Physiol. Genomics AJP: Endocrinology and Metabolism
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Physiol. Genomics 27: 295-308, 2006. First published August 1, 2006; doi:10.1152/physiolgenomics.00318.2005
1094-8341/06 $8.00
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplementary Tables and Figure
Right arrow All Versions of this Article:
27/3/295    most recent
00318.2005v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (2)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Cerutti, C.
Right arrow Articles by Gustin, M.-P.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Cerutti, C.
Right arrow Articles by Gustin, M.-P.
Received 21 December 2005; accepted in final form 26 July 2006.
Physiological Genomics 27:295-308 (2006)
1094-8341/06 $8.00 © 2006 American Physiological Society

Transcriptional alterations in the left ventricle of three hypertensive rat models

Catherine Cerutti , Mazen Kurdi , Giampiero Bricca , Wassim Hodroj , Christian Paultre , Jacques Randon and Marie-Paule Gustin

Equipe d’Accueil 3740 Génomique fonctionnelle dans l'athérothrombose, Université Lyon 1, Lyon, France


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
Left ventricular hypertrophy (LVH) is commonly associated with hypertension and represents an independent cardiovascular risk factor. The aim of this study was to test the hypothesis that the cardiac overload related to hypertension is associated to a specific gene expression pattern independently of genetic background. Gene expression levels were obtained with microarrays for 15,866 transcripts from RNA of left ventricles from 12-wk-old rats of three hypertensive models [spontaneously hypertensive rat (SHR), Lyon hypertensive rat (LH), and heterozygous TGR(mRen2)27 rat] and their respective controls. More than 60% of the detected transcripts displayed significant changes between the three groups of normotensive rats, showing large interstrain variability. Expression data were analyzed with respect to hypertension, LVH, and chromosomal distribution. Only four genes had significantly modified expression in the three hypertensive models among which a single gene, coding for sialyltransferase 7A, was consistently overexpressed. Correlation analysis between expression data and left ventricular mass index (LVMI) over all rats identified a larger set of genes whose expression was continuously related with LVMI, including known genes associated with cardiac remodeling. Positioning the detected transcripts along the chromosomes pointed out high-density regions mostly located within blood pressure and cardiac mass quantitative trait loci. Although our study could not detect a unique reprogramming of cardiac cells involving specific genes at early stage of LVH, it allowed the identification of some genes associated with LVH regardless of genetic background. This study thus provides a set of potentially important genes contained within restricted chromosomal regions involved in cardiovascular diseases.

hypertension; cardiac hypertrophy; angiotensin; microarray


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
HEART REMODELING in the form of left ventricular hypertrophy (LVH) is very common during hypertension and is considered as the first step toward myocardial infarction or heart failure (23). Indeed, the development of LVH in hypertension is often regarded first as an adaptation to increased workload, while the transition to heart failure reflects the loss of efficacy of this process. Various factors are involved in the development of LVH including cardiac workload and neuro-humoral factors such as the cardiac sympathetic drive and the renin-angiotensin-aldosterone system (RAAS), which may affect the cardiac workload through increases in vascular tone and volemia and directly stimulate hypertrophy through specific cardiomyocyte receptors. In addition, the large variety of genes and pathways involved in cardiac growth alterations in animal models such as transgenic mice and genetic inheritance in humans supports the evidence for a genetic contribution to LVH (1, 8, 12, 19). Natural history of LVH during hypertension may thus be viewed as the succession of three phases: development, stabilization, and failure, where workload and environmental and genetic factors interfere to produce the phenotype. The first phase is mostly studied using aortic banding models, which produce an abrupt and sustained increase in left ventricle pressure associated with variable activation or inhibition of the RAAS and the sympathetic drive according to the position of the stenosis relative to the renal arteries and the aortic arch. Alternatively, comparisons of spontaneously hypertensive rats with their control strain and/or studies of pharmacologically induced alterations in blood pressure and their consequences on LVH have been carried out.

LVH has already been analyzed using microarrays, studying acute or chronic LVH models (9, 16, 22, 30) or taking into account factors such as sex or temporal progression toward cardiac failure (20, 28). Genes of interest have been proposed from the lists of differentially expressed genes. However, until now, to the best of our knowledge, no direct comparison of heart transcriptome has been undertaken simultaneously in different spontaneously hypertensive rat models, the only comparative LVH study being done in transgenic mice (2).

The aim of the present work was to test the hypothesis that high blood pressure (BP) results in a cardiac-specific gene expression pattern independently of other factors including genetic ones. For this purpose, we analyzed global gene expression in the left ventricle of three different hypertensive rat models and their respective controls at an early stage of LVH development using high-density microarrays. We used two genetically derived rat hypertensive inbred strains: the spontaneously hypertensive rat of the Okamoto strain (SHR) selected from a Wistar-Kyoto (WKY) population, and the Lyon hypertensive strain (LH) with its Lyon low blood pressure (LL) control strain. These latter represent two inbred strains belonging to the extremes of the BP distribution of the original population (27). The last model is a transgenic rat harboring heterozygous for the mouse Ren2 gene [TGR(mRen2)27], which leads to a fulminant hypertension, and its nontransgenic littermate, providing a model where a single pathogenic mechanism involving renin is involved in all phenotypic alterations (17).


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
Animals
All procedures were performed in accordance with the National Research Council's Guide for the Care and Use of Laboratory Animals. They were conducted by an investigator licensed to experiment on animals (GB). Male rats from the three different models were studied at 12 wk of age. SHR (n = 5) and WKY (n = 5) rats were obtained from IFFA-Credo (Les Oncins, France). LH (n = 5) and LL rats (n = 5) were obtained from the breeding colony of Centre National de Recherche Scientifique (CNRS) FRE 2678, Lyon, France. Transgenic heterozygous TGR(mRen2)27 rats (TGR+/–), resulting from the insertion of the mouse Ren2 transgene within a Sprague-Dawley Hanover genetic background, were obtained from our own colony (n = 6) maintained by crossbreeding transgene positive males with nontransgenic sibling females. Nontransgenic littermates (TGR–/–) were used as controls (n = 6).

