Molecular and physiological characterization of RV remodeling in a murine model of pulmonary stenosis

Abstract

Right ventricular (RV) dysfunction is a common long-term complication in patients after the repair of congenital heart disease. Previous investigators have examined the cellular and molecular mechanisms of left ventricular (LV) remodeling, but little is known about the stressed RV. Our purpose was to provide a detailed physiological characterization of a model of RV hypertrophy and failure, including RV-LV interaction, and to compare gene alterations between afterloaded RV versus LV. Pulmonary artery constriction was performed in 86 mice. Mice with mild and moderate pulmonary stenosis (PS) developed stable hypertrophy without decompensation. Mice with severe PS developed edema, decreased RV function, and high mortality. Tissue Doppler imaging demonstrated septal dyssynchrony and deleterious RV-LV interaction in the severe PS group. Microarray analysis showed 196 genes with increased expression and 1,114 with decreased expression. Several transcripts were differentially increased in the afterloaded RV but not in the afterloaded LV, including clusterin, neuroblastoma suppression of tumorigenicity 1, Dkk3, Sfrp2, formin binding protein, annexin A7, and lysyl oxidase. We have characterized a murine model of RV hypertrophy and failure, providing a platform for studying the physiological and molecular events of RV remodeling. Although the molecular responses of the RV and LV to afterload stress are mostly concordant, there are several key differences, which may represent targets for RV failure-specific therapy.

the right ventricle is uniquely at risk in patients with complex congenital heart disease involving right-sided obstructive lesions (e.g., tetralogy of Fallot, tetralogy/pulmonary atresia) and in patients with systemic right ventricles (1, 40, 46). Increased stress on the right ventricle in the form of increased hemodynamic loading (pressure and/or volume) may result in abnormalities in cardiac structure, function, metabolism, coronary perfusion, neurohormonal activation, and molecular signaling. These stresses lead to both adaptive and maladaptive structural and molecular remodeling of the right ventricle and deleterious effects on the left ventricle through ventricular-ventricular interaction and may limit long-term survival (3, 7, 29, 42). For many of these patients, detrimental conditions for the right ventricle exist throughout life, even after successful repair or palliation. As surgical techniques for the primary repair of complex forms of congenital heart disease improve, long-term survival and quality of life will depend on our ability to preserve long-term right ventricular (RV) function. Decisions regarding the ideal time for surgical reintervention will need to be based on more quantitative rather than qualitative measures.

Developing a better understanding of the molecular mechanisms of RV remodeling will assist in developing new therapeutic modalities to diagnose, prevent, and treat RV dysfunction. Although there are considerable data on the molecular events underlying afterload-induced left ventricular (LV) remodeling, there is little information for the right ventricle and some evidence to suggest that the stress responses of the right and left ventricles may be different. In the past, it was believed that differences in global structure and loading conditions represented the main differences between the right and left ventricles. However, recent work has shown an increasing divergence in the anterior and primary heart field pathways leading to the differentiation of RV versus LV cardiomyocytes during early development (46) and chamber-specific differences in cell signaling and calcium handling, suggesting fundamental differences between the right and left ventricles at the cellular level as well (10, 32, 34, 54, 72). Recent studies suggest that standard pharmacotherapies used to treat LV dysfunction may not be as effective in patients with dysfunction of a systemic right ventricle (18). The development and characterization of a murine model of the afterloaded right ventricle will add significantly to our knowledge of chamber-specific molecular responses to stress and provide the basis for the further dissection of molecular pathways using the wide range of transgenic and gene knockouts for specific signaling components.

We utilized a murine model of RV pressure overload hypertrophy and RV failure using pulmonary artery (PA) constriction (PAC). The purpose of the current study was to, first, provide a detailed physiological characterization of this model, including a documentation of the effects of RV-LV interaction using tissue-Doppler imaging (TDI). TDI has emerged as a leading technique for the assessment of ventricular dyssynchrony using data obtained from myocardial motion (17, 57). TDI-derived strain has been shown to be an excellent index of regional myocardial function because it is less influenced by overall cardiac motion (24, 25, 61). Although the usefulness of TDI has been shown in the assessment of LV function in mice (58), data on RV function are lacking. We also sought to provide the first evaluation of genome-wide transcriptional alterations associated with RV pressure loading as well as a preliminary comparison of gene changes in the afterloaded RV versus LV.

METHODS

Model of PAC.

All mice were male FVB, aged 8 to 10 wk. Anesthesia was induced with pentobarbital sodium (50 mg/kg ip). Mice were then intubated transtracheally with a 20-gauge angiocath and ventilated artificially using a Harvard rodent ventilator (Harvard Apparatus, Holliston, MA) at a rate of 120–140 breaths/min with a tidal volume of 10 μl/g body wt. Maintenance anesthesia was provided with 1% to 2% isoflurane. A right lateral thoracotomy was then performed. The main pulmonary trunk was identified under the left atrial appendage and banded with a 7-0 suture, tied tight against a 27-gauge needle, which was then removed. The chest was then closed with 7-0 sutures around adjacent ribs, and the skin was closed with 5-0 suture. Air was evacuated to avoid the need for a chest tube postoperatively. Our laboratory has previously shown the importance of using true sham controls in studies of gene expression in the afterloaded LV (73), so all age- and strain-matched controls in this study were subjected to the same operative procedure, including the dissection of the main PA, with the sole exception of the placement of the band.

Echocardiography and TDI.

Echocardiographic images were acquired with a GE Vivid 7 ultrasound system (GE Healthcare, Milwaukee, WI) equipped with both 13- and 10-MHz transducers. Mice subjected to pulmonary banding and sham-operated controls underwent echocardiography at 7 days postoperative to evaluate the status of the surgical intervention and obtain hemodynamic data. Doppler signals analyzed included maximum pulse wave Doppler and velocity-time integral (VTI), with angle correction, in the RV outflow tract (RVOT) to estimate the peak pressure gradient (PPG) and mean pressure gradient (MPG) between the RV and PA. Cardiac output was calculated as VTI × RVOT area × heart rate (HR) (70).

Mice were divided into three groups based on echocardiographic findings. Mice with a PPG of <20 mmHg were excluded from further study (n = 2). Mice with a PPG of 20–35 mmHg and without RV enlargement (Fig. 1A) were included in the mild pulmonary stenosis (PS) group. Mice with PPG between 35 and 60 mmHg but without evidence of flattening of the interventricular septum and without evidence of heart failure were included in the moderate PS group. Mice with an RV-PA PPG between 35 and 60 mmHg and with evidence of RV enlargement and either a flat interventricular septum or concave septal shift encroaching into the left ventricle (Fig. 1, B and C) were included in the severe PS group. As will be seen below, these mice also had evidence of right-sided heart failure [increased RV end-diastolic pressure, tricuspid regurgitation (Fig. 1D), and peripheral edema]. Peripheral edema was assessed both visually and by increases in body weight (BW) and liver weight.

Fig. 1.

A: systolic 2-dimensional image from the parasternal short-axis view at the level of the left ventricular (LV) papillary muscles (PM) in a mouse with mild pulmonary stenosis (PS). Mice with moderate PS showed similar 2-dimensional echo characteristics. B: comparable systolic 2-dimensional image in a mouse with severe PS showing a markedly enlarged right ventricular (RV) chamber and flattening of the interventricular septal (IVS). C: comparable systolic image in a mouse with severe PS showing bowing of the IVS into the LV cavity. D: color Doppler echocardiogram from parasternal short-axis view at the level of aortic valve in a mouse with severe PS showing significant tricuspid regurgitation (TR).

LV fractional shortening (FS), LV diastolic dimension, and HR were calculated from M-mode echocardiograms in the parasternal short-axis view at the level of papillary muscles. RVOT dimension and RVOT FS were measured from the RVOT view at the level of the aortic valve. TDI was performed to assess RV function and RV-LV interaction. TDI images were collected in the apical four-chamber view at frame rates of 477 frames/s and at depths of 1 cm. The region of interest was placed at the tricuspid valve annulus to evaluate RV free wall maximum velocity, and analysis was performed offline with the use of commercially available Echo Pac software (GE Healthcare).

ECG and invasive hemodynamic parameters.

ECGs were recorded to assess intra-RV conduction using Chart for Windows v4.1 (AD Instruments, Colorado Springs, CO). For invasive hemodynamic evaluation, a 1.4 F transducer-tipped micromanometer catheter (Millar Instruments, Houston, TX) was inserted via the right jugular vein to determine RV pressure and change in pressure over time (dP/dt) as described previously (56, 68). Steady-state data were measured over an average of 10 beats.

Organ weight and histopathology.

Whereas all echocardiographic measurements were performed at 7 days, all morphometric and gene expression analysis studies were performed at 10 days after PAC. This 3-day break was introduced to minimize gene changes secondary to the stress of anesthesia at the time echocardiography (73). At 10 days after PAC, the heart and liver were excised, and their weights were determined. The heart was separated into right and left ventricles and the septum. From the mid-RV and mid-LV, transverse 5-μm sections were cut and stained with wheat germ agglutinin and 4-6-diamidino-2-phenylindole to determine the myocyte cross-sectional area. The cardiomyocyte cross-sectional area was measured with the use of a Leica imaging system (Leica Microsystems, Exton, PA). At least 60 cardiomyocytes were examined in each heart (n = 3) for a total of 180 cardiomyocytes for each condition, and the data were averaged (2).

Gene microarrays.

Samples were obtained from the RV free wall 10 days after either PAC or sham operation. Mice with severe PS were used for gene expression studies (n = 16). Total RNA was isolated using the RNeasy Mini Kit (Qiagen, Valencia, CA), and RNA was then reverse transcribed to double-stranded cDNA. Labeled cRNA was synthesized by the incubation of 1 g cDNA with biotin-labeled ribonucleotides and RNA polymerase for 5 h at 37°C using the BioArray High Yield RNA transcript labeling kit (Enzo Diagnostics, Farmingdale, NY). Biotin-labeled cRNAs were then fragmented by heating and hybridized onto microarrays composed of 38,467 70mer oligonucleotide probes representing over 25,000 genes, essentially the entire mouse transcriptome. The array includes 35,302 well-annotated probes targeting mouse genes and alternative exons, as well as multiple spots for biological and technical controls. For more details on the microarray platform and our methods for quality control, see Wagner et al. (71), Zhao et al. (73), and the Stanford Functional Genomics Web site (60). Microarrays were scanned on an Agilent G2565AA microarray scanner. Quantitative RT-PCR (QRT-PCR) was performed for each of the major differentially expressed genes in discussion for verification. Gene expression data from RV samples after PAC were then compared with LV samples obtained from mice after 10 days of aortic banding [transverse aortic constriction (TAC)], obtained from previous studies from our laboratory (73).

