Transcription Factor GATA4 Is Activated but Not Required for Insulin-like Growth Factor 1 (IGF1)-induced Cardiac Hypertrophy

Background: The transcription factor GATA4 is essential in pathological cardiac hypertrophy.

Results: The physiological stimulus IGF1 also increased GATA4 activity but did not require GATA4 for the induction of hypertrophy.

Conclusion: In contrast to pathological stimuli, IGF1 activates but does not require GATA4 for induction of hypertrophy.

Significance: Therapeutic modulation of hypertrophy to a physiological pattern by IGF1 can be achieved independent of GATA4.

Abstract

Insulin-like growth factor 1 (IGF1) promotes a physiological type of cardiac hypertrophy and has therapeutic effects in heart disease. Here, we report the relationship of IGF1 to GATA4, an essential transcription factor in cardiac hypertrophy and cell survival. In cultured neonatal rat ventricular myocytes, we compared the responses to IGF1 (10 nmol/liter) and phenylephrine (PE, 20 μmol/liter), a known GATA4 activator, in concentrations promoting a similar extent of hypertrophy. IGF1 and PE both increased nuclear accumulation of GATA4 and phosphorylation at Ser¹⁰⁵ (PE, 2.4-fold; IGF1, 1.8-fold; both, p < 0.05) and increased GATA4 DNA binding activity as indicated by ELISA and by chromatin IP of selected promoters. Although IGF1 and PE each activated GATA4 to the same degree, GATA4 knockdown by RNA interference only blocked hypertrophy by PE but not by IGF1. PE induction of a panel of GATA4 target genes (Nppa, Nppb, Tnni3, Myl1, and Acta1) was inhibited by GATA4 knockdown. In contrast, IGF1 regulated only Acta1 in a GATA4-dependent fashion. Consistent with the in vitro findings, Gata4 haploinsufficiency in mice did not alter cardiac structure, hyperdynamic function, or antifibrotic effects induced by myocardial overexpression of the IGF1 receptor. Our data indicate that GATA4 is activated by the IGF1 pathway, but although it is required for responses to pathological stimuli, it is not necessary for the effects of IGF1 on cardiac structure and function.

Introduction

Insulin-like growth factor 1 (IGF1)³ is an important regulator of body and organ size during postnatal development (1) and regulates the pluripotent properties of human embryonic stem cells (2). Its production is stimulated by growth hormone and in response to physical exercise (3).

In the heart, IGF1 promotes a physiological pattern of hypertrophy, which is distinguishable from the pathological pattern induced by stimuli such as pressure overload (4). Key features are increased muscle mass with supranormal contractile function, absence of histopathology, and normal survival (5). Downstream signaling requires activation of the phosphoinositide 3-kinase (PI3K p110α) and AKT pathway (4–7). Via activation of this pathway IGF1 promotes cellular growth (8), anti-apoptotic (9), anti-fibrotic (5, 10), and positive inotropic effects (11). In addition, IGF1 has an ability to inhibit signaling cascades that are activated in models of pathological hypertrophy (e.g. downstream of GPCR) (10).

Based on these beneficial cellular actions, the IGF1/PI3K pathway has been considered as a therapeutic agent in heart disease, where it could potentially modulate hypertrophy toward a more physiological type and improve myocyte survival and function. Indeed, marked beneficial effects of IGF1/PI3K have been reported in animal models of dilated cardiomyopathy, decompensated eccentric hypertrophy, diabetic cardiomyopathy, pressure overload, and myocardial infarction (reviewed in Ref. 12). Growth hormone and IGF1 treatment have also been evaluated in human heart failure patients, and a recent meta-analysis of 12 randomized controlled trials suggested that this treatment can improve left ventricular systolic function (13).

GATA4 is a transcription factor with an essential role in cardiac development and in hypertrophy to diverse pathological stimuli (reviewed in (14)). Furthermore, GATA4 is required to maintain normal cardiac function and is up-regulated in human heart failure (15–17). Similar to IGF1, it also acts as a cell survival factor (15, 18). Collectively, this background suggested that GATA4 and the IGF1 pathway may be linked. IGF1 inhibits glycogen synthase kinase-3β (GSK3β) (19) and activates Mek1 and ERK1/2 in cardiac myocytes in vitro (20). Because both GSK3β and ERK1/2 influence GATA4 transcription factor activity (21, 22), we hypothesized that GATA4 might be downstream of IGF1 and an important mediator of its growth and protective effects in cardiomyocytes. Our work presented here confirmed that GATA4 is activated by IGF1. However, GATA4 was not required downstream of IGF1 for the induction of myocyte hypertrophy in vitro, or IGF1 receptor-induced physiological hypertrophy and protective effects in vivo. This reveals an important difference in the underlying mediators of pathological versus physiological type of cardiac hypertrophy.

