded mouse hearts. Our investigation examines the time course of cellular and molecular events leading to transition from inflammation to myocardial fibrosis in mice undergoing TAC protocols. Despite the absence of significant cardiomyocyte loss, pressure overload triggers a transient inflammatory reaction, associated with chemokine and cytokine upregulation and recruitment of macrophages, but not neutrophils, in the murine heart. Induction of pro-inflammatory cytokines activates inhibitory mediators, such as TGF-β, that suppress inflammation, but also promote interstitial and perivascular fibrosis. Because pressure overload is not associated with significant cardiomyocyte loss, TGF-β-mediated matrix deposition has no reparative function and is maladaptive. Fibrotic remodeling of the ventricle is initially associated with marked hypertrophy without ventricular enlargement or systolic functional impairment, followed by the development of chamber dilation and systolic dysfunction. Characterization of the inflammatory and fibrotic response in the murine pressure-overloaded heart will facilitate mechanistic studies using genetically targeted animals to explore the pathogenesis of cardiac fibrosis.
Animal experiments were approved by the Baylor College of Medicine Institutional Review Board. All animals received humane care in compliance with the “Principle of laboratory and animal care” (NIH publication No 86-23, revised 1985). Male and female, 3-5 month-old C57/BL/6 mice (Jackson laboratories) were anesthetized by an intraperitoneal injection of sodium pentobarbital (60 μg/g). Aortic banding was achieved by creating a constriction between the right innominate and left carotid arteries. A 6-0 suture was tied twice around a blunt 3-mm segment of a 27-gauge needle, which was positioned adjacent to the aorta and was removed after placement of the ligature. The degree of pressure overload was measured by right-to-left carotid artery flow velocity ratio after constricting the transverse aorta. Only mice with a flow ratio from 5:1 to 10:1 were used for analysis. At the end of the experiment, the heart was excised, fixed in zinc-formalin, and embedded in paraffin for histological studies, or frozen for RNA/protein isolation. Animals used for histology underwent 3, 7, and 28 days of banding (n=8/group). Mice used for RNA extraction underwent 3 (n=7) or 7 days (n=4) of banding. An additional group of mice (n=7) was used for protein extraction after 7 days of TAC. As a control, a “sham” operation without aortic constriction was performed on age-matched mice (histology n=6, RNA n=6, protein n=8).
Zinc-formalin-fixed sections were cut at 5 μm and stained immunohistochemically with the following antibodies: mouse anti-α–smooth muscle actin (α-SMA) (Sigma, St. Louis, MO) rat anti-mouse macrophage Mac-2 (Cedarlane, Burlington NC), rat anti-neutrophil (Serotec, Raleigh NC), rat anti-mouse CD31 (Pharmingen, San Diego, CA), and rabbit anti-mouse tenascin-C antibody (R&D Systems). Staining was performed using a peroxidase-based technique with the Vectastain ELITE kit (Vector, Burlingame, CA) and developed with diaminobenzidine+nickel (Vector) (Zymek et al. 2006), (Gersch et al. 2002). Subsequently the sections were counterstained with eosin. The MOM kit (Vector) was used for α-SMA staining. For CD31 staining, the Tyramide Signal Amplification kit (Perkin Elmer, Boston, MA) was used on trypsin-pretreated sections (Zymek et al. 2006). All staining techniques were validated using appropriate negative and positive controls. Negative controls included sections stained after omission of the primary antibody. Sections from mouse infarcts that contain large numbers of neutrophils, macrophages, myofibroblasts and vascular cells (Dewald et al. 2004), and express large amounts of tenascin-C (Bujak et al. 2007) were used as positive controls. Assessment of neutrophil and macrophage density was performed by counting the number of neutrophils and Mac-2-immunoreactive cells respectively (Dewald et al. 2005). Quantitation of the tenascin-C-stained area was performed in four fields from two sections from each animal. Tenascin-C staining was expressed as a percentage of the total myocardial area. The collagen network was identified using picrosirius red staining (Dewald et al. 2005). Assessment of the percentage of the collagen-stained area was performed in basal, mid-ventricular and apical segments of the left ventricle and in the right ventricle using ImagePro software. Nine high-power fields were used for each segment. Microvascular density was assessed by counting the number of CD31-positive profiles in the endocardial and epicardial segment of the left ventricular free wall, the septum and the right ventricle. Four high-power fields were used for analysis of each segment. In order to quantitate the development of perivascular fibrosis in pressure-overloaded hearts, arterioles that were cut in cross-section were scanned. The media and the adventitia were traced and the ratio of the adventitial to the medial area (A/M) was measured. Ten random arterioles were used for analysis in each animal.
