Hedgehog Signaling Induces Cardiomyogenesis in P19 Cells*

Sonic Hedgehog (Shh) is a critical signaling factor for a variety of developmental pathways during embryogenesis, including the specification of left-right asymmetry in the heart. Mice that lack Hedgehog signaling show a delay in the induction of cardiomyogenesis, as indicated by a delayed expression of Nkx2-5. To further examine a role for Shh in cardiomyogenesis, clonal populations of P19 cells that stably express Shh, termed P19(Shh) cells, were isolated. In monolayer P19(Shh) cultures the Shh pathway was functional as shown by the up-regulation of Ptc1 and Gli1 expression, but no cardiac muscle markers were activated. However, Shh expression induced cardiomyogenesis following cellular aggregation, resulting in the expression of factors expressed in cardiac muscle including GATA-4, MEF2C, and Nkx2-5. Furthermore, aggregated P19 cell lines expressing Gli2 or Meox1 also up-regulated the expression of cardiac muscle factors, leading to cardiomyogenesis. Meox1 up-regulated the expression of Gli1 and Gli2 and, thus, can modify the Shh signaling pathway. Finally, Shh, Gli2, and Meox1 all up-regulated BMP-4 expression, implying that activation of the Hedgehog pathway can regulate bone morphogenetic protein signals. Taken together, we propose a model in which Shh, functioning via Gli1/2, can specify mesodermal cells into the cardiac muscle lineage.

Sonic Hedgehog (Shh) is a secreted glycoprotein that is involved in a plethora of developmental processes (20,21). It is a key signal in determining left-right asymmetry that includes situating the heart on the left side of the body (22)(23)(24). The Shh signaling pathway is mediated by Gli1/2/3, a family of zincfinger transcription factors that are regulated at the protein and expression level by Shh (25)(26)(27). The transmission of Shh signaling is accomplished via its receptor Patched1 (Ptc1) and Smoothened (Smo). The current model proposes that Shh binds to Ptc1 and, upon complex formation, both proteins are brought into the cell by endocytosis and degraded by the lysosomal pathway. This allows the trafficking of Smo, situated in the Golgi apparatus, to the plasma membrane. Smo then mediates the activation of intracellular factors resulting in the positive regulation of the Gli2/3 protein. In turn, these factors regulate the expression of Shh target genes such as Ptc1 and Gli1 (28).
Evidence suggests a role for the Shh signaling pathway in heart morphogenesis. ShhϪ/Ϫ and Gli2Ϫ/ϪGli3Ϫ/ϩ mice both display a heart-looping defect (23,29), and Ptc1 Ϫ/Ϫ mice exhibit heart morphological defects (30). Furthermore, several studies also support an involvement for Shh and Gli factors in regulating cardiomyogenesis. Mice lacking the Smo gene (31) were embryonic lethal, likely due to defects in cardiac looping. Upon examination of the cardiac crescent in these mice, Nkx2-5 expression was found to be severely down-regulated. Upon fusion of the pre-cardiac fields, Nkx2-5 expression returned to wild-type levels. Finally, in the same study homozygous Ptc1 mutant mice displayed a widened expression domain of Nkx2-5 in the cardiac crescent in comparison to wild-type mice. Therefore, Shh signaling via Ptc1 and Smo seems to be critical for maintaining wild-type levels of Nkx2-5 expression in the cardiac crescent.
Shh and Gli factors are expressed throughout the mesoderm during gastrulation (22,32). Particularly, Shh is expressed in Henson's node and other early mid-line structures (31). This indicates that Shh and Gli factors are present at the correct time in embryonic heart development to be involved in specifying early mesodermal cells to progress into the cardiac lineage.
