Author contributions: H.T. and L.H.: performed the experiments, analyzed the data, and wrote the manuscript; L.C. and K.H.L.: performed the experiments; A.M.S. and M.K.: supervised, provided funding, and revised the manuscript; B.M.A.: conception, designed and supervised the study, analyzed the data, wrote and revised the manuscript. H.T. and L.H. contributed equally to this article.
Disclosure of potential conflicts of interest is found at the end of this article.
First published online in STEM CELLSEXPRESS November 18, 2011.
Delta like-1 (Dlk1)/preadipocyte factor-1 (Pref-1)/fetal antigen-1 (FA1) is a novel surface marker for embryonic chondroprogenitor cells undergoing lineage progression from proliferation to prehypertrophic stages. However, mechanisms mediating control of its expression during chondrogenesis are not known. Thus, we examined the effect of a number of signaling molecules and their inhibitors on Dlk1 expression during in vitro chondrogenic differentiation in mouse embryonic limb bud mesenchymal micromass cultures and mouse embryonic fibroblast (MEF) pellet cultures. Dlk1/Pref-1 was initially expressed during mesenchymal condensation and chondrocyte proliferation, in parallel with expression of Sox9 and Col2a1, and was downregulated upon the expression of Col10a1 by hypertrophic chondrocytes. Among a number of molecules that affected chondrogenesis, transforming growth factor-β1 (TGF-β1)-induced proliferation of chondroprogenitors was associated with decreased Dlk1 expression. This effect was abolished by TGF-β signaling inhibitor SB431542, suggesting regulation of Dlk1/FA1 by TGF-β1 signaling in chondrogenesis. TGF-β1-induced Smad phosphorylation and chondrogenesis were significantly increased in Dlk1−/− MEF, while they were blocked in Dlk1 overexpressing MEF, in comparison with wild-type MEF. Furthermore, overexpression of Dlk1 or addition of its secreted form FA1 dramatically inhibited TGF-β1-induced Smad reporter activity. In conclusion, our data identified Dlk1/FA1 as a downstream target of TGF-β1 signaling molecule that mediates its function in embryonic chondrogenesis. The crosstalk between TGF-β1 and Dlk1/FA1 was shown to promote early chondrogenesis during the embryonic endochondral ossification process. STEM CELLS 2012; 30:304–313.
Endochondral ossification is a process where mesenchymal cells condense, proliferate, and differentiate into chondrocytes, produce cartilaginous matrix, and undergo maturation and hypertrophy. Chondrocyte hypertrophy is followed by matrix mineralization and eventually substitution of cartilage by bone .
Several signaling pathways were demonstrated to be involved in embryonic endochondral ossification processes at different sequential stages. Transforming growth factor β (TGF-β), Notch, and Wnt signaling pathways are activated during mesenchymal condensation while bone morphogenetic protein (BMP), insulin growth factor (IGF), and fibroblast growth factor (FGF) signaling pathways modulate chondroprogenitor proliferation and differentiation. TGF-β1 is one of the most important growth factors in promoting chondrocyte proliferation via activation of β-catenin signaling and primary chondrogenesis through Smad3 and chromatin remodeling [2, 3]. A negative feedback loop of parathyroid hormone related peptide (PTHrP) and Indian hedgehog (Ihh) controls the proliferation and prehypertrophy of chondrocytes. PTHrP maintains chondrocyte proliferation and prevents premature hypertrophy while Ihh accelerates the differentiation and hypertrophy of chondrocytes [4–6]. Chondrocyte hypertrophy is regulated by BMP, FGF, and Wnt pathways . Moreover, p38/mitogen activated protein kinase (MAPK) has been implicated in regulation of chondrogenesis by inducing aggrecan expression and maintenance . The balance of signaling pathways determinates the rate of differentiation in each step of the cascade by regulating chondroblast proliferation, maturation, and hypertrophy [9, 10].
