Familial hypertrophic cardiomyopathy (FHC) is a disease characterized by ventricular hypertrophy, fibrosis, and aberrant systolic and/or diastolic function. We previously developed two transgenic mouse models that carry FHC-associated mutations in α-tropomyosin (TM): FHC α-TM175 mice show patchy areas of mild ventricular disorganization and limited hypertrophy, whereas FHC α-TM180 mice exhibit severe hypertrophy and fibrosis and die within 6 mo. To obtain a better understanding of the molecular mechanisms associated with the early onset of cardiac hypertrophy, we conducted a detailed comparative analysis of gene expression in 2.5-mo-old control, FHC α-TM175, and α-TM180 ventricular tissue. Results show that 754 genes (from a total of 22,600) were differentially expressed between the nontransgenic (NTG) and the FHC hearts. There are 178 differentially regulated genes between NTG and the FHC α-TM175 hearts, 388 genes are differentially expressed between NTG and FHC α-TM180 hearts, and 266 genes are differentially expressed between FHC α-TM175 and FHC α-TM180 hearts. Genes that exhibit the largest increase in expression belong to the “secreted/extracellular matrix” category, and those with the most significant decrease in expression are associated with “metabolic enzymes.” Confirmation of the microarray analysis was conducted by quantitative real-time PCR on gene transcripts commonly associated with cardiac hypertrophy.
- familial hypertrophic cardiomyopathy mutations
familial hypertrophic cardiomyopathy (FHC) is associated with mutations in contractile proteins of the cardiac sarcomere, including myosin heavy and light chains (MHC, MLC), actin, tropomyosin (TM), troponin T and I, and myosin binding protein C. The disease is characterized by myocyte disarray, asymmetric ventricular hypertrophy, fibrosis, and cardiac arrhythmias that often result in heart failure and death. The mechanism(s) by which mutations in sarcomeric contractile proteins initially trigger the hypertrophic response is poorly understood.
To obtain a better understanding of the molecular mechanisms associated with the development of FHC, numerous laboratories have generated transgenic and knockout mouse models expressing mutations in structural and sarcomeric proteins. α-TM is an essential contractile protein involved in the regulation of sarcomeric function through its interaction with other thin filament components. Eight distinct missense mutations (Glu62Gln, Ala63Val, Lys70Thr, Val95Ala, Asp175Asn, Glu180Gly, Glu180Val, Leu185Arg) in α-TM are associated with FHC, which manifest variable degrees of cardiac hypertrophy (5, 7, 9, 10, 15, 24). In previous work, we generated transgenic mouse models for two FHC α-TM mutations (Asp175Asn and Glu180Gly) that lie within the internal troponin T binding region (14, 18, 19). Physiological analyses of hearts from both model systems show impairment of contractility and relaxation. Interestingly, histological analysis reveals that the FHC α-TM175 mice show patchy areas of mild ventricular disorganization and hypertrophy, without premature mortality or a significant difference in heart weight-body weight ratio. The FHC α-TM180 mice develop severe cardiac hypertrophy with an increased heart weight-body weight ratio, substantial interstitial fibrosis, and atrial and ventricular calcification and mineralization that culminate in lethality before 6 mo. Thus the two FHC models bearing single amino acid point mutations in the same α-TM region have similar physiological dysfunction in cardiac sarcomere performance, yet vastly different pathological profiles with resulting mortality. This situation offers a unique opportunity to examine the molecular changes in cardiac gene expression that occur during the initial phases of cardiac hypertrophy that result in different phenotypes.
