Elucidating the molecular mechanism of cardiac remodeling using a comparative genomic approach

Maria Mirotsou1, Coran M.H. Watanabe2, Peter G. Schultz2, Richard E. Pratt1 and Victor J. Dzau1

1 Cardiovascular Research, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts 02115
2 Department of Chemistry, The Scripps Research Institute, La Jolla, California 92037


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
It is proposed that analysis of global gene expression would provide an understanding of the molecular mechanisms of cardiac remodeling. However, previous studies have only provided "snapshots" of differential gene expression. Furthermore, the differences in gene expression between regions of the heart that can result in sampling variability have not been characterized. In this study, we employed the Affymetrix GeneChip technology to evaluate the patterns of expression in two different in vivo models of cardiac remodeling and in two different regions (left ventricle free wall and intraventricular septum) of the heart. Mice underwent transverse aortic constriction (TAC), myocardial infarction (MI), or sham operation, and RNA from the left ventricle free wall and the septum was isolated 1 wk later. Histological analysis showed profound myocyte hypertrophy and fibrosis in both the septum and the left ventricle free wall of the TAC model, whereas, in the MI model, only the left ventricle exhibited hypertrophy. These differences were also reflected in the expression analysis. In conclusion, our analysis shows that regional differences in gene expression exist in the heart. Moreover, common pathways that are coregulated in both models exist, and these might be central to the hypertrophic phenotype regardless of the initial hypertrophic stimuli.

left ventricle; septum; myocardial infarction; aorta constriction; cardiac hypertrophy; gene expression


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
DESPITE INTENSE EFFORTS of molecular, biochemical, physiological, and clinical investigation, the molecular mechanisms involved in the development of cardiac hypertrophy and the transition to failure remain incompletely understood. Previous efforts in studying candidate genes have identified a limited list of genes whose patterns of expression are altered during the development of hypertrophy or failure (2, 3, 5, 10). However, these studies have involved only a small fraction of the genes potentially expressed in the heart. Moreover, these studies have employed different species and models, and few genes have been compared across multiple models to determine the consistency of the response.

With the development of the microarray technology and the advances in bioinformatics, the candidate gene method has been replaced with a broader, genomic scale approach. However, analysis at this level can yield hundreds of genes that are differentially expressed between control and experimental tissues, without indication of their pathophysiological importance. Indeed, previous attempts of expression profiling in experimental or human cardiac failure have yielded "snapshots" of differential gene expression between normal vs. failing myocardium (14, 18, 32, 34). It is unclear whether these differences were important in mediating pathophysiology or merely represent secondary phenomena. Adding to the complexity of the phenotype is recent evidence indicating regional heterogeneity in both the biochemistry and the function of the normal and the hypertrophic heart (20, 38). These findings suggest the importance of the evaluation of regional differences in the understanding and treatment of cardiomyopathies.

To further restrict the set of genes to those specifically involved in the hypertrophic phenotype, we compared two different models of cardiac hypertrophy and remodeling to define genes concordantly regulated during the development of the disease state. Our hypothesis is that the cardiac hypertrophic response in different experimental models is mediated, to a large extent, by a common set of genes that play convergent pathophysiological roles. The divergently expressed genes in different hypertrophic conditions may mediate model-specific or condition-specific processes such as concentric or eccentric hypertrophy, etc. A similar approach has recently been employed by Gerritsen et al. (16) to define genes involved in angiogenesis.

In this report, we have performed expression profiling to compare two different murine models of cardiac hypertrophy and remodeling: myocardial infarction (MI) and transverse aortic constriction (TAC). The former results in increased regional wall stress to the noninfarcted myocardium in response to impaired ventricular function in the area of MI (15, 19). The latter causes pressure overload that confers global increase in wall stress (29). We also examined the pattern of gene expression in the left ventricle free wall (LV) and intraventricular septum in normal and the two models of cardiac remodeling. This comparison might provide useful insights in the understanding of cardiac remodeling with potential clinical implications in regard to biopsy sampling.


    MATERIALS AND METHODS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 

Animal surgery.
All animal procedures were approved by the Harvard Standing Committee on Animals. Male C57BL/6 mice (Jackson Laboratories) underwent TAC (29) or ligation of the coronary artery (MI) (15, 19) at 8 wk of age under ketamine and xylazine anesthesia. Additional mice underwent the identical surgical procedure without placement of the band (TAC sham) or without placement of the ligation (MI sham) to serve as age- and sex-matched sham-operated controls. Animals were euthanized 1 wk following surgery, and the hearts were collected and weighed to calculate the heart-to-body weight ratio. These samples were used for histology (Mason’s trichrome and hematoxylin-eosin staining) or RNA extraction.

