SEARCH

SEARCH BY CITATION

Keywords:

  • Tbx5;
  • heart development;
  • Holt-Oram Syndrome;
  • CKB;
  • nppa;
  • gelsolin;
  • hey2;
  • prcad;
  • microarray;
  • atrial septum;
  • ventricular septum;
  • retina

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Tbx5 is a member of the T-box family of transcription factors and is associated with Holt–Oram syndrome (HOS), a congenital disorder characterized by heart and limb defects. Although implicated in several processes during development, only a few genes regulated by Tbx5 have been reported. To identify candidate genes regulated by Tbx5 during heart development, a microarray approach was used. A cardiac-derived mouse cell line (1H) was infected with adenoviruses expressing Tbx5 or β-galactosidase and RNA was isolated for analysis using an Affymetrix gene chip representing over 39,000 transcripts. Real-time reverse transcriptase-polymerase chain reaction confirmed Tbx5 induction of a subset of the genes, including nppa, photoreceptor cadherin, brain creatine kinase, hairy/enhancer-of-split related 2, and gelsolin. In situ hybridization analysis indicated overlapping expression of these genes with tbx5 in the embryonic mouse heart. In addition, the effect of HOS-associated mutations on the ability of Tbx5 to induce target gene expression was evaluated. Together, these data identify several genes induced by Tbx5 that are potentially important during cardiac development. These genes represent new candidate gene targets of Tbx5 that may be related to congenital heart malformations associated with HOS. Developmental Dynamics 235:2868–2880, 2006. © 2006 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Tbx5 belongs to the T-box family of transcription factors and is required for the embryonic development of the heart and forelimbs (Koshiba-Takeuchi et al.,2000; Bruneau et al.,2001). Mutations in TBX5 are associated with Holt–Oram syndrome (HOS), a congenital disorder characterized by limb malformations, cardiac septal defects, and conduction system anomalies (Basson et al.,1997; Li et al.,1997). Consequently, the function of Tbx5 in the formation of the heart and limbs has been the subject of several investigations (Logan,2003; Plageman and Yutzey,2005). During development, tbx5 is expressed in the embryonic heart and forelimbs, suggesting that altered Tbx5 function at these locations underlies abnormal development (Chapman et al.,1996). Because tbx5 encodes a transcription factor, mutations of the gene are presumed to disrupt development by altering the transcriptional function of Tbx5 protein, thereby affecting target gene expression. Strikingly, Tbx5 cooperatively regulates cardiac transcription with Nkx2.5 and Gata4, two proteins encoded by genes with mutations associated with atrial septal defects similar to what is observed in HOS (Basson et al.,1997; Li et al.,1997; Schott et al.,1998; Garg et al.,2003; Linhares et al.,2004; Plageman and Yutzey,2004). To understand how mutations in TBX5 cause developmental defects, it is important to know what genes Tbx5 regulates during embryogenesis.

Several investigations have examined the role of Tbx5 in various aspects of cardiac development (Plageman and Yutzey,2005). The expression of tbx5 throughout the cardiac primordia and ability to induce precocious beating in multipotent P19Cl6 cells implicates Tbx5 in the regulation of cardiomyocyte differentiation (Liberatore et al.,2000; Hiroi et al.,2001). Subsequent to the formation of the heart tube, tbx5 expression becomes restricted to the posterior regions that are fated to form the atria and left ventricle (Bruneau et al.,1999; Liberatore et al.,2000). Experiments eliminating or ectopically expressing tbx5 at these stages suggest a role for Tbx5 in cardiac chamber specification and morphogenesis (Liberatore et al.,2000; Bruneau et al.,2001). Investigation of Tbx5 function by genetic manipulation in mice has demonstrated a requirement for Tbx5 in atrial and ventricular septation, a process that is also disrupted in HOS (Bruneau et al.,2001; Takeuchi et al.,2003). Mice heterozygous null for tbx5 display atrioventricular block, a disruption in conduction system function that resembles what is observed in HOS (Bruneau et al.,2001; Moskowitz et al.,2004). In addition to processes related to cardiogenesis, embryonic limb development and retinal patterning are affected in loss or gain of function studies of Tbx5 (Koshiba-Takeuchi et al.,2000; Agarwal et al.,2003; Rallis et al.,2003). Taken together, it is clear that Tbx5 is an important regulator of many processes during embryonic development; however, the full complement of genes regulated by Tbx5 are only beginning to emerge.

A limited number of Tbx5 target genes have been identified through analysis of genetically altered mice with reduced Tbx5 function that exhibit HOS phenotypes. Tbx5 directly activates the promoters of connexin 40 (Cx40), natriuretic peptide precursor A (nppa), and fibroblast growth factor 10 (fgf10), and expression of these genes is significantly reduced in mice that lack tbx5 (Bruneau et al.,2001; Rallis et al.,2003). fgf10 is required for limb bud outgrowth, and reduced fgf10 expression in the absence of Tbx5 is associated with the abnormal development of the forelimbs in Tbx5 mutant mice (Rallis et al.,2003). Proper electrophysiological function of the heart is dependent on the gap junction protein Cx40 (Kirchhoff et al.,1998; Simon et al.,1998), and reduction in its expression is hypothesized to affect the conduction system in HOS (Bruneau et al.,2001; Moskowitz et al.,2004). Additionally, tbx5 null mice have alterations in atrial chamber development with concomitant reduction in expression of chamber specific genes, such as nppa (Bruneau et al.,2001). Although reduction in the expression of the genes fgf10, Cx40, and nppa may underlie the HOS conduction system and limb phenotypes, the downstream targets of Tbx5 during atrial and ventricular septal development have yet to be identified.

