Effect of Cell-Based Intercellular Delivery of Transcription Factor GATA4 on Ischemic Cardiomyopathy
Recent loss-of-function studies highlight the importance of the transcription factor GATA4 in the myocardial response to injury in the adult heart. However, the potential effects of gain-in-function of GATA4 overexpression, and transcription factors in general, is hindered by the fact that transcription factors are neither secreted nor taken up by cells. Chimeric proteins incorporating motifs of cell-penetrating proteins are secreted from cells and internalized by surrounding cells. We engineered a chimeric protein consisting of GATA4 and the cell-penetrating protein VP22. Cardiac fibroblasts stably transfected with the GATA4:VP22, GFP:VP22, or green fluorescent protein (GFP) constructs were transplanted into Lewis rats 1 month after left anterior descending ligation. GATA4:VP22 expression in the border zone was associated with increased cardiac myosin expression and cardiac myocyte size (30 μm versus 19 μm, P<0.01). Compared with the GFP:VP22 control group, 6 weeks after cardiac fibroblast transplantation (10 weeks after myocardial infarction), animals that received GATA4:VP22-expressing cardiac fibroblasts demonstrated increased cardiac function and less negative remodeling. These data demonstrate a novel strategy for transcription factor delivery to injured myocardium and indicate that the delivery of GATA4 locally to the infarct border zone induces multiple local effects in the border zone cardiac myocytes resulting in beneficial ventricular remodeling and improved global left ventricular function.
GATA4 is critical for the viability and hypertrophic response of cardiac myocytes following myocardial injury.1,2 Consistent with these findings, GATA4 has been shown to directly induce the expression of BCL2 in cardiac myocytes in vitro.3 These findings suggest that the local overexpression of GATA4 could lead to local cardiac hypertrophy and improved cardiac myocyte survival. The goal of this study was to assess, for the first time, the effects of local GATA4 overexpression in the infarct border zone at a time remote from myocardial infarction in a model of ischemic cardiomyopathy. To do so, because transcription factors are neither secreted nor internalized by surrounding cells, we used a novel method of local GATA4 delivery to the infarct border zone.
We achieved local GATA4 delivery by combining cell-based gene therapy with a cell-penetrating protein (CPP). Cell-based gene therapy has been shown to be an effective strategy for stimulating angiogenesis and improving heart function.4–6 Moreover CPPs, by serving as vectors for the transmembrane intercellular delivery of fused proteins, have emerged as a tool to modulate biological activities.7–11 Cardiac fibroblasts were engineered in culture to overexpress a chimeric protein encoding GATA4 and the CPP VP22. We recently demonstrated the utility of this strategy to deliver the nonsecreted marker protein green fluorescent protein (GFP) to myocardial tissue.12 In this previous study, we demonstrated that VP22 chimeric proteins are delivered to ≈1 mm2 of myocardium in a direction radial to the needle track. In the present study, we quantified the effects of sustained GATA4:VP22 release into the infarct border zone in a rat model of ischemic cardiomyopathy induced by left anterior descending (LAD) ligation 1 month before cell delivery. We aim to show that the cell-based sustained delivery of nonsecreted functional proteins offers a potential novel strategy to study the effects of nonsecreted proteins in adult hearts at a time remote from myocardial injury, as well as a potentially interesting strategy for the optimization of cardiac function following myocardial injury.
Materials and Methods
GATA4:VP22 constructs were made by inserting GATA4 (generously provided by Dr J. D. Molkentin at Cincinnati Children’s Hospital Medical Center, Ohio) into C termini of VP22 in frame in the plasmid pVP22/myc-His (Invitrogen, Carlsbad, Calif). The cytomegalovirus promoter drove GATA4 expression in the vector pVP22/GATA4. The anticipated molecular mass for fusion protein GATA4:VP22 was ≈80 kDa. Plasmids of pVP22/GFP and GFP were constructed and transfected into rat cardiac fibroblasts as described previously.12
Luciferase GATA4 promoter reporter construct containing GATA4 binding site was made using the GATA motif (A/T)GATA(A/G).13 The insert containing (A/T) GATA (A/G) motif with 3 repeats was pGL3 vector (Promega Corp, Madison, Wis) at the XhoI and HindIII sites. The DNA sequence insert from 5′ to 3′ ligated into pGL3 was as depicted in Figure 1C.
