This study utilized a transgenic mouse model that expresses an inducible dominant-negative mutation of the transforming growth factor (TGF)-β type II receptor (DnTGFβRII) to define the structural and functional responses of the left ventricle (LV) to pressure-overload stress in the absence of an intact TGF-β signaling cascade. DnTGFβRII and nontransgenic (NTG) control mice (male, 8–10 wk) were randomized to receive Zn2+ (25 mM ZnSO4 in drinking H2O to induce DnTGFβRII gene expression) or control tap H2O and then further randomized to undergo transverse aortic constriction (TAC) or sham surgery. At 7 days post-TAC, interstitial nonmyocyte proliferation (Ki67 staining) was greatly reduced in LV of DnTGFβRII+Zn2+ mice compared with the other TAC groups. At 28 and 120 days post-TAC, collagen deposition (picrosirius-red staining) in LV was attenuated in DnTGFβRII+Zn2+ mice compared with the other TAC groups. LV end systolic diameter and end systolic and end diastolic volumes were markedly increased, while ejection fraction and fractional shortening were significantly decreased in TAC-DnTGFβRII+Zn2+ mice compared with the other groups at 120 days post-TAC. These data indicate that interruption of TGF-β signaling attenuates pressure-overload-induced interstitial nonmyocyte proliferation and collagen deposition and promotes LV dilation and dysfunction in the pressure-overloaded heart, thus creating a novel model of dilated cardiomyopathy.
- cardiac hypertrophy
- Smad protein
- extracellular matrix
pressure-overload stress (e.g., hypertension or aortic stenosis) results in excessive cardiac fibrosis, changes in left ventricular (LV) geometry, and disruption in LV contractility, ultimately leading to heart failure (16, 3). Our previous studies (20) have shown that transforming growth factor (TGF)-β plays an important role in pressure-overload-induced cardiac hypertrophy, remodeling, and fibrosis. TGF-β expression is upregulated in the LV of mice subjected to pressure-overload stress induced by transverse aortic constriction (TAC; Ref. 36). The pressure-overload-induced increase in TGF-β is accompanied by LV hypertrophy, fibrosis, and remodeling that is exaggerated in the absence of antifibrotic and growth-inhibiting effects mediated by the atrial natriuretic peptide (ANP) signaling pathway. Thus the ANP null mouse (Nppa−1/−1) exhibits a profibrogenic phenotype that appears to be related to unopposed TGF-β signaling (36, 20, 12).
TGF-β1, the major isoform in the heart (35, 6), stimulates cardiac fibroblast transformation (to myofibroblasts) and proliferation, as well as extracellular matrix (ECM) production (22, 28). Mice overexpressing TGF-β1 have cardiac hypertrophy and interstitial fibrosis (29), and TGF-β1 expression is increased in hearts subjected to pressure overload (35, 33, 6, 20). TGF-β signals through the membrane bound heterotrimeric TGF-β receptors type I and II (TGFβRI and TGFβRII). When the TGFβRI and TGFβRII are activated, downstream signaling molecules Smad2 and Smad3 are phosphorylated on the C-terminal serine residues. Phosphorylated Smad2 and Smad3 (pSmad2 and Smad3) bind to Smad4 and translocate to the nucleus (32). The Smad complex then binds to response elements in the promoter regions of the ECM genes and activates profibrogenic factors by upregulating gene transcription.
Our present study utilized a novel mouse model that expresses an inducible dominant negative mutation of the TGF-β receptor type II gene (DnTGFβRII), and thus cannot activate the TGF-β/Smad signaling cascade (31), to characterize the effect of blocking TGF-β signaling on the phenotype of the pressure-overloaded LV. We observed that induction of DnTGFβRII expression attenuates pressure-overload-induced interstitial nonmyocyte proliferation and collagen deposition and promotes LV dilation and dysfunction in the mouse, resulting in a novel model of dilated cardiomyopathy.
Transgenic mice and animal preparation.
Male DnTGFβRII transgenic mice originally generated by Dr. Rosa Serra (31) and nontransgenic (NTG) C57/BL6 mice were studied. All mice were raised in our resident colony, which was founded with pathogen-free breeding pairs. Genotypes were identified by PCR assay of genomic DNA from tail snips after weaning (2). The DnTGFβRII mouse expresses a cytoplasmically truncated TGFβRII receptor that competes with endogenous receptors for heterodimeric complex (TGFβRI and TGFβRII) formation and is thus a dominant-negative mutant (31). The DnTGFβRII lacks the cytoplasmic kinase domain and has no intrinsic activity. Overexpression of DnTGFβRII is under the control of a metallothionein-derived promoter that can be induced by Zn2+.
