Cardiac Overexpression of Myotrophin Triggers Myocardial Hypertrophy and Heart Failure in Transgenic Mice* □ S

, Cardiac hypertrophy and heart failure remain leading causes of death in the United States. Many studies have suggested that, under stress, myocardium releases factors triggering protein synthesis and stimulating myocyte growth. We identified and cloned myotrophin, a 12-kDa protein from hypertrophied human and rat hearts. Myotrophin (whose gene is localized on human chromosome 7q33) stimulates myocyte growth and par-ticipates in cellular interaction that initiates cardiac hypertrophy in vitro . In this report, we present data on the pathophysiological significance of myotrophin in vivo , showing the effects of overexpression of cardio-specific myotrophin in transgenic mice in which cardiac hypertrophy occurred by 4 weeks of age and progressed to heart failure by 9–12 months. This hypertrophy was associated with increased expression of proto-onco-genes, hypertrophy marker genes, growth factors, and cytokines, with symptoms that mimicked those of human cardiomyopathy, functionally and morphologi-cally. This model provided a unique opportunity to analyze gene clusters that are differentially up-regulated during initiation of hypertrophy versus

protein synthesis and cardiomyocyte growth. In vivo and in vitro studies have shown that many growth factors contribute to the cardiac hypertrophy process (1,2). Studies from our laboratory and others have demonstrated that factors other than high blood pressure are responsible for initiating cardiac hypertrophy in the setting of hypertension (3,4). We hypothesized that mechanical or humoral factors act on the myocardium, inducing one or more factors that trigger protein synthesis and myocardial cell growth. Using spontaneously hypertensive rats, we identified and characterized a 12-kDa protein, myotrophin, that stimulates myocyte growth (5). The myotrophin gene was mapped, for the first time, to human chromosome band 7q33 (6). We have shown that myotrophin stimulates transcription of proto-oncogenes (e.g. c-myc, c-fos, and c-jun), (7) ␤-myosin heavy chain (␤-MHC), 1 atrial natriuretic factor (ANF), and connexin (7). Large increases in myotrophin correlate with the onset of hypertrophy in spontaneously hypertensive rats and humans (8).
The purpose of our study was to evaluate whether or not increased myotrophin expression induces cardiac hypertrophy in vivo. We developed transgenic (Tg) mice overexpressing myotrophin in the heart under the transcriptional regulation of the ␣-MHC promoter. Myotrophin overexpression in these mice resulted in cardiac hypertrophy that led to heart failure. This process was associated with mechanistic changes, including significant increases in expression of proto-oncogenes, hypertrophy marker genes, and growth factor genes, as well as increased collagen deposition. These changes contributed to substantial alterations in the expression and organization of sarcomeric and structural proteins. Using these mouse models, we also documented the changes in gene expression during initiation of cardiac hypertrophy versus during its progression to heart failure. Our Tg mouse model, as a result of myotrophin overexpression, very closely mimicked the symptoms associated with the human hypertensive heart and provided a unique opportunity to study molecular changes along with changes in growth factors and cytokines during the initiation of hypertrophy and its transition to heart failure. Identification of genes whose expression is altered during the initiation and transition phases might suggest novel strategies to limit hypertrophy and its progression to heart failure.

EXPERIMENTAL PROCEDURES
Animals-Animal studies were approved by the institutional Animal Research Committee according to internal policies and guidelines of the National Institutes of Health for the humane care and use of animals in research.
Generation of Transgenic Mice-An ␣-MHC-myotrophin transgene was constructed using a recombinant myotrophin (myo) in pcDNA3myo-vector. We designed a 72-base oligomer carrying the 5Ј skeletal ␣-actin untranslated region and used it to generate a chimeric myo-cDNA, placed adjacent to the cytomegalovirus promoter of pcDNA3. The ␣-MHC promoter-containing vector was provided by Dr. Jeffrey Robbins (University of Cincinnati). This vector has an ␣-MHC promoter and a heterologous 3Ј-untranslated region containing the human growth hormone poly(A) site. The ␣-MHC-myo-recombinant vector was digested using the NotI restriction enzyme. The resulting linear DNA fragment containing the ␣-MHC promoter-myo coding sequence and human growth hormone poly(A) was isolated and purified using Qiagen gel extraction kit (Qiagen, Valencia, CA). Pronuclear injection was performed at the University of Cincinnati transgenic animal facility, using standard techniques.