Inflammatory gene expression in murine hearts was assessed using an RPA (RiboQuant; Pharmingen) (Dewald et al. 2005), (Gersch et al. 2002). mRNA expression levels of the chemokines Macrophage Inflammatory Protein (MIP)-1α, MIP-1β, MIP-2, Monocyte Chemoattractant Protein (MCP)-1, Interferon-γ-inducible Protein (IP)-10, lymphotactin, T cell activation protein (TCA)-3 and eotaxin, the cytokines Tumor Necrosis Factor (TNF)-α, Interleukin (IL)-1β, IL-6, and IL-10, and TGF-β1, -β2, and -β3, and the matricellular protein Osteopontin were determined as previously described (Dewald et al. 2005). The signals were quantified using Image QuaNT software and normalized to the ribosomal protein L32 mRNA. Unprotected probes were loaded as size markers.
Protein was isolated from whole hearts (sham, 7 days TAC). Western blotting with rabbit anti-Smad2, anti-p-Smad2, anti-p-Smad1/5, and anti-Smad1/5 (Cell Signaling, Beverly, MA) was performed as previously described (Bujak et al. 2007). Protein extracted from TGF-β-stimulated mouse cardiac fibroblasts was used as a positive control (Bujak et al. 2007), (Bujak et al. 2008b).
Short axis M-mode echocardiography was performed prior to the surgical procedure and before the end of each experiment (3, 7 or 28 days of TAC) using a 13 MHz probe (Sequoia C256; Acuson, Mountain View, CA) (Bujak et al. 2008a). The following parameters were measured as indicators of function and remodeling: left ventricular end-diastolic diameter (LVEDD), left ventricular end-systolic diameter (LVESD), fractional shortening (FS=[LVEDD-LVESD]×100/LVEDD), interventricular septal (IVS) thickness, posterior wall (PW) thickness and left ventricular mass (LV mass= 1.05(IVS thickness + LVEDD + PW thickness)3 − LVEDD3 (Collins et al. 2003).
Statistical analysis was performed using ANOVA followed by t-test corrected for multiple comparisons (Student-Newman-Keuls). Paired t-test was used to compare echocardiographic endpoints before instrumentation and after TAC. Data were expressed as mean ± SEM. Statistical significance was set at 0.05.
Pressure overload triggers a transient inflammatory response in the myocardium associated with increased expression of pro-inflammatory cytokines. Sham-operated hearts exhibited negligible mRNA expression of TNF-α, IL-1β and IL-6. After 3 days of TAC significant TNF-α and IL-1β mRNA upregulation was noted in the pressure-overloaded heart; in contrast, IL-6 expression remained low (Fig. 1). After 7 days of aortic banding, TNF-α and IL-1β expression returned to sham levels.
Several members of the chemokine family were upregulated in the pressure overloaded heart. Sham-operated hearts exhibited low level expression of MCP-1 (MCP-1:L32 ratio:0.043±0.0021) and MIP-2 mRNA (MIP-2:L32 ratio 0.023±0.02), but negligible expression of IP-10, MIP-1α, MIP-1β, TCA-3 and eotaxin message (<1% of L32 levels). After 7 days of TAC significant induction of MIP-2, IP-10 and MCP-1 was noted (Fig. 2). In contrast, cardiac mRNA levels of MIP-1α, MIP-1β, lymphotactin, TCA-3 and eotaxin remained low (ratio to L32<0.01).