Systolic BP was measured by tail-cuff plethysmography (IITC, Woodland Hills, CA) the day before death. Rats were anesthetized with pentobarbital sodium, and hearts were extracted, weighed, and immediately dissected in atria, right ventricular free wall, and left ventricle containing both the free wall and the septum. Each tissue sample was quickly dried, weighed, flash frozen in liquid nitrogen, and stored at –80°C for later extractions. Left ventricular mass index (LVMI) was calculated as the ratio of left ventricle mass to body weight.

Total RNA Isolation and Real-Time RT-PCR Analysis
Total RNA was extracted from left ventricular tissue using Trizol Reagent (Invitrogen, Carlsbad, CA) and treated with DNase (Deoxyribonuclease I, Amplification Grade, Invitrogen). Quality and concentration of total RNA were assessed by capillary electrophoresis (Agilent Bioanalyzer 2100; Agilent Technologies, Waldbronn, Germany). Reverse transcription was performed using random hexamers as primers (pdN6; Amersham Biosciences, Piscataway, NJ) and RT (Superscript II RNase H, Invitrogen), using 2 µg of total RNA in a total volume of 50 µl. Quantitative real-time RT-PCR was performed for selected gene transcripts with a MyiQ thermal cycler (Bio-Rad Laboratories, Hercules, CA). Real-time RT-PCR was performed using the resulting cDNA (0.5 µl) as template, iQ SYBR Green Supermix (Bio-Rad Laboratories), and the appropriate set of primers (Table 1) in 96-well microtiter plates. Two-step RT-PCR real-time amplifications were carried out as follows: 3 min at 95° followed by 10 s at 95° and 45 s at 60°. For each sample, PCR was performed in duplicate. Dilutions of control template were used to generate the standard curves for each target gene in triplicate. Cycle threshold values were calculated for the different products with Optical System Software v1.0 (Bio-Rad Laboratories). Expression levels obtained from the standard curves were normalized against 18S rRNA. All genes to be validated were quantified on the same tissue and/or mRNA samples as those analyzed by microarray.


View this table:
[in this window]
[in a new window]

  		Table 1. Quantitative real-time PCR primers

 
Microarray Experiments
Affymetrix GeneChip Rat Expression Array RAE230A [15,866 genes and expressed sequence tags (ESTs)] was used (Affymetrix, Santa Clara, CA) and processed on the Affymetrix platform of the Strasbourg Genopole. Each gene or EST is represented on the chip by 11 probe pairs of perfectly matched (PM) or mismatched (MM) 25-bp oligonucleotides. Expression data were obtained from Affymetrix Microarray Suite 5.0 software. For each probe set, the average PM-MM value was taken into account in association with information about signal detection (Present, Marginal, Absent) resulting from statistical comparison of the 11 PM-MM values to the background intensity. This comparison indicated whether a particular cRNA was actually detected (Present), marginally detected (Marginal), or not detected (Absent). Data have been submitted to ArrayExpress database with accession number E-MEXP-357.

Gene Expression Data Analysis
Only genes called Present in the transcriptome of all tissue samples (6,400 genes) and a few genes called Present in all the rats of one group and Absent in all rats of the corresponding complementary group (52 genes) were included in this analysis. Normality of the distributions of expression values obtained over the 32 rats was tested for each gene with the Shapiro-Wilk W-test using an alpha error risk of 0.01. Expression data distributions were found normal for 6,017 of the 6,452 studied genes. Therefore, no further data transformation was applied.

Data clustering.
Hierarchical clustering of rats or of subsets of genes was performed using the 6,452-gene or the 32-rat expression vectors, respectively. The expression profile data clustering and analysis software (EPCLUST; http://ep.ebi.ac.uk/EP/EPCLUST/, European Bioinformatics Institute) was used. The Pearson correlation coefficient (r) was chosen to compute distances between expression vectors (d = 1 – r), and the unweighted pair group method with arithmetic mean (UPGMA), based on average distance clustering, was used to build the hierarchical tree.

Statistical analysis of expression data.
For each gene, the comparison of expression data between hypertensive and control rats was done with the significance analysis of microarrays (SAM) 1.21 software (Stanford University Labs, Stanford, CA), designed to deal with multiple comparisons. The false discovery rate expressed as q-value was used to evaluate statistical significance, and its threshold was set at 0.10. No threshold was set for fold change. Genes differentially expressed were investigated using 1) a multiclass analysis to test differences between several groups of rats, either between the three normotensive rat models to evaluate genetic variability or between the six groups of rats to detect genes with steady expression, and 2) a two-class analysis within each pair of hypertensive and normotensive groups to specify expression changes in each model. The lists of differentially expressed genes obtained were then compared with identify shared or model-specific markers of LVH.

Relationships Between Gene Expression and LVMI Phenotype
Linear correlation between LVMI and expression level was tested for each of the 15,866 genes spotted on the microarrays using data from the 32 studied rats to identify a set of genes whose expression changes are associated with LVMI changes. The Pearson correlation coefficient was used, and statistical significance was set at P < 0.001 (|r| >0.554 for n = 32) to reduce false positive detection.