RNA extraction.
The heart was dissected, and the base of the LV and the intraventricular septum were isolated and frozen in liquid nitrogen for RNA extraction. The base was chosen to avoid the infarcted region, which is located in the apex. RNA was extracted using Trizol reagent (GIBCO BRL). RNA concentration was determined by optical density at 260 nm, and quality was confirmed by denaturing agarose gel electrophoresis. Eight micrograms of total RNA from individual hearts was used for microarray analysis.

Oligonucleotide arrays.
A total of 18 left ventricular and septum samples were used for microarray analysis. RNA was harvested from left ventricle and septum of mice that underwent transverse aorta constriction (TAC, n = 2), MI (n = 3) or sham operation (TAC sham or MI sham, n = 4) 1 wk after surgery. Affymetrix Oligonucleotide Murine Genome 11k arrays (GeneChips A and B, see http://www.affymetrix.com, catalog no. 900188 and 900189), which contain ~12,483 known genes and expressed sequences tags (ESTs), were used in these studies. RNA processing, hybridization, and initial analyses were conducted as suggested by Affymetrix. In brief, total RNA was reverse transcribed to double-stranded cDNA (Life Technologies), and biotinylated cRNA was generated by an in vitro transcription reaction (Enzo). Labeled cRNA was purified (Qiagen), fragmented by alkaline treatment, and hybridized to a GeneChip array overnight at 45°C. The array was washed, stained with streptavidin-phycoerythrin, and scanned. Each array was scanned twice (Affymetrix Gene Array scanner), and an average intensity for each probe pair was generated. Data were first analyzed using Affymetrix Microarray Suite (MAS 4.01) to assess quality of RNA and hybridization.

Microarray analysis.
The software DNA-Chip Analyzer (dChip) (24) was used to implement model-based expression calculations (raw data have been submitted in http://www.ncbi.nlm.nih.gov/geo/ under series number GSE 415; GSE is a Gene Expression Omnibus database accession number), two-group comparison, and clustering. To identify genes that change in at least one of the experimental comparisons (LV MI vs. LV sham, LV TAC vs. LV sham, septum MI vs. septum sham, septum TAC vs. septum sham, LV MI vs. septum MI, LV TAC vs. septum TAC, LV sham vs. septum sham), the gene expressions of samples of the same group were pooled and an unpaired t-test was applied. The t-statistic is computed as (mean1 - mean2)/sqrt[SE(mean1)2 + SE (mean2)2], and the P value is computed based on the t-distribution [degree of freedom (df) = group.size1 + group. size2 - 1]. Given this, in our analysis the smallest df used in any comparison was 5 (group.size1 = 2, group.size2 = 4, df = 4 + 2 - 1 = 5). A two-tailed P < 0.05 was used to filter the genes that were significantly changing between the groups. In addition, as verification to the dChip analysis, we also used the method of false discovery rate [Tusher et al. (35); Significance Analysis of Microarrays (SAM) software], which have been proposed for analysis of small samples. For clustering, a hierarchical clustering algorithm (average linkage method) was applied. Briefly, the log2-transformed expression values across all samples were standardized to a mean of 0 and standard deviation of 1 (row standardization). Replicate arrays from each experimental group were pooled, and the average values were used to estimate correlations between the samples and to define the clustering nodes. An overview of the protocol for the data analysis is presented in Fig. 1.



View larger version (23K):
[in this window]
[in a new window]
 
Fig. 1. Analysis scheme. Analysis A: regional differences in gene expression in normal heart. From the 12,483 probes sets of the array, we filtered for transcripts that were present (expressed) in at least 50% of the left ventricle (LV) sham samples (2 of 4 samples) and or in at least 50% of the septum sham samples (2 of 4 samples). As shown, 4,908 transcripts were present in at least 50% of the LV samples, and 4,574 transcripts were present in at least 50% of the septum samples, with 4,125 transcripts expressed in both regions. To identify transcripts that were differentially expressed between the two regions, we applied a t-test analysis as described in MATERIALS AND METHODS. We found 106 transcripts to be differentially regulated between the two regions with 95% confidence. Of those transcripts, 70 were also shown to be different by an independent analysis using the significance of microarray analysis (SAM) method. Analysis B: gene expression differences after myocardial infarction (MI) or transverse aortic constriction (TAC). From the 12,483 probes sets of the array, we filtered for transcripts that were present (expressed) in at least 11% of the LV sham samples (2 of 18 samples); 6,836 transcripts passed this filtering criterion, and these were subjected to statistical analysis by the unpaired t-test method. The filtering criteria at these steps were present in at least one the samples compared, 95% statistical significance, and change as defined by the t-test in at least one of the following comparisons: LV MI vs. LV sham, LV TAC vs. LV sham, septum MI vs. septum sham, septum TAC vs. septum sham, LV MI vs. septum MI, LV TAC vs. septum TAC, LV sham vs. septum sham. We found that 1,263 transcripts passed this filtering step. These were then queried for transcripts that were changing much between the different groups (variation across all samples between 0.03 and 10) but whose gene expression levels were consistent within replicate arrays (variation across replicate samples between 0–10). The cutoff values for these parameters are according to the suggestions of dChip manual. Finally, 421 transcripts passed these criteria, and they were subjected to functional categorization and hierarchical clustering as described in MATERIALS AND METHODS.