Upregulation of pro-inflammatory mediators was associated with infiltration of the pressure-overloaded myocardium with leukocytes. A modest increase in myocardial neutrophil density was noted after 3 days of TAC, but did not reach statistical significance (Fig. 3). On the other hand, macrophage density was markedly increased in the pressure-overloaded heart after 7 days of TAC (Fig. 3B, E). Partial resolution of the macrophage infiltrate was observed after 28 days of aortic banding (Fig. 3C-E).
In the pressure-overloaded heart the inflammatory response was associated with induction of matricellular proteins followed by the development of diffuse fibrosis. Perivascular and interstitial infiltration with spindle-shaped α-SMA-positive myofibroblasts was noted after 7 days of TAC (Fig. 4A). The matricellular proteins osteopontin and tenascin-C were markedly induced in the pressure-overloaded ventricle. Induction of osteopontin mRNA was noted after 3 days of TAC (Fig. 4B). Extensive deposition of tenascin-C (Fig. 4C-D), a marker of active interstitial remodeling (Frangogiannis et al. 2002), was observed after 7 days, but significantly decreased after 28 days (Fig. 4C) of aortic banding. Sirius red-stained area significantly increased in the pressure-overloaded heart after 3-28 days of aortic banding, peaking at 7 days (Fig. 4F, G). Marked expansion of the cardiac interstitium was noted in both the left and right ventricular myocardium and was comparable in basal, mid-ventricular and apical segments (Fig. 4G).
Inflammatory and fibrotic changes in the pressure-overloaded heart were predominantly localized in perivascular areas. After 7 days of TAC, cardiac arterioles showed marked perivascular inflammation and expansion of the adventitia (Fig. 5B-E), associated with macrophage (Fig. 5C) and myofibroblast infiltration. The ratio of the arteriolar adventitial to medial area (A:M) markedly increased after 7-28 days of aortic banding (Fig. 5F). After 28 days of TAC, deposition of dense collagen fibers was noted in the arteriolar adventitia (Fig. 5E).
Because TGF-β plays a crucial role in the transition from inflammation to fibrosis we examined the time course of TGF-β isoform expression in the pressure-overloaded heart. Fibrotic remodeling of the ventricle was associated with marked upregulation of TGF-β1, -β2 and -β3 mRNA after 7 days of TAC (Fig. 6A-C) and with activation of TGF-β signaling evidenced by markedly increased levels of p-Smad2 and p-Smad1 (Fig. 6D-E).
In sham-operated hearts, the density of CD31-positive vessels (Fig. 7) was significantly lower in the right ventricular myocardium when compared to the left ventricle (Fig. 7D). The number of microvascular profiles was comparable in subendocardial and subepicardial areas of the left ventricle (Fig. 7D). After 3 days of TAC, microvascular density significantly increased in all left ventricular myocardial segments. Microvascular density in both right and left ventricular myocardium peaked after 7 days and remained elevated after 28 days of aortic banding (Fig. 7D).
The inflammatory and fibrotic changes in the pressure-overloaded heart were associated with alterations in chamber dimensions and progressive impairment of left ventricular function. After 3-7 days of TAC, the hearts exhibited marked concentric hypertrophy (Fig. 8), but had preserved systolic function (Table 1). LV mass was markedly increased; however, chamber dimensions (LVEDD and LVESD) remained normal, and fractional shortening was preserved (Table 1). After 28 days of aortic banding, LV mass increased further, and the hypertrophied heart developed systolic dysfunction (suggested by a significant reduction in FS) and dilative remodeling (indicated by a marked increase in LVESD) (Fig. 8C, Table 1).
The availability of genetically targeted mice has revolutionized biomedical research providing unique opportunities to study the role of specific mediators in complex pathobiological processes. However, dissection of the mechanistic basis of a pathophysiologic condition requires knowledge of the time course of cellular and molecular events in carefully designed mouse models of disease. Our study describes in detail the cardiac inflammatory and fibrotic response in a mouse model of cardiac pressure overload due to TAC. The findings provide the necessary foundations for mechanistic studies using genetically targeted animals to explore the role of inflammatory mediators in the development of fibrous tissue deposition and cardiac dysfunction in the pressure-overloaded heart.
