Author + information
- Received June 26, 2003
- Revision received November 13, 2003
- Accepted November 14, 2003
- Published online May 5, 2004.
- Alan R Collins, PhD*,
- Janet Schnee, MD*,
- Wei Wang, MD*,
- Sarah Kim, BS*,
- Michael C Fishbein, MD†,
- Dennis Bruemmer, MD*,
- Ronald E Law, PhD*,
- Susanne Nicholas, MD, PhD*,
- Robert S Ross, MD, FACC‡ and
- Willa A Hsueh, MD*,* ()
- ↵*Reprint requests and correspondence:
Dr. Willa A. Hsueh, University of California, Los Angeles, School of Medicine, Division of Endocrinology, Diabetes and Hypertension, Warren Hall, 900 Veteran Avenue, Suite 24-130, Los Angeles, California 90095-7073, USA.
Objectives Osteopontin (OPN) is upregulated in left ventricular hypertrophy and is stimulated by angiotensin II (AngII). Our objective was to determine whether mice deficient in OPN would be protected from AngII-induced cardiac fibrosis.
Background Interstitial fibrosis can lead to myocardial dysfunction and ultimately heart failure. Osteopontin activates integrins that regulate cell adhesion, migration, and growth, thus implicating OPN in the process of cardiac fibrosis.
Methods Osteopontin null (OPN−/−) mice (n = 18) and wild-type controls (n = 20) were infused with AngII (2.5 or 3.0 μg/kg/min) for four days or three weeks via osmotic mini-pumps. Hearts were assessed morphometrically and histologically, including quantitative assessment of fibrosis via optical microscopic imaging analysis. Cardiac fibroblasts derived from these mice were evaluated for adhesion and proliferation. Cardiac transcript expression for cytokines, extracellular matrix (ECM), integrin, and atrial natriuretic peptide were assessed.
Results Osteopontin−/−mice exhibited less cardiac fibrosis (0.7%) than wild-type mice (8.0%) (p < 0.01) and lowered heart/body weight ratios (0.10% vs. 0.23%) (p < 0.01) after three weeks of AngII infusion. Expression of transforming growth factor-beta, fibronectin, and collagen was not different between OPN−/−and wild-type mice, despite the decrease in ECM accumulation in the OPN−/−mice. Adhesion to ECM substrates decreased by 30% to 50% in cardiac fibroblasts of OPN−/−mice but was restored in OPN−/−cells by the addition of recombinant osteopontin.
Conclusions Osteopontin mediates cardiac fibrosis, probably through the modulation of cellular adhesion and proliferation. Because OPN is increased in cardiac hypertrophy and its lack attenuates fibrosis, understanding of OPN function is essential to extend our knowledge about molecular determinants of cardiac hypertrophy and failure.
Cardiac hypertrophy involves an increase in myocyte volume associated with a disproportionate increase in the volume of interstitial fibroblasts and extracellular matrix relative to that of myocytes (1). Little is known about fibroblast changes that occur during cardiac hypertrophy, but progressive fibrosis impairs myocyte contractility, oxygenation, and metabolism, thus contributing to ventricular dysfunction and perhaps ultimately heart failure (1,2). Osteopontin (OPN) is a large-acid phosphoprotein adhesion molecule that is secreted by both cardiac interstitial fibroblasts and myocytes (3–5)and is upregulated in left ventricular hypertrophy (LVH) and failure in humans and animal models (4–6). Osteopontin contains the arginine-glycine-aspartate tripeptide integrin binding motif and, by acting as an integrin ligand, OPN can activate cell signaling pathways and gene expression and thereby regulate cell differentiation and function (2,6). We found that OPN expression in the heart is potently regulated by angiotensin II (AngII) and that OPN is an important factor controlling cardiac fibroblast growth, adhesion to extracellular matrix, and collagen gel contraction (4). Therefore, we hypothesized that OPN may be an important cardiac profibrotic factor. The availability of mice with targeted disruption of the Sppl gene that encodes osteopontin null (OPN−/−) mice allowed us to test this hypothesis and to elucidate mechanisms involved in cardiac fibrosis (7).
Osteopontin knockout mice on a Black Swiss background breeding pairs were the generous gift of R. Johnson and C. Giachelli with the permission of L. Liaw (7). Osteopontin+/+littermate controls were used throughout the study. For AngII stimulation, adult mice weighing 18 to 22 g were anesthetized with isoflurane. Osmotic mini-pumps (Alza Corp., Palo Alto, California) containing either AngII (2.5 or 3 μg/kg/min) in phosphate-buffered saline (PBS), or PBS vehicle alone, were implanted subcutaneously for either four days (acute) or three weeks (chronic).
