|Home | About | Journals | Submit | Contact Us | Français|
The TGFβ family member Nodal has been implicated in heart induction through misexpression of a dominant negative version of the type I Nodal receptor (Alk4) and targeted deletion of the co-receptor Cripto in murine ESCs and mouse embryos; however, whether Nodal acts directly or indirectly to induce heart tissue or interacts with other signaling molecules or pathways remained unclear. Here we present Xenopus embryological studies demonstrating an unforeseen role for the DAN family protein Cerberus within presumptive foregut endoderm as essential for differentiation of cardiac mesoderm in response to Nodal. Ectopic activation of Nodal signaling in non-cardiogenic ventroposterior mesendoderm, either by misexpression of the Nodal homologue XNr1 together with Cripto or by a constitutively active Alk4 (caAlk4), induced both cardiac markers and Cerberus. Mosaic lineage tracing studies revealed that Nodal/Cripto and caAlk4 induced cardiac markers cell non-autonomously, thus supporting the idea that Cerberus or another diffusible factor is an essential mediator of Nodal-induced cardiogenesis. Cerberus alone was found sufficient to initiate cardiogenesis at a distance from its site of synthesis. Conversely, morpholino-mediated specific knockdown of Cerberus reduced both endogenous cardiomyogenesis and ectopic heart induction resulting from misactivation of Nodal/Cripto signaling. Since the specific knockdown of Cerberus did not abrogate heart induction by the Wnt antagonist Dkk1, Nodal/Cripto and Wnt antagonists appear to initiate cardiogenesis through distinct pathways. This idea was further supported by the combinatorial effect of morpholino-medicated knockdown of Cerberus and Hex, which is required for Dkk1-induced cardiogenesis, and the differential roles of essential downstream effectors: Nodal pathway activation did not induce the transcriptional repressor Hex while Dkk-1 did not induce Cerberus. These studies demonstrated that cardiogenesis in mesoderm depends on Nodal-mediated induction of Cerberus in underlying endoderm, and that this pathway functions in a pathway parallel to cardiogenesis initiated through the induction of Hex by Wnt antagonists. Both pathways operate in endoderm to initiate cardiogenesis in overlying mesoderm.
Heart induction in vertebrates is initiated during early embryogenesis by specification of paired cardiac primordia within the mesoderm on either side of the dorsal midline. Several classes of secreted proteins have been shown to be capable of inducing heart differentiation in non-cardiac mesoderm, including Dickkopf-1 and Crescent, both of which inhibit the canonical Wnt/β-catenin signaling pathway (Marvin et al., 2001; Schneider and Mercola, 2001) and act through the non-cell-autonomous induction of the homeodomain protein Hex in the endoderm underlying the cardiac mesoderm and regulation of downstream diffusible proteins (Foley and Mercola, 2005a). Nodal proteins are members of the TGFβ superfamily of signaling molecules that are important for numerous developmental processes, including the formation of axial mesoderm [Xenopus, (Agius et al., 2000; Jones et al., 1995) zebrafish (Toyama et al., 1995), mouse (Conlon et al., 1994a; Zhou et al., 1993a) and chick (Bertocchini et al., 2004; Bertocchini and Stern, 2002)] and induction and patterning of the endoderm (Agius et al., 2000; Chang and Hemmati-Brivanlou, 2000; David and Rosa, 2001; Henry et al., 1996). Nodal signaling is transduced to the cytoplasm through a heteromeric receptor complex consisting of the activin type II receptor, the Alk4 type I receptor (Chen et al., 2004; Reissmann et al., 2001) and the EGF-CFC protein Cripto (Gritsman et al., 1999; Yeo and Whitman, 2001). In Xenopus, overexpression of several Nodal family members or a constitutively activated form of Alk4 (caAlk4) has been shown to induce cardiac mesoderm (Reissmann et al., 2001; Takahashi et al., 2000). Disruption of this signaling complex by the use of dominant negative Alk receptors or the genetic disruption of the gene encoding Cripto blocks spontaneous cardiomyocyte differentiation in murine embryonic stem cells (ESCs) (Parisi et al., 2003; Sonntag et al., 2005; Xu et al., 1998a; Xu et al., 1999b), Xenopus embryos (Reissmann et al., 2001) and zebrafish (Griffin and Kimelman, 2002; Reiter et al., 2001). Although these results suggested a potential role for Nodal signaling in heart induction, a mechanistic understanding was elusive in part because previous studies were not designed to isolate a specific cardiogenic function from the broader inductive and patterning effects of Nodal on mesendoderm. In particular, whether Nodal signaling specifically affects cardiogenesis, potential downstream signaling mediators, and potential cooperation with other signaling pathways that pattern anterior mesendoderm all remained to be resolved.
