Noggin inactivation affects the number and differentiation potential of muscle progenitor cells in vivo

Inactivation of Noggin, a secreted antagonist of Bone Morphogenetic Proteins (BMPs), in mice leads, among others, to severe malformations of the appendicular skeleton and defective skeletal muscle fibers. To determine the molecular basis of the phenotype, we carried out a histomorphological and molecular analysis of developing muscles Noggin−/− mice. We show that in 18.5 dpc embryos there is a marked reduction in muscle fiber size and a failure of nuclei migration towards the cell membrane. Molecularly, the absence of Noggin results in an increased BMP signaling in muscle tissue as shown by the increase in SMAD1/5/8 phosphorylation, concomitant with the induction of BMP target genes such as Id1, 2, 3 as well as Msx1. Finally, upon removal of Noggin, the number of mesenchymal Pax7+ muscle precursor cells is reduced and they are more prone to differentiate into adipocytes in vitro. Thus, our results highlight the importance of Noggin/BMP balance for myogenic commitment of early fetal progenitor cells.

Yet not all Pax7 + cells proliferate and differentiate but some remain in an undifferentiated state and become activated in the postnatal life following triggers like trauma or physical exercise. These cells are called satellite cells when they can be morphologically identified as mononucleated cells residing between the myofiber plasma membrane and the basal lamina (from 17.5 dpc onwards). They are considered to form the stem cell niche responsible for the growth and restoration of the muscle 9 .
The dermomyotomal and sclerotomal somitic populations are subject to the intricate crosstalk of several signaling cascades including WNT, Sonic hedgehog (SHH), and Bone Morphogenetic proteins (BMPs), ensuring a regulated differentiation of these lineages. WNT signaling from the overlying epidermis and the roof plate of the neural tube induces the expression of dermomyotome specific genes, while SHH signaling from the notochord and the floor plate of the neural tube induces sclerotomal gene expression 10 . In addition, BMP expression in the epidermis, the roof plate of the neural tube and the lateral plate mesoderm prevent the differentiation of myogenic lineage 11 . Differently, Noggin, present in the roof and floor plate of the neural tube, blocks this BMP action and therefore allows for the myogenic precursors to differentiate 12,13 . This balance between multiple signaling pathways results, among others, in the restricted expression of myogenic regulatory factors (MRF) and Pax genes in myogenic cell populations 14 .
Besides its role during the patterning of the somite, BMP signaling also affects the differentiation of myofibers. The effect of BMP signaling was shown to depend on the developmental stage and the progression along the myogenic program. Whereas the differentiation of embryonic myoblasts was shown to be insensitive to BMP signals, the fetal myoblasts and the Pax7 + precursors require a decrease of the BMP signaling in order to allow further myogenic differentiation 8 .
We have reported before that the Noggin null (Noggin −/− ) mice presented a muscle defect during the final stages of in utero development 4 . From early stages onwards the size of the muscles was reduced. At 18.5 dpc, the Noggin −/− embryos displayed a dramatic muscle phenotype characterized by disorganized myofibers that failed to align. In addition, the number of multinucleated myofibers was reduced, more mononucleated myoblasts were found and the nuclei failed to migrate towards the plasma membrane in the Noggin −/− muscles. The defects appear late during the development since we could not detect them at 16.5 dpc, indicating that the phenotype was induced between 16.5 dpc and 18.5 dpc (fetal myogenesis). While we made that observation, the molecular basis of that phenotype was not investigated. Here we focused on the molecular mechanisms underlying the phenotype. Our results suggest that the absence of Noggin, and consequently BMP hypersignaling, leads to the decrease of the number of muscle progenitor cells and their phenotypic shift towards adipocytes in vitro.

Results
Noggin −/− muscle fibers display a reduced fiber thickness. The limb muscles of Noggin −/− mice are strongly affected, with some of them missing and others severely malformed. Thus, to carry out a meaningful phenotypic analysis that would focus on the primary molecular differences rather than consequences of severe muscle malformation, we needed to select anatomically the same muscle in both genotypes. We performed H&E staining and compared the forelimb anatomy using the Jatlasviewer (www.emouseatlas.org/emap/analysis_tools_ resources/software/jatlasviewer.html). We identified musculus flexor carpi ulnaris (indicated in red in Fig. 1A,B) as the muscle clearly identifiable and the least malformed in both wild type and Noggin null genotypes (Fig. 1C,E). For the analysis, we focused on three different developmental stages: one without an apparent defect (16.5 dpc), a dramatic defect (18.5 dpc) and the stage in between (17.5 dpc) ( Fig. 1C-E').