Determination of the Ratio of Heart Weight to Body Weight-Mice were euthanized by CO 2 and autopsied immediately for gross signs of heart failure, inflammatory lesions, or congenital defects. The hearts were then removed, washed with 1ϫ phosphate-buffered saline, blotted dry, and weighed in a Mettler precision balance (Mettler-Toledo, Inc., Columbus, OH). Hypertrophy was measured using the heart weight (HW)/body weight (BW) in mg/g (3).
Hybridization Analysis-Four founder mice were identified by Southern blot analysis of genomic DNA (10 g) prepared from tail biopsies. Northern blots were performed with ϳ20 g of total cardiac RNA. The labeled myotrophin cDNA probe was generated using random primers, as described previously (9).
Quantitation of Myotrophin Protein-Western blots were performed using 10 g of total protein from 9-month-old wild-type (WT) and transgenic (Tg) mice hearts via standard techniques using polyclonal anti-myotrophin antibody and was normalized using glyceraldehyde-3phosphate dehydrogenase antibody (Novus Biologicals Inc., Littleton, CO). Myotrophin from the hearts of Tg and WT mice was quantitated as described previously (8).
RNase Protection Assay-Total RNA from WT and Tg mice hearts (4 weeks and 9 months old; n ϭ 5) was extracted as described previously (10), and 15 g was used in an RNase protection assay using templates specific for growth factors and cytokines according to the manufacturer's protocol (RiboQuant; BD PharMingen; MCK-3B template set). After RNase digestion, protected fragments were resolved on 6% denaturing polyacrylamide gels and quantified using a PhosphorImager. We normalized the value of each hybridized signal to that of an internal control, glyceraldehyde-3-phosphate dehydrogenase.
Determination of Cardiac Function-Two-dimensional echocardiography was performed at the Image Analysis Core of the Cleveland Clinic's Department of Cardiovascular Medicine. Lightly sedated Tg and WT mice (4-and 36-week-old), using 0.2 ml of Avertin (99% tertamyl alcohol and 99% 2,2,2-tribromoethanol; Aldrich), were evaluated using the M-mode views on a transthoracic study, measuring left ventricular (LV) systolic and diastolic dimensions, interventricular septum and LV posterior wall thickness, and left atrial chamber diameter. End-diastolic and end-systolic frames were defined as those showing the largest and smallest areas, respectively. Data were correlated with timing of the QRS complex. Digitized images were obtained using an ultrasound system (Hewlett-Packard Sonos 5500; Palo Alto, CA).
Histology-All hearts were fixed, paraffin-embedded, and cut into 4-m sections. Sections (taken from same areas of the heart of WT and Tg mice) were stained with hematoxylin/eosin for structural analysis using standard techniques, and myocyte dimensions were quantitated by image scanning using the Image Pro Plus software program.
Isolation of Cardiac Myocytes from Hearts of Adult FVB and Tg Mice Overexpressing Myotrophin (11)-Hearts were taken out from heparininjected mice and were cannulated via aorta (n ϭ 6). Hearts were perfused with perfusion buffer (glucose (1 g), NaHCO 3 (0.58 g), and pyruvic acid (0.27 g), pH 7.3) with 95% O 2 and 5% CO 2 on a Langendorff apparatus. After perfusing the heart for 10 min in EGTAsupplemented perfusion buffer, hearts were digested using collagenase (2 mg/ml) for 28 min, with gradual enhancement of CaCl 2 . After 28 min of digestion with collagenase, the heart was taken out and incubated in a diluted collagenase solution for 10 min in a shaking water bath at 37°C. The ventricles were separated from the atria, triturated for 30 s, and subsequently filtered through cheesecloth. The filtrate was centrifuged at 400 rpm for 2 min, the supernatant was removed, and the pellet was resuspended in 4% bovine serum albumin solution and observed under a phase-contrast microscope. Preparations with 80 -85% beating rod-shaped cells were used for experimental purposes.