Delta like-1/preadipocyte factor-1/fetal antigen-1 (Dlk1/Pref-1/FA1) is a transmembrane protein of the Notch/Delta/Serrata family . The extracellular domain of Dlk1 is proteolytically cleaved by ADAMT17 to produce a bioactive FA1 fragment . Dlk1/Pref-1 was originally identified as a negative regulator of adipocyte differentiation [13, 14], and more recent data suggests a role for Dlk1 in many other differentiation processes during embryogenesis and in postnatal organisms, including hematopoiesis , myogenesis , and osteogenesis [17, 18]. Deficiency of Dlk1 expression in both human and mouse was demonstrated to be responsible for a variety of developmental defects, including growth retardation, obesity, and skeletal malformations [19, 20]. Recently, we have identified Dlk1/FA1 as a novel chondrogenic surface marker for chondroprogenitor cells during mouse limb bud development in vivo and human embryonic stem cell differentiation into the chondrocyte lineage in vitro . Our further studies demonstrated an inhibitory role for Dlk1 on insulin-induced chondrogenesis by Akt-dependent mechanism . In addition, Dlk1/FA1 was shown to promote early commitment of mesenchymal stem/stromal cells (MSCs) into the chondrocytic lineage through regulation of Sox9 expression .
Here, we identified the upstream signaling molecules that modulate Dlk1/FA1 expression and function during chondrogenesis using in vitro models of embryonic limb bud micromass and mouse embryonic fibroblast (MEF) pellet cultures derived from wild type (WT), Dlk1−/−, and Dlk1-overexpressing mice. Dlk1 expression was upregulated during the chondrocyte proliferation phase and dramatically abolished upon chondrocyte hypertrophy in both of these in vitro models. Our data identified Dlk1 as a downstream target of TGF-β signaling whose expression is critical for mediating the function of the TGF-β1 pathway in chondrogenesis.
MATERIALS AND METHODS
Cell Isolation and Culturing
Mouse limb bud cells were isolated from E11.5 C57Bl/6 wild-type mice fetuses as previously described . Briefly, limbs were dissected from the embryos under a dissecting microscope and digested with 3 mg/ml Dispase (Invitrogen, Taastrup, Denmark, www.invitrogen.com) for 20 minutes at 37°C. Cells were filtered through a prewashed 40-μm cell strainer (Becton Dickinson, Brøndby, Denmark, www.bd.com) to generate a single-cell suspension. Cells were suspended at a density of 2 × 107 cells per milliliter in Dulbecco's modified Eagle's medium (DMEM)/F12 + 10% fetal bovine serum (FBS; PAA, Pasching, Austria, www.paa.com) and plated in 10 μl of media as a micromass. Cells were incubated for 2 hours at 37°C prior to careful addition of 500 μl medium (DMEM/F12 supplemented with 50 μg/ml vitamin C [Sigma-Aldrich, Brøndby, Denmark, www.sigmaaldrich.com/denmark.html], 10 nm β-glycerophosphate [Sigma-Aldrich], and 10% FBS). Media was changed every 2nd day.
MEFs were isolated from E13.5 WT, Dlk1−/−, and Dlk1 transgenic mouse embryos [25, 26]. Fetuses were eviscerated and beheaded before being finely minced and digested with 0.25% trypsin (Invitrogen) for 5 minutes at 37 °C. Cells were plated in D-MEM (Invitrogen) supplemented with 10% FBS and 1% penicillin/streptomycin (P/S; Invitrogen), and cultured until confluence. Cells were routinely trypsinized for 5 minutes at 37 °C and replated at a density of 6,000 cells per centimeter square. Passage 2–3 cells were used in experiments.
MEF cells were induced to the chondrogenic lineage using a pellet micromass culture method, as previously described for MSC chondrogenesis [27–29]. Briefly, 200,000 cells were centrifuged in a 15 ml polypropylene tube at 500g for 5 minutes. Pellets were cultured in 500 μl chondrogenic medium (D-MEM supplemented with 10 ng/ml of TGF-β3 [R&D Systems, Abingdon, U.K.], 100 nM dexamethasone [Sigma-Aldrich], 50 μg/ml of vitamin C, 40 μg/ml of L-proline [Sigma-Aldrich], 100 μg/ml of sodium pyruvate [Sigma-Aldrich], and 50 mg/ml of ITS + Premix [BD Biosciences, Brøndby, Denmark]). MEF cultures were supplemented with 100 ng/ml BMP-2 (Peprotech, London, UK, www.peprotech.com). Pellets were cultured for up to 21 days with medium change every 3–4 days.