Bioinformatics Tools
Probe annotations were obtained from Affymetrix NetAffx Analysis Center in October 2005. Functional annotation of transcripts was based on Gene Ontology (GO) term assignments provided by the Rat Genome Database (RGD; http://rgd.mcw.edu/) developed by the Medical College of Wisconsin (Milwaukee, WI) (24). Information concerning quantitative trait loci (QTLs) relative to BP or cardiac mass phenotypic traits was obtained from RGD. Caryoscope software (http://dahlia.stanford.edu:8080/caryoscope/index.html) developed at Stanford University was used to position on chromosomes the selected genes and their expression level.

Standard Statistical Analysis
Data are given as means ± SE, and statistical analyses were performed with Statistica 6.0 software (StatSoft, Tulsa, OK). The comparisons of cardiovascular variables and of global microarray data among all the groups were made with the nonparametric Kruskal-Wallis test, followed by the nonparametric Mann-Whitney U-test for post hoc tests. Comparisons of mRNA levels obtained by quantitative RT-PCR (qRT-PCR) used the nonparametric Mann-Whitney U-test. Comparisons of gene distributions across the chromosomes were made with a chi-square test for distributions including more than five genes on each chromosome.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 GRANTS
 REFERENCES
 
Cardiovascular Parameters and Expression Data in the Left Ventricle
Hemodynamic and heart characteristics in the different groups of rats are summarized in Table 2. SHR and TGR+/– rats had similar high levels of systolic BP. In the Lyon strains, LH rat hypertension was less severe, and LL rats had the lowest systolic BP. LVMI was significantly higher in all of the three hypertensive groups compared with their controls, with less marked LVH in LH rats than in SHR and TGR+/– rats. Although differences in weight between hypertensive and control rats could be noticed in LH and TGR+/– rats, LVMI values were in accordance with previous reports in the three models (15, 18). LVMI was highly positively correlated to BP among the different groups (Fig. 1A). About 50–60% of the transcripts were declared Present in each group, and their average signals showed <10% variation between the groups (Table 2). The 6,452 selected transcripts represented 40.7% of the coding sequences spotted on the microarrays.


View this table:
[in this window]
[in a new window]

  		Table 2. Cardiovascular characteristics and Affymetrix GeneChips expression data of the 12-wk-old hypertensive rats of the 3 models

 

Figure 1
View larger version (13K):
[in this window]
[in a new window]

  		Fig. 1. General characteristics of the 6 groups of rats. A: relationship between systolic blood pressure and left ventricular mass index calculated as the ratio of left ventricular weight to body weight. Each point is the mean ± SE of the 2 variables in each group of rats. B: hierarchical clustering of rats according to gene expression of the 6,452 genes declared Present in all the rats. Left: expected clustering grouping first all hypertensive and all normotensive rats. Right: observed clustering. The vertical line indicates the distance at which all rats of each model are clustered. Strains: TGR+/–, transgenic heterozygous TGR(mRen2)27; TGR–/–, nontransgenic control; WKY, Wistar-Kyoto; SHR, spontaneously hypertensive; LH, Lyon hypertensive; LL, Lyon low-blood pressure control.

 
Assessment of Interstrain Variability of Transcript Levels
Comparing the three groups of normotensive rats, SAM multiclass analysis identified 4,129 (64%) of the 6,452 selected transcripts showing significant differential expression between the three control groups, pointing out the large extent of interstrain variability of gene expression. Cluster analysis of all rats according to expression of the 6,452 selected genes also highlighted interstrain variability as shown by the hierarchical tree (Fig. 1B). Indeed, following the horizontal axis expressing the distance between expression profiles, hypertensive and normotensive rats from the same model were first assembled at a low distance and finally clustered according to their genetic background.

Conversely, multiclass SAM analysis comparing the six groups of rats allowed detection of genes with low intergroup variations corresponding to the largest q-values. A set of 191 genes with q-value >25% was selected to be considered as housekeeping genes in myocardium and may subsequently serve for interexperimental data normalization (Supplemental Table 1; Supplemental Materials are available in the online version of this article).

Differentially Expressed Genes in Hypertensive Rats from Each Model
Variability in gene expression was also observed for several LVH markers (Fig. 2). At this early stage of LVH development, only certain markers were differentially expressed in the various hypertension models. Atrial and brain natriuretic peptides (Nppa and Nppb) showed increased expression in TGR+/– and LH rats, reaching statistical significance only for Nppa in TGR+/– rats. The ATPase, Ca++ transporting, slow twitch 2, Atp2a2, alias Serca-2, transcript was less abundant in SHR compared with WKY rats but showed no difference in the two other comparisons. The myosin heavy chain shift was not detected.


Figure 2
View larger version (35K):
[in this window]
[in a new window]

  		Fig. 2. Expression intensities of classical so-called markers of left ventricular hypertrophy obtained in each group of hypertensive rats. Data are means ± SE. *P < 0.05: hypertensive vs. control rats for each model.

 
The main features of the genes differentially expressed in TGR+/–, SHR, and LH rats compared with their respective controls (TGR–/–, WKY, and LL rats) are given in Fig. 3. Mean expression ratios of differentially expressed genes were similar in the three models (SHR vs. WKY, 1.21 ± 0.05; LH vs. LL, 1.27 ± 0.14; TGR+/– vs. TGR–/–, 1.28 ± 0.05; ANOVA, P = 0.92). Expression ratios >2 or <0.5 were observed for a larger number of genes in SHR (45 genes) and LH (37 genes) rats than in TGR+/– rats (4 genes) (Fig. 3A). A large number of genes were differentially expressed in SHR vs. WKY or in LH vs. LL rats, whereas in TGR+/– vs. TGR–/– rats, this number was dramatically lower (14- to 18-fold). The Venn diagram (Fig. 3B) summarizes these results, showing that only four genes were common to the three comparisons. Three of them, follistatin-like 1 (Fstl1), coding for a binding protein for transforming growth factor superfamily (antagonist of activin signaling), phosphoglucomutase 1 (Pgm1), involved in the early glucose metabolism and storage, and an EST (accession no. BI295828) highly similar to synaptopodin 2-like predicted gene (Synpo2l), had increased expression in TGR+/– rats but decreased in either SHR or LH rats. The fourth gene, Siat7A, coding for sialyltransferase 7A, adding neuraminidic acid residues to glycanes, was overexpressed in the three hypertensive groups with similar mean expression ratios (1.31 to 1.33). Expression changes of these four genes are given in Fig. 4 and have been confirmed by qRT-PCR. Validation of microarray expression is given as well for Nppa and Atp2a2 genes.