Immunocytochemical Staining Using myo Antibody (12)-The cells were plated on slides coated with laminin and were fixed in 4% paraformaldehyde dissolved in 1ϫ TBS for 10 min. Cells were then permeabilized with 0.1% Triton X-100 in 1ϫ TBS for 5 min. The cells were blocked with heat-inactivated 4% horse serum and 0.1% sodium azide for 2 h. Cells were incubated overnight with primary antibody (8) (polyclonal anti-myotrophin) at 4°C, washed extensively with 0.1% Triton X-100 in 1ϫ TBS, and incubated with fluorescein isothiocyanateconjugated goat anti-rabbit IgG (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) for 2 h at room temperature in the dark. The slides were extensively washed with 0.1% Triton X-100 in 1ϫ TBS and once with 1ϫ TBS. The cells were mounted on coverslips with 4Ј,6diamidino-2-phenylindole and were observed under a fluorescent microscope.
Statistical Analysis-Results were expressed as mean Ϯ S.E. Data was analyzed by two-way analysis of variance, and differences between groups were determined by a least-square means test (SUPERNOVA). A value of p Ͻ 0.05 was considered significant.
Gene Array Studies-Gene array analyses utilized RNA from heart tissue isolated from WT and Tg mice aged 4 weeks (initiation of hypertrophy) or 36 weeks (end-stage transition from hypertrophy to heart failure). Five WT and five Tg animals from each age group were examined. Total cellular RNA (10 g) was reverse transcribed, and doublestranded cDNA (1 g) was transcribed into cRNA (Enzo Bioarray RNA transcript labeling kit; Affymetrix, Inc., Santa Clara, CA). Biotin-11-CTP and biotin-16-UTP were incorporated into cRNA during synthesis. Approximately 20 g of biotinylated cRNA was fragmented in buffer at 94°C for 35 min and added to murine genome U74Av2 arrays, containing ϳ12,683 gene probe sets (Affymetrix), and hybridized at 40°C for 16 h. Arrays were washed, stained with streptavidinphycoerythrin, and read using a confocal microscope scanner with a 560-nm filter.
Data Analysis-Data were analyzed using the Microarray Suite and Data Mining Tools software packages (Affymetrix). Genes identified as up-regulated were consistently increased in each pairwise comparison between the animals of interest (such as 9-month-old Tg mice) and all other animals of both age groups (p values varying from 0 to 0.0025). Up-regulated genes were screened to eliminate those with "marginal" or "absent" absolute calls in the induced samples. Genes marked as downregulated were consistently decreased among all pairwise comparisons (p values ranging from 0.90 to 1). -Fold change was calculated by converting average signal log ratio values (change in the expression level of a transcript between the control and experimental samples) for each probe set to a whole number. This was accomplished by raising 2 to the power of the signal log ratio value (2 (signal log ratio) ).
Gene clusters with similar expression patterns were identified using the self-organizing map (SOM)-clustering algorithm (10) of the Data Mining Tool software package (version 3.0; Affymetrix). Signal values for all genes were imported from the Microarray Suite (version 5.0) into the Data Mining Tool after publication using MicroDB software (Affymetrix) (13). SOM clustering was performed using default filtering values and parameters set to identify up to 81 possible clusters. After clustering, results again were screened to eliminate genes that failed to attain a "present" call in any of the pertinent samples.
Detailed tables for the changes in gene expression between failing and nonfailing hearts are provided as Supplemental Material (Supplemental Table I).

Generation of Tg Mice
Four founders were identified from the live births resulting from pronuclear injection (42 mice) with the myotrophin transgene. All four lines were expanded by crossing with non-Tg mice. The transmission rate of ␣-MHC myotrophin transgenic mice showed a 45% transmission frequency, determined by Southern analysis, typical of a Mendelian inheritance (line 1, 81/156 (52%); line 2, 7/32 (22%); line 3, 20/51 (39%); line 4, 20/51 (39%)). The ventricles and atria were enlarged significantly in all four lines of Tg mice by 4 weeks of age compared with the age-matched WT mice. The heart weight/body weight (HW/BW) ratio also increased significantly in all Tg mice during the progression to hypertrophy (Fig. 1, A-B). All four lines of mice displayed myotrophin overexpression and developed significant hypertrophy, which eventually led to heart failure.
At ϳ36 weeks of age, the Tg mice, which overexpress cardiospecific myotrophin, developed symptoms of heart failure, including lethargy, edema, pulmonary effusion, and lack of alertness. The kidneys of the Tg mice did not differ from those of WT mice (Table I).