Microarray analysis.

Statistical analysis was performed using Stanford microarray database software, with subsets of the data exported to the Institute for Genome Research Multiple Array Viewer (55), significance analysis of microarrays (SAM), and classification software such as prediction analysis of microarrays (62, 67). Clustering algorithms used include two-dimensional hierarchical clustering analysis, K-means clustering, ingenuity pathway assist (Ingenuity Systems, Redwood City, CA), and self-organizing maps (59, 63). These analyses identify smaller clusters of genes with distinct expression patterns that highlight distinct characteristics of subsets of the experimental samples. Gene ontology (GO) overrepresentation analysis was used to identify biological processes which were up- or downregulated. Groups of genes identified as differentially regulated were analyzed for GO class overrepresentation using Fisher's exact test.

Model of TAC.

RV gene expression during PAC was compared with LV gene expression during TAC as published previously (73). Anesthesia was induced with 3% isoflurane and maintained with 1.5% isoflurane. TAC was performed via a left thoracotomy incision, avoiding the pleural space and, hence, the need for artificial ventilation, as described by Rockman et al. (52). A 7-0 silk suture was placed around the transverse aorta between the left common carotid artery and the brachiocephalic trunk and tied tight around both the aorta and a 27-gauge needle, which was then removed, yielding a reproducible degree of constriction. LV myocardium was obtained at 10 days postoperatively and hybridized to Affymetrix U74Av2 mouse genome arrays (Affymetrix, Santa Clara, CA), containing 12,488 known genes and expressed sequences tags (ESTs). Details of the analysis of these samples have been published previously (73).

Statistical analysis.

The XL stat software (Addinsoft) was used. Values are expressed as means ± SD. Statistical significance of differences in mean values of physiological parameters from sham-operated and PAC (mild, moderate, and severe PS) mice were assessed by one-way ANOVA with Fisher's paired least significant difference post hoc testing using Statview software (SAS, Cary, NC). A probability value ≤0.05 was considered significant.

Animal care.

All procedures were performed in accordance with National Institutes of Health standards and were approved by the Administrative Panel on Laboratory Animal Care at Stanford University.

RESULTS

Survival and evidence of right heart failure.

PAC was performed on 86 male FVB mice. In the early phase of our studies, four mice died during the procedure due to hemorrhage from the PA. Subsequent survival has been 100%. Echocardiographic evaluation to assess RV-PA PPG was performed on postoperative day 7 in 82 mice. Two mice with an RV-PA PPG of <20 mmHg were excluded, leaving 80 mice for further study. Thus, after a short initial learning curve, PAC can be successfully performed in 92% of mice.

Thirty-nine mice were categorized into the severe PS group with evidence of RV enlargement and either a flat interventricular septum or concave septal shift (Fig. 1, B and C). Twenty-nine mice were categorized into the moderate PS group without a flattening of the interventricular septum, and 12 mice were categorized into the mild PS group (Fig. 1A). Within each group, one subgroup was evaluated by weekly echocardiography and survival analysis performed by a log-rank analysis of Kaplan-Meier curves (Fig. 2). There was a significant difference in survival between the mild, moderate, and severe PS groups. All mice with mild PS survived to 100 days, which was the end of our study. In contrast, the mean survival was 50.8 days in the moderate PS group and only 19.6 days in the severe PS group (P < 0.001).

Fig. 2.

Kaplan-Meier survival curves comparing mice with mild PS (n = 5), moderate PS (n = 13), and severe PS (n = 16) after pulmonary artery (PA) constriction (PAC). Survival half-time was 19.6 days in severe PS vs. 50.8 days in moderate PS. None of the mice with mild PS died by 100 days, which was the end of our study. There was a significant difference between moderate PS and severe PS survival (P < 0.001 by log-rank analysis).

There was no significant difference between groups in BW or in HR at 7 days postoperative (Table 1); however, all mice with severe PS developed signs of right heart failure with peripheral edema by 7 days postoperative (Fig. 3A). The edema was moderate in 35 mice and severe in four mice. No mice with mild or moderate PS developed peripheral edema.

Table 1.

Echocardiographic parameters of PAC and sham-operated controls at 1 wk postoperative

	Control	Mild	Moderate	Severe	ANOVA
n	16	5	13	16
BW, g	27.7±3.6	27.9±1.4	27.8±1.6	27.7±3.6	NS
HR, beats/min	449.2±36.2	457.4±37.4	451.8±34.0	462.0±38.2	NS
LVDd, mm	3.18±0.15	3.16±0.19	3.14±0.24	1.89±0.21*	P < 0.001
LV FS, %	44.65±2.46	43.26±2.82	44.94±3.30	63.01±5.49*	P < 0.001
RVOTDd, mm	1.19±0.07	1.20±0.07	1.40±0.11*	1.71±0.14*	P < 0.001
RVOT FS, %	32.77±4.44	33.55±6.77	29.52±3.11	16.39±3.92*	P < 0.001
PPG/RV-PA, mmHg	2.55±1.56	26.35±3.47*	41.45±4.39*	47.54±3.67*	P < 0.001
VTI, cm	5.59±0.51	5.74±0.38	4.19±0.41*	1.86±0.22*	P < 0.001
CO, ml/min	28.08±4.97	29.77±4.57	29.46±6.41	19.80±3.48*	P < 0.001

Values are means ± SD; n, no. of mice/group.

^*P < 0.001 vs. control. PAC, pulmonary artery (PA) constriction; BW, body weight; HR, heart rate; LVDd, left ventricular (LV) diastolic dimension; FS, fractioning shortening; RVOTDd, right ventricular (RV) outflow tract (RVOT) diastolic dimension; PPG, peak pressure gradient; VTI, velocity-time integral; CO, cardiac output; NS, not significant.

Fig. 3.

A: a mouse with generalized edema 2 wk after banding (right) compared with a sham-operated control (left). B: representative ECGs before (a) and after (b) PAC. Mice with severe PS showed ECG findings of a widened QRS with rsR′ pattern, consistent with an incomplete right bundle branch block, as well as ST-T elevations within 7 days of PAC. Mice with mild or moderate PS did not show these ECG findings.

Electrocardiographic findings 1 wk after PAC.

ECGs were performed 1 wk after PAC in five mice with mild PS, in 13 mice with moderate PS, and in 16 mice with severe PS. Eleven of the 16 (69%) mice with severe PS developed a widened QRS with an rsR′ pattern [incomplete right bundle branch block (iRBBB)] along with ST-T elevations (Fig. 3B). None of the mice with mild or moderate PS developed ECG abnormalities.

Echocardiographic findings 1 wk after PAC.

Characteristics of each PAC group and sham-operated controls at postoperative day 7 are shown in Table 1. Mice with moderate PS showed a mild enlargement of the RV (as assessed by the short-axis RVOT diameter) versus that of mice with mild PS and sham-operated controls. In contrast, mice with severe PS showed a markedly enlarged RV chamber (increased by 44%) and either flattening (Fig. 1B) of the septum or bowing of the septum into the LV, with the LV assuming a crescent-shaped appearance in the short-axis (Fig. 1C). In all mice with severe PS, there was also evidence of tricuspid regurgitation (Fig. 1D), which was not present in mice with mild or moderate PS.

RV-LV interaction and LV dyssynchrony.

Color M-mode echocardiograms in mild and moderate PS showed that interventricular septal (IVS) motion was coordinated with the LV free wall, whereas there was paradoxical septal motion during systole in severe PS (Fig. 4A). Combined with the right bundle branch block, this resulted in intraleft ventricular dyssynchrony, similar to that seen in many patients with repaired tetralogy of Fallot. The delay between anterior LV free wall and posterior septal maximal excursion averaged 33.3 ± 15.6 ms in mice with severe PS. TDI also showed significant septal-LV free wall dyssynchrony, which was present only in mice with severe PS (Fig. 4B). Left ventricle diastolic dimension (LVDd) was significantly lower in severe PS versus the other groups (Table 1). Because of the marked decrease in LV volume, M-mode-derived LV FS was increased; however, RVOT FS was reduced in the severe PS group (see Serial changes in LV and RV function), and cardiac output (CO), derived by Doppler echocardiography, was significantly reduced in mice with severe PS (Table 1).

Fig. 4.

A: representative color M-mode echocardiogram from a mouse with moderate PS (e) compared with severe PS (f). In moderate PS, the IVS motion is coordinated with the LV posterior wall (LVPW), whereas in severe PS, there is paradoxical IVS motion during systole and a time delay between LVPW motion and IVS motion (arrows). B: representative tissue Doppler image (TDI) showing LV dyssynchrony in a mouse with severe PS. Doppler regions of interest are marked on the 2-dimensional echo image (left) with a green circle (IVS) and yellow circle (LVPW). The corresponding wall motion plots show the time differential between the IVS (green plot and green arrow) and the LVPW (yellow plot and yellow arrow).

Serial changes in LV and RV function.

LV FS was increased at 1 wk in severe PS, as LV dimensions were decreased secondary to the septal shift (Fig. 5, top). LV FS was not different from control in mild and moderate PS. LV FS did not change significantly over the course of 5 wk in any of the three groups. In contrast, RVOT FS was significantly reduced at all time points in the severe PS group (P < 0.001 vs. mild and moderate PS). In the moderate PS group, RVOT FS was initially not different from control, but by 4 wk after PA banding began to decrease (P < 0.01 vs. control; Fig. 5, middle). In the mild PS group, RVOT FS remained at normal control levels (based on data obtained from nonbanded mice of similar strain and age in our laboratory) throughout the study period.