In the present study, we hybridized RNA from 2.5 mo-old ventricular tissue of nontransgenic (NTG) control, α-TM175, and α-TM180 hearts to Affymetrix MOE430A GeneChip arrays containing 22,600 well-characterized mouse genes/expressed sequence tags. Samples used for hybridization consisted of pooled and nonpooled RNA extracts from the three groups. Comprehensive statistical analysis shows that 754 genes were differentially expressed between the NTG and the FHC hearts at a 1.3-fold change (P ≤ 0.01). Of these genes, 178 genes were differentially expressed in the ventricles between NTG and FHC α-TM175 hearts; 388 genes were differentially expressed between NTG and FHC α-TM180 hearts; and 266 genes are differentially expressed between FHC α-TM175 and FHC α-TM180 hearts. Transcripts that exhibit the greatest increase belong to the “secreted/extracellular matrix” category, which may reflect the increased fibrosis and myocyte disarray associated with the α-TM180 hearts. Other significant changes in transcript levels occurred in genes classically associated with cardiac hypertrophy [i.e., atrial natriuretic factor (ANF), β-MHC, sarco(endo)plasmic reticulum Ca2-ATPase (SERCA)]. To confirm the microarray analysis, we conducted quantitative RT-PCR analyses on many different mRNAs; results from this analysis were in strong concordance with the microarray data. Thus this study provides significant insight into the diverse array of genes that are active during the early signaling of mild and severe cardiac hypertrophy.
MATERIALS AND METHODS
FHC α-TM175 and α-TM180 mice.
A comprehensive description on the generation of FHC α-TM175 and α-TM180 transgenic mice has been documented in our previous studies (14, 18, 19). In brief, the cardiac-specific α-MHC promoter was ligated to the α-TM cDNA encoding Asp175Asn or Glu180Gly mutations for α-TM175 and α-TM180, respectively. Expression of these mutant proteins in the transgenic hearts was at similar levels in both mouse models, and the mice exhibit similar physiological changes in cardiac performance (14, 18). Ventricular tissue RNA from 2.5-mo-old mice was prepared. The quality of the total RNA was analyzed with an Agilent 2001 Bioanalyzer to ensure that samples prepared for microarray hybridizations meet the Affymetrix guidelines.
Histological analysis of transgenic and control hearts.
Hearts from α-TM175, α-TM180, and NTG littermate male and female mice from 2.5-mo-old were isolated and fixed in 10% neutral buffered formalin. Dehydration was accomplished through alcohol and xylene gradients, followed by embedding in paraffin. Sections (5 μm) were prepared and stained with hematoxylin and eosin or with trichrome stain to assess fibrosis.
Microarray hybridization and data analysis.
Microarray hybridizations were performed using the Affymetrix MOE430A GeneChip array, which contains over 22,600 probe sets representing transcripts and variants from >14,000 well characterized mouse genes. Ten hybridization experiments were performed in which each genotypic group (NTG, α-TM175, and α-TM180) was represented by two or more nonpooled (NP) individual RNA extracts and one pooled (P) sample that resulted from combining 20 individual heart extracts using equal numbers of male and female mice. (One of the NP NTG gene chip samples and one of the NP α-TM175 samples did not work in the microarray experiments.) With the FHC α-TM175 and α-TM180 mice, we have not observed sex-biased differences in the FHC phenotype. Ten micrograms of total RNA were used to synthesize double-stranded cDNA with the SuperScript kit (Invitrogen), incorporating a T7 promoter sequence by using oligo dT-T7 primer. Biotin-labeled cRNAs were then generated from the cDNA and hybridized to Affymetrix MOE430A GeneChip arrays; hybridizations were performed at 45°C for 16 h in a GeneChip 640 hybridization oven. Arrays were stained with phycoerythrin-conjugated streptavidin (Molecular Probes, Eugene, OR), and hybridization signals were amplified using antibody amplification with goat IgG (Sigma-Aldrich) and antistreptavidin biotinylated antibody (Vector Laboratories, Burlingame, CA), as described in the Affymetrix GeneChip Expression Analysis Manual. Images were scanned using a GeneArray scanner (Agilent Technologies, Palo Alto, CA) and GeneChip cel files were subsequently processed by RMAExpress 0.1 (http://rmaexpress.bmbolstad.com) using the default options. RMA data were then loaded into GeneSpring 7.0 software (Silicon Genetics, Redwood City, CA), and all the samples were then normalized to the mean of the pooled control (NTG). Raw and RMA experimental data were deposited in the Gene Expression Omnibus database (http://www.ncbi.nlm.nih.gov/geo/) under the GSE Series accession number GSE4678. Gene expression in the TM mutant mouse hearts was examined relative to the average NTG heart. To do this, we first used RMAexpress, which normalizes gene expression in three steps: a background adjustment, quantile normalization (4), and finally summarization. Log-scaled gene expression levels for each probe set of each sample are then transformed to ratios relative to mean of expression level values in the NTG heart. Thus a ratio of 1 represents an equivalent level of expression in TG and NTG hearts. Using the average expression values within a genotype, we filtered genes that had a 1.3-fold change between two specified genotype groups. From this resulting list, a Welch t-test with a P value cut-off of 0.01 was performed to identify genes that were significantly regulated between groups being compared. Since there were three genotype groups, three comparisons were done: NTG vs. α-TM175, NTG vs. α-TM180, and α-TM175 vs. α-TM180. These results were pooled and subjected to hierarchical clustering using a Pearson correlation where genes were discarded that had no data in half of the conditions.
