1 Institute for Experimental Medical Research
2 Department of Cardiothoracic Surgery, Ullevaal University Hospital, Oslo N-0407, Norway
3 Cardiovascular Division, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts 02215
4 Research Forum, Ullevaal University Hospital, Oslo N-0407, Norway
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
extracellular matrix; gene expression; hypertrophy
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The signaling pathways that initiate growth and remodeling have become a focus of great interest. Elucidating these is thought to lead to the development of substances specifically aimed at preventing the development of pathological remodeling. Mechanical stimuli, circulating hormones, as well as autocrine and paracrine factors are possibly involved in initiating myocardial growth. Transduction of mechanical stimuli may involve molecules that bind extracellular matrix molecules to the cytoskeleton, but the specific molecules involved in mechanotransduction have not yet been identified. Moreover, autocrine and paracrine factors, as well as intracellular signal regulators involved in initiating cardiac remodeling are only beginning to be unraveled.
The aim of our study was to identify essential genes involved in the early phase of cardiac remodeling. We (11) and others (10) have previously shown development of cardiac hypertrophy already 1 wk after induction of MI in the mouse. In the present study the gene expression in noninfarcted left ventricular tissue from mice 1 wk after MI was compared with data obtained from sham-operated mice. A cDNA filter hybridization array containing 1,176 genes, representing a wide variety of signaling molecules, was used. Differentially expressed genes were confirmed by Northern blotting. Our study demonstrates differential expression of a large number of extracellular matrix-associated genes and resulted in the identification of a subset of genes, including syndecan-1, -2, -3, and -4, WT-1, fibronectin, collagen 6A, and basic fibroblast growth factor (FGF) receptor 1. Western blot analysis showed increased protein levels of syndecan-3 and -4. Since the syndecans link the cytoskeleton to the extracellular matrix and function as required coreceptors for FGF, we suggest that they are involved in initiating cardiac remodeling.
![]() |
METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
One week after primary surgery the animals were anesthetized before the hearts were removed, blotted dry, dissected into left ventricular noninfarcted area, infarcted area, and right ventricle, and weighed. The tissue was then immediately snap frozen in liquid nitrogen, or fixed overnight in 4% formalin if intended for histological analysis. The investigation conforms with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH publication No. 85-23, revised 1996).
Histological analysis.
Hearts fixed in formalin were processed into paraffin wax. Six 4-µm thick slices were cut along the short axis from the mid-ventricular level of each tissue sample. Slices were either stained with hematoxylin-eosin for morphology or van Gieson or Masons trichrome for identification of collagen. Tissue samples from four mice subjected to coronary artery ligation and three sham-operated mice were evaluated.
Immunohistochemistry.
The excised hearts were placed in ice-cold phosphate-buffered saline (PBS, pH 7.4), quickly embedded in Tissue-Tek OCT compound (Sakura Finetek, Torrance, CA), and frozen in dry ice-cooled isopropanol before storing at -70°C. Sections 5-µm thick were cut using the Cryostat 1720 from Leitz, Wetzlar, Germany, and collected on SuperFrost Plus slides (Menzel-Gläser, Braunschweig, Germany). Sections were fixed for 5 min in acetone at -20°C and air dried. After rinsing in Tris-buffered saline (TBS, pH 7.6) containing 0.05% Tween 20, sections were incubated with the Dako peroxidase blocking reagent (Dako, Glostrup, Denmark) for 15 min at 37°C followed by TBS rinse. Sections were then incubated for 30 min at 37°C with affinity-purified antibodies (rabbit) against collagen type I (1:100 dilution) and III (1:1,200 dilution) (Rockland, Gilbertsville, PA), respectively. After three washes in TBS the primary antibody was labeled for 30 min using the Dako EnVision+ system (peroxidase, rabbit, K4010) and 3,3'-diaminobenzidine tetrahydrochloride as the chromogen. Sections were counterstained with hematoxylin for 12 min, mounted, and examined under a Leitz Aristoplan microscope.
cDNA filter array analysis.
The procedure followed instructions in Atlas Pure Total RNA Labeling System provided by Clontech (Palo Alto, CA). In brief, noninfarcted left ventricular tissues from MI mice and left ventricular tissues from sham mice were homogenized, and total RNA was isolated and DNase I treated. The total RNA was added to biotinylated oligo(dT) magnetic beads. The cDNA was then synthesized by means of reverse transcription using MMLV reverse transcriptase and colorless [32P]ATP. Filter membranes (Atlas Mouse 1.2 Array, Clontech) were prehybridized in ExpressHyb (Clontech) with salmon testis DNA (Sigma D7665), before labeled cDNA was hybridized to the filter membranes overnight. Each pair of membranes was exposed on the same screen for 7 days and read with a phosphor imager (Fujifilm Bio-Imaging Analyzer BAS-1800; Fuji Photo Film, Tokyo, Japan).