EXPERIMENTAL PROCEDURES

Neonatal Cardiomyocyte Culture

Neonatal rat ventricular cardiomyocytes (NRVM) were isolated from 3-day-old Wistar rats as described previously (23). Cells were preplated for 1 h to remove nonmyocytes and then plated on gelatinized (2%) cell culture dishes and cultured overnight in DMEM-F12 medium with 10% fetal calf serum and penicillin/streptomycin (100 units/ml) and then changed to serum-free medium containing 80% DMEM-F12 and 20% M199 supplemented with penicillin/streptomycin. On day 2, cells were infected for 8 h with adenovirus for RNA interference against Gata4 (AdG4i described below) and then incubated in virus-free medium. Agonist stimulation was started at day 4 post infection with either PE (20 μmol/liter) or IGF1 (10 nmol/liter). For cardiomyocyte size measurements, cells were fixed with methanol/acetone (ratio 7:1, at −20 °C) and stained with anti-Desmin antibody (1:4, Biomeda, catalog no. V2022) and Alexa Fluor 488 secondary antibody (1:200, Molecular Probes, Invitrogen, catalog no. A21206). Nuclei were stained with ToPro-3 (1:500, Molecular Probes, catalog no. T3605).

Protein Synthesis Assay

NRVM grown on 12-well plates were treated with or without agonists for 24 h as described above and then switched to leucine-free media (Invitrogen/Invitrogen) supplemented with 1 μCi [³H]leucine (Amersham Biosciences) per ml. After a 2-h incubation, cells were treated with 5% trichloroacetic acid (1.5 ml per well) at 4 °C for 30 min and then lysed with NaOH (0.1 n) at 37 °C for 30 min. [³H]leucine incorporation was measured using a scintillation counter in a 500-μl sample volume.

Replication-deficient Adenoviruses for AdG4i

The generation of the adenoviral construct was reported previously (24). Briefly, the following sequences were cloned downstream of the human H1 RNA polymerase III promoter in pENTR-H1/Zeo (underlined letters indicate sequences directed against Gata4): sense, ATCCCCGGAGGGGATTCAAACCAGATTCAAGAGATCTGGTTTGAATCCCCTCCTTTTTGGAA; antisense, AGCTTTTCCAAAAAGGAGGGGATTCAAACCAGATCTCTTGAATCTGGTTTGAATCCCCTCCGGG.

Control RNAi adenovirus (AdCon) was generated by cloning an RNAi sequence that did not have any sequence matches to expressed sequences in the rat genome. NRVM were infected with a multiplicity of infection of 15 plaque-forming units (pfu).

Nuclear Extracts and ELISA for GATA4 DNA Binding Activity

NRVM plated on culture dishes of 10 cm were treated with no agonist (control), 20 μmol/liter PE, or 10 nmol/liter IGF1 for 3 h. Nuclear proteins were extracted by selective lysis buffers (Active Motif TransAM Nuclear extract kit, catalog no. 40010), and 15 μg of nuclear extracts were applied to a GATA4 transcription factor assay kit (TransAM, catalog no. 46496) according to the manufacturer's instructions. At the end of the protocol, bound GATA4 protein was detected by spectrophotometry.

Chromatin Immunoprecipitation

NRVM were plated and treated with agonists as described above. Chromatin preparation was performed by an enzymatic method (Cell Signaling, catalog no. 9003) according to the manufacturer's directions. For immunoprecipitation, a goat primary antibody against GATA4 (Santa Cruz Biotechnology, sc-1237) was used. Quantitative real-time PCR was performed as described below. Sequences and a more detailed description of procedures are given in the supplemental data.