Meox1/2 are homeobox-containing transcription factors (33) that have been investigated for their roles in skeletal myogenesis and sclerotome formation (34 -36), as they are expressed in the developing somite (33,37). In addition, Meox1/2 have a distinctive expression pattern in several stages of heart development in the mouse. Meox1/2 proteins are expressed in the lateral plate mesoderm at E7.5-E8 (38), with Meox1 mRNA detected just prior to the protein expression (33). In later heart stages Meox1 mRNA is found in the outflow tract of the premature heart at E10.5 (33), whereas the Meox2 protein is expressed in the linear heart tube at E8.0 and E12.5 in the atrial/ventricular myocytes (38). Meox1 protein is also found in the primitive streak where pre-cardiac cells are first thought to be specified to the heart lineage (37). Moreover, there is a MEF2 binding site in the Meox2 promoter, and MEF2C and Meox2 have very similar spatio-temporal expression patterns in the developing heart. Interestingly, mice lacking both Meox1/2 genes showed no gross abnormality of the heart at p.c. 16.5 (35).
Recent results in our laboratory utilizing the P19 cell system have shown that Meox1 and Gli2 are involved in a positive regulatory loop (39) in differentiating P19 cells. P19 cells comprise a pluripotent embryonal carcinoma stem cell line that can differentiate into cardiomyocytes in the presence of dimethyl sulfoxide (Me 2 SO) and with cellular aggregation (40). To further examine the role of Shh, Gli2, and Meox1 in the context of embryonic stem cells, we expressed each of these factors in P19 cells and examined the differentiation potential of each cell line. Aggregated P19 cells expressing either Shh, Gli2, or Meox1 were capable of differentiating into cardiac muscle in the absence of Me 2 SO. Therefore, these factors appear to play a role in the specification of mesodermal cells into the cardiac muscle lineage.

MATERIALS AND METHODS
Plasmid Constructs-Meox1 cDNA was sub-cloned into the phosphoglycerate kinase (PGK) vector as described previously (39). The coding domain of Shh was amplified by PCR utilizing the oligonucleotides 5Ј-AAACTCGAGCCACCATGCTGCTGCTGCTGGCC-3Ј and 3Ј-AAAC-TCGAGTCAGCTGGACTTGACCGC-5Ј at an annealing temperature of 67.7°C. The 5Ј and 3Ј oligonucleotides contained XhoI sites (boldface) to facilitate cloning. Amplified cDNA was then sub-cloned into pCR-Blunt Vector (Invitrogen). Shh cDNA was subsequently sub-cloned into the PGK vector containing the pgk-1 promoter (41). The PGK-Puro and PGK-LacZ constructs were described previously (42).
Cell Culture and Transfections-P19 cells were cultured as described (43) in ␣-minimum essential media supplemented with 5% Cosmic calf serum (HyClone, Logan, UT) and 5% fetal bovine serum (CanSera, Rexdale, Ontario, Canada). Stable cell lines were isolated as described previously (15, 44 -47). PGK-Meox1 stable cell lines were prepared as described previously (39). Briefly, CMV-Gli2 and PGK-Shh were stably transfected using the FuGENE™ 6 transfection reagent per the manufacturer's instructions (Roche Applied Science). A mixture containing 2.04 g of CMV-Gli2, PGK-Shh, or PGK vector alone for controls, and 0.09 g of PGK-puro, 0.17 g of PGK-LacZ, and 0.77 g of B17 were used to create the Gli2 and Shh cell lines. For all stable transfections, a mixture of DNA and the FuGENE™ 6 reagent was added to 2.5 ϫ 10 5 cells in 35-mm tissue culture dishes. Transfection efficiency for each experiment was assessed by ␤-galactosidase assays as described (48). Cells were selected for puromycin resistance (2 g/ml) for 7-9 days. Resulting colonies were tested for expression by slot blot analysis, and stable cells lines were isolated and termed P19(Gli2), P19(Shh), or P19(control) cells.
P19(control), P19(Gli2), P19(Shh) and P19(Meox1) cells were differentiated in the absence of Me 2 SO, and P19 cells were differentiated in the presence of 0.8% Me 2 SO as described (15, 44 -47, 49). Differentiation involved aggregation of cells for 4 days in Petri dishes (in the absence or presence of Me 2 SO) followed by plating into tissue culture dishes on day 4. Cells were harvested for total RNA on day 6.