 
Real-time RT-PCR.
Real-time RT-PCR (RT-PCR) reactions were carried out in iCycler IQ Real-Time Detection Systems (Bio-Rad). SuperScript One-step RT-PCR with Platinum Taq kits (Invitrogen) were used for all RT-PCR amplification in a total volume of 50 µl, which contained 200 ng total RNA, 5 mM MgSO4, 500 nM forward and reverse primers, and 200 nM fluorogenic probe. RT-PCR amplification for each RNA sample was performed in triplicate. One control without reverse transcriptase for each RNA sample and one control with distilled H2O instead of RNA for each primer and probe set were also performed. Relative gene expression analysis is carried out using the standard curve method (17, 36). Details on the primers used for the PCR reactions are given in Supplemental Table S2, available at the Physiological Genomics web site.1


    RESULTS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 

Regional differences in gene expression in the normal heart.
Microarray gene profiling was performed using RNA extracted from the septum and the left ventricle of the sham-operated animals (TAC sham left ventricle n = 2, MI sham left ventricle n = 2, TAC sham septum n = 2, and MI sham septum n = 2). The expression data were filtered as indicated in Fig. 1A. From the 4,125 transcripts that were found expressed in both regions (Supplementary Table S1), 1,721 were called present in all samples (100% P calls in all 8 samples). The vast majority of these transcripts have been reported to be expressed in the heart. However, our analysis also identified a set of 441 transcripts representing novel or unannotated ESTs. From the set of transcripts expressed in the heart, 693 and 360 transcripts were uniquely present in the left ventricle or septum, respectively.

To identify genes differentially expressed between the left ventricle and the septum of the sham samples we used the dChip (24) software (Fig. 1A). We found that 106 genes exhibited region-specific differential expression in the normal adult heart. These data were also analyzed using the SAM software (35), which confirmed 70 of the 106 probe sets as significantly different between the normal ventricle and the septum with a false detection ratio of 0 (Table 1). Examples of genes differentially expressed between the septum and the left ventricle of the sham animals are the dopamine receptor 2, arginase 1, and cardiac atria myosin light chain. Interestingly, four GTP-related genes (the Rab GDP-dissociation inhibitor ß; the Rho GDP-dissociation inhibitor-ß; the guanine nucleotide binding protein, {alpha}-inhibiting 2; and a homolog of the guanine nucleotide binding protein G) were all found to be expressed higher in the LV than the septum (Table 1).


View this table:
[in this window]
[in a new window]
 
Table 1. Genes that are differentially expressed between LV and septum of the normal heart

 
To our knowledge, this is the first report of such a large scale analysis (~12,000 transcripts) showing regional differences in the profiles of gene expression in the normal heart. Although the functional consequences of this observation are unclear, these variations in profiles might reflect differences in the function or hemodynamic characteristics of these two regions.

Gene expression following MI or TAC.
Both TAC and MI resulted in cardiac hypertrophy as evidenced by an increase in the heart-to-body weight ratio of the mice (27% and 17% increase in heart mass respectively, P < 0.05). This was confirmed by histological analysis following Mason’s trichrome staining, which revealed significant hypertrophy and early fibrosis in both the left ventricle and septum of the TAC-operated hearts (Fig. 2). In contrast, in the MI model, only the left ventricle exhibited hypertrophy and mild fibrosis, whereas the septum of the MI animals did not appear different from the sham control septum.



View larger version (71K):
[in this window]
[in a new window]
 
Fig. 2. Hypertrophy and fibrosis in the heart 1 wk after MI or TAC. Cardiac tissue sections from MI-, TAC-, TAC sham-, and MI sham-operated animals (n = 2 for each group) were stained with Mason’s trichrome staining 1 wk after the operation. The cardiac muscles are stained red, whereas blue staining indicates the fibrosis areas. Images A, C, E, and G show the septum in the different models, whereas images B, D, F, and H show the left ventricle.

 
The genes that were differentially expressed between the different experimental groups were identified by querying for transcripts that were present in at least 2 of the 18 samples compared and were differentially expressed in at least one of the comparisons with P < 0.05. Using these criteria, we identified 1,263 probes that were differentially expressed between different regions or models (MI or TAC). Approximately 850 probe sets represented known genes, whereas the remainder were unannotated ESTs. These genes were then subjected to further filtering to eliminate probe sets with high variability among replicas and genes that showed low variance among all samples (see Fig. 1B for details). We found that 421 genes passed these criteria. Functional categorization of these transcripts revealed that many shared common physiological roles and were grouped into categories based on their biological roles (Table 2). These transcripts were then subjected to hierarchical clustering using dChip (24). A graphical representation of the clustering of the 421 genes is shown in Fig. 3. As evidence to the quality of the data, replicate genes within the data set were found in the same clusters (i.e., fibronectin, two of two probe sets on the array; collagens, three of three probe sets on the array; heat shock protein 47 kDa, two of three probe sets on the array).