Expression of DnTGFβRII was induced by giving 25 mM ZnSO4 in the drinking water to DnTGFβRII mice (DnTGFβRII+Zn2+) beginning 1 wk before TAC or sham surgery (36) and continuing throughout the study. NTG mice drinking either ZnSO4 water (NTG+Zn2+) or double-distilled H2O (NTG+H2O) served as controls. DnTGFβRII mice drinking double-distilled H2O (DnTGFβRII+H2O) with TAC or sham surgery were used as additional controls. These eight groups of mice (Sham-NTG+H2O, TAC-NTG+H2O, Sham-NTG+Zn2+, TAC-NTG+Zn2+, Sham-DnTGFβRII+H2O, TAC-DnTGFβRII+H2O, Sham-DnTGFβRII+Zn2+, and TAC-DnTGFβRII+Zn2+) were fed a standard diet (Harlan-Teklad) and were housed in rooms maintained at a constant humidity (60 ± 5%), temperature (24 ± 1°C), and light cycle (6:00 AM-6:00 PM).
Induction of DnTGFβRII expression was confirmed by detecting the DnTGFβRII receptor mRNA in heart, lung, and liver of DnTGFβRII mice given ZnSO4 (or H2O as a negative control) using RT-PCR (primers: 5′-ATC GTC ATC GTC TTT GTA GTC-3′ and 5′-TCC CAC CGC ACG TTC AGA AG-3′) as described previously (2). All protocols were approved by the Institutional Animal Care and Use Committee of the University of Alabama at Birmingham and were consistent with the Public Health Service Policy on Humane Care and Use of Laboratory Animals (Office of Laboratory Animal Welfare, August 2002) and the Guide for the Care and Use of Laboratory Animals published by National Institutes of Health (NIH Publication No. 96-01, revised in 2002).
Male DnTGFβRII and NTG mice, 8–10 wk of age, were anesthetized with an intraperiotoneally administered mixture of ketamine (80 mg/kg) and xylazine (12 mg/kg), and TAC was performed as described previously (36). The aortic band was located between the proximal left carotid artery and the brachiocephalic arteries on the ascending aorta. Pressure gradients across the TAC were 50–60 mmHg, and similar among genotypes, as described previously (36). Sham-operated mice served as controls.
Separate groups of mice were killed at 7, 28, and 120 days after TAC with an overdose of ketamine/xylazine followed by cervical dislocation. Hearts, lungs, liver, and kidneys were quickly removed and the LV and right ventricle (RV) were dissected carefully and weighed. LV sections were divided into two portions: the apical portion was fixed with 4% paraformaldehyde, embedded in paraffin, and sectioned for histological analysis; the basal portion was immediately frozen in liquid N2 and stored at −80°C for biochemical analysis.
RT-PCR and Western blotting and analyses.
For RT-PCR, total RNA was prepared from snap-frozen tissue using TRIzol reagent (Invitrogen), treated with DNAase I to remove genomic DNA, and then purified using an RNA purification kit (Invitrogen) as described previously (19–21). The protein- and DNA-free RNA was reverse transcribed to cDNA using the SuperScriptIII First-Strand Synthesis System (Invitrogen). cDNA of was amplified by PCR using a Bio-Rad iCycler with specific primers for DnTGFβRII allele (5′-ATC-GTC-ATC-GTC-TTT-GTA-GTC-3′ and 5′-TCC-CAC-CGC-ACG-TTC-AGA-AG-3′).
For Western blot analysis, LVs were homogenized in a tissue protein extraction reagent (T-PER) described previously (19–21). Protein samples were separated by 10% SDS-PAGE and transferred to polyvinylidene difluoride membrane. Blots were probed with anti-FLAG or anti-phospho-Smad3 (pSmad3) primary antibodies and a horseradish peroxidase-conjugated secondary antibody (Pierce). Autoradiographs were quantitated by densitometry (NIH ImageJ). pSmad3 protein levels were normalized using α-tubulin protein levels as an internal standard.