Figure 3
View larger version (13K):
[in this window]
[in a new window]

  		Fig. 3. Genes differentially expressed in TGR+/–, SHR, and LH rats compared with TGR–/–, WKY, and LL rats, respectively. A: box plots representing distributions of expression ratios for each comparison. B: Venn diagram showing the no. of differentially expressed genes for each comparison and the no. of common genes.

 

Figure 4
View larger version (43K):
[in this window]
[in a new window]

  		Fig. 4. Validation of microarray gene expression (array; A) by quantitative real-time RT-PCR (qPCR) of 6 genes in TGR–/– (n = 6), TGR+/– (n = 6), WKY (n = 5), SHR (n = 5), LL (n = 5), and LH (n = 5) rats, using similar mRNA samples for both methods. For each, i.e., TGR+/– (left), SHR (middle), and LH (right), less-shaded or open bars represent controls (TGR–/–, WKY, and LL rats), and shaded or solid bars are TGR+/–, SHR, and LH rats, respectively. mRNA level for arrays is based on signal intensities and for qRT-PCR on calculated expression x108 for Siat7A and x109 for the other genes. Data are means ± SE. *P < 0.10 [significance analysis of microarrays (SAM)] and {dagger}P < 0.05 (Mann-Whitney test): TGR+/– vs. TGR–/–, SHR vs. WKY, or LH vs. LL.

 
Among the 67 distinct genes showing expression changes in TGR+/– rats (Table 3), only a few genes involved in cellular growth (connective tissue growth factor, Ctgf; transforming growth factor beta 2, Tgfb2), extracellular matrix (ECM) (thrombospondin 2, Thbs2), or cytoskeleton (tropomyosin 1), as well as Nppa, had increased expression by >50%. A large number of genes involved in metabolism showed weaker expression increases. Only 16 genes showed decreased expression, by 16 to 38%, including 9 genes involved in metabolism. Hierarchical clustering of these genes (TGR clustering, Fig. 5) showed that 1) the four probes targeting the same EST (BF561368 similar to C11orf17 protein, targeting RGD1306959 unknown gene) were clustered at a very low distance, 0.08, corresponding to a high Pearson correlation coefficient of 0.92, and 2) cutting the hierarchical tree at a distance of 0.292, corresponding to a Pearson correlation coefficient of 0.708 (P = 0.01, n = 12), led to six clusters containing at least five genes. Interestingly, Siat7A was clustered with Tgfb2, cyclin D1 (Ccnd1), CD151 antigen (Cd151), and melanoma cell adhesion molecule (Mcam), which are involved in various cell processes from cell adhesion to regulation of cell cycle, and Nppa was found close to Fstl1 and to two serine protease inhibitors, Serpinf1 and Serping1.


View this table:
[in this window]
[in a new window]

  		Table 3. Genes with significant expression changes in TGR+/– rats compared with TGR–/– rats

 

Figure 5
View larger version (36K):
[in this window]
[in a new window]

  		Fig. 5. Hierarchical clustering of genes differentially expressed in TGR+/– rats. Rows indicate genes, and columns indicate rats. Expression values of each gene are encoded into 6 levels using quantiles of each distribution. The hierarchical tree is represented according to distance between genes, computed as d = 1 – r (where r is Pearson correlation coefficient). The vertical blue line defines the statistical significance threshold of the clusters (r = 0.708, n = 12, P = 0.01).

 
Genes differentially expressed in both TGR+/– rats and SHR or LH rats may identify pathways common to the different models (Table 3). Distinct genes were detected for SHR and LH rats. Of the 20 genes common to SHR and TGR+/– rats, 9 were overexpressed in both, including four and a half LIM domains 1 (Fhl1) and arginine vasopressin-induced 1 (Avpi1) genes. Interestingly, one predicted gene, similar to C11orf17 protein, had twofold increased expression in both SHR and TGR+/– rats and was moreover correlated with LVMI, pointing thus to its potential implication in LVH development. Several genes, including Nppa and Ccnd1, exhibited opposite variations of expression in SHR and TGR+/– rats. In a similar way, of the 14 transcripts common to TGR+/– and LH rats, Hspb3, member 3 of the heat shock 27-kDa protein family, Gadd45a (growth arrest and DNA damage-inducible 45 alpha), and Prss23 (protease serine 23) were overexpressed in both models, whereas Ech1 (enoyl-CoA-hydratase) was underexpressed.

qRT-PCR Validation of Microarray Expression Changes
Microarray expression changes were validated by qRT-PCR for the following six genes: Nppa, Atp2a2, Siat7A, Pgm1, Fstl1genes, and Synpo2l predicted gene. Figure 4 gives the mRNA levels obtained with both methods for each of the six groups of rats. All expression changes that were found significant with microarray analysis were also found significant with qRT-PCR, in the same direction with much more important changes. In addition, unchanged expression of Nppa in LH rats was detected as significantly increased with qRT-PCR. Similarly, unchanged expression of Atp2a2 in TGR+/– and LH rats was detected as significantly reduced in TGR+/– but increased in LH rats.