Cardiac cell lines have been used extensively to study cardiomyocyte differentiation and gene regulation (Delcarpio et al.,1991; Brunskill et al.,2001; Monzen et al.,2002). The 1H cell line is derived from transgenic embryonic mouse hearts expressing the SV-40 large T antigen under the control of the murine nkx2.5 cardiac promoter and expresses several early cardiac genes, including nkx2.5, gata4, dHand, and smooth muscle α-actin (Brunskill et al.,2001). Because the 1H cell line is cardiac derived and expresses transcription factors, such as Nkx2.5 and Gata4, which cooperatively regulate cardiac gene expression with Tbx5, it is an appropriate cellular environment for the examination of Tbx5 transcriptional targets. To identify genes induced by Tbx5, microarray analysis was conducted with RNA isolated from 1H cells infected with adenovirus expressing tbx5 (Ad-Tbx5) or β-galactosidase (Ad-β-gal). Several genes were identified that are induced by Tbx5 transcriptional activity and overlap with Tbx5 expression in the developing heart. These genes represent new potential targets of Tbx5 that may be related to congenital heart malformations associated with HOS.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Identification of Tbx5-Induced Genes by Microarray Analysis

An Affymetrix microarray screen was conducted with RNA isolated from the cardiac-derived 1H cell line expressing exogenous Tbx5 to identify genes induced by Tbx5. A recombinant adenovirus containing the entire murine tbx5 coding sequence was generated (Ad-Tbx5) for ectopic expression of Tbx5. Confirmation of adenovirus-mediated Tbx5 protein expression was achieved by Western blot analysis of NIH 3T3 cells infected with Ad-Tbx5 or a β-galactosidase expressing adenovirus (Ad-β-gal) as a negative control (Fig. 1A). Tbx5-specific polyclonal antiserum was used to detect an approximately 57.8-kDa protein, the expected size of Tbx5 (Fig. 1A). To determine whether introduced Tbx5 activates expression of known Tbx5 target genes, RT-PCR was performed on RNA isolated from 1H cells infected with Ad-Tbx5 or Ad-β-gal (Fig. 1B). In these studies, expression of nppa was increased approximately fivefold in response to exogenous Tbx5 (Fig. 1B), demonstrating that 1H cells activate known Tbx5 target genes when Tbx5 is overexpressed. Therefore, this system is appropriate for further identification of Tbx5 transcriptional targets.

thumbnail image

Figure 1. Exogenous Tbx5 induces the expression of nppa in the 1H cell line. A: Western blot of NIH 3T3 cell lysates 48 hr after infection with β-galactosidase (Ad-β- gal) or Tbx5 (Ad-Tbx5) expressing adenoviruses. Use of the Tbx5-specific polyclonal antisera reveals a band approximately 57.8 kDa, the expected size of the Tbx5 protein. B: Reverse transcriptase-polymerase chain reaction (RT-PCR) analysis of two independent samples of 1H cell RNA 48 hr after infection with Ad-β-gal or Ad-Tbx5 using primers specific to β-galactosidase, tbx5, nppa, and gapdh.

Download figure to PowerPoint

To identify transcripts induced by Tbx5, RNA was isolated from 1H cells 48 hr after infection with Ad-Tbx5 or Ad-β-gal and hybridized to the Affymetrix 430 2.0 mouse genomic array. The microarray analysis was performed on RNA isolated from two independent Ad-Tbx5 infected groups and two Ad-β-gal infected groups. The detected fluorescence was converted into numeric values by the Robust Multichip Average (RMA) algorithm followed by normalization to the average fluorescence for each array (Bolstad et al.,2003; Boes and Neuhauser,2005). The average expression level for each transcript was calculated from the biological duplicates, and the average fold change between the Ad-β-gal and Ad-Tbx5 groups was determined. Transcripts in the Ad-Tbx5–infected group that were significantly increased greater than 1.3-fold over the Ad-β-gal–infected group (P < 0.05) and have reported cardiac expression are listed in Table 1. A complete list of genes induced 1.2-fold or greater by Tbx5 can be found in the data supplement.

