Activation of Hedgehog Signaling Inhibits Osteoblast Differentiation of Human Mesenchymal Stem Cells§

Authors

  • Magali Plaisant,

    1. Institute of Signaling, Biology, Development and Cancer, Université de Nice Sophia-Antipolis, CNRS UMR6543, France
    Search for more papers by this author
  • Coralie Fontaine,

    1. Institute of Signaling, Biology, Development and Cancer, Université de Nice Sophia-Antipolis, CNRS UMR6543, France
    Search for more papers by this author
  • Wendy Cousin,

    1. Institute of Signaling, Biology, Development and Cancer, Université de Nice Sophia-Antipolis, CNRS UMR6543, France
    Search for more papers by this author
  • Nathalie Rochet,

    1. GEPITOS, Université de Nice Sophia-Antipolis, CNRS, UFR de Médecine, France
    Search for more papers by this author
  • Christian Dani,

    1. Institute of Signaling, Biology, Development and Cancer, Université de Nice Sophia-Antipolis, CNRS UMR6543, France
    Search for more papers by this author
  • Pascal Peraldi

    Corresponding author
    1. Institute of Signaling, Biology, Development and Cancer, Université de Nice Sophia-Antipolis, CNRS UMR6543, France
    • Centre de Biochimie (UMR 6,543 CNRS), Université de Nice-Sophia Antipolis, Faculte des Sciences, 28 Avenue Valrose, 06108 Nice Cedex 2, France
    Search for more papers by this author
    • Telephone: 33-493377704; Fax: 33-4-92076404


  • Disclosure of potential conflicts of interest is found at the end of this article.

  • Author contributions: M.P.: collection and assembly of data, data analysis and interpretation, manuscript writing; C.F.: collection of data, data analysis and interpretation; W.C.: collection of data; N.R.: data analysis and interpretation, provision of study material; C.D.: conception and design, data analysis and interpretation, financial support; P.P.: conception and design, collection and assembly of data, data analysis and interpretation, manuscript writing.

  • §

    First published online in STEM CELLSExpress December 18, 2008.

Abstract

Mesenchymal stem cells within the bone are responsible for the generation of osteoblasts, chondrocytes, and adipocytes. In rodents, Indian hedgehog has been shown to play a role in osteoblast differentiation. However, evidence for a direct function of hedgehog (Hh) in human osteoblastic differentiation is missing. Using different models of human mesenchymal stem cells we show that Hh signaling decreases during osteoblast differentiation. This is associated with a decrease in Smoothened expression, a key partner that triggers Hh signaling, and in the number of cells displaying a primary cilium, an organelle necessary for Hh signaling. Remarkably, treatment of human mesenchymal stem cells with sonic hedgehog or two molecules able to activate Hh signaling inhibits osteoblast differentiation. This inhibition is visualized through a decrease in mineralization and in the expression of osteoblastic genes. In particular, activation of Hh signaling induces a decrease in Runx2 expression, a key transcriptional factor controlling the early stage of osteoblast differentiation. Consistently, the activation of Hh signaling during the first days of differentiation is sufficient to inhibit osteoblast differentiation, whereas differentiated osteoblasts are not affected by Hh signaling. In summary, we show here, using various inducers of Hh signaling and mesenchymal stem cells of two different origins, that Hh signaling inhibits human osteoblast differentiation, in sharp contrast to what has been described in rodent cells. This species difference should be taken into account for screening for pro-osteogenic molecules. STEM CELLS2009;27:703–713

INTRODUCTION

Mesenchymal stem cells reside in a variety of tissues and can differentiate into several cell types such as osteoblasts, adipocytes, myocytes, and chondrocytes. Within the bone, these cells give rise to osteoblasts, responsible for the mineralization of the bone, to chondrocytes that produce and maintain the cartilaginous matrix, and to adipocytes, which are found in abundance in the bone marrow. The fate of these cells is tightly controlled by various growth factors, hormones, and morphogens, and it has been proposed that a perturbation of the balance could be responsible for osteoporosis [1, 2]. Among these factors, hedgehog (Hh) plays an important role in the formation of the skeleton. During skeletogenesis, Indian hedgehog (Ihh) and sonic hedgehog (Shh) are involved in patterning the axial, appendicular, and facial skeleton. During endochondral ossification, when bone is formed after the generation of a cartilaginous scaffold, Ihh is expressed by prehypertrophic chondrocytes and coordinates growth and differentiation of chondrocytes [3]. In rodents, Ihh is also involved in osteoblast differentiation, although its precise function, to induce or to sustain the differentiation process, remains debated [4, 5]. Nevertheless, evidence for a direct function of Hh signaling in osteoblast differentiation has been provided by in vitro experiments using established rodent cell lines such as C3H10T1/2, KS483, and MC3T3-E1 [6–10].