Myotrophin Overexpression Was Associated with Histologic Lesions in Heart Tissues of Tg Mice
The LV heart walls of Tg mice (18 weeks old) were severely thickened compared with those of WT mice ( Fig. 1C (I-III)) and showed concentric hypertrophy ( Fig. 1C (I)). Both right and left ventricles of the Tg mice were enlarged and displayed increased septal thickness compared with those of the WT mice. Histology examination showed typical large nuclei in the Tg group, confirming myocyte hypertrophy ( Fig. 1C (II)). Multiple foci of classic myocyte disarray were observed in the Tg mouse tissue, a change not present in the WT hearts ( Fig. 1C (II)). Fibrotic foci accompanied by dystrophic calcification were also observed in the Tg mice but were absent in WT hearts ( Fig. 1C (III)). Small foci of apparent myocyte slippage surrounded the coronary vessels.

Myotrophin Is Overexpressed in Heart Tissue of Tg Mice
Myotrophin mRNA ( Fig. 2A (1)) and protein ( Fig. 2B (1 and 2)) were increased in the myocardium of all four Tg lines through four generations compared with WT mice. When regression analysis was done between mRNA expression and HW/BW in WT and Tg mice (age varying from 18 to 24 weeks) from all four lines, a linear correlation between myotrophin gene expression and HW/BW was observed ( Fig. 2A (2); y ϭ 14.916x ϩ 8.9615 and r 2 ϭ 0.9691 for Tg; y ϭ 15.209x Ϫ 15.744 and r 2 ϭ 0.9227 for WT). Fluorescein isothiocyanate-tagged myotrophin was abundant and distinctly visible in the myocytes from 24-week-old Tg mice from all lines, compared with age-matched WT ( Fig. 2B (3)). However, myotrophin mRNA expression in the kidneys, livers, and lungs of Tg mice did not differ from that in WT mice (data not shown). As a consequence of myotrophin gene overexpression, expression of both hypertrophy marker genes (ANF and ␤-MHC) and proto-oncogenes (c-fos, c-jun, and c-myc) were also up-regulated in all four generations of the Tg mice lines (Fig. 2, C and D).

Myotrophin Overexpression Leads to Myocyte Hypertrophy in Tg Mice
To document myotrophin overexpression-induced changes in myocytes, we quantitated myocyte dimension by hematoxylin/ eosin staining of heart tissue and image scanning using the Image Pro Plus software program. The cross-sectional area of myocytes in both the 4-week-old and 9-month-old mice was significantly increased (Fig. 3A). The myocyte cross-section increased from 350 to 781 per m 2 (p Ͻ 0.01) in the 4-week-old transgenic mice. The increase was even larger in the 9-monthold mice (534 -2164) per m 2 (p Ͻ 0.001).
We also quantitated myocyte dimension by isolating myocytes from WT and Tg mice from 9-month-old mice heart. The cross-sectional area of myocytes in 9-month-old mice was significantly increased (Fig. 3B) from 2431 Ϯ 712 per m 2 in WT to 6297 Ϯ 1280 per m 2 in Tg mice (p Ͻ 0.001). All myocytes were hypertrophied, and no atrophy was observed. These data provide evidence that cardiac hypertrophy was present in the transgenic animals in all four lines and four generations as early as 4 weeks of age and that this hypertrophy worsened in the older Tg animals (Fig. 3). The cross-sectional areas were quantified in 30 myocytes from each mouse (WT n ϭ 5; Tg n ϭ 8, representing all four lines).