Table 1. Tbx5-Induced Genes With Reported Cardiac Expression
Gene nameGene symbolFold changeGenbank #Cellular functionCardiac function/expression
nonmuscle myosin II-AMyh95.84NM_022410Cytoskeleton organization/cell adhesionMouse embryonic cardiac expression (Murakami et al.,1993)
sentrin specific peptidase 2Senp23.87NM_029457Proteolysis and peptidolysisMouse adult cardiac expression (Best et al.,2002)
phospholipase A2, group IIEPla2g2e3.80NM_012044Lipid catabolismMouse adult infarcted heart expression (Masuda et al.,2005)
natriuretic peptide precursor ANppa3.64NM_008725HormoneChamber myocardium marker, mouse embryonic cardiac expression (Claycomb,1988)
photoreceptor cadherinprCAD3.52NM_130878Cell adhesionPhotoreceptor survival (Rattner et al.,2001)
creatine kinase, brainCKB3.03NM_021273Creatine kinase activityEnergy metabolism, mouse embryonic cardiac expression (Wessels et al.,2000)
mesoderm specific transcriptMest2.68NM_008590UnknownTrabeculation, mouse embryonic cardiac expression (King et al.,2002)
zinc finger, CCHC domain containing 5Zcchc52.43NM_199468Putative transcription factorMouse embryonic cardiac expression (Brandt et al.,2005)
Notch-regulated ankyrin repeat proteinNrarp2.11NM_025980Cell signalingMouse, human adult cardiac expression (Krebs et al.,2001)
ScleraxisScx2.00NM_198885Transcription factorTendon development, cardiac valve marker, mouse embryonic expression (Cserjesi et al.,1995)
cyclin-dependent kinase inhibitor 1Cp57(kip2)1.96NM_009876Cell cycleCell proliferation inhibitor, mouse embryonic expression (Kochilas et al.,1999)
hey2Hey21.94NM_013904Transcription factorVentricular septation, mouse embryonic ventricular expression (Kokubo et al.,1999; Leimeister et al.,1999)
formin-family protein FHOS2FHOS21.92NM_175276Cytoskeleton organizationMouse adult cardiac expression (Kanaya et al.,2005)
prenylcysteine oxidase 1Pcyox11.90NM_025823Prenylcysteine catabolismMouse adult cardiac expression (Beigneux et al.,2002)
clusterinClu1.85NM_013492Cell deathMouse embryonic cardiac valve expression (Witte et al.,1994)
Wilms tumor homologWt11.77NM_144783Transcription factorEpicardial development, mouse embryonic cardiac expression (Carmona et al.,2001)
gelsolinGsn1.76NM_146120Actin filament severingMouse embryonic cardiac expression (Arai and Kwiatkowski,1999)
Muscleblind-like 1Mbnl11.74NM_020007Alternative splicingMouse embryonic cardiac expression (Kanadia et al.,2003)
popeye domain containing 2Popdc21.73NM_022318Cell adhesionMouse embryonic cardiac expression (Andree et al.,2000)
solute carrier family 5, member 1Slc5a11.71NM_019810Glucose/sodium ion transportHuman adult cardiac expression (Zhou et al.,2003)
inhibin beta-BInhbb1.68NM_008381Cell signalingMouse embryonic cardiac expression (Moore et al.,1998)
bone morphogenetic protein 4Bmp41.65NM_007554Cell signalingCardiac cushion development, mouse embryonic cardiac expression (Keyes et al.,2003)
Zinc finger protein, multitype 1Zfpm1 (Fog1)1.58NM_009569Transcription factorHeart looping, mouse cardiac cushion expression (Gitler et al.,2003)
suppressor of cytokine signaling 2Socs21.58NM_007706Cell signalingHuman adult cardiac expression (Dey et al.,1998)
transforming growth factor, beta receptor IITgfbr21.55NM_009371Cell signalingMouse adult cardiac expression (Lawler et al.,1994)
actin, alpha 1, skeletal muscleActa11.48NM_009606Cytoskeleton organizationMuscle contraction mouse embryonic cardiac expression (Sassoon et al.,1988)
procollagen, type VI, alpha 2Col6a21.47NM_146007Cell adhesionMouse embryonic cardiac valve expression (Marvulli et al.,1996)
titinTtn1.43NM_011652Muscle contractionMouse embryonic heart expression (Schaart et al.,1989)
nuclear factor of activated T-cells, cytoplasmic, calcineurin-dependent 1Nfatc11.41NM_016791Transcription factorMouse embryonic cardiac valve expression (de la Pompa et al.,1998)
odd-skipped related 1Osr11.40NM_011859Putative transcription factorAtrial septation/venous valve development, mouse embryonic cardiac expression (Wang et al.,2005)
Troponin T2, cardiacTnnt21.32NM_011619Muscle contractionMouse embryonic cardiac expression (Wang et al.,2001)
Limb-bud and heartLbh1.31NM_029999Putative transcription factorMouse embryonic cardiac expression (Briegel and Joyner,2001)
bone morphogenetic protein 6Bmp61.30NM_007556Cell signalingEndocardial cushion development, mouse embryonic cardiac expression (Kim et al.,2001)

As expected, one of the genes with the largest fold change was nppa (3.64-fold), demonstrating that the microarray screen can detect changes in known downstream targets of Tbx5 (Table 1). Among the genes listed in Table 1, are those implicated in many aspects of heart development that have also been associated with Tbx5 and are involved in cardiac septation (hey2, odd-skipped related 1), proliferation inhibition (p57kip2), myogenic differentiation (skeletal alpha-actin, titin, cardiac troponin T2), and epicardial development (WT1). Several of the genes also have reported function or expression in regions outside of the heart where Tbx5 is expressed, such as the forelimb and retina (Table 1; Koshiba-Takeuchi et al.,2000; Agarwal et al.,2003; Rallis et al.,2003). In this report, the genes nppa, phospholipase A2 group IIE (pla2g2e), photoreceptor cadherin (prCAD), brain creatine kinase (CKB), hairy/enhancer-of-split related 2 (hey2), and gelsolin were chosen for further study based on previously reported cardiac function or expression and relatively large statistically significant fold changes detected in the microarray.

Confirmation of Microarray Data by Quantitative PCR

Real-time RT-PCR was performed on RNA isolated from Ad-Tbx5– or Ad-β-gal–infected 1H cells used in the microarray analysis to verify whether expression is induced in response to Tbx5 (Fig. 2A). The fold increase in expression of nppa, pla2g2e, prCAD, CKB, hey2, and gelsolin, determined by quantitative RT-PCR (qRT-PCR; Fig. 2A, black bars) is comparable to the microarray data (white bars, Fig. 2A). No increase in the expression of gapdh was observed in the Ad-Tbx5–infected group relative to the Ad-β-gal–infected group by microarray or qRT-PCR, indicating that there was not a general increase of transcription in response to exogenous Tbx5. In addition, the real-time RT-PCR did not indicate an increase in the top two microarray targets: non-muscle myosin II-A and sentrin specific peptidase 2 (data not shown). To confirm that the putative target gene induction by Tbx5 occurs in normal fetal heart muscle cells, primary cardiomyocytes isolated from E14.5 mouse embryos were placed in culture and infected with Ad-β-gal or Ad-Tbx5. Using the same primers as Figure 2A, real-time RT-PCR was performed on RNA isolated from the cardiomyocyte cultures 48 hr postinfection (Fig. 2B). Expression of nppa (4.99-fold), pla2g2e (6.22-fold), prCAD (7.59-fold), CKB (2.87-fold), hey2 (2.77-fold), and gelsolin (1.36-fold) all increased in response to exogenous Tbx5. Of interest, the fold increase was greater in the cardiomyocytes than 1H cells, especially for prCAD (∼124-fold) and CKB (∼18-fold). The large fold activation is likely due to low basal expression of these genes in the fetal cardiac myocytes. Cardiomyocyte gapdh expression was not affected with the ectopic expression of Tbx5. These data demonstrate that Tbx5 induces the expression of nppa, pla2g2e, prCAD, CKB, and hey2 in 1H cells and in mouse embryonic cardiomyocytes.