Our study demonstrates that the Xenopus Nodal homologue XNr1 is sufficient to specify an ectopic heart field in non-cardiac mesoderm. Most importantly, mosaic analysis of the induced heart tissue showed that cells expressing either XNr1 with its co-receptor Cripto or caAlk4 were precluded from joining the heart field, suggesting that Nodal signaling cell-autonomously inhibits cardiogenesis while simultaneously stimulating production of a diffusible intermediary that induces cardiogenesis in adjacent cells. Gain and loss of function interventions showed that the secreted Cerberus protein is produced by the cells that respond to Nodal and is essential to initiate cardiogenesis in adjacent cells but is not required for heart induction by Wnt antagonists. Cerberus mRNA is induced directly by Nodal (Osada and Wright, 1999; Piccolo et al., 1999; Yamamoto et al., 2003) in a spatiotemporal domain that localizes precisely to the region of dorsoanterior endoderm required for cardiogenesis (Schneider and Mercola, 1999b). Our study, therefore, illuminates a complex genetic cascade for heart specification that involves signaling through parallel pathways that antagonize Nodal and Wnt activity in the endoderm resulting in production of diffusible signals such as Cerberus.
Embryos were fertilized in vitro, dejellied in 2% cysteine–HCl, pH7.8, and maintained in 0.1x MMR (Peng, 1991) Embryos were staged according to Nieuwkoop and Faber (Nieuwkoop and Faber, 1994). Dorsoanterior marginal zone (DMZ) or ventroposterior marginal zone (VMZ) explants were dissected at stage 10.25–10.5, when the blastopore was clearly discernible. Explant dissections were performed in 0.75x MMR using a fine tungsten needle and processed immediately or cultured in 0.75x MMR until sibling controls had reached appropriate stages. For gene expression analysis, tissues were flash frozen for subsequent RNA isolation or fixed in MEMPFA for in situ hybridization as below.
In situ hybridization was performed according the protocol of Harland (Harland, 1991). Digoxygenin-labelled probes were transcribed from the following linearized plasmid templates (restriction digest, polymerase): pBS-Cerberus (EcoRI, T7); pXαMhc (HindIII, T7); pGEM3Z-Nkx2.5 (XbaI, T7); pGemT-Tbx5 (Not1, T7) and pXTnIc (Not1, T7). Following in situ hybridization, most explants were paraffin embedded and sectioned for analysis.
Synthetic, capped mRNA for injection was transcribed from plasmids pSP6-nls-βgal, pCS2-Dkk1, pCS2-XNr1, pCS2-CA-Alk4, and pCS2-XCer and pCS2-XCer-Short (kind gifts from Eddy deRobertis) using SP6 and T7 mMessage kits (Ambion). All cDNAs used encode Xenopus proteins with the exception of the hAlk4-CA T206D construct that contains a mutated, constitutively active form of the human Alk4 gene. The antisense Cer morpholino oligonucleotide was designed against bases −35 to −11 upstream of the AUG (5′-CTAGACCCTGCAGTGTTTCTGAGCG-3′) as designed and validated by Silva (Silva et al., 2003) and 2 pmol was injected. The rescue was carried out using a pCS2-XCer, which lacks the sequence corresponding to the morpholino. The antisense Hex morpholino was 5′-GGTGCTGGTACTGCATGTCGATTCC-3′ (Foley and Mercola, 2005b; Smithers and Jones, 2002) and injected at 1pmol. The standard control morpholino provided by Gene Tools was also injected at 2 pmol. Depending on the experiment, either a 10,000 MW Alexa 546 lysinated dextran (AD546, Molecular Probes) or mRNA encoding nls-β-gal were co-injected as lineage labels.