Since the histological appearance of the muscle suggested defective fiber thickness, we performed an F-Actin staining, using Phalloidin, and measured thickness of muscle fibers by measuring the muscle fiber diameter using ImageJ software. No difference in the overall morphology or in fiber thickness could be detected at 16.5 dpc in both genotypes. At 17.5 dpc Noggin −/− myofibers were slightly but significantly thinner and this trend persisted until 18.5 dpc where the difference was apparent ( Fig. 1F-H'). Quantitative analysis of this observation confirmed the histological phenotype (Fig. 1I).
The loss of Noggin increases BMP signaling in embryonic muscles. Canonical BMP signaling phosphorylates SMAD1/5/8 and this phosphorylation status is used to detect active BMP signaling. Since Noggin is a BMP antagonist, its absence should result in augmented BMP signaling. Therefore we investigated the activation of the SMAD1/5/8 proteins. The immunohistochemical analysis for the phosphorylated SMAD1/5/8 (P-SMAD1/5/8) displayed an increase in the number of positive cells in the muscle already from 16.5 dpc onwards as quantified in Fig. 2A and shown in Fig. 2B-C' . To confirm that observation we carried out Western Blot (WB) analysis for both P-SMAD1/5/8 and IDs proteins levels (Fig. 2D,E respectively) in the same muscle. The results of the P-SMAD1/5/8 WB has been quantified (Fig. 2F).
To verify the presence of active BMP signaling in the Noggin −/− musculus flexor carpi ulnaris, we investigated the expression of BMP immediate target genes. Using qRT-PCR at 16.5 dpc, we detected apparent upregulation of Msx2 and significant upregulation of Id3 and Msx1 (Fig. 2G).
Thus, at 16.5 dpc we detected the onset of molecular changes preceding the histological defects detectable 24 hours later.
The loss of Noggin does not affect the rate of proliferation. One of the functions of BMP signaling is to promote the exit from the cell cycle 15 . Thus one might expect that in the absence of Noggin, the cellular proliferation could be impaired leading to a decrease in the size of the muscle 16 . To address this issue, we analyzed the mitotic activity of the limb muscle cells using a KI67 antibody. KI67 is a nuclear protein present mainly in the G2, M, and at the end of the S phase and its presence is associated with active cell proliferation 17 . Despite the differences in muscle size and fiber thickness, no significant difference in cell proliferation could be detected in Noggin −/− muscles between the 16.5 dpc (no observable muscle phenotype) and 18.5 dpc (a strong phenotype) (Fig. 3).
Noggin −/− muscles have less Pax7 + progenitor cells. The cell proliferation was not affected in the muscle of Noggin null mice and the onset of muscle phenotype was relatively late during embryogenesis. Thus we hypothesized that perhaps not the muscle induction but rather muscle growth could be affected by excessive BMP signaling. During the muscle growth, Pax7 + cells provide the reservoir of stem cells that selectively enter myogenic differentiation and contribute to the growth of the muscle fiber during late embryonic development and in postnatal muscle repair processes 18,19 . Since BMP has a known negative effect on myogenesis (reviewed in ref. 20), we hypothesized that perhaps the number of Pax7 + cells in the late embryonic muscle could be reduced in the absence of Noggin. To investigate that, we carried out immunohistochemical analysis of the developing limb muscle for the expression of Pax7. Following the staining, we counted the number of positive cells using ImageJ software. Indeed, the number of Pax7 + cells was significantly reduced in Noggin −/− muscles throughout the stages investigated (Fig. 4). Impairment of myogenic potential in Noggin null progenitors. The cell proliferation was not affected in Noggin −/− mice, but the number of Pax7 + progenitor cells was reduced. To exclude a defective proliferation, we analyzed the levels of P-ERK 21,22 . To accomplish that, fetal myoblast progenitors from Noggin −/− and wild type embryos were isolated from 16.5 dpc limbs and cultured under myogenic conditions. No significant differences were detected in the levels of P-ERK, suggesting no defects in the proliferation rate were present in the muscle progenitors of Noggin −/− mice (Fig. 5A,B).
