Focal Adhesion Kinase Signaling Regulates Cardiogenesis of Embryonic Stem Cells*

The signaling steps that induce cardiac differentiation in embryonic stem (ES) cells are incompletely understood. We examined the effect of adhesion signaling including Src and focal adhesion kinase (FAK) on cardiogenesis in mouse ES cells using α-myosin heavy chain promoter-driven enhanced green fluorescent protein or luciferase as reporters. Cardiac transcription factors including Nkx2.5 and Tbx5 mRNA were first expressed at day 4 in hanging drop embryoid bodies, and adhesion of embryoid bodies to surfaces at or before that day strongly inhibited differentiation of ES cells to cardiomyocytes. Since adhesion signaling could suppress cardiogenesis through Src kinases, embryoid bodies were exposed to the small molecule PP2, known as a Src family kinase inhibitor. PP2 during embryoid body adhesion dramatically increased cardiomyocyte differentiation and decreased mRNA expression of neuronal cellular adhesion molecule and α-fetoprotein, neuroectodermal, and endodermal markers, respectively. Surprisingly, although there was an interaction between Src and FAK in cardiogenesis, the procardiogenic effect of PP2 appeared incompletely explained by Src kinase inhibition, since another Src family kinase inhibitor, SU6656, failed to induce cardiogenesis. Instead, PP2 specifically inhibited adhesion-induced FAK phosphorylation. In ES cells stably expressing FAK-related nonkinase, which functions as a dominant negative FAK, cell migration from embryoid bodies was inhibited, whereas α-myosin heavy chain expression and myosin-stained cardiomyocytes were increased, suggesting that reducing cell motility may contribute to cardiogenesis. These data indicate that FAK is a key regulator of cardiogenesis in mouse ES cells and that FAK signaling within embryoid bodies can direct stem cell lineage commitment.

Embryonic stem (ES) 3 cells are totipotent cells derived from the inner cell mass of the preimplantation embryo (1)(2)(3). Aggregation of ES cells into spheres called embryoid bodies triggers their differentiation into all three germ layers: ectoderm, mesoderm, and endoderm (1)(2)(3). ES-de-rived cardiomyocytes are similar to neonatal cardiomyocytes with respect to expression of contractile proteins, ion channels, and connexins (1) and can electrically couple with other myocardial cells after cell transplantation in vivo (4,5). Cardiogenesis of ES cells is a model for studying differentiation of cardiac myocytes as well as a possible source of cardiomyocytes for cell therapy. However, the mechanisms of cardiogenesis are incompletely understood, and current differentiation methods are inefficient in both mouse and human ES cells. Chemical compounds and endogenous substances such as Me 2 SO, 5-azacytidine, retinoic acid, ascorbic acid, nitric oxide, oxytocin, and dynorphin B promote cardiogenesis in ES cells (6 -13), but the mechanisms of induction by these factors have not been clarified.
Extracellular matrix proteins and their predominant cell surface receptors, the integrins, are known to transduce adhesion signals through kinases such as focal adhesion kinase (FAK) and the Src family tyrosine kinases (14 -18). These signaling pathways also participate in tissue regeneration (19,20) and stem cell differentiation. In the embryo, mesoendodermal development is disturbed in fibronectin, laminin ␥1 chain, or ␣5 integrin knock-out mice (21)(22)(23). Prudhomme et al. (24) reported that adhesion of mouse ES cells to fibronectin promotes initial differentiation, whereas adhesion to laminin enhances squamous epithelial cell differentiation in ES cells (25). Laminin 5 and ␣3␤1 integrin signaling induce osteogenic gene expression in human mesenchymal stem cells (26). However, the relationship between adhesion signaling and cardiac differentiation in ES cells is incompletely understood. Therefore, we explored adhesion signaling pathways, including Src and FAK, in the differentiation of mouse ES cells to cardiomyocytes. Here we report that embryoid body adhesion before the initiation of the cardiac differentiation program inhibits cardiogenesis. We explored signaling pathways downstream of adhesion, including the Src tyrosine kinase pathway. We found that exposure of ES cells to PP2, a Src family kinase inhibitor, during embryoid body adhesion dramatically promotes selective cardiogenesis, as shown by increases in ␣-MHC and cardiac troponin T expression and synchronized beating cardimyocytes. Surprisingly, we found that the effect of PP2 was incompletely explained by Src kinase inhibition, and PP2 exerts its effect in part by inhibition of adhesioninduced FAK activation. In ES cells stably expressing FAK-related nonkinase, which functions as a dominant negative FAK, cell migration was inhibited, whereas cardiogenesis was increased. These findings indicate that FAK signaling is a key negative regulator of cardiogenesis in mouse ES cells, and inhibiting FAK activation can promote cardiogenesis.