At 7 days after TAC, cell proliferation and apoptosis were assessed in paraffin-embedded LV cross sections using Ki-67 and terminal deoxynucleotidyl transferase (TUNEL) reagent (Vector Laboratories), respectively, according to the manufacturer's instruction. Proliferative or apoptotic indices were determined by counting the number of Ki67- or TUNEL-stained nuclei in ×400 microscopic fields of posterior wall and septum of each LV. The identity of the samples was masked to the two examiners to avoid bias. Twelve randomly selected high-power fields from each mouse were examined and quantitated.
LV cardiomyocyte area was measured in 28-day samples of LV as previously described (26). Morphometric analysis of each heart section was performed with a computer-based morphometric system (Motic Image Plus 2.0). Four hearts from each experimental group were included in the histological analysis. At least five hematoxylin-eosin stained cross sections of each heart were examined. Eighty cardiomyocytes from each LV were measured in nucleated transverse sections, and the average area of the 80 myocytes was calculated. All morphometric analyses were carried out by a single examiner, who was blinded with respect to the experimental group to which each sample belonged.
At 28 and 120 days after TAC, LV interstitial collagen volume percentage, at the level below the mitral valve, was measured in picrosirius red (0.1%)-stained cross sections as described previously (36, 13, 12), using a microscopic system with a green (540 nm) filter to enhance contrast for computer imaging analysis. Quantitative morphometric analysis of collagen content was carried out by light microscopy with a Qimaging QiCam digital camera (Qimaging) interfaced with a computer system running Metamorph 6.2v4 software (Universal Imaging). At least 12 randomly selected images (×400) from the posterior wall and septum of each LV were analyzed. The identity of the samples was masked to the two examiners to avoid bias.
At 120 days after TAC or sham surgery, echocardiography was performed on isofluorane-anesthetized mice with a 15-MHz transducer (Phillips) and a commercially available ultrasound system (Phillips Sonos 5500) as described previously (11, 26, 13, 12). LV end-diastolic dimension (LVEDD), LV end-systolic dimension (LVESD), and septal [intraventricular septal (IVS)] and posterior wall (PW) diastolic thickness were measured by two-dimensional-guided M-mode echocardiography from the parasternal long-axis view. LV end-diastolic volume (EDV), end-systolic volume (ESV), fractional shortening (FS), and ejection fraction (EF), and cardiac output (CO) were calculated as EDV = 7 × LVEDD3/(2.4 + LVEDD), ESV= 7 × LVESD3/(2.4 + LVESD), FS= (LVEDD − LVESD)/LVEDD × 100, EF= (EDV − ESV)/EDV × 100, and CO = heart rate (HR) × (EDV − ESV)/1000, respectively (11). A single examiner, blinded to genotype and treatment, interpreted all studies.
Results were expressed as means ± SE. Analyses were carried out using the SigmaStat statistical package. Tissue weight and echocardiographic measurements were normalized by analysis of covariance with body weight as the covariate (Packard 1988). Our primary statistical test was ANOVA, one-way ANOVA to evaluate the differences in mean values due to main effects (genotype, TAC, or Zn2+), and two-way ANOVA to test their interactions. P < 0.05 was considered statistically significant.
Drinking 25 mm ZnSO4 water induced mutated DnTGFβRII gene expression and attenuated TGF-β/smad signaling in heart.
We confirmed that drinking Zn2+ water for 1 wk induced expression of mutated DnTGFβRII mRNA in tissues of DnTGFβRII+Zn2+ but not DnTGFβRII+H2O mice by RT-PCR analysis (Fig. 1A). We also detected using Western blot analysis the FLAG epitope protein linked to the mutated DnTGFβRII protein in LV of DnTGFβRII+Zn2+ but not DnTGFβRII+H2O mice (Fig. 1B). Further, we found that phospho-Smad3 (pSmad3) levels in LV of DnTGFβRII+Zn2+ mice were lower than in LV of DnTGFβRII+H2O mice 1 wk after TAC (Fig. 1C). These data provide strong evidence that there was interruption of the TGF-β/Smad signaling pathway in the DnTGFβRII mice given Zn2+ in drinking water.
TAC induced LV hypertrophy in all groups.