Cardiomyocyte necrosis activates innate immune mechanisms triggering a potent inflammatory reaction that ultimately results in reparative fibrosis. Evolving evidence suggests that cardiac insults that do not cause cardiomyocyte death, such as pressure overload, are also capable of inducing an inflammatory reaction in the myocardium. Several distinct mechanisms may be involved in activation of the inflammatory response in the pressure overloaded heart. Neurohormonal activation may directly stimulate inflammatory pathways through generation of reactive oxygen species and activation of the NF-κB system. Angiotensin II and aldosterone enhance NF-κB activity exerting direct pro-inflammatory effects on the myocardium (Muller et al. 2000), (Sun et al. 2002). On the other hand pressure overload may induce alterations in cardiac extracellular matrix through activation of Matrix Metalloproteinases (MMPs). Matrix fragments may serve as “danger signals” that activate Toll-Like Receptors (TLRs), triggering inflammatory mediator synthesis in the pressure-overloaded myocardium (Beg 2002).
Although the inflammatory reactions following infarction and pressure overload share many common characteristics, they also have significant differences. Compared with the rapid upregulation of a broad spectrum of inflammatory mediators in myocardial infarction (Dewald et al. 2004), pressure overload induces only a subgroup of inflammatory chemokines and cytokines. IL-6, MIP-1α and MIP-1β expression is markedly upregulated in healing infarcts (Dewald et al. 2004), but not in pressure-overloaded hearts. In addition, abundant infiltration of the myocardium with neutrophils is noted in infarction, but not following aortic constriction. The intense inflammatory reaction in the infarcted myocardium probably reflects the rapid activation of several distinct proinflammatory pathways (such as the complement cascade, reactive oxygen and TLR-mediated pathways) following cardiomyocyte necrosis. In contrast, in the pressure-overloaded ventricle, the late induction of chemokine synthesis (Fig. 2) appears to be more consistent with delayed pro-inflammatory effects of neurohormonal mediators.
What is the role of inflammatory mediators in fibrotic remodeling of the pressure-overloaded ventricle? Pro-inflammatory cytokines are highly pleiotropic and may regulate the fibrotic process through several distinct pathways. Both TNF-α and IL-1β decrease collagen synthesis and enhance MMP expression and activity in cardiac fibroblasts (Siwik and Colucci 2004), (Li et al. 2002), (Brown et al. 2007) inducing matrix degradation (Siwik et al. 2000). However, TNF-α and IL-1β also stimulate expression of growth factors (such as TGF-β) that promote matrix deposition. The relative significance of these properties in the pathogenesis of cardiac fibrosis remains poorly understood. Transgenic mice with cardiac-specific overexpression of TNF-α develop heart failure (Bryant et al. 1998) associated with increased collagen deposition and denaturation, and enhanced MMP activity (Li et al. 2000). Enhanced fibrosis in TNF-α–overexpressing animals is associated with increased expression of TGF-β isoforms (Sivasubramanian et al. 2001). These findings suggest that TNF-α mediates fibrotic changes, while enhancing matrix-degrading activity. TNF-α null mice on the other hand exhibited reduced fibrosis and decreased MMP-9 activity following aortic constriction suggesting that this cytokine is an important profibrotic mediator in the pressure-overloaded heart (Sun et al. 2007). In contrast, the role of IL-1 signaling in fibrotic remodeling of the pressure-overloaded heart has not been investigated.
Chemokines, such as MCP-1 may also play an important role in the pathogenesis of cardiac fibrosis (Frangogiannis et al. 2007). MCP-1 mRNA expression was induced in a rat model of pressure overload due to suprarenal aortic constriction. Chronic MCP-1 neutralization attenuated fibrosis, but not cardiomyocyte hypertrophy, and ameliorated diastolic dysfunction without affecting systolic function (Kuwahara et al. 2004). In the cardiomyopathic heart MCP-1 may mediate its pro-fibrotic effects through recruitment of mononuclear cells that may serve as an important source of fibrogenic mediators. In addition, MCP-1 may directly modulate fibroblast phenotype and activity (Kruglov et al. 2006), (Gharaee-Kermani et al. 1996) and may be involved in recruitment of fibrocytes, a circulating population of cells that share leukocyte and mesenchymal markers, and are capable of myofibroblast differentiation (Quan et al. 2004), (Wynn 2008).