Fig. 5.

Serial evaluation of LV and RV function in mice with PAC. Top: LV fractional shortening (FS) was increased at 1 wk (w) in severe PS (primarily due to reduction in LV cavity size) and normal in mild and moderate PS. There was no significant difference between 1 and 4 wk in each group. Middle: RV outflow tract (RVOT) FS (in %) was decreased at 1 wk in severe PS and normal in mild and moderate PS. There was no further decline in RV FS in severe PS mice; however, RV FS began to decrease in mice with moderate PS by 3 wk. Bottom: RV wall velocity was decreased at 1 wk in severe PS and normal in mild and moderate PS. As with RVOT FS, mice with moderate PS showed a decrease in RV wall velocity by 3 wk. †P < 0.01 vs. 1 wk; *P < 0.01 vs. mild and moderate. Data obtained from 12 control mice before any surgical procedure are also shown.

Although used with good results in humans and larger animals, TDI has not been utilized extensively in the mouse. We found that RV wall velocity obtained from TDI can be performed with good reliability in this species. Inter- and intraobserver variability for TDI were 8.4 ± 3.1% and 7.2 ± 4.2%, respectively. Serial measurements of RV wall velocity were recorded in mild, moderate, and severe PS groups (Fig. 5, bottom). RV wall velocity was significantly reduced at all time points in severe PS group compared with control (P < 0.001). RV wall velocity started out at control levels and then significantly fell at 4 and 5 wk in the moderate PS group (P < 0.05). RV wall velocity remained at control levels in the mild PS group.

Invasive hemodynamics.

To correlate RV wall velocity obtained from TDI with invasive measurement of RV function, 23 mice representing varying degrees of RV banding (5 controls without PAC, 2 with mild PS, 9 with moderate PS, and 7 with severe PS) were catheterized at different time points after PAC. Echocardiographic and hemodynamic studies were done concurrently under the same anesthetic conditions. RV systolic pressure and dP/dt increased immediately after the placement of the PA band (Fig. 6A). RV end-diastolic pressure (RVEDP) was increased at 6 h after PAC (vs. pre, P < 0.01) and recovered by 24 h. At 10 days after PAC, mice with mild and moderate PS had mild elevations in RVEDP. In contrast, those with severe PS had significant elevations in RVEDP (P < 0.01; Fig. 6B). There was an excellent correlation (r = 0.833; P < 0.0001) between noninvasive measurements of RV free wall velocity and invasive measurements of RV maximal dP/dt (dP/dt_max; Fig. 6C), indicating that RV free wall velocity is a reasonable surrogate for measurements of RV dP/dt when serial measurements are required.

Fig. 6.

Invasive hemodynamics. A: maximal change in pressure over time (dP/dt_max) increases immediately after PA banding (arrow). Tracing shows continuous measurements over 5 min. B: RV end-diastolic pressure (RVEDP) was significantly elevated at 6 h after banding and returned to baseline levels by 24 h. At 10 days (d), in mice with severe PS, RVEDP was significantly increased compared with that of pre-PAC (*P < 0.01) and compared with that of mild and moderate PS (*P < 0.01). Prebanding, n = 3 at 6 h, n = 3 at 24 h, n = 3 at 10 day; mild, n = 4 at 10 day; moderate, n = 5 at 10 day; severe, n = 5 at 10 day. C: there was a good correlation between RV free wall velocity obtained from TDI and dP/dt_max measured at various time intervals from 6 h to 10 day (r = 0.833; P < 0.0001; n = 23).

Pathology and morphometrics.

Table 2 shows heart weight data from each of the three groups at 10 days postoperative. In the severe PS group, heart weight and heart weight-to-BW ratio were markedly increased. All of this increase in weight was due to the increased weight of the RV, which increased by 80%, with no significant change in LV weight. The RV-to-BW ratio was also significantly increased in the moderate PS group. Additional evidence of RV failure in the severe PS group was manifested by an increase in liver weight/BW by 22%, which was not present in the mild and moderate PS groups.

Table 2.

Morphometric characteristics of PAC and sham-control mice at 10 days postoperative

	Sham	Mild	Moderate	Severe	ANOVA
n	16	7	7	16
BW, g	27.74±1.98	28.66±1.08	28.10±0.88	27.34±3.49	NS
HW, mg	122.93±10.78	127.04±9.41	133.60±4.60	142.79±15.86*	<0.001
HW/BW, mg/g	4.45±0.34	4.43±0.26	4.76±0.22	5.25±0.50*	<0.001
LVW/BW, mg/g	1.95±0.25	1.79±0.13	1.81±0.09	1.87±0.24	NS
RVW/BW, mg/g	0.94±0.10	1.05±0.08	1.41±0.15*	1.70±0.28*	<0.001
Sep/BW, mg/g	1.18±0.18	1.12±0.11	1.15±0.14	1.29±0.21	NS
Liver/BW, mg/g	36.35±7.10	35.01±2.91	36.20±3.41	44.33±6.83*	<0.001

Values are means ± SD; n, no. of mice/group. These sham and severe pulmonary stenosis (PS) groups were used for gene microarray studies.

^*P < 0.001 vs. control. HW, heart weight; LVW, LV weight; RVW, RV weight; Sep, interventricular septal weight.

A microscopic evaluation showed that the RV cavity was significantly dilated at 10 days in the severe PS group (Fig. 7A). The average area of RV myocytes was significantly increased in moderate PS versus control (340.9 ± 20.7 vs. 203.0 ± 6.2 μm²; P < 0.01) and even further increased in severe PS (454.5 ± 10.2 μm²; P < 0.01; Fig. 7, B and C). In all mice with PS, the degree of PS correlated well with the RV-to-BW ratio (r = 0.692) and to histological evidence of cardiomyocyte hypertrophy (r = 0.768). There was no evidence of fibrosis or inflammation.

Fig. 7.

A: low-power microscopic examination (Masson trichrome stain) showing increased RV free wall and cavity dimensions in severe PS compared with sham control. B: high-power microscopic examination (wheat germ agglutinin with 4-6-diamidino-2-phenylindole) showing increased myocyte cross-sectional areas in severe PS compared with sham control. C: cell area was significantly increased in moderate and severe PS groups compared with sham. *P < 0.05 vs. sham control.

Gene expression in the afterload-stressed RV.

RV samples from 16 severe PS mice at 10 days postoperative and from 16 sham controls were hybridized to gene microarrays, and data were analyzed using SAM and Fisher's exact test algorithms to rigorously identify differentially expressed genes and GO biological processes. Relative gene expression was expressed as the fold difference in gene expression in PAC RV versus sham RV. One-hundred ninety-six genes were upregulated and 1,114 genes were downregulated in the PAC RV compared with those of controls (Tables 3 and 4). Fisher's exact test was applied to the 8,773 unique GO annotated genes on the array to identify statistically significantly enriched and depleted GO groups in the PAC RV (Table 5). Among the most significantly upregulated processes were phosphate transport, regulation of coagulation, inorganic anion transport, and cell adhesion, which includes most of the collagens as well as other extracellular matrix (ECM) genes. Downregulated processes were dominated by energy pathways.

Table 3.