Pellets were fixed in 4% paraformaldehyde for 15 minutes at room temperature. Samples were embedded in paraffin and sectioned at 5 μm. Proteoglycans were visualized by Toluidine Blue and Alcian Blue stainings. Sox9, type II collagen, and Dlk1/FA1 depositions were detected by immunohistochemistry. Rehydrated sections were digested with bovine testicular hyaluronidase (2,000 U/ml) in phosphate-buffered saline (PBS; pH 5.5) for 60 minutes at 37°C to facilitate antibody access. Nonspecific antibody binding was blocked by 3% bovine serum albumin (Sigma-Aldrich) in PBS. Sections were incubated with a monoclonal antibody against chicken type II collagen (1:400, 6B3 Chemicon Millipore, Copenhagen, Denmark, www.millipore.com), rabbit polyclonal antibody against rat FA1 (21), or Sox9 (1:200, R&D) for 1 hour at room temperature. After incubating with “Ready-to-use” Post Blocking (Dako, Glostrup, Denmark, www.dako.com) for 20 minutes, sections were incubated with Poly–horseradish peroxidase anti-mouse/rabbit IgG for 30 minutes. 3,3′-Diaminobenzidine tetrahydrochloride (Dako) chromogen was used for detection of the primary antibody. Sections were counterstained with hematoxylin. In control sections, the primary antibody was replaced with normal serum.
Signaling Pathway Analysis
MEFs were plated in six-well plates at a density of 10,000 cells per centimeter square in D-MEM + 10% FBS, and 1% P/S. To study the effect of TGF-β signaling, medium was supplemented with one of the activators: TGF-β1 (1–100 ng/ml; R&D Systems, Abingdon, UK, www.rndsystems.com), TGF-β3 (1–100 ng/ml; R&D), Activin A (1–100 ng/ml; R&D), Activin B (1–100 ng/ml; R&D), BMP-2 (10–100 ng/ml; R&D) or specific inhibitors: Follistatin  (5–500 ng/ml; R&D), Noggin  (5–500 ng/ml; R&D), SB431542  (10 nM to 10 μM; Sigma-Aldrich), SB505124  (10–1000 nM; Sigma-Aldrich), SD-208  (100 nM to 10 μM; Tocris Bioscience, Bristol, UK, www.tocris.com), and LY 364947 (1 μM; Tocris Bioscience). Cells were harvested for RNA isolation after 1–7 days of culture.
Chondrogenic differentiation of MEF was performed in a pellet culture system as described above. Chondrogenic medium was supplemented with 10 μM SB43154 or 10 ng/ml TGF-β1. Samples were collected for RNA extraction and histological analysis after 5, 10, and 20 days of culture.
Limb bud cells were cultured as described above. Standard culture medium was supplemented with TGF-β1 (20 ng/ml), SB431542 (10 μM), BMP-2 (100 ng/ml), gamma secretase inhibitor IX N-(N-(3,5-difluorophenacetyl-L-alanyl))-S-phenylglycine t-butyl ester (DAPT)  (100 nM; Calbiochem, www.merck-chemicals.com), or the MAPK inhibitor PD169316  (2 μM; Calbiochem, www.merck-chemicals.com). Samples were collected for RNA extraction at days 5, 10, and 15.
Transfection and Sma and Mad–Related Protein (SMAD) Luciferase Assay
Transfections and luciferase assay were performed using Fugene 6 (Roche Diagnostics, Mannheim, Germany, www.roche.com), Cignal SMAD reporter Assay Kit (QIAGEN Nordic, Copenhagen, Denmark) and the Dual-Luciferase assay (Promega Biotech AB, Nacka, Sweden), respectively, following the protocols recommended by the manufacturers. Briefly, 293-H cells were transfected with SMAD reporter system. After 24 hours of the transfection, cells were either transfected with pCD2-Dlk1 plasmid (gift from Dr. J. Battey, NIH, Bethesda, MD) or treated with purified FA1 (1 μg/ml) in the presence of 10 ng/ml of recombinant human TGF-β1 protein (R&D) and assayed for luciferase activity after 18 hours.