Interpretation of array results.
The integrated signal intensities of the spots on the scanned filters were determined using AtlasImage 1.0a (Clontech). We performed a pair-wise comparison of signal intensities of filters made from MI hearts and sham hearts. A total of six experiments were included in the analysis. Filters were rejected if the background intensity was not homogenous. Spots were visually inspected to exclude artifacts and bleeding. The average background intensity of the membrane was subtracted from the signal intensity of each spot. Spots with intensities within 10% above background were regarded as not expressed. Spot intensities were normalized to the median intensity of all expressed spots on the membrane. The gene expression ratios for each experiment were ranked, and the frequency of a given gene present in defined percentiles for all experiments was used to identify differentially expressed genes. The binomial probability function was used to calculate the probability of obtaining false positives (type II error). For a data set of 1,176 genes, the genes with MI/sham ratios above the 85th percentile or below the 15th percentile in three or more experiments had a probability of type II error <0.048. Genes fulfilling these criteria were regarded as differentially expressed.
Northern blotting.
Northern blotting was performed as described earlier (34). In brief, mRNA extraction was carried out using Dynabeads oligo(dT)25 (Dynal, Oslo, Norway). Three micrograms of poly(A)+ RNA was size-fractioned on a formaldehyde-agarose gel and transferred to a nylon membrane (Nytran N transfer membrane; Schleicher and Schuell, Dassel, Germany). The membrane was subsequently hybridized with 32P-labeled cDNA probes. Autoradiography of the filters was performed in the same way as with the filter arrays. Densitometric analyses were carried out using Image Gauge V3.12 (Fuji Photo Film) software. The intensity of each band was normalized to its respective GAPDH signal. The mean signal for the sham hearts was set equal to 100. Differential expression found in MI hearts was expressed as mean ± SE percent difference compared with sham-operated animals.
The mouse cDNA probe for WT-1 was a kind gift from Andreas Schedl (School of Biochemistry and Genetics, University of Newcastle upon Tyne, Newcastle upon Tyne, UK), the fibronectin probe was from Richard Hynes (Howard Hughes Medical Institute, Centre for Cancer Research, Department of Biology, Massachusetts Institute of Technology, Cambridge, MA), the syndecan-1 and -4 probes were from Arie Horowitz (Dartmouth Medical Center, Hanover, NH), the syndecan-2 probe was from Guido David (Center for Human Genetics, Campus Gasthuisberg, Leuven, Belgium), the syndecan-3 probe was from Merton Bernfield (Division of Newborn Medicine, Harvard Medical School, Boston, MA), and the FGF receptor 1 probe was from Peter A. Cattini (Department of Physiology, University of Manitoba, Winnipeg, Manitoba, Canada).
Western blotting.
Whole tissue lysates were obtained from the noninfarcted region of mouse hearts or from sham-operated hearts by RIPA solution (Boston Bioproducts, Ashland, MA) containing protease inhibitor cocktail (Roche, Mannheim, Germany). The protein concentration was determined according to the Bradford method (Bio-Rad, Hercules, CA). Forty micrograms of total proteins were fractionated by 8% (for syndecan-3) or 12% (for syndecan-4) SDS-polyacrylamide gel electrophoresis and transferred to PVDF membranes (Millipore, Bedford, MA). Ponceau Red staining (Boston Bioproducts) was used to stain the membranes for 1 min to show loading of proteins. The membranes were incubated with primary antibodies overnight at 4°C. Syndecan-3 was detected by polyclonal rabbit anti-syndecan-3 antibody (1:1,000 dilution; a kind gift from Dr. Sarka Tumova, Neuroscience Center, University of Helsinki, Helsinki, Finland). Syndecan-4 was detected by a 1:3,000 dilution of polyclonal chicken anti-syndecan-4 antibody (Aves Laboratories, Tigard, OR; a kind gift from Dr. Arie Horowitz, Dartmouth Medical Center, Hanover, NH). Following the primary antibody incubation, the membranes were incubated for 1 h at room temperature with horseradish peroxidase-conjugated appropriate goat anti-rabbit IgG (1:4,000; Calbiochem, San Diego, CA) for syndecan-3 or goat anti-chicken IgY for syndecan-4 (1:4,000, Aves Laboratories). Immune complexes were visualized using the ECL detection system (Amersham Biosciences, Buckinghamshire, UK). The film images of syndecan-3, syndecan-4, and their respective Ponceau Red-stained blots were scanned and quantitated with ImageQuant software (Alpha Innotech, San Leandro, CA). The measurements for syndecan-3 and syndecan-4 were subsequently normalized against the respective Ponceau Red measurements. For syndecan-3 analysis five MI hearts and five sham hearts were used, and for syndecan-4 analysis six MI hearts and six sham hearts were used. For calculation of changes in percent, the mean value of the sham hearts was set equal to 100. Differential expression found in MI hearts was expressed as mean ± SE percent change compared with sham hearts.