Effect of GATA4:VP22 on Gene Expression in Mesenchymal Stem Cells
Cell culture 6-well plates and inserts (BD Biosciences) were used to test the effect of intercellular delivery of GATA4 on mesenchymal stem cells (MSCs) in vitro. MSCs isolated from the bone marrow of Lewis rat14 (5000 cells) were placed in one well of the cell culture plates. Rat cardiac fibroblasts (5000 cells) stably expressing GFP, GFP:VP22, or GATA4:VP22 were added to the cell culture insert cup. The bottom layer of the cup consisted of a porous polyethylene terephthalate membrane with 1-μm pores. Analyses of these cells are detailed in the online data supplement at http://circresahajournals.org.
Myocardial Infarction and Cell Delivery
The Animal Research Committee approved all animal protocols, and all animals were housed in the Association for the Assessment and Accreditation of Laboratory Animal Care animal facility of the Cleveland Clinic. Myocardial infarction and cell delivery were performed as described previously5,6 and as detailed in the online data supplement.
We performed echocardiography as detailed previously15,16 and as described in the online data supplement.
Quantification of Cardiac Myosin Expression in Myocardium
To quantify cardiac myosin expression in the infarct border zone and areas remote from myocardial infarction, immunofluorescence was performed by incubation with monoclonal cardiac myosin heavy chain (Chemicon) with secondary donkey anti-mouse Alexa fluor 594 (Invitrogen). Fluorescence images were obtained with a Leica TCS SP2 AOBS confocal laser scanning system (Leica Co, Wetzlar, Germany).
Means of fluorescence intensity (FI) of cardiac myosin in the myocardium were measured from 4 chosen fields per animal, either in the periinfarct zone (FIp) or remote areas (FIr) using ImagePro Plus software at a magnification of ×63. Relative fluorescence intensity (R) was obtained by dividing mean of fluorescence intensity in periinfarct zone by the one in remote areas per animal.
All data are presented as means±SD. Comparisons between groups for continuous variables were made by Student t test or variance analysis followed by post hoc Tukey’s honestly significantly different test for multiple comparisons, where appropriate. For segmental radial strain data, we used repeated measures analysis of variance to analyze the change of segmental radial strains during a 6-week treatment period, followed by analysis of contrasts. Results were considered statistically significant if P<0.05. For animal survival analysis Kaplan–Meier curves were constructed to assess survival rates during follow up. Log rank statistics was used to compare survival of GATA4 treatment group with the survival of 2 merged control groups. The control groups were merged to eliminate the need for multiple comparisons.
GATA4:VP22 Is Delivered Intercellularly and Binds in the Nucleus
A plasmid encoding GATA4:VP22 was constructed that included a myc tag in the C terminus of the fusion proteins (Figure 1A). We transfected this construct into HeLa cells and rat cardiac fibroblasts (RCFs). Western blot using antibodies against myc or GATA4 epitopes showed that the chimeric protein was expressed with the predicted molecular mass in RCFs with the GATA4:VP22 construct being 84 kDa (Figure 1B).
To verify that the chimeric GATA4 protein retained its ability to activate GATA4 responsive promoters, we used a luciferase reporter assay with a GATA4 responsive promoter in HeLa cells. A GATA4 binding sequence containing 5′-(A/T)GATA(A/G)-3′ was inserted into the pGL3–luciferase construct (Figure 1C). The reporter construct was transiently transfected into wild-type HeLa cells or HeLa cells stably expressing empty vector, GATA4, or GATA4:VP22. The results of luciferase assay demonstrated that GATA4:VP22 chimeric protein maintained the ability to activate a GATA4 responsive promoter (Figure 1D).
To further determine whether the chimeric GATA4 protein was active, we evaluated the effects of coculturing MSCs in the presence of GFP-, GFP:VP22-, and GATA4:VP22-expressing cardiac fibroblasts. In these experiments, we separately cultured the MSCs and the cardiac fibroblasts by culturing the MSCs on the cell culture well and growing the cardiac fibroblasts on a well insert. In this configuration, the MSCs were exposed to proteins secreted by the cardiac fibroblasts. Western blot analysis of the MSCs after 3 and 7 days after culture demonstrated that the MSCs had increased expression of the GATA4 inducible gene troponin I. The upregulation of troponin I in the presence of GATA4:VP22-expressing, but not GFP:VP22-expressing, cardiac fibroblasts is consistent with the GATA4:VP22 construct having native GATA4 activity. Interestingly, we also observed an increase in the cardiac transcription factor Nkx2.5 as well as the structural protein α-actin.