10–12 explants were pooled from each round of injections and RNA was produced using Qiagen Rneasy Lipid Tissue Kit (Qiagen). First strand synthesis was carried out using a poly-dT primer and Superscript II Reverse Transcriptase (Invitrogen). Real time PCR was performed on a Roche Light Cycler using the Light Cycler FastStart DNA master SYBR Green I kit (Roche). Quantification was carried out by normalizing levels to amount of total cDNA using the ubiquitously expressed EF1α transcript. The EF1α primer spans an intron and PCR in the absence of RT thus confirmed that RNA samples were free from genomic DNA contamination. Q-RT-PCR was performed on 6 biological samples and run data shown as the average of two technical replicates.
Primer Pairs were as below:
Cerberus:(+) 5′-AAGCCTGGTGCCAAGAT-3′; (−) 5′-ATTCAGCGTCAGGTGGTT-3′
EF1A: (+) 5′-TACCCTCCTCTTGGTCGTT-3′; (−) 5′-GGTTTTCGCTGCTTTCTG-3′
Hex: (+) 5′-GTG GCTACTTACCGGAC-3′; (−) 5′-CCTTTCCGCTTGTGCA-3′
Synthetic mRNAs encoding XNr1 and Cripto were co-injected at doses ranging from 5 pg each to 100 pg each into the non-cardiac ventroposterior mesoderm (VMZ) of 4–16 cell stage Xenopus embryos. When embryos reached early gastrula stage (Nieuwkoop and Faber st.10.25–10.5), VMZ regions were explanted and cultured in isolation until age-matched siblings reached st. 23–25. 25 pg or higher of each co-injected mRNA consistently yielded robust expression of the heart field markers Nkx2.5 and Tbx5 visualized by in situ hybridization (Nkx2.5, 51% of explants were positive, n=109; Tbx5, 49%, n=108) (Figure 1B,D). In subsequent experiments, 50 pg of each mRNA was injected unless noted otherwise. When cultured to st. 30–32, similarly processed explants expressed transcripts encoding the cardiac contractile proteins αMhc and TnIc (αMhc, 27%, n=26; TnIc, 48%, n=72) (Figure 1C,E). Although Xnr-1/Cripto induced robust expression of both early and late cardiac markers, beating heat tubes were rarely observed and, when examined in histological section, we did not observe the characteristic architecture of tube with open lumen that is typical of native myocardium or VMZ explants induced by Dkk1 (Schneider and Mercola, 2001) at this developmental stage (note absence of lumen in figures 1C,E, and data not shown). Similar results were obtained with caAlk4 (Nkx2.5, 14%, n=135; TNIc 19%, n=48) (Figure 1F,G)
Immunoprecipitation and chemical cross-linking experiments show that Cripto, the prototypic member of the epidermal growth factor-related Cripto-FRL1-Criptic (EGF-CFC) family, interacts with Nodal and Alk4, suggesting its role as a co-receptor necessary to recruit Nodal to receptor complexes (Gritsman et al., 1999; Yeo and Whitman, 2001). Cripto has also been reported to signal independently of Nodal and Alk4 to activate the MAP kinase and Akt pathway (Bianco et al., 2003; Yan et al., 2001). To determine if Cripto has Nodal-independent cardiac-inducing activity, we injected Cripto mRNA alone as above and assessed the induction of cardiac markers. We injected 50 pg, which in our hands was the optimal dose for co-injection with Xnr-1, as well as two higher doses, 100 pg and 200 pg. None of the doses tested showed any induction of cardiac gene expression or morphological evidence of cardiac differentiation (n=76, data not shown); therefore, we conclude that the cardiogenic function of Cripto is to potentiate Nodal/XNr1signaling.
To test whether XNr1/Cripto signaling acts directly on the cardiac mesoderm, we co-injected Xnr-1 and cripto mRNAs along with the fluorescent Alexa-dextran 546 (AD546) as a lineage label into single ventral blastomeres of 8–16 cell stage embryos (Figure 1A). VMZ explants were prepared as above and cultured until age-matched siblings had reached either stage 25–28 or stage 30–32, when they were processed for in situ hybridization and examined for localization of heart field markers Nkx2.5 and Tbx5, or the cardiac contractile protein genes αMhc and TnIc relative to the lineage label in histological section. In all cases, the lineage label, which marks the progeny of the injected cells, was located adjacent to but not overlapping the induced cardiac tissue (Figure 1B-E). Similarly, caAlk4 induced cardiac mesoderm cell non-autonomously, although in some explants we observed minor overlap indicating that activation of the Nodal signaling pathway does not obligatorily preclude heart induction (Figure 1F,G).