Western Blotting

Whole protein lysates were extracted from NRVM plated on six-well plates and processed as described previously (40 μg per lane) (23). For separated nuclear and cytosolic protein fractions, 30 μg were used per lane. Immunoblotting was performed with the following primary antibodies: GAPDH (1:10000, Bio Trend, catalog no. 4699-9555), GATA4 (1:200, Sigma Aldrich, catalog no. G8794), GATA4 phospho-Ser¹⁰⁵ (1:200, Abcam, catalog no. 5245), ERK1/2 (1:1000, Cell Signaling, catalog no. 9102), phospho-ERK1/2 at Thr²⁰² and Tyr²⁰⁴ (1:7000, Cell Signaling, catalog no. 9101), AKT (1:1000, Cell Signaling, catalog no. 9272), AKT phospho-Ser⁴⁷³ (1:1000, Cell Signaling, catalog no. 9271); GSK3β (1:1000, BD Biosciences, catalog no. 610201/02); GSK3β phospho-Ser⁹ (1:2000, Cell Signaling, catalog no. 9336). Band densitometry was performed as described (23); for nuclear versus cytosolic fractions, the ratio of band densities was calculated.

Quantitative Real-time PCR

RNA was extracted from NRVM on six-well plates by an on column extraction (Qiagen RNeasy kit, catalog no. 74104). Quantitative real-time PCR (qRT-PCR) was performed on an ABI 7900 Sequence Detector (Applied Biosystems) as described previously (15). Primer and probe sequences are provided in supplemental Table 2.

In Vivo Experiments

Generation of transgenic mice with cardiac specific overexpression of the IGF1 receptor (IGFR) (5) and germ line-deleted mice with heterozygous knock-out for Gata4 (G4D) (15) was described previously. Male mice on a uniform F1 genetic background from matings between a C57BL6/J and a FVB/N parent were used in this study. Aortic banding by transverse aortic constriction (TAC), echocardiography, Masson's trichrome stain, and cell size measurements in dissociated cardiomyocytes were performed as published previously (15). All protocols were approved by the Institutional Animal Care and Use Committees and were consistent with the Guidelines for the Care and Use of Laboratory Animals published by the National Institutes of Health.

Statistics

Values are given as mean ± S.E. Statistical comparisons were performed by one-way analysis of variance followed by Fisher LSD test or by Welch's t test.

RESULTS

IGF1- and PE-induced Hypertrophy Is Associated with Activation of GATA4

We compared the growth-promoting effects of 10 nmol/liter IGF1 and 20 μmol/liter PE on neonatal rat ventricular cardiomyocytes. Per each group, a total of >600 cells from five independent cultures were examined for cell size. Both agonists resulted in significant increases in cell size (PE by 48 ± 7% and IGF1 by 62 ± 9%) and in protein synthesis (Fig. 1, A and B). These effects were associated with increased GATA4 binding activity as indicated by two types of evidence: (a) ELISA assay and (b) chromatin IP (ChIP) with consecutive qRT-PCR for quantification of specific GATA4 binding to the promoter regions of previously reported GATA4 target genes.

FIGURE 1.

Hypertrophy and GATA4 activation by IGF1 and PE. A and B, IGF1 (10 nmol/liter) and PE (20 μmol/liter) induced similar degrees of hypertrophy. A, Representative cultured cardiomyocytes stained for sarcomeres (desmin, green) and nuclei (ToPro3, blue), imaged by fluorescence microscopy (bar = 50 μm). B, protein synthesis measured by leucine incorporation after agonist stimulation for 24 h (n = 5). C and D, IGF1 (10 nmol/liter, 3 h) and PE (20 μmol/liter, 3 h) both induced significant increases in GATA4-binding activity. C, ELISA for GATA4 DNA-binding activity (n = 5). Specificity of the assay was confirmed by competition of GATA4 binding by addition of unbound WT versus mutant (Mut) oligonucleotide. D, binding of GATA4-sensitive promoters after chromatin IP measured by qRT-PCR. *, p < 0.05 versus control (Con).

By ELISA, we found that PE and IGF1 induced similar quantitative increases in GATA4 DNA-binding activity (PE, 75 ± 11%; IGF1, 77 ± 13%; both p < 0.05). Specificity of the assay was confirmed by competitive binding studies, in which oligonucleotides containing a wild-type, but not a mutant, GATA-binding site attenuated binding to the immobilized probe (Fig. 1C).

Chromatin IP also revealed significant increases in GATA4 binding by both agonists, PE and IGF1. Atrial natriuretic peptide (Nppa) and brain natriuretic peptide (Nppb) showed moderate responses, and binding to myosin light chain (Myl1) and α-skeletal Actin (Acta1), increased profoundly by both agonists (Fig. 1D and supplemental data).