Blood pressure (BP) was measured by indirect tail cuff plethysmography (Visitech Systems, Apex, North Carolina). At least 10 measurements were obtained at each time point and the mean and standard deviation were recorded. Animals were habituated to the apparatus by BP measurement during the week before initiation of the experimental protocol. Blood pressures were obtained daily until a plateau level was attained (four days) and then weekly for the duration of the study.
After sacrifice by pentobarbital overdose and removal of the osmotic mini-pump, the animal was weighed. The heart was next removed, weighed, and was either snap frozen in liquid nitrogen and maintained at −70°C or prepared for histologic analysis.
Quantification of interstitial fibrosis
Coronal sections of the heart were made at the equator of the ventricles and the tissue fixed in 10% paraformaldehyde in PBS. Paraffin sections (4 μm) were used for Mallory's trichrome staining. Quantification of interstitial fibrosis was performed using digital microscopic analysis. The microscopic image was displayed on a high-resolution monitor and digitized by a video frame grabber (PCVISION Plus, Imaging Technology, St. Laurent, Quebec, Canada) running on an IBM-compatible computer. A morphometric analysis program (Image Pro, Media Cybernetics, Silver Spring, Maryland) was used to determine the area of fibrosis, defined as the area of blue trichrome stained fibers relative to the entire specimen. Contiguous high-power fields comprising an entire LV section for each sample were analyzed.
Analysis of gene expression from heart tissue
Total ribonucleic acid was isolated from tissue using the single-step acid guanidinium thiocyanate-phenol-chloroform extraction method of Chomczynski and Sacchi (8)as modified by Gibco-BRL using Trizol reagent (Gibco-BRL, Rockville, Maryland). Ribonucleic acid was size fractionated by electrophoresis through a denaturing agarose gel and transferred to nitrocellulose. [32P]-dCTP–labeled c-deoxyribonucleic acid probe was then allowed to hybridize to the Northern blot. Hybridization signals were normalized to signal obtained from Cricetulus griscus ribosomal protein S2 (clone CHO-B Gene Bank accession number L22552), a constitutively expressed gene, to correct for any variation in loading and transfer. Quantification of Northern blots was performed using a scanner with densitometric software (Scion Image, Frederick, Maryland). The complementary deoxyribonucleic acid probes for known markers of cardiac hypertrophy including cardiac ankyrin repeat protein (CARP), atrial natriuretic peptide (ANP), and OPN were used, as well as probes for the profibrotic factor, transforming growth factor (TGF)-beta, the ECM protein genes fibronectin and collagen I, and beta1 integrin.
Cardiac fibroblast cell culture and treatment
Murine cardiac fibroblast cultures were prepared from neonatal OPN−/−and WT mouse pup hearts (4). In brief, one- to four-day-old mouse pup hearts were washed, minced, and subjected to digestion with 4 μg/ml collagenase type II (Worthington, Lakewood, New Jersey) and 2 μg/ml dispase (Boehringer-Mannheim, Indianapolis, Indiana). After 20 to 30 min incubation at 37°C, the supernatant was collected, and undigested tissue was subjected to a second incubation in the digestive cocktail. After three rounds of incubation, the cells from the three collected supernatants were pooled, washed, and preplated for 60 min in DMEM/F12 supplemented with 5% fetal bovine serum. The adherent fibroblasts were then grown until confluent in DMEM/F12 supplemented with 10% fetal bovine serum. The cells were passaged at 4 × 105cells/10-cm plate and experiments were performed in the first or second passage after starvation in serum-free DMEM/F12 media supplemented with insulin (5 μg/ml), transferrin (5 μg/ml), and selenium (5 ng/ml) (ITS, Sigma, St. Louis, Missouri).
Adhesion and proliferation assays
Adhesion was measured to a variety of matrices including collagen I, fibronectin, laminin, and vitronectin at concentrations of 10 μg/ml. After 18 to 24 h of serum starvation, fibroblasts were treated with 20% fetal calf serum or left untreated. After a further 48 h incubation the cells were trypsinized, washed, and counted. They were then seeded at 10,000 cells/well in a 96-well plate that had been coated with test substrate and then blocked with BSA. After a 1-h incubation at 37°C the wells were washed and adherent cells were quantified by fixation and staining with toluidine blue, followed by determination of optical density at 595 nm. Bovine collagen I (Sigma), murine fibronectin (Calbiochem, La Jolla, California), murine laminin (Gibco BRL), and rat vitronectin (Sigma) were each used at 10 μg/ml in PBS for 1 h at room temperature.