The preceding results suggested that Nodal/Cripto and Alk4 signaling initiates cardiogenesis indirectly by triggering production of a diffusible inducing factor. Previous studies had mapped the spatial expression of two genes, Cerberus (Cer) and the homeodomain transcription factor Hex, coincident with the region of deep dorsoanterior endoderm that stimulates heart induction (Schneider and Mercola, 1999b). Whereas the transcriptional repressive function of Hex is essential for heart induction by the Wnt antagonist Dkk1 (Foley and Mercola, 2005a), Cerberus has not been placed in a heart-inducing cascade. To test if Cerberus is a component of the Nodal pathway, two ventral blastomeres of 4–8 cell stage embryos were injected with XNr1 and Cripto mRNAs or with activated Alk4 mRNA and VMZ explants were isolated when embryos reached early gastrula stages either immediately at stage 10.25–10.5 or when age-matched siblings had reached stages 11, 13 or 19. Both injected XNr1/Cripto and caAlk4 induced Cer transcription (Figure 2). Interestingly, the kinetics recapitulated the transient pulse of expression seen during normal embryonic development with maximal levels of induction at early gastrula stages (st. 10.5–11), declining after gastrulation (st. 13) such that ectopic transcripts were no longer detectable by early neurula stages (st. 19) (Figure 2A and 2G-I). In a mosaic assay, injection of single ventro-posterior blastomeres with XNr1/Cripto and either β-galactosidase mRNA or AD546 as a lineage label followed by isolation of VMZ explants at st. 10.25–10.5 and examination on histological section revealed that the majority of ectopic Cer occurred in the progeny of the injected cells (Figure 2B-E). Induction was however, not strictly cell-autonomous. We observed instances of cells adjacent to the lineage-labeled cells that also expressed Cer, as best visualized using a nuclear localized lacZ to trace cell lineage (Figure 2D-E). Interestingly, ectopic Cer never occurred in the deep endoderm (Figure 2D,E red arrowheads) or in the superficial pigmented layer (Figure 2F), despite many instances of staining with the lineage label indicating descent from blastomeres injected with the Nodal pathway activator reflecting, possibly, either a requirement for additional mesendoderm-specific signaling or the presence of signals that alter Nodal responsiveness as has been demonstrated by (Engleka et al., 2001). By contrast, induction of Cer by caAlk4 was almost entirely cell non-autonomous (Figure 2G-I) suggesting responsiveness to Nodal signaling may be required for the maintenance of Cerberus expression. Ectopic Cer-positive cells in the VMZ exhibited a remarkable involution behavior reminiscent of DMZ deep mesendoderm during gastrulation (Figure 2G-I), indicating that ectopic Nodal pathway activation elicited a normal developmental response.
To test Cerberus involvement in heart induction, Cer mRNA was co-injected with the lineage label AD546 into one ventral blastomere at the 8–16 cell stage and explants were isolated at st. 10.25–10.5 and cultured until age-matched siblings had reached st.23–25 when they were analyzed for expression of the early cardiac markers Nkx2.5 and Tbx5. Injected explants showed robust expression of early cardiac markers (Nkx2.5, 69% of explants were positive, n=58; Tbx5, 72%, n=113). Interestingly, Cerberus-injected explants did not show expression of markers for differentiated cardiac mesoderm (data not shown). The lineage label was most often distant from cells expressing the induced Nkx2.5 or Tbx5 (Figure 3A,B), consistent with the idea that localized production of the secreted Cerberus protein induces cardiac mesoderm in nearby tissue. The distance from lineage label to ectopic cardiac mesoderm was generally further than that seen in VMZ explants injected with XNr1/Cripto or caAlk4, perhaps suggesting a thresholded response to Cerberus or that the cells that produce Cerberus and cells that express early heart markers might migrate away from each other during the time course of the experiment.