Genes Correlated with LVMI
Correlation analysis between LVMI and gene expression was performed over the 15,866 genes spotted on the microarrays. It disclosed 57 and 50 transcripts positively and negatively correlated to LVMI, respectively (Table 4). Most of these genes (63%) did not show any statistically significant difference whatever the expression comparison between hypertensive and normotensive rats. Figure 6 gives scatterplots and regression lines obtained for some genes, showing positive or negative correlations between expression and LVMI.


View this table:
[in this window]
[in a new window]

  		Table 4. Genes with expression highly correlated with left ventricular mass index through all the 32 studied rats

 

Figure 6
View larger version (14K):
[in this window]
[in a new window]

  		Fig. 6. Examples of regression plots between expression intensity and LVMI for several positively (A) and negatively (B) correlated genes. r is Pearson correlation coefficient.

 
The positively correlated transcripts were mostly genes involved in metabolism, cytoskeleton, ECM, or cell growth, with Siat7A and Tgfb2 showing the highest correlation coefficients. The negatively correlated transcripts were mostly metabolism and cytoskeleton genes, as well as signal transduction genes such as G protein ß3-subunit (Gnb3), RASD member 2 (Rasd2), and MAP kinase kinase (Map2k6), or alpha1B adrenergic receptor (Adra1b). Interestingly, when clustering all gene transcripts correlated to LVMI, these four genes were grouped with a high correlation, suggesting the existence of a coordinated regulation at different levels of the G protein-coupled MAP kinase pathway (Supplemental Figure).

Chromosomal Localization of Selected Genes
Using Caryoscope software, all differentially expressed genes and all genes correlated to LVMI were positioned along the chromosomes. Differentially expressed genes were widespread on the different chromosomes, representing 4–11% of spotted genes for each chromosome in SHR and LH strains, with the largest percentage for chromosome 20 (10.3% and 11.7%, respectively), followed by chromosomes 9 and 18 in SHR and chromosomes 6 and 16 in LH rats (7.5–9.9%). Comparison of the distributions of the differentially expressed genes and of the whole set of probes of the array across the chromosomes showed that they were different for both SHR (P = 0.003) and LH rats (P = 0.026). Therefore, the large percentages of genes observed on some chromosomes were not due to hazard. The number of genes selected by the TGR+/– comparison or the correlation between expression and LVMI was much lower, leading to percentages of spotted genes between 0.2 and 1.6%. The largest percentage was obtained for chromosome 13 for TGR+/– rats and for chromosome 11 for LVMI correlation.

Twenty preferential chromosomal regions were selected based on the joint presence of differentially expressed genes in the three hypertensive models and of genes correlated to LVMI (Fig. 7). These regions, spanning 0.8 to 8.3 Mbp, were shared out in 11 of the 21 chromosomes. In each selected region, the proportion of detected genes ranged from 9 up to 50% of the known genes in the region, with 22% as mean value. Interestingly, most of these regions were within BP or cardiac mass QTLs identified in SHR or LH rats available from RGD. The fraction of regions within QTLs was higher than the calculated probability of inclusion of the regions in QTLs by chance for 8 of the 11 concerned chromosomes (Supplemental Table 2). It was lower for chromosomes 4 and 11, in which our selected regions were not within any QTL, and no QTLs have been identified on chromosome 15 for the studied rat strains. Chromosomes 1, 2, 5, and 10 were overrepresented with three regions each. Interestingly, Siat7A belongs to the 106.1- to 110.0-Mbp region on chromosome 10.


Figure 7
View larger version (19K):
[in this window]
[in a new window]

  		Fig. 7. Chromosomal mapping of all the genes identified from between-group comparisons and from correlation analysis between expression data and LVMI. Blood pressure and cardiac mass QTL regions already identified in SHR or LH rats are indicated. For each comparison, expression data are represented as log ratios, with negative values on the left of the chromosomal line and positive values on the right. Similarly, negative and positive correlation coefficients with LVMI (rLVMI) are at left and right, respectively. Selected chromosomal regions are represented with gray background. *Chromosomes for which the selected regions were not related with any QTL.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
GRANTS
 REFERENCES
 
In this study, by examining global gene expression in the left ventricle of young adult rats of three models representative of polygenic and monogenic hypertension with Affymetrix GeneChips, we showed 1) Siat7A as an important factor overexpressed in all three studied models, the expression of which is positively correlated with LVMI; 2) a set of genes presenting quantitative relationships between LVMI and expression levels across the different models studied; and 3) specific chromosomal locations where several genes associated with LVH are concentrated.

The use of young adult animals allowed us to investigate an early stage of LVH development, before the onset of any heart dysfunction. The high correlation between LVMI and systolic BP over the six groups of rats clearly indicated that, independently of the genetic background, cardiac muscle mass adapts to the workload. However, the overall variability of heart transcriptome is mainly model dependent, hypertension being clearly hidden behind the strain factor as shown by the rat clustering. This result suggests that each hypertensive model develops its own transcriptional adaptations in the frame of its genetic background.