Intercellular Delivery of GATA4 Improved Left Ventricular Function of Infarcted Hearts
We investigated the effects of intercellular delivery of GATA4 1 month after myocardial infarction on left ventricular (LV) remodeling and function. LV function and contractility were evaluated by detailed echocardiography. RCFs stably expressing GATA4:VP22 or control groups of RCFs stably expressing GFP:VP22 or GFP were directly injected into the periinfarct zones of Lewis rats 1 month after LAD ligation in 5 divided doses around the infarct zone.5,6
Baseline echocardiography was performed within 2 days before cell transplantation (4 weeks after LAD ligation) and 4 and 6 weeks after cell transplantation (total, 10 weeks after LAD ligation). Four weeks after cell transplantation, the percentage change in LV fractional shortening compared with baseline was significantly increased (37% increase) in the GATA4:VP22 group compared with a decline in function in the GFP:VP22 (−11%, P=0.013) and the GFP (−9%, P=0.027) control groups (Figure 2A). Six weeks after cell transplantation, the improvement in fractional shortening was maintained in hearts that received GATA4:VP22 (54% increase) compared with the GFP:VP22 (−15%, P=0.036) and GFP (−27%, P=0.007) control groups (Figure 2A).
Systolic and diastolic thickness of the anterior and inferior walls and shortening fraction at baseline as well as 4 and 6 weeks after cell transplantation are presented in Tables 1 and 2⇓, respectively. The data in Table 1 demonstrate that between baseline and 4 and 6 weeks, there was a trend toward greater pathological hypertrophy of the inferior wall in those animals that received control cardiac fibroblasts compared with those that received GATA4:VP22-expressing cardiac fibroblasts. Despite the increase in wall thickness, or perhaps because of the pathological nature of the wall thickening in the noninfarcted myocardium, the wall thickening of the inferior wall was significantly increased 6 weeks after cell transplantation with GATA4:VP22 (65%) compared with the control groups (GFP:VP22, 35%, P=0.021; GFP 26%, P=0.002; Figure 2B). In contrast, there was no significant difference in anterior wall thickening compared with baseline in any groups (Figure 2B).
We observed a significant difference in segmental radial LV strains 4 weeks after myocardial infarction (baseline measurements).15 The radial strains were lowest in the infarct zone (anterior and lateral segments), higher in the border (posterior and anteroseptal segments), and greatest in the zones remote from myocardial infarction (inferior and septal segments) (P<0.001). Six weeks after cell transplantation, there were no differences observed among groups in the infarct zone (p=NS); however, radial strains improved significantly in the border zones, and the noninfarcted zones in those animals that received GATA4:VP22-expressing cells compared with the GFP:VP22 group (P=0.045) and GFP group (P=0.031) (Figure 2C).
Cell-Based Intercellular Delivery of GATA4 Resulted in Remodeling of the Infarcted Hearts and Increased the Survival Rate of Infarcted Rats
To further investigate whether the intercellular delivery of GATA4 influenced the morphology of the heart, Masson’s trichrome staining was performed to quantify the extent of LV fibrosis. Anatomic assessments of Masson’s trichrome staining are shown in Figure 3A. The percentage area of collagen deposition in the left ventricle was significantly reduced 6 weeks after cell transplantation of GATA4:VP22-expressing RCFs (19%) compared with GFP:VP22 (39%, P=0.001), GFP (31%, P=0.035), and PBS (33%, P=0.003), and the percentage of the LV circumference in which there was collagen deposition was also reduced significantly in the GATA4:VP22 compared with GFP:VP22, GFP, and PBS (data not shown).
To evaluate the effect of using CPP VP22 delivery of GATA4 in infarcted rat hearts, the survival rates of each group were analyzed during the 6 weeks following cell transplantation. Figure 4 shows that the group that received GATA4:VP22 had a 100% of survival rate under our experimental conditions, whereas there was a 30% mortality rate in the control groups. The survival in the GATA4:VP22 experimental group was significantly better in a treatment group compared with the combination of the 2 control groups (P=0.026).