Real-time reverse transcription PCR analysis.
Select gene transcripts were subject to real-time quantitative polymerase chain reaction (RT-PCR) analysis. Gene-specific real-time RT-PCR primers (Table 1) were synthesized using either the data from the PrimerBank database (25), and/or by using the Primer3 program (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi) and verified for base complementarity (http://www.basic.northwestern.edu/biotools/oligocalc.html). cDNA was synthesized for 50 min at 50°C in a 20-μl reaction containing 1× First-Strand Buffer, 5 μg total RNA, 50 ng of random hexamers, 2 μM dNTPs, 40 units RNase inhibitor, and 200 units Superscript III reverse transcriptase (RT) (Invitrogen). Real-time PCR was performed in a 20-μl reaction, 96-well format [0.2 μl cDNA; 250 nM each primer; 1× DyNAmo HS SYBR Green Master mix (Finnzymes)] using an Opticon 2 real-time PCR machine (MJ Research). Three samples were measured in each experimental group in triplicate, with a minimum of two independent experiments. The relative amount of target mRNA normalized to GAPDH was calculated according to the method described by Pfaffl (17).
RESULTS AND DISCUSSION
Previous work in our laboratory established two transgenic mouse models expressing mutant FHC α-TM proteins (14, 18, 19). Both FHC mutations (Asp175Asn and Glu180Gln) are located in the internal TnT binding region of α-TM. The transgenic mice expressing these proteins exhibit cardiac abnormalities similar to those observed in human patients, including ventricular hypertrophy, fibrosis, diastolic dysfunction, and increased myofibrillar sensitivity to Ca2+. Although the α-TM protein is altered in the same region, two distinct phenotypes ensue: FHC α-TM 175 mice exhibit only mild hypertrophy and fibrosis affecting ∼5% of the myocardium (14) (Fig. 1), whereas FHC α-TM180 mice have severe pathological alterations and fibrosis (19) (Fig. 1). The similarities and disparities that exist between these two model systems present the unique opportunity to identify changes in gene expression that contribute to their different phenotypes. To examine the gene expression profiles of the two mouse models, we conducted extensive microarray analyses using NTG, FHC α-TM175, and FHC α-TM180 left ventricular tissue mRNA.
Total RNA from 2.5-mo-old mice was isolated and purified for use in gene expression matrices. This postnatal time was selected because the FHC α-TM180 mice at this time point are beginning to morphologically show features associated with cardiac hypertrophy (increased heart weight-body weight ratio) and are also beginning to manifest impaired diastolic dysfunction. Both individual (NP) and pooled (P) tissue samples from NTG, α-TM175, and α-TM180 ventricles were used to independently synthesize cDNA, followed by the generation of biotin-labeled cRNA. This cRNA was hybridized to the Affymetrix GeneChip Mouse Genome 430A array containing 22,600 probe sets representing 14,000 well-characterized mouse genes. Statistical analysis of the digitized data was performed with the GeneSpring 7.0 software.