Immunofluorescence-For assaying myosin heavy chain (MyHC) and cardiac ␣-actin expression, cells were plated onto gelatin-coated coverslips on day 4 and fixed on day 6 in Ϫ20°C methanol. Cells were rehydrated in PBS, MyHC expression was detected utilizing the monoclonal MF20 antibody supernatant (50), and cardiac ␣-actin expression was detected utilizing the Ac1-20.4.2 antibody (Research Diagnostics Inc.). Nuclei were detected by Hoechst stain as described previously (47). In brief, 100 l of a 1:1 dilution of the MF20 supernatant or 100 l of a 1:10 dilution of Ac1-20.4.2 in 2% albumin and 2% sheep serum PBS was incubated on coverslips for 1 h at room temperature. Coverslips were then washed three times for 5 min each time in PBS and incubated with 100 l of a 1:100 dilution of goat anti-mouse Cy3-linked antibody (Jackson ImmunoResearch) for 1 h at room temperature. After the PBS washes, coverslips were mounted, and immunofluorescence for MyHC was visualized with a Zeiss Axioskop microscope. Images were captured on a Sony 3 charge-coupled device camera and processed utilizing Axiovision, Adobe Photoshop7, and Canvas8 software. Immunofluorescence for cardiac ␣-actin was visualized with an Olympus BX50 microscope. Images were captured on a Roger Scientific Cool Snap camera and processed utilizing Image Pro-Plus 5.1 (Media Cybernetics) and Canvas 9 software.
For assaying GATA-4 expression, cells were plated as above and fixed on day 6 with 3.7% formaldehyde at room temperature for 15 min. Cells were then perforated with 0.2% Triton X-100 for 5 min at room temperature. Cells were blocked for 30 min at room temperature with 2% horse serum and 2% albumin PBS. Expression was detected utilizing a goat anti-GATA-4 antibody (sc-1237; Santa Cruz Biotechnology). Briefly, 100 l of a 1:100 dilution of anti-GATA-4 antibody in blocking solution was incubated on coverslips for 1 h at room temperature. Coverslips were washed as described above and incubated with 100 l of a 1:800 dilution of donkey anti-goat Cy3-linked antibody (Jackson ImmunoResearch) at room temperature for 1 h. After PBS washes, immunofluorescence was visualized, and images were processed on cardiac ␣-actin-stained coverslips as described above.
Northern Blot Analysis-Total RNA was isolated from each differentiation utilizing the LiCl method, as described (44). 12 g of total RNA were separated on a 1% agarose and formaldehyde gel, transferred onto Hybond-N (Amersham Biosciences) by capillary action, and crosslinked by UV light irradiation. Blots were then hybridized to DNA probes labeled with (␣-32 P)dNTP by multi-prime labeling (Amersham Biosciences) for 16 h at 42°C. Probes for BMP-4, Nkx2-5, GATA-4, MEF2C, Meox1, Gli2, and18S (for standardization of loading) have been described previously (47). The probes used were a 908-bp BamHI fragment of mouse Shh cDNA, an 841-bp EcoRI fragment of mouse Ptc1 cDNA, a 1.6-kb EcoRI fragment of Gli1 cDNA, an 0.8-kb EcoRI fragment of Gli3 cDNA, a 2.1-kb EcoRI fragment of a mouse Zic1 expressed sequence tag clone (EST) (clone identifier 6418777; Invitrogen), and a 992-bp SstI fragment of a mouse Zic3 EST clone (clone identifier 5120056; Invitrogen). The Ptc1 probe utilized for Southern blot analysis of RT-PCR products consisted of full-length mouse Ptc1 cDNA. Northern blots were then washed five times for 5 min each in low stringency wash (2ϫ SSC, 0.2% SDS) at 42°C and for 15 min in high stringency wash (0.1ϫ SSC, 0.2% SDS) at 65°C. Blots were visualized by autoradiography. Quantitative analyses were performed using National Institutes of Health Image 1.63 software.