The comparisons of expression levels between hypertensive and normotensive rats for each model showed that expression changes were low, since a very limited number of transcripts displayed expression ratios higher than 2 or lower than 0.5. This may be related to the underestimation of expression changes found with the microarrays, since qRT-PCR gave similar but much more marked changes for all the validated genes. Statistical significance could be detected in certain cases for which microarray analysis did not reveal any change. In addition, microarrays may show limited sensitivity for low-abundance mRNAs, such as those previously studied by our team (14, 15). However, SHR and LH strains exhibited more differentially expressed genes and much higher or lower extreme expression ratios than TGR+/– rats. This may result both from the selection process of SHR and LH rat strains and from the polygenic character of hypertension in these rats contrarily to TGR+/– rats, which differ from their controls only by one additional gene. Our results pointed out several classical markers of LVH that were detected in at least one group of hypertensive rats. However, Nppa was clearly underexpressed in SHR compared with WKY rats, and this decrease was confirmed by qRT-PCR. This result has already been shown in SHR between 8 and 20 wk of age (3). Reduced expression of Atp2a2 was observed in SHR with microarrays and in TGR+/– rats with qRT-PCR only, in accordance with previous studies in mice with angiotensin II treatments (16) or, to a lesser extent, with aortic constriction (30) or in double-transgenic rats harboring the human angiotensinogen and renin genes (29). The classical shift in myosin heavy chain composition was not detected (6, 25) with the microarrays. This may be a consequence of the loss of linearity between signal detected and mRNA amount for high expression levels.

Among the 6,452 studied transcripts, Siat7A was the only gene that was repeatedly overexpressed in the three hypertensive models and correlated to LVMI. Thus this sialyltransferase could be considered as a potential specific marker for hypertension-induced LVH. Interestingly, serum sialic acid has been proposed as a marker of cardiovascular morbidity (13), and heart sialic acid content was shown to be increased during heart failure (7). TGR clustering showed that Siat7A was closely associated with other overexpressed genes, such as Tgfb2 and Mcam, contributing to ECM, thus suggesting that these genes share common causes and may be involved in LVH initiation. Positive correlation between LVMI and expression level observed for Tgfb2 and several ECM genes (thrombospondin 2 and 4, Thbs2 and Thbs4; or tissue inhibitor of metalloproteinase 1, Timp1) may be similarly interpreted.

Fstl1 was found overexpressed in TGR+/– rats but underexpressed in the two hypertensive inbred strains. By use of microarrays, expression of Fstl1 has been shown to be increased in 20-mo-old SHR (20) and in some mouse models (2, 30) but decreased in another mouse model with less severe LVH (2). Fst1 was closely associated with Nppa and two serine peptidase inhibitors in TGR clustering. Because Nppa is now considered a counterregulatory mechanism during LVH (4, 5) and Fstl1 belongs to a family of inhibitors of transforming growth factor signaling, the whole cluster may act as a counterregulation of LVH.

Taken as a whole, our results could suggest that small changes within regulatory networks rather than large changes in a single molecular complex may be common to different hypertension models. Indeed, detailed analysis of the genes selected from the different comparisons highlights the fact that similar regulations may be achieved through different target genes. For example, Ccnd1 was underexpressed in SHR rats, possibly contributing to a blockade in the G1B phase of the cell cycle, which is typical of cardiomyocyte hypertrophy. The same blockade may be mediated in the two other models by an increase of Gadd45a mRNA, whose encoded protein is known to bind cyclin D1 and to block its function. Likewise, monoaminergic transmission through a catecholaminergic pathway may be downregulated, as suggested by the increased expression of Maoa in SHR and LH rats and the negative correlations of Adra1b, Gnb3, and Map2k6 expression with LVMI.

Testing LVMI as a linear predictor of mRNA level took into account the interstrain variability and thus allowed the selection of genes whose expression is quantitatively related to LVMI regardless of the genetic background. Whatever the direction of causality, 1) these genes are candidate markers of LVH in hypertension, and 2) genetic variations of expression of these genes, if translated into modified protein activity, could interfere with LVH development in hypertension. Some genes such as Timp1, Dscr1, Thbs4, and Thbs2 had expression levels correlated with LVMI and were, moreover, differentially expressed in at least one group of hypertensive rats, in accordance with other microarray studies (10, 16, 20, 21). Another set of genes was uncovered that did not exhibit any significant difference in expression between hypertensive and normotensive rats and thus may also represent good "candidates" more closely associated with LVMI changes, regardless of the genetic background. For such genes, the high degree of correlation with LVMI is not explained by large variations between hypertensive and normotensive rats but rather by both intra- and intergroup variability, the latter mainly accounting for correlation. These genes may not yet have significant expression changes at this early stage of LVH development.

Positively correlated genes such as Tgfb2 and ECM genes may be causal in LVH development. By contrast, the negatively correlated genes relative to cardiac signaling pathways (Map2k6, Gnb3, Rasd2, and Adra1b) may be involved in a negative regulation following LVH formation. Most of them have already been associated with LVH (26, 28). These transcripts could be directly associated with a G protein-coupled pathway involved in cardiomyocyte growth. In addition, the presence of genes coding for proteins involved in RNA and DNA editing such as NonO/p54nrb (Sfpq), ribonucleoprotein B (Snrpb2), and Siah-binding protein (Siahbp1) suggests that these splicing-related genes may have a global activating/inhibiting influence on the chromosomal regions identified in this study.

Various altered gene expressions across the different models may arise from the genetic effects on transcript abundance of cis- and trans-acting factors. Hubner et al. (11) identified cis- and trans-acting expression QTLs (eQTLs) in kidney and fat of SHR and Brown-Norway recombinant inbred hypertensive rat strains. Interestingly, the Adra1b region on chromosome 10 was described as a cis-acting eQTL in the kidney. Moreover, 6 of our 20 chromosomal regions of interest contained cis-acting eQTLs described in both kidney and fat (11). Thus the genes identified in this study and located within these cis-acting regions may be particularly interesting and should be investigated further.

In conclusion, we have shown that hypertension-induced LVH at an early stage is not associated with a unique pattern of cardiac cell reprogramming but with minor changes in gene expression that seem to be maintained within the interstrain variability. The only common gene overexpressed in the three rat models is Siat7A, coding for a sialyltransferase. Genes whose expression is correlated with LVMI also may represent other genes associated with LVH development. Finally, chromosomal regions gathering together genes with altered expression in relation to hypertension or LVH were identified within BP and cardiac mass QTLs, thus pointing to candidate regions for hypertension.


   GRANTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
REFERENCES
 
This work was supported by a grant from Fondation de France and by Ministère de la Recherche (for Affymetrix experiments).


   ACKNOWLEDGMENTS
 
We thank Alain Bataillard for providing LH and LL rats. We are grateful to Nicole Dizerens for technical help and John McGregor and Liliane McGregor for helpful discussions.


   FOOTNOTES
 
Address for reprint requests and other correspondence: C. Cerutti, EA 3740 Génomique fonctionnelle dans l'athérothrombose, Faculté de Médecine Laennec, Université Lyon 1, 7 rue Guillaume Paradin, 69372 Lyon Cedex 08, France (e-mail: cerutti{at}univ-lyon1.fr)

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


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 

  1. Arnett DK, de las Fuentes L, and Broeckel U. Genes for left ventricular hypertrophy. Curr Hypertens Rep 6: 36–41, 2005.
  2. Aronow BJ, Toyokawa T, Canning A, Haghighi K, Delling U, Kranias E, Molkentin JD, and Dorn JW 2nd. Divergent transcriptional responses to independent genetic causes of cardiac hypertrophy. Physiol Genomics 6: 19–28, 2001.[Abstract/Free Full Text]
  3. Bell D, Kelso EJ, Argent CCH, Lee GR, Allen AR, and McDermott BJ. Temporal characteristics of cardiomyocyte hypertrophy in the spontaneously hypertensive rat. Cardiovasc Pathol 13: 71–78, 2004.[Web of Science][Medline]
  4. Booz GW. Putting the brakes on cardiac hypertrophy: exploiting the NO-cGMP counter-regulatory system. Hypertension 45: 341–346, 2005.[Abstract/Free Full Text]
  5. Bubikat A, De Windt LJ, Zetsche B, Fabritz L, Sickler H, Eckardt D, Godecke A, Baba HA, and Kuhn M. Local atrial natriuretic peptide signaling prevents hypertensive cardiac hypertrophy in endothelial nitric-oxide synthase-deficient mice. J Biol Chem 280: 21594–21599, 2005.[Abstract/Free Full Text]
  6. Chassagne C, Wisnewsky C, and Schwartz K. Antithetical accumulation of myosin heavy chain but not alpha-actin mRNA isoforms during early stages of pressure-overload-induced rat cardiac hypertrophy. Circ Res 72: 857–864, 1993.[Abstract/Free Full Text]
  7. Crook JR, Goldman JH, Dalziel M, Madden B, and McKenna WJ. Increased ventricular sialylation in patients with heart failure secondary to ischemic heart disease. Clin Cardiol 20: 455–458, 1997.[Web of Science][Medline]
  8. Deschepper CF, Masciotra S, Zahabi A, Boutin-Ganache I, Picard S, and Reudelhuber TL. Functional alterations of the Nppa promoter are linked to cardiac ventricular hypertrophy in WKY/WKHA rat crosses. Circ Res 88: 223–228, 2001.[Abstract/Free Full Text]
  9. 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]
  10. Heymans S, Schroen B, Vermeersch P, Milting H, Gao F, Kassner A, Gillijns H, Herijgers P, Flameng W, Carmeliet P, Van de Werf F, Pinto YM, and Janssens S. Increased cardiac expression of tissue inhibitor of metalloproteinase-1 and tissue inhibitor of metalloproteinase-2 is related to cardiac fibrosis and dysfunction in the chronic pressure-overloaded human heart. Circulation 112: 1136–1144, 2005.
  11. Hubner N, Wallace CA, Zimdahl H, Petretto E, Schulz H, Maciver F, Mueller M, Hummel O, Monti J, Zidek V, Musilova A, Kren V, Causton H, Game L, Born G, Schmidt S, Muller A, Cook SA, Kurtz TW, Whittaker J, Pravenec M, and Aitman TJ. Integrated transcriptional profiling and linkage analysis for identification of genes underlying disease. Nat Genet 37: 243–253, 2005.[CrossRef][Web of Science][Medline]
  12. Kato N, Hyne G, Bihoreau MT, Gauguier D, Lathrop GM, and Rapp JP. Complete genome searches for quantitative trait loci controlling blood pressure and related traits in four segregating populations derived from Dahl hypertensive rats. Mamm Genome 10: 259–265, 1999.[CrossRef][Web of Science][Medline]
  13. Knuiman MW, Watts GF, and Divitini ML. Is sialic acid an independent risk factor for cardiovascular disease? A 17-year follow-up study in Busselton, Western Australia. Ann Epidemiol 14: 627–632, 2004.[CrossRef][Web of Science][Medline]
  14. Kurdi M, Cerutti C, Randon J, McGregor L, and Bricca G. Macroarray analysis in the hypertrophic left ventricle of renin-dependent hypertensive rats: identification of target genes for renin. J Renin Angiotensin Aldosterone Syst 5: 72–78, 2004.[Abstract/Free Full Text]
  15. Kurdi M, Randon J, Cerutti C, and Bricca G. Increased expression of IL-6 and LIF in the hypertrophied left ventricle of TGR(mRen2)27 and SHR rats. Mol Cell Biochem 269: 95–101, 2005.[CrossRef][Web of Science][Medline]
  16. Larkin JE, Frank BC, Gaspard RM, Duka I, Gavras H, and Quackenbush J. Cardiac transcriptional response to acute and chronic angiotensin II treatments. Physiol Genomics 18: 152–166, 2004.[Abstract/Free Full Text]
