SPARC mediates early extracellular matrix remodeling following myocardial infarction

Abstract

Secreted protein, acidic, and rich in cysteine (SPARC) is a matricellular protein that functions in the extracellular processing of newly synthesized collagen. Collagen deposition to form a scar is a key event following a myocardial infarction (MI). Because the roles of SPARC in the early post-MI setting have not been defined, we examined age-matched wild-type (WT; n=22) and SPARC-deficient (null; n=25) mice at day 3 post-MI. Day 0 WT (n=28) and null (n=20) mice served as controls. Infarct size was 52 ± 2% for WT and 47 ± 2% for SPARC null (P=NS), indicating that the MI injury was comparable in the two groups. By echocardiography, WT mice increased end-diastolic volumes from 45 ± 2 to 83 ± 5 μl (P < 0.05). SPARC null mice also increased end-diastolic volumes but to a lesser extent than WT (39 ± 3 to 63 ± 5 μl; P < 0.05 vs. day 0 controls and vs. WT day 3 MI). Ejection fraction fell post-MI in WT mice from 57 ± 2 to 19 ± 1%. The decrease in ejection fraction was attenuated in the absence of SPARC (65 ± 2 to 28 ± 2%). Fibroblasts isolated from SPARC null left ventricle (LV) showed differences in the expression of 22 genes encoding extracellular matrix and adhesion molecule genes, including fibronectin, connective tissue growth factor (CTGF; CCN2), matrix metalloproteinase-3 (MMP-3), and tissue inhibitor of metalloproteinase-2 (TIMP-2). The change in fibroblast gene expression levels was mirrored in tissue protein extracts for fibronectin, CTGF, and MMP-3 but not TIMP-2. Combined, the results of this study indicate that SPARC deletion preserves LV function at day 3 post-MI but may be detrimental for the long-term response due to impaired fibroblast activation.

matricellular proteins are extracellular matrix (ECM) components that contribute accessory, but not structural, roles to modify cell-ECM interactions (12). Secreted protein, acidic, and rich in cysteine (SPARC, osteonectin, BM40) is a collagen binding matricellular protein that is robustly expressed in fibroblasts and endothelial cells and at low levels in cardiac myocytes. In addition, it has also been reported that SPARC expression is associated with α-smooth muscle actin-positive myofibroblasts and CD45-positive leukocytes. SPARC stimulates cell signaling, adhesion, survival, proliferation, and migration in several cell types (4, 14). SPARC regulates postsynthetic procollagen processing and assembly into fibrils, providing a third layer of regulation for collagen deposition in addition to mechanisms involving transcription or degradation (14).

SPARC levels increase in hearts of rats subjected to β-adrenergic receptor stimulation (11) and in the left ventricles (LVs) of patients with LV hypertrophy (15). Bradshaw et al. (3) have reported that SPARC deletion attenuates pressure-overload induced collagen accumulation, resulting in improved diastolic function.

Schellings et al. (16) have shown that SPARC deletion results in a fourfold higher incidence of mortality following MI, due to increased rates of rupture and heart failure over the first 14 days. The increased incidence of rupture was associated with decreased deposition of mature collagen fibers. Further, an infusion of transforming growth factor-β (TGF-β) in SPARC null mice at day 2 pre-MI decreased rupture rates and increased collagen deposition. At the same time, TGF-β infusion did not alter infarct healing, suggesting that SPARC works through TGF-β-dependent and independent pathways. However, the roles of SPARC in the early post-MI setting have not been defined. In this study, we evaluated the early MI response in SPARC null mice to isolate critical SPARC-dependent functions in remodeling and to explore the hypothesis that SPARC deletion alters the cardiac fibroblast response to MI, which might contribute to a more complete understanding of factors contributing to the higher incidence of cardiac rupture associated with MI.

MATERIALS AND METHODS

All animal procedures were conducted according to the Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85–23, revised 1996) and were approved by the Institutional Animal Care and Use Committees at the Medical University of South Carolina and the University of Texas Health Science Center at San Antonio.

Mice.