thumbnail image

Figure 2. Quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR) confirmation of microarray results. A: Real-time qRT-PCR was performed on RNA used in the microarray analysis using primers specific for nppa, pla2g2e, prCAD, CKB, hey2, gelsolin, and gapdh. The average fold increase of the 1H cells infected with adenovirus expressing tbx5 (Ad-Tbx5) group over the 1H cells infected with adenovirus expressing β-galactosidase (Ad-β-gal) group is depicted for the microarray data (white bars, n = 2) and the qRT-PCR analysis (black bars, n = 4). B: Real-time qRT-PCR was performed on RNA isolated from embryonic ventricular mouse cardiomyocytes after a 48-hr infection with Ad-β-gal or Ad-Tbx5. The average fold-increase of the Ad-Tbx5 group over the Ad-β-gal group is depicted (n = 4). Expression levels for qRT-PCR were normalized to L7, and error bars represent the standard error.

Download figure to PowerPoint

Expression Analysis of Tbx5-Induced Genes

To determine whether the genes induced by Tbx5 in cultured cells are expressed in the developing heart, RT-PCR was performed on RNA isolated from the atria, ventricles, forelimb, hindlimb, liver, and stomach of E14.5 mouse embryos. The genes analyzed include tbx5, prCAD, pla2g2e, CKB, hey2, gelsolin, and L7 as a loading control (Fig. 3). As anticipated, relatively high tbx5 expression is detected in the atria, lung, and forelimb and weak expression is detected in the ventricles (Fig. 3). Although previous reports suggested prCAD expression is specific to the developing retina, prCAD expression was detected in the atria at E14.5 (Fig. 3; Rattner et al.,2001). Expression of pla2g2e, CKB, hey2, and gelsolin also was enriched in the heart when compared with the limbs, lungs, liver, and stomach (Fig. 3). CKB and gelsolin expression is found in both the atria and ventricles, whereas the expression of pla2g2e is detected in the atria and hey2 in the ventricles. These data demonstrate that the genes prCAD, pla2g2e, CKB, hey2, and gelsolin all are expressed in the embryonic heart with tbx5.

thumbnail image

Figure 3. Cardiac expression of Tbx5-induced genes in embryonic mice. Reverse transcriptase-polymerase chain reaction (RT-PCR) was performed on equivalent amounts of RNA isolated from the developing heart, limbs, lung, liver, and stomach at embryonic day (E) 14.5 using primers specific for tbx5, pla2g2e, prCAD, CKB, hey2, gelsolin, and L7. PCR reactions were terminated at various cycles consistent with the linear range of amplification and the reaction products were separated by electrophoresis and visualized by ethidium bromide.

Download figure to PowerPoint

Further expression analysis of the Tbx5-induced genes was performed by in situ hybridization on cryosectioned embryonic hearts. The expression patterns of each gene were compared with tbx5 to identify regions of overlapping expression at E10.5, a critical time point in heart chamber morphogenesis and early septal development (Fig. 4). tbx5 is expressed in the atrial and left ventricular myocardium and is absent in the outflow tract myocardium, as has been previously reported (Fig. 4A; Bruneau et al.,1999; Liberatore et al.,2000). CKB (Fig. 4B) and gelsolin (Fig. 4D) are expressed throughout the myocardium and overlap with tbx5 in the myocardium of the atria and left ventricle, but not the outflow tract. hey2 expression (Fig. 4C) is detected in the endothelial cells of the atrioventricular and outflow tract cushions and in the compact myocardium of the left ventricle, where it overlaps with tbx5. Although pla2g2e and prCAD expression was detected by RT-PCR in the atria at embryonic day (E) 14.5, expression of these genes was not detected in the atrial myocardium by in situ hybridization at E10.5 (data not shown). Because of the proposed role of Tbx5 in septal development, expression was also evaluated at the onset of atrial and ventricular septation at E10.5. Strong expression of tbx5 is detected in the myocardium of the atrial (Fig. 4E) and ventricular septum (Fig. 4I). The atrial septum expresses CKB (Fig. 4F), gelsolin (Fig. 4H), and lacks any detectable hey2 expression (Fig. 4G). Expression of CKB (Fig. 4J), hey2 (Fig. 4K), and gelsolin (Fig. 4L) are all detected in the ventricular septum. Overlapping expression of tbx5 with CKB, hey2, and gelsolin in the developing septa is consistent with a role for Tbx5 inducing the expression of these genes during heart chamber formation.

thumbnail image

Figure 4. Overlapping expression of tbx5 with Tbx5-induced genes detected in the microarray in early septum development. AL: In situ hybridization was performed on cryosections of embryonic day (E) 10.5 mouse hearts using probes specific for tbx5 (A,E,I), CKB (B,F,J), hey2 (C,G,K), and gelsolin (D,H,L). Cardiac expression patterns for each gene are visualized in the whole heart (A–D), in the primary atrial septum (PAS; E–H), and ventricular septum (VS; I–L). OFT, outflow tract; RA, right atria; LA, left atria; RV, right ventricle; LV, left ventricle.