EXPERIMENTAL PROCEDURES
Materials and Reagents-Glasgow minimum essential medium, ES cell-qualified fetal bovine serum, KNOCKOUT SR, Superscript II, salmon sperm DNA, and the pcDNA3.1/V5-His TOPO vector were from Invitrogen. Leukemia inhibitory factor was from Chemicon. The pEGFP-1 vector was from BD Biosciences. The pGL3-Basic vector and Bright-Glo luciferase assay system were from Promega. Fugene 6 and Expand high fidelity PCR enzyme were from Roche Applied Science. The QuantiTect SYBR Green reverse transcription-PCR kit was from Qiagen. Gelatin type A, BrdUrd, TRI reagent, JumpStart REDTaq ReadyMix, protease inhibitor mixture, and rabbit antibody to actin were from Sigma. The mouse monoclonal antibody to BrdUrd (G3G4), cardiac troponin T, and sarcomeric myosin (MF20) were from the Developmental Studies Hybridoma Bank. The rabbit polyclonal antibodies to phospho-Src (Tyr 416 ), ERK1/2, and phospho-ERK1/2 were from Cell Signaling. The rabbit polyclonal antibodies to MHC, FAK (C-903), and the carboxyl terminus of FAK (C-20) were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Horseradish peroxidase-conjugated goat anti-mouse and rabbit IgG were from Bio-Rad. Alexa Fluor 568-conjugated goat anti-mouse IgG and Hoechst 33342 were from Molecular Probes, Inc. (Eugene, OR). Protein G-agarose and the mouse monoclonal antibody to phosphotyrosine were from Upstate Biotechnology, Inc. Me 2 SO was from the American Type Culture Collection (Manassas, VA). PP2, SU6656, AG1296, and Genistein were from Calbiochem. PDGF-BB was from R & D Systems.
ES Cell Culture and Differentiation-CGR8 murine ES cells were grown on gelatin-coated dishes without feeder cells in Glasgow mini-mum essential medium supplemented with 15% KNOCKOUT SR and leukemia inhibitory factor. Cells were passaged every 3 days. To induce differentiation, cells were first enzymatically dissociated and cultured as hanging drops for embryoid body formation as described previously (13). Differentiation medium with 10% ES cell-qualified fetal bovine serum without leukemia inhibitory factor was added and then exchanged every other day. After embryoid body formation (days 3-7), cells were plated on 0.25 mg/cm 2 gelatin-coated dishes. In each experiment, spontaneously beating cardiomyocytes within embryoid bodies were observed at days 7-8.
Vector Construction and Stable Transfection-The ␣-MHC promoter-driven EGFP vector was made and tested previously in our laboratory (13). For the ␣-MHC promoter-luciferase vector, pGL3-Basic and neomycin-resistant pEGFP-1 vectors were digested with KpnI and XbaI, and the luciferase coding region of pGL3-basic and pEGFP-1 without EGFP were ligated to make luciferase expression vectors. The 5.5-kb fragment of the ␣-MHC promoter region was excised from ␣-MHC-pBluescript SKϩ with SacI and HindIII and then subcloned into the luciferase expression vector. For the FRNK expression vector, the 1.1-kb FRNK coding region (corresponding to amino acids 691-1053 of mouse FAK; GenBank TM accession number M95408) was generated by PCR from mouse lung cDNA. The PCR primers used were as follows (5Ј to 3Ј): CGCCTCGAGCGGATGAGGGAATCCAGAAGA (forward) and CGCAAGCTTGCTCAGTGTGGCCGTGTCTGC (reverse). The resultant PCR product was subcloned into the pcDNA3.1/V5-His TOPO vector with the cytomegalovirus promoter, and correct orientation was confirmed by sequencing. To obtain stable transformants, ES cells were transfected with Fugene 6 and selected for 14 days with 500 g/ml Geneticin as previously described (13).