Statistics.
All appropriate data are presented as means ± SE. Simple comparisons between groups used for Northern and Western blot analysis were made using the Mann-Whitney Rank Sum Test based on raw data obtained by densitometry from the blots.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
|
Array analysis of gene expression.
Of a total of 1,176, we found that 641 genes were expressed in our cDNA filter arrays. Twenty-three genes were, according to the criteria described in the METHODS, significantly upregulated in noninfarcted left ventricular myocardium compared with ventricular tissue from sham-operated mice, and five genes were only expressed after MI. Thirteen genes demonstrated decreased expression (Table 2).
|
Furthermore, we found FGF receptor 1 to be upregulated after MI. Its required coreceptor syndecan-3, one of a family of four transmembranous heparan sulfate proteoglycans, was also upregulated. Syndecan-1 and -2 showed a trend toward upregulation after MI, but the changes did not reach the level of significance according to our strict criteria for significant upregulation. Syndecan-4 was not present on our filter array. Together with syndecan-3 we found a coordinate upregulation of WT-1, which is known to be a transcriptional factor for syndecans.
Northern blot analysis of gene expression.
In accordance with our array results, Fig. 2A clearly depicts an increase in expression of fibronectin in noninfarcted left ventricular tissue from MI mice compared with left ventricular tissue from sham-operated mice. The blot for FGF receptor 1 also shows an increased expression in noninfarcted left ventricular tissue (Fig. 2B). Densitometric assessment of fibronectin and FGF receptor 1 showed a mean increase of 262 ± 41% (P < 0.05) and 69 ± 11% (P < 0.05), respectively.
|
Results for syndecan-3 were also in accordance with the array results and showed a mean increase in expression of 45 ± 16% (P < 0.05) in MI hearts (Fig. 2D). Moreover, syndecan-1 and -4 both showed increased expression in MI hearts compared with sham hearts. As expected the syndecan-1 probe hybridized to two distinct bands, corresponding to 3.4 and 2.6 kb, resulting from processing at alternate polyadenylation sites (8). An increased expression of 90 ± 61% (P < 0.05) and 160 ± 28% (P < 0.05) for the two syndecan-1 bands, respectively (Fig. 2H), and 30 ± 9% (P < 0.05) for syndecan-4 (Fig. 2E) was calculated. The syndecan-2 probe identified three distinct bands corresponding to 3.5, 2.4, and 1.2 kb, known to be produced by differential polyadenylation (22). All three bands depicted an increased signal for MI hearts of 27 ± 9% (P < 0.05), 34 ± 11% (P < 0.05), and 77 ± 31% (P < 0.05), respectively (Fig. 2C).
Western blot analysis of protein levels.
The protein levels of syndecan-3 and -4 were significantly increased in the noninfarcted region after MI (Fig. 3). Immunoblot analysis for syndecan-4 showed a statistically significant increase of 64% (P < 0.05) in MI hearts compared with sham controls. For syndecan-3 we found a 72% increase (P < 0.05).
|
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The cDNA filter array strategy provides an efficient means of simultaneously analyzing a large number of RNA transcripts. We have found that repeated array analysis and stringent criteria for defining genes as differentially expressed are essential for any deduction based on array results. We were able to confirm all upregulated genes tested, by means of Northern blotting, indicating a high degree of reliability of our array and data analysis.