There were no significant differences in body weight among groups. There were significant increases in heart weight in both the genotypes treated with either ZnSO4 or H2O at all time points following TAC. Both genotypes demonstrated cardiac hypertrophy, as evidenced by increased total heart and individual chamber weights as early as 7 days after TAC (Table 1). Similiarly, cross-sectional areas of cardiomyocytes from LV of all genotype treatment groups were significantly greater at 28 days post-TAC compared with sham controls (Table 1). RV weights were unchanged following TAC in all groups. Survival rates were 80% in the first 5 days postsurgery and then 100% for the remainder of the study in all groups.
Normalized LV weight did not differ between genotypes after sham surgery and was increased in all mice following TAC. There were no significant differences in LV weight of DnTGFβRII+Zn2+ mice compared with the other three groups at any time point following TAC.
TAC-induced nonmyocyte proliferation in LV was attenuated in DnTGFβRII+Zn2+ mice.
Nonmyocyte proliferation assessed by Ki67 staining at 7 days post-TAC was increased in the LV posterior wall and septum of NTG+H2O, NTG+Zn2+, and DnTGFβRII+H2O mice (Figs. 2 and 3). In contrast, TAC-induced nonmyocyte proliferation in LV of DnTGFβRII+Zn2+ was dramatically reduced, indicating that blocking TGF-β signaling attenuated nonmyocyte proliferation. Sham-operated mice from all four genotype/treatment groups showed minimal nonmyocyte proliferation in posterior wall and septum of LV (Figs. 2 and 3).
Apoptosis was minimal in all groups.
TUNEL staining revealed very few apoptotic cells in any of the LV specimens from the eight experimental groups (data not shown), suggesting that TAC-induced pressure-overload stress did not increase apoptotic cell death in heart at 7 days after TAC.
TAC-induced increase in collagen deposition in LV was attenuated in DnTGFβRII+Zn2+ mice.
At 28 and 120 days post-TAC, NTG+H2O, NTG+Zn2+, and DnTGFβRII+H2O mice developed threefold increases in total collagen protein content in LV compared with their sham-operated controls (Figs. 4 and 5). This effect appeared to be maximal at 28 days. Interstitial collagen levels were not significantly changed from sham control levels in LV of DnTGFβRII+Zn2+ mice at either time point post-TAC. There was no significant increase in percent LV collagen volume in TAC-DnTGFβRII+Zn2+ at 28 or 120 days after TAC compared with their sham-operated controls (Fig. 5). These results indicate that TGF-β signaling is necessary for interstitial collagen deposition in response to pressure-overload stress.
TAC-induced LV dilation and dysfunction but not LV hypertrophy were exacerbated in DNTGFβRII+Zn2+ mice.
Echocardiographic examination of the four TAC groups and sham-treated NTG+H2O mice was carried out at 120 days post-TAC or sham surgery (Figs. 6 and 7). LVEDD and EDV were not changed in NTG mice but were increased in both groups of DnTGFβRII mice 120 days after TAC. The TAC-induced increases in LVEDD and EDV were significantly greater in DnTGFβRII+Zn2+ mice compared with their H2O controls. LVESD and ESV were significantly increased in all genotype/treatment groups after TAC. Similar to LVEDD and EDV, the TAC-induced increases in LVESD and ESV were significantly greater in DnTGFβRII+Zn2+ mice. EF and FS did not differ between genotypes after sham surgery and were reduced in both genotypes after TAC. The TAC-induced decreases in EF and FS were significantly exacerbated by ZnSO4 treatment in DnTGFβRII mice. Cardiac output and heart rate were not different among experimental groups.
In the present study, we have utilized a novel DnTGFBRII mouse model of inducible inhibition of the TGF-β/Smad signaling cascade to define the contribution of TGF-β signaling to the phenotype of cardiac remodeling and fibrosis in response to pressure overload. The DnTGFβRII mouse model offers important advantages for studying the contribution of TGF-β signaling to pressure-overload-induced cardiac hypertrophy and remodeling. TGF-β signaling is essential for the embryonic development of the heart, and homozygous deletion of the TGF-β gene leads to embryonic lethality, which has prevented the successful development of a TGF-β knockout model (18). The mutation in the DnTGFβRII mouse can be induced in the mature animal; thus it does not disrupt critical TGF-β signaling pathways in the developing heart, while it does provide the ability to selectively inhibit the downstream signaling of TGF-β at the receptor level later in life. Also, there is no need to repeatedly inject agents into the mice or monitor crests and troughs of drug levels. Our transgenic model works on the genomic level, rendering it superior to models that require pharmacological interventions.