Next, to address the differentiation potential of progenitor cells, Noggin −/− cells were analyzed by qRT-PCR at 0 and 3 days of differentiation in vitro (Fig. 5C). At day 0, no significant differences in the expression levels of myogenic markers between Noggin +/+ and Noggin −/− cells were detected, with the exception of MyoD mRNA. After 3 days in culture, the expression of the myogenic progenitor markers, Pax3 and Pax7, decreased significantly irrespective of the genetic background. The expression levels of myogenic markers like MyoD, Myf5 and Myogenin increased in the wild type while decreased significantly in the Noggin −/− mouse cells (Fig. 5C). Thus as expected 23,24 , under the in vitro differentiation conditions used, the Pax3 + and Pax7 + cells did not maintain growth. At the same time, in the absence of Noggin, myogenic differentiation was repressed.
To ensure that the analyzed cell population was free of other cells that might affect the outcome of the experiment, we tested for the presence of cell types commonly contaminating such cultures, such as fibroblasts or chondrocytes. No large differences in the expression levels of Collagen II (Coll II) and SOX9 25   nuclei in Noggin −/− cultures was reduced to respectively 4% ± 2.26 and 15.8% ± 7.8 of Noggin +/+ values (Fig. 6A). After three days under the differentiation condition the number of MyoD + cells was significantly reduced in Noggin −/− fetal myoblasts as compared to controls (8% ± 4 of Noggin +/+ cells; Fig. 6A). The myoblasts during the later stages of differentiation fuse to form myotubes (Suppl Fig. 1A). Thus we determined the fusion index calculated as the percentage of nuclei inside of MyHC + and α -Sarcomeric Actinin + myotubes to the number of total nuclei of Noggin −/− cultures and showed that it was also markedly decreased ( Fig. 6A and Suppl. Fig. 1B).
To semi-quantify these observations we carried out a protein expression analysis by WB. While MyoD and MyHC protein levels were detectable in Noggin +/+ cells, they were almost undetectable in Noggin −/− cells at early (d0) and late (d3) stage of differentiation ( Fig. 6B-E).
In addition, the levels of embryonic MyHC (eMyHC) strongly expressed in Noggin +/+ cells late stage, were barely detectable in Noggin −/− cells at both time points of differentiation (Suppl Fig. 1C).
Aberrant adipogenic potential in Noggin −/− progenitors. It has been reported that BMP signaling pathway is also involved in adipocyte commitment and differentiation in vitro and in vivo 29,30 . To investigate this aspect of cellular differentiation in Noggin −/− cells, we tested their adipocyte commitment in vitro by growing them in adipogenic medium. Noggin +/+ progenitor cells were able to shift towards an adipocyte fate, even if they were still able to generate myotubes (Suppl Fig. 1D). Noggin −/− cells, however, only differentiated towards adipocytes, even more efficiently compared to Noggin +/+ cells (1.7 ar ± 0.2 vs. 1 ar ± 0.03) as shown by quantification of Oil-Red-O staining (Fig. 7A,B).
The molecular signature of these cells further supported their adipogenic phenotype. Noggin −/− cells showed a significant increase in the expression of the majority of the adipocyte genes after 6 days of adipogenic induction as shown by qRT-PCR data ( Fig. 7C and Table 1-please note the logarithmic scale on the Figure). Notably, at day 0 Perilipin (166% ± 1.5) and Adiponectin (151% ± 1.6) were already present at increased levels as compared to the Noggin −/− cells (Fig. 7C).

Discussion
We have previously reported on a muscle defect in Noggin −/− mice 4 . Here we investigated the molecular basis of that phenotype. Interestingly, the defect has a late onset, indicating that the initial muscle induction and formation occurs normally. At the later stages however, when the muscle acquires more mass, it becomes sensitive to the absence of Noggin. Our analysis indicates that the loss of Noggin in the mouse embryonic muscle resulted in an increased BMP signaling together with a decrease in the relative number of Pax7 + cells. Intriguingly, the differentiation defect persisted in in vitro setting suggesting that it was cell autonomous. Specifically the cells lost the ability to enter the myogenic pathway while maintained the adipogenic one.