Download figure to PowerPoint

In previous studies, murine expression of prCAD was reported in the developing photoreceptors of the postnatal retina; however, embryonic retinal expression was not evaluated (Rattner et al.,2001). Because tbx5 is expressed in the embryonic retina (Gibson-Brown et al.,1998a,b), overlapping expression with prCAD was investigated by in situ hybridization of cryosections of the developing eye. At E14.5, tbx5 and prcad expression was detected in a subset of cells of the outer neuroblastic zone of the retinal margin (arrowheads in Fig. 5C,D). These data are consistent with Tbx5 inducing the expression of prCAD in the outer neuroblastic zone of the retina. The induction of prCAD by Tbx5 in 1H cells and cardiomyocytes and the detection of prCAD transcript by RT-PCR in E14.5 atria (Fig. 3) prompted further investigation of cardiac prCAD expression by in situ hybidization. Expression of tbx5 at E14.5 is strong throughout the atrial myocardium, including the atrial septum and venous valves (Fig. 5E,G). prCAD expression is absent in most of the atrial myocardium (Fig. 5F); however, it is detected in the venous valves and in the atrial septum (Fig. 5H). Overlapping expression of prCAD in the heart and eye is consistent with Tbx5 induction of its expression in both structures and may indicate a yet uncharacterized role for prCAD in venous valve development or atrial septation.

thumbnail image

Figure 5. Overlapping expression patterns of tbx5 and prCAD in the heart and eye at embryonic day (E) 14.5. A–H: In situ hybridization was performed for tbx5 (A,C,E,G) and prCAD (B,D,F,H) on cryosections of E14.5 mouse embryos. AD: Eye expression of tbx5 and prCAD overlaps in the outermost cells of the outer neuroblastic zone (ONZ) of the developing retina (R; arrowheads in C,D). The insets in A and B are shown at a higher magnification in C and D. E,F: Cardiac expression of tbx5 and prCAD overlaps in the right and left venous valves (RVV, LVV). G,H: Expression of prCAD in the primary atrial septum (PAS; H) overlaps with expression of tbx5 (G). RA, right atria; LA, left atria; RV, right ventricle; LV, left ventricle; L, lens.

Download figure to PowerPoint

Tbx5-Induction Is Compromised by HOS-Associated Mutations

Mutations in the human TBX5 protein coding sequence are associated with the heart defects found in patients with HOS. Several reports have described that HOS-associated Tbx5 mutant proteins have deficient transcriptional regulatory activity in vitro (Ghosh et al.,2001; Hiroi et al.,2001; Fan et al.,2003a,b; Plageman and Yutzey,2004). To determine whether these mutations compromise the ability of Tbx5 to induce genes identified in the mircroarray, adenoviruses were generated expressing the HOS-associated mutant alleles Tbx5(R237Q) and Tbx5(R279ter) and used to infect 1H cells. The R237Q mutation affects Tbx5 transcriptional activity by disrupting DNA binding and inhibiting its association with Nkx2.5 (Fan et al.,2003b). The R279ter mutation is missing a strong activation domain due to a premature stop codon and cannot activate transcription in vitro (Plageman and Yutzey,2004). RT-PCR performed on RNA isolated from infected 1H cells demonstrated that exogenous tbx5 expression from the recombinant adenovirus is similar among the wild-type or mutant Tbx5 adenovirus (Fig. 6A). Expression levels of Tbx5-induced genes were quantified using real-time RT-PCR and normalized to L7. The average fold increase from the Ad-Tbx5–, Ad-Tbx5(R237Q)–, or Ad-Tbx5(R279ter)–infected groups over the Ad-β-gal–infected group was calculated from four to six independent samples (Fig. 6B). A significant decrease in the induction of nppa, CKB, and hey2 was observed when comparing either the Ad-Tbx5(R237Q)– or Ad-Tbx5(R279ter)–infected cells with the Ad-Tbx5–infected cells, whereas induction of pla2g2e and prCAD were only significantly decreased in the Ad-Tbx5(R279ter)–infected cells (Fig. 6B). The induction of gelsolin was not significantly altered by either mutant allele (Fig. 6B). These data demonstrate that HOS-associated mutations decrease the ability of Tbx5 to induce expression of selected genes and that different HOS-associated mutations have variable effects on the induction of individual genes in 1H cell cultures.

thumbnail image

Figure 6. Holt–Oram syndrome (HOS) associated mutant Tbx5 proteins are deficient in their ability to induce gene expression. A: Reverse transcriptase-polymerase chain reaction (RT-PCR) was performed on RNA isolated from 1H cells 48 hr postinfection with Ad-β-gal, Ad-Tbx5, or either of the HOS mutant alleles (Ad-Tbx5(R237Q), Ad-Tbx5(R279ter)) using primers specific to tbx5 and L7. B: Real-time quantitative RT-PCR (qRT-PCR) was performed on the RNA used in A, using primers specific to nppa, pla2g2e, prCAD, CKB, hey2, and gelsolin. Expression levels were normalized to L7, and the average fold increase over Ad-β-gal was calculated (n = 4). Asterisks represent fold changes that are significantly different than the wild-type (P < 0.05).