Because of the increase in the prevalence of osteoporosis, screening for osteogenic molecules that might be used as therapeutic agents for bone disease has been performed in rodent cell lines. This has led to the identification of purmorphamine, a 2,6,9-trisubstituted purine, as an activator of osteoblast differentiation in C3H10T1/2 cells [11]. Subsequently, purmorphamine was identified as an activator of Hh signaling through direct binding to Smoothened (Smo) [12]. Smo is a key element in Hh signaling. Schematically, Hh signaling is initiated by the binding of Hh to its receptor, Patched (Ptc). Upon binding, Ptc relieves its suppression on Smo. Smo is then localized into the primary cilium of the cell, an organelle playing a critical role in Hh signaling [13, 14]. There, Smo activates an intracellular cascade that results in the stabilization of Gli2. This transcription factor translocates into the nucleus and induces the transcription of Hh target genes, such as Gli1, a reliable marker of Hh signaling activity [15–17].

Although a direct effect of Hh in rodent osteoblast differentiation has been documented, its function in human cells has not been investigated. In humans, mutation in the Hh pathway, such as in brachydactyly type A1 (mutation in Ihh) [18] and Gorlin syndrome (mutation of Ptc associated with higher Hh signaling, also known as basal cell nevus syndrome), are associated with skeletal malformations [19]. However, the effect of homozygous mutation of Ihh is thought to be associated primarily with a deregulation of chondrocyte homeostasis [20]. So far, no particular defect has been observed in bone ossification, and evidence for a direct effect of Hh on osteoblast differentiation in humans is still missing. Moreover, the effect of the Hh signaling pathway on human mesenchymal stem cell differentiation has never been investigated, probably because of the scarcity of appropriate cell models.

Recently, we isolated human mesenchymal stem cell lines derived from adipose tissue. These cells are called human multipotent adipose-derived stem (hMADS) cells or adipose stem cells according to the international nomenclature [21]. hMADS cells exhibit a normal karyotype, multipotency at the clonal level, and high self-renewal [22, 23]. They differentiate with high efficiency into osteoblasts [24] and thus represent a useful tool to investigate the function of signaling pathways during osteoblast differentiation.

Here, using both hMADS cells and mesenchymal stem cells from bone marrow, we show that osteoblast differentiation of human mesenchymal stem cells is associated with a decrease in Hh signaling. Moreover, in contrast to what has been observed in rodent cell lines, activation of Hh signaling inhibits human osteoblast differentiation.

MATERIALS AND METHODS

Cell Culture and Osteoblast Differentiation

Characterization of hMADS cells and their ability to differentiate into osteoblasts has been extensively described [22–24]. All experiments were performed on hMADS-2 and confirmed on hMADS-1 and hMADS-3 cells. Human mesenchymal stromal cells isolated from bone marrow (hBMSCs) were purchased from Cambrex (Emerainville, France, http://www.cambrex.com). Cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (FCS), 2.5 ng/ml human fibroblast growth factor (hFGF)-2 (for hMADS cells only), 60 μg/ml penicillin, and 50 μg/ml streptomycin. hFGF-2 was removed at confluence. Two days after confluence (day 0), cells were incubated in an osteogenic medium: α-modified Eagle's medium (α-MEM) containing 1% FCS supplemented with 200 μM L-ascorbic acid phosphate, 10 mM β-glycerophosphate, 100 nM dexamethasone, and 10 ng/ml epidermal growth factor. Media were changed every other day and cells were used at the indicated days.

Human primary osteoblasts (hOSTs) were purchased from Cambrex. Cells were maintained in α-MEM (Cambrex) supplemented with 10% FCS, 60 μg/ml penicillin, and 50 μg/ml streptomycin. When indicated, the culture medium was supplemented with 200 μM ascorbic acid phosphate, 10 mM β-glycerophosphate, and 100 nM dexamethasone (mineralization medium).

Purmorphamine (Calbiochem, San Diego, http://www.emdbiosciences.com) was dissolved in dimethylsulfoxide and used at 2 μM. 20(S)-hydroxycholesterol and 22(S)-hydroxycholesterol (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) were dissolved in methanol and chloroform, respectively, and oxysterol cocktail was used at 10 μM (1:1).