We used male and female C57/BL6/SV129 wild-type (WT; n=23) and SPARC null (n=29) mice at 4–6 mo of age for the MI study. Day 0 control WT (n=28) and SPARC null (n=20) mice were used as naïve controls. To induce MI, the mice were anesthetized with 2% isoflurane and the left anterior descending coronary artery was permanently ligated using minimally invasive surgery as described previously (20). Because rupture predominantly occurs at days 3–7 post-MI in mice (7), we evaluated the 3-day post-MI time point.

At 3 days post-MI, the mice were killed under isoflurane and the heart and lung were removed. The LVs and right ventricles were divided and weighed individually. The LV was divided into apex, midcavity, and base, and the three LV sections and right ventricles were stained with 1% 2,3,5 triphenyltetrozolium chloride (Sigma) and photographed for infarct size determination by measuring infarct length as a percentage of the entire LV length (5). Infarct and remote regions were taken from the apex and base, respectively, and were individually snap frozen and stored at −80°C. The midcavity section was fixed in 10% zinc-formalin and paraffin-embedded for histological examination. The lungs were removed, and wet and dry weights were determined.

Echocardiographic measurements.

For the echocardiography analysis, 0.5–2% isoflurane in a 100% oxygen mix was used to anesthetize the mice. Electrocardiogram and heart rate were monitored using a surface electrocardiogram. Images were acquired using the Vevo 770 High-Resolution In Vivo Imaging System (Visual Sonics) and were taken at heart rates >400 beats/min to achieve physiologically relevant measurements. Measurements were taken from the two-dimensional parasternal long-axis and short axis (m-mode) recordings from the midpapillary region. Echocardiographic studies were performed before death of day 0 control mice and at day 0 and 3 days post-MI for mice in the MI groups. For each parameter, three images from consecutive cardiac cycles were measured and averaged. End-diastolic radius to free wall (infarct) thickness was calculated as an estimate of LV wall stress (1).

Histology.

The LV midcavity section was embedded in paraffin, sectioned at 5 μm, and stained using hematoxylin and eosin for routine assessment. Myocyte cross-sectional area was determined by measuring myocyte circumference of at ≥10 myocytes per section in the remote region of hematoxylin and eosin stained sections. From the circumference, areas were calculated. Macrophages were stained using a rat anti-Mac-3 monoclonal antibody (Cedarlane Laboratories; clone M3/84; 1:100 dilution) followed with the Vectastain Elite ABC kit (Vector Laboratories).

Fibroblast isolation and ECM arrays.

Cardiac fibroblasts were isolated by enzymatic digestion with liberase blendzyme 1 (Roche). Fibroblasts were incubated in DMEM with 10% FBS and 1% antibiotic-antimycotic solution (Cellgro; 30–004-CI). Cells were used at passages 2–4 to reduce phenotypic conversion due to prolonged culturing. To obtain basal ECM levels, fibroblasts were serum starved for 48 h. Fibroblasts were collected using TrypLE Express Cell Dissociation Reagent (Invitrogen; 12604) and snap frozen until analysis.

For the ECM arrays, total RNA was isolated from the cells using TRIzol reagent plus Total RNA purification kit (Invitrogen). The cDNA was synthesized using SABiosciences RT² first strand kit (C-03). The RT² qPCR Primer Array for Extracellular Matrix and Adhesion Molecules (SuperArray APMM-013A) was used for the gene array. This array uses SYBR Green-based quantitative real-time PCR assay to determine the gene expression of 84 ECM and adhesion molecule genes in one 96-well plate. Results were analyzed based on the ΔΔCt method with normalization of raw data to three housekeeper genes (Gusb, Hprt1, and Hsp90ab1) and were reported as 2^−ΔCt values.

Fibroblast ECM gene pattern analysis.

Fibroblast ECM gene array data collected from the six groups of mouse tissue described above were analyzed using Agilent Genespring GX 11.0 to identify differentially expressed genes and organize the genes into clusters. All data were combined as a single data set and loaded to Genespring GX 11.0. The data set (expressed as 2^−ΔCt units) was inputted into Genespring, without any normalization algorithm or baseline transformation. By default, Genespring converts the raw data into log scale if no normalization algorithm is selected. A total of 22 significantly differentially expressed genes was found, based on ANOVA analysis of the data set. Hierarchical clustering was performed with the Linkage set as Centroid, i.e, the distance between two clusters was calculated as the average distance between their respective centroids. All 22 differentially expressed genes were clustered in the dendrogram.