Immunoprecipitation and Western Analysis-Western analysis was performed as previously described (13). ES cells or embryoid bodies were lysed with modified radioimmune precipitation buffer (phosphate-buffered saline, pH 7.4, 1% Triton X-100, 0.25% sodium deoxycholate, and 1 mM EDTA) with protease inhibitor mixture, 1 mM phenylmethylsulfonyl fluoride, 1 mM sodium fluoride, and 1 mM sodium orthovanadate. After the protein concentrations of cell lysates were determined by Bradford assay, samples were subjected to electrophoresis in 7.5-12% SDS-polyacrylamide gels. Membranes were incubated with primary antibodies overnight (anti-actin diluted 1:2000; anti-FAK diluted 1:200; others diluted 1:1000) and sequentially detected with horseradish peroxidase-conjugated goat anti-mouse or rabbit IgG (1:3000, 1 h) and enhanced chemiluminescence. For FRNK, the rabbit antibody to the carboxyl terminus of FAK (C-20) was used as the pri-mary antibody. For immunoprecipitation, cell lysates were precleared with protein G-agarose beads, and samples containing the same amount of protein were incubated with rabbit antibody to FAK overnight, followed by incubation with protein G-agarose beads for 2 h. Immunocomplexes were then washed three times with ice-cold lysis buffer, resuspended in 2ϫ Laemmli sample buffer, boiled, and subjected to SDS-polyacrylamide gel electrophoresis.
Cell Proliferation Assay-BrdUrd labeling was performed as previously described (27). Day 3 embryoid bodies were dissociated with trypsin-EDTA, and single cells were cultured on gelatin-coated chamber slides in differentiation medium for 48 h with or without Me 2 SO or Src inhibitors. Then cells were incubated with 10 M BrdUrd for 30 min. After fixation, permeabilization, and blocking, cells were incubated sequentially with anti-BrdUrd antibody G3G4 (1:1000) and Alexa Fluor 568-conjugated goat antimouse IgG (1:400). Slides were double-stained with Hoechst 33342 for nuclear staining. For cell counting, embryoid bodies were plated on gelatincoated dishes at day 4 and trypsinized at day 4 or 5, and the cell number was counted manually.
Cell Migration Assay-Day 4 embryoid bodies were plated on gelatincoated dishes and incubated in differentiation medium containing 25 mM HEPES with or without Me 2 SO or Src inhibitors for 24 h. Subsequently, dishes were mounted on an inverted microscope (IX-70; Olympus) on a temperature-controlled stage (TC-344B; Warner Instrument Corp.). Experiments were performed at 37°C and a thin layer of mineral oil was used to minimize evaporation of culture media. After a brief equilibration period, images were automatically acquired at ϫ10 mag- nification every 2 min for 4 h (corresponding to 120 frames) using a digital CCD camera (CoolSNAP HQ; Roper Scientific) and stored on a computer for subsequent cell migration analysis using MATLAB software (The MathWorks). Single cells were tracked over time using a custom-written normalized cross-correlation algorithm based on the nuclear position, and single cell tracking coordinates were subsequently used to compute migration distance and velocity for each cell.
Statistical Analysis-Data were shown as mean Ϯ S.D. and analyzed by Student's t test or 1-way analysis of variance with post hoc analysis, with p Ͻ 0.05 considered significant.  NOVEMBER 25, 2005 • VOLUME 280 • NUMBER 47

Cardiac Differentiation Programs Begin in Hanging Drop Embryoid
Bodies at Day 4-First, to explore when cardiac differentiation programs initiate within embryoid bodies, mRNA expression of heart-related transcription factors was analyzed by quantitative PCR (Fig. 1). Embryoid bodies were adhered to gelatin 6 days after hanging drop embryoid body formation, a typical protocol for differentiation of ES cells. Brachyury, an early mesodermal marker (28), was transiently expressed from days 3-5. GATA 4 was expressed earliest among the transcription factors examined and increased rapidly after day 4. The expression of Nkx2.5, Tbx5, and MEF2C, which have crucial roles in cardiac development in the embryo (29,30), began after day 4, followed by expression of the gene for the sarcomeric protein ␣-MHC. Nkx2.5 is known to express early within the embryonic heart (31,32), and the observed sequential gene expression pattern in embryoid bodies was consistent with the cardiac developmental process. Collectively, these data indicate that cardiac differentiation programs in embryoid bodies start at day 4, before the typical embryoid body adhesion step.