GATA4 Is Activated by Similar Signaling Downstream of IGF1 and PE

GATA4 can be activated by phosphorylation at serine 105 (21) and inhibited by GSK3β, which promotes the nuclear export of GATA4 (22). We hypothesized that PE and IGF1 increase GATA4-binding activity by both phosphorylation at Ser¹⁰⁵ and inhibition of GSK3β via phosphorylation at Ser⁹. Consistent with this hypothesis, we found that both agonists promoted nuclear accumulation of GATA4, as indicated by a significant increase in the ratio of nuclear to cytoplasmic GATA4 (Fig. 2A). Both agonists also increased phosphorylation of GATA4 at Ser¹⁰⁵ (IGF1, 1.8-fold; PE, 2.4-fold; Fig. 2B), whereas total GATA4 protein level remained unchanged (data not shown). Phosphorylation of the upstream kinase ERK1/2 was robustly increased by PE and to a lesser extent by IGF1 (Fig. 2, C and D). In contrast, IGF1 induced stronger phosphorylation of GSK3β at Ser⁹ (1.6-fold versus 1.2-fold with PE; Fig. 2E). Only IGF1 and not PE was capable of inducing phosphorylation of AKT at Ser⁴⁷³ (3.4-fold; Fig. 2F). Total protein levels of ERK1/2, GSK3β, and AKT did not change during exposure to IGF1 or PE (data not shown).

FIGURE 2.

IGF1- and PE-dependent signaling for GATA4 activation. A, Western blot analysis of cytoplasmic and nuclear protein extracts after 3 h of stimulation, the ratio of nuclear to cytoplasmic GATA4 levels increased after IGF1 (10 nmol/liter) and PE (20 μmol/liter) (n = 4). B–F, quantitation of Western blots for phosphorylated protein levels in whole protein lysates after 5 min of stimulation with IGF1 or PE, normalized to corresponding total protein and to GAPDH (n = 6 experiments). B, phosphorylated GATA4 (Ser¹⁰⁵). C, phosphorylated ERK1 (Thr²⁰² and Tyr²⁰⁴). D, phosphorylated ERK2 (Thr¹⁸⁵ and Tyr¹⁸⁷). E, phosphorylated GSK3β (Ser⁹). F, phosphorylated AKT (Ser⁴⁷³). G, representative Western blots. *, p < 0.05 versus control (Con).

GATA4-dependent Gene Expression in PE- and IGF1-induced Hypertrophy

Having shown that GATA4 bound directly to a panel of downstream target genes, we were next interested whether GATA4 was also required for agonist-induced changes in gene expression. Therefore, we used a loss-of-function approach for GATA4 and treated NRVM either with a virus promoting RNA interference for Gata4 (AdG4i) or AdCon. Infection of neonatal cardiomyocytes with AdG4i resulted in knockdown of GATA4 >90% at day 4 post-infection (Fig. 3A) and significantly down-regulated the mRNA expression of four of six GATA4 targets that we investigated: atrial natriuretic peptide (Nppa), brain natriuretic peptide (Nppb), β-myosin heavy chain (Myh7), and troponin I (Tnni3). In contrast, α-skeletal Actin (Acta1) and myosin light chain (Myl1) were strongly up-regulated (Fig. 3B). We then analyzed the response of these genes to PE versus IGF1 stimulation. In AdCon-treated NRVM, PE induced a time-dependent up-regulation of five of six of the examined genes (Nppa, Nppb, Tnni3, Acta1, Myl1; all p < 0.05) (Fig. 4A). This response was GATA4-dependent, as it was diminished or abolished after AdG4i (Fig. 4B). Interestingly, the response of these genes to IGF1 stimulation was markedly different: IGF1 significantly down-regulated expression of Nppa, Nppb, Myh7, and Tnni3, but up-regulated Acta1 and Myl1 (Fig. 4C). With the exception of Acta1, this response was not GATA4-dependent, as knockdown of GATA4 did not alter the effect of IGF1 (Fig. 4, C and D). These data suggest that, despite their similar activation of GATA4 binding, PE and IGF1 differentially regulate expression of GATA4 downstream genes.

FIGURE 3.

Loss-of-function model against Gata4. A, Western blot for GATA4 protein levels after infection (p.i.) with AdG4i (multiplicity of infection of 15) or control virus (AdCon) for indicated number of days. AdG4i resulted in almost complete knockdown of GATA4. B, relative mRNA expression of GATA4 downstream genes determined by qRT-PCR in cardiomyocytes infected with AdCon or AdG4i (n = 7 per group). AdG4i decreases expression of four of six downstream targets. *, p < 0.05 versus AdCon.