View this table:
[in this window]
[in a new window]
 
Table 2. Functional annotation of genes that are differentially expressed during cardiac remodeling

 


View larger version (44K):
[in this window]
[in a new window]
 
Fig. 3. Clusters of gene expression data. The average expression profile of each cluster is illustrated on the right. A: transcripts upregulated both after MI or TAC, but the response is higher in the TAC. This set is enriched for genes encoding for immune response proteins. B: transcripts upregulated both after MI or TAC, but the response in the septum is higher in the TAC than in the MI. This set is enriched for extracellular matrix and cell communication genes. C: transcripts upregulated both after MI or TAC, but the response is higher in the MI. This set is enriched for cell cycle genes. D: transcripts downregulated both after MI or TAC. This set is enriched for enzyme encoding genes.

 
Visual inspection of the data revealed several interesting observations. First and consistent with the morphological analysis, there was a strong overall concordance in the levels of expression between the LV and the septum in the TAC model. In contrast, there was discordance in the response of the MI LV and the MI septum. Secondly, several clusters existed that were enriched for particular functional categories. For example, cluster A contained genes encoding immune response proteins and genes involved in cytoskeleton organization and biogenesis, cluster B was highly enriched for extracellular matrix genes and genes involved in cell adhesion, and cluster C was enriched for cell cycle and signal transduction genes. It is also interesting to note that these three main clusters also exhibited differential patterns of expression in the two models. Specifically, cluster A contained transcripts that were concordantly upregulated in both models but whose magnitude of response was greater in the TAC compared with the response to MI (Fig. 3). Genes in this cluster included the four and a half LIM domains 1, ANP, BNP, tubulins, and various histocompatibility antigens. Similarly, cluster C consisted of genes with a greater response in the MI than in the TAC model. Among those were the cyclin D3, cyclin-dependent kinase 4, tropomodulin 3, epidermal growth factor receptor pathway substrate 8, capping protein {alpha}1, growth arrest specific 1 and growth arrest specific 2.

Notably, consistent with the histological data, sets of genes could be identified that were upregulated in both regions of the TAC animals and upregulated in the LV of the MI, but with little or no upregulation in the MI septum (Fig. 3, cluster B, and Table 3). Among these genes were many extracellular matrix genes encoding several procollagen isoforms, fibrillin, biglycan, and fibronectin. Interestingly, this cluster also included many genes involved in cell differentiation and proliferation such as the stromal cell-derived factor 3, transgelin 2, secreted frizzled-related sequence protein 2, mesoderm-specific transcript, homeobox protein Mox-1, nidogen-1 and nidogen-2, and osteoblast-specific factor 2. Thus, in agreement with our hypothesis, a set of genes does exist that exhibit concordant patterns of expression in the two models of ventricular hypertrophy and might reflect common biological phenomena, mechanisms, or pathways that mediate the hypertrophic response regardless of the initiating event.


View this table:
[in this window]
[in a new window]
 
Table 3. Hypertrophy induced genes that are concordantly regulated between MI LV, TAC LV, and TAC septum

 
Validation of the microarray data by real-time RT-PCR.
Real-time RT-PCR was used to measure the relative expression of seven selected genes as a means to validate the microarray data. Since the main aim of this paper was to identify genes that were concordantly regulated in both models, we chose to validate genes whose microarray profile indicated that they were upregulated in both models. Of the seven genes that were chosen for real-time PCR analysis, five, actin {alpha}1-skeletal muscle, four and a half LIM domains 1, orosomucoid 1, osteoblast-specific factor 2 (fasciclin I-like), and aplysia ras-related homolog 9 (RhoC), matched this profile The other two, guanylate nucleotide binding protein 3 and tyrosine 3-monooxgenase, showed no significant variation with the microarray analysis and were used arbitrarily as controls for false negatives. As shown in Fig. 4 and Table 4, the results from this analysis were in agreement with the microarray data in terms of both the direction of change and for most of the genes the magnitude of the change (Table 4).



View larger version (13K):
[in this window]
[in a new window]
 
Fig. 4. Real-time RT-PCR validation. Graphical representation of the expression values for one of the seven genes validated, as estimated by real-time RT-PCR and by microarray analysis. For the real-time PCR, the quantities of the RNA for each gene were normalized to the quantity of 18S in the same sample, which was also measured by real-time PCR. For the microarray data, the expression values are given in the linear scale after background correction, normalization, and scaling down (scaling factor 500). The correlation coefficient between the expression values as estimated by the two methods is indicated.