A requirement for Cerberus was evaluated by injecting the antisense morpholino oligonucleotide (Cer MO), (Silva et al., 2003) with XNr1 and Cripto mRNAs into 1 ventral blastomere of the 4-cell stage embryo at doses that give robust cardiac induction. Cer Mo significantly decreased the incidence of heart induction by XNr1/Cripto (7.4%, n=121, when co-injected with Cer MO versus 34.3% explants were positive in absence of Cer MO, n=137, Figure 3C,G,G′). In separate experiments, Cer Mo also decreased XNr1/Cripto induction of αMHC (6.3% in presence of Cer MO, n=16, versus 32.8% positive in absence, n=22), and Nkx2.5 (7.3% in presence of Cer MO, n=82, versus 44.6% positive in absence, n=148, Figure 3F,F′).
Having demonstrated that Cerberus is essential for ectopic cardiogenesis induced by Nodal pathway activation, we then tested for a role in endogenous heart development by injecting 2 pmol of Cer MO into the 2 dorsal blastomeres of 4-cell stage embryos. As for ectopic induction, 2 pmol doses were able to significantly reduce expression of each of the cardiac markers we examined relative to controls injected with either control morpholino or injected with an inert mRNA, including: Nkx2.5 (58.5% positive in presence of Cer MO, n=41, versus 94%, n=42, for controls, Figure 3D, D′); Tbx5 (65.4% in presence of Cer MO, n=52, versus 78.9%, n=38, for controls); TnIc (35.9%, n=39, in presence of Cer MO versus 75%, n=20, for controls, Figure 3 E, E′), and αMHC (59.8%, n=82, versus 68.8%, n=32, for controls). Consist with the observations of Silva et al. (Silva et al., 2003), head structures were essentially unaffected by morpholino injection (see Figure 3D′, E′), supporting the idea that Cerberus plays specific roles in mesendoderm patterning despite potency as an ectopic head inducer. Co-injection of 25 pg of pCS2-XCer plasmid (which lacks the sequence targeted by Cer MO) with 2 pmol of the Cer MO rescued the expression of Nkx2.5 (Figure 3D′ inset) (85.7%, n=42), confirming selectivity of Cer MO for the endogenous mRNA. Taken together, our results implicate Cerberus as an essential effector of the Nodal pathway for cardiogenesis.
Since Dkk-1 also induces ectopic cardiac gene expression in VMZ explants (Schneider and Mercola, 2001), we next examined involvement of Cerberus in this pathway. Dkk-1 was injected into 1 ventral blastomere at the 4-cell stage either alone or in combination with 2 pmol of Cer Mo. As shown in Figure 3C,H,H′, Cer MO did not interfere with the ability of Dkk-1 to induce cardiac marker expression (55.3%, n=38, of explants expressed TnIc when injected with Dkk-1 alone as opposed to 71%, n=62, when Dkk-1 was co-injected with Cer MO).
This result suggests that the Nodal and Dkk1 pathways are distinct at the level of Cerberus. We had shown that the homeodomain transcription factor Hex is an essential component of heart induction by Dkk1 (Foley and Mercola, 2005b), raising the possibility that cardiogenesis induced by Nodal pathway activation might also require Hex. Ectopic Hex transcripts were not induced in VMZ tissue by targeting injections of XNr-1/Cripto mRNAs, at doses ranging form 25 pg each up to 100 pg each, into single blastomeres of the 4-cell stage embryo. VMZ explants were dissected at st. 10.5 and snap frozen until analysis of Hex transcript levels by quantitative RT-PCR normalizing to that for EF1α. Induction of Hex transcript was not seen in response to injections of XNr1/Cripto (data not shown). Similarly, Cerberus also did not induce Hex, and previous results had shown that Dkk1 does not induce Cer (Foley and Mercola, 2005b). We then used the specific morpholinos to verify that the pathways are distinct by loss-of-function experiments. Embryos were injected into two dorsovegetal blastomeres at the 8 cell stage with either 2 pmol of Cer Mo or 1 pmol of Hex MO, levels which submaximally abrogate early cardiac induction (Figure 4E; 84.1%, n=44, for Cer MO and 52.2%, n=46, for Hex MO as compared to 95.4%, n=109, for Control MO). Co-injection of Cer MO and Hex MO at these conditions more severely attenuated heart induction than either alone (Figure 4E; 34%, n=47), indicative of a genetic interaction. Together, these experiments demonstrate that Nodal pathway activation does not overlap with Dkk1 induction of cardiogenesis, at least at the levels of the downstream effectors Cerberus and Hex.