FIGURE 4.

Role of GATA4 in IGF1- and PE-induced gene expression. A–D, relative mRNA expression of GATA4 downstream genes determined by qRT-PCR in cardiomyocytes infected with AdCon or AdG4i (n = 7 per group). A, PE (20 μmol/liter) up-regulated expression of five of six targets after 24 h. B, AdG4i abolished PE-dependent up-regulation of gene expression. C, IGF1 (10 nmol/liter) up-regulated Acta1 and Myl1 expression, but down-regulated the others. D, AdG4i blocked up-regulation of Acta1 but not Myl1 by IGF1. *, p < 0.05 versus control (Con); #, p < 0.05 versus AdCon.

GATA4 Is Essential for PE- but Not for IGF1-induced Hypertrophy

To test the consequences of GATA4-dependent gene expression, we investigated the functional role of GATA4 for the induction of hypertrophy after IGF1 or PE and examined cell size and protein synthesis in the presence of AdG4i. Knockdown of GATA4 almost completely prevented the PE-induced increase in cell size and in protein synthesis but did not significantly affect the hypertrophic response to IGF1 (Fig. 5, A–C). These data indicate that GATA4 is essential for PE-induced hypertrophy but not for IGF1-induced hypertrophy.

FIGURE 5.

Requirement of GATA4 in IGF1 and PE-induced hypertrophy. A–C, AdG4i blocked PE- but not IGF1-induced hypertrophy. A, cultured cardiomyocytes after infection with AdG4i (multiplicity of infection of 15) and stimulation with IGF1 (10 nmol/liter) or PE (20 μmol/liter) for 48 h. Staining for sarcomeres (desmin, green) and nuclei (ToPro3, blue), imaged by fluorescence microscopy. Bar, 50 μm. B, quantitative results for myocytes cell area from n = 5 cultures. C, protein synthesis indicated by leucine incorporation, agonist stimulation for 24 h (n = 5). *, p < 0.05 versus AdCon. Con, control.

Role of GATA4 in IGF1-induced Hypertrophy in Vivo

As a sensitive approach to screen for a other roles of GATA4 in IGF1-induced hypertrophy, we studied the effects in vivo. We used mice with a heterozygous deletion of Gata4 exon 2 (abbreviated G4D), which express 50% less GATA4 than wild-type mice. As reported previously (15), these mice were susceptible to heart failure after pathological stimulation by pressure overload. We crossed G4D mice to transgenic mice with cardiomyocyte restricted overexpression of the IGFR (5). In mice followed up to an age of 18 months, mortality was not different between G4D-IGFR, IGFR, and wild-type mice. In accordance to our in vitro data, heart weight to body weight ratios (HW/BW) were increased significantly in IGFR mice (5.5 ± 0.1 mg/g; n = 5) compared with wild-type littermate controls (4.6 ± 0.1 mg/g; n = 7, p < 0.05), and this cardiac hypertrophy was not affected by heterozygous knock-out for Gata4 (HW/BW ratio 5.5 ± 0.1 mg/g in G4D-IGFR; n = 5; Fig. 6A). At the cellular level, IGFR cardiomyocytes showed significant hypertrophy, which was not attenuated in G4D-IGFR cardiomyocytes (Fig. 6B). In fact, G4D-IGFR cardiomyocytes were even larger than IGFR cardiomyocytes, which was associated with a decrease in cardiomyocyte number (hypoplasia) observed in both G4D and G4D-IGFR hearts (supplemental Fig. 4). Because hypoplasia might lead to hypertrophy of remaining cardiomyocytes for compensation of wall stress (15), this could explain why cell size was largest in G4D-IGFR hearts. IGFR mice had hypercontractile systolic function, as measured by echocardiographic fractional shortening (61 ± 3% (IGFR) compared with 53 ± 3% (wild-type); p < 0.05). This was unchanged by Gata4 heterozygosity (G4D-IGFR, 64 ± 3%; Fig. 6C). Neither of the mouse models developed cardiac fibrosis under basal conditions, as determined by Masson's trichrome staining of histological sections. However, induction of pressure overload by TAC induced significant fibrosis in WT, which was further exacerbated in G4D mice (p < 0.05 versus WT). In contrast, TAC-induced myocardial fibrosis was substantially less in IGFR and G4D-IGFR mice (Fig. 6, D and E). These data suggest that activation of the IGF1 pathway in vivo has strong, beneficial effects on the cardiac phenotype that are insensitive to GATA4 levels.