Western Blot Assays
Antibodies specific for Smad2 (total or phosphorylated) were obtained from Cell Signaling Technology. α-Tubulin was purchased from Santa Cruz Biotechnology (Aarhus, Denmark, www.scbt.com). For Western blot analysis, cells were collected at specific time points post-treatment and washed in cold PBS buffer before being lysed in cell lysis buffer (10 mM Tris-HCl, pH 7.4, 150 mM sodium chloride, 1% NP-40, 0.1% sodium dodecyl sulfate, 1 mM EDTA, 1 mM phenyl-methylsulfonyl fluoride, 1 mM NaF, 1 mM Na3VO4), supplemented with protease inhibitor cocktail (Roche Diagnostics). A total of 20 μg of protein was separated on 8%–12% NuPAGE Novex Bis-Tris gel systems (Invitrogen) followed by transfer to polyvinylidene fluoride membrane (Millipore, USA). The membrane was blocked and probed with antibodies as indicated and incubated with peroxidase-conjugated secondary antibody (Santa Cruz Biotechnology). Proteins were visualized by enhanced chemiluminescence (Thermo Scientific, Denmark, www.thermoscientific.com).
RNA Extraction and Real-Time Polymerase Chain Reaction (PCR) Analysis
RNA was extracted using a single-step method of TRIzol (Invitrogen). RNA concentration and purity were measured using a Nanodrop Spectrophotometer and cDNA was synthesized from 1 μg of total RNA using revertAid H minus first strand cDNA synthesis kit (Fermentas, St. Leon-Rot, Germany, www.fermentas.com) according to the manufacturer's instructions. Quantitative real-time PCR was performed with a 7500 Real-Time system using Fast SYBR Green Master Mix (Applied Biosystems, CA, www.appliedbiosystems.com) with primers listed in Table 1. After normalization to β-actin mRNA, relative expression levels and fold induction of each target gene were calculated using a comparative CT method [(1/2) formula, where ΔCT is the difference between CT-target and CT-reference] with Microsoft Excel 2007.
Table 1. Primer sequences
Comparisons between groups were analyzed using analysis of variance and two-tailed unpaired Student's t test.
Expression of Dlk1 During Chondrogenic Differentiation of Limb Bud Cells
The micromass culture system is an established model to study embryonic limb development and endochondral ossification in vitro. In this culture system, the E11.5 limb bud-derived mesenchymal cells have the ability to differentiate spontaneously in vitro into mature chondrocytes and to recapitulate the endochondral ossification program . As shown in Figure 1A, based on cell morphology and gene expression analysis, the differentiation program in the micromass culture was divided into four phases: (a) mesenchymal cell condensation (days 1–3; characterized by expression of Sox9 and Col2a1), (b) chondrocyte proliferation (days 4–8; S100, Sox9, Col2a1, Acan, and Comp), (c) chondrocyte prehypertrophy and hypertrophy (days 8–13; Col10a1, PTHrP, and Ihh), and (d) matrix mineralization (day 10 onward; Alpl and Bglap) (Fig. 1A, 1B).
Gene expression analysis by real-time PCR revealed that during chondrogenesis, Dlk1 was expressed with a pattern similar to early chondrogenic markers Sox9 and Col2a1 (Fig. 1B). The expression increased until day 6 and thereafter rapidly decreased at day 7 to be completely abolished upon upregulation of the late chondrogenic genes PTHrP, Ihh, and Col10a1 (Fig. 1B). Similarly, Dlk1/FA1 protein was observed in proliferating chondrocytes but was abolished upon hypertrophy at day 10 as assessed by immunohistochemistry (Fig. 1C). In addition, enzyme-linked immunosorbent assay (ELISA) measurement of a soluble active form of Dlk1 (FA1) in the conditioned medium revealed increased production of FA1 during the chondrocyte proliferation phase, followed by reduced amount of FA1 upon chondrocyte hypertrophy, thus resembling the pattern of gene expression (Fig. 1D).
Expression of Dlk1 During Chondrogenic Differentiation of MEFs
To examine the expression of Dlk1 in another in vitro model of chondrogenesis, we used MEFs, which can be induced by TGF-β and BMP-2 to differentiate into the chondrogenic lineage in three-dimensional (3D) pellet cultures . Real-time PCR analysis and immunohistochemistry revealed that MEFs differentiated into cells with chondrocyte morphology that expressed cartilage-specific genes Sox9, Col2a1, and Acan and secreted proteoglycans and type II collagen into the matrix (Fig. 2). Dlk1 was highly expressed during the early phase of chondrogenic differentiation, reaching the highest expression at day 5 and decreased gradually to be undetectable at day 20 (Fig. 2A). Chondrocyte-like cells were observed at day 10 and by day 15 pellets consisted uniformly of type II collagen expressing chondrocytes (Fig. 2B).