For pattern analysis, expression levels of the 22 identified differentially expressed genes were normalized with respect to the averaged expression levels of the WT control group for WT samples and the null day 0 group for null samples. Normalized expression was defined as expression levels in infarct region of the LV (LVI) and noninfarct region (LVC) divided by expression levels in the respective control. Normalized gene expression of LVC and LVI samples from WT tissues were grouped into six patterns: increased-decreased, increased-sustained, increased-increased, decreased-increased, decreased-sustained, and decreased-decreased. Expression levels of these genes from null samples were also analyzed to illustrate the effects of SPARC deletion.

LV protein extraction.

Soluble proteins were extracted from LVI and LVC by homogenizing the sample in PBS containing 1× complete protease inhibitor cocktail (Roche). After centrifugation, the soluble supernatant was collected. The insoluble proteins present in the pellet were extracted by further homogenization in Sigma reagent 4 (7 M urea, 2 M thiourea, 40 mM Trizma base, and the detergent 1% C7BzO) and 1× complete protease inhibitor cocktail (Roche). Protein concentrations were determined using the Bradford assay. Insoluble protein extracts were diluted 1:40 with water for Bradford assay compatibility. Total protein (10 μg) for each fraction of all samples were run on one-dimensional SDS gels, and the gels were stained with Coomassie blue to confirm protein concentration and loading accuracy.

Immunoblotting of LV tissue extracts.

Soluble or insoluble protein differences were determined by immunoblotting using antibodies against α-smooth muscle actin (Abcam; ab5694; 1:1,000), CD31 [platelet endothelial cell adhesion molecule-1 (PECAM-1); Abcam; ab28364; 1:1,000], connective tissue growth factor (R&D Systems; AF660; 1:1,000), fibronectin (Millipore; AB1954; 1:10,000), heat shock protein-47 (hsp-47; Epitomics; 3198; 1:1,000), matrix metalloproteinase-3 (MMP-3; Abcam; ab53015; 1:1,000), matrix metalloproteinase-9 (MMP-9; Abcam; ab38898; 1:1,000), periostin (Abcam; ab14014; 1:1,000), SPARC (R&D Systems; MAB942; 1:1,000), tissue inhibitor of metalloproteinase-2 (TIMP-2; Chemicon; AB8107; 1:1,000), and TGF-β (Sigma; AV37156, 1:1,000). Total protein (10 μg for each sample) was loaded on 26-well 4–12% Criterion Bis-Tris gels (Bio-Rad). Equal protein transfer was verified using reversible total membrane stain (Pierce; 24580) on the nitrocellulose membranes. Immunoblotting was performed as previously described (6). Molecular Imaging Software (Kodak) was used to measure densitometry, which was normalized to the mean intensity of the total membrane stain.

Statistical analyses.

Data are reported as means ± SE, and groups were analyzed using one-way ANOVA followed by Student-Newman-Keuls post hoc test. A P < 0.05 was considered significant.

RESULTS

Survival, echocardiographic, and necropsy analyses.

By immunoblotting, SPARC protein levels were increased to approximately twofold in the infarct region of WT mice at day 3 (Fig. 1). The 3-day post-MI survival rate was 96% for WT (n=1 mouse died) and 86% for SPARC null mice (n=4 mice died; P=0.37, by Fisher's exact t-test). One null death could be attributed to rupture, while one WT and three null deaths were attributed to arrhythmias or acute congestive heart failure.

Fig. 1.

Secreted protein, acidic, and rich in cysteine (SPARC) protein levels increase in the left ventricles (LV) at day 3 post-myocardial infarction (MI). Top: representative immunoblot images. Bottom: quantification of wild-type (WT) day 0 (n=12), WT day 3 remote (LVC; n=11), and WT day 3 infarct (LVI; n=11). Left: soluble fraction. Right: insoluble fraction. *P < 0.05 vs. control day 0. ^#P < 0.05 vs. remote LVC.

