Transforming Growth Factor-β Signaling Inhibits the Osteogenic Differentiation of Mesenchymal Stem Cells via Activation of Wnt/β-Catenin Pathway
Article information
Abstract
Background
Due to the contradictory and temporally variable effects of transforming growth factor-β (TGF-β) and the Wnt/β-catenin pathways on osteogenic differentiation in different stem cell types, we sought to examine the activity of these pathways as well as their interaction during the osteogenic differentiation of the osteo-induced adipose-derived mesenchymal stem cells (AD-MSCs).
Methods
The osteo-induced AD-MSCs were treated with TGF-β1 (1 ng/mL) either alone or together with its antagonist SB-431542 (10 μM) or that of the Wnt antagonist, inhibitor of Wnt production 2 (IWP2) (3 μM), every 3 days for 21 days. Cells were then analyzed for calcium deposit, bone matrix production, and the osteogenic markers gene expression.
Results
Our results showed firstly that, either of the pathways is active since the mRNA expressions of their respective target genes, PAI-1 and Cyclin D1 were detectable although the latter was at a very low level. Secondly that, treatment with TGF-β1 decreased levels of calcium deposit, bone matrix production and the osteogenic markers gene expression. Accordingly, osteogenesis was induced in those treated with SB either alone or together with the TGF-β1, pointing to inhibitory effect of TGF-β pathway on osteogenic differentiation. Thirdly that following treatment with IWP2 and TGF-β1, the inhibitory effect of TGF-β1 on bone matrix production was reversed. Fourthly, there was constantly low expression of Wnt3amRNA but progressively increasing that of its endogenous antagonist Dkk-1mRNA throughout.
Conclusions
Together these results suggest that TGF-β1 requires the active Wnt/β-catenin signaling pathway to exert its inhibitory effects on osteogenic differentiation of AD-MSCs.
INTRODUCTION
Bone-related defects, diseases, and their repair costs are among major public health concerns. Mesenchymal stem cells (MSCs) are an attractive cell source for tissue regeneration as they have the potential of self-renewal and multilineage differentiation as well as non or hypo-immunogenic characteristics.[1] Adipose-derived MSCs (AD-MSCs) are promising sources of MSCs that can differentiate into mesodermal and non-mesodermal lineages. They are easily accessible in large quantities with minimal invasive isolation procedures, ideally applicable for stem cell-based therapies and tissue engineering.[2] Due to their therapeutic promises, it is important to understand the mechanism underlying the osteogenic differentiation of the stem cells and the signaling pathways regulating their differentiation. When MSCs start to differentiate into osteoblasts, osteogenic markers such as alkaline phosphatase (ALP) is secreted by early osteoblasts to induce matrix production. Later during mineralization phase osteocalcin (OCN) is secreted by late osteoblasts. Another osteogenic marker is plasminogen activator inhibitor-1 (PAI-1) whose expression has been reported essential for the differentiation of MSCs into osteocytes.[3] Therefore, in the present study, we considered the evaluation of these markers during the osteogenic differentiation. Transforming growth factor-β1 (TGF-β1) and Wnt signaling pathways have been widely involved in multiple biological and pathological processes during mammalian bone formation,[4] stem cell maintenance, and tissue homeostasis.[5] TGF-β1 ligands bind to their heteromeric complexes of type I and type II receptors. The phosphorylated type I receptor phosphorylate R-Smads which form a complex with co-Smad and Smad4 and then translocate into the nucleus to direct transcriptional response. Having known that SMAD7 and PAI-1 are specific gene targets of TGF-β1,[6] their expressions were reported to be induced by TGF-β1 during osteogenic differentiation of bone marrow cells.[7] Regarding the Wnt proteins, they bind to their frizzled receptors and lipoprotein related protein-5/6 co-receptors, which leads to inactivation of glycogen synthase kinase-3β (GSK-3β). This results in the stabilization of β-catenin, followed by its translocation into the nucleus, which activates the Wnt target gene.
The role of TGF-β signaling in osteogenic differentiation has been controversial, that is, its inhibitory role has been highlighted by some researchers in many types of stem cells from different sources,[8,9] whereas its inducing effects reported by others.[10–12] These controversies have also been reported during development, possibly stage-dependent, e.g., Roelen and Dijke (2003) [13] showed that chondroblastic and osteoblastic differentiation was promoted by TGF-β at early stages, but inhibited later. The controversies have also been evidenced for the role of canonical Wnt signaling pathway during in vivo bone formation in development and adults as well as in vitro during osteogenic differentiation of mesenchymal stem cells.[14] The interaction between the two pathways during osteogenic differentiation is also controversial, that is, TGF