Regulation of Dlk1 Expression by Different Chondrogenic Signaling Pathways
To study upstream regulators of Dlk1 expression and function during chondrogenesis, we examined the effect of stimulating or inhibiting signaling pathways that are known to affect chondrogenesis on the expression of Dlk1 in limb bud micromass culture.
As shown in Supporting Information Figure S1A, treatment of limb bud cultures with BMP-2 (100 ng/ml) markedly promoted chondrocyte maturation and hypertrophy, as evidenced by increased gene expression of Col10a1, Ihh, and Bglap while slightly upregulating Dlk1 expression at day 3 (Supporting Information Fig. S1A). Treatment with PD169316, a specific inhibitor for p38 MAPK pathway, inhibited chondrocyte differentiation and hypertrophy by decreasing the expression of Sox9, Col2a1, and Ihh, without affecting Dlk1 expression during the differentiation time course (Supporting Information Fig. S1B). Suppression of Notch signaling by the gamma secretase inhibitor DAPT markedly enhanced markers of chondrocyte maturation, terminal differentiation, and matrix mineralization as assessed by upregulation of Ihh, Col10a1, and Bglap, without affecting the expression of Dlk1 and early chondrogenic markers (Supporting Information Fig. S1C).
Dlk1 Is a Downstream Target of TGF-β Signaling Pathway
Among different examined members of TGF-β superfamily on limb bud cultures (including TGF-β3, Activin A, Activin B, BMP-2, or BMP-4, data not shown), we identified TGF-β1 as a possible regulator of Dlk1 expression. Treatment of limb bud cultures with TGF-β1 resulted in maintaining the differentiated chondrocytes in the proliferative phase, while inhibiting their maturation and matrix mineralization, as shown by downregulation of Col10a1, Ihh, and Bglap (Fig. 3A). This effect of TGF-β1 was associated with marked downregulation of Dlk1 expression without affecting other early chondrogenic markers Sox9 and Col2a1. Conversely, blocking TGF-β1 signaling by SB431542 (a selective inhibitor of TGF-β type I receptors [also known as activin receptor-like kinases; ALKs] ALK4, ALK5, and ALK7 ) dramatically inhibited chondrogenic differentiation of limb bud cells in association with upregulation of Dlk1 by sixfold over control nontreated cells and downregulation of all chondrogenic markers Col2a1, Sox9, Col10a1, and Ihh (Fig. 3B). No changes in Dlk1 expression were detected when BMP or activin signaling pathways were blocked with Noggin or follistatin, respectively (data not shown). These data suggested Dlk1/FA1 as a mediator for TGF-β1 function during chondrogenesis.
We further examined the effect of TGF-β1 signaling on Dlk1 expression in MEF in monolayer cultures and during chondrogenic differentiation in pellet micromass cultures. As shown in Figure 4A and 4B, Dlk1 was significantly downregulated in response to TGF-β1 and markedly upregulated by SB431542 during the whole time course of treatment of MEFs in monolayer culture as well as during chondrogenesis in 3D culture. Furthermore, treatment of MEF cells for 3 days with other specific small-molecule inhibitors of ALKs, SB505124 (ALK4, ALK5, ALK7 ), LY 364947 (ALK-5 ), and SD-208 (ALK-5 ) significantly upregulated the expression of Dlk1 (Fig. 4C).