Embryoid Body Adhesion before Initiation of Cardiac Differentiation Inhibits Cardiogenesis-To determine if embryoid body adhesion affects cardiogenesis, ␣-MHC promoter-driven EGFP cells were cultured as embryoid bodies and then adhered on gelatin at various time points, and EGFP-positive cardiomyocytes were observed at day 10. The number of EGFP-positive cells was significantly decreased if embryoid bodies were adhered at day 4 or earlier ( Fig. 2A). This finding was also confirmed by ␣-MHC promoter-driven luciferase activity (Fig. 2B). These data suggest that signals downstream of embryoid body adhesion inhibit cardiogenesis.
PP2 Exposure during Embryoid Body Adhesion Promotes Selective Cardiogenesis-Src tyrosine kinase and FAK are kinases downstream of integrin signaling (15)(16)(17). Therefore, we examined whether exposure of PP2, a pyrazolopyrimidine derivative known as an Src family kinase inhibitor, could promote cardiogenesis during embryoid body adhesion. Exposure of 10 M PP2 from day 4 dramatically enhanced ␣-MHC promoter-driven luciferase activity (4-fold) compared with control embryoid bodies adhered at day 4 ( Fig. 3A) and also with embryoid bodies adhered at day 6 (data not shown). Days 4 -6 was the critical period for PP2 to exert its inductive effect. In contrast, PP2 exposure from day 6 onward had no effect, and from times before day 4, PP2 inhibited subsequent cardiogenesis. Total protein content of embryoid bodies was not changed significantly by PP2 exposure from day 4 (data not shown). PP2 exposure during days 4 -6 induced more ␣-MHC promoter-driven EGFP-positive cells than control, and the proportion of EGFP-positive cells was 5-10% (Fig. 3B). Furthermore, differentiated cardiomyocyte clusters following PP2 treatment beat more synchronously than Me 2 SO-treated cells (supplemental videos, DMSO.avi and PP2.avi). These effects were achieved by concentrations of PP2 greater than or equal to 5 M, and 10 M was most effective (Fig. 3C). Recently, one of the Src family kinases (cYes) was reported to be important for mouse and human ES cell self-renewal, and cYes inhibition combined with a low concentration of retinoic acid leads to multilineage differentiation (27). To exclude the possibility that PP2 promotes differentiation to other cell lineages of ectodermal, mesodermal, and endodermal origin, NCAM, PECAM-1, and ␣-fetoprotein (␣-FP) mRNA expression was analyzed by quantitative PCR; NCAM, PECAM-1, and ␣-FP are markers of neuroectoderm, endothelial cells of mesoderm, and endoderm, respectively (33,34). At day 10, NCAM expression was decreased and ␣-FP was strikingly diminished, whereas PECAM-1 was not affected by PP2 (Fig. 3D). These data demonstrate that PP2 exposure during embryoid body adhesion promotes selective cardiogenesis, possibly by regulating stem cell lineage commitment.