Cerberus is a multifunctional protein that antagonizes Wnt, Nodal and BMP signaling. In order to begin to understand the basis for heart induction by Cerberus, we injected mRNA that encodes a truncated form of the protein known as Cerberus-Short (Cer-S), which does not antagonize Wnts or BMPs but retains the ability to block Nodal signaling. Cer-S was unable to initiate cardiogenesis, since only 1 explant out of 78 showed ectopic Nkx2.5 expression and only 1 out of 52 expressed ectopic Tbx5 when injected into single blastomeres of VMZ and assayed as for Cerberus above. Similarly, VMZ explants injected with a truncated form of the BMP receptor (tBR), which blocks BMP signaling specifically (Graff et al., 1994) also did not show consistent ectopic Nkx2.5 (1.5%, n=65) or Tbx5 (2.5%, n=40). Since neither a Nodal nor BMP antagonist induced Nkx25 or Tbx5, these results implicate Wnt antagonism as a component of the cardiogenic function of Cerberus, but this hypothesis cannot be verified directly since the Wnt antagonizing activity is not known to be retained on a fragment that lacks Nodal- and BMP-inhibiting activities. Cerberus within the embryo might function in combination with other signals and these can be recapitulated in part by injection of tBR, which induces secondary body axes that are both devoid of anterior structures (Graff et al., 1994; Suzuki et al., 1994) and lack early cardiac gene expression (Nkx2.5, 0%, n=18). In contrast to injection of either alone, co-injection of Cer-S complemented tBR and robustly induced Nkx2.5 in 44.2% (n=52) and Tbx5 in 50% (n=32) of injected VMZ explants (Figure 4F-I), implicating inhibition of Nodal signaling as a component of Cerberus cardiogenic activity. The contribution of multiple functions of Cerberus to heart field induction are discussed below.
Our results revealed multiple ways in which Cerberus interacts with the Nodal signaling pathway to induce cardiogenic mesoderm in Xenopus. Although Nodal family members have been implicated in heart induction, their role in this process has not been distinguished from their more general ability to induce mesoderm and endoderm. For example, high doses of recombinant Activin, which has overlapping affinity with Nodal for cell surface receptors, can induce cardiogenesis in Xenopus animal cap and chick epiblast tissue (both primitive ectoderm) (Logan and Mohun, 1993; Yatskievych et al., 1997); however, since both mesoderm and endoderm form in these experiments, whether cardiogenesis results from a direct effect of receptor activation or by an interaction between these tissues has not been resolved. One clue as to how this family of molecules might specify specific mesodermal fates came from the finding that Activin (Green and Smith, 1990; Symes et al., 1994), as well as Nodal and its homologues can act in a concentration and/or duration-sensitive manner to specify fate and behavioral differences of induced mesendoderm in zebrafish (Gritsman et al., 1999; Reiter et al., 2001), Xenopus (Agius et al., 2000; Chang and Hemmati-Brivanlou, 2000; Faure et al., 2000; Lee et al., 2001) and mouse (Lowe et al., 2001; Meno et al., 1999; Norris et al., 2002), suggesting a model whereby pre-chordal and cardiac mesoderm, as well as endoderm, require higher, earlier and/or more sustained levels of Nodal signaling than does formation of more posterior mesoderm. Our results suggest a spatially and/or temporally complex model in which Nodal initiates cardiogenesis through induction of a diffusible signal that we showed to consist of Cerberus by both gain and loss of function experiments.