Dlk1 Functions Downstream TGF-β1 Signaling to Mediate Its Effect on Chondrogenesis
Our results suggest that Dlk1/FA1 is acting downstream of TGF-β1 signaling to mediate, at least in part, its effects on chondrogenesis. In this model, Dlk1/FA1 functions as a negative regulator of chondrogenesis whose expression is downregulated in response to TGF-β1 signaling to allow chondrocyte differentiation. To examine this hypothesis, we studied the effect of absence and overexpression of Dlk1 on TGF-β1-induced chondrogenesis using MEF cultures established from embryos of Dlk1 knockout (Dlk1−/− MEF) and Dlk1-overexpressing mice (Dlk1-Tg-MEF). WT, Dlk1−/− MEF, and Dlk1-Tg-MEF were induced to differentiate into chondrocytes in micromass pellets in the presence of TGF-β1. Interestingly, chondrocyte differentiation was significantly enhanced in Dlk1−/− MEF in response to TGF-β1 compared to both WT-MEF and Tg-MEF as assessed by increased pellet size, more intense staining for proteoglycans by Toluidine Blue and Alcian Blue and for type II collagen, and significant upregulation of early and late chondrogenic markers (Col2a1, Acan, and Col10a1) after 20 days of culture (Fig. 5A, 5B). Overexpression of Dlk1 was shown to inhibit MEF differentiation into chondrocytes in response to TGF-β1 as revealed by markedly downregulated Col2a1, Acan, and Col10a1 mRNA expression (Fig. 5A, 5B).
To obtain insight into the mechanism by which Dlk1/FA1 mediates the function of TGF-β1 in chondrogenesis, we studied the effect of Dlk1 overexpression and knockout on TGF-β1-induced phosphorylation of Smad2 in MEF cells. As shown in Figure 6A, treatment of Dlk1−/− MEF with TGF-β1 resulted in marked stimulation of Smad2 phosphorylation compared with WT-MEF, while TGF-β1-induced phosphorylation of Smad2 was reduced in Tg-MEF cells. Furthermore, overexpression of Dlk1 or addition of its soluble form FA1 dramatically inhibited TGF-β1-induced SMAD reporter luciferase activity in 293-H cells by 50% and 75%, respectively (Fig. 6B).
We have recently identified Dlkl/FA1 as a novel surface marker for chondroprogenitor cells during embryonic chondrogenesis and demonstrated the Akt pathway-dependent inhibition of insulin induced in vitro chondrogenesis by Dlk1/FA1 [21, 22]. However, the mechanisms by which Dlk1 is regulated and its interaction with other signaling pathways during chondrogenic differentiation are not well known. In this study, we have identified Dlk1 as a novel downstream target of the TGF-β signaling pathway that mediates its function by promoting early chondrogenesis. Additionally, we observed altered Smad2 phosphorylation with Dlk1/FA1 overexpression suggesting that Dlk1/FA1 inhibits activation of TGF-β signaling pathway.
We used mouse embryonic limb bud culture as an in vitro system that spontaneously differentiates in culture without any exogenous chondrogenic inducers and recapitulates the sequential stages of chondrogenesis from condensing mesenchyme to matrix mineralization [24, 39, 40]. The expression profile in the limb bud model corroborated our previous findings where Dlk1 demonstrated a restricted spatiotemporal expression profile during in vivo chondrogenesis . In the limb bud system, Dlk1 was expressed during mesenchymal condensation and chondrogenic differentiation while the expression was abolished upon chondrocyte hypertrophy. In another chondrogenic model, we obtained a similar expression pattern of Dlk1/FA1 during the in vitro induction of MEF micromass pellets into chondrocytes by TGF-β3 and BMP-2. Thus, our data identified Dlk1/FA1 as a specific early chondrogenic marker with an expression pattern similar to Col2a1 and Sox9 during embryonic chondrogenesis.
The in vitro expression pattern of Dlk1 is consistent with the recently published role of Dlk1 in chondrogenesis, which acts by promoting chondrocyte proliferation and inhibiting chondrocyte maturation to hypertrophy. This was demonstrated by increased hypertrophic zone in the growth plate and upregulated expression of the prehypertrophic and hypertrophic chondrocyte markers Ihh and Col10a1 in Dlk1−/− embryos . In this regard, the function of Dlk1 in maintaining progenitor cell proliferation and inhibiting their differentiation has been established in other differentiation models including adipogenesis, osteogenesis, and hematopoiesis [14, 17, 18, 41].