Our analysis shows that the loss of Noggin during mouse embryonic development results in an increased BMP signaling in developing muscle. Molecularly it has been shown that the inhibitory effect of BMP signaling on the myogenic program is a result of the sequestering of the E-proteins by the immediate BMP target genes (Id1, 2, 3) 31,32 . Thereby the ID proteins block the MRF-mediated activation of the myogenic program, which is dependent on the binding of the MRFs to these E-proteins. Furthermore, Msx1 was shown to block cellular differentiation by preventing cell cycle exit and to antagonize the myogenic activity of Pax3 in migrating limb muscle precursors 33,34 . Since the BMP target genes are upregulated in the Noggin −/− mice, the sequestering of the E-proteins, the inhibition of cell cycle exit and/or the loss of differentiation of the migrating limb precursors are all potential mechanisms underlying the Noggin −/− muscle phenotype.
The BMP signaling was shown to regulate the patterning of the somite and the myogenic program 3 , thus a muscle defect in BMP antagonist mutants could have been anticipated. However, the extent of the muscle defect in the Noggin −/− mice and the late onset was far greater than reported for other BMP antagonist mutants 35 . Reshef et al. did suggest that the epaxial musculature in the Noggin −/− mice was largely absent in the posterior part of the embryo, but no further research was conducted on the muscles of the Noggin −/− mice 36 . The work presented by Lassar's group did in fact lay a foundation for our work. Reshef et al. have shown that the precise regulation of BMP signaling by Noggin and other antagonists, controls the entry of Pax3 + cells into myogenic differentiation program defined by the expression levels of MyoD and Myf5. This regulatory loop is therefore essential for the myogenesis within the somites in vitro and in ovo. We have extended the work of Reshef et al. by showing that the same regulatory loop used also at later stages of development in mouse. Specifically, we show that, in Noggin −/− embryos, not only the muscle fibers exhibit a morphological defect but the maintenance of the fetal myogenic progenitor pool is impaired as well. In fact, we show for the first time that Noggin −/− myogenic progenitors are unable to properly differentiate into the myogenic lineage, but display a more efficient commitment to adipogenic lineage.
In vitro, BMP signaling was shown to balance the proliferation and differentiation of muscle satellite cells. Upon differentiation, BMP signaling is downregulated resulting in the activation of the myogenic program and formation of myofibers 37,38 . In contrast, when BMP was applied ectopically in the chicken wing, the number of muscle fibers and Pax7 + cells increased 14,16 . However, when BMP was applied during earlier stages of chicken development, similar chondrogenesis and myogenesis defects were seen as in the Noggin −/− mice 39 . These results strongly suggest that the timing and the localization of the BMP signal and the differentiation status of the target cells determines the molecular response explaining the apparent contradiction in the results published by us and among these research groups.
It was suggested that cells in the somitic environment could have different intrinsic properties as they respond differently to the same signals 40 . For example, embryonic myoblasts were shown to be insensitive to BMP signaling allowing them to stop proliferating and differentiate into the primary myofibers. For fetal myoblasts and satellite cells, however, it was shown that the level of BMP signaling has to be reduced, by either reducing the expression of BMP ligands, or antagonizing the signaling, in order to let differentiation and fusion into secondary  fibers occur 16 . This potentially explains the Noggin −/− muscle phenotype, as the increased BMP signaling in the muscle would have no effect on the differentiation of the primary fibers, but it could inhibit the generation of the secondary fibers. Therefore fetal myoblasts, which normally give rise to the secondary fibers, maintain the committed but undifferentiated state. Indeed, we observe mononucleated myoblasts in the Noggin −/− muscle, concomitant with a reduction of multinucleated myofibers. The limited number of primary fibers that is formed in the absence of Noggin results in a muscle that fails to maintain its organization.