PP2 Inhibits Adhesion-induced FAK Phosphorylation and Increases Tbx5 mRNA Expression-Since PP2 is a Src family kinase inhibitor, we tested whether SU6656, another Src family kinase inhibitor (27,35), also enhances cardiogenesis in ES cells. Day 4 embryoid bodies were adhered, and adhesion-induced Src family kinase activation (defined as tyrosine 416 autophosphorylation (36)) was detected by Western analysis. Src family kinases were activated up to 3 h after adhesion in embryoid bodies treated with Me 2 SO, whereas they were strongly inhibited by SU6656 as well as PP2 (Fig. 4A). Surprisingly, however, SU6656 exposure did not promote cardiogenesis in embryoid bodies even at high concentrations (Fig. 4B). This result was confirmed by Western analysis of MHC and cardiac troponin T, showing that only PP2 markedly enhanced expression of these cardiac proteins (Fig. 4E). Then, we exam-  NOVEMBER 25, 2005 • VOLUME 280 • NUMBER 47 ined whether stimulation with the Src activator PDGF-BB affects cardiogenesis in embryoid bodies. PDGF-BB at a concentration of 50 ng/ml activated Src in Me 2 SO-treated control cells (Fig. 4C) and significantly inhibited ␣-MHC mRNA expression (Fig. 4D, Me 2 SO: 50.4 Ϯ 25.5% compared with control without PDGF stimulation, n ϭ 3, p Ͻ 0.05). This inhibitory effect of PDGF was also observed in SU6656-pretreated cells, but not in PP2-treated cells, both at the mRNA and the protein level (Fig. 4D, SU6656: 126.4 Ϯ 22.5 to 72.9 Ϯ 12.6% by stimulation with PDGF (50 ng/ml), n ϭ 3, p Ͻ 0.05; Fig. 4E, upper panel), although SU6656 inhibits PDGF-induced activation of Src equal to or greater than PP2. Instead, PP2 effectively blocked PDGF-induced FAK activation (Fig. 4C, right). We further tested whether inhibition of PDGF receptor activation promotes cardiogenesis, but exposure of the PDGF receptor kinase inhibitor AG1296 with or without SU6656 did not have an effect similar to PP2 (data not shown).

Focal Adhesion Kinase and Cardiogenesis
Next, the cardiac gene expression profile in cells treated with either Me 2 SO, PP2, or SU6656 was compared by quantitative PCR (Fig. 5). We found that the gene expression profile of SU6656 was almost identical to that of Me 2 SO-treated control regarding cardiac genes analyzed here. An increase in ␣-MHC expression over control by PP2 was observed from day 7, and interestingly, among cardiac transcription factors examined, only Tbx5 was significantly increased by PP2 from day 5 (3.5and 3.4-fold as control at day 5 and day 6, respectively), prior to the increase in gene expression for other transcription factors and ␣-MHC. This is noteworthy, because Tbx5 is a transcription factor that is critical for cardiogenesis in the embryo (37) and synergistically promotes cardiogenesis with Nkx2.5 in embryonic carcinoma cells (38). Collectively, these results suggest that Src participates in cardiac differentiation, whereas the observed cardiogenic effect of PP2 is incompletely explained by Src inhibition, and suggest that PP2 exposure during embryoid body adhesion specifically increases Tbx5 mRNA expression to promote cardiogenesis.
Several groups previously reported that PP2 inhibits phosphorylation of FAK (39 -41, 62), the epidermal growth factor receptor (42), PDGF receptor ␤ (42), or c-Kit (43) in various cell types. Therefore, we examined whether PP2 specifically inhibits adhesion-induced FAK and ERK1/2 phosphorylation. Adhesion-induced FAK phosphorylation was almost completely blocked by PP2 exposure but not by SU6656, whereas ERK1/2 activation did not differ significantly between PP2 and SU6656 (Fig. 6A). Moreover, cell migration outside of the embryoid body for 2 days after adhesion was strongly inhibited in embryoid bodies treated with PP2 (Fig. 6B), consistent with previous reports that FAK activity plays a role in cell migration (44 -47). It is also recognized that altering cellular ability to move could drive the change of the cell to committed status. These data indicate that PP2 specifically inhibits adhesion-induced FAK phosphorylation and that FAK inhibition may be the mechanism by which PP2 induces cardiogenesis in embryoid bodies.

Stable Expression of FRNK Inhibits Cell Migration and Promotes Cardiogenesis in Embryoid
Bodies-To establish the effect of FAK inhibition on cardiogenesis, ES cells stably expressing FAK-related nonkinase (FRNK) or GFP were generated and analyzed for cardiogenesis. FRNK is the 41-kDa FAK carboxyl-terminal domain that functions as a competitive inhibitor of FAK signaling (47,48) and inhibits FAK phosphorylation in various cell types, such as fibroblasts (49), vascular smooth muscle cells (44,50), and rat neonatal cardiomyocytes (51). At day 4, embryoid bodies from FRNK-expressing ES cells retained FRNK expression (Fig. 7A, top left). In FRNK-expressing embryoid bodies, adhesion-induced FAK phosphorylation was inhibited, and Src activity was also suppressed compared with control (Fig. 7A, bottom left), indicating that an interaction between FAK and Src exists in this system.