Reverse Transcription Polymerase Chain Reaction-Total RNA was harvested on day 6 utilizing the LiCl method, followed by purification with the RNeasy® mini kit (Qiagen, Mississauga, Ontario, Canada). Approximately 1 g of DNase I-treated RNA was used to synthesize first strand DNA, utilizing the Superscript first strand synthesis kit (Invitrogen). Random hexamers (50 ng) were utilized per the manufacturer's protocol. Two microliters of first strand DNA was then used for PCR reactions.
First strand reactions were tested for linearity with each set of oligonucleotides, and negative controls included RT experiments in the absence of RNA or reverse transcriptase enzyme and PCR in the absence of a first strand reaction. Southern blot analysis was carried out to analyze RT-PCR products by utilizing probes for tubulin, Nkx2-5 as described previously (15) and Shh as described for Northern blot analysis. The probe for ␤-actin consisted of an ϳ520-bp fragment of the PCR product amplified with the primers described above.
Enzyme-linked Immunosorbent Assay (ELISA)-P19(Shh), P19(control), and P19 cells were grown to confluence in 100-mm tissue culture dishes. The media were then retrieved, and the concentration of Shh was meas-ured therein using the R&D Systems DuoSet ELISA development kit (DY461) and the SpectraMax Plus 384 microplate reader and software (Molecular Devices).

Shh Induces Cardiomyogenesis in Aggregated P19 Cells-
Because previous studies have implicated a role for Shh signaling in heart morphogenesis, we tested whether Shh signaling was sufficient to induce cardiomyogenesis in P19 cells. As such, we isolated clonal populations of P19 cells overexpressing Shh termed P19(Shh) cells. Subsequently, we aggregated P19(Shh) cells and P19(control) cells in the absence of Me 2 SO. On day 6 of the differentiation, cultures were examined for the presence of cardiomyocytes. Staining with an anti-MyHC antibody showed the formation of cardiomyocytes in P19(Shh) cells but not in P19(control) cells (Fig. 1, compare panel D with panel  B). Although MyHC is expressed in both cardiac and skeletal myocytes, the morphology indicated a rounded cardiomyocyte phenotype as opposed to a bipolar skeletal myocyte phenotype. Hoechst dye staining was included as a control for cell number in both stains (Fig. 1, A and C). Eleven stable P19 cell lines overexpressing Shh with the capacity to induce cardiomyogenesis over controls were isolated.
Using an enzyme-linked immunosorbent assay, we determined the concentration of extracellular Shh in the media of confluent cultures of the three P19(Shh) clones that were used for further analysis, as well as the P19(control) and parental P19 cells. The concentration of Shh found in the extracellular media of P19(Shh) cells was found to be between ϳ8 and 19 g/ml (data not shown), whereas the concentration of Shh in the extracellular media of P19(control) and P19 cells was found to be lower than the detectable limit of the assay, which was 7.5 pg/ml.
To further analyze the cardiac muscle phenotype, the induction of gene expression was monitored by Northern blot or RT-PCR analysis (Fig. 2, panels I and II, respectively), and the induction of protein expression was monitored via immunofluorescence (Fig. 3). Northern blot analysis showed that Shh transcripts were overexpressed in P19(Shh) cells, whereas P19(control) cells, both positive and negative, showed no detectable Shh expression at this exposure level (Fig. 2, panel I,  section A). GATA-4, BMP-4, MEF2C, and Nkx2-5 gene expressions were found to be up-regulated in P19(Shh) cells on day 6 but not in P19(control) cells (Fig. 2, panel I, sections B and C and panel II, sections L and M, respectively), confirming the presence of cardiomyocytes.