Cytokine and Growth Factor Gene Up-regulation Is Associated with Disease Stage in Tg Mice
We examined the relative expression of growth factors and cytokines in the Tg mice representing all four lines, using RPA ( Fig. 4, n ϭ 5), compared with age-matched WT mice. We studied two age groups of animals: 4-week-old mice, which represented the onset of hypertrophy, and 36-week-old mice, which represented the chronic phase of hypertrophy, during its transition to heart failure. A novel finding was the age-associated changes in expression of different cytokines. As shown in Fig. 4A, at 4 weeks of age, some of the cytokine transcripts were induced in Tg hearts, compared with age-matched WT. Expression of LT-␤, TGF-␤ 2 , and TGF-␤ 3 were significantly up-regulated in 4-week-old Tg mice compared with age-matched WT (p Ͻ 0.05). In the 36-week-old Tg mice, interleukin-6, macrophage migration inhibitory factor, tumor necrosis factor-␣, interferon-␥, and different isoforms of the transforming growth factor-␤ family (TGF-␤ 1 , -␤ 2 , and -␤ 3 ) were significantly elevated, compared with the age-matched WT mice (p Ͻ 0.01). However, the percentage increase in cytokine transcripts was comparatively higher in 36-week-old Tg than in 4-week-old Tg mice. Interestingly, expression levels of interleukin-6, tumor necrosis factor-␣, interferon-␥, TGF-␤ 2 , and macrophage migration inhibitory factor did not change in the young Tg animals during initiation of hypertrophy compared with the agematched WT mice (Fig. 4B). These data suggest that the cytokine-/growth factor-mediated hypertrophic process is different in young and old Tg mice, especially during transition to heart failure.  fractional shortening (FS), was not changed in the 4-week-old Tg mice, compared with WT (FS ϭ 50 Ϯ 0.01% in WT versus 48 Ϯ 0.04% in the Tg group (p ϭ not significant)).
Echocardiographic data from 9-month-old Tg mice revealed statistically significant changes compared with the age- mm, p Ͻ 0.01) and increased left ventricular posterior wall thickness (0.090 Ϯ 0.007 versus 0.075 Ϯ 0.016 mm, p Ͻ 0.05) in the Tg mice. Furthermore, we found a large amount of pleural effusion in the Tg mice, which suggested that hypertrophy had already advanced to heart failure.
These data suggest that in the young Tg mice, cardiac function was not compromised, despite the presence of hypertrophy, whereas in the 9-month-old Tg mice, cardiac function was significantly compromised.

DNA Microarray Results
Changes in Gene Expression at the Initiation of Hypertrophy-To identify candidate genes that mediate physiological responses to myotrophin overexpression, oligonucleotide gene array analyses were performed on heart samples from Tg and age-matched WT controls. Cardiac RNAs from five transgenic and five WT animals at each age (4 weeks and 9 months) were used in the gene profiling studies. To identify genes with expression patterns that correlated with initiation of hypertrophy (4 weeks) or transition to heart failure (9 months), two strategies were used. Pairwise comparisons between the experimental animals of interest and all other animal samples were used to identify genes with consistent up-or down-regulation at a particular developmental time. In addition, SOM clustering was used to identify gene clusters with similar expression patterns that might reflect similar modes of regulation within the pertinent samples. Genes up-regulated by more than 1.8fold are included in Tables II-IV. A detailed list of up-regulated and down-regulated genes is included as Supplemental Material (Supplemental Table I).
Eighty genes were consistently up-regulated in all pairwise combinations when the 4-week-old Tg mice were compared with all other mice (9-month-old Tg and WT and 4-week-old WT) (Table II). When just one of the pairwise combinations was varied with the others, 179 genes were induced. Of those, 39 genes were clustered in three major functional categories: extracellular matrix and cytoskeleton, cell signaling, and growth factors/transcriptional regulators. Eleven of 30 up-regulated expressed sequence tags (ESTs) had some assigned function. Among these, sarcolemmal protein SLAP, actin cross-linking protein 7, talin, glycogenin 1, and Cdc 5-like protein were elevated during the initiation of cardiac hypertrophy.
When 4-week-old Tg animals were compared with their agematched WT animals only, a slightly different picture emerged (Table III). Seventy-four genes were up-regulated in all pairwise comparisons between 4-week-old Tg versus WT. These genes were clustered into six functional categories: extracellular matrix, myofibrillar and cytoskeletal protein, cellular signaling factors, growth and transcription factors, cell defense, and protein expression regulators. Forty-five known genes were down-regulated in 4-week-old Tg animals compared with the WT animals. Down-regulated genes were clustered primarily as cell signaling or mitochondrial proteins (Table III).
Changes in Gene Expression during Transition to Heart Failure-Pairwise comparisons were also used to identify genes that decreased in expression when comparing 9-month-old Tg hearts with all other samples (Table II). One hundred thirtythree genes were consistently elevated in failing hearts compared with nonfailing WT or younger Tg hearts. Fifty-one of these genes were functionally clustered into six different categories: cell signaling; growth and transcription factors; extracellular matrix and cytoskeletal protein, cell defense; apoptosis; and protein expression regulators and metabolic enzymes. Of 82 ESTs, only 11 had unknown functions.