Another explanation of the observed phenotype could be based on a possibility of lineage switch. In fact, the myogenic potential of C2C12 cells can be fully blocked by the addition of BMP ligands, which then promote osteochondrogenic differentiation of this cell line 41,42 and reviewed in ref. 43. Thus we hypothesize that a number of muscle progenitor cells are drawn into the chondrogenic lineage in the Noggin −/− limb 44,45 . This would deplete the pool of muscle precursors in a way that initially enough cells are left to maintain the muscle organization, but at later stages, when the muscles grow rapidly, there are not enough cells to supply the muscle. In this view, the loss of Noggin induces SMAD-dependent increase in BMP signaling that results in an inhibition of fetal myoblast differentiation and a reduction in the number of satellite cells, which causes a severe late-onset muscle defect. At the same time, the prechondrogenic pool increases, leading to excessive chondrogenesis.
To test this hypothesis we investigated the molecular and cellular properties of mononuclear myogenic progenitors isolated from Noggin +/+ and Noggin −/− littermates. Interestingly, the transcript levels of early myogenic markers (Pax3 and Pax7) had the same dynamics in Noggin +/+ and Noggin −/− progenitors. However, protein levels of Myf5 and MyoD were significantly reduced in Noggin −/− progenitors after 3 days of myogenic induction. Indeed, it has been previously reported that no significant differences were detected for Pax3, Pax7 and Myf5 expression in Noggin −/− embryonic muscles 13 . The regulation of Pax7 protein levels is complex and not yet completely defined. In principle we cannot exclude that in our conditions, as recently reported 46 and reviewed in ref. 23, a post-translational modification could induce the ubiquitination and consequently a cleavage of Pax7.
In this view a complete inhibition of the MyoD and possibly Myogenin proteins is probably the main cause of the lack in differentiation of early fetal myoblasts in our cultures as it was previously reported 47 .
One of the key findings of this work is that MyoD and consequently embryonic and adult MyHC protein levels are strongly decreased in Noggin −/− progenitors. Recently, the pattern of expression of MyHC isoforms has been fully characterized and defined as a transition form from one cohort of MyHC isoforms (embryonal, neonatal and type I MyHC) to another (fast type II isoforms), occurring during muscle cell differentiation 48 . Thus, eMyHC dramatically decreased in Noggin −/− cells at both time points considered, could not support further expression of the adult MyHC isoform.
Intriguingly, the response of the Noggin −/− muscle progenitors to adipocytogenic conditions dramatically differs from the wild type cells. The absence of Noggin and thus overexposure to BMP signaling during embryonic development induces the muscle progenitor cells to be significantly more responsive to adipogenic activation than the corresponding wild type cells. Indeed, Adiponectin and Perilipin genes normally expressed in adult white adipose tissue 49 , are also highly transcribed in these aberrant progenitor cells.
These results are in line with previous reports showing the participation of BMP4 in adipocyte lineage commitment and the rescue induced with Noggin exposure in vivo 50 . Moreover, it has been recently demonstrated that murine myoblasts (C2C12) display lipid accumulation upon BMP6 stimulation 51 .
Taken together our results stress the importance of balance between Noggin and BMPs for myogenic commitment of early progenitor cells during limb development. The absence of Noggin could induce a concomitant BMP over-signaling, likely responsible for the impaired myogenic commitment of Noggin −/− progenitors. Indeed, BMP2, in cooperation with SMAD1/4 and C/EBPα , can induce the expression of PPARγ 2, key transcription factor for adipocyte differentiation 52 . Our data also indicate the power of cell memory, where the level of BMP exposure is remembered during later differentiation stages, even in vitro, and even when the actual signal is not present anymore.
Further studies are necessary to define the role of other BMP inhibitors 53 , e.g. Chordin or Follistatin, Xnr3 or Cerberus, which could eventually compensate the absence of Noggin and be involved in the aberrant differentiation ability observed in Noggin −/− muscle progenitors 54 . It could be also interesting to investigate if the alteration of Noggin/BMP balance is involved in the adipogenic conversion of skeletal muscle precursors observed in chronic muscle degenerations.