FIGURE 6.

Role of GATA4 in IGF1-induced hypertrophy in vivo. A, HW/BW ratio was similarly increased in IGFR and G4D-IGFR but decreased in G4D. B, dissociated cardiomyocytes showed hypertrophy in G4D and IGFR and an additive effect of both in G4D-IGFR. C, fractional shortening was similarly increased in IGFR and G4D-IGFR, but decreased in G4D. D, area fraction of cardiac fibrosis in sham- and TAC-operated mice, IGFR, and G4D-IGFR hearts exhibited similar antifibrotic effects after TAC. E, representative histological sections of hearts after TAC, fibrotic areas appear in blue (Masson's trichrome stain). A–D, numbers inside bars indicate number of samples per group. For fibrosis staining, n = 3–7 per group. *, p < 0.05 versus WT; #, p < 0.05 versus sham.

DISCUSSION

IGF1 and GATA4 are two potent mediators of hypertrophy and cellular survival. Yet, the relationship between IGF1 and GATA4 has not previously been elucidated. Our data indicate that GATA4 activation occurs downstream of IGF1, similar to its activation downstream of other hypertrophic agonists such as PE. However, GATA4 was not essential for IGF1-induced hypertrophy in vitro and in vivo, and IGF1 could exert important structural and functional effects independent of GATA4.

Activation of GATA4 and Signaling in IGF1 Pathway

IGF1 stimulates the PI3K-AKT-GSK3β pathway. Activation of this pathway by the cytokine interleukin-18 (25) and also by carbamylated erythropoietin (26) has been linked to an increase in GATA4 activity in cardiomyocytes. Therefore, we hypothesized that GATA4 might be activated downstream of IGF1 and important for the induction of hypertrophy. To test our hypothesis, we compared the effects of IGF1 to PE, a prototypic agonist for pathological hypertrophy, which is known to require GATA4 (27). We found that IGF1 and PE, in concentrations that induced similar degrees of hypertrophy, stimulated GATA4 DNA-binding activity to a similar extent (Fig. 1). Furthermore, both agonists induced GATA4 phosphorylation at Ser¹⁰⁵ and nuclear accumulation of GATA4 (Fig. 2). Phosphorylation of GATA4 at Ser¹⁰⁵ increases GATA4 DNA-binding activity as well as transactivation potential, and this is essentially regulated by the kinase ERK1/2 (21). In contrast, GATA4 nuclear accumulation is regulated mainly by the activity of the kinase GSK3β (22), which can be inhibited by the kinase AKT. Accordingly, GATA4 nuclear accumulation is enhanced under conditions of AKT overexpression (28). In our experiments, only IGF1 induced phosphorylation of AKT, which was not present in PE-induced hypertrophy. This is in line with previous data on insulin and PE (29) and with in vivo data showing that AKT is essential only in physiological but not in pathological hypertrophy (6).

Functional Role of GATA4 in IGF1-induced Hypertrophy

Effects on Gene Expression

In our study, we investigated genes that were previously shown to be downstream targets of GATA4 (30–31). Knockdown of GATA4 by siRNA confirmed GATA4 dependence in four of six targets (Fig. 3), two others (Acta1 and Myl1) were up-regulated possibly due to compensatory activation of other transcription factors having binding sites on their promoters (31, 32).

Our chromatin IP data confirmed direct binding of GATA4 to Nppa, Acta1, Myl1, and weakly to NppA. Interestingly, both agonists, IGF1 and PE, resulted in an overall similar pattern of activation of GATA4 binding (Fig. 1) but had very different effects on gene expression: PE up-regulated Nppa, Nppb, Tnni3, Acta1, and Myl1, and GATA4 knockdown significantly diminished or completely abolished this up-regulation (Fig. 4). Thus, PE-induced expression of this set of genes was GATA4-dependent.

In contrast, IGF1 only up-regulated Acta1 in a GATA4-dependent fashion. Myl was up-regulated irrespective of GATA4 expression, and all other genes (Nppa, Nppb, Myh7, and Tnni3) were down-regulated by IGF1. This difference to PE is striking in the light of the similar activation pattern of GATA4 binding. A possible explanation lies in differences in the co-activation of other transcription factors and co-factors that act in concert with GATA4. One of those might be Nuclear factor of activated T-cells (NFAT) which was shown to be activated selectively in pathological hypertrophy, but not in physiological hypertrophy induced by exercise or IGF1 (33).