Echocardiographic, necropsy, and infarct size analyses for WT and null mice are shown in Table 1. Day 0 naïve WT and null mice were used as controls. SPARC null mice demonstrated reduced body weight, compared with age-matched WT mice at baseline. Both WT and null post-MI groups showed increased dilation compared with the respective day 0 controls; however, the increased dilation in the null mice post-MI was attenuated compared with the WT post-MI group. Likewise, ejection fraction decreased in both MI groups, with the decrease in ejection fraction being attenuated in the null post-MI mice compared with the WT post-MI group. The improved ejection fraction indicates that SPARC null mice showed better LV function at day 3 post-MI compared with the WT at day 3 post-MI.

Table 1.

Echocardiography and necropsy results

	Wild Type		SPARC Null
Echocardiography
	Day 0	Day 3 MI	Day 0	Day 3 MI
Sample sizes, n	28	22	20	25
Heart rate, beats/min	465 ± 9	494 ± 11	445 ± 9	452 ± 7‡
End-diastolic volume, ml	45 ± 2	83 ± 5*	39 ± 3	63 ± 5†‡
End-systolic volume, ml	20 ± 2	67 ± 4*	14 ± 2	46 ± 4†‡
Ejection fraction, %	57 ± 2	19 ± 1*	65 ± 2*	28 ± 2†‡
Stroke volume, ml	25 ± 1	16 ± 2*	25 ± 2	16 ± 1†
Free wall (infarct) wall thickness (systole), mm	1.05 ± 0.03	0.54 ± 0.04*	1.02 ± 0.05	0.56 ± 0.04†
End diastolic radius to infarct wall thickness, mm/mm	2.55 ± 0.05	4.34 ± 0.20*	2.33 ± 0.08	3.90 ± 0.21†
Necropsy
	Day 0 Control	Day 3 MI	Day 0 Control	Day 3 MI
BW, g	25.1 ± 0.7	22.7 ± 1.1*	20.1 ± 0.6*	18.0 ± 0.7‡
LV mass, mg	86 ± 3	102 ± 4*	67 ± 2*	86 ± 3†‡
LV to BW, mg/g	3.4 ± 0.1	4.8 ± 0.1*	3.4 ± 0.1	4.8 ± 0.1†
Lung wet weight to BW mg/g	5.0 ± 0.1	9.3 ± 0.7*	5.0 ± 0.2	8.3 ± 0.7†
Lung dry weight to BW, mg/g	1.1 ± 0.1	2.0 ± 0.2*	1.2 ± 0.1	1.8 ± 0.2
Infarct size, %		52 ± 2		47 ± 2

Data are means ± SE. SPARC, secreted protein, acidic, and rich in cysteine; LV, left ventricle; BW, body weight; MI, myocardial infarction.

^*P < 0.05 vs. wild-type day 0 control;

^†P < 0.05 vs. null day 0 control;

^‡P < 0.05 vs. wild-type day 3 MI.

The LV mass-to-body weight ratios increased in both WT and null post-MI groups, and there was no difference between the MI groups. The decrease in lung wet weight-to-body wet ratios in the null MI, compared with the WT MI, indicates that the null mice have less edema post-MI. Wall thinning and infarct sizes were similar between the two groups, indicating that functional responses were not due to initial differences in the severity of injury.

Morphometric analyses.

We measured the outside circumference and myocyte cross-sectional areas from hematoxylin and eosin stained sections (Fig. 2). The LVs in the WT MI group showed increased dilation compared with the null, consistent with the echocardiographic findings. Myocyte cross-sectional areas were 137 ± 5 μm² for WT day 0 (n=28) and increased to 169 ± 4 μm² for WT day 3 (n=22; P < 0.05). Similarly, myocyte areas were 132 ± 6 μm² for null day 0 (n=20) and increased to 163 ± 3 μm² for null day 3 (n=25; P < 0.05). Compensatory hypertrophic response appeared normal in null mice, consistent with findings from studies of pressure-overload-induced hypertrophy in SPARC null mice (3). Macrophages were quantified by Mac-3 immunohistochemistry. WT macrophage levels in the infarct region were 2.22 ± 0.17% (n=22), and null macrophage levels in the infarct region were 2.47 ± 0.15% (n=24; P=NS).

Fig. 2.