These data demonstrated that, in addition to improving the LV function, cell transplantation of RCFs stably expressing GATA4:VP22 could also improve the morphology of infarcted hearts by reducing fibrosis of infarcted myocardium. The analysis of survival rates indicated that this strategy is well tolerated and could improve the survival rate of infarcted rats, suggesting that cell-based sustained intercellular delivery of GATA4 might have prevented negative remodeling, a process that occurs late after myocardial infarction. Furthermore, these survival data as well as our histological findings suggest that there were no negative consequences or immune response to sustained VP22 expression in the heart.
Cardiac Myocyte Hypertrophy
GATA4 is known to be involved in cardiac hypertrophy, which can protect the heart by reducing wall stress and the subsequent sparing of energy and oxygen consumption,17 thereby compensating for decreased cardiac function following myocardial infarction1 or pressure overload.2 Therefore we examined the extent of cardiac hypertrophy by measuring myocyte cross-sectional areas in the infarct border zone and areas remote from the infarct zone 6 weeks following cell transplantation with GATA4:VP22-expressing RCFs or the control cell groups.
We observed greater cardiac myocyte hypertrophy in the infarct border zone in those animals that received GATA4:VP22-expressing RCFs (Figure 5A) compared with the control groups and noninfarcted animals. Microscopic analysis of hematoxylin/eosin-stained sections revealed that the diameter of cross-section of the cardiac myocytes in the periinfarct zone of those animals that received RCF GATA4:VP22 was significantly increased (30 μm) compared with all other control groups (GFP:VP22, 19 μm, P=0.001; GFP, 19 μm; P=0.001) (Figure 5B). No difference was observed in cardiac myocyte size among groups or compared with normal hearts in areas remote from the infarct zone (cross-sectional diameter: PBS, 14.5±2.0 μm; RCFs expressing GFP, 14.6±0.7 μm; GFP:VP22, 14.9±2.0 μm; GATA4:VP22: 15.0±2.3 μm).
GATA4 Responsive Gene Expression
We investigated the effect of intercellular delivery of GATA4 on the expression of cardiac myosin in the infarct border zone. Double-labeled immunohistochemical staining was used to identify the original area of transplanted cells within the infarct zone. The VP22 chimeric proteins have myc tags in the C terminus. Using myc staining, we traced injected cells in the periinfarct zone 6 weeks after cell transplantation. In the periinfarct zone around the needle track from cell injection, there were GATA4:VP22- and GFP:VP22-expressing cells (Figure 6A). High-power magnification of the insert in the GATA4:VP22 image in Figure 6A reveals the presence of myc tag in the cytoplasm of the cardiac myocytes (green or yellow−red+green channels) as wells as the presence of myc-tagged protein in the nucleus (green or pink−blue+green channels).
In sections from animals stained for cardiac myosin heavy chain (antibody recognized both α and β isoforms), using constant laser power and exposure times, we found in the animals that received GATA4:VP22-expressing RCFs a relatively higher expression of cardiac myosin heavy chain (GATA4:VP22; Figure 6C), as well as hypertrophied cardiac myocytes, compared with control groups GFP:VP22 and GFP (GFP:VP22 and GFP; Figure 6A).
GATA4 has been shown to induce antiapoptotic proteins in the heart following myocardial infarction.3 In vitro, the antiapoptotic protein BCL2 has been shown to be a direct target of GATA4 and could mediate the prosurvival effects of GATA4.2,14,19 We observed a significant increase in the level of BCL2 expression in the infarct border zone in animals that received GATA4:VP22 (Figure 6D). No increase in BCL2 expression was observed in areas remote from the infarct (data not shown) or in the GFP:VP22 control group (Figure 6D).
Our study provides the first data to be derived from the effects of transcription-factor delivery by a cell-penetrating peptide using cell-based gene transfer in ischemic cardiomyopathy. We show that gain-of-function of the transcription factor GATA4 within the infarct border zone at a time remote from myocardial infarction via cell-based gene transfer using the cell-penetrating peptide VP22 can significantly impact LV remodeling and cardiac function.
Our data demonstrate that the chronic overexpression of GATA4 in the infarct border zone had several effects ultimately leading to global improvement in cardiac function. We found that intercellular delivery of GATA4-induced cardiac myosin overexpression and hypertrophy (Figure 5) of cardiomyocytes around the injection site in the periinfarct zone compared with injections with control cardiac fibroblasts (Figure 6). Thus sustained reconstitution of the transcription factor GATA4 in the infarct zone may exert favorable effects on myocardium repair and preservation. All of these findings were associated with enhanced LV function outside the infarct zone (Figure 2) and improved ventricular morphology (Figure 3). Thus, consistent with the role of GATA4 in embryonic and injured adult heart, where it regulates several distinct processes (such as cardiogenesis, antiapoptosis, and morphogenesis),13,20,21–24 the intercellular delivery of GATA4 in the infarcted adult heart may improve the LV function through multiple effects.