135 genes significantly upregulated >1.5-fold in PAC RV versus sham RV^*

Gene Name	Gene Description	GenBank ID	Score(d)	Fold Change
Postn	Periostin	AV084876	3.056	16.04
Fnbp4	Formin binding protein 4	AA240645	4.962	11.61
Lox	Lysyl oxidase	AV014751	3.472	7.68
Mfap4	Microfibrillar associated protein 4	BG074573	4.417	7.48
Sfrp2	Secreted frizzled related protein 2	AV021712	2.267	7.48
Uchl1	Ubiquitin carboxy-terminal hydrolase	BG074009	3.031	7.23
Aspn	Asporin	AV020793	3.454	6.21
Col3a1	Procollagen, type III, alpha 1	BG073709	2.808	4.16
Dkk3	Dickkopf homolog 3	BG075561	5.263	4.16
Serpinb1a	Serine (or cysteine) peptidase inhibitor, clade B, member 1a	BG073257	2.816	4.13
Col5a2	Procollagen, type V, alpha 2	AV089281	3.118	4.02
Col5a1	Procollagen, type V, alpha 1	BG067011	3.574	3.88
Col3a1	Procollagen, type III, alpha 1	BG074327	3.437	3.86
Fstl1	Follistatin-like 1	AV024220	5.335	3.71
Set	Set translocation, myeloid leukemia-associated	AV031220	4.968	3.63
Traf4	Tnf receptor associated factor 4	AV024412	4.870	3.60
Sparc	Secreted acidic cysteine rich glycoprotein	AV094848	4.116	3.51
Ccnd2	Cyclin D2	AV087918	4.030	3.51
Mgp	Matrix Gla protein	BG074366	4.218	3.50
Fstl1	Follistatin-like 1	BG073316	4.191	3.36
Sssca1	Sjogren's syndrome/scleroderma autoantigen 1 homolog	AV023779	3.433	3.35
Lpp	LIM domain containing preferred translocation partner in lipoma	BG068912	3.110	3.31
Nupr1	P8 protein	AV087068	3.237	3.30
Rps20	Ribosomal protein S20	AV170826	6.675	3.23
Bgn	Biglycan	BG073809	5.770	3.17
Prss23	Protease, serine, 23	AV073989	4.318	3.17
Ubox5	U-box domain containing 5	BG072221	3.956	3.16
Rbp1	Retinol-binding protein 1	AV074050	3.565	3.15
Sparc	Secreted acidic cysteine rich glycoprotein	AV104148	2.859	3.15
Anxa7	Annexin A7	BG076013	4.491	3.03
Pola1	Polymerase (DNA directed), alpha 1	AV095001	4.679	3.03
Nbl1	Neuroblastoma candidate region, suppression of tumorigenicity 1	AV078033	5.377	2.83
Col14a1	Procollagen, type XIV, alpha 1	AV017616	2.809	2.80
Tacstd1	Tumor-associated calcium signal transducer 1	AV089835	5.118	2.80
Col3a1	Collagen, type III, alpha-1	BG072787	2.802	2.80
Serpinb1a	Serine (or cysteine) peptidase inhibitor, clade B, member 1a	AV061227	2.578	2.77
Gpx3	Glutathione peroxidase 3	AV038358	6.227	2.76
Cd48	CD48 antigen	AV056452	2.879	2.75
Rbp1	Retinol-binding protein 1	AV146205	3.359	2.71
Rkhd2	Mex3 homolog C (C. elegans)	AV024424	4.460	2.70
Gata1	GATA binding protein 1	AV024112	3.820	2.67
Shc1	Src homology 2 domain-containing transforming protein C1	BG070010	3.863	2.65
Pdgfrl	Platelet-derived growth factor receptor-like	AV013190	3.206	2.64
Gpx3	Glutathione peroxidase 3	AV137417	3.153	2.63
Cyb5r3	Cytochrome b5 reductase 3	BG067095	2.821	2.63
Loxl1	Lysyl oxidase-like 1	AV094998	5.053	2.60
Hexb	Hexosaminidase B	BG069642	2.816	2.57
Proz	Protein Z, vitamin K-dependent plasma glycoprotein	AV078387	2.846	2.56
Fndc1	Fibronectin type III domain containing 1	AV010532	2.959	2.54
Clu	Clusterin	AV149922	6.506	2.53
Nt5dc2	5′-nucleotidase domain containing 2	AW547246	3.488	2.53
Rtn4	Reticulon 4	BG064276	3.027	2.48
Col4a1	Procollagen, type IV, alpha 1	AA162273	4.510	2.46
Snx10	Sorting nexin 10	AV095218	3.147	2.43
Fhl1	Four and a half LIM domains 1	AV083596	3.399	2.42
Clu	Clusterin	AV074721	4.009	2.39
Traf5	Tnf receptor-associated factor 5	AV091488	4.721	2.37
Anxa1	Annexin A1	AV037865	3.434	2.33
Nupr1	P8 protein	AV087698	3.090	2.31
Cd24a	CD24a antigen	AV105800	2.997	2.31
Rbp1	Retinol-binding protein 1	AV140184	8.093	2.30
Zfp87	Zinc finger protein 87	BG075308	3.210	2.30
Rbbp7	Retinoblastoma-binding protein 7	AW544081	6.911	2.29
Vim	Vimentin	AV113424	4.562	2.27
Zfp364	Zinc finger protein 364	BG072444	3.726	2.26
Col4a2	Procollagen, type IV, alpha 2	AV010312	2.755	2.26
Ryr2	Ryanodine receptor 2	BG074010	2.827	2.24
H13	Histocompatibility 13	AV055621	2.646	2.22
Dstn	Destrin	BG073428	3.185	2.21
Synpo21	Synaptopodin 2-like	AV015246	5.243	2.21
Cola2	Collagen alpha-2(I) chain precursor	AV009300	2.797	2.20
Arl6ip2	ADP-ribosylation-like factor 6-interacting protein 2	AV015499	3.031	2.18
Plp2	Proteolipid protein 2	AV133831	4.596	2.15
Cdv3	Carnitine deficiency-associated gene expressed in ventricle 3	AV072373	3.324	2.11
Dstn	Destrin	AV050410	4.280	2.10
Angpt2	Angiopoietin 2	AA020573	4.174	2.09
Rhoc	Ras homolog gene family, member C	AV140333	3.965	2.08
Arpc1b	Actin related protein 2/3 complex, subunit 1B	AV000246	4.745	2.07
Anxa5	Annexin A5	AV087971	2.870	2.04
Gnb1	Guanine nucleotide binding protein, beta 1	AV078383	2.626	2.03
Pmp22	Peripheral myelin protein 22	AV087039	3.746	2.02
Serpine2	Serine (or cysteine) peptidase inhibitor, clade E, member 2	AV017162	3.403	2.01
Anxa3	Annexin A3	AV218319	5.386	1.98
Mif	Similar to macrophage migration inhibitory factor	BG064410	3.676	1.96
Anxa4	Annexin A4	AV103319	2.660	1.96
Rdh5	Biogenesis of lysosome-related organelles complex-1, subunit 1	AV083165	3.014	1.95
Cd63	Cd63 antigen	AV093530	3.026	1.94
Clu	Clusterin	BG072209	2.952	1.94
Vim	Vimentin	AV123111	2.969	1.93
Atp6v1a	ATPase, H+ transporting, lysosomal V1 subunit A	BG064589	2.882	1.93
Arf2	ADP-ADP-ribosylation factor 2	AV030860	2.722	1.92
Plk2	similar to XP_001102530.1 polo-like kinase 2 (Drosophila) isoform 2	AV049483	3.516	1.91
Nid2	Nidogen 2	AV013588	2.732	1.90
Tmem176b	Transmembrane protein 176B	AV013352	4.584	1.88
Cnn2	Calponin 2	AV025199	3.051	1.88
Prkar1a	Protein kinase, cAMP-dependent, regulatory, type I, alpha	BB566556	4.444	1.87
Cst3	Cystatin 3	AV153101	3.212	1.86
Anxa5	Annexin A5	AA137915	2.698	1.86
Rell1	RELT-like 1	AV031438	3.319	1.84
Cd34	CD34 antigen	AV086521	3.544	1.83
Pmp22	Peripheral myelin protein 22	AV113888	3.100	1.83
Azin1	Antizyme inhibitor 1	AV030927	4.074	1.82
Cald1	Caldesmon 1	BG064630	2.635	1.81
Dstn	Destrin	AV087224	2.677	1.80
Ccnd2	Cyclin D2	AV140268	3.939	1.79
Gfm2	G elongation factor, mitochondrial 2	AV095236	3.684	1.77
Pik3ca	Phosphatidylinositol 3-kinase, catalytic, alpha polypeptide	BG076256	3.014	1.77
Mtdh	Metadherin	AV083741	2.761	1.76
Tusc3	strongly similar to XP_001094069.1 similar to tumor suppressor candidate 3 isoform a isoform 4	AV031846	2.741	1.74
Pkm2	Pyruvate kinase, muscle, 2	AV094449	3.977	1.70
Actn4	Actinin alpha 4	AI836968	3.079	1.69
Hn1	Hematological and neurological expressed sequence 1	AV094890	3.892	1.69
Ifitm2	Interferon-induced transmembrane protein 2	AV049395	2.671	1.66
Ppgb	Beta-galactosidase protective protein	AV088011	2.791	1.65
Ntan1	N-terminal Asn amidase	AV058809	2.852	1.63
Tmem43	Transmembrane protein 43	BG073951	2.763	1.61
Akt1	V-AKT murine thymoma viral oncogene homolog 1	AV058304	2.705	1.60
Pigq	Phosphatidylinositol glycan anchor biosynthesis, class Q	AV006019	3.635	1.59
Ugp2	UDP-glucose pyrophosphorylase 2	AV086208	3.047	1.58
Sec13l1	SEC13 homolog	BG065187	3.118	1.57
Eif4ebp1	Eukaryotic translation initiation factor 4E-binding protein 1	AV087438	2.733	1.56
Wif1	WNT Inhibitory factor 1	AV032229	3.266	1.56
Gmfb	Glia maturation factor, beta	AV162369	2.669	1.55
S100a16	S100 calcium binding protein A16	AV088022	2.573	1.55
Gusb	Beta-glucuronidase	AV111448	3.314	1.52
Gas6	Growth arrest-specific 6	BG076011	2.738	1.52
Rdbp	Complement factor B	AV133629	2.624	1.52
Btg2	B-cell translocation gene 2	AV086968	2.673	1.52
Taf10	Transcribed locus, strongly similar to XP_001109041.1 similar to integrin-linked kinase isoform 3	BG072693	2.902	1.50
Sptlc1	Serine palmitoyltransferase, long chain base subunit 1	AV062462	2.930	1.47
Nptn	Neuroplastin	AV088324	3.211	1.46
Mgmt	Methylguanine-DNA methyltransferase	AV087599	2.696	1.44
Ehd1	EH-domain containing 1	AV024484	2.590	1.39
Jund1	Jun proto-oncogene related gene d1	AV014760	2.668	1.31
Dazap2	DAZ-associated protein 2	AV082449	3.466	1.31

The significance analysis of microarrays (SAM) algorithm was employed to identify genes with statistically different expression levels between PA-banded and sham tissues. SAM incorporates a false discovery rate (FDR) correction for multiple testing errors and calculates a d statistic for each gene based on the ratio of change in gene expression to SD in the data for that gene (Ref. 18).

^*Out of a total of 196 significantly upregulated genes and expressed sequence tags (ESTs).

Table 4.