 

View this table:
[in this window]
[in a new window]
 
Table 4. Validation of fold changes by real-time PCR

 

    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
Previous profiling analysis of the myocardium following induction of hypertrophy revealed complex patterns of expression where typically hundreds of genes were differentially expressed between the normal and hypertrophied myocardium (14, 18, 32, 34). These data, while interesting, were unable to identify those genes that play important pathophysiological roles vs. those that represent secondary phenomena. Here, we propose a strategy that compares altered gene expression patterns between different models of cardiac hypertrophy and remodeling. We hypothesized that among the genes that exhibit differential patterns of expression, some would be genes that are concordantly regulated and that these genes may play a causal role in the hypertrophic response.

To test this hypothesis, we employed the microarray technology to evaluate the patterns of expression that occur in response to two different models of cardiac pressure overload in vivo. TAC produces pressure overload as a stimulus for the development of concentric hypertrophy, resembling the clinical scenario observed in patients with hypertension or aortic stenosis (29). In contrast, coronary artery ligation results in the death of the region of the myocardium perfused by the LAD, resulting in hypertrophy of the remaining viable left ventricle (15, 19). In addition, we investigated the differences in expression between the different regions (septum and LV) of the normal heart and the heart after overload. It is gratifying to note that several of the genes identified have been previously presented to be associated with hypertrophy using more traditional methods. These included ANP, BNP, actins, and collagens (1, 3, 6, 7, 26, 37). Moreover, validation of the transcriptional effects by RT-PCR suggests significant agreement with the microarray data (~85%).

In the normal heart, the septum and left ventricle are morphologically and functionally similar, and both regions consist primarily of cardiac myocytes. Consistent with this, the patterns of gene expression in these two regions are also very similar, with 85–90% of the genes expressed in one region also expressed in the other. However, distinct differences in the profiles of expression were also observed. Strikingly, there appears to be differential expression among several genes, including cardiac atria myosin light chain, dopamine receptor 2, and GTP- and ATP-related proteins. These differences most likely reflect physiological and hemodynamic differences between the two regions. For instance, the atrioventricular bundle (AVB) is localized only the septum. Molecular characterization of the mouse cardiac conduction is rudimentary at present; however, increasing information regarding cardiac function and embryonic heart development indicates that distinct profiles of gene expression do exist. Indeed, Franco and colleagues (12, 13) showed that in AVB, ß-MHC and MLC2v are expressed, whereas {alpha}-MHC, MLC2a, Cx43, desmin, or {alpha}-SMA are not. In addition, another study showed that enrichment of G protein {alpha}-subunit mRNAs is found in the atrioventricular-conducting system of the mammalian heart (11).

In this study, histological analysis of the samples revealed that 1 wk following surgery, the pathology of the base of the heart was different between the MI- and the TAC-operated animals. Specifically, both the septum and the LV in the TAC-operated animals showed diffuse myocyte hypertrophy and fibrosis, whereas in the MI-operated animals, hypertrophy and fibrosis were found only in the LV but not in the septum. These model-specific differences in regional responses were also shown in the microarray expression data. Cluster analysis revealed that patterns of gene expression were altered more substantially in the LV than in the septum of the MI model. These regional effects most likely reflect differences in the mechanical stimuli resulting from the two different insults to the heart. The impairment of myocardial function in the area of MI causes stress in the neighboring myocardium that has to increase its contractility capacity to maintain cardiac function. Thus, following MI, the effect is initially a local effect, which gradually affects the whole ventricle. In contrast, the TAC model from the start exerts a more global effect since the entire ventricle must respond to the pressure overload caused by the constricted aorta.

The region-specific differences in the hypertrophic response allow us to query the data set for genes that exhibit a pattern of expression that parallels the morphological patterns. As shown in Fig. 3, hierarchical clustering reveals several distinct patterns. Cluster B contains genes whose expression is upregulated in the models in regions where hypertrophy was evident but to a much lesser extent in the septum of the MI animals. This cluster contains 92 genes, listed in Table 3. As anticipated, among these genes are several genes encoding extracellular matrix and cell adhesion molecules, such as several different isoforms of procollagen, tenascin, fibronectin, fibrillin, sparc (related to osteonectin), and osteoblast-specific factor 2 (periostin). The osteoblast-specific factor 2 (osf-2) functions as a cell adhesion molecule for preosteoblasts and is thought to be involved in osteoblast recruitment, attachment, and spreading (22). Recent data show, however, that osf-2 mRNA is expressed in the developing mouse embryonic and fetal heart and that it is localized to the endocardial cushions that ultimately divide the primitive heart tube into a four-chambered heart (22). These data suggest that osf-2 is involved in the recruitment, attachment, and spreading of endocardial cells into the overlying extracellular matrix during cardiac morphogenesis. Whether osf-2 plays a similar role during cardiac hypertrophy is yet to be determined.