Approximately 50 genes were down-regulated in Tg hearts compared with all other samples. Most of these genes were FIG. 4. A, a typical autoradiogram of RNase protection assays measuring cytokine expression in hearts from Tg and WT mice. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a loading control. An RNase protection assay was performed in five different Tg and WT mice, representing all four lines. This figure represents a typical finding from five independent experiments. B, estimation of the expression pattern of the cytokine transcript level normalized with glyceraldehyde-3-phosphate dehydrogenase in 4-week-old and 9-month-old WT and Tg mice hearts, representing all four lines. identified in the aforementioned pairwise comparisons and represented several functional groups: extracellular and cytoskeleton proteins, cellular signaling, or mitochondrial enzymes. The 21 down-regulated ESTs included AP-4-related protein, histone H1.2, mitochondrial ribosomal protein MRPs15, electron transfer flavoprotein, and transcription elongation factor A (SII).
Pairwise comparison of 9-month-old Tg and age-matched WT heart samples yielded 197 genes that were significantly and consistently up-regulated, 84 of which were clustered in eight functional categories: cell signaling, extracellular matrix, and cytoskeleton or protein expression. Apart from these groups, some genes were categorized as transcriptional and growth factors, cell defense proteins, apoptosis-related proteins, or proteins involved in cell division. Twenty-eight genes were identified as known ESTs among 113 ESTs that showed increased expression in failing mouse hearts. In addition to the up-regulated genes, 206 genes were consistently down-regulated across the pairwise comparisons. Forty-seven were clustered into functional groups, including cell signaling, matrix and cytoskeleton, or mitochondrial enzymes. Of 159 down-regulated ESTs, 16 had known functions (Table III). Fig. 6 summarizes SOM clustering analysis of the maximally changed genes in 9-month-old Tg versus all as well as 4-weekold Tg versus all. When SOM clustering was performed using absolute gene expression values from all samples from 9-month-old Tg animals compared with either age-matched WT or 4-week-old WT or Tg mice, definitive clusters of candidate genes up-or down-regulated during the transition from hypertrophy to heart failure emerged (Fig. 6, a and b). Those maximally up-regulated include fibronectin, VCAM1, slow myosin heavy chain, matrix metalloproteinase 3, ceruloplasmin, apolipoprotein D, and MRP8. Approximately 80 genes were expressed at a higher level in three 4-week-old Tg animals when compared with all other animals (Fig. 6c). Included within this group were skeletal muscle actin, MLC3F, calsequestrin, immediate early genes, SLAP, glycogenin, skeletal muscle tropomyosin, talin, and disintegrin. All genes from this cluster were identified in the pairwise comparisons noted above. Interestingly, SOM analysis did not identify clusters of genes consistently down-regulated at this early developmental time point (data not shown).

Comparison of Gene Expression between Heart Failure and Initiation Stage: Old Tg (9 Months Old) Versus Young Tg (4 Weeks Old)-
The genes expressed more highly in 9-month-old Tg animals included a subset of genes that specifically increased in expression between 4 weeks and 9 months, as the Tg animals progressed from hypertrophy to heart failure. Pairwise comparison of 9-month-old Tg and 4-week-old Tg samples identified 276 genes that were specifically up-regulated in Tg comparisons but not in WT animals (thereby excluding age-regulated genes) (Table IV). Of these up-regulated genes, 44 genes were classified as extracellular matrix and cytoskeleton, growth and transcription factors, cell signaling factors, cell defense, apoptotic, protein expression regulators, or mitochondrial proteins. Eleven of the ESTs with known functions included calcium-binding protein A15, casein kinase I, insulinlike growth factor 3, and Rab-6 (a ras oncogene family protein).