To determine whether the up-regulation of GATA-4 transcript levels corresponded to an increase in protein expression, we examined the expression of the GATA-4 protein in differentiated P19(Shh) cells by immunofluorescence with an anti-GATA-4 antibody. Nuclear expression of GATA-4 was observed in P19(Shh) cells but not in P19(control) cells (Fig. 3, compare  panel D with panel B). A Hoechst stain was included as a control for cell number and location of the nucleus. To further confirm the presence of muscle-specific structural proteins found in embryonic cardiac cells, we also stained for the expression of cardiac ␣-actin. As expected, cardiac ␣-actin was To further dissect the Shh signaling pathway, Ptc1 and Gli1/ 2/3 transcripts were examined for changes in gene expression by Northern blot and RT-PCR analysis (Fig. 2). Ptc1 and Gli1 transcripts were up-regulated on day 0 and 6 in P19(Shh) cells as compared with P19(control) cells (Fig. 2, panel I, sections D and E, respectively). Examining Gli2/3 transcripts by densitometry showed that they were not significantly up-regulated on day 6 in P19(Shh) cells in comparison to P19(control) cells (Fig. 2, panel I, sections F and G, respectively). Examining Zic1 and Zic3 transcripts by densitometry also showed no consistent change in expression (Fig. 2, panel I, sections H and I, respectively). Previous results have shown that Gli2 can activate Meox1 expression in P19 cells (39). Hence, we examined the expression profile of Meox1 in P19(Shh) cells. It was shown that Meox1 transcript levels were not enhanced over controls in P19(Shh) cells (Fig. 2, panel I, section J). 18S and tubulin transcripts were assayed for loading controls (Fig. 2, panel I,  section K and panel II, section N, respectively). Therefore, Shh up-regulated the expression of Ptc1 and Gli1 prior to cellular aggregation, indicating the normal functioning of the Shh pathway in P19 cell monolayer cultures. Subsequent to aggregation, activation of the Shh pathway resulted in the up-regulation of GATA-4, MEF2C, Nkx2-5, and BMP4 transcripts as well as the nuclear GATA-4 protein, leading to cardiomyogenesis.
Gli2 Induces Cardiomyogenesis in Aggregated P19 Cells-To identify downstream components of the Shh signaling pathway, we examined the function of Gli2 in aggregated P19 cells. Previous studies in our laboratory (39) showed that Gli2 could induce low levels of skeletal myogenesis in aggregated P19 cells. Interestingly, in the same experiments these cell lines also differentiated into much higher levels of cardiac muscle. To study these results further, we created seven stable P19 cell lines expressing high levels of Gli2 termed P19(Gli2). P19(Gli2) and P19(control) cells were aggregated without Me 2 SO, and cultures were inspected for the formation of cardiomyocytes on day 6 of the differentiation. Immunofluorescence with an anti-MyHC antibody showed that all seven P19(Gli2) cell lines differentiated into cardiomyocytes and that P19(control) cells did not. A representative P19(Gli2) cell line is shown (Fig. 1,  compare panel F with panel B). Hoechst dye staining was included as a control for cell number (Fig. 1E).
To confirm the presence of cardiomyocytes, we examined the expression of several genes by Northern blot or RT-PCR analysis (Fig. 4, panels I and II, respectively). Analysis by Northern blot showed high levels of exogenous Gli2 transcripts in P19(Gli2) cells that were not present in P19(control) cells (Fig.  4, panel I, section A). Furthermore, GATA-4, BMP-4, Nkx2-5, and MEF2C expression levels were up-regulated on day 6 in P19(Gli2) cells and not in P19(control) cells, aggregated in the absence of Me 2 SO (Fig. 4, panel I, sections B and C and panel II, sections J and K, respectively).
Next, we examined expression levels of Gli1/3 and Ptc1 to determine changes in gene expression in the Shh signaling pathway. Gli1 transcripts were up-regulated on days 0 and 6 in P19(Gli2) cells but not in P19(control) cells (Fig. 4, panel I,  section D). Furthermore, quantization by densitometry showed that Gli3 was up-regulated 4 Ϯ 1-fold (S.D.; n ϭ 4) on day 6 in P19(Gli2) cells when compared with P19(control) cells (Fig. 4,  panel I, section E). Ptc1, however, was shown not to be consistently up-regulated on day 6 in P19(Gli2) cells when compared with P19(control) cells (Fig. 4, panel I, section F).