  17. Mullins JJ, Peters J, and Ganten D. Fulminant hypertension in transgenic rats harbouring the mouse Ren-2 gene. Nature 344: 541–544, 1990.[CrossRef][Medline]
  18. Ohta K, Kim S, and Iwao H. Role of angiotensin-converting enzyme, adrenergic receptors, and blood pressure in cardiac gene expression of spontaneously hypertensive rats during development. Hypertension 28: 627–634, 1996.[Abstract/Free Full Text]
  19. Pravenec M, Zidek V, Musilova A, Vorlicek J, Kren V, St Lezin E, and Kurtz TW. Genetic isolation of a blood pressure quantitative trait locus on chromosome 2 in the spontaneously hypertensive rat. J Hypertens 19: 1061–1064, 2001.[CrossRef][Web of Science][Medline]
  20. Rysa J, Leskinen H, Ilves M, and Ruskoaho H. Distinct upregulation of extracellular matrix genes in transition from hypertrophy to hypertensive heart failure. Hypertension 45: 927–933, 2005.[Abstract/Free Full Text]
  21. Schroen B, Heymans S, Sharma U, Blankesteijn WM, Pokharel S, Cleutjens JP, Porter JG, Evelo CT, Duisters R, van Leeuwen RE, Janssen BJ, Debets JJ, Smits JF, Daemen MJ, Crijns HJ, Bornstein P, and Pinto YM. Thrombospondin-2 is essential for myocardial matrix integrity: increased expression identifies failure-prone cardiac hypertrophy. Circ Res 95: 515–522, 2004.[Abstract/Free Full Text]
  22. Strom CC, Kruhoffer M, Knudsen S, Stensgaard-Hansen F, Jonassen TEN, Orntoft TF, Haunso S, and Sheikh SP. Identification of a core set of genes that signifies pathways underlying cardiac hypertrophy. Compar Funct Genomics 5: 459–470, 2004.[CrossRef]
  23. Swynghedauw B. Molecular mechanisms of myocardial remodeling. Physiol Rev 79: 215–262, 1999.[Abstract/Free Full Text]
  24. Twigger SN, Pasko D, Nie J, Shimoyama M, Bromberg S, Campbell D, Chen J, Cruz Nd Fan C, Foote C, Harris G, Hickmann B, Ji Y, Jin W, Li D, Mathis J, Nenasheva N, Nigam R, Petri V, Reilly D, Ruotti V, Schauberger E, Seiler K, Slyper R, Smith J, Wang W, Wu W, Zhao L, Zuniga-Meyer A, Tonellato PJ, Kwitek AE, and Jacob HJ. Tools and strategies for physiological genomics: the Rat Genome Database. Physiol Genomics 23: 246–256, 2005.[Abstract/Free Full Text]
  25. van Bilsen M and Chien K. Growth and hypertrophy of the heart: towards an understanding of cardiac specific and inducible gene expression. Cardiovasc Res 27: 1140–1149, 1993.[Web of Science][Medline]
  26. Vecchione C, Fratta L, Rizzoni D, Notte A, Poulet R, Porteri E, Frati G, Guelfi D, Trimarco V, Mulvany MJ, Agabiti-Rosei E, Trimarco B, Cotecchia S, and Lembo G. Cardiovascular influences of {alpha}1b-adrenergic receptor defect in mice. Circulation 105: 1700–1707, 2002.
  27. Vincent M and Sassard J. Lyon strains. In: Handbook of Hypertension. Genetic Models of Hypertension, edited by Swales JD. Oxford: Blackwell Scientific, 1994, p. 455–457.
  28. Weinberg EO, Mirotsou M, Gannon J, Dzau VJ, Lee RT, and Pratt RE. Sex dependence and temporal dependence of the left ventricular genomic response to pressure overload. Physiol Genomics 12: 113–127, 2003.[Abstract/Free Full Text]
  29. Wellner M, Dechend R, Park JK, Shagdarsuren E, Al Saadi N, Kirsch T, Gratze P, Schneider W, Meiners S, Fiebeler A, Haller H, Luft FC, and Muller DN. Cardiac gene expression profile in rats with terminal heart failure and cachexia. Physiol Genomics 20: 256–267, 2005.[Abstract/Free Full Text]
  30. Zhao M, Chow A, Powers J, Fajardo G, and Bernstein D. Microarray analysis of gene expression after transverse aortic constriction in mice. Physiol Genomics 19: 93–105, 2004.[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 Am. J. Physiol. Heart Circ. Physiol.Home page
J. Peters, T. Schluter, T. Riegel, B. S. Peters, A. Beineke, U. Maschke, N. Hosten, J. J. Mullins, and R. Rettig
Lack of cardiac fibrosis in a new model of high prorenin hyperaldosteronism
Am J Physiol Heart Circ Physiol, November 1, 2009; 297(5): H1845 - H1852.
[Abstract] [Full Text] [PDF]

Home page
 CirculationHome page
R. Sarwar and S. A. Cook
Genomic Analysis of Left Ventricular Remodeling
Circulation, August 4, 2009; 120(5): 437 - 444.
[Full Text] [PDF]

Home page
 Physiol. GenomicsHome page
M.-P. Gustin, C. Z. Paultre, J. Randon, G. Bricca, and C. Cerutti
Functional meta-analysis of double connectivity in gene coexpression networks in mammals
Physiol Genomics, June 1, 2008; 34(1): 34 - 41.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplementary Tables and Figure
Right arrow All Versions of this Article:
27/3/295    most recent
00318.2005v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (2)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Cerutti, C.
Right arrow Articles by Gustin, M.-P.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Cerutti, C.
Right arrow Articles by Gustin, M.-P.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Visit Other APS Journals Online
Copyright © 2006 by the American Physiological Society.