Shh-conditioned medium was obtained from an HEK 293 cell line stably transfected with Shh-N expression and neomycin resistance constructs (obtained from American Type Culture Collection, #CRL-2782). The Shh-N-producing HEK 293 cells were grown to 80% confluency in DMEM containing 10% (vol/vol) FBS and 400 μg/ml G418. The medium was replaced with DMEM containing 1% FBS, and after 2 days of secretion, the medium was collected and filtered through a 0.22-μm membrane. Control medium was obtained from HEK 293 cells.

RNA Extraction and Analysis

Total RNA was extracted with the TRI-Reagent kit (Euromedex, Souffelweyersheim, France, http://www.euromedex.com) according to the manufacturer's instructions. Total RNA was subjected to real-time quantitative reverse transcription-polymerase chain reaction (RT-PCR) analysis as described elsewhere [25]. Primers were designed using Primer Express software (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com) and validated by testing PCR efficiency using standard curves (85% ≤ efficiency ≤ 115%). Gene expression was quantified using the comparative CT (threshold cycle) method; TATA-binding protein (TBP) was used as the reference. Stability of TBP expression along osteoblastic differentiation was verified by comparing (a) the CT of TBP and (2) the expression of TBP using hypoxanthine-guanine phosphoribosyltransferase as a reference in undifferentiated and differentiated cells (supporting information Fig. 5). The list of human primers used is provided in supporting information Table 1 and murine primers are provided in supporting information Table 2.

Determination of Alkaline Phosphatase Activity, Calcium Level, and Alizarin Red Staining for Mineralization

Alkaline phosphatase (ALP) activity was assayed spectrophotometrically as previously described [26]. The calcium level was measured with the QuantiChrom Calcium Assay Kit (BioAssay Systems, Hayward, CA, http://www.bioassaysys.com) according to the manufacturer's instructions. To examine mineralization, cells were fixed with paraformaldehyde (4% in phosphate-buffered saline [PBS]) and stained with Alizarin Red for 20 minutes at room temperature.

Immunostaining

Cells were seeded on coverslips and treated as described. Cells were rinsed twice with PBS and fixed with 4% paraformaldehyde in PBS for 20 minutes at room temperature. Fixed cells were incubated in PBS with 3% bovine serum albumin (BSA), 0.1% Tween-20, and 0.1% Triton X-100 for 30 minutes at room temperature. Cells were then incubated with monoclonal acetylated tubulin antibody (T6793; Sigma) diluted 1:1,000 in PBS with 3% BSA, 0.1% Tween-20 for 2 hours at room temperature, followed by AlexaFluor 594-conjugated donkey anti-mouse IgG (1:600; Invitrogen, Cergy-Pontoise, France, http://www.invitrogen.com) for 1 hour at room temperature. Nuclei were counterstained with Hoechst 33258 (0.5 μg/ml; Invitrogen). Images were taken on a Zeiss microscope (Carl Zeiss MicroImaging, Inc., Thornwood, NY, http://www.zeiss.com).

Ten representative fields were examined for each condition and the total number of cilia was counted along with the total number of nuclei. The percentage of cells with cilia was evaluated in three independent experiments realized in triplicate.

Statistical Analysis

Statistically differences between groups were analyzed by Student's t-test and are indicated on figures as follow: *p ≤ .05, **p ≤ .01, ***p ≤ .001.

RESULTS

Hh Signaling Is Downregulated During Osteoblast Differentiation

hMADS cells represent an attractive model to study osteoblastic differentiation of human mesenchymal stem cells [22–24]. First, we investigated the modulation of Hh signaling during osteoblast differentiation through Gli1 expression levels. Indeed, Gli1 is considered to be an adequate marker of Hh pathway activity [15–17, 27–29]. hMADS cells were treated with an osteogenic medium for 10 days, RNA was extracted, and the expression of genes of interest was monitored by real-time quantitative RT-PCR. Differentiation was assessed by the increase in the expression of two classic osteoblast marker genes: ALP and Runx2 (supporting information Fig. 1A). As illustrated, osteoblast differentiation of hMADS cells was associated with a 96% lower Gli1 expression level (Fig. 1 A). The expression levels of three other genes—those encoding insulin-like growth factor binding protein 5, Dkk2, and insulin-like growth factor I [30, 31], the expression of which has been shown to be downregulated during inhibition of Hh signaling—were also found to dramatically decrease during differentiation (Fig. 1C). We then compared the level of Gli1 in undifferentiated hMADS cells and in mature osteoblasts in primary culture (Fig. 1B). As observed, Gli1 was 94% less expressed in mature osteoblasts than in undifferentiated hMADS. We also analyzed Hh signaling in C3H10T1/2 mouse mesenchymal stem cells. Under osteoblastic conditions Hh signaling decreased, as observed through Gli-1 and Dkk2 expression (see supporting information Fig. 2).