250 genes significantly downregulated >2.0-fold in PAC RV versus sham RV^*

Gene Name	Gene Description	Gene ID	Score(d)	Fold Down
Fkbp4	FK506 binding protein 4	BG065656	−2.193	6.54
Mospd1	Motile sperm domain containing 1	BG068741	−2.395	5.38
Acaa2	Acetyl-Coenzyme A acyltransferase 2 (mitochondrial 3-oxoacyl-Coenzyme A thiolase)	BG072552	−2.856	5.28
Sybl1	Synaptobrevin-like 1	AV113528	−4.873	5.00
Ces3	Carboxylesterase 3	BG072503	−2.156	4.93
A2bp1	Ataxin 2-binding protein 1	AV029909	−6.284	4.86
Stk16	Serine/threonine kinase 16	AV119666	−2.037	4.83
Dgat2	Diacylglycerol O-acyltransferase 2	AI847556	−2.031	4.61
Aes	Amino-terminal enhancer of split	BG074671	−2.295	4.09
Hod	HOP homeobox	AI840878	−3.112	4.07
Ppil1	Peptidylprolyl isomerase (cyclophilin)-like 1	AV015645	−2.154	3.92
Hod	HOP homeobox	AV065655	−2.097	3.60
Serpinb11	Serine (or cysteine) peptidase inhibitor, clade B (ovalbumin), member 11	AV082348	−2.206	3.53
Prdx3	Peroxiredoxin 3	BG074871	−2.300	3.50
Dci	Dodecenoyl-Coenzyme A delta isomerase (3,2 trans-enoyl-Coenyme A isomerase)	AV106338	−3.290	3.49
Dci	Dodecenoyl-Coenzyme A delta isomerase (3,2 trans-enoyl-Coenyme A isomerase)	BG069423	−8.829	3.43
Tuba8	Tubulin, alpha 8	AA063914	−3.529	3.35
Hadh2	Hydroxysteroid (17-beta) dehydrogenase 10	BG073539	−2.043	3.32
Hibadh	3-hydroxyisobutyrate dehydrogenase	AI854120	−2.658	3.28
Aarsd1	Alanyl-tRNA synthetase domain containing 1	AI847872	−2.278	3.27
Acat1	Acetyl-Coenzyme A acetyltransferase 1	AV084664	−2.558	3.24
Auh	AU RNA binding protein/enoyl-coenzyme A hydratase	AV095181	−2.413	3.21
Ndufb7	NADH dehydrogenase (ubiquinone) 1 beta subcomplex, 7	BG075930	−2.110	3.18
Sf3b2	Splicing factor 3b, subunit 2	AV065784	−2.965	3.17
Csnk2b	Casein kinase 2, beta polypeptide	BG075196	−2.013	3.16
Ythdf2	YTH domain family 2	AV084848	−1.994	3.13
Hadhb	Hydroxyacyl-Coenzyme A dehydrogenase/3-ketoacyl-Coenzyme A thiolase/enoyl-Coenzyme A hydratase (trifunctional protein), beta subunit	BG064090	−2.257	3.12
Ndufa3	NADH-ubiquinone oxidoreductase 1 alpha subcomplex, 3	AV054731	−2.424	3.12
Tuba4	Tubulin, alpha 4A	AI840604	−7.166	3.09
Arhgap12	Rho GTPase activating protein 12	BG064038	−1.991	3.07
Atp5 g2	ATP synthase, H+ transporting, mitochondrial F0 complex, subunit c (subunit 9), isoform 2	AV056821	−3.260	3.06
Hdlbp	High density lipoprotein (HDL) binding protein	AV056261	−2.815	3.04
Eno3	Enolase 3, beta muscle	AI841640	−6.989	3.04
Mapk14	Mitogen activated protein kinase 14	AA544997	−5.407	3.03
Mrpl32	Mitochondrial ribosomal protein L32	AV035121	−2.995	3.02
Hod	HOP homeobox	AV081983	−2.911	3.00
Hadhsc	Hydroxyacyl-Coenzyme A dehydrogenase	AV013144	−2.652	2.92
Rnf6	Ring finger protein 6	AV015385	−2.035	2.92
Gpsn2	Glycoprotein, synaptic 2	AV106079	−2.173	2.91
Acaa2	Acetyl-Coenzyme A acyltransferase 2 (mitochondrial 3-oxoacyl-Coenzyme A thiolase)	AV084156	−5.371	2.90
Hadha	Hydroxyacyl-Coenzyme A dehydrogenase/3-ketoacyl-Coenzyme A thiolase/enoyl-Coenzyme A hydratase (trifunctional protein), alpha subunit	BG074757	−2.412	2.90
Dcps	Decapping enzyme, scavenger	BG063230	−4.204	2.90
Hrc	Histidine rich calcium binding protein	BG073810	−2.371	2.88
Pex11b	Peroxisomal biogenesis factor 11b	AV006229	−4.785	2.86
Pipox	Pipecolic acid oxidase	AV069402	−3.867	2.85
Ckmt2	Creatine kinase, mitochondrial 2	AV085004	−2.245	2.85
Acadm	Acyl-Coenzyme A dehydrogenase, medium chain	AI840666	−3.584	2.85
Etfb	Electron transferring flavoprotein, beta polypeptide	AV086609	−6.881	2.84
Sdha	Succinate dehydrogenase complex, subunit A, flavoprotein (Fp)	AV074725	−2.909	2.84
Tmem176b	Transmembrane protein 176B	BI076417	−2.922	2.83
Hspd1	Heat shock protein 1 (chaperonin)	BG064728	−3.015	2.83
Ak2	Adenylate kinase 2	AV088727	−2.390	2.81
Synj2	Synaptojanin 2	AV083725	−2.900	2.81
Tcea3	Transcription elongation factor A (SII), 3	BG071507	−7.615	2.77
Dscr3	Down syndrome critical region gene 3	AV111492	−2.132	2.77
Ndufv1	NADH dehydrogenase (ubiquinone) flavoprotein 1	BG076088	−2.697	2.77
Cox8b	Cytochrome c oxidase, subunit VIIIb	AV083848	−2.834	2.76
Fkbp4	FK506 binding protein 4	BG064128	−3.876	2.76
Etfa	Electron transferring flavoprotein, alpha polypeptide	AA139719	−6.798	2.76
Vdac1	Voltage-dependent anion channel 1	AV095022	−2.426	2.76
Fahd1	Fumarylacetoacetate hydrolase domain containing 1	AV061746	−2.902	2.74
Tuba4	alpha-tubulin TUBA4	AV032310	−4.484	2.73
Ech1	Enoyl coenzyme A hydratase 1, peroxisomal	AV000529	−4.862	2.72
Txndc1	Thioredoxin domain containing 1	BG064124	−2.833	2.70
Zan	Zonadhesin	AV326603	−3.831	2.68
Eed	Embryonic ectoderm development	BG068740	−2.137	2.66
Zfp644	Zinc finger protein 644	BG069858	−3.025	2.66
Idh2	Isocitrate dehydrogenase 2 (NADP+), mitochondrial	AV006267	−3.493	2.65
Mrps17	Mitochondrial ribosomal protein S17	BG073811	−3.625	2.64
Kif4	Kinesin family member 4	AV355663	−2.785	2.64
Txn2	Thioredoxin 2	AA116866	−2.667	2.62
Nudt4	Nudix (nucleoside diphosphate linked moiety X)-type motif 4	AI854103	−2.131	2.60
Fyco1	FYVE and coiled-coil domain containing 1	AV006218	−3.682	2.59
Dld	Dihydrolipoamide dehydrogenase	AI847502	−2.849	2.59
Hfe2	Hemochromatosis type 2 (juvenile) (human homolog)	AV087892	−3.091	2.59
Rabepk	Rab9 effector protein with kelch motifs	BG075188	−2.783	2.59
Lrpprc	Leucine-rich PPR-motif containing	AV162299	−3.973	2.57
Acsl1	Acyl-CoA synthetase long-chain family member 1	AV005791	−4.637	2.56
Slc2a4	Solute carrier family 2 (facilitated glucose transporter), member 4	AV005800	−3.023	2.56
Idh3b	Isocitrate dehydrogenase 3 (NAD+) beta	AI838687	−3.678	2.56
Hspd1	Heat shock 60-KD protein 1	BG073067	−3.334	2.55
Armet	Arginine-rich, mutated in early stage tumors	BG063580	−2.050	2.55
Slc25a11	Solute carrier family 25 (mitochondrial carrier, oxoglutarate carrier), member11	AV089747	−3.883	2.54
Acadm	Acyl-Coenzyme A dehydrogenase, medium chain	AV086733	−7.705	2.54
Zbtb20	Zinc finger and BTB domain containing 20	BG073885	−4.558	2.53
Acads	ACYL-CoA dehydrogenase, short-chain	AV093663	−2.256	2.52
Tceb1	Transcription elongation factor B (SIII), polypeptide 1	BG071546	−3.545	2.52
Suclg2	Succinate-CoA ligase, GDP-forming, beta subunit	AV087975	−2.577	2.52
Etfa	Electron transferring flavoprotein, alpha polypeptide	BG074876	−3.079	2.52
Fkbp4	FK506 binding protein 4	AV111500	−2.614	2.52
Sec31l1	SEC31 homolog A (S. cerevisiae)	BG074808	−2.125	2.52
Atp5d	ATP synthase, H+ transporting, mitochondrial F1 complex, delta subunit	AI841365	−2.169	2.52
Eif5b	Eukaryotic translation initiation factor 5B	BG067345	−2.028	2.47
Nnt	Nicotinamide nucleotide transhydrogenase	AV006306	−6.495	2.47
Il12b	Interleukin 12b	AA267353	−2.568	2.46
Atp5f1	ATP synthase, H+ transporting, mitochondrial F0 complex, subunit b, isoform 1	AA119394	−3.970	2.46
Exosc10	Exosome component 10	BG063453	−2.972	2.46
Prss2	Transcribed locus, weakly similar to NP_062562.1 factor VII precursor, isoform b	AV005299	−2.481	2.45
Uqcrc1	Ubiquinol-cytochrome c reductase core protein I	AI841290	−2.836	2.43
Magmas	Mitochondria-associated protein involved in granulocyte-macrophage colony-stimulating factor signal transduction	AV066653	−2.234	2.43
Atrx	Alpha thalassemia/mental retardation syndrome X-linked homolog	BG068630	−3.518	2.42
Slc35b1	Solute carrier family 35, member B1	AV060614	−2.117	2.42
Mrps28	Mitochondrial ribosomal protein S28	AV107441	−2.906	2.41
Cox5b	Cytochrome c oxidase, subunit Vb	AV066262	−2.118	2.41
D19Ertd678e	Coiled-coil domain containing 86	AV094491	−3.703	2.41
Ndufa3	NADH dehydrogenase (ubiquinone) 1 alpha subcomplex, 3	AV104160	−2.149	2.40
Mrpl15	Mitochondrial ribosomal protein L15	AV133566	−2.081	2.40
Mas1	MAS1 oncogene	AV028487	−2.911	2.40
Gsn	Gelsolin	AV006010	−3.244	2.39
Fusip1	FUS-interacting protein 1	AV057448	−2.949	2.38
Extl3	Exostoses (multiple)-like 3	AI840748	−2.089	2.37
Vamp8	Vesicle-associated membrane protein 8	AV053401	−2.194	2.37
Gtpbp2	GTP binding protein 2	BG071130	−2.897	2.37
Asnsd1	Asparagine synthetase domain containing 1	AV134139	−2.278	2.36
BC038328	Zinc finger protein 708	AV133851	−2.237	2.36
Vdac1	Voltage-dependent anion channel 1	AV037195	−2.069	2.36
Arhgap12	Rho GTPase activating protein 12	AV083597	−2.178	2.35
Atp5b	ATP synthase, H+ transporting mitochondrial F1 complex, beta subunit	AV005999	−2.755	2.35
Ndufs3	NADH-ubiquinone oxidoreductase Fe-S proetin 3	AV066283	−2.025	2.34
Hspd1	Heat shock 60-KD protein 1	BG065342	−2.757	2.33
Arl8b	ADP-ribosylation factor-like 8B	AV041040	−2.236	2.33
Kcnv1	Potassium channel, subfamily V, member 1	AV088564	−2.224	2.32
Anxa6	Annexin A6	AV094561	−3.076	2.32
Cacna1e	Calcium channel, voltage-dependent, alpha-1E subunit	AA855859	−4.131	2.31