Lysyl oxidase (Lox) and the related protein lysyl oxidase-like (Loxl) protein were also in this gene list. lox and loxl are responsible for the catalysis of lysine-derived cross-links in fibrillar collagens and elastin. These proteins are upregulated in several models of fibrosis such as carbon tetrachloride-induced liver fibrosis (21, 28). More recently, novel roles have been attributed to lysyl oxidases including diverse biological functions such as developmental regulation, tumor suppression, cell motility, and cellular senescence (8).

In addition, two isoforms of nidogen (1 and 2) are present in this gene list. Nidogen-1 (entactin) acts as a bridge between the extracellular matrix molecules laminin-1 and type IV collagen and thus participates in the assembly of basement membranes, and recent data suggest that it also regulates laminin-1 controlled gene expression (27, 3). Evidence also suggests that the nidogens could modulate angiogenesis. In vitro studies have shown that nidogen-laminin complexes enhanced the stabilization of microvessels in a dose-dependent manner (30).

Another intriguing gene on this list was mesenchyme homeobox 1, Mox-1. In cultured cells, Mox-1 is thought to be involved in the skeletal muscle lineage; however, during early embryogenesis, at the primitive streak stage, Mox-1 is expressed in mesoderm, posterior to the heart and in midgestation (4). It is found in the heart cushions and truncus arteriosus (4). Thus we speculate that Mox-1 may be expressed in newly formed or recruited myocytes.

Interestingly, a significant number of cell cycle genes were identified [cyclin B, cyclin D3, cyclin-dependent kinase 4 (cdk4), growth arrest protein 1 and 2]. Cyclin B1 is expressed predominantly in the G2/M phase of cell division, whereas cyclin D is normally expressed during the G1 phase of the cell cycle. The complexes formed by CDK4 and the cyclin D are involved in the control of cell proliferation during the G1 phase. Both CDK4 and cyclin D have been shown to induce hypertrophic growth in cardiac myocytes in vitro (33). The growth arrest specific gene 1 (Gas1) has been implicated in inhibiting cell cycle progression in cell culture studies (9). Recent experiments with Gas1 indicate that it may act as a growth-inducing gene, challenging its previous function as a gene specifically involved in growth arrest and highlighting a role in the regulation of cell growth and regulation (25). Similarly, Gas2 has been suggested to play an important role in regulating chondrocyte proliferation and differentiation (23). These results are consistent with recent data suggesting that proliferation of cardiac myocytes may occur following injury.

In summary, our data demonstrate, for the first time, significant region-specific differences in gene expression in the heart. Moreover, the results highlight the role of extracellular matrix as a basic mechanism in the response to cardiac hypertrophy. Importantly, we have identified several genes such as Mox-1, interferon, nitrogen, and perforin not previously associated with hypertrophy. Finally, to our knowledge this is the first study comparing different models of cardiac hypertrophy as a means of identifying basic mechanisms involved in cardiac remodeling regardless of the initial stimuli.


    ACKNOWLEDGMENTS
 
Present address of C. M. H. Watanabe: Texas A&M University Department of Chemistry College Station, Texas 77843.

C. D. Sigmund served as the review editor for this manuscript submitted by Editors R. E. Pratt and V. J. Dzau.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grants HL-58516 and HL-73219 (to V. J. Dzau) and HL-61661 (to R. E. Pratt). M. Mirotsou is a recipient of an AHA Northeastern Affiliate Fellowship.


    FOOTNOTES
 
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).

Address for reprint requests and other correspondence: V. J. Dzau or R. E. Pratt, Brigham and Women’s Hospital, Dept. of Medicine, 75 Francis St., Boston, MA 02115 (E-mail: vdzau{at}partners.org or rpratt{at}rics.bwh.harvard.edu).

10.1152/physiolgenomics.00071.2003.

1 The Supplementary Material for this article (Supplemental Tables S1 and S2) is available online at http://physiolgenomics.physiology.org/cgi/content/full/00071.2003/DC1. Back