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Tbx5 is implicated in several processes during heart development, but the complete set of genes regulated by Tbx5 during cardiac differentiation and chamber formation is still not fully known. To identify genes regulated by Tbx5 during heart development, microarray analysis was performed on RNA isolated from cardiac-derived 1H cells with ectopic expression of Tbx5. Several Tbx5-induced genes were identified, many of which are expressed and function during heart development. Real-time RT-PCR confirmed that Tbx5 induces the expression of nppa, pla2g2e, prCAD, CKB, hey2, and gelsolin in 1H cells as well as in E14.5 cardiomyocytes. In situ hybridization of mouse hearts demonstrated that expression of CKB, hey2, and gelsolin overlaps with tbx5 in the atrial and/or ventricular septum at the onset of septation. Overlapping expression of tbx5 with prCAD, a gene previously reported to be expressed exclusively in the retina, was also detected in the venous valves and atrial septum. When the ability of Tbx5 and HOS-associated mutant alleles to induce expression was compared, it was observed that mutant Tbx5 could not induce expression to the same level as wild-type. Although several genes induced by Tbx5 were identified, it remains to be determined if expression is regulated directly by Tbx5 or if intermediate genes are necessary for induction. Together these analyses identify several cardiac expressing genes induced by Tbx5 that are candidate downstream targets of Tbx5 in heart development and may be related to the congenital heart defects of HOS.

The microarray screen was used to identify several genes that are expressed in the same regions or function in the same processes during cardiac development as Tbx5. A role for Tbx5 in cardiomyocyte differentiation is supported by its early expression throughout the cardiac primordia before differentiation and lack of heart formation in transgenic frogs expressing a dominant-negative form of Tbx5 (Horb and Thomsen,1999; Liberatore et al.,2000). Overexpression of Tbx5 in P19Cl6 cells, a derivative of multipotential P19 cells partially committed to the cardiac lineage, accelerates cardiac differentiation by promoting precocious beating and inducing the expression of early cardiac markers (Hiroi et al.,2001; Fijnvandraat et al.,2003). The microarray screen reported here indicates that the contractile protein genes skeletal alpha-actin, titin, and cardiac troponin T2 are slightly induced by Tbx5 in 1H cells (1.48-, 1.43-, 1.32-fold, respectively), consistent with previous observations that Tbx5 promotes differentiation of the cardiac lineage by regulating the expression of contractile protein genes (Table 1). Tbx5 also induces the expression of CKB (Fig. 2), a gene involved in energy metabolism expressed throughout the myocardium at E8.5 (Lyons et al.,1991; Shen et al.,2003). CKB is one of the genes most affected by Tbx5 in the microarray analysis of 1H cells and is dramatically induced in primary E14.5 cardiomyocytes, consistent with Tbx5 regulation of CKB during myogenesis. Tbx5 is also implicated in the inhibition of cardiomyocyte proliferation and proepicardial migration and induces the expression of the cell-cycle arrest gene p57kip2 and the epicardial marker WT1 (Kochilas et al.,1999; Moore et al.,1999; Hatcher et al.,2001,2004). Tbx5 induction of these genes may contribute to reported functions of Tbx5 in differentiation, proliferation, and epicardial development.

The atrial and ventricular septal defects associated with HOS and the phenotype of tbx5 heterozygous mutant mice are consistent with a critical role for Tbx5 in cardiac septation (Bruneau et al.,1999). The restricted expression of tbx5 in the left but not right ventricles is hypothesized to control the position of the ventricular septum (Takeuchi et al.,2003). Although it has been demonstrated that Tbx5 is required for normal septation, the downstream targets regulated by Tbx5 during septation are not fully characterized. Previous studies indicate that Tbx5 functions in the regulation of SALL4 and MYH6, two genes with reported mutations associated with cardiac septal defects in the human population (Al-Baradie et al.,2002; Kohlhase et al.,2002; Ching et al.,2005; Koshiba-Takeuchi et al.,2006). However, it is unclear whether Tbx5 can directly regulate the expression of these genes in the developing heart in vivo. The microarray data listed in Table 1 includes two Tbx5-induced genes, hey2, and odd-skipped related 1, that are also associated with septal defects (Donovan et al.,2002; Gessler et al.,2002; Sakata et al.,2002; Wang et al.,2005). Loss of either gene in mice results in atrial or ventricular septal defects similar to those found in tbx5+/− mice (Donovan et al.,2002; Gessler et al.,2002; Sakata et al.,2002; Wang et al.,2005). Real-time RT-PCR confirmed that Tbx5 induces hey2 in both 1H cells and ventricular cardiomyocytes (Fig. 2), and in situ hybridization demonstrated that expression of tbx5 and hey2 overlap at the onset of ventricular septation (E10.5; Fig. 4). Additionally, expression of hey2 is dramatically reduced in the hearts of tbx5 null mice (Bruneau et al.,2001). Collectively, these data suggest that the ventricular septal defects found in hey2 and tbx5 mutant mice are caused by disruptions in the same transcriptional pathway.