Several downstream pathways have been implicated in mediating the function of Dlk1, including IGF-1, MAPK, and the extracellular signal–regulated kinase mitogen-activated protein kinase (MEK/ERK) pathway [42–45]. Dlk1 was recently shown to interact with fibronectin and initiate the MEK/ERK pathway through activation of integrin downstream signaling and thereby inhibit adipogenesis . Furthermore, inhibition of adipogenesis, osteogenesis, and chondrocyte maturation by Dlk1 was suggested to be mediated by Sox9 transcription factor . Dlk1 was shown to upregulate Sox9 which in turn inhibits Runx2 expression and thereby osteoblast differentiation. Similarly, Sox9 was shown to suppress the activity of CCAAT-enhancer-binding protein β (C/EBPβ) and C/EBPδ, resulting in the inhibition of adipogenesis. Our recent data showed an inhibitory role for Dlk1 on phosphatidylinositol 3-kinases/Akt signaling pathway activity and subsequent suppression of chondrogenesis, while no effect was observed in ERK1/2 or p38 MAPK pathways . It is thus likely that the mechanisms by which Dlk1 functions in differentiation are cell type dependent.
Signaling pathways upstream of Dlk1 have only been studied in relation to adipocyte differentiation. In this context, the expression of Dlk1 was shown to be negatively regulated by several factors known to mediate adipogenesis including glucocorticoid, dexamethasone, growth hormone, cytokine, IL-11 and BMP-7 [41, 47–49]. Our detailed analysis on the possible regulation of Dlk1 in response to different chondrogenesis-related signaling molecules identified TGF-β1 as a specific inhibitor of Dlk1 expression during chondrogenic differentiation of limb bud and MEF cells. Downregulation of Dlk1 by TGF-β1 treatment was associated with delayed chondrocyte hypertrophy and inhibition of matrix mineralization by maintenance of the chondrocytic phenotype by expression of the early chondrogenic markers Sox9 and Col2a1. In contrast to the inhibition of Dlk1 expression by TGF-β1 to maintain chondrocytes in the premature stage, blocking of TGF-β signaling with selective inhibitors of ALK-2, ALK-5, and ALK-7 receptors strongly inhibited mesenchyme condensation and chondrogenesis in parallel with fourfold stimulation of Dlk1 expression.
TGF-β1 is known for its role in promoting early chondrogenic differentiation in several in vitro models including limb bud cells [50, 51], chondroblasts [52, 53], ATDC5 mesenchymal cells , and MSC [55, 28]. Smad activation has been shown to mediate TGF-β signaling in chondrogenesis. TGF-β1 promotes chondrocyte proliferation through the activation of β-catenin signaling and primary chondrogenesis through Smad3 signaling and chromatin remodeling [2, 3]. Smad3 stimulates transcriptional activity of Sox9 in a TGF-β1-dependent manner and, together with Smad2, Smad4, and Sox9, forms a transcriptional complex in the enhancer region of Col2a1 to activate its expression [3, 56]. Conversely, the Smad2/3 pathway inhibits chondrocyte maturation and hypertrophy in vitro and in vivo [40, 57–59]. Deficiency in the functional TGF-β1 pathway by disruption of Smad3 or the type II receptor leads to premature chondrocyte hypertrophy and degenerative joint diseases resembling osteoarthritis in mice [60, 61].
Since both Dlk1 and TGF-β are known to play a role in chondrogenesis, we investigated their potential interaction in chondrogenesis. In this regard, our data supports the notion that Dlk1 mediates the effect of TGF-β1 on chondrogenesis based on the following findings: (a) TGF-β1 alone was able to significantly enhance chondrocyte differentiation in MEF lacking Dlk1 while this effect was abolished in MEF overexpressing Dlk1; (b) activation of Smad phosphorylation by TGF-β1 was increased in MEF lacking Dlk1, while it was blocked in MEF overexpressing Dlk1; (c) overexpression of Dlk1, or addition of its secreted form FA1, inhibited TGF-β1-stimulated SMAD reporter activity.
Taken together, our data showed that the downregulation of Dlk1 by TGF-β1 is critical for mediating the stimulatory effect of TGF-β1 on chondrocyte differentiation. Thus, the crosstalk between TGF-β1 signaling pathway and Dlk1/FA1 exists for mediating early embryonic chondrogenesis.
We are grateful to Lone Christiansen, Bianca Jørgensen, Tina Nielsen, and Tuula Oivanen for excellent technical assistance. This study was supported by grants from NovoNordisk Foundation in Denmark and Finnish Cultural Foundation, Academy of Finland (projects 203446 and 205581) in Finland. H.T. was supported by Turku Graduate School of Biomedical Sciences, Turku University Foundation, and Finnish Cultural Foundation.
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.