The advantage of comparing gene expression profiles in closely related models of cardiac hypertrophy is the potential to identify model-specific molecular events. Therefore, we tested the hypothesis that model-specific gene expression changes could provide insight into cardiac hypertrophy and the different pathophysiology of the α-TM175 vs. α-TM180 mice. Results reveal there is significant divergence in gene expression among the NTG and the two FHC models (Fig. 2), with a total of 754 genes being differentially expressed (Supplemental Table 1). (The online version of this article contains supplemental data.) Results from the three comparisons (NTG vs. α-TM175 = 178 genes, NTG vs. α-TM180 = 388 genes, and α-TM175 vs. α-TM180 = 266 genes) revealed few genes commonly regulated across the model systems (Fig. 3). Of the 754 differentially expressed genes, 16 were specifically regulated in α-TM175, 24 were specifically regulated in α-TM180, and 34 were co-regulated in both α-TM175 and α-TM180. Only two genes were co-regulated in all the three comparisons made. The lists of these model-specific genes are shown in Tables 2⇓⇓–5 with the corresponding fold increase/decrease.
With the goal of discovering model-specific patterns, hierarchical clustering (Pearson correlation) was applied to the expression profiles of 754 genes with a pool of the three statistical group comparisons being performed. These data are represented by a “gene tree,” in which genes of most similar expression patterns are closest to one another within a node of the tree (Fig. 4). There are five distinct gene groups of interest among the FHC and NTG heart samples: 1) increased expression in α-TM175 only (group I), 2) increased expression in α-TM180 only (group V), 3) upregulated in both α-TM175 and α-TM180 (group III), 4) downregulated in both α-TM175 and α-TM180 (group II), and 5) decreased expression in α-TM175 only (group IV).
Further classification of the 754 genes according to molecular function as determined from the gene ontology (GO) listing found in the GeneSpring software shows the biggest category of differentially expressed genes were those belonging to the structural protein/extracellular matrix group (305/754). Because some transcripts have multiple functions, the sum total of molecular functions in Fig. 5 is greater than the number of genes being analyzed (1,212 functions vs. 754 genes). The next largest set of genes is associated with enzymes/metabolism (273 transcripts), followed by a substantial number of transcripts with unknown function (195 mRNAs). Another classification according to GO cellular component shows that 109 genes of the 754 belong to the extracellular matrix group.
As the heart progresses from normal to either the α-TM175 or α-TM180 phenotype, the heart is transitioning from a physiologically and morphologically normal state to being dysfunctional. This impaired condition is manifest by decreased rates of contraction and/or relaxation that occur in the hearts of these two mouse models at 2.5 mo of age. There are 34 genes that are differentially expressed in both the α-TM175 and α-TM180 mouse hearts, when compared with the NTG hearts (Table 2). These are transcripts that change their expression during the initial onset of cardiac hypertrophy and altered sarcomeric function. The associated genes can be classified into three broad categories: 1) decreases in expression of genes associated with energy metabolism, 2) increases in expression of genes associated with either the extracellular matrix or inflammation, and 3) genes with other heterogenous properties (Table 2). The associated changes in gene expression are consistent with the morphological and functional alterations. A decreased rate of contraction/relaxation without an increase in heart rate appears to be reflected by a reduction in energy production; this may reflect the downregulation of genes associated with energy metabolism. Also, with the development of fibrosis in the α-TM175 and α-TM180 hearts, one would expect increases in expression of extracellular matrix genes, as occurs in these hearts. Additional changes in gene expression that are very interesting are increased levels of osteoblast-specific factor 2 (OSF-2) and microfibrillar associated protein 5, which also occurs in other models of cardiac hypertrophy (1, 3).
In addition to genes whose expression was commonly altered, the α-TM175 and α-TM180 hearts also underwent changes that were specific to each model. An additional 16 genes were differentially expressed only in the α-TM175 hearts (Table 3). The expression of these select genes was altered when hypertrophy was confined to localized regions and the associated immunological response is more limited than what is observed in the more global effects of extensive cardiac disease found in the α-TM180 mice. Three of these genes are associated with the inflammatory response (interleukin 11, immunoresponse gene 1, vasoactive intestinal peptide receptor 1), and two are associated with the extracellular matrix (procollagen type XII α1 and desmocollin 2, a plasma membrane protein that associates with cadherin and other adhesion molecules). Five of the genes have undefined functions (RIKEN cDNAs), and others possess various functions, such as being a transcription factor (i.e., even skipped homeotic gene 2 homolog), chromatin binding (i.e., H1 histone family member), and involvement with protein degradation (i.e., ubiquitin-specific protease 3).