    References
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 

  1. Adachi S, Ito H, Tamamori M, Tanaka M, Marumo F, and Hiroe M. Skeletal and smooth muscle {alpha}-actin mRNA in endomyocardial biopsy samples of dilated cardiomyopathy patients. Life Sci 63: 1779–1791, 1998.[ISI][Medline]
  2. Aihara Y, Kurabayashi M, Saito Y, Ohyama Y, Tanaka T, Takeda S, Tomaru K, Sekiguchi K, Arai M, Nakamura T, and Nagai R. Cardiac ankyrin repeat protein is a novel marker of cardiac hypertrophy: role of M-CAT element within the promoter. Hypertension 36: 48–53, 2000.[Abstract/Free Full Text]
  3. Black FM, Packer SE, Parker TG, Michael LH, Roberts R, Schwartz RJ, and Schneider MD. The vascular smooth muscle {alpha}-actin gene is reactivated during cardiac hypertrophy provoked by load. J Clin Invest 88: 1581–1588, 1991.[ISI][Medline]
  4. Candia AF, Hu J, Crosby J, Lalley PA, Noden D, Nadeau JH, and Wright CV. Mox-1 and Mox-2 define a novel homeobox gene subfamily and are differentially expressed during early mesodermal patterning in mouse embryos. Development Cambridge 116: 1123–1136, 1992.
  5. Chien KR, Knowlton KU, Zhu H, and Chien S. Regulation of cardiac gene expression during myocardial growth and hypertrophy: molecular studies of an adaptive physiologic response. FASEB J 5: 3037–3046, 1991.[Abstract/Free Full Text]
  6. Contard F, Koteliansky V, Marotte F, Dubus I, Rappaport L, and Samuel JL. Specific alterations in the distribution of extracellular matrix components within rat myocardium during the development of pressure overload. Lab Invest 64: 65–75, 1991.[ISI][Medline]
  7. Cornelius T, Holmer SR, Muller FU, Riegger GA, and Schunkert H. Regulation of the rat atrial natriuretic peptide gene after acute imposition of left ventricular pressure overload. Hypertension 30: 1348–1355, 1997.[Abstract/Free Full Text]
  8. Csiszar K. Lysyl oxidases: a novel multifunctional amine oxidase family. Prog Nucleic Acid Res Mol Biol 70: 1–32, 2001.[ISI][Medline]
  9. Del Sal G, Ruaro ME, Philipson L, and Schneider C. The growth arrest-specific gene, gas1, is involved in growth suppression. Cell 70: 595–607, 1992.[ISI][Medline]
  10. Di Nardo P, Minieri M, Carbone A, Maggiano N, Micheletti R, Peruzzi G, and Tallarida G. Myocardial expression of atrial natriuretic factor gene in early stages of hamster cardiomyopathy. Mol Cell Biochem 125: 179–192, 1993.[ISI][Medline]
  11. Eschenhagen T, Laufs U, Schmitz W, Scholz H, Warnholtz A, Weil J, and Schafer HJ. Enrichment of G protein {alpha}-subunit mRNAS in the AV-conducting system of the mammalian heart. J Mol Cell Cardiol 27: 2249–2263, 1995.[ISI][Medline]
  12. Franco D and Icardo JM. Molecular characterization of the ventricular conduction system in the developing mouse heart: topographical correlation in normal and congenitally malformed hearts. Cardiovasc Res 49: 417–429, 2001.[ISI][Medline]
  13. Franco D, Markman MM, Wagenaar GT, Ya J, Lamers WH, and Moorman AF. Myosin light chain 2a and 2v identifies the embryonic outflow tract myocardium in the developing rodent heart. Anat Rec 254: 135–146, 1999.[ISI][Medline]
  14. Friddle CJ, Koga T, Rubin EM, and Bristow J. Expression profiling reveals distinct sets of genes altered during induction and regression of cardiac hypertrophy. Proc Natl Acad Sci USA 97: 6745–6750, 2000.[Abstract/Free Full Text]
  15. Gehrmann J, Frantz S, Maguire CT, Vargas M, Ducharme A, Wakimoto H, Lee RT, and Berul CI. Electrophysiological characterization of murine myocardial ischemia and infarction. Basic Res Cardiol 96: 237–250, 2001.[ISI][Medline]
  16. Gerritsen ME, Soriano R, Yang S, Ingle G, Zlot C, Toy K, Winer J, Draksharapu A, Peale F, Wu TD, and Williams PM. In silico data filtering to identify new angiogenesis targets from a large in vitro gene profiling data set. Physiol Genomics 10: 13–20, 2002.[Abstract/Free Full Text]
  17. Heid CA, Stevens J, Livak KJ, and Williams PM. Real time quantitative PCR. Genome Res 6: 986–994, 1996.[Abstract]
  18. Hwang DM, Dempsey AA, Lee CY, and Liew CC. Identification of differentially expressed genes in cardiac hypertrophy by analysis of expressed sequence tags. Genomics 66: 1–14, 2000.[ISI][Medline]
  19. Kanno S, Lerner DL, Schuessler RB, Betsuyaku T, Yamada KA, Saffitz JE, and Kovacs A. Echocardiographic evaluation of ventricular remodeling in a mouse model of myocardial infarction. J Am Soc Echocardiogr 15: 601–609, 2002.[ISI][Medline]