Similar pairwise comparisons identified 30 cardiac genes down-regulated in 9-month-old Tg animals compared with 4-week-old Tg animals. These genes clustered into several functional groups: extracellular matrix and cytoskeleton proteins, mitochondrial enzymes, cell signaling factors, or cell cycle regulators. Fifteen known ESTs in this group included cyclophilin D, tropomyosin 5, exportin 1, and Ras-related protein RAL1. DISCUSSION This study reinforces the proofs we have previously presented that myotrophin is a significant causal factor in the hypertrophy/heart failure continuum. Data presented here document the effects of myotrophin protein overexpression at molecular, cellular, morphological, and functional levels in a specially developed line of Tg mice. These data indicate that myotrophin overexpression initiates cardiac hypertrophy, eventually progressing to heart failure, a process associated with changes in expression of proto-oncogenes, ANF, ␤-MHC, and cytokines. Importantly, using this model and the new tools of state-of-the-art DNA microarray analysis ( Fig. 6 and Tables II and III), we have elucidated patterns of gene up-regulation and down-regulation that may be involved during initiation of cardiac hypertrophy and progression to heart failure in humans.  Table I. We documented several novel mechanistic changes that occur during the transition from hypertrophy to heart failure. We confirmed that myotrophin overexpression resulted from increased myotrophin mRNA and protein levels in all lines and generations of Tg mice (Fig. 2). Importantly, the increase in myotrophin triggered a significant increase in cytokines and growth factors such as LT-␤, tumor necrosis factor-␣, interfer-on-␥, interleukin-6, TGF-␤ 1 , TGF-␤ 2 , TGF-␤ 3 , and macrophage migration inhibitory factor, in 9-month-old Tg mice, when chronic hypertrophy advanced to heart failure (Fig. 4), whereas, at 4 weeks of age, three genes (LT-␤, TGF-␤ 2 , and TGF-␤ 3 ) were up-regulated in Tg mice, compared with the age-matched WT. Echocardiographic data showed significant hypertrophy of left ventricle and septum (asymmetric hyper-  trophy) in 36-week-old Tg mice hearts, a typical change observed in human hypertrophy. Mice afflicted with hypertrophy also had severely compromised cardiac function associated with pleural effusion, a common occurrence during human heart failure. However, this compromised function did not occur in the hypertrophied hearts of young 4-week-old Tg mice despite presence of hypertrophy. Our data also suggest that atrial enlargement arises from mitral and tricuspid valve regurgitation, which occurs because the ventricular cavity enlarges, causing an incomplete sealing in these valves. This cluster of symptoms mimics human cardiomyopathic hypertrophy with end-stage heart failure. Although other Tg models have been reported (14,15), none have studied the progression of hypertrophy that advances to heart failure in the manner we have described. Previously, using isolated myocytes, we have shown that the mode of action of myotrophin protein is mediated through protein kinase C and NF-B signaling pathways (16). This in vivo model, overexpressing myotrophin, provided us with the opportunity to dissect out the role of myotrophininduced signaling pathways for the initiation process of cardiac hypertrophy and its progression to heart failure. Work is in progress to determine protein kinase C and NF-B cascade in Tg hearts at 4 weeks, 16 weeks, and 9 months compared with their age-matched WT. This model provided the opportunity to further the genomewide screening of cardiac tissue as a tool to identify new genes altered during initiation, progression, and transition from hypertrophy to heart failure. This study is unique because this type of experiment is not possible in humans. The alteration of several hypertrophy-associated genes reported in recent gene array studies in human failing heart (17) were found to be similar to the murine heart failure model overexpressing myotrophin, reported in this study. For the first time, we documented the alterations of gene clusters that participate during the initiation of hypertrophy and during the transition from hypertrophy to heart failure (Table II). Our data also suggest that the initiation of hypertrophy utilizes a transcriptional program involving specific sets of genes, which are distinct from those that operate during the transition phase. Characterization of the expression of these novel genes during initia-tion and transition phases could provide new insights into cardiac remodeling. In failing hearts, natriuretic peptide precursors type A and B were ranked as the top candidate genes when compared with either age-matched WT or with the total of all young WT and Tg samples. SOM clustering analysis identified the maximally changed gene expressions during both initiation and progression of hypertrophic process. The functions of many of them are yet to be defined during hypertrophy to heart failure. This gene cluster continues to be expressed during the progressive deterioration of cardiac function (18). In addition, significantly increased expression of extracellular matrix proteins, like collagen type I and type VIII, fibronectin, C-type lectin, and matrix metalloproteinases, was observed in failing hearts (19). Growth factors like TGF-␤ 2 , TGF-␤ 3 , tumor necrosis factor-␣, insulinlike growth factor, and hypoxia-inducing factor 1␣ can be important during the transition from hypertrophy to heart failure. Genes involved in fatty acid metabolism (e.g. apolipoprotein-D and -E) and glucose metabolisms (glucose-6-phosphate dehydrogenase) were up-regulated, whereas lactate dehydrogenase and fatty acid-binding proteins were downregulated. Several mitochondrial enzymes were consistently and significantly down-regulated in failing hearts compared with either WT or 4-week-old Tg hearts, a finding that may explain the reduced cardiac energy production during heart failure. Induction of apoptotic proteins such as CIDE-A, Bcl2binding protein NIP3, and caspase-9 in failing hearts signifies active programmed cell death. Metallothionein 1 and 2 are stress-inducible, metal-binding proteins whose antioxidant function and regulation of apoptosis in the heart were reported previously (20). Unlike Tan et al. (17), we found increased expression of metallothionein proteins as well as such other cell cycle regulators as GAS 6, cyclin 1, and histone H1. This seeming discrepancy indicates that cell death and cell regeneration can occur simultaneously in the failing heart. No such genes were found when 36-week-old Tg mice were compared with 4-week-old Tg mice, although NIP3 and caspase-9 were up-regulated in old Tg animals. Apoptosis was not evident during the initiation phase of hypertrophy, yet induction of cyclin and cdk genes started as early as 4 weeks of age in Tg mice hearts. 2 Our data showed that initiation of hypertrophy was associated with induction of fewer genes. Compared with genes from failing hearts, interestingly, no genes were down-regulated during the initiation of hypertrophy (4-week-old Tg mice). Among the cytoskeletal proteins, myosin alkali light chain, skeletal muscle calsequestrin, ␤-tropomyosin, and transcriptional activators like mitogen-activated protein kinase/extracellular signal-regulated kinase kinase kinase 3, STAT6, Pod1, and DP1 were up-regulated in the hearts of 4-week-old Tg mice. Combining all of these findings, the data are expected to provide information for changes in cardiac metabolism during onset of hypertrophy and its transition to heart failure.
In conclusion, this new Tg mouse model of hypertrophy resulting from myotrophin overexpression, leading to heart failure, is important because of its similarity to end-stage heart failure in human beings in both molecular (gene expression of hypertrophy marker genes, proto-oncogenes, cytokines, and growth factors) and physiological parameters (pleural effusion and lethargy). Heart failure occurred in all of our Tg mice sacrificed to date (n ϭ 150), and the average life span was 8 -11 months in the F 1 , F 2 , F 3 , and F 4 generations. Several growth factors and cytokines were increased during the initiation FIG. 6. Gene expression patterns in WT and Tg animals. SOM cluster analysis was used to identify gene clusters that exhibited similar expression patterns. Affymetrix signal values from genes representing three of the identified expression patterns were normalized to sample 1 (a 9-month-old WT animal), and the relative expression values were plotted for each animal. The sample ID numbers represent the 9-month-old WT animals (1 and 2), 9-month-old Tg mice (3-5), 4-weekold WT mice (6 and 7), and 4-week-old Tg mice (8 -10). An abbreviated gene name is provided for each gene assessed. phase and during the transition of hypertrophy to heart failure. Comparison of the gene array data between the initiation of hypertrophy and its transition to heart failure involves differential activation of functional gene clusters. Up-regulation of growth factors, calcium-binding proteins, proteins regulating programmed cell death, and extracellular/cytoskeletal proteins as well as down-regulation of mitochondrial proteins and cytoskeletal/myofibrillar proteins mark the transition phase, which is associated with severely compromised heart function, thereby differing from the mechanisms of the initiation process of hypertrophy as well as those operating in nonfailing WT hearts.
The intricate, multifaceted process of heart failure, especially its transition from longstanding hypertrophy to heart failure, involves many factors. Eventual heart failure is probably the result of cross-talk between neurohumoral mechanisms and growth factors. Our new genetic data, added to our prior findings based on molecular and biochemical data, convincingly demonstrate that myotrophin is a factor that not only initiates hypertrophy but is also associated with the progression to heart failure. We expect that this new mouse model will provide the key to elucidate further molecular mechanisms that occur during advancement of hypertrophy to heart failure and will facilitate the design of effective therapies.