Tbx5 is implicated not only in heart development, but also in the morphogenesis of the eye. Normally, Tbx5 is expressed in the dorsal aspect of the optic cup, but misexpression of tbx5 in the ventral side alters retinal patterning and axon projection (Koshiba-Takeuchi et al.,2000). Microarray analysis and real-time RT-PCR demonstrate that Tbx5 induces the expression of prCAD, a gene that is crucial to the structural integrity of the outer segment of the adult retina (Rattner et al.,2001). It was previously reported that postnatally prCAD is expressed exclusively in the outer nuclear layer of the retina. However, expression in the embryonic mouse retina was not determined (Rattner et al.,2001). In this report, in situ hybridization revealed expression of prCAD in the outer neuroblastic zone of the retina at E14.5 (Fig. 5B,D), and this expression overlaps with tbx5 (Fig. 5A,C). Although vision is not impaired in HOS, older patients display alterations in the scotopic b-wave, an electrophysiological measurement used to assess retinal function (Gruenauer-Kloevekorn et al.,2005). Of interest, prCAD mutant mice also had abnormal scotopic b-waves, suggesting that these alterations in retinal electrophysiology might be related to reduced prCAD expression associated with Tbx5 loss of function (Rattner et al.,2001). Surprisingly, prCAD transcript was also detected in the embryonic heart by RT-PCR and in situ hybridization analysis. The majority of the myocardium lacks prCAD expression however, expression was detected specifically in the venous valves and atrial septum (Fig. 5). In support of this observation, an online expression pattern database of mouse embryos (http://www.genepaint.org) also reveals the expression of prCAD in the venous valves (Visel et al.,2004). Expression of prCAD specifically in the venous valves and atrial septum is similar to what is observed for odd-skipped related 1, a gene that is implicated in the development of the atrial septum, which also was among the Tbx5-induced genes (Table 1; Wang et al.,2005). Although there were no reported cardiac defects in the prCAD mutant mice, atrial septal defects can be asymptomatic in the human population and may have been overlooked in the mutant mice (Rattner et al.,2001; Andrews et al.,2002). The induction of prCAD by Tbx5 and their overlapping embryonic expression patterns are consistent with Tbx5 regulation of prCAD during the development of the eye and heart.

HOS-associated mutations affect transcriptional function of Tbx5 by disrupting DNA binding, preventing association with other transcription factors, or by eliminating functional domains (Fan et al.,2003b; Plageman and Yutzey,2004). For example, the missense mutation Tbx5(R237Q) inhibits DNA binding and association with Nkx2.5, whereas the nonsense mutation Tbx5(R279ter) prematurely stops translation and causes Tbx5 to lack a transcriptional activation domain (Fan et al.,2003b; Plageman and Yutzey,2004). Previously, we demonstrated that these HOS-associated mutations decrease the ability of Tbx5 to activate the nppa promoter in vitro (Plageman and Yutzey,2004). In this report, we compare the ability of wild-type or mutant Tbx5 to induce endogenous gene expression in vivo (Fig. 6B). Both the R237Q and R279ter mutation decreased the ability of Tbx5 to induce expression of nppa, CKB, and hey2, whereas induction of prCAD, pla2g2e, and gelsolin by Tbx5 was reduced by the R279ter nonsense mutation, but was not significantly affected by the R237Q missense mutation. The decrease in gene expression induction suggests that the developmental defects in HOS could be caused by insufficiency of the expression of Tbx5-induced genes, including nppa, CKB, prCAD, and hey2. Although conflicting reports exist regarding whether the severity of the HOS phenotype can be correlated with a specific mutation, our data suggest that the R237Q mutation, and the R279ter mutation vary in their ability to disrupt Tbx5 transcriptional function affecting multiple target genes (Basson et al.,1999; Brassington et al.,2003). Consequently, distinct phenotypes resulting from either mutation may occur because expression of the transcriptional targets of Tbx5 is variably affected.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Recombinant Adenovirus Construction and Infection

The Ad-Tbx5, Ad-Tbx5(R237Q), and Ad-Tbx5(R279ter) adenoviruses were generated from the expression plamids pAC-CMV-Tbx5, pAC-CMV-Tbx5(R237Q) and pAC-CMV-Tbx5(R279ter) (Plageman and Yutzey,2004) using previously described methods (Gomez-Foix et al.,1992). The mouse cardiac 1H cell line (Brunskill et al.,2001) was propagated in Dulbecco's modified Eagle's medium (DMEM; Cellgro), supplemented with 10% fetal bovine serum (Hyclone), 100 units/ml penicillin/streptomycin (Invitrogen), and 1% chick embryo extract (Sera Laboratories International, Ltd). When the cells reached approximately 70% confluency, the medium was removed and replaced with DMEM containing 1.85 × 107 plaque forming units/ml, and 100 units/ml penicillin/streptomycin. This adenovirus concentration is sufficient to infect >90% of the 1H cells as determined by staining with 1 mg/ml X-gal (Amresco) after infection with Ad-β-gal. After 90 min, the medium was replaced and cells were incubated for 48 hr. Subsequently, RNA was extracted using the TRIzol reagent (Invitrogen) and BCP phase separation reagent (Molecular Research Center, Inc.) and precipitated overnight at −20° with isopropanol. Mouse E14.5 primary cardiomyocytes were isolated and cultured using a previously described protocol (Evans et al.,2003). RNA was extracted 48 hr after infection in the same manner as for 1H cells.

Western Blot Analysis

Protein lysate preparation and Western blot analysis was performed as previously described in (Parsons et al.,2004). Briefly, protein was isolated from NIH 3T3 cells 48 hr after infection with Ad-β-gal or Ad-Tbx5. The blot was incubated with a 1:1,000 dilution of a custom synthesized, affinity purified rabbit α-Tbx5 antibody raised to the peptide corresponding to amino acids 10–23 of the murine Tbx5 protein (Zymed Laboratories, Inc.). The blot was subsequently incubated with a 1:1,000 dilution of a goat α-rabbit alkaline phosphatase (AP) antibody (Santa Cruz Biotechnology, Inc.) and visualized using chemifluorescence as described in (Parsons et al.,2004).