Figure 1.

Hedgehog signaling is downregulated during osteoblast differentiation. hMADS cells were treated in the presence or absence of an osteogenic medium. Shown are results of quantitative RT-PCR of Gli1 (A) and IGF-1, IGFBP-5, and Dkk2 (C) after 10 days of differentiation. (B): Expression of Gli1 in undifferentiated hMADS cells and hOSTs. (D): Quantitative RT-PCR of Ihh, Dhh, Ptc, Gli2, Gli3, and Smo. Data are mean ± standard error of triplicates from an experiment representative of three independent experiments. Abbreviations: Dhh, desert hedgehog; hMADS, human multipotent adipose-derived stem; hOST, human primary osteoblast; IGF, insulin-like growth factor; IGFBP, IGF binding protein; Ihh, Indian hedgehog; Ptc, Patched; RT-PCR, reverse transcription-polymerase chain reaction; Smo, Smoothened; TBP, TATA-binding protein.

Thus, compared with undifferentiated hMADS cells, hMADS cells under differentiation conditions as well as mature osteoblasts have a lower level of Hh signaling activity.

We then analyzed key components of the Hh signaling cascade in hMADS cells during differentiation (Fig. 1D). The expression of the ligands Ihh and desert hedgehog (Dhh) remained unchanged during differentiation, and levels of Shh were too low to allow quantification (not shown). No change in Ptc, Gli2, or Gli3 expression levels was detected, whereas osteoblast differentiation was associated with a 60% decrease in Smo expression.

The primary cilium plays a major role in Hh signaling [13, 14]. We analyzed whether hMADS cells possess such an organelle and if it was modified during osteoblast differentiation. hMADS cells were treated for 5 days with a control or an osteogenic medium. Cells were fixed and labeled with an antibody to acetylated α-tubulin to reveal the ciliary axoneme. As shown in Figure 2, 92% of untreated hMADS cells exhibited a primary cilium, which is similar to what has been observed in other cell lines [32, 33]. After 5 days in an osteogenic medium, the percentage of cells having a primary cilium was reduced to 48%, which is comparable with what has been observed in MC3T3-E1 osteoblasts [34]. Because the rate of cell proliferation influences the incidence of detectable primary cilia [32], we performed a double-labeling of cells with antiacetylated tubulin and Ki67 (a marker of cell proliferation). Fifteen percent Ki67+ cells were detected under differentiation conditions, compared with 5% under control conditions. However, this 10% difference cannot account for the 40% difference observed between undifferentiated and differentiated cells. This is supported by the detection of Ki67 cells, which do not possess any detectable primary cilium (supporting information Fig. 4). Then, we analyzed human osteoblast primary culture (Fig. 2C, 2D). Quantification showed that only 33% of these cells displayed a primary cilium, which is comparable with the percentage determined for hMADS cells under differentiation conditions.

Figure 2.

Immunolabeling of primary cilia in hMADS cells and hOSTs. hMADS cells were incubated in ctrl medium (A) or osteogenic medium (B). (C): hOSTs were placed in ctrl medium. Antiacetylated tubulin antibody (red) labeled the primary cilia (arrows). Nuclei were stained by Hoechst 33258. The insets represent a magnification of the picture. (D): Quantification of cells with primary cilia. Abbreviations: ctrl, control; hMADS, human multipotent adipose-derived stem; hOST, human primary osteoblast.

In conclusion, the decrease in the expression of Smo and in the number of cells exhibiting a primary cilium could be responsible for the decrease in Hh signaling observed during osteoblast differentiation.

Stimulation of Hh Signaling Impairs Osteoblast Differentiation of hMADS Cells

We then analyzed the effect of activation of Hh signaling on osteoblast differentiation. First, we used Shh-conditioned medium obtained from an HEK 293 cell line stably transfected with Shh. Control medium was obtained from HEK 293 cells. First, cells were treated without (none) or with conditioned medium from control (ctrl) or Shh-producing cells (Shh). The activation of Hh signaling was confirmed through the induction of Gli1 expression (Fig. 3A). The effect of Hh was tested after 10 days of osteoblastic treatment through gene expression of osteoblastic markers (Fig. 3B). The control conditioned medium did not affect osteoblast differentiation. Surprisingly, conditioned medium from Shh-secreting cells was found to inhibit osteoblast differentiation, as observed through the expression of ALP, Runx2, osteoprotegerin, and osteonectin.