Proliferation was measured by bromodeoxyuridine (BrdU) incorporation using florescent activated cell sorting. After 18 to 24 h of serum starvation, fibroblasts were treated with 20% fetal calf serum or left untreated. After a further 20-h incubation, BrdU was added to a final concentration of 60 mM and the incubation continued for an additional 2 h. Fibroblast proliferation rates were determined by analyzing incorporation of BrdU by flow cytometry with Fast Immune Anti-BrdU with Dnase (Becton Dickinson, Franklin Lakes, New Jersey) according to the manufacturer's directions. Data were analyzed with Facscan software.
Differences among means in BP and interstitial fibrosis between groups were performed using analysis of variance with the Student-Newman-Keuls test for differences among means. Data are presented as means ± SEM.
Blood pressure and markers of cardiac hypertrophy
Mice lacking OPN expression had lower baseline BP (101 ± 2 mm Hg, n = 18) compared to WT mice (116 ± 7 mm Hg, n = 20) (p < 0.05) (Fig. 1A). AngII infusion at 2.5 μg/kg/min effected similar changes from baseline BPs in both groups (55 mm Hg for WT and 48 mm Hg for OPN−/−mice (Fig. 1B). Despite the similar rise in BP, OPN−/−had lower absolute final BP (147 ± 3 mm Hg, n = 18) versus WT (171 ± 10 mm Hg, n = 20) (p < 0.05) (Fig. 1A). To confirm that our subsequent results were not related simply to this absolute BP difference, a separate group of OPN−/−animals (n = 10) were infused with a higher dose of AngII for 3 weeks (3 μg/kg/min), which increased BP to 185 ± 4 mm Hg, an absolute level even higher than the WT animals attained at the lower AngII dose, and one that produced a statistically increased change in BP from basal values, vs. the other two groups (Figs. 1A and 1B).
Heart morphometry was assessed in the mice after vehicle or AngII infusion. Despite lower basal BPs, OPN−/−mice (n = 18) had similar baseline heart/body weight ratio compared to WT (n = 20) (Fig. 2). Although AngII-mediated increases in BP were similar, the OPN−/−mice had a blunted hypertrophic response versus WT even when dosing was increased to 3.0 μg/kg/min, an infusion dose that caused an absolute BP level above that of WT infused mice.
To assess for early molecular changes during AngII-mediated stimulation, Northern blot analyses were performed at four days after initiation of infusion (when similar differences from basal pressure were detected between groups). We demonstrated an upregulation of both cardiac ankyrin repeat protein (CARP) and ANP (Fig. 3A), known markers of the ventricular hypertrophic response.
After three weeks of AngII infusion CARP expression returned to basal levels while ANP remained modestly elevated (Fig. 3B). Osteopontin expression was barely detected in sham-infused WT mice, but was markedly upregulated in WT mice after four days of AngII infusion. After three weeks of AngII infusion, OPN expression decreased in WT mice (Fig. 3C). No OPN was detected in hearts of OPN−/−mice before or after AngII infusion (data not shown). These results suggest that AngII-mediated stress modulates cardiac OPN expression, a response necessarily absent in OPN−/−animals.
Lack of OPN expression attenuates cardiac fibrosis
Because AngII infusion can modulate cardiac morphology, we compared the morphometry of OPN−/−and WT mice (Fig. 4). At baseline, no histologic differences were detected. After three weeks of AngII infusion, OPN−/−mice had minimal interstitial fibrosis as evidenced by Mallory's trichrome staining, whereas WT mice had marked interstitial fibrosis. On the basis of quantitative image analyses, the WT mice demonstrated a 10-fold (0.7 ± 0.1% to 7.6 ± 1.8%, n = 12) increase in interstitial fibrosis in response to AngII as compared to a more modest four-fold (0.4 ± 0.2% to 1.6 ± 0.7%) response in the OPN−/−mice, whether they were infused at 2.5 (n = 15) or 3 (n = 8) μg/kg/min AngII.
To evaluate for potential etiologies of this difference in the AngII-stimulated fibrotic response, we evaluated extracellular matrix and cytokine expression in the mice. As shown in Figure 5, transcripts for fibronectin, collagen I, beta-1 integrin, and the profibrotic cytokine TGF-beta increased to a similar extent in AngII-infused WT and OPN−/−hearts.