Microarray Hybridization and Analysis

Two independent RNA samples were isolated from 1H cells infected with either Ad-β-gal or Ad-Tbx5, as described above, and submitted to the Cincinnati Children's Hospital Research Foundation Affymetrix Microarray Core. A total of 10 μg of RNA from each group was used to make double-stranded cDNA using the One-Cycle cDNA Synthesis Kit (Affymetrix). The cDNA was subsequently used to synthesize a biotin-labeled cRNA using the IVT Labeling Kit (Affymetrix). The cRNA was chemically fragmented and hybridized to the Mouse Genome 430 2.0 Array (Affymetrix) using standard protocols. Arrays were washed and stained using the Fluidics Station 450 (Affymetrix) and scanned using the GeneChip Scanner 3000 (Affymetrix). The “*.cel” files were loaded into GeneSpring Gx 7.3 software (Silicon Genetics) and subjected to Robust Multichip Average (RMA) analysis (Bolstad et al.,2003; Boes and Neuhauser,2005). The data were normalized to the average fluorescent value emitted from each microarray, and the value from each gene was averaged from biological duplicates. The fold change was calculated and Student's t-test was performed to determine statistically significant differences in gene expression between Ad-Tbx5 and Ad-β-gal infected 1H cells. A complete list of genes significantly induced by greater than 1.2-fold (P < 0.05) was placed in a table in the data supplement.

Real-time qRT-PCR

cDNA was generated with oligo(dT) primers and the SuperScript first-strand synthesis kit (Invitrogen) from 1 μg of RNA. Quantitative real-time RT-PCR was performed as described previously (Plageman and Yutzey,2004). The oligonucleotide sequences for each PCR primer pair used and the resulting product size are as follows: tbx5 5′-CAAACTCACCAACAACCACC-3′ 3and 5′-GTGATTCTGGTACGAAGTC-3′ (187 bp), β-galactosidase 5′-GTCACACTACGTCTGAACGT-3′ and 5′-CTGCACCATTCGCGTTACG-3′ (471 bp), nppa 5′- TGCCGGTAGAAGATGAGGTC-3′ and 5′-AGCAGCTGGATCTTCGTAGG-3′ (240 bp), pla2g2e 5′-CCAGTGGACGAGACGGATTG-3′ and 5′-AGCTCTCTTGTCACACTC-3′ (177 bp), prCAD 5′-TGACCCCAGCACTAGAAGTGT-3′ and 5′-CAGTATCACGACTTTCTCTGCC-3′ (154 bp), CKB 5′-AGTTCCCTGATCTGAGCAGC-3′ and 5′-GAATGGCGTCGTCCAAAGTAA-3′ (117 bp), hey2 5′-CCAGCGGGAGGCAGCAGTG-3′ and 5′-GTGGATAGGCGACATGGGGTTGAC-3′ (156 bp), gelsolin 5′-TCACGGGTGATGCCTATGTCA-3′ and 5′-TAGTCATCCAGTTGCACAGTAAAG-3′ (145 bp), gapdh 5′-AAGGTCGGTGTGAACGCATT-3′ and 5′-GAAGATGGTGATGGGCTTCC-3′ (219 bp), L7 5′-GAAGCTCATCTATGAGAAGGC-3′ and 5′-AAGACGAAGGAGCTGCAGAAC-3′ (202 bp). The conditions for the PCR reactions were as follows: 94°, 30 sec; 61°, 30 sec; 72°, 30 sec. Gene expression levels were quantified based on the threshold cycle (C(t)) calibrated to a standard curve generated using 1H cell cDNA and normalized to L7. In PCR reactions visualized by ethidium bromide staining, the number of cycles was optimized to determine the presence or absence of an amplified product of appropriate size.

Probe Generation and In Situ Hybridization

Digoxigenin UTP-labeled antisense RNA probes were generated from pGEMT-vector plasmids containing nucleotide sequences from mouse tbx5, CKB, hey2, gelsolin, and prCAD. The plasmid used to generate the tbx5 riboprobe was previously described (Chapman et al.,1996). RT-PCR was performed on E14.5 heart RNA using the following primers: CKB, 5′-GTGGCGGGCGACGAGGAGAGTTAC-3′ and 5′-CCAGGTGGGGCAGCTTGATGTG-3′; hey2, 5′-GACAACTACCTCTCAGATTATGGC-3′ and 5′-CGGGAGCATGGGAAAAGC-3′; gelsolin, 5′-GCGCGGCCCAGCACTAT-3′ and 5′-AAGGGGTTCTCATCAGCCACT-3′; and prCAD, 5′-CCCCTGGCCACAAAGTAACA-3′ and, 5′-AATCCCAAGAAGAGAAAAACAGTC-3′. The resulting PCR fragments were cloned into pGEMT-Vector (Promega). Plasmids were verified by sequencing, linearized, and used to generate riboprobe with T7 polymerase (tbx5, CKB, hey2, and prCAD) or Sp6 polymerase (gelsolin). E10.5 and E14.5 mouse embryos were isolated and fixed overnight at 4°C in 4% paraformaldehyde/phosphate buffered saline followed by infiltration with sucrose and embedded in OCT (Sakura; Stern,1993). Fourteen micrometer frozen sections were placed on Superfrost/Plus slides (Fisherbrand), and in situ hybridization was performed as previously described in (Toresson et al.,1999). BM Purple AP substrate (Roche) was incubated with sections for 16–96 hr to visualize the hybridized probes.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

We thank Christina Alfieri, Heather Evans-Anderson, Alex Lange, and Joy Lincoln for technical assistance and scientific discussions. We also thank Sue Kong, Sarah Williams, and the CCHMC Affymetrix Microarray Core for assisting with the generation and analysis of the microarray data. K.E.Y. and T.F.P. were funded by the NIH and T.F.P received an AHA-Ohio Valley Affiliate Pre-Doctoral Fellowship.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information
FilenameFormatSizeDescription
DVDY20923SupplTable1.xls110KSupplementary Table.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.