The majority of previous studies that focused on GATA4 function have assessed its role in the regulation of cardiac development and differentiation of myocyte in embryos.25 Because it controls a number of critical functions in the heart, including embryonic cardiogenesis and proper cardiac morphogenesis early in the heart development, most of studies have focused on GATA4 knockout mice with failed cardiogenesis and early lethality. The importance of GATA4 in regulating heart failure in the adult heart is less understood, although loss-of-function studies have recently demonstrated that GATA4 regulates cardiac hypertrophic response and myocyte survival following myocardial injury in the adult heart.1,2,20 These studies showed that deletion of GATA4 specifically in heart, using Cre/lox-dependent conditional gene-targeting approach, leads to a progressive and dose-dependent deterioration in cardiac function and dilation in adulthood, and compromises cardiac hypertrophy and myocyte viability.1 In line with this previous work showing detrimental effects of cardiac-specific deletion of GATA4 in the heart, our studies demonstrate improved LV function with local GATA4 overexpression late after myocardial infarction in the adult infarcted myocardium (Figures 2 and 3⇑).
In our study, improvement in LV function was mediated by an increase of radial strain in both infarct border zone and remote-infarct zone in the GATA4 treatment group compared with controls (Figure 2C). It appears somewhat unexpected that improvements in radial strain was not confined to the treated (border) zones, but extended into the areas remote for infarct. This finding suggests that the strain not only reflects local properties of the myocardium but may also be determined by the interaction of these local properties and regional stress.26 In end-systole, stress–strain relationships are linear, with the slope of this relationship reflecting regional contractility.27 After myocardial infarct, regional stress increases in both remote and infarct zones.28 Decreases in stress in response to the remodeling of fibrotic myocardium improves regional fractional shortening (a measure related to radial strain) of remote regions.29 Therefore, a regional improvement of myocardial properties in any segment can lead to better global contraction, smaller end-systolic volume, and therefore reduction of stresses and improvement in strains26 in all LV regions. The only regions in which improvement will not occur, as seen in our study, are regions with complete or near-complete fibrosis, with the strains being virtually nonexistent because of high-tensile properties of fibrous tissue and lack of contractile tissue.28
Moreover, this improved LV function was associated with higher survival rate in the GATA4 treatment group (Figure 4). With that said, we and others have demonstrated the potential for cell based therapies to alter the arrhythmogenic potential of the myocardium.30,31 The cardiac myocyte hypertrophy induced by GATA4:VP22, combined with the heterogeneity of cardiac myocytes size induced by local GATA4 delivery, could theoretically increase arrhythmogenic risk of the remodeled ventricle. Future studies will be required to determine whether local alterations in cardiac myocyte gene expression alters the arrhythmogenic potential of the infarct border zone.
As we know, the limited mitotic capacity of adult cardiomyocyte restricts the repair of the ischemic myocardium, leading to replacement by fibrotic tissue, which disrupts proper contractile function and results in decreased cardiac performance. Cell-based intercellular delivery of GATA4 might prevent the postinfarct remodeling and death rate by direct protection through antiapoptotic mechanisms (Figure 6C) and by preserving ventricular morphology and reducing fibronecrosis (Figure 3) and by inducing cardiac myosin overexpression in cardiomyocytes (Figure 6A and 6B) and myocardial hypertrophy in the infarct border zone (Figure 5). Because there was no functional improvement in the infarct zone, GATA4 might only function in the survival and remodeling of spared, but not dead myocardium.
Cell-penetrating peptides have been previously used to deliver a vast range of different biologically active proteins in different tissues.9,32 Our data provide the first evidence using cell-penetrating peptide to modulate remodeling in ischemic cardiomyopathy. As compared with other delivery vectors, such as electroporation and microinjection that are impractical to use in vivo or liposome encapsulation and receptor-mediated endocytosis that are limited by the lack of targeting and low yield of delivery, cell-based delivery of peptide-based chimeric proteins is, to date, the only method that succeeds in delivering a cargo without disturbing the plasma membrane and is applicable in vivo.7 From the analysis of survival rate at 6 weeks beyond the time an antibody response could have been mounted (Figure 4), we provide evidence that cell-based intercellular delivery could provide sustained release of functional proteins in the injured heart without causing toxic side effects.
