Tenascin-C (TN-C) is an extracellular matrix glycoprotein with high bioactivity. It is expressed at low levels in normal adult heart, but upregulated under pathological conditions, such as myocardial infarction (MI). Recently, we (Ref. 34) reported that MI patients with high serum levels of TN-C have a greater incidence of maladaptive cardiac remodeling and a worse prognosis. We hypothesized that TN-C may aggravate left ventricular remodeling. To examine the effects of TN-C, MI was induced by ligating coronary arteries of TN-C knockout (KO) mice under anesthesia and comparing them with sibling wild-type (WT) mice. In WT+MI mice, TN-C expression was upregulated at day 1, peaked at day 5, downregulated and disappeared by day 28, and the molecule was localized in the border zone between intact myocardium and infarct lesions. The morphometrically determined infarct size and survival rate on day 28 were comparable between the WT+MI and KO+MI groups. Echocardiography and hemodynamic analyses demonstrated left ventricular end-diastolic diameter, myocardial stiffness, and left ventricular end-diastolic pressure to be significantly increased in both WT+MI and KO+MI mice compared with sham-operated mice. However, end-diastolic pressure and dimension and myocardial stiffness of KO+MI were lower than those of the WT+MI mice. Histological examination revealed normal tissue healing, but interstitial fibrosis in the residual myocardium in peri-infarcted areas was significantly less pronounced in KO+MI mice than in WT+MI mice. TN-C may thus accelerate adverse ventricular remodeling, cardiac failure, and fibrosis in the residual myocardium after MI.
- extracellular matrix
left ventricular (lv) remodeling after myocardial infarction (MI) is a clinically important process, since it may result in dilatation, hypertrophy, and poor prognosis. It is well recognized that ventricular remodeling is accompanied by changes in the structure and composition of the myocardial extracellular matrix (ECM), and the significance of several molecules related to ECM turnover, especially matrix metalloproteinase (MMP)-2 and MMP-9, has received a great deal of attention (reviewed in Ref. 36).
Tenascins (TN) are a family of four multimeric ECM glycoproteins, each with distinct features; they are named TN-C, -X, -R, and -W (44). TN-C, which was found to be the first member of the family expressed during embryonic development, as well as in wound healing and cancer invasion in various tissues, may regulate cell behavior and matrix organization during tissue remodeling. In vitro studies suggest that TN-C controls the balance of cell adhesion and deadhesion, as well as cell motility, proliferation, differentiation, and survival (6).
TN-C is not expressed in healthy adult hearts, but is upregulated under many pathological conditions (9, 16, 18, 26, 28, 39), such as acute MI (15, 30, 34, 38, 45), and some cases of dilated cardiomyopathy (39, 43) associated with tissue injury and inflammation (16, 26, 28, 43). After MI, TN-C appears during the acute stage at the interface between infarcts and intact myocardium, where tissue remodeling most actively occurs (15). Our laboratory previously reported that TN-C may weaken the links between cardiomyocytes and connective tissue and accelerate the recruitment of myofibroblasts to injured sites; therefore, it might play an important role in myocardial tissue healing (15, 38).
Although TN-C molecules are deposited in extracellular spaces and regulate local cell behavior, soluble forms are released into the bloodstream. Interestingly, TN-C serum levels are elevated in acute MI patients and so provide a possible predictor of ventricular remodeling and poor prognosis after infarction (34). We and other groups have reported increased serum TN-C in patients who have suffered heart failure, reflecting ventricular remodeling and poor prognosis (10, 12, 25, 40); this suggests an adverse effect on the progression of ventricular remodeling. Conversely, upregulation of TN-C might reflect complementary responses to heart failure, as with the natriuretic peptide system, since various remodeling modulators, such as proinflammatory cytokines, growth factors, angiotensin II, hypoxia, reactive oxygen species, and acidosis, in addition to mechanical stretch, increase synthesis of TN-C by cardiac fibroblasts in vitro (5, 17, 28, 47).
To investigate whether TN-C exacerbates ventricular remodeling, we examined its effects by targeted deletion of TN-C in an experimental model of MI in mice. TN-C knockout (TNKO) mice were independently generated by two different groups and described as showing no distinct morphological phenotypes (8, 33). However, recent detailed studies have shown differences in specific cell behavior and response, particularly in various disease models, such as attenuated fibrotic change in immune-mediated hepatitis (7), allergic inflammation in bronchial asthma (27) and arthritis (24), reduced neointimal hyperplasia after vascular surgery (48), and delayed repair of articular cartilage injury (31). In the present study, we compared cardiac remodeling in TNKO and wild-type (WT) mice after permanent coronary ligation by echocardiographic and hemodynamic measurements and histological analysis.