Figure 3.

Shh-conditioned medium inhibits osteoblast differentiation. Shh-conditioned medium was obtained from an HEK 293 cell line stably transfected with Shh-N expression. Human multipotent adipose-derived stem cells were treated with the osteogenic cocktail alone (none) or supplemented with conditioned medium from control cells (Ctrl) or Shh-producing cells (Shh). Culture media were changed every other day. After 10 days, Gli1 (A) and ALP, Runx2, osteonectin, and osteoprotegerin (B) gene expression was analyzed using real-time quantitative reverse transcription-polymerase chain reaction. Data are mean ± standard error of a triplicate from a representative experiment. Abbreviations: ALP, alkaline phosphatase; Ctrl, control; Shh, sonic hedgehog; TBP, TATA-binding protein.

Conditioned medium contains undefined molecules that could interfere with the effect of Hh on osteoblast differentiation. Moreover, the reproducibility between the different batches can not be controlled over a long period. So, to work under defined conditions we decided to use purmorphamine, a purine derivative specifically activating Hh signaling through direct Smo binding. Purmorphamine has been shown to be specific to Hh signaling using high-density microarray [12, 35]. Moreover, purmorphamine was originally isolated as an activator of osteoblast differentiation in murine cells. We have previously shown that Shh and purmorphamine at 2 μM have a similar ability to induce Gli1 expression in hMADS cells and that this drug is not toxic or mitogenic for hMADS cells [25].

hMADS cells were treated with an osteogenic medium, in the presence or absence of purmorphamine, and RNA was extracted after 5, 10, 15, or 21 days. A decrease in Gli1 expression during osteoblast differentiation was observed after 5 days of treatment (Fig. 4A). However, at each time point, cells were still responsive to purmorphamine. Moreover, at 2 μM, purmorphamine did not produce a massive activation of the Hh pathway, suggesting that its effect cannot be attributed to a supraphysiologic activation of the pathway. Then, we investigated ALP mRNA expression. Inhibition of ALP expression was detected as early as 5 days after treatment and was maintained for at least 21 days (Fig. 4B). Accordingly, ALP activity was reduced after 15 days of treatment (Fig. 4C). This was accompanied by a decrease in mineralization, as shown by Alizarin Red staining (Fig. 4D), and of calcium content (Fig. 4E). A 50%–70% decrease in the expression of the osteoblast markers Runx2, osteonectin, osteopontin, and osteoprotegerin was also observed (Fig. 4F). To rule out that the inhibition of osteoblast differentiation mediated by Hh signaling in human cells was specific to our experimental conditions, we tested the effect of purmorphamine on C3H10T1/2 mouse mesenchymal stem cells. Under the same conditions, and as previously reported [11], purmorphamine was indeed pro-osteoblastic in rodent cells, as observed through the induction of ALP and Runx2 (supporting information Fig. 3). Altogether, the analysis of mineralization, ALP activity, and expression of osteoblast markers indicate that chronic activation of the Hh signaling pathway by purmorphamine impairs osteoblast differentiation of human mesenchymal stem cells.

Figure 4.

Purmorphamine treatment inhibits osteoblast differentiation. hMADS cells were treated with an osteogenic medium in the presence or absence of purmorphamine (2 μM) from day 0 to completion of experiment (AF) or during the indicated times (G, H). (A): Gli1 mRNA expression. (B): ALP mRNA expression. (C): ALP activity. (D): Alizarin Red staining. (E): Measure of calcium by quantitative colorimetric determination. Results are representative of three independent experiments. (F): Gene expression analyzed by quantitative RT-PCR. The data presented are mean ± SE of three independent experiments performed in triplicate. (G, H): hMADS cells were differentiated for 15 days in the absence (−) or presence of purmorphamine from day 0 to day 15 (0-15) or from day 0 to day 7 (0-7). (G): Quantitative RT-PCR of ALP expression. (H): Alizarin Red staining. Data are mean ± SE of triplicates from an experiment representative of three independent experiments. Abbreviations: ALP, alkaline phosphatase; ctrl, control; hMADS, human multipotent adipose-derived stem; RT-PCR, reverse transcription-polymerase chain reaction; SE, standard error; TBP, TATA-binding protein.