Quantitative PCR analysis showed that ␣-MHC mRNA expression was significantly increased (ϩ70%) in FRNK-expressing embryoid bodies but strongly decreased in GFP-expressing embryoid bodies (Fig. 7B). Interestingly, adhesion-induced FAK phosphorylation was strongly activated in these GFP-expressing embryoid bodies compared with control (Fig. 7A, bottom right). Increases in Tbx5 expression were not detected in this experiment using ES cells stably expressing FRNK. Furthermore, when day 14 embryoid bodies were immunostained with anti-sarcomeric myosin antibody MF20, the proportion of cardiomyocytes was increased in FRNK-expressing embryoid bodies as well as PP2-treated embryoid bodies when compared with control. This confirmed the procardiogenic effect of FRNK in ES cells (Fig. 7C).
FAK can affect cell proliferation and migration. Thus, next we explored if FAK inhibition affects cell proliferation and cell migration in ES cells. In cells from day 5 embryoid bodies, the percentage of BrdUrdpositive cells was significantly decreased with PP2 exposure as well as SU6656 (Fig. 8A, top left). In FRNK-expressing cells, the percentage of BrdUrd was increased, whereas the actual cell counts were significantly decreased compared with control both at day 4 and 5 (Fig. 8A, top right). On the other hand, when observed by time lapse microscopy, cell migration distance was significantly decreased both in PP2-treated and FRNK-expressing cells, but not in SU6656-or Me 2 SO-treated cells (Fig.  8B). These data indicate that FAK inhibition blocks cell migration and promotes cardiogenesis in embryoid bodies.

DISCUSSION
In the present study, we utilized heart-specific ␣-MHC promoterdriven EGFP or luciferase as reporters for analyzing cardiac differentiation from embryoid bodies of mouse ES cells. We first demonstrated that embryoid body adhesion before the initiation of cardiogenic transcriptional programs inhibits cardiogenesis and that exposure of PP2, a compound known as an Src family kinase inhibitor, during embryoid body adhesion promotes selective cardiogenesis. Furthermore, we showed that this cardiogenic effect of PP2 appears unexplained solely by Src kinase inhibition; instead, the procardiogenic effect is at least partly due to inhibition of adhesion-induced FAK activation. In embryoid bodies, heart-related transcription factors Nkx2.5, Tbx5, and MEF2C mRNA started to express after the mesodermal induction marker Brachyury expression, then followed by ␣-MHC expression. This result confirmed the established finding that the process of ES cell differentiation to cardiomyocytes recapitulates that of early cardiogenesis in the embryo. When embryoid body adhesion occurred before expression of these transcription factors, cardiogenesis was strongly inhibited, suggesting that adhesion-related signals interfere with the initial cardiac differentiation program.
While we investigated the role of integrin signaling on cardiogenesis, we found that PP2, an inhibitor of Src kinase that is an immediate downstream kinase in integrin signaling, dramatically promotes cardiogenesis in ES cells. PP2 did not induce apoptosis or cell death within embryoid bodies, since total protein content did not change significantly. The effect of PP2 was not observed in ES cell monolayer culture without embryoid body formation. ES cell aggregation itself can decrease Nanog transcription factor expression and induce primitive endoderm differentiation (52), suggesting that PP2 acts only after the initial germ layer specification of ES cells. Furthermore, PP2 enhanced cardiogenesis in embryoid bodies, whereas differentiation to cell lineages of neuroectodermal and endodermal origins was strongly inhibited, indicating that PP2 may control general stem cell lineage commit-ment. PP2 may direct cells toward cardiac progenitors rather than promote proliferation of cardiac progenitors, because Tbx5, which was specifically increased with PP2, can inhibit proliferation of chick embryonic cardiomyocytes (53). It is also noteworthy that PP2 did not affect cardiogenesis after day 6, when Tbx5, Nkx2.5, and MEF2C expression levels plateau in embryoid bodies, when stem cells have already been directed to cardiac progenitors. Differentiated cardiomyocytes with PP2 expressed greater amounts of cardiac contractile proteins, such as myosin heavy chain and cardiac troponin T, and contracted more synchronously with neighboring cells than control, suggesting that PP2 treatment may enhance not only cardiac differentiation but also cardiomyocyte maturation. This finding could possibly be explained by an increase in Tbx5, which targets genes including ␣-MHC (54) and connexin 40 (37).