SPARC deletion attenuates LV dilation at day 3 post-MI. A: representative photomicrographs of hematoxylin and eosin stained sections from WT and SPARC null day 0 and day 3 post-MI left ventricles. B: quantification of the outside circumference, showing that absence of SPARC attenuates dilation at day 3 post-MI. Sample sizes are n=28 for WT day 0, n=22 for WT day 3, n=20 for null day 0, and n=25 for null day 3. *P < 0.05 vs. control (WT or null at day 0). ⁺P < 0.05 vs. WT day 3.

Fibroblast ECM gene array.

Excluding SPARC (which was absent in the null mice), there were 22 genes encoding ECM and adhesion molecules that were differentially expressed in fibroblasts isolated from day 0 and the remote and infarct regions of day 3 post-MI LV. Table 2 summarizes the 22 statistically significant changes. Of interest, fibroblasts isolated from the remote vs. infarct regions for either set showed similar expression patterns, indicating that fibroblasts from these two regions share similar ECM and adhesion molecule responses at the individual cell level.

Table 2.

ECM genes differentially expressed between WT and SPARC null cardiac fibroblasts

	Compared with WT Day 0		Compared with Null Day 0
Gene	WT LVC	WT LVI	Null Day 0	Null LVC	Null LVI
Adamts2				↓	↓
Cdh2				↓	↓
Vcam1	↓	↓	↓
Mmp3	↓	↓	↓
Sgce				↓	↓
Ctnnb1				↓	↓
Itgav			↑
Tgfb1				↑
Itgam	↑
Col4a1			↑	↓	↓
Itgb1			↑	↓	↓
Col4a2					↓
Fn1			↑	↓	↓
Timp2			↑	↓	↓
Adamts5				↓	↓
Lama2	↓	↓		↓	↓
Lamc1			↑	↓
Postn			↑	↓	↓
Itga3			↑	↓
Col3a1			↑
CCN 2 (CTGF)					↓
CD-31 (PECAM-1)	↓	↓	↓

WT, wild type; LVI and LVC, infarct and noninfarct region of the LV, respectively; CTGF, connective tissue growth factor; PECAM-1, platelet endothelial cell adhesion molecule-1. Arrows indicate direction of significantly different changes at P < 0.05 for the group listed.

In SPARC null day 0 fibroblasts ECM gene array, MMP-3, VCAM1, and CD31 (PECAM-1) levels were reduced, whereas Col3a1, Col4a1, Fn1, Lamc1, Postn, TIMP-2, and several integrins (α₃, α_v, and β₁) were elevated compared with gene expression levels in WT day 0 fibroblasts. We conclude that regulation of these genes is influenced by SPARC deletion.

We measured collagens Iα1, IIα1, IIIα1, IVα1, IVα2, IVα3, Vα1, and VIα1 in our screen. Of these collagen subtypes, collagens IIIα1 and IVα1 were the only collagens that were statistically different among groups. Collagen III mRNA levels were actually decreased, while collagen IV mRNA levels were increased, in WT post-MI fibroblasts. Both collagen III and collagen IV levels decreased in post-MI fibroblasts from the null mice. The fact that SPARC deletion was accompanied by a decrease in collagen expression may have a negative impact on scar formation at later time points.

By pathway analysis, six patterns of change emerged (Fig. 3B) from the fibroblast ECM gene array results from which the following conclusions were drawn: 1) SPARC regulates collagen production (particularly collagen III and IV) in addition to deposition; 2) SPARC effects the expression of other ECM genes, particularly fibronectin and periostin, which has not been previously reported in cardiac fibroblasts; 3) SPARC regulates different ECM genes in separate patterns, as we see six distinct patterns of change; and 4) CD31 (PECAM-1), collagen III, and CCN2 [connective tissue growth factor (CTGF)] are likely indirectly regulated by SPARC, as the pathway analysis showed these genes to be the least integrated in the cluster (Fig. 3A). Based on differential expression of genes found in the array analysis, we measured myocardial tissue levels of TIMP-2, fibronectin, CTGF, and MMP-3 by immunoblotting (Fig. 4).

Fig. 3.