MATERIALS AND METHODS
The investigation was performed in line with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (publication no. 85-23, revised 1996) and was approved by our Institutional Animal Research Committee. The original TNKO mouse (33) was backcrossed with BALB/c inbred mice for more than 10 generations. TN-C-null (−/−) male mice, aged between 8 and 10 wk old, were used in the experiments. WT (+/+) littermates were used as controls.
The MI model involved ligation of the left coronary artery of male TNKO mice (KO+MI group) and sibling WT mice (WT+MI group) under anesthesia, in accordance with the methods described by Michael et al. (23). A sham operation without coronary artery ligation was also performed for both WT (WT+sham) and TNKO (KO+sham) mice.
For survival analysis of the WT+MI (n = 35) and KO+MI (n = 30) mice, the cages were inspected daily to record the deaths of animals during the 4-wk study period.
Echocardiographic and hemodynamic measurements.
Four weeks after ligation, echocardiographic studies were performed, as previously described (11), on surviving WT+MI (n = 16), KO+MI (n = 16), WT+sham (n = 11), and KO+sham (n = 11) mice under light anesthesia with tribromoethanol-amylene hydrate [Avertin; 2.5% (wt/vol), 8 ml/g ip] and spontaneous respiration. After the echocardiographic measurements, LV pressure was measured in accordance with the methods described by Williams et al. (46). Two investigators (M. Ikeuchi and H. Matsusaka.), who were not provided with information about the experimental groups, performed in vivo LV function studies that included echocardiography and LV pressure measurements. Our laboratory's recent validation study demonstrated that intraobserver and interobserver variability with our echocardiographic measurements for LV cavity dimensions and fractional shortening were small, and measurements made in the same animals on separate days were highly reproducible, as previously described (11, 35).
To determine the myocardial stiffness constant, the σ-ln(1/H) value of the LV regional wall was evaluated with a spherical model of the ventricle to calculate mean wall stress (σ) using the equation σ = PD/4H, where P is LV pressure, D is LV short-axis diameter, and H is the mean wall thickness of LV anterior and posterior walls (13, 22). The diastolics = ln(1/H) data points from the point of minimal wall stress to the end-diastole point were fitted to a single exponential curve with zero asymptote σ = C exp [K ln(1/H)], where K is the myocardial stiffness constant.
At day 28, all surviving mice were killed after echocardiographic and hemodynamic examinations, and their hearts were removed for histological analysis. WT-MI mouse hearts were excised at days 1, 2, 3, 5, 7, and 14 after coronary ligation (n = 3 for each time point). KO-MI mouse hearts at day 5 (n = 3) were also excised for histological comparison. The entire LV from the apex to the base was cut into three transverse sections, fixed in 4% paraformaldehyde and embedded in paraffin. Sections were cut to a thickness of 3 μm.
Infarct size at day 28 was analyzed as described by Pfeffer et al. (32). Infarct length was measured along the endocardial and epicardial surfaces in each of the LV sections, and the values from the three sections were summed. Total LV circumference was calculated as the sum of endocardial and epicardial segment lengths from all LV sections. Infarct size (in percent) was calculated as total infarct circumference divided by total LV circumference. The reliability of this measurement was confirmed, as previously described (11).
To detect interstitial collagen fibers, picrosirius red staining was performed using a Scion imaging system, as previously described (28). Collagen volume fraction was measured in six fields of both border zones and remote areas of LVs for each heart in the MI and corresponding sham-operated groups.
Immunostaining of tissue sections was performed as previously described (15). In brief, after treatment with pepsin for 10 min for antigen retrieval, sections were incubated with anti-TN-C polyclonal rabbit antibody overnight at 4°C and subsequently with peroxidase-conjugated anti-rabbit IgG Fab′ (1:500 MBL, Nagoya, Japan) for 1 h. After washing, diaminobenzidine/H2O2 solution was used to demonstrate antibody binding.
LV myocardial tissues of WT+sham and WT+MI mice on 1, 3, 5, 7 and 14 days after coronary ligation were homogenized in Isogen (NipponGene, Toyama, Japan). Total RNA was isolated, and RT-PCR was performed as previously described (29). The forward and reverse primers for TN-C were 5′-GTTTGGAGACCGCAGAGAAGAA-3′ and 5′-TGTCCCCATATCTGCCCATCA-3′, respectively. The expected size of the PCR fragment was 344 bp. β-Actin was used as an internal control, with forward and reverse primers of 5′-GTGGGGCGCCCCAGGCACCA-3′ and 5′-CTCCTTAATGTCACGCACGATTTC-3′, respectively, and the expected size of the PCR fragment was 571 bp.
Western blot analysis.