Materials and Methods
Embryo processing. All the mouse experiments were carried out in accordance with European guidelines on animal research and approved by the ethics committee at the University of Leuven. Noggin heterozygous mice were maintained in CD1 background as described before 4 . Mice were sacrificed and embryos were collected at 16.5, 17.5 and 18.5 days post coitus (dpc). Forelimbs were isolated using forceps and subsequently embedded in Tissue-Tek (Laborimpex), snap frozen in liquid nitrogen and stored at − 80 °C. Hind limbs were processed for beta-galactosidase (beta-gal) genotyping. Frozen sections (5 μ m) for immunological and histological staining (Haematoxylin and eosin; standard protocol) were made using a cryostat (Prosan). Beta-galactosidase genotyping. Targeted Noggin inactivation was achieved by replacing Noggin exon with LacZ expression cassette 3 permitting the genotyping by beta-gal staining. The well-described phenotypic differences were used to discriminate between heterozygous and homozygous animals ( 4,55 ). Hind limbs were fixed in 4% PFA in PBS for 10 minutes and washed twice with LacZ1 solution (PBS, 2mM MgCl 2 , 5 mM EGTA) for 10 minutes at RT, followed by another washing step with LacZ2 (PBS, 0.1 M Na 3 PO 4 , 2 mM MgCl 2 , 0.01% sodium deoxycholate, 0.02% NP-40; pH 7.2-7.4) for 5 minutes at RT. Staining was performed by applying staining solution (5 mM K 3 Fe(CN) 6 , 5 mM K 4 Fe(CN) 6 , 0.5 mg/ml, X-gal in LacZ2 buffer for 2-3 hours at 37 °C or ON at RT. After washing steps with PBS and clearing with 2% KOH, limbs were treated with an increasing gradient of glycerol/KOH before storage.
Immunohistochemistry. Frozen sections were fixed with methanol for 10 minutes at − 20 °C. After quenching in 3% H202/MQ and washing in TBST, antigen retrieval was performed with sodium citrate buffer (pH6), followed by additional washing with TBST. Sections were blocked with Normal Goat Serum (Millipore) (1/5 in TBST) and O/N incubated with the primary antibody (P-SMAD1/5/8, Cell Signaling, 9511S; 1/100) at 4 °C. After washing, the biotinylated secondary antibody was applied for 30 minutes at RT. The signal was amplified with the ABC reagent (Vectastain, dilution according to manufacturer's protocol) for 30 minutes at RT. Phalloidin staining was performed using manufacturers protocol (Sigma). After washing, the color was developed using DAB (DAKO, dilution according to manufacturer's protocol) and sections were mounted using Pertex. The P-SMAD1/5/8 protocol was used with following modifications. Frozen sections were fixed with ethanol and methanol for 5 minutes and quenched in 0.5% H202/MeOH. Washing steps were performed with PBS, a KI67 antibody (DAKO TEC3; 1/100) or Pax7 antibody (Hybridoma bank, Iowa, USA; 1/400) was used and blocking was done with 10%NGS/PBS. ImageJ software was used to measure muscle fiber thickness and the quantification of the percentage of P-SMAD1/5/8, KI67 or Pax7 + cells. The significance of the difference in expression was analyzed using 2way ANOVA test with Bonferroni comparison and significance was achieved with p-values less than 0.05 (Student's T-test).

Gene expression analysis.
For muscle tissue, RNA was isolated by scraping the muscles of at least 10 frozen sections 16.5 dpc embryonic limbs using the High Pure RNA Isolation Kit (Roche Applied Science). At least three RNA isolations were performed per condition and subsequently reverse transcribed using Primescript RT Reagent (Takara; manufacturers protocol). qRT-PCR was performed in duplicate with gene specific primers ( Table 2) with SYBR Premix Ex Taq II (Takara; manufacturers protocol) using Rotor-gene 6000 detection system (Corbett Research, Westburg). Gene expression was normalized to the housekeeping gene Hprt1 and presented as a ratio to control embryos. The significance of the difference in expression was analyzed using the Student's T-test.
For the cells, Noggin +/+ and Noggin −/− early fetal myoblasts, chondrocytes and fibroblasts were used for RNA extraction with RNA Mini Kit, removing genomic DNA traces by Turbo DNase. 1ug RNA was reverse-transcribed with SuperScript III kit and qPCR was performed in 384 well plates (10ul final volume; thermal profile, 95 °C 15″ − 60 °C 45″ (40x); ViiA 7 qPCR plate reader), using Platinum Sybr Green Mix, 1ul 1:4 diluted cDNA and 250 nM primers (all kits, reagents and plate reader by Life Technologies, Carlsbad, CA, USA). Proliferating MEF cells as well as fibroblasts from adult mice or 3T3-L1 preadipocytes (ATCC, Manassas, VA, USA) were used as positive controls.