  20. Kaprielian R, Sah R, Nguyen T, Wickenden AD, and Backx PH. Myocardial infarction in rat eliminates regional heterogeneity of AP profiles, Ito K+ currents, and [Ca2+]i transients. Am J Physiol Heart Circ Physiol 283: H1157–H1168, 2002. First published May 2, 2002; 10.1152/ajpheart.00518.2001.[Abstract/Free Full Text]
  21. Kim Y, Peyrol S, So CK, Boyd CD, and Csiszar K. Coexpression of the lysyl oxidase-like gene (LOXL) and the gene encoding type III procollagen in induced liver fibrosis. J Cell Biochem 72: 181–188, 1999.[ISI][Medline]
  22. Kruzynska Frejtag A, Machnicki M, Rogers R, Markwald RR, and Conway SJ. Periostin (an osteoblast-specific factor) is expressed within the embryonic mouse heart during valve formation. Mech Dev 103: 183–188, 2001.[ISI][Medline]
  23. Lee KK, Tang MK, Yew DT, Chow PH, Yee SP, Schneider C, and Brancolini C. gas2 is a multifunctional gene involved in the regulation of apoptosis and chondrogenesis in the developing mouse limb. Dev Biol 207: 14–25, 1999.[ISI][Medline]
  24. Li C and Hung Wong W. Model-based analysis of oligonucleotide arrays: expression index computation and outlier detection. Proc Natl Acad Sci USA 98: 31–36, 2001.[Abstract/Free Full Text]
  25. Liu Y, May NR, and Fan CM. Growth arrest specific gene 1 is a positive growth regulator for the cerebellum. Dev Biol 236: 30–45, 2001.[ISI][Medline]
  26. Magga J, Vuolteenaho O, Tokola H, Marttila M, and Ruskoaho H. B-type natriuretic peptide: a myocyte-specific marker for characterizing load-induced alterations in cardiac gene expression. Ann Med 30, Suppl 1: 39–45, 1998.
  27. Miosge N, Holzhausen S, Zelent C, Sprysch P, and Herken R. Nidogen-1 and nidogen-2 are found in basement membranes during human embryonic development. Histochem J 33: 523–530, 2001.[ISI][Medline]
  28. Murawaki Y, Kusakabe Y, and Hirayama C. Serum lysyl oxidase activity in chronic liver disease in comparison with serum levels of prolyl hydroxylase and laminin. Hepatology Baltimore 14: 1167–1173, 1991.
  29. Nakamura A, Rokosh DG, Paccanaro M, Yee RR, Simpson PC, Grossman W, and Foster E. LV systolic performance improves with development of hypertrophy after transverse aortic constriction in mice. Am J Physiol Heart Circ Physiol 281: H1104–H1112, 2001.[Abstract/Free Full Text]
  30. Nicosia RF, Bonanno E, Smith M, and Yurchenco P. Modulation of angiogenesis in vitro by laminin-entactin complex. Dev Biol 164: 197–206, 1994.[ISI][Medline]
  31. Pujuguet P, Simian M, Liaw J, Timpl R, Werb Z, and Bissell MJ. Nidogen-1 regulates laminin-1-dependent mammary-specific gene expression. J Cell Sci 113: 849–858, 2000.[Abstract/Free Full Text]
  32. Stanton LW, Garrard LJ, Damm D, Garrick BL, Lam A, Kapoun AM, Zheng Q, Protter AA, Schreiner GF, and White RT. Altered patterns of gene expression in response to myocardial infarction. Circ Res 86: 939–945, 2000.[Abstract/Free Full Text]
  33. Tamamori Adachi M, Ito H, Nobori K, Hayashida K, Kawauchi J, Adachi S, Ikeda MA, and Kitajima S. Expression of cyclin D1 and CDK4 causes hypertrophic growth of cardiomyocytes in culture: a possible implication for cardiac hypertrophy. Biochem Biophys Res Commun 296: 274–280, 2002.[ISI][Medline]
  34. Tan FL, Moravec CS, Li J, Apperson Hansen C, McCarthy PM, Young JB, and Bond M. The gene expression fingerprint of human heart failure. Proc Natl Acad Sci USA 99: 11387–11392, 2002.[Abstract/Free Full Text]
  35. Tusher VG, Tibshirani R, and Chu G. Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci USA 98: 5116–5121, 2001.[Abstract/Free Full Text]
  36. Winer J, Jung CK, Shackel I, and Williams PM. Development and validation of real-time quantitative reverse transcriptase-polymerase chain reaction for monitoring gene expression in cardiac myocytes in vitro. Anal Biochem 270: 41–49, 1999.[ISI][Medline]
  37. Yamamoto K, Burnett JC, Jougasaki M, Nishimura RA, Bailey KR, Saito Y, Nakao K, and Redfield MM. Superiority of brain natriuretic peptide as a hormonal marker of ventricular systolic and diastolic dysfunction and ventricular hypertrophy. Hypertension 28: 988–994, 1996.[Abstract/Free Full Text]
  38. Yanagida R, Sugawara M, Kawai A, and Koyanagi H. Regional differences in myocardial work of the left ventricle in patients with idiopathic dilated cardiomyopathy: implications for the surgical technique used for left ventriculoplasty. J Thorac Cardiovasc Surg 122: 600–607, 2001.[Abstract/Free Full Text]