The time frame of the Hh effect was analyzed. hMADS cells were treated or not with purmorphamine for 15 days, or only during the first 7 days. As shown previously, purmorphamine chronic treatment decreased the expression of ALP mRNA (Fig. 4G) and mineralization (Fig. 4H). A similar inhibition was observed when cells were treated with purmorphamine only during the first 7 days of differentiation.

These results indicate that, in contrast to its effect in rodent cells, purmorphamine inhibits osteoblast differentiation in human mesenchymal stem cells.

Oxysterols Inhibit hMADS Osteoblast Differentiation

We then investigated whether the inhibition of osteoblast differentiation could be reproduced by other Hh activators. We used oxycholesterols, which activate Hh signaling through a mechanism different from the one induced by purmorphamine and have been shown to activate osteoblast differentiation in rodent cells. Oxysterols were used at 10 μM, a concentration that has been shown to maximally stimulate osteoblast differentiation and Hh signaling in murine cells [36–40]. First, we verified whether oxysterols were able to activate Hh signaling in our cells. hMADS cells were treated or not with an oxysterol cocktail 20(S)-hydroxycholesterol and 22(S)-hydroxycholesterol or with purmorphamine for 48 hours. A comparison of Gli1 expression levels shows that oxysterols and purmorphamine induced a similar activation of Hh signaling (Fig. 5A). hMADS cells were treated with an osteogenic cocktail in the presence or absence of oxysterols or purmorphamine. After 15 days, osteoblastic differentiation was analyzed. Oxysterols induced a decrease in mineralization as shown by Alizarin Red staining (Fig. 5B). Consistent with these observations, the level of the osteoblastic markers ALP, Runx2, osteoprotegerin, osteopontin, and osteonectin was decreased by oxysterols (Fig. 5C). This confirms that activation of Hh signaling inhibits osteoblastic differentiation.

Figure 5.

Oxysterols inhibit osteoblast differentiation. hMADS cells were placed in the presence or absence of osteogenic medium for 15 days and treated with or without a combination of 20(S) and 22(S) oxysterols (5 μM each) or with purmorphamine (2 μM). (A): Expression of Gli1 using quantitative reverse transcription-polymerase chain reaction in undifferentiated cells after 48 hours with purmorphamine or oxysterols. (B): Alizarin Red staining performed after 15 days of differentiation. (C): Osteoblast marker gene expression. Data are mean ± standard error of triplicates from an experiment representative of three independent experiments. Abbreviations: ALP, alkaline phosphatase; ctrl, control; hMADS, human multipotent adipose-derived stem; TBP, TATA-binding protein.

Hh Signaling Inhibits hBMSC Osteoblastic Differentiation

To determine whether the downregulation of the Hh pathway was confined to hMADS cells, we investigated the effect of a stimulation of Hh signaling in primary culture of hBMSCs. hBMSCs were treated with an osteogenic medium in the presence or absence of purmorphamine. Osteoblast differentiation was again associated with a decrease in Gli1 expression and purmorphamine was able to activate the Hh signaling pathway in these cells (Fig. 6A). Purmorphamine addition to the osteogenic medium induced a 50% decrease in ALP activity and mRNA expression (Fig. 6B, 6C). Runx2, osteopontin, osteoprotegerin, and osteonectin expression levels were also found to be significantly decreased (Fig. 6D). These data indicate that activation of the Hh signaling pathway impairs osteoblast differentiation of hBMSCs.

Figure 6.

Hedgehog signaling inhibits osteoblast differentiation of hBMSCs. hBMSCs were placed in the presence or absence of osteogenic medium for 15 days. Purmorphamine (2 μM) was added or not from day 0 to day 15. (A, B, D): Marker gene expression by quantitative reverse transcription-polymerase chain reaction. (C): ALP activity. Results are representative of three independent experiments (mean ± standard error of triplicates). Abbreviations: ALP, alkaline phosphatase; ctrl, control; hBMSC, human bone marrow mesenchymal stromal cell; TBP, TATA-binding protein.