One striking feature of our study was the high relative abundance of extracellular matrix-associated genes showing regulated expression. Fibronectin, fibrillin, osteopontin, biglycan, and collagen, all known important proteins of the extracellular matrix, demonstrated either increased expression or induction following MI. This suggests a marked remodeling and de novo deposition of extracellular matrix during the early phase following MI. Our histological findings indicate the same. Collagen staining of sections of the left ventricle, as well as immunohistochemical detection of collagen type I and III, showed increased deposition of extracellular matrix proteins in noninfarcted left ventricular tissue. The observed upregulation of FGF receptor 1, angiotensin converting enzyme, and tumor necrosis factor- receptor 2 may influence this deposition (18, 26, 32). There is also an activation of the collagen degrading system parallel to the increased extracellular matrix deposition in noninfarcted myocardium (35). In line with this notion, we found an increase in the mRNA abundance for the metalloproteinase gelatinase A, indicating a change in quality as well as quantity of the extracellular matrix.
Altered extracellular matrix composition in the noninfarcted region of the left ventricle, which in part results from increased expression of genes encoding extracellular matrix proteins, may have profound effects on the diastolic properties of the myocardium by affecting its elasticity. Increased stiffness of the left ventricle may contribute to the characteristic abnormalities of diastolic function, including reduced passive left ventricular filling (12, 21). The extracellular matrix remodeling process has also been associated with degradation of matrix components and depletion of collagen. This process may contribute to left ventricular dilatation (37).
BMP-1 is one of a family of bone morphogenetic proteins first isolated from osteogenic material and shown able to initiate ectopic bone growth (40). Interestingly, our histological examination of left ventricular tissue 8 wk after ligation of the left coronary artery in the rat has even revealed a transformation of papillary musculature into osteogenic material (data not shown). It has become increasingly clear that BMP-1 differs considerably from the other bone morphogenetic proteins and that the task of BMP-1 might be to activate latent forms of the bone morphogenetic proteins (40, 29). Especially BMP-2 and BMP-4 have been thought to function as growth factors, stimulating the proliferation and orchestrating the deposition of mesenchymal cells (9). In addition to this, it is known that BMP-1 processes procollagen into collagen (14) and probiglycan into biglycan (27). This means that BMP-1 potentially possesses a dual role that could make it important in the remodeling of the extracellular matrix following MI.
Osteopontin is an adhesive glycophosphoprotein that was first identified in bone tissue (7). We observed induction of osteopontin expression in the noninfarcted region. It is known that osteopontin interacts with integrins and with extracellular matrix proteins such as collagen and fibronectin, and thus a role for this protein in extracellular matrix remodeling in myocardial failure has been suggested (31). Since osteopontin is also involved in calcification (23), its induction might have contributed to the accumulation of osteogenic material previously mentioned.
The expression of ß1-integrin was also found to be induced in the noninfarcted region following MI in our study. It is known to be localized in costameres (43) and intercalated discs of cardiomyocytes (36). The integrins have been shown to have several important functions in cell-extracellular matrix adhesion, and may act as mechanotransducers. Mice with cardiac-specific inactivation of ß1-integrin might therefore develop dilated cardiomyopathy because of a reduced cell-extracellular matrix linkage ability (28). The increased expression of integrins in the noninfarcted region may indicate that they have a role as mechanotransducers in the remodeling response.
Also, the expression of fibrillin, a constituent of the extracellular matrix, was induced. Fibrillins form the structural framework of an essential class of extracellular microfibrils that endow dynamic connective tissues with long-range elasticity (25). Their importance is emphasized by the linkage of fibrillin mutations to Marfan syndrome and related connective tissue disorders that are associated with severe cardiovascular, ocular, and skeletal defects (24).
Increased expression of the syndecan family following MI indicates a role for these transmembranous proteoglycans in the postischemic heart. The expression of syndecan-3 was significantly increased after MI on our arrays, and Northern blotting showed an upregulation of all four members of the family. Western blotting furthermore, showed a 64% increase in syndecan-3 and a 72% increase in syndecan-4 protein levels, which were found to be statistically significant. We chose to measure the protein levels of syndecan-3 because of its relatively strong mRNA expression in cardiac tissue and of syndecan-4 because we consider it interesting in the context of myocardial remodeling, particularly since it is found in cardiomyocytes. The increases in the transmembranous proteins syndecan-3 and syndecan-4 were most likely higher in the membrane of syndecan producing cells than measured in whole tissue. Such upregulation of the syndecans in the noninfarcted area after MI has not previously been reported. However, one study (19) showed that a rapid increase in syndecan-1 and -4 gene expression occurred in the ischemic region of the heart, and recently plasma levels of syndecan-4 were found to be elevated in patients with acute MI (16).