In conclusion, we implemented a new carrier system that combines cell transplantation with cell-penetrating peptide VP22 to induce sustained local intercellular delivery of functional transcription factors such as GATA4 to the injured heart. This sustained delivery of GATA4 resulted in multiple changes in the myocardium, including alterations in cardiac fibrosis and local hypertrophy of cardiac myocytes, with the final result of improved cardiac remodeling and function.
Sources of Funding
This work was supported by NIH grant HL74400, the Shalom Foundation, the Wilson Foundation, and the State of Ohio.
M.S.P. and M.K. are named as coinventors on a pending patent filed by the Cleveland Clinic Foundation that relates to the use of cell penetrating peptides for modulation of LV remodeling.
Original received November 5, 2006; revision received April 10, 2007; accepted May 1, 2007.
Oka T, Maillet M, Watt AJ, Schwartz RJ, Aronow BJ, Duncan SA, Molkentin JD. Cardiac-specific deletion of Gata4 reveals its requirement for hypertrophy, compensation, and myocyte viability. Circ Res. 2006; 98: 837–845.
Bisping E, Ikeda S, Kong SW, Tarnavski O, Bodyak N, McMullen JR, Rajagopal S, Son JK, Ma Q, Springer Z, Kang PM, Izumo S, Pu WT. Gata4 is required for maintenance of postnatal cardiac function and protection from pressure overload-induced heart failure. Proc Natl Acad Sci U S A. 2006; 103: 14471–14476.
Kobayashi S, Lackey T, Huang Y, Bisping E, Pu WT, Boxer LM, Liang Q. Transcription factor gata4 regulates cardiac BCL2 gene expression in vitro and in vivo. FASEB J. 2006; 20: 800–802.
Askari A, Unzek S, Goldman CK, Kiedrowski M, Forudi F, Ellis SG, Thomas JD, Topol EJ, DiCorleto PE, Penn MS. Cellular, but not direct adenoviral delivery of VEGF results in improved LV function and neovascularization in dilated ischemic cardiomyopathy. J Am Coll Cardiol. 2004; 43: 1908–1914.
Molkentin JD. The zinc finger-containing transcription factors GATA-4, -5, and -6. Ubiquitously expressed regulators of tissue-specific gene expression. J Biol Chem. 2000; 275: 38949–38952.
Popovic ZB, Benejam C, Bian J, Mal N, Drinko J, Lee K, Forudi F, Reeg R, Greenberg NL, Thomas JD, Penn MS. Speckle tracking echocardiography correctly identifies segmental left ventricular dysfunction induced by scarring in a rat model of myocardial infarction. Am J Physiol Heart Circ Physiol. In press.
Rota M, Boni A, Urbanek K, Padin-Iruegas ME, Kajstura TJ, Fiore G, Kubo H, Sonnenblick EH, Musso E, Houser SR, Leri A, Sussman MA, Anversa P. Nuclear targeting of Akt enhances ventricular function and myocyte contractility. Circ Res. 2005; 97: 1332–1341.
Misao J, Hayakawa Y, Ohno M, Kato S, Fujiwara T, Fujiwara H. Expression of bcl-2 protein, an inhibitor of apoptosis, and Bax, an accelerator of apoptosis, in ventricular myocytes of human hearts with myocardial infarction. Circulation. 1996; 94: 1506–1512.
Liang Q, De Windt LJ, Witt SA, Kimball TR, Markham BE, Molkentin JD. The transcription factors GATA4 and GATA6 regulate cardiomyocyte hypertrophy in vitro and in vivo. J Biol Chem. 2001; 276: 30245–30253.
Grepin C, Nemer G, Nemer M. Enhanced cardiogenesis in embryonic stem cells overexpressing the GATA-4 transcription factor. Development. 1997; 124: 2387–2395.
Watt AJ, Battle MA, Li J, Duncan SA. GATA4 is essential for formation of the proepicardium and regulates cardiogenesis. Proc Natl Acad Sci U S A. 2004; 101: 12573–12578.
Molkentin JD, Lin Q, Duncan SA, Olson EN. Requirement of the transcription factor GATA4 for heart tube formation and ventral morphogenesis. Genes Dev. 1997; 11: 1061–1072.