The major findings of this study are that under pressure-overload stress, inhibition of TGF-β signaling results in dramatic reductions in nonmyocyte proliferation and collagen content and subsequent development of LV dilation and systolic dysfunction in DnTGFβRII mice. It is well established that under stress the nonmyocyte cells in the LV produce ECM proteins to “control” the amount of remodeling in the overloaded ventricle (20, 3, 5). These proliferating interstitial cells have been characterized as cardiac myofibroblasts by α-smooth muscle actin staining in our previous study (20). Our current findings are consistent with our previous observations that TGF-β stimulates proliferation and myofibroblast transformation of isolated mouse cardiac fibroblasts (20) and that nonmyocyte proliferation and ECM deposition account for pressure-overload-induced LV enlargement in mice subjected to TAC (36). In the current study, there was minimal apoptotic staining (TUNEL stain) of either fibroblasts or myocytes in LV of all experimental groups after 7 days of pressure overload, indicating that apoptosis does not play a major role in the LV remodeling observed in these animals.
In the current study, inhibition of the TGF-β/Smad signaling pathway attenuated the deposition of collagen in our pressure-overload model. In contrast to the other three treatment groups, there was no evidence of pressure-overload-related increases in LV collagen levels in DnTGFβRII+Zn2+ hearts at any time point post-TAC. By 120 days of pressure overload, there was evidence of exaggerated LV dilation and systolic dysfunction, as indicated by increased LV ESV and LVESD, LVEDV and LVEDD, and decreased FS and EF by echocardiographic examination in DnTGFβRII+Zn2+ mice. Taken together, these data suggest that interstitial collagen maybe required to maintain ventricular structure under chronic pressure-overload stress. This study adds to a growing body of knowledge concerning the role of TGF-β signaling in the pathogenesis of cardiac remodeling/fibrosis in response to various forms of stress in mouse models. A previous study (37) in C57/BL6 mice subjected to TAC demonstrated marked increases in TGF-β expression (mRNA for TGF-β1, -β2, and -β3) and signaling (pSmad2 and pSmad1 protein levels). At 3 days post-TAC, the hearts exhibited marked concentric hypertrophy but had no impairment in systolic function by echocardiography. However, after 28 days of TAC the LV was dilated and FS was reduced and collagen deposition significantly increased, indicating the development of heart failure and remodeling in the presence of increased TGF-β expression/signaling (37). This study provided a platform for our investigation of the role of TGF-β signaling in the pathogenesis of pressure-overload-induced cardiac pathology.
Several recent studies using a myocardial infarction model have examined the importance of the TGF-β signaling pathway in determining functional and structural adaptations to ischemic cardiac injury. Inhibition of TGF-β signaling with a nonselective TGF-β1, -β2, and -β3 inhibitor (1D11) has been shown to have detrimental effects on the structure and function of the LV of C57BL/6J mice with myocardial infarction induced by coronary artery ligation (14). Mice underwent intraperitoneal injection of 1D11 or control vehicle every other day beginning 1 wk before or 5 days after coronary artery ligation. Survival was significantly reduced in both groups of 1D11-treated mice (<50%), compared with the vehicle-treated control mice (81%). Deaths were attributed to heart failure. Collagen 1 mRNA levels were attenuated at 3 days after ligation in the 1D11-pretreated group vs. control group. In contrast, at 8 wk collagen deposition assessed by immunohistochemistry was unchanged in both infarct and noninfarct regions by 1D11 treatment compared with control. However, there was significantly increased LV dilation over time (7, 21, 56 day echo) and a nonsignificant worsening of fractional shortening in both 1D11 groups, consistent with our present findings (14). The cardiac effects of a novel oral selective TGFβRI antagonist (SD-208) have also been examined in a murine coronary artery ligation model (10). All mice survived the 30-day experimental period, during which SD-208 treatment reduced collagen deposition and cardiac mass compared with controls, supporting the importance of TGF-β signaling in cardiac remodeling.