A: clustered dendrogram display of gene levels from the cardiac fibroblast ECM gene array. Results were analyzed based on the ΔΔCt method with normalization of raw data to three housekeeper genes (Gusb, Hprt1, and Hsp90ab1) and are reported as the log ratios of the 2^−ΔCt values. Cardiac fibroblasts were isolated from day 0 controls and the remote (LVC) and infarct (LVI) regions of day 3 MI samples from both WT and null mice. Of the 84 extracellular matrix and adhesion molecule genes evaluated, 22 were statistically different among groups by ANOVA analysis, and these were used for the cluster analysis. Dendrogram and colored image were produced by Genespring GX 11.0. Color scale ranges from blue for log ratios −6.9 and below to red for log ratios 6.9 and above. Each gene is represented by a single row of colored boxes. Each tissue is represented by a single column. B: gene levels in LVC and LVI were normalized to day 0 control values and plotted according to 6 patterns. Top: WT levels for each section. Bottom: null levels. Sample sizes are n=4 for WT day 0, n=6 for WT day 3 remote LVC, n=3 for WT day 3 infarct LVI, n=7 for null day 0, n=6 for null day 3 remote LVC, and n=5 for null day 3 infarct LVI.

Fig. 4.

Myocardial levels of fibronectin, CCN2 [connective tissue growth factor (CTGF)], and matrix metalloproteinase-3 (MMP-3) by immunoblotting. Left: soluble fraction levels. Right: insoluble fraction level. A: soluble fibronectin levels increase 5-fold in the infarct region of WT LV, and this increase is attenuated in the null. Insoluble fibronectin levels also increase in the infarct region of the WT mice. Null mice show increased levels in the day 0 controls compared with the WT controls. B: soluble CCN2 (CTGF) levels increase in both the remote and infarct regions of WT LV, while CTGF levels in the null are highest at day 0 and in the remote region. Insoluble CTGF levels also increase in the infarct region of the WT mice, while the null mice show increased levels only in the day 0 control group. C: soluble active MMP-3 levels decrease in the infarct regions of both WT and null mice. Interestingly, active MMP-3 levels in the null are highest at the day 0 control time point. Insoluble active MMP-3 is reduced in the infarct region of the null mice compared with the null day 0 values. Sample sizes are n=12 for WT day 0, n=11 for WT day 3 LVC, n=11 for WT day 3 LVI, n=11 for null day 0, n=9 for null day 3 LVC, and n=9 for null day 3 LVI. *P < 0.05 vs. control (WT or null at day 0). ⁺P < 0.05 vs. WT at that same time point. ^#P < 0.05 vs. remote at that same time point.

Immunoblotting of post-MI LV tissue.

TIMP-2 protein levels in the LV tissue did not differ among any of the groups, which contrasted with the changes in mRNA levels seen in the remote and infarct fibroblasts. Normalized densitometry values for TIMP-2 were 119 ± 14 U for WT day 0 controls (n=12), 128 ± 10 U for the WT LVC remote region (n=11), 143 ± 13 U for the WT LVI infarct region (n=11), 125 ± 14 U for null day 0 controls (n=11), 105 ± 12 U for the null LVC remote region (n=9), and 95 ± 20 U for the null LVI infarct region (n=9; P=NS).

Soluble fibronectin levels were lower in SPARC null infarct regions, compared with WT infarct regions (Fig. 4A). While levels were significantly elevated in both infarct regions compared with the day 0 controls, the response was blunted in the absence of SPARC. Insoluble fibronectin was higher at baseline in null hearts, consistent with the increased gene expression detected in SPARC null fibroblasts over that of WT fibroblasts taken from day 0 hearts (Table 2 and Fig. 3). These results suggest that in addition to influencing expression and deposition of collagens, SPARC may also play a significant role in the regulation of fibronectin post-MI.

Levels of CTGF were higher in the remote region in SPARC null hearts post-MI, compared with levels in the remote region of WT hearts post-MI (Fig. 4B). Although levels of CTGF significantly and linearly increased in WT hearts post-MI, CTGF levels at day 0 and in the remote region of the null mice were found to be higher than those of WT. The fact that WT fibroblasts did not show an increase in CTGF gene expression post-MI suggests that the primary source of CTGF is another cell type present in vivo, for example, macrophages. In the insoluble fraction, CTGF levels were increased in the infarct region of WT but not null, LV. The lack of an increase in levels of CTGF in response to MI in SPARC null heart might contribute to subsequent ruptures observed at later time points post-MI.