Stimulation with PDGF-BB activated Src and led to inhibition of cardiogenesis in control ES cells, and Src activity was blocked by stable expression of FRNK. These results indicate that there is an interaction between Src and FAK in cardiogenesis in this system. On the other hand, 1 M PP2 did not induce the cardiogenic effect, although the IC 50 of Src kinase by PP2 is 0.1 M (55). Moreover, another Src family kinase inhibitor SU6656 failed to mimic the effect of PP2 and also failed to  block the inhibitory effect of PDGF. Although Src inhibitors used here do not allow one to distinguish the effects of each Src family kinase, the observed cardiogenic effect of PP2 appears to be incompletely explained by Src inhibition. It should be noted that in this experimental setting with serum, FAK can interact not only with Src but with other kinases, such as phosphoinositide 3-kinase and Rho small GTPases (56 -59). PP2 has been reported to inhibit phosphorylation of FAK and several receptor tyrosine kinases such as epidermal growth factor receptor, PDGF receptor ␤, c-Kit, or Bcr/Abl in rat neonatal cardiomyocytes, vascular smooth muscle cells, lung fibroblasts, and some cancer cell lines. FAK is considered a key step in integrin and receptor tyrosine kinase signaling (60). We found that PP2 specifically inhibits adhesion-induced FAK, but not ERK1/2, phosphorylation in ES cells. FRNK is a competitive inhibitor of FAK and is known to be expressed solely in vascular smooth muscle cells of aorta and lung (44). In the present experiment, stable overexpression of FRNK led to increased ␣-MHC mRNA expression and MF20-positive cells in embryoid bodies. However, we cannot exclude the possibility that PP2 inhibited another target as well as FAK. In contrast, adhesion at early stages (Figs. 6A and 7A) and stable expression of GFP (Fig. 7A) led to FAK activation in embryoid bodies, with subsequent inhibition of cardiogenesis. These results further indicate that FAK inhibits cardiogenesis. It was previously reported that FAK is not essential for differentiation of mouse ES cells (61). These authors found "no evidence of defects" in multilineage differentiation in FAK knock-out ES cells, and embryoid bodies were analyzed up to 7 days, which is earlier than the time when cardiogenesis peaks in our experiments. In addition, heart-specific mRNA or protein expression was not presented in their study.
Inhibition of FAK can lead to changes in cell proliferation or migration, and it is possible that altering cellular ability to proliferate or to move in response to paracrine stimuli drives cell commitment. Thannickal et al. (62) reported that 10 M PP2 prevents lung myofibroblast differentiation by inhibition of adhesion-induced FAK phosphorylation. In this experiment, cell migration was inhibited both in PP2treated cells and FRNK-expressing cells, indicating that inhibition of cell migration, but not cell proliferation, may promote cardiogenesis in ES cells. Also, FAK inhibition has its effect possibly by inhibition of phosphoinositide 3-kinase, one of the downstream targets of FAK, since phosphoinositide 3-kinase signaling is reported to regulate mouse ES cell self renewal (59) as well as initial cardiac differentiation of mouse embryonic carcinoma cells by maintaining canonical Wnt/␤-catenin signaling (63). Interestingly, although their two-dimensional model of embryonic carcinoma cells was different from our embryoid body system, they observed that prolonged activation of Wnt signaling after day 4 inhibited full differentiation into spontaneous beating cardiomyocytes, as we observed in the situation of continuous FAK activation described above.
These findings could lead to more efficient production of cardiomyocytes from ES cells and suggest that understanding FAK signaling could help promote cardiogenesis of adult somatic stem cells. Future studies to define cardiomyocyte types induced by PP2, downstream targets of FAK, and the mechanisms of the initial embryoid body interactions may be useful.