However, activation of inflammatory pathways in the pressure-overloaded heart is a transient event, followed by resolution of the leukocyte infiltrate and development of fibrosis. Induction of pro-inflammatory mediators triggers “stop signals” capable of suppressing acute inflammation. In the reparative response following myocardial infarction, TGF-β appears to play an important role as an inhibitory mediator, suppressing pro-inflammatory cytokine synthesis (Bujak and Frangogiannis 2007), (Ikeuchi et al. 2004), while promoting extracellular matrix deposition (Bujak and Frangogiannis 2007), (Lutgens et al. 2002). Although in the infarcted myocardium the anti-inflammatory and profibrotic actions of TGF-β may bridge the inflammatory and the reparative phase of healing, in the pressure-overloaded heart, TGF-β-mediated matrix deposition, in the absence of significant cardiomyocyte loss, is clearly maladaptive. Activation of TGF-β signaling pathways in the pressure-overloaded heart may suppress inflammation, but at a heavy cost, resulting in fibrotic remodeling of the ventricle. Thus, dissection of the signaling pathways responsible for the anti-inflammatory and pro-fibrotic actions of TGF-β is crucial for designing optimal therapeutic strategies that attenuate fibrosis without interfering with resolution of inflammation. We demonstrated activation of both the Smad2/3 and the Smad1/5 pathways in the pressure overloaded myocardium (Fig. 6). Smad3 signaling appears to play an essential role in remodeling of the infarcted heart, in part by mediating TGF-β-induced expression of matrix proteins by border zone fibroblasts (Bujak et al. 2007). On the other hand, activation of the Smad1 pathway has been suggested as a Smad3-independent mode of TGF-β signaling that may operate in chronic fibrotic disorders (Pannu et al. 2007). The role of these pathways in mediating anti-inflammatory and profibrotic actions in the pressure-overloaded heart remains unknown.
Cardiac fibrosis following pressure overload is associated with extensive vascular remodeling. Microvascular density significantly increases after 3-28 days of TAC (Fig. 7); increased perfusion may be needed to meet the demands of the hypertrophied myocardium and to ensure delivery of nutrients to the cells involved in fibrotic remodeling. Thus, angiogenesis accompanies the development of fibrosis and may be essential for fibroblast growth (Strieter et al. 2007). In addition, myocardial arterioles actively participate in the fibrotic process. Pressure overload results in adventitial inflammation followed by extensive periarteriolar fibrosis (Fig. 5), reflecting the prominent role of the proinflammatory and profibrotic effects of local angiotensin II on vascular and adventitial cells (McEwan et al. 1998), (Lee et al. 1995).
Our quantitative analysis seems to indicate that collagen deposition peaks after 7 days and may somewhat decrease after 28 days of TAC (Fig. 4G). However, these findings should not be interpreted as indicative of reversal of fibrosis. In order to study the extent of collagen deposition in sections from the pressure-overloaded heart we quantitatively assessed the area stained for Sirius red. Because Sirius red also binds to non-collagenous proteins containing basic amino acids, the technique is not specific for collagen unless polarized light microscopy is performed (Junqueira et al. 1979), (Nielsen et al. 1998). In addition, quantitation of the collagen content in the heart using a hydroxyproline assay was not performed. Although our findings accurately illustrate the expansion of the cardiac interstitial space following TAC, quantitative analysis of the Sirius red-stained area does not directly reflect the amount of collagen in the pressure-overloaded heart.
The time course of cellular events in fibrotic remodeling of the pressure-overloaded heart suggests a transition from an inflammatory to a fibrotic response. Early inflammatory mediator upregulation may be triggered by neurohormonal pathways and ultimately results in activation of signals that suppress inflammation, while promoting matrix deposition. The detailed characterization of cardiac fibrosis due to aortic constriction in the mouse provides us with a great tool to investigate the mechanisms involved in fibrotic remodeling of the pressure-overloaded heart.
Supported by NIH R01 HL-76246 and HL-85440.