MMP-3 levels decreased in WT and null infarct LV, consistent with the observed decrease in gene expression found in fibroblasts (Fig. 4C). In fibroblasts, MMP-3 gene levels were significantly lower in null day 0 compared with WT day 0 cells; however, samples from LV day 0 showed increased active MMP-3 levels in null vs. WT LV tissue. MMP-3 is an upstream activator of several MMPs, including MMP-9 (13). Therefore, baseline ECM turnover might be higher in the absence of SPARC, whereas the decrease in MMP-3 post-MI is predicted to favor a net accumulation of ECM at day 3.

Because SPARC has been shown to associate with MMP-9 and MMP-3 levels were found to be altered in cardiac fibroblasts and LV tissue of null mice, we also measured levels of MMP-9. MMP-9 was found to be increased in the infarct regions, but levels were not different between WT and null groups (P=NS).

To determine if SPARC deletion influenced fibroblast and endothelial responses post-MI, we measured α-smooth muscle actin, hsp-47, periostin, and TGF-β for fibroblast activation markers and CD31 (PECAM-1; for endothelial cell numbers). Levels of α-smooth muscle actin and TGF-β did not change among groups, indicating that the 3-day post-MI time point is before fulminant fibroblast activation. Levels of hsp-47 and CD31 were decreased in the infarct region of WT but not null mice (Fig. 5), suggesting that the early decrease in ECM and angiogenic responses are attenuated by SPARC deletion. Periostin levels increased from 9,044 ± 586 U in WT day 0 samples to 12,922 ± 585 and 15,191 ± 768 U in WT day 3 post-MI remote and infarct regions, respectively (both P < 0.05). Periostin similarly increased post-MI LV of SPARC null mice from 10,514 ± 712 U in day 0 controls to 14,009 ± 794 and 15,328 ± 759 U in null day 3 post-MI remote and infarct region, respectively (both P < 0.05). Although periostin levels trended higher in the SPARC null heart tissue, similar to findings in SPARC null fibroblasts, differences in periostin levels did not reach statistical significance.

Fig. 5.

Heat shock protein (HSP-47) and CD31 [platelet endothelial cell adhesion molecule-1 (PECAM-1)] levels decrease in the infarct region of WT but not SPARC null mice at day 3 post-MI by immunoblotting. A: soluble fraction of the LV extracts were blotted for Hsp-47. B: insoluble fraction of the LV extracts were blotted by CD31. For both, levels decreased in the infarct region of the WT LV, and this decrease was attenuated in the null LV extracts. Sample sizes are n=12 for WT day 0, n=11 for WT day 3 LVC, n=11 for WT day 3 LVI, n=11 for null day 0, n=9 for null day 3 LVC, and n=9 for null day 3 LVI. *P < 0.05 vs. control (WT or null at day 0). ^#P < 0.05 vs. remote at that same time point.

DISCUSSION

The goal of this study was to examine the role of SPARC in early remodeling events post-MI. SPARC has been shown to be a key cofactor in collagen assembly, and previous work (16) has shown the SPARC deletion results in increased rupture rates post-MI. We examined the functional consequences of SPARC gene deletion on the first 3 days of remodeling with respect to echocardiographic parameters and ECM synthesis in cardiac fibroblasts. The significant and unique findings of this study were that MI induction in SPARC null mice resulted in 1) improved functional remodeling parameters, including attenuated decreases in ejection fraction and less severe dilation; and 2) altered fibroblast phenotypes in terms of ECM and cell adhesion molecule expression. This is the first demonstration that SPARC regulates fibroblast function in the early post-MI setting.