The α-TM180 hearts exhibit severe cardiac hypertrophy, fibrosis, and impaired function by 4 mo of age. However, by 2.5 mo postpartum, the onset of severe pathological change is already manifesting in these hearts, which are reflected in 24 altered gene transcript levels (Table 4). All of these transcripts have increased expression over control hearts except one, kallikrein 24, a vascular growth factor that has a role in apoptosis and angiogenesis by cleaving substrates, such as growth factors, hormones, or extracellular matrix components. Six of the genes are associated with the extracellular matrix, two are associated with apoptosis (annexin A5 and reticulon 4), and two are associated with the immune response (fibrinogen-like protein 2 and Tnf receptor-associated factor 3). Interestingly, myotrophin, which is associated with cardiac hypertrophy and cardiomyopathy (20), is significantly increased in the α-TM180 hearts by 2 mo, as is disintegrin and metalloprotease domain 10, which is involved in the release of angiotensin-converting enzyme.
Only two genes, transforming growth factor (TGF)-β3 and dihydropyrimidinase-related protein-3 are specifically co-regulated in all the three comparisons (Table 5). To identify genes that may be co-regulated with TGF-β3, a Pearson similarity measure (0.9) was applied to the expression profile of TGF-β3, which resulted in 237 genes (Fig. 6) from the 754 total genes (Fig. 2). These 237 genes show similar transcriptional regulation as TGF-β3, as well as passing stringent statistical filters, and their expression profiles are proportional to the severity of the pathological phenotype (i.e., mild in α-TM175 hearts and moderate to severe in α-TM180 hearts).
To verify the results of regulated gene expression detected in the microarray analysis, we conducted real-time quantitative RT-PCR analysis on a select number of transcripts associated with cardiac hypertrophy using ventricular RNA isolated from the left ventricles of NTG, α-TM175, and α-TM180 mice. The following RNAs were examined: ANF, α- and β-MHC, skeletal actin, TGF-β1 and -β2, tropomodulin, troponin I, OSF-2, SERCA2a, and calsequestrin (primer-specific sequences for TGF-β3 are unavailable). Results from the RT-PCR analysis show that there are dramatic increases in levels of expression of ANF, β-MHC, skeletal actin, TGF-β2, OSF-2, and calsequestrin; increased expression of these mRNAs is often associated with cardiac hypertrophy (Table 6). In addition, there were significant decreases in expression of α-MHC, tropomodulin, and SERCA2a, whereas troponin I levels were not significantly altered. These RT-PCR results are in general concordance with the microarray data and serve to verify the results.
A high degree of clinical heterogeneity has been observed between patients carrying different FHC missense mutations in α-TM (7, 9, 24, 26). The Asp175Asn and Glu180Gly mutations both result in a charge change in the amino acid, thereby potentially disrupting the coil-coiled structure of the TM dimer or its interaction with actin or troponin T (13). The results of our present study have identified and dissected the genes associated with changes in sarcomeric performance from those genes associated with the development of cardiac hypertrophy. These results show that decreased synthesis of metabolic enzymes is an early response following impairment of sarcomeric function. Altered expression of genes associated with energy metabolism is a common response in hearts undergoing hypertrophy and failure. We hypothesize that the FHC α-TM mutations indirectly cause an inefficient usage of ATP-hydrolysis to support contractile work. There appears to be a more pronounced effect with the FHC α-TM180 mutation than the α-TM175 mutation because of the more severe cardiac pathology and the increased number of altered transcripts. In addition, there is a marked inability to increase contractile performance upon acute ionotropic challenge (β-adrenergic stimulation) in the FHC α-TM180 vs. FHC α-TM175 isolated hearts (14, 18). Similar results showing altered metabolic energy profiles and reduced contractile reserve have been demonstrated in FHC cardiac MHC and cardiac troponin T mouse models (8, 23). Coupled with this altered metabolic enzyme response is the induction of extracellular matrix/cytoskeletal and inflammatory gene transcripts. Additional genes in these classes are similarly activated or repressed as hypertrophy and fibrosis dramatically increases.