Ggnbp2	Gametogenetin binding protein 2	AI841660	−2.008	2.31
Slc25a22	Solute carrier family 25 (mitochondrial carrier, glutamate), member 22	AV030675	−4.129	2.29
Atp5c1	ATP synthase, H+ transporting, mitochondrial F1 complex, gamma polypeptide 1	BG074111	−2.175	2.29
Hdlbp	High density lipoprotein (HDL) binding protein	AV012382	−4.346	2.29
Hsp110	Heat shock protein 110	AV031906	−2.049	2.28
Uqcrq	Ubiquinol-cytochrome c reductase, complex III subunit VII	AV006451	−2.836	2.27
Hod	Homeobox only domain	AV068725	−2.289	2.27
Mrpl37	Mitochondrial ribosomal protein L37	AV087163	−3.247	2.27
Amotl1	Angiomotin-like 1	AV085162	−2.246	2.26
Reep5	Receptor accessory protein 5	AV006204	−4.009	2.26
Pcmt1	Protein-L-isoaspartate (d-aspartate) O-methyltransferase 1	BG073964	−2.019	2.25
Vps16	Vacuolar protein sorting 16	BG072913	−2.276	2.25
Pygm	Muscle glycogen phosphorylase	AV006273	−2.789	2.25
Adh7	Alcohol dehydrogenase 7 (class IV), mu or sigma polypeptide	AV087904	−2.085	2.25
Eno3	Enolase 3, beta muscle	AV006198	−4.186	2.25
Ndufb9	NADH-ubiquinone oxidoreductase 1 beta subcomplex, 9	AV086909	−3.681	2.24
Tssk2	Testis-specific serine kinase 2	AV042468	−2.455	2.23
Vdac3	Voltage-dependent anion channel 3	AV104389	−3.656	2.23
Idh3b	Isocitrate dehydrogenase 3 (NAD+) beta	AV170829	−5.063	2.22
Rapgef4	Rap guanine nucleotide exchange factor (GEF) 4	AA388005	−3.119	2.21
Hadhsc	Hydroxyacyl-Coenzyme A dehydrogenase	BG074015	−2.336	2.21
Pygm	Muscle glycogen phosphorylase	AI324008	−3.632	2.21
Ghr	Growth hormone receptor	BG072812	−3.885	2.21
Ech1	Enoyl-CoA hydratase, peroxisomal	AV088052	−4.343	2.21
Usp15	Ubiquitin specific peptidase 15	AV039076	−3.803	2.21
Agl	Amylo-1,6-glucosidase, 4-alpha-glucanotransferase	AV016572	−2.862	2.21
Impa1	Inositol (myo)-1(or 4)-monophosphatase 1	AA152733	−2.280	2.21
Atp5 g1	ATP synthase, H+ transporting, mitochondrial F0 complex, subunit c (subunit 9), isoform 1	BG073112	−2.273	2.20
Rbpsuh	Recombination signal-binding protein suppressor of hairless, drosophila, homolog of	AV094997	−3.744	2.20
Hadh2	Hydroxyacyl-Coenzyme A dehydrogenase, type II	AV077923	−2.823	2.20
Aqp1	Aquaporin 1	AI838965	−3.040	2.20
Rpe	Ribulose-5-phosphate-3-epimerase	AV085827	−2.305	2.19
Mtap7	Microtubule-associated protein 7	AV041269	−2.660	2.19
Slc25a4	Solute carrier family 25 (mitochondrial carrier, adenine nucleotide translocator), member 4	AI841357	−7.234	2.19
Tagln2	Transgelin 2	AV084804	−2.140	2.18
Ptov1	Prostate tumor over expressed gene 1	BG073526	−2.280	2.17
Hadhb	Hydroxyacyl-Coenzyme A dehydrogenase/3-ketoacyl-Coenzyme A thiolase/enoyl-Coenzyme A hydratase (trifunctional protein), beta subunit	AV088068	−3.908	2.17
Mllt4	Myeloid/lymphoid or mixed lineage leukemia, translocated to, 4	AV010160	−3.132	2.17
5Arc	Activity regulated cytoskeletal-associated protein	BG519376	−3.426	2.16
Park7	Parkinson disease (autosomal recessive, early onset) 7	BG075688	−2.226	2.16
B3 gat3	Beta-1,3-glucuronyltransferase 3 (glucuronosyltransferase I)	AV085924	−2.795	2.16
Il11ra1	Interleukin 11 receptor, alpha chain 1	AV083410	−2.887	2.16
Rab14	RAB14, member RAS oncogene family	AV055902	−7.178	2.16
Ldhb	Lactate dehydrogenase B	AV006303	−2.436	2.16
Spop	Speckle-type poz protein	BG072804	−5.541	2.16
Cox7c	Cytochrome c oxidase, subunit VIIc	BG063960	−2.744	2.15
NGFB	Nerve growth factor, beta polypeptide	W46522	−2.071	2.15
Capza3	Capping protein (actin filament) muscle Z-line, alpha 3	AV039134	−3.081	2.14
Ctsf	Cathepsin F	AV085152	−1.999	2.14
Atg7	Autophagy-related 7	AW539206	−2.411	2.14
Sbk1	SH3-binding kinase 1	BG063893	−5.401	2.13
Nexn	Nexilin	BG072510	−3.305	2.13
Atp5a1	ATP synthase, H+ transporting, mitochondrial F1 complex, alpha subunit, isoform 1	AI841258	−2.089	2.13
Cox7c	Cytochrome c oxidase, subunit VIIc	AV015972	−4.189	2.13
Ide	Insulin degrading enzyme	AV006006	−2.321	2.12
Orc5l	Origin recognition complex, subunit 5	AV133637	−4.868	2.12
Nol7	Nucleolar protein 7	BG066700	−2.335	2.12
Plekhg1	Pleckstrin homology domain containing, family G (with RhoGef domain) member 1	AV084852	−3.538	2.12
Vldlr	Very low density lipoprotein receptor	AV018220	−4.117	2.12
Fasr	Fas receptor	AV134967	−2.567	2.12
Hspa4	Heat shock protein 4	BG065493	−2.672	2.11
Lpl	Lipoprotein lipase	AA049917	−2.519	2.11
Cops7a	COP9 (constitutive photomorphogenic) homolog, subunit 7a (Arabidopsis thaliana)	AV087105	−3.114	2.11
Acadvl	Acyl-Coenzyme A dehydrogenase, very long chain	AI839605	−2.388	2.10
Camk4	Calcium/calmodulin-dependent protein kinase IV	AV028684	−2.577	2.10
Atp5 g1	ATP synthase, H+ transporting, mitochondrial F0 complex, subunit c (subunit 9), isoform 1	AV112952	−3.004	2.10
Pex19	Peroxisome biogenesis factor 19	BG072565	−2.160	2.10
Chpt1	Synaptonemal complex protein 3	AV062575	−6.037	2.10
Mgst3	Glutathione S-transferase, microsomal, 3	AV056432	−2.480	2.10
Akt2	Thymoma viral proto-oncogene 2	BG073895	−2.055	2.10
Fdft1	Farnesyldiphosphate farnesyltransferase 1	AV095240	−2.501	2.09
Sdhc	Succinate dehydrogenase complex, subunit C, integral membrane protein, 15-KD	AV103143	−4.511	2.09
Raly	HnRNP-associated with lethal yellow	AV171686	−2.495	2.09
Pink1	PTEN-induced putative kinase 1	AV087434	−2.613	2.09
Etfdh	Electron transfer flavoprotein dehydrogenase	AV087820	−2.759	2.09
Mrpl36	Mitochondrial ribosomal protein L36	AV061668	−2.173	2.09
Bzw2	Basic leucine zipper and W2 domains 2	BG064378	−2.187	2.08
Aga	Aspartylglucosaminidase	AV056679	−2.161	2.08
Ldhb	Lactate dehydrogenase B	AV032980	−2.810	2.08
Tcof1	Treacher Collins Franceschetti syndrome 1, homolog	BG064056	−3.181	2.08
Mrpl45	Mitochondrial ribosomal protein L45	BG074455	−2.815	2.08
Upf3b	UPF3 regulator of nonsense transcripts homolog B	BG066789	−2.032	2.07
Sdha	Succinate dehydrogenase complex, subunit A, flavoprotein (Fp)	AV087966	−3.372	2.07
Lzts2	Leucine zipper, putative tumor suppressor 2	BG063967	−3.201	2.07
Atp5f1	ATP synthase, H+ transporting, mitochondrial F0 complex, subunit b, isoform 1	AI836064	−6.161	2.07
Seh1l	SEH1-like (S. cerevisiae)	AV025108	−2.554	2.07
Ebpl	Emopamil binding protein-like	AV094562	−2.563	2.06
Ptp4a3	Protein tyrosine phosphatase, type 4A, 3	AV087012	−2.477	2.06
Aptx	Aprataxin	AV085680	−2.627	2.06
Mdh1	Malate dehydrogenase 1, NAD (soluble)	BG067903	−2.618	2.06
Mapk14	Mitogen-activated protein kinase 14	BG074842	−2.419	2.06
Ncf1	Neutrophil cytosolic factor 1	AV074152	−2.431	2
Parn	Poly(A)-specific ribonuclease (deadenylation nuclease)	AA408467	−2.817	2.06
Smurf1	SMAD specific E3 ubiquitin protein ligase 1	BG076120	−2.769	2.05
Rgs2	Regulator of G-protein signaling 2	BG067321	−3.334	2.05
Pnn	Pinin	AV135835	−2.050	2.05
Cacnb2	Calcium channel, voltage-dependent, beta 2 subunit	BG076402	−2.234	2.05
Cs	Citrate synthase, mitochondrial	AV006265	−3.596	2.05
Stat1	Signal transducer and activator of transcription 1	AA170538	−2.378	2.04
Tgfb1	Transforming growth factor, beta 1	AA049522	−2.692	2.04
Mfng	MFNG O-fucosylpeptide 3-beta-N-acetylglucosaminyltransferase	AA116377	−2.013	2.04
Hdlbp	High density lipoprotein (HDL) binding protein	BG064469	−2.713	2.04
Gstm1	Glutathione S-transferase, MU-1	AV149857	−2.256	2.04
Pik3r4	Phosphatidylinositol 3 kinase, regulatory subunit, polypeptide 4, p150	AV074828	−2.566	2.04
Abl1	V-abl Abelson murine leukemia oncogene 1	BG064435	−2.074	2.04
Plrg1	Pleiotropic regulator 1, PRL1 homolog (Arabidopsis)	C80566	−3.591	2.03
Abcd3	ATP-binding cassette, sub-family D (ALD), member 3	AV140550	−2.499	2.03
Drg2	Developmentally regulated GTP binding protein 2	AV084970	−2.727	2.03
Chmp2a	Chromatin modifying protein 2A	BG075465	−2.278	2.02
Ccbl2	Cysteine conjugate-beta lyase 2	AV103749	−2.449	2.02
Hspd1	Chaperonin	BB610862	−2.971	2.02
Farslb	Phenylalanyl-tRNA synthetase, beta subunit	BG064281	−2.473	2.02
Etfdh	Electron transferring flavoprotein, dehydrogenase	BG070637	−2.472	2.02
Farp2	FERM, RhoGEF and pleckstrin domain protein 2	AA146115	−2.968	2.01
Hnrpa1	Heterogeneous nuclear ribonucleoprotein A1	BG072533	−2.335	2.01
Mbnl2	Muscleblind-like 2	BG072107	−2.174	2.01
Hbld2	Iron-sulfur cluster assembly 1,	AI851055	−2.122	2.01
Smpd2	Sphingomyelin phosphodiesterase 2, neutral	AV005649	−2.332	2.01
Fyn	Fyn proto-oncogene	AV051790	−4.000	2.01
Abcb7	ATP-binding cassette, subfamily B, member 7	BG074307	−2.087	2.01
Ndufb10	NADH dehydrogenase (ubiquinone) 1 beta subcomplex, 10	AI836006	−4.215	2.01
Ndufs6	NADH-ubiquinone oxidoreductase Fe-S proetin 6	AV071159	−2.816	2.01
Alad	Aminolevulinate, delta-, dehydratase	BG063937	−2.467	2.00
Aco2	Aconitase, mitochondrial	AV085145	−2.255	2.00
Stx17	Syntaxin 17	AV015430	−2.102	2.00
Gabarapl1	Gamma-aminobutyric acid (GABA(A)) receptor-associated protein-like 1	AV150053	−2.393	2.00