Hh Signaling Does Not Affect Mature Osteoblasts

We then investigated whether purmorphamine could affect differentiated human osteoblasts. To this end, we used mature osteoblast primary culture. Osteoblasts were treated for 7 days in the presence or absence purmorphamine. The increase in Gli1 expression after purmorphamine treatment shown in Figure 7A indicates that Hh signaling was still functional in these cells. Purmorphamine did not significantly affect mineralization, ALP activity, or calcium content (Fig. 7B–7D). Consistently, the expression levels of ALP, Runx2, osteonectin, osteopontin, and osteoprotegerin were not affected (Fig. 7E). These results indicate that Hh signaling does not affect mature osteoblasts.

Figure 7.

Purmorphamine does not affect human primary osteoblasts. Human primary osteoblasts were placed in mineralization medium in the absence or presence of purmorphamine for 7 days. (A, C, E): Gene expression analyzed by quantitative reverse transcription-polymerase chain reaction. (B): Alizarin Red staining. (C): ALP activity. (D): Calcium content. Data are mean ± standard error of triplicates from an experiment representative of three independent experiments. Abbreviations: ALP, alkaline phosphatase; ctrl, control; TBP, TATA-binding protein.

DISCUSSION

In rodents and humans the involvement of Hh signaling in skeletal development is well established [3, 41]. In both species, Ihh participates in endochondral ossification through regulation of chondrocyte homeostasis. Ihh also has a direct effect on osteoblast differentiation in rodents. However, so far the function of Hh signaling in human osteoblast differentiation has not been established. We have taken advantage of a human mesenchymal stem cell line developed in our laboratory to address this question in detail.

First, we investigated Hh signaling during osteoblast differentiation. In rodents, the modulation of Hh signaling during osteoblast differentiation is controversial. In vitro it increases during the differentiation of KS483 cells [9], whereas in vivo results [4] and our experiment in C3H10T1/2 cells suggest that Hh signaling decreases during osteoblast differentiation. Undifferentiated mesenchymal stem cells are endowed with a basal level of Hh signaling that is probably sustained by an autocrine effect of Ihh and Dhh, which are expressed in these cells. Then, Hh signaling decreases during osteoblast differentiation. Accordingly, a low level of Hh signaling was observed in mature osteoblasts. The decrease in Smo expression during osteoblast differentiation could be, in part, responsible for the decrease in Hh signaling. Indeed, experiments using small hairpin RNA have shown that downregulation of Smo expression leads to a decrease in Hh activity [42]. Osteoblast differentiation of hMADS cells was also associated with a decrease in the number of cells displaying a primary cilium. Consequently, primary cultures of human osteoblasts have a low percentage of cilium-positive cells. Primary cilia are small antenna-like organelles found on most vertebrate cells and emerge from the older of the two centrioles after mitosis. They are involved in various biological processes such as cell division, signal processing, and cellular movement, and can act as mechanosensors [13, 14]. The cilium has been shown to be necessary for Hh signaling, so the disappearance of the primary cilium during osteoblast differentiation could be responsible for Hh signaling downregulation. This decrease in detectable primary cilium could be linked to an absence of primary cilium or a dramatic decrease in the size of the cilium, which would make it more difficult to detect. However, in both cases, it is likely that the signaling pathways controlled by this organelle would be affected. In addition to Smo, several other receptors, such as somatostatin receptor 3, 5-HT6 serotonin receptor, and platelet-derived growth factor receptor α, are also located at the primary cilium [14]. Regulation of the signaling pathway through modulation of the primary cilium structure could be an original mechanism of signal transduction. Unraveling the mechanisms leading to Smo downregulation and loss of the primary cilium could provide us with important information about the regulation of signaling pathways in mesenchymal stem cells.

We show that activation of Hh signaling in human mesenchymal stem cells inhibits osteoblast differentiation. Activation of Hh signaling induced a 50% decrease in expression of Runx2. This phenomenon is probably the main event responsible for the antiosteogenic effect of Hh. Indeed, Runx2 is a key transcription factor that directs multipotent mesenchymal cells to an osteoblastic lineage and inhibits them from differentiating into adipocytes and chondrocytes. Runx2 is considered as the earlier master gene that triggers osteoblast differentiation [43]. Thus, it is likely that the decrease in Runx2 expression observed here led to the decrease in “later” genes such as those encoding osteoprotegerin, osteopontin, osteonectin, and ALP. Ultimately this leads to a decrease in mineralization, one of the more distal events of osteoblast differentiation.