This study was restricted to an examination of changes in gene expression in 2.5-mo-old FHC mice. This time point was chosen to detect initial transcript alterations during early stages of cardiac hypertrophy, before severe cardiomyopathic effects of the FHC mutations. Although beyond the scope of this investigation, it would be very informative to obtain similar comparisons in gene expression at multiple time points with the FHC α-TM175 and α-TM180 model systems. In this manner, one could assess whether temporal activation of identical gene sets is operative in the development of hypertrophy and/or whether specific gene sets are turned “on and off” as cardiac hypertrophy progresses through various pathological stages.
The role that members of the TGF-β family play in the development of cardiac hypertrophy has recently been addressed in several studies (20, 21). TGF-β1 has been implicated in mediating hypertrophic cardiomyocyte growth through angiotensin II and an inflammatory response. The TGF-β2 and -β3 members are involved in cell proliferation, repair processes, and production/maintenance of extracellular matrices. Both TGF-β2 and TGF-β3 knockout mice die perinatally and reveal important roles for these TGF-β isoforms in myocardial development (2). Interestingly, in our model systems, the expression of TGF-β genes were generally increased in the hypertrophic hearts, except TGF-β1 expression levels, which were not altered in the FHC α-TM175 mice and were minimally increased (1.1-fold) in the FHC α-TM180 hearts. However, TGF-β2 expression was increased in both model systems (1.4- and 4.1-fold in the FHC α-TM175 and 180 mice, respectively) where many extracellular matrix proteins are also upregulated. TGF-β3 expression was increased in the FHC hearts (1.3- and 1.8-fold increase in the FHC α-TM175 and 180 mice, respectively). Interestingly, increased levels of TGF-β1 and TGF-β2 correlate with increased collagen content in dilated cardiomyopathy (16). Most studies addressing the roles of TGF-βs in cardiac hypertrophy have focused on TGF-β1 and do not distinguish TGF-β1 from TGF-β2 or TGF-β3. However, this current study suggests previously unidentified roles for TGF-β2 and TGF-β3 in cardiac hypertrophy. Whether ablation of TGF-β1, 2, or 3 expression in the FHC α-TM180 mice would be able to suppress the cardiac hypertrophic phenotype is an interesting area for future investigation.
One might expect that, since the two mutations are at a distance of only five amino acids from each other in the α-TM molecule, an overlapping set of genes might be activated in response to the hypertrophic process. What is surprising is that the two transgenic models have such a diverse array of differentially expressed transcripts. From the gene matrix analysis, we find that relatively few transcripts are commonly increased or decreased in both FHC models. Wide arrays of genes alter their expression in different models of cardiac hypertrophy. For example, a direct comparison among four different mouse models of hypertrophy resulting from perturbations in protein kinase C-ε activation peptide, calsequestrin, calcineurin, and Gαq demonstrates that transcriptional alterations are highly specific to individual genetic causes of hypertrophy (1). Furthermore, comparison of the differentially regulated transcripts in the FHC α-TM models with the alterations seen in the four models described above (and additional models of hypertrophy and heart failure) (3, 11, 12, 27) fails to show specific gene sets that are commonly regulated in response to cardiac disease. The most commonly activated transcripts are those of ANF, β-MHC, and those associated with the cytoskeleton/extracellular matrix. Thus identification of gene-specific targets may be difficult and suggests that general categories of genes may prove to be better targets for therapeutic intervention.
This work was supported in part by National Heart, Lung, and Blood Institute Grant HL-71952 to D. F. Wieczorek.
We thank Jon Neumann for production of the transgenic mice and Maureen Luehrmann and Angel Whitaker for daily care of the animals. We also like to thank Bhuvaneswari Sakthivel for assistance in the bioinformatics and Dr. M. Azhar for insightful discussions.
Address for reprint requests and other correspondence: D. F. Wieczorek, Dept. of Molecular Genetics, Biochemistry, & Microbiology, Univ. of Cincinnati College of Medicine, Cincinnati, OH 45267-0524 (e-mail:)
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).
- Copyright © 2006 the American Physiological Society