The fact, that some targets did not only fail to be up-regulated by IGF1 but were even down-regulated might be explained by an alternate transcriptional pathway involving Foxo3a, a Forkhead transcription factor that regulates a set of “atrogenes,” so named for their anti-hypertrophic action (34). IGF1 was shown to inhibit Foxo3a activity, resulting in down-regulation of expression of Foxo3a target genes by IGF1 (34). The Nppa and Nppb promoter regions also have several Forkhead transcription factor binding sites (35), which may explain their down-regulation by IGF1.

Effects on Cell Size and Protein Synthesis

We found that knockdown of GATA4 blocked PE-induced hypertrophy (Fig. 5), as assessed by both cell size and protein synthesis. Previously, a dominant-negative GATA4-engrailed repressor fusion protein was shown to block PE-induced hypertrophy in vitro (27). Our loss-of-function study reinforces this result and extends it by excluding potential off target effects inherent to overexpression of a dominant-negative protein.

In contrast to PE, IGF1-induced hypertrophy did not require GATA4. IGF1 can up-regulate mRNA translation directly via mediators such as mTOR, S6 kinase, and eukaryotic initiation factor eIF4E and may in general stimulate protein synthesis largely independent of transcriptional changes (36). Our data indicate that GATA4 plays a non-essential or redundant role in IGF1-induced hypertrophy.

Cardiac Responses to IGF1 in Vivo Are Not Sensitive to GATA4 Dosage

We previously showed that a physiologically relevant reduction in GATA4 expression (50% in G4D mice) resulted in a significantly altered response to TAC, a pathological hypertrophic stimulus (15). On this background, it appears striking that the hypertrophic response to IGF1 activation in vivo was not influenced by a similar reduction in GATA4 expression. Heart size, cardiac function, and histology were unaltered in G4D-IGFR compared with IGFR mice (Fig. 6). In fact, cardiomyocytes hypertrophied even more in G4D-IGFR compared with IGFR hearts. This is likely due to congenital reduction of cardiomyocyte number in G4D and G4D-IGFR fetal hearts (supplemental Fig. 4), resulting in compensatory hypertrophy of remaining cardiomyocytes. On top of this baseline hypertrophy in G4D hearts, IGFR stimulated additional hypertrophy in G4D-IGFR hearts, indicating that IGF1-induced hypertrophy is insensitive to GATA4 dosage. It is possible that a more complete reduction of GATA4 levels might have unmasked a requirement for GATA4 downstream of IGF1. However, our in vitro data, in which >90% knockdown of GATA4 protein did not have a significant influence on the hypertrophic response of IGF1, makes this possibility less likely.

Stimulation of the IGF1 pathway improved contractile function and ameliorated contractile dysfunction and cardiac fibrosis under pressure overload (5, 15). In contrast, reduction of GATA4 levels by 50% exacerbated these adverse effects (15). Remarkably, the beneficial effects of IGF1 pathway stimulation were preserved in G4D-IGFR mice, suggesting that these beneficial actions of IGF1 are insensitive to GATA4 dosage. The explanation for this independence might be either that GATA4 is not involved or that loss of GATA4 might be compensated for by other factors in the PI3K-AKT cascade.

Therapeutic Implications

Accumulating evidence suggests that activation of the IGF1 pathway has beneficial effects in heart disease and can modulate a preexisting pathological hypertrophy toward a more physiological pattern and improve survival (5, 9, 10). However, a therapeutic strategy targeted and restricted to the heart is mandatory to avoid tumorigenic effects of IGF1 stimulation in other organs (37). The findings of our study, that IGF1 stimulates GATA4 activity but is able to exert beneficial effects independent of GATA4, suggests that cardiac-restricted stimulation of the IGF1 pathway, remains an appropriate therapeutic strategy in clinically relevant situations when GATA4 activity is down-regulated, such as in doxorubicin cardiotoxicity (18).

Conclusion

A comprehensive knowledge of the participating and interacting factors in the IGF1 pathway is required to increase its applicability as a therapeutic approach in heart disease. In this regard, IGF1 and GATA4 show an interesting relationship: downstream signaling from IGF1 to GATA4, but functional independence of the IGF1 pathway and GATA4.