The SAM algorithm was employed to identify genes with statistically different expression levels between PA-banded and sham tissues. SAM incorporates an FDR correction for multiple testing errors and calculates a d statistic for each gene based on the ratio of change in gene expression to SD in the data for that gene (Ref. 18).

^*Out of a total of 1,114 significantly downregulated genes and ESTs.

Table 5.

Gene ontology category transcripts upregulated in severe PS RV versus sham RV

Gene Ontology Category	Total Genes	Changed Genes	P Value (Log10)
Phosphate transport	38	7	−4.78
Negative regulation of coagulation	5	3	−3.97
Inorganic anion transport	51	7	−3.92
Regulation of coagulation	6	3	−3.68
Cell adhesion	239	15	−3.63
Anion transport	64	7	−3.30
Skeletal development	84	7	−2.59
Organismal physiological process	634	25	−2.58
Neuromuscular physiological process	4	2	−2.54
Ion transport	224	12	−2.40
Tyrosine kinase signaling pathway	91	7	−2.40
Coagulation	32	4	−2.28
Negative regulation of Wnt receptor signaling pathway	6	2	−2.15
Negative regulation of apoptosis	81	6	−2.04
Negative regulation of programmed cell death	81	6	−2.04
Sensory perception	107	7	−2.01
Organismal movement	7	2	−2.01
Histone acetylation	8	2	−1.89
Regulation of apoptosis	199	10	−1.89
Apoptosis	261	12	−1.88
Regulation of programmed cell death	200	10	−1.87
Death	300	13	−1.81
Ionic insulation of neurons by glial cells	9	2	−1.79
Myelination	9	2	−1.79
Programmed cell death	269	12	−1.79
Regulation of physiological process	1,366	41	−1.79
Smooth muscle contraction	10	2	−1.70
Adult locomotory behavior	10	2	−1.70
Negative regulation of histone acetylation	1	1	−1.65
Regulation of Wnt receptor signaling pathway	11	2	−1.62
Enzyme-linked receptor protein signaling pathway	130	7	−1.59
Cell surface receptor-linked signal transduction	397	15	−1.52
Calcium ion homeostasis	32	3	−1.47

Some interesting genes related to cardiac function were increased in severe PS versus sham RV. For each of these genes, altered expression was verified by QRT-PCR (Fig. 8). Periostin (16.0 fold) was the most dramatically increased in the PAC RV. This gene is highly expressed in the myocardium in patients with heart failure (30). Periostin expression is positively regulated by transforming growth factor (TGF)-β (27), and Ingenuity Pathway Assist gene pathway analysis demonstrated a pattern of both the up- and downregulation of multiple transcripts in the TGF-β signaling pathway, which could be the subject of future study once verified by PCR. Other transcripts with marked upregulation in the PAC RV included lysyl oxidase (LOX; 7.7 fold), microfibrillar-associated protein 4 (7.5 fold), and secreted acidic cysteine rich glycoprotein (3.5 fold), related to ECM proteins and their cross-linking enzyme.

Fig. 8.

Comparison of RT-PCR vs. array for 10 differentially expressed genes in PAC RV vs. sham RV. For each of these genes, there was excellent agreement between both methods.

We next compared gene expression data from mice after PAC with data from our previous study of mice after 10 days of LV afterload stress using TAC (73). Many genes were upregulated in both the PAC RV and TAC LV, since many of the same processes of matrix remodeling, metabolic changes, and actin cytoskeletal alterations must necessarily occur in both models of afterload stress. However, there were several transcripts that showed significant differential regulation in the afterload stressed RV but not in the LV (Table 6). Those which were upregulated in the PAC RV but not in the TAC LV included three from the upregulated GO biological process Wnt signaling: Dickkopf 3, Sfrp2, and Wif1. Other important RV-specific upregulated genes include annexin A7, clusterin/apolipoprotein J, neuroblastoma suppression of tumorigenicity 1 (Nbl1), formin binding protein (Fnbp4), and LOX.

Table 6.

Transcripts that showed significant upregulation in the afterload-stressed RV but not in the afterload-stressed LV

Gene	Function	Fold Change
Fnbp4	Regulated by p53 and involved in apoptosis	11.6
Uchl1	Involved in the processing of ubiquitin precursors	7.2
Dkk3	Inhibitor of Wnt signaling pathway	4.2
Nbl1	TGF-β antagonist that modulates BMP2 signaling	2.8
Tacstd1	Ga733 tumor-associated antigen gene family	2.8
Clusterin/apolipoprotein J	Involved in apoptosis	2.5
Annexin A7	Calcium/phospholipid-binding protein	3
Wif1	Inhibitor of Wnt signaling pathway	1.6
Sfrp2	Inhibitor of Wnt signaling pathway	7.5
LOX	Regulates cell migration and actin polymerization	7.7

DISCUSSION

RV dysfunction is a common cause of long-term morbidity in many patients with congenital heart defects, especially those producing RV pressure overload (RVOT obstructions such as tetralogy of Fallot) and those with systemic right ventricles (e.g., hypoplastic left heart syndrome) (21, 41, 42). For many of these patients, despite incredible progress in surgical repair or palliation, long-term survival and quality of life will depend on our ability to preserve long-term RV function. Although there are considerable data on the molecular events underlying afterload-induced LV remodeling, there is little information on molecular events in the afterloaded RV. The complex geometry of the RV and differences in physiology with respect to the LV, however, make RV failure difficult to assess both clinically and in the laboratory. The response of the RV to increased afterload consists of both hypertrophy and enlargement, dilation of the tricuspid valve annulus leading to tricuspid regurgitation, and leftward displacement of IVS, leading to alterations in LV diastolic function, which can be significantly compromised by deleterious ventricular-ventricular interaction.

We have adapted a previously described model of RV afterload stress (51), with updated and detailed physiological and genomic characterization. This model shows many characteristics of both acute and chronic RV failure encountered in clinical settings: RV dilation, right bundle branch block, tricuspid regurgitation, leftward septal shift, right-sided heart failure, and decreased survival. Although there are a few previous reports of PAC in mice and rats (8, 20, 51, 53, 64), none has characterized the physiological response as clearly. We analyzed our PAC model by dividing subjects into three groups based on the degree of pressure gradient between the RV and PA and the presence or absence of IVS shift. Our results show that mice with severe PS have significantly decreased survival compared with those with mild PS, with the moderate group having a survival intermediate between the other two groups. These data suggest that our mild and moderate PS groups represent a reasonable model of chronic RV dysfunction and that our severe group represents a good model of acute decompensated RV function. This reconstructed natural history is also similar to those models of LV outflow tract obstruction (TAC) associated with LV failure (37). Although it could be argued that mice in our severe group would not meet the clinical criteria used in humans to qualify for severe PS based on outflow tract gradient alone, mice in this group did have evidence of RV failure, including significant RV dilation, decreased RV wall motion, elevation of RVEDP, decreased CO, clinical evidence of right-sided heart failure, and decreased survival. In these animals, the lack of a higher RV-PA gradient may reflect the lower CO in the severe PS group, as well as the temporal difference between an artificial animal model (with the acute onset of RV outflow obstruction) versus that in humans with congenital heart disease (with chronic RV outflow obstruction allowing the gradual development of RV hypertrophy and chronic adaptation).