Cardiac fibroblast adhesion and growth
Another potential etiology for the altered fibrotic response of the OPN−/−mice may be altered cell adhesion or proliferation. Thus, we next evaluated for an alteration in adhesive capabilities of fibroblasts derived from the OPN−/−and WT mice. Cardiac fibroblast adhesion to a variety of ECM proteins including collagen I, fibronectin, laminin, and vitronectin was decreased in cells from OPN−/−mice compared to WT (Fig. 6). Overall, adhesion was decreased by 30% to 50% in OPN−/−cells compared to the OPN+/+fibroblasts. When OPN−/−fibroblasts were grown in the presence of recombinant OPN, adhesion was nearly restored to that of the WT controls (Fig. 6). These data show an important role of OPN−/−in cardiac fibroblast adhesion to ECM.
The BrDU incorporation was used to assess the proliferative responses of OPN−/−as compared to WT fibroblasts. Serum-stimulated growth was blunted in cardiac fibroblasts from OPN−/−compared to WT-derived fibroblasts (Fig. 7). Addition of recombinant OPN only partly restored the growth responses of OPN−/−fibroblasts to that of WT cells.
The present investigation demonstrates that OPN is a critical factor in the development of AngII-mediated cardiac fibrosis. Although AngII infusion resulted in similar increases in systolic BP and in ventricular markers of hypertrophy, OPN−/−mice demonstrated less increase in heart weight/ body weight ratio and four-fold less interstitial fibrosis than their WT counterparts. These histopathologic changes were associated with equivalent increases in TGF-beta, fibronectin, collagen I, and beta-1 integrin expression in WT and OPN−/−hearts, suggesting similar increases in ECM production. However, cardiac fibroblasts from neonatal OPN−/−mice demonstrated less proliferative responses and reduced adhesion to a variety of extracellular matrices, including collagen, fibronectin, laminin, and vitronectin compared to cells from neonatal WT mice. Thus, OPN appears to participate in growth and adhesion of cardiac fibroblast cells. Moreover, addition of recombinant OPN restored the adhesive properties of OPN−/−cardiac fibroblasts. There is a trend, but the difference was not statistically significant for recombinant OPN to restore growth in OPN−/−cells. Because an important mechanism by which OPN affects these functions is through activation of cell surface integrins, controlled interference with integrin activation may be one approach to reduce cardiac fibrosis and potentially prevent deterioration of ventricular performance in the face of deleterious hypertrophic stimuli.
Tissue remodeling in the heart is analogous to the wound healing response, with production of extracellular matrix, realignment of cells, and a restoration of normal amounts of matrix by enzymatic degradation by matrix metalloproteinases. Fibrosis results from excess accumulation of ECM. Angiotensin II can promote abnormal remodeling in cardiovascular and renal tissues because it promotes increased ECM production with a decrease in ECM degradation (1,5,9). In the heart AngII has multiple profibrotic actions: it stimulates cardiac fibroblast growth; increases expression of TGF-beta, a cytokine-stimulating collagen, and other ECM protein production; and increases several factors regulating cardiac fibroblast adhesion to ECM, including OPN and integrins (10,11). These effects are mediated by the AngII type 1 receptor (4,10,12,13), whereas the AngII type 2 receptor has been implicated in the inhibition of cardiac fibrosis (14,15)and mortality after experimental myocardial infarction in mice (16)and is upregulated in failing human hearts (17). Interestingly, the AT2 receptor has been detected on the cardiac fibroblast in humans (17)but has not been detected on mouse cardiac fibroblasts (11).
AngII also enhances tissue accumulation of ECM (16). Physiologic doses of AngII (10−10M) stimulate OPN production in both cardiac fibroblasts and in cardiac myocytes, suggesting that OPN may mediate some of these endogenous AngII effects (4,5). This conclusion is supported by the present study, in which genetic ablation of OPN attenuated cardiac fibroblast growth and adhesion and ultimately slowed the development of AngII-mediated cardiac interstitial fibrosis in vivo. These results suggest that though AngII stimulates TGF-beta and fibronectin expression in WT and OPN−/−mice, the decrease in cardiac fibroblast growth contributes to the decrease in fibrosis. These data demonstrate that OPN affects critical mechanisms that increase ECM accumulation and that OPN is an important cardiac profibrotic factor.