Protein was extracted from homogenized tissue in Isogen in accordance with the manufacturer's instructions. Samples (10 μg/lane) were subjected to SDS-PAGE with 2–15% gradation polyacrylamide gel, transferred onto Immobilon membranes (Millipore, Bedford, MA), and immunostained with anti-TN-C antibody (1 μg/ml) using the indirect immunoperoxidase method, as previously described (15). Immunoreactivity was detected with the ECL system (Amersham, Arlington Heights, IL).
All data are expressed as means ± SE. Survival was analyzed by the Kaplan-Meier method. For between-group analysis, multiple-comparison tests were performed using one-way ANOVA followed by Bonferroni's post hoc test.
Expression of TN-C after MI.
In WT-MI mice, the level of TN-C mRNA was upregulated beginning at day 1 after injury, downregulated at day 7, and not detectable at day 14 (Fig. 1A). TN-C protein also became detectable at day 1, peaked at day 5, then was downregulated until disappearance at day 14 (Fig. 1B). Immunostaining for TN-C (Fig. 1C) was observed at the borders between intact myocardial tissues and necrotic areas at day 1. Strong staining for TN-C was detected in developing granulation tissue around days 3–5 and disappeared by day 28. No immunoreactivity was detected in remote areas at any stage.
Myocardial repair in TNKO mouse.
Myocardial healing associated with granulation tissue formation appeared to proceed normally in KO mice, and there was no obvious difference from WT mice on routine histological analysis (Fig. 2). No immunoreactivity for TN-C was observed in any tissue sections collected from TNKO mice.
Early operative mortality was comparable between KO and WT mice, and 15/35 (43%) WT+MI and 12/30 (40%) KO+MI mice died within 24 h after ligation. Between weeks 1 and 4 after surgery, four WT+MI (11.4%) and one KO+MI (3%) mice died of heart failure. One KO+MI mouse died of heart rupture at day 8, while no rupture was observed in the WT+MI group. At day 28, no significant difference in the survival rate was found between the WT and KO groups (49 vs. 57%; Fig. 3A).
Mice at day 28 after coronary ligation.
See Table 1. The surviving mice at day 28 were killed after echocardiographic and hemodynamic measurements. WT+MI lung-to-body weight ratios and right ventricle (RV)-to-body weight ratios of WT+MI were significantly increased compared with those of the WT+sham mice. No significant difference in lung-to-body weight ratios and RV-to-body weight ratios was found between the KO+sham and KO+MI group. RV-to-body weight ratios of KO+MI were significantly lower than in the WT+MI mice.
Infarct sizes determined by morphometric analysis 28 days after ligation were comparable (59 ± 9.3 vs. 55.5 ± 7.5%, P = not significant) between the WT+MI (n = 6) and KO+MI (n = 6) mice (Fig. 3B).
Cardiac function, stiffness, and remodeling.
See Table 2. There were no significant differences in LV size and function between sham-operated KO and WT mice. In WT+MI mice, LV end-diastolic diameter increased significantly, accompanied by a reduced ejection fraction compared with those in WT+sham mice. Myocardial stiffness and LV end-diastolic pressure also increased in WT+MI mice. In KO+MI mice, LV end-diastolic diameter and LV end-diastolic pressure showed similar significant increases compared with those of KO+sham mice (P < 0.01 and P < 0.05, respectively). However, end-diastolic pressure and dimension and myocardial stiffness were lower than in the WT+MI mice, suggesting that TN-C plays an important role in LV remodeling and stiffness.
Quantitative analysis of the percentage of interstitial collagen volume in the residual myocardium at border zones revealed a significant increase in the WT+MI mice compared with the WT+sham mice (11.3 ± 1.0 vs. 0.9 ± 0.4%, P < 0.01). In the KO+MI case, percentage of interstitial collagen volume was also greater than that in KO+sham mice (7.3 ± 2.3 vs. 1.2 ± 0.4%, P < 0.01), but was significantly reduced compared with that in WT+MI mice (P < 0.01). In remote zones, no significant differences in percentage of fibrotic area were found between either WT+sham and KO+sham mice (0.8 ± 0.3 vs. 1.1 ± 0.4%, P = not significant) or WT+MI and KO+MI mice (2.3 ± 1.4 vs. 2.0 ± 0.9%, P = not significant) (Fig. 4).
The present study clearly demonstrated that TN-C is involved in the progression of adverse ventricular remodeling after MI. Targeted deletion improved cardiac function and myocardial stiffness, in association with the reduction in interstitial fibrosis in the border zone myocardium, where TN-C expression was localized, while it did not appear to change in remote areas of the residual myocardium.