Zic proteins bind to the same DNA sequences as the Gli family of transcription factors and have also been shown to physically interact with them (51). Hence, it was interesting to investigate whether the expression level of two genes from the Zic family, Zic1 and Zic3, were altered because of the overexpression of Gli2 in P19 cells. Results showed that Zic1 and Zic3 transcripts were down-regulated when comparing day 0 to day 6 in P19(Gli2) cells and P19(control) cells (Fig. 4, panel I,  sections G and H, respectively). Meox1, in accordance with a previous report from our laboratory (39), was activated on day 6 in P19(Gli2) clones and not in P19(control) cells (data not shown). 18S and ␤-actin expression were assayed for loading controls (Fig. 4, panel I, section I and panel II, section L, respectively). Therefore, prior to cellular aggregation Gli2 upregulated the expression of Gli1. Subsequent to aggregation, Gli2 up-regulated the expression of Nkx2-5, GATA-4, MEF2C, and BMP4, leading to cardiomyogenesis.
Meox1 Induces Cardiomyogenesis in Aggregated P19 Cells-In previous studies we showed that Meox1-activated Gli2 expression was not sufficient to initiate skeletal myogenesis in aggregated P19 cells (39). Here we show that Meox1 is sufficient to activate cardiomyogenesis in P19 cells. We reexamined stable P19 cell lines overexpressing Meox1, termed P19(Meox1), to further our understanding of the role of Meox1 in cardiomyogenesis. P19(Meox1) and P19(control) cells were aggregated in the absence of Me 2 SO and inspected for the development of cardiomyocytes on day 6. Immunostaining with an anti-MyHC antibody showed that P19(Meox1) cell lines differentiated into cardiomyocytes (Fig. 1H) and not P19(control) cells (Fig. 1B). Hoechst dye staining was included as a control for cell number (Fig. 1G). Altogether, three P19(Meox1) cell lines were isolated that triggered cardiomyogenesis in aggregated P19 cells.
To examine the factors activated in Meox1-induced cardiomyogenesis, Northern blot analysis and RT-PCR were utilized (Fig. 5, panels I and II, respectively). As shown by Northern blot analysis, P19(Meox1) cells expressed high levels of the Meox1 transcript in comparison with P19(control) cells (Fig. 5,  panel I, section A). GATA-4, BMP-4, Nkx2-5, and MEF2C were up-regulated on day 6 in P19(Meox1) cell lines and not in P19(control) cells (Fig. 5, panel I, sections B and C and panel II, sections I and J, respectively).
To deduce a possible mechanism by which Meox1 may be inducing cardiomyogenesis, we hypothesized that Meox1 may be exerting its activity via the Shh signaling pathway. Thus, we looked at the temporal gene expression of Ptc1 and Gli1/2/3. Ptc1 and Gli3, as analyzed by densitometry (data not shown), had no significant change in transcript levels on day 6 when comparing P19(Meox1) cells and P19(control) cells (Fig. 5, panel I, sections D and E, respectively). However, Gli1 was up-regulated on day 6 in P19(Meox1) clones and not in P19(control) cells (Fig. 5, panel I, section F structural genes, resulting in the formation of cardiomyocytes. Meox1 is placed parallel to Gli1/Gli2 in the model (Fig. 6) because mice lacking Meox1/2 still form heart muscle (35). However, it is possible that a delay in cardiac morphogenesis occurred prior to E16.5 in mice lacking Meox1/2. Furthermore, these mice display a similar phenotype to mice lacking Hedgehog signaling (31), suggesting that Meox proteins may mediate the response to Hedgehog signals (35). We had shown previously that Meox1 can up-regulate Gli2 expression (39). The results presented here implicate Meox as a mediator of Hedgehog signaling by regulating Gli1/2 expression in a pathway leading to cardiomyogenesis.