Fetal myoblasts isolation and myogenic differentiation. Early fetal myoblasts were isolated from forelimbs of E16.5 Noggin −/− and control embryos (after careful removal of bones and skin) as previously described 40 . Briefly, tissues were dissociated one hour at 37 °C with 0.06% Collagenase (from Clostridium Histolyticum, Sigma-Aldrich St. Louis, Mo, USA, for all reagents excepted when notified) and 0.04% Pancreatin in PBS (Gibco-Invitrogen, Carlsbad, CA, USA). After centrifugation, cells were seeded in collagen-coated dishes in presence of Dulbecco's modified Eagle's medium (DMEM)-high glucose and 20% fetal bovine serum (FBS; Gibco-Life Technologies, Carlsbad, CA, USA), supplemented with 1% Chicken Embryo Extract (Sera Laboratories International, UK), 100 mg/ml Sodium Pyruvate, 100 IU/ml penicillin and streptomycin, 2 mM L-glutamine. The plates were incubated at 37 °C, 5% CO 2 , 5% O 2 and the medium refreshed every two days. When cells reached confluence, the differentiation was induced by shifting medium to DMEM high glucose supplemented with 2% horse serum (Gibco-Invitrogen, Carlsbad, CA, USA) and 100 mg/ml Sodium Pyruvate (differentiation medium; DM).

Adipogenic differentiation.
Early fetal myoblasts were dispended on collagen coated dishes and expanded till confluence, in presence of STEMPRO Adipogenesis Differentiation Medium (Gibco-Lyfe Technologies, Carlsbad, CA, USA) to enhance adipogenesis, as described from the manufacturer. 3T3-L1 preadipocytes (ATCC) used as control for adipocyte differetiation were grown in DMEM supplemented with 10% (v/v) fetal calf serum (Gibco-Lyfe Technologies, Carlsbad, CA, USA), 25 mM HEPES (pH 7.0), 100 U/ml penicillin and streptomycin, and 25 μ g/ml Fungizone Antimycotic (Gibco-Lyfe Technologies, Carlsbad, CA, USA). To induce adipocyte differentiation, confluent cells were cultured in DMEM with 10% (v/v) FBS (Gibco-Lyfe Technologies, Carlsbad, CA, USA), 25 mM HEPES (pH 7.0), 100 U/ml penicillin and streptomycin, and 25 μ g/ml fungizone, plus 0.5 mM isobutylmethyl-xanthine, 1 μ M dexamethasone, and 1 μ g/ml insulin (MDI). MDI was replaced 48 h after, with insulin (1 μ g/ml), and the medium changed every two days. . Secondary Alexa Fluor donkey antibodies (Life Technologies, Carlsbad, CA, USA) were diluted at 4 μ g/ml in PBS supplemented with 2.5% w/v BSA and the nuclei counterstained with 10 ug/ml Hoechst. Fluorescence images were acquired at 10x magnification via Eclipse Ti inverted microscope (Nikon). Fusion index (number of myotube nuclei vs total number of nuclei) such as the percentage of myotube positive for a specific marker were calculated for a minimum of four random fields from Noggin +/+ and Noggin −/− embryo cells of different embryos (n = 7, three different litters).
Oil-Red-O staining. Lipid accumulation in early fetal myoblasts was assessed by Oil-Red-O staining.
Briefly, cells were fixed for 15 min in 4% PFA, washed three times in PBS, and stained for 30 min with the lipophilic dye Oil-Red-O in 65% isopropanol, followed by washing steps and drying. Bright field Oil-Red-O images were captured at 10x and 20x magnification using an Eclipse Ti inverted microscope (Nikon). Lipids

Statistical analysis.
All results were expressed as mean ± SD, with the exception of gene expression (mean ± SEM). Representative WB show independent samples. When two groups were compared, a Student's t-test was used. When three or more groups were compared a two-way ANOVA was used. The statistical significance of the differences between percentage values was assessed using a Kruskal-Wallis one-way analysis of variance by ranks test. All statistical tests were performed via Prism software (GraphPad).