The parallel changes in ventricular ANP, CARP, and OPN in WT mice also support our conclusion that increased OPN expression is a molecular marker of LVH. When stressed, the myocyte response includes an increase in ANP production that limits myocyte and fibroblast growth, an alteration in phospholamban expression that improves cardiac contractility, and a change in the adult pattern of contractile protein gene expression to that of a fetal pattern (18–20). Indeed, in a variety of animal models of cardiac hypertrophy (aortic banding, Goldblatt hypertension, and the spontaneously hypertensive rat) and in humans, ventricular ANP expression correlated with OPN expression (4). The present study demonstrates that LVH in the mouse was also associated with upregulation of OPN expression, although unlike in the rat, expression was transient, as was the increased ANP and CARP expression. We have shown that both the cardiomyocyte and fibroblast are sources of OPN in the heart (5), but the effects of OPN on the myocyte are unknown. The OPN production by the myocyte may be a paracrine mechanism by which the myocyte communicates with neighboring fibroblasts to regulate their growth and fibrotic responses (6). Blocking antibodies against OPN inhibit rat cardiac fibroblast adhesion to matrix proteins, and thus growth, as adhesion is necessary for cell growth (4). Similarly, cardiac fibroblasts communicate with myocytes through expression and secretion of paracrine factors such as endothelin, an important regulator of cardiomyocyte hypertrophic responses (21).
Despite the upregulation of CARP and ANP, OPN−/−animals have a tendency to develop less increase in heart weight in response to AngII infusion compared to WT, in which the increase over sham infusion was significant. Even at higher levels of AngII infusion (3 μg/kg/min) with a marked increase in BP compared to sham-infused, there was only a tendency for an increase in heart weight that was not statistically significant. Whether this result was related to the decreased fibrosis and possibly decreased myocyte hypertrophy is unknown. Nevertheless, these observations suggest that OPN not only is a marker of the hypertrophic response to AngII and hypertension but also may mediate the response.
In agreement with our study in the heart, OPN−/−mice demonstrate altered remodeling responses in other tissues. For example, skin incision in OPN−/−mice was associated with more cell debris, greater disorganization of matrix, and smaller collagen fibril diameter in the wound than in WT mice (7). Osteopontin-null mice were also studied in a model of obstructive uropathy caused by ureteral ligation. Obstructive uropathy is associated with an increase in renal OPN expression, tubular atrophy, and increased interstitial inflammation and fibrosis (22). The uropathic OPN−/−mice had decreased renal macrophage levels, decreased TGF-beta expression, and less accumulation of collagens I and IV, resulting in less interstitial fibrosis compared to WT kidney (22). Likewise, after myocardial infarction, OPN−/−mice had increased left ventricular dilation and reduced collagen content compared with WT mice (23). The WT mice had a seven-fold increase in type I collagen in the infarcted region, whereas OPN−/−mice had decreased expression of collagen I messenger ribonucleic acid and no evident collagen accumulation at the infarct site (23).
Taken together, the present observations suggest that OPN may modulate the AngII-mediated fibrotic response and collagen accumulation in tissue injury, possibly as a result of alterations in cell proliferation and adhesion. Further studies are necessary to determine whether modulation of OPN is a beneficial strategy to control adverse cardiac remodeling.
☆ This study was supported by the following grants: NIH R01 HL66915 (to Dr. Hsueh) and HL57872 from the National Heart, Lung, and Blood Institute (to Dr. Ross). The first two authors contributed equally to this work. Dr. Karl T. Weber served as Guest Editor for this manuscript.
- angiotensin II
- atrial natriuretic protein
- blood pressure
- cardiac ankyrin repeat protein
- extracellular matrix
- left ventricular hypertrophy
- osteopontin null
- phosphate-buffered saline
- ribonucleic acid
- transforming growth factor
- wild type
- Received June 26, 2003.
- Revision received November 13, 2003.
- Accepted November 14, 2003.
- American College of Cardiology Foundation
- Liaw L,
- Almeida M,
- Hart C.E,
- Schwartz S.M,
- Giachelli C.M
- Graf K,
- Do Y.S,
- Ashizawa N,
- et al.
- Hsueh W.A,
- Law R.E,
- Do Y.S
- Gaedeke J, Peters H, Noble NA, Border WA. Angiotensin II, TGF-beta and renal fibrosis. Contrib Nephrol 2001:153–60.
- Kawano H,
- Do Y.S,
- Kawano Y,
- et al.
- Peng J,
- Gurantz D,
- Tran V,
- Cowling R.T,
- Greenberg B.H
- Ichihara S,
- Senbonmatsu T,
- Price E Jr..,
- Ichiki T,
- Gaffney F.A,
- Inagami T
- Tsutsumi Y,
- Matsubara H,
- Ohkubo N,
- et al.
- Shubeita H.E,
- McDonough P.M,
- Harris A.N,
- et al.
- Trueblood N.A,
- Xie Z,
- Communal C,
- et al.