Expression of Shh in P19 monolayer cultures was sufficient to activate the Shh pathway, as shown by the up-regulation of Ptc1 and Gli1 expression. Interestingly, this activation was not sufficient to induce cardiomyogenesis. Consequently, it appears that another factor activated by cellular aggregation is necessary for Shh to initiate cardiomyogenesis. Previous studies have shown that Brachyury T and Wnt5b expressions are up-regulated by cellular aggregation in the absence of Me 2 SO and that BMP4 expression is not (52). Furthermore, Wnt11 plays a role in cardiac induction in the anterior lateral plate mesoderm (53,54). In addition, Wnt3a/8 signaling via ␤-catenin was sufficient to induce cardiomyogenesis in P19CL6 cells (55). Therefore, it is likely that Shh requires the activation of another pathway, such as the Wnt signaling pathway, to mediate the induction of cardiomyogenesis.
Mice lacking the Smo gene did not up-regulate Ptc1 expression, indicating a total lack of Hedgehog signaling (31). These mice still formed heart muscle, although the expression of Nkx2-5 was delayed significantly. Therefore, although the results presented here indicate that Shh is sufficient to initiate cardiomyogenesis in aggregated P19 cells, the data from knockout mice indicate that this pathway is not essential and can be bypassed by other factors/pathways.
In P19(Meox1), P19(Gli2), and P19(Shh) cells there was an up-regulation of the cardiac muscle transcription factors GATA-4, MEF2C, and Nkx2-5. Shh has been shown to activate the expression of Nkx2-2 and Nkx6 -2, leading to the activation of GATA-2 and GATA-3 in a distinct subclass of serotonergic neurons (56). Furthermore, Shh can regulate Nkx-6.1 expression in the ventral spinal cord (57) and Nkx3.1 in caudal somites (58). Whether the Shh signaling pathway is acting directly or indirectly on the cardiac muscle transcription factors remains to be determined. Meox1, Gli2, or Shh expression in P19 cells each resulted in the up-regulation of BMP-4 expression. Hedgehog expression in various endodermal tissues is adjacent to BMP expression in mesenchymal sites, resulting in the hypothesis that BMPs are general targets of Hedgehog signaling (59). In support of this hypothesis, Gli factors can directly regulate the expres-sion of BMP-4 in promoter assays (60). In addition, BMP-4 is directly regulated by GATA-4 (61). Whether Shh functions to initiate cardiomyogenesis by activating BMP-4 remains to be determined.
In P19(Gli2) cells there was an induction of Gli1 expression but not Gli3 expression. This is in agreement with previous studies illustrating that Gli2 can activate the expression of Gli1 (62,63). In P19(Shh) cells there was also an up-regulation of Gli1 expression but no significant change in Gli2/3 expression. Although Gli2 is the primary mediator of Shh signaling, it is not regulated by Shh at the level of transcription. The literature points to the Wnt class of signaling factors in activating the expression of Gli2/3, whereas Shh is thought to regulate Gli2/3 activity during somitogenesis (64,65). Moreover, Gli1 expression, in accordance with other studies, is activated by Shh (66,67).
With the discovery of the Zic family of genes as regulators of Gli function (68), it was interesting to examine their expression pattern in P19(Gli2) and P19(Shh) cells. During the differentiation of each cell line there was no reproducible change in either Zic1 or Zic3 expression in comparison with the differentiation pattern of control cells. Mice lacking Zic3 are defective in heart looping, positioning, and morphogenesis (69). Mice lacking Zic1 and/or Zic2 show no gross heart defects (70).
In summary, Shh signaling via Gli1/2 is sufficient to initiate the expression of cardiac muscle transcription factors, leading to cardiomyogenesis in aggregated P19 cells. This is the first observation of the use of Shh to direct stem cells toward a cardiac muscle lineage, and it has implications for potential approaches to cell therapy.