Cerberus is a multifunctional protein that inhibits Wnt, BMP and Nodal activity. Consequently, the unexpected finding that the truncated Cer-S protein, which contains the Nodal inhibitory activity, retained the ability to initiate ectopic cardiogenesis in secondary body axes induced by tBR points to the possibility that Cerberus participates in a feedback mechanism to establish the appropriate level and/or timing of Nodal signaling needed for heart field specification (diagrammed in Figure 4J). Moreover, the feedback inhibition of Nodal might account for the tight control of endogenous Cer transcripts, which accumulate only at the onset of gastrulation in a pattern that coincides precisely, both temporally and spatially, with the domain of heart-inducing activity within the deep dorsoanterior endoderm (Schneider and Mercola, 1999a). Tight control of endogenous Cer expression is likely to be essential for heart induction because persistent expression adjacent to the heart field after gastrulation would be expected to interfere with BMP, which is required for continued cardiogenesis in Xenopus [reviewed in (Foley and Mercola, 2004)], and this might explain why Cerberus injection induces only markers of the heart field (Tbx5 and Nkx2.5) but not later genes encoding contractile proteins indicative of cardiomyogenesis.
The fact that Cer-S injected alone into non-cardiogenic mesoderm did not induce ectopic cardiac tissue suggests that the truncated protein lacks an additional activity that contributes to heart induction. Since the BMP antagonist tBR alone did not initiate cardiogenesis while co-injection with Cer-S was sufficient, it seems likely that BMP antagonism is involved. BMP is known to be required for heart induction, but Xenopus studies place its function slightly later to maintain the induced state (Shi et al., 2000). Finally, we cannot rule out the involvement of the Wnt antagonist activity of Cerberus. Importantly, the finding that Cerberus did not induce Hex expression, which is induced by inhibition of canonical Wnt signaling in VMZ, and that Dkk1 induction of cardiac gene expression was not blocked by Cer MO indicate that the Dkk1/Wnt antagonism and Nodal pathways operate in parallel at least at the level of these downstream effectors (Figure 4J).
Results from prior studies of mouse embryonic and ESC cardiogenesis are consistent with the idea of a diffusible signal downstream of Nodal, but the nature of the factor was not elucidated. Nodal−/− (Conlon et al., 1994b; Zhou et al., 1993b) and Smad2−/− (Waldrip et al., 1998) mouse embryos fail to develop a primitive streak and die before anterior structures, including heart, form and they also do not express markers of the anterior visceral endoderm (AVE), an extraembryonic tissue that shares gene expression profiles and inductive properties with the deep dorsoanterior endoderm in Xenopus. An elegant series of chimera experiments by Robertson and colleagues (Brennan et al., 2001; Varlet et al., 1997) overcame the early lethality and showed that Nodal signaling within the AVE and epiblast is essential to specify anterior structures. Although mouse embryos lacking Cripto produce mesoderm, they lack anteroposterior axial development and do not form cardiomyocytes (Ding et al., 1998; Xu et al., 1999a). Similarly, Cripto−/− ESCs also form mesoderm and cardiogenesis is blocked (Xu et al., 1998b). Comparison of the phenotypes of Cripto−/− and Nodal−/− embryos indicates that Cripto is essential for Nodal to signal in both the epiblast and the AVE that overlies the epiblast. Moreover, considerable evidence shows that the AVE is required for induction of heart (Arai et al., 1997) and other anterior structures (Beddington and Robertson, 1999), suggestive that Nodal signaling from this tissue initiates a Cripto-dependent cascade that is essential for heart induction. In favor of a diffusible signal downstream of Nodal, Cripto−/− mESCs introduced into wild type mouse blastocysts efficiently contribute to the heart (Xu et al., 1999a), demonstrating that secreted or membrane localized factors from the wild type cells act on the mutant cells. Although a shed version of Cripto has been postulated as a potential mediator (Parisi et al., 2003), the murine data could also be explained by the involvement of the murine Cerberus homologue mCer1/Cerberus-like (Belo et al., 1997; Biben et al., 1998; Rhinn et al., 1998). Whether mCer1 acts downstream of Nodal/Cripto and Alk4 signaling in a cascade that contributes to cardiogenic differentiation of mesoderm in the mouse is under investigation.
We thank Hanh Nguyen for expert technical assistance with in situ hybridization. AF gratefully acknowledges the support of an NRSA fellowship (F32 HL69595) and a postdoctoral fellowship from the California Institute for Regenerative Medicine (CIRM). MM gratefully acknowledges support from NIH/NHLBI (R01 HL059502 and R01 HL067079).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.