The mechanisms responsible for regulation of syndecan gene expression are not known in detail. However, it has been shown that factors secreted by inflammatory cells, such as tumor necrosis factor- (13) and PR-39 (6), affect the expression of syndecans. Moreover, both monocytes and tissue-resident macrophages are known to express syndecan-1 (42) and syndecan-4 (41) core proteins. Presently, the relative contribution of inflammatory cells and other cells with regard to altered syndecan gene expression in the noninfarcted region is uncertain. However, in a study examining the peri-infarct region (19), it was found that although infiltrating macrophages accounted for a substantial increase in syndecan-1 and -4 expression, increased expression was also noted in the levels of syndecan-1 mRNA in endothelial cells and syndecan-4 mRNA in cardiac myocytes.
All four syndecans are expressed in fibroblasts (15), whereas only syndecan-4 has so far been identified in cardiomyocytes (19). The syndecans are implicated in several signal transduction cascades that regulate cell proliferation, and they are also at the cross section of mechanical signaling, interacting with molecules thought to be involved in mechanotransduction such as integrins (39). The syndecans bind growth factors and extracellular matrix molecules via their extracellular glycosaminoglycan chains (1). There is evidence that the syndecans have a role in focal adhesion formation, and the localization of syndecan-4 to focal adhesions in a variety of cell types (39) further supports syndecan facilitation of organized cell-matrix interaction. Via their small cytoplasmic domain, the syndecans interact with the cytoskeleton and potential downstream signal transducers. Syndecan-4 has been shown to interact with and activate PKC (30), which has previously been shown to be involved in myocyte growth (33). In addition to PKC
, the syndecan-4 V region binds phosphatidylinositol 4,5-bisphosphate (PIP2), which itself is capable of activating PKC
, and potentiates PKC
activity induced by PIP2 10-fold (38). The syndecans are known to bind to FGF, for which they are required coreceptors, contributing to cell proliferation (2). Furthermore, the syndecans enable the binding of cytoskeletal actin filaments to two of the extracellular matrix proteins (2) that we found upregulated in heart failure, namely, fibronectin and type VI collagen. The increased expression of all members of the syndecan family together with upregulation of genes encoding interacting proteins (ß1-integrin, collagen 6A, fibronectin, and FGF receptor 1) increases, in our opinion, the possibility that these play a biological role in the heart following MI. Lastly, a transcription factor for syndecans, WT-1 (4), also demonstrated increased expression. Strikingly, WT-1 knockout mice have been shown to develop dilated cardiomyopathy (17), possibly indicating a role for WT-1 in cardiac remodeling through increasing the transcription of syndecans.
A relatively moderate increase in syndecan expression has been shown to have biological effects on fibroblast migration (20). Moreover, it has been reported that mice heterozygous for a disrupted syndecan-4 gene had statistically significant delayed wound healing and impaired angiogenesis in the granulation tissue compared with wild-type littermates (5). These studies indicate that changes in syndecan expression observed in vivo have effects on biological processes related to those seen in the heart during postischemic remodeling. However, few studies have addressed this particular issue, and further work is needed to assess the role of the syndecans in the myocardium following MI.
In conclusion, this study shows that there is an upregulation of a large number of extracellular matrix-associated genes in the noninfarcted region following MI in the mouse, indicating the importance of remodeling of the extracellular matrix in the early phase after MI. The observed upregulation of BMP-1 may have a particularly important role in this context, since it processes procollagen into collagen and activates latent growth factors. Moreover, we have identified a subset of genes with increased expression following MI, including syndecans 14, WT-1, fibronectin, collagen 6A, and FGF receptor 1. Western blotting showed significantly increased protein levels of syndecan-3 and -4. Since the syndecans link the cytoskeleton to the extracellular matrix and function as required coreceptors for FGF, we suggest a role for the syndecans in cardiac remodeling following MI.
![]() |
ACKNOWLEDGMENTS |
---|
This work was supported by Anders Jahres Fund for Promotion of Science, Professor Carl Sembs Medical Research Fund, Rakel and Otto Bruuns Fund, and the Norwegian Research Council.
![]() |
FOOTNOTES |
---|
Address for reprint requests and other correspondence: A. Vanessa Finsen, Institute for Experimental Medical Research, Ullevaal Univ. Hospital, Kirkeveien 166, Oslo, Norway (E-mail: alexandra.finsen{at}ioks.uio.no).
10.1152/physiolgenomics.00144.2002.
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|