In our study, the SPARC null mice showed improved remodeling parameters at day 3 post-MI. In the previous study by Schellings et al. (16), they observed no difference in LV function between the day 3 post-MI WT and SPARC null mice. There are two key differences in our experimental designs that may explain this difference. Our mice were on a mixed background strain (C57/BL6 and SV129) while the mice in the study of Schellings et al. were on a pure C57/BL6 background. The C57/BL6 strain has been shown to be a fibrosis-prone strain (10), and differences in rupture rates among strains have been reported (8, 18). In addition, the mice in our study were 4–6.5 mo of age (the equivalent of 30- to 45-yr-old humans) while the mice in the study of Schellings et al. (16) were 2.5–4.5 mo of age (the equivalent of 20- to 30-yr-old humans). The strain and age differences, therefore, could account for the slight differences in MI response seen in the SPARC null mice between the two studies. Additional studies comparing the effect of SPARC deletion on post-MI remodeling across species and ages are warranted. Trombetta and Bradshaw (17) recently reported that SPARC null mice show attenuated aging in periodontal ligaments, indicating that age superimposed on MI in the setting of SPARC deletion is likely to have a different outcome. It would be interesting to determine whether the absence of SPARC switches from a positive early regulator to a negative late regulator of remodeling, given that LV rupture is prevalent later. If so, the causes of this transition, while currently not clear, would potentially be very interesting. For instance, these studies may help to discern if cardiac function parameters early post-MI predict cardiac rupture or whether rupture is caused strictly by sudden structural failure.

SPARC deletion altered cardiac fibroblast phenotypes. Our results suggest that the effect of SPARC may be to alter the remodeling kinetics by changing cell response, perhaps in addition to or independent of a direct ECM role. In our study, we define impaired fibroblast activation as the decreased ability of the fibroblast to increase ECM production in the post-MI setting. The fibroblast CTGF results suggest that SPARC deletion takes CTGF out of the MI wound healing equation. SPARC deletion was compensated by increased CTGF levels at baseline, which suggests that collagen synthesis is higher at baseline and is confirmed by the higher gene expression of collagens IIIα1 and IVα1 in Table 2. In WT mice, CTGF was only needed post-MI, while in SPARC null mice CTGF was needed all of the time. One explanation for this decline may be that cardiac fibroblasts in SPARC null hearts are constantly trying to assemble more ECM as collagen levels are reduced in uninjured hearts of SPARC-null mice but are not successful. High levels of CTGF in day 0 and remote null fibroblasts suggest that collagen and CTGF may be important mediators of myocyte-fibroblast interactions (2, 9).

The switch in integrin isoforms in the remote region is interesting. The WT mice showed increased integrin α_m in the remote region post-MI, whereas the null mice showed decreased β₁ and decreased α_3. This result is consistent with the fact that macrophage numbers were not different between WT and null post-MI infarcts, which is consistent with results reported by Schellings et al. (16) for the day 7 time point. By day 14, Schellings et al. (16) do report a decreased number of macrophages in the nulls compared with the WT, indicating that SPARC may regulate macrophage viability and chronic immune responses at later times post-MI.

In addition to altering fibroblast responses, SPARC deletion likely also affects macrophage and endothelial cell responses, which would contribute to the attenuated remodeling phenotype observed. SPARC is expressed in macrophages and endothelial cells, and SPARC has been shown to interact with the scavenger receptor stabilin 1 to regulate macrophage clearance (19). Therefore, SPARC deletion likely influences the inflammatory and angiogenic responses in the MI setting. CD31 levels decreased at day 3 post-MI in the WT LV, but this decrease was attenuated in the absence of SPARC. This indicates that SPARC deletion serves to preserve blood vessel numbers in the infarct region, rather than stimulating an angiogenic response. Additional studies are needed to clarify the role of SPARC in these two cell types.

In conclusion, this study is the first to examine the role of SPARC on fibroblast ECM production in the post-MI setting. While SPARC deletion resulted in improved function at day 3 post-MI, the absence of SPARC also resulted in a blunted fibroblast ECM response. Differential timing of SPARC inhibition, therefore, is predicted to yield different outcomes in terms of infarct scar phenotype.

GRANTS

Support for this study was provided by National Institute of Biomedical Imaging and Bioengineering Grant 1R03-EB-009496 and National Heart, Lung, and Blood Institute Grant SC2-HL-101430 (to Y. Jin); National Heart, Lung, and Blood Institute Grants 2P01-HL-48788 and HL-094517 and a Veteran's Administration Merit Award (to A. D. Bradshaw); and National Heart, Lung, and Blood Institute Grant R01-HL-75360, the American Heart Association AHA 0855119F, and the Max and Minnie Tomerlin Voelcker Fund (to M. L. Lindsey). S. M. McCurdy was supported in 2008 by the American Physiological Society Summer Undergraduate Research Fellowship.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).