In the mouse MI model, both TN-C mRNA and TN-C protein became detectable at day 1 after coronary ligation, peaked at day 5, then were downregulated and disappeared by day 28, when infarcted areas were replaced by scar tissue. During the healing process, TN-C was only localized in the border zones and was not detected in the intact myocardium. This spatiotemporal expression pattern is comparable with previous results observed in an experimental rat model (15) and a human myocardial sample obtained at autopsy (45).
It has been proposed that TN-C plays an important role in tissue repair in various organs based on its specific expression that is closely associated with injury and inflammation. A number of studies have demonstrated that TN-C promotes proliferation and migration of parenchymal epithelial cells (37, 49) and angiogenesis (1). Furthermore, we previously reported that TN-C accelerates migration and differentiation of myofibroblasts (38), which play an important role in wound healing by synthesizing collagens and exerting strong contractive forces to promote wound healing. These functions expedite tissue healing, may protect against cardiac rupture, and prevent ventricular dilatation after MI.
On the other hand, TN-C may also promote adverse ventricular remodeling. For example, TN-C weakens the adhesion of cardiomyocytes to connective tissue (15) and upregulates expression and activity of MMP-2 and -9 (14, 20, 27, 42). Furthermore, there is a growing body of evidence suggesting that TN-C enhances inflammatory responses (24) with activation of NF-κB (27) and cytokine upregulation (7). Although these functions are useful for clearing damaged tissue and releasing residual cardiomyocytes from connective tissue for rearrangement, they might cause progressive degradation of ECMs and slippage of myocytes within the LV wall.
Moreover, a system that compensates for a lack of TN-C apparently exists (8, 18, 21, 33). The recruitment of myofibroblasts to injured sites of myocardium is delayed in TNKO mice, but approaches that of WT mice by day 3 after injury (38); however, the compensatory mechanism has not been identified. Therefore, it is difficult to theoretically determine whether TN-C is harmful or beneficial overall for tissue reconstruction after MI.
In the present study, during the acute stage, tissue healing in TNKO mice seemed to proceed normally with no difference in survival rate compared with that in WT mice. However, ventricular remodeling was significantly reduced in the TNKO mice, and cardiac function was improved at day 28 after coronary ligation. It seems that TN-C plays critical roles in the modulation of responses to myocardial injury and exerts harmful effects on the heart at later stages. The results correspond to our laboratory's recent clinical finding that patients with high serum levels of TN-C in the acute stage within 1 wk after infarction have a greater incidence of ventricular remodeling 6 mo later and a worse long-term prognosis (34).
Histologically, interstitial fibrosis of residual myocardium at day 28 in TNKO animals was markedly reduced compared with that in the WT mice in peri-infarct areas, where TN-C is normally deposited during the acute stage. Thus it is suggested that the deletion of TN-C may alleviate the progression of fibrosis, which could contribute to improvement of myocardial stiffness.
It has been proposed that TN-C promotes fibrosis because of its upregulated expression in various fibrogenic processes (2, 6, 19, 28). Direct support for this is the finding that locally applied TN-C accelerates the recruitment of myofibroblasts and collagen fiber formation in aneurysmal cavities in a rat model (41). Furthermore, deficiency of TN-C attenuated the expression of collagen I and III in an immune-mediated chronic hepatitis model (7) and deposition of proteoglycan in neointima after aortotomy in mice (48).
Although multiple molecular pathways are involved in the progression of ventricular remodeling and fibrosis, transforming growth factor (TGF)-β/Smad3 signaling may be a cascade of major importance (reviewed in Ref. 3). Loss of Smad3 results in a very similar phenotype to that evident in our TNKO mice, demonstrating reduced interstitial fibrosis and attenuation of remodeling associated with downregulation of TN-C (4). Since TGF-β is a strong inducer of TN-C in cardiac fibroblasts (26), TN-C may be a key molecule in the cascade of TGF-β/Smad3 signaling in ventricular remodeling.
In conclusion, TN-C may aggravate unfavorable responses during myocardial repair and could, therefore, be a target molecule for prevention of adverse ventricular remodeling after MI.
This study was supported in part by a Grant-in-aid for scientific research (no. 21590927) from the Ministry of Education, Culture, Sports, Science and Technology of Japan (to K. Imanaka-Yoshida), a research grant for intractable diseases from the Ministry of Health, Labor and Welfare of Japan (to K. Imanaka-Yoshida and M. Hiroe), and a grant from the Japan Foundation for the Promotion of the International Medical Center of Japan (to M. Hiroe and K. Imanaka-Yoshida).
No conflicts of interest are declared by the author(s).
The authors express deep gratitude to Dr. Shinobu Arai for helpful advice on the statistical analysis and to Akiyo Sekimoto, Mari Hara, and Miyuki Namikata for technical assistance.
- Copyright © 2010 the American Physiological Society