In all groups of mice with PS, the degree of PS correlated well with RV weight-to-BW ratio and to histological evidence of cardiomyocyte hypertrophy. Previous reports in PA-banded rats (8, 20, 53) describe the doubling of RV weight after banding for 3 wk (8), although no histological data were presented. There are two previous reports describing the technique of PAC in mice (51, 64); however, these reports provide minimal details concerning physiological variables.

An accurate estimation of RV diastolic dimension by ultrasound is difficult not only in mice but in humans as well. We used RVOT dimension and RVOT FS to quantify RV dilation and function (43). These data show that RV function is preserved in mild and moderate PS and significantly decreased in severe PS. All mice with severe PS had evidence of tricuspid regurgitation, which also differentiated them from the mild and moderate groups. Tricuspid regurgitation is used clinically as a sign of RV dilation and dysfunction in humans with congenital heart disease and RV outflow obstruction (42, 50). Increased RV size, whether due to increased preload (atrial septal defect) or afterload (sleep apnea, primary pulmonary hypertension, pulmonary embolus), leads to iRBBB in humans. Mice with severe PS showed iRBBB, also differentiating them from mice with mild or moderate PS. Other signs of right heart failure in the severe PS group included increases in liver weight to BW and the development of peripheral edema.

The clinical assessment of RV failure has been based primarily on echocardiography, although magnetic resonance angiography is playing an increasingly important role (31). Recent human studies show that TDI holds promise for providing an accurate assessment of RV function, e.g., peak systolic velocity at the basal tricuspid annulus (9, 26, 39). In this study, we demonstrate that RV wall velocity can be obtained noninvasively in mice using TDI and that TDI-derived RV wall velocity correlates well with dP/dt_max obtained from invasive hemodynamic studies. This suggests that TDI can be useful in the serial evaluation of murine models or RV failure, in which the chronic micromanometer catheterization of the RV is not feasible. The relationship between LV wall velocity and dP/dt_max in mice has already been established and shows a similarly strong correlation (58).

One of the major problems in RV failure is the negative ventricular-ventricular interaction mediated in part through septal shift, encroaching into the LV cavity, and impairing LV diastolic function (6, 28, 69). Our model of severe PS accurately recapitulates this process, with elevation of RVEDP, shift of the septum into the LV, and decreased CO. Using echocardiography, we were able to quantify the degree of intra-LV dyssynchrony by using TDI to measure the offset between septal and LV free wall contraction. Also interventricular dyssynchrony could be measured from the time difference of peak systolic velocity at the base of the tricuspid and mitral valves.

There have been multiple previous studies of increased RV afterload in many mammalian models, including cats (15, 47), dogs (22, 45), pigs (4, 16), rabbits (5, 19, 23), and rats (8, 11, 20, 33, 36, 53). The advantages of a well-characterized murine model, particularly in the ability to alter murine gene expression, are well known. Rockman et al. (51) published a similar model of PAC in the mouse, although the degree of physiological characterization was less and the examination of gene expression changes was limited to a standard panel of heart failure genes.

There have been few prior studies of gene expression changes in the stressed RV. Several have examined RV gene expression associated with pulmonary hypertension (11, 33). In rats with monocrotaline-induced pulmonary hypertension, Buermans et al. (11) compared gene expression by microarray in compensated or decompensated RV hypertrophy. Ventricles destined to progress to failure showed an activation of proapoptotic pathways, particularly related to mitochondria, whereas the group with compensated hypertrophy showed blocked pro-death effector signaling via p38-MAPK, through the upregulation of MAPK phosphatase-1. Kogler et al. (33) also using a monocrotoline model, studied the altered regulation of calcium regulatory genes, finding a downregulation of sarco(endo)plasmic reticulum Ca²⁺-ATPase (SERCA)2a, phospholamban (PLB), and the ryanodine receptor. LekanneDeprez et al. (36) used a PA banding model in the rat to investigate alterations in the expression of selected genes during the transition from compensated RV hypertrophy to failure. PA banding resulted in an induction of atrial natriuretic peptide, a moderate increase in collagen III α1, and a decrease in SERCA2 and PLB (36). However, in contrast to the current study, these authors did not utilize a genome-wide approach.

We found that in the afterload stressed RV, 196 genes were upregulated and 1,114 genes were downregulated compared with sham-operated controls. The importance of comparing surgical subjects with sham controls has been well established. Our laboratory has previously described the marked and prolonged gene expression changes that can be introduced even by sham operation (73).

The RV and LV share many common pathways, which are either up- or downregulated during the development of pressure overload, since many of the same processes of matrix remodeling, metabolic changes, and actin cytoskeletal alterations must necessarily occur in both. GO analysis, to identify significantly enriched and depleted groups, showed the most significantly upregulated processes were in the categories of phosphate transport, regulation of coagulation, inorganic anion transport, and cell adhesion. Downregulated processes were dominated by energy pathways. Periostin, the transcript with the greatest fold change (16.0 fold) in the PAC RV, is also increased in the TAC LV. Periostin is a secreted protein involved in bone formation and cell-cell adhesion. It is upregulated in the Syrian hamster model of heart failure (69) and expressed in cardiac fibroblasts and implicated in cardiac dysfunction (38). The overexpression of periostin causes LV dilation and an increase in collagen deposition (30). In contrast, the inhibition of periostin improves LV function in Dahl salt-sensitive rats (30). Periostin knockout mice also show less fibrosis and hypertrophy after TAC or myocardial infarction (44). LOX (upregulated 7.7 fold), microfibrillar-associated protein 4 (7.5 fold), and secreted acidic cysteine rich glycoprotein (3.5 fold) are additional ECM proteins that may contribute to progressive ventricular remodeling, dilation, and heart failure (48).

Despite these commonly expressed gene programs, there are several important differences between the afterload stressed RV and LV. Several transcripts showed a significant differential upregulation in the RV but not in the LV, including three from the Wnt signaling pathway (12, 14), recognized as one of the major families of developmentally regulated signaling molecules. These included the Wnt inhibitors Dickkopf 3 (4.2 fold), Sfrp2 (7.5 fold), and Wif1 (1.6 fold). Other important RV-specific upregulated genes include annexin A7 (3 fold), which has been associated with the regulation of calcium handling in cardiomyoctes (13), and clusterin/apolipoprotein J (2.5 fold), which has a nearly ubiquitous expression pattern in human tissues. Clusterin is differentially regulated in many severe physiological disturbances including cell death, aging, cancer progression, and various neurological diseases (65). Clusterin protects cardiomyocytes against ischemic cell death via a complement independent pathway (35); however, the function of clusterin remains an enigma, due to its intriguingly distinct and often opposite functions in different cell types.

Several additional genes were differentially regulated in the RV versus LV. Nbl1 is upregulated in the PAC RV (2.8 fold) but slightly downregulated in the TAC LV. Nbl1 is a TGF-β antagonist that modulates bone morphogenetic protein (BMP)2 signaling and prevents cells from entering the G1/S phase of the cell cycle. In LV hypertrophy, the activation of the first part of the G1 phase (cyclin D and E) occurs but without progression to the S phase. Nbl1 null mice have no gross phenotype unless crossbred with a Noggin (BMP antagonist) heterozygote, where there are skeletal but no known cardiac defects (49). The Fnbp4/formin binding protein was upregulated 11.6 fold in the RV compared with only twofold in the TAC LV. Formin is an actin regulator that plays a role in limb morphogenesis. In thymocytes, Fnbp4 is regulated by p53 and involved in cell death pathways. LOX/Lysyl oxidase (5.4 fold) regulates cell migration and actin polymerization through the focal adhesion kinase/Src signaling complex. LOX is also involved in the cross-linking of fibrillar collagens and elastin and is upregulated in myocardial ischemia (66).

There are several limitations in our methodology, some of which hold true for any small animal model of human disease. Although murine studies are a powerful platform by which to study genetic alterations in cardiac disease, substantial species differences must be acknowledged and findings confirmed in larger mammalian models. We chose a model of PA banding because residual RVOT obstruction is one of the major long-term sequelae in patients with repaired congenital heart disease, including tetralogy of Fallot, corrected transposition, and other complex congenital heart lesions where the RV is at risk. Other models, including pulmonary hypertension, add additional variables (altered pulmonary vascular pathology) aside from increased RV afterload, which could result in different gene expression patterns in the stressed RV. Other experimental manipulations required to produce pulmonary hypertension in rodents (e.g., hypoxia) could also have independent effects on the RV myocardium.

There are other limitations inherent to any microarray-based study. These transcriptional profiles provide a snapshot look at gene expression changes, which will vary with time after banding and with a multitude of other physiological changes. They also do not account for the myriad of posttranscriptional changes, which for many signaling pathways involved in hypertrophy and heart failure are the dominant regulators. Gene microarrays are also less sensitive in detecting small changes in gene expression that may still be biologically meaningful. We have used a strategy designed to minimize the number of false negative calls by using four replicate arrays per condition, but it is inevitable that some genes with small but meaningful changes in gene expression will not be identified using our stringent statistical analyses. Our laboratory has also shown previously that the accuracy of gene microarrays is more limited for genes with very low expression levels (73). Finally, there are limitations to our study in our comparison of gene expression in the RV with the LV. Although we obtained samples at similar time points after PAC and TAC, it is not possible to totally equalize the degrees of afterload stress between the two ventricles, given that the LV normally is exposed to systemic arterial pressure and the RV to the considerably lower PA pressure. Further investigation of RV-LV differential gene expression is thus warranted, with multiple additional time points as well as varied degrees of afterload stress, well beyond the scope of the present methodological study. Ultimately, one or more of these genes may provide utility as a biomarker for the early prediction of clinical outcome as well as a potential therapeutic target for patients with RV failure. A physiologically well-characterized murine model of PS should provide a valuable platform for these studies.

GRANTS

This work was supported in part by National Heart, Lung, and Blood Institute Grant HL-061535 (to D. Bernstein).