The inhibition of osteoblast differentiation of human mesenchymal stem cells by activation of Hh signaling observed here is in sharp contrast to what has been reported in rodent cells. Indeed, experiments performed in established rodent mesenchymal cell lines showed that Hh increases the expression of osteoblastic markers and mineralization, probably through Runx2 expression [6–10]. The time frame of Hh action in vitro and in vivo is a matter of debate. Although it has been proposed that, in vitro, only immature multipotent cells are sensitive to the osteoblastic effect of Hh [8], others report otherwise [7]. There is also some controversy over in vivo effects because Rodda and McMahon showed that Ihh is a transient requirement for osteoblast differentiation [5], whereas Mak et al. [4] proposed that this signaling pathway also regulates mature osteoblasts. In human mesenchymal stem cells, we showed that Hh affects the early stage of osteoblast differentiation. Moreover, mature osteoblasts were insensitive to purmorphamine. This is consistent with an effect of Hh signaling on Runx2 expression, which is involved early in osteoblast commitment.

We provide here several lines of evidence that argue that the antiosteogenic effect of Hh signaling is not confined to one type of molecule or one cell line. Indeed, in addition to Shh, two molecules able to activate Hh signaling, purmorphamine and oxysterols, inhibit osteoblast differentiation. Interestingly, both were first identified as potent osteogenic molecules in rodent cells and recognized as activators of Hh signaling afterward [11, 12, 36–40]. Moreover, they activate Hh signaling through different mechanisms: purmorphamine binds directly to Smo [35], whereas oxysterols do not bind to Smo [39] but trigger its translocation to the cilia [44]. Finally, as illustrated in Figures 3 and 4, and as previously described [25], these molecules do not produce a massive overactivation of the Hh pathway. The antiosteoblastic effect of Hh signaling is not specific to a single cell line. This was also observed in two other hMADS cell lines, isolated from distinct donors (data not shown), and was reproduced in hBMSCs. Finally, we have verified that, under the same experimental conditions, purmorphamine stimulates osteoblast differentiation of rodent mesenchymal stem cells. Together, this indicates that the distinct Hh effect observed in mesenchymal murine versus human cells is an interspecific difference. Together with our previous observation that Hh signaling differentially affects adipocyte differentiation of human and rodent mesenchymal stem cells [25], this points to a profound difference in the function of Hh signaling in rodent and human mesenchymal stem cells.

What could be the involvement of Hh signaling in osteoblast homeostasis of human endochondral bone? In rodents, overexpression of Ihh in chondrocytes promotes osteoblast differentiation [45]. On the other hand, analysis of the mouse phenotype with an activation of the Hh pathway through Ptc1 haploinsufficiency has given conflicting results. Ptc−/+ animals were reported to have greater bone mass [46]. In contrast, deletion of one allele of Ptc, specifically in mature osteoblasts, gives rise to osteopenic mice, caused by a stimulation of osteoclasts [4]. In humans, brachydactyly type A1 and Gorlin syndrome are caused by an alteration in Hh signaling [18–20, 47]. These syndromes are associated with skeletal malformations that are thought to be the result of an imbalance in chondrocyte homeostasis. However, so far there is no evidence that osteoblast differentiation is directly affected. Finally, although in the adult mouse growth plate chondrocytes are the major sources of Ihh, the growth plate disappears by 20 years of age in humans, raising questions about the involvement of this morphogene in adult bone.

Together, the effect of Hh activators as therapeutic drugs for osteopenia and osteoporosis in humans should be seriously evaluated. If in humans, the inhibitory effect of Hh activation on osteoblast differentiation observed here is recapitulated in vivo, as well as the activation of osteoclasts described in rodents [4], Hh signaling stimulating drugs could have disastrous effects on bone structure.

CONCLUSION

Our results show that Hh signaling decreases during osteoblast differentiation of human mesenchymal stem cells. This is associated with two original phenomena: a decrease in Smo expression and a loss of the primary cilium. Activation of Hh signaling inhibits osteoblast differentiation, probably by targeting Runx2 expression, an early master gene of osteoblast differentiation. This work reveals a striking contrast in the control of osteoblast differentiation in human and rodents and suggests caution in the use of Hh signaling as a target for pro-osteogenic molecules.

Acknowledgements

We are grateful to Drs. Sophie Giorgetti-Peraldi, Annie Ladoux, Ez-Zoubir Amri, Didier Pisani, and Pr. G. Ailhaud for helpful comments. This work was supported by grants from the Centre National de la Recherche Scientifique and the Fondation pour la Recherche Medicale, Equipe FRM. M.P. and C.F. were supported by the Fondation pour la Recherche Medicale.

DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST

The authors indicate no potential conflicts of interest.

Ancillary