Other studies (17, 27) have used adenoviral overexpression of the extracellular domain of the human TGFβRII to block TGF-β signaling in mouse infarct models. Administration of the TGFβRII plasmid 7 days before coronary ligation was associated with increased mortality and LV dilation, but when the same plasmid was administered on the day of ligation and 7 days later, there was significant attenuation of LV remodeling and dysfunction compared with controls (17). Similarly, in another study, administration of an adenoviral TGFβRII plasmid 3 days after coronary ligation was associated with attenuation of LV dilation and reduced cardiac fibrosis compared with controls (27). In contrast, administration of the plasmid 4 wk after coronary ligation, after a scar had already formed, had no effect on LV remodeling. These findings suggest that activation of TGF-β signaling appears to be protective during early ischemic myocardial damage but not during the late phase after myocardial infarction. Thus the timing of interventions that inhibit TGF-β signaling is critical when evaluating differing results. Collectively, in infarct models, TGF-β antagonists reduce collagen deposition and cause varying degrees of cardiac dysfunction depending on the time course and dose of treatment and time points that are evaluated. Early treatments appear to have protective effects on cardiac function, whereas later treatments have detrimental effects.
The effects of the oral TGFβRI antagonist NP-40208 on transgenic mice with cardiac-restricted overexpression of tumor necrosis factor (MHCsTNF) have also been examined (30). MHCsTNF mice develop a heart failure phenotype characterized by progressive myocardial fibrosis, an increase in TGF-β levels, and LV diastolic dysfunction as shown by an increase in the chamber stiffness constant using the Langendorff method. In this study, transgenic and control mice aged 4–12 wk were treated with NP-40208 for 4–12 wk, followed by harvesting of their hearts. Nuclear Smad2/3 increased in the MHCsTNF group compared with littermate control mice, consistent with TGF-β activation. This effect was attenuated by administration of NP-40208, indicating TGF-β signaling blockade. MHCsTNF mice treated with NP-40208 had a small but statistically significant decrease in heart weight-to-body weight ratio and significantly attenuated LV collagen deposition compared with MHCsTNF control mice, as shown by picrosirius red staining, confirming that the TGF-β pathway plays an important role in the development of myocardial fibrosis.
The current findings complement our previous observations that the antifibrotic hormone atrial natriuretic peptide (ANP) is upregulated in heart under pressure or volume overload stress and is functionally active in modulating increases in fibrosis and remodeling. Mice with homozygous deletion of the pro-ANP gene (Nppa−/−) exhibit cardiac fibrosis under resting conditions (13, 11, 36, 26) and develop exaggerated fibrosis and remodeling after pressure or volume overload compared with wild-type control mice (13, 36, 26). These abnormalities appear in attenuated form in heterozygous Nppa+/− mice subjected to pressure-overload stress resulting from TAC, indicating that even partial ANP deficiency results in adverse cardiac remodeling under hemodynamic stress (13). The profibrogenic phenotype that results from eliminating the antifibrotic ANP/cGMP/PKG signaling pathway appears to be related to unopposed TGF-β signaling in response to pressure-overloaded stress. Our laboratory (24, 36, 13, 12) has shown that ANP and TGF-β play important counterregulatory roles in pressure-overload-induced cardiac remodeling and fibrosis. We (20) have also reported previously a direct interaction between the ANP/cGMP/PKG (antifibrotic) and TGF-β (profibrotic) pathways in cardiac fibroblasts. TGF-β accelerates myofibroblast transformation and ECM production in cardiac fibroblasts, and activation of the ANP signaling cascade inhibits these processes by interrupting TGF-β1 signaling at the level of Smad3 phosphorylation.
In summary, we observed that under pressure-overload stress, inhibition of TGF-β signaling attenuates interstitial collagen content and myofibroblast proliferation, leading to LV dilation and dysfunction over a 4-mo follow-up period. These findings underscore the importance of the TGF-β signaling pathway in the pressure-overloaded mouse heart and demonstrate that blockade of this signaling pathway in our novel DnTGFβRII mouse model results in cardiac pathology by promoting LV dilation and dysfunction. Future studies of TGF-β signaling in the stressed heart are warranted to achieve a better understanding of the molecular pathogenesis of the injury response with the ultimate goal of developing novel therapeutic strategies.
This work was supported, in part, by National Heart, Lung, and Blood Institute Grants HL-080017 and HL-044195 (to Y. F. Chen) and HL-07457 and HL-75211 (to S. Oparil); Cardiovascular Pathophysiology Predoctoral Training Grant T32 HL-007918 (to J. A. Lucas); and American Heart Association Greater Southeast Affiliate Grants 0455197B (to Y. F. Chen) and 0425455B and 0765398B (to D. Xing).
No conflicts of interest are declared by the author(s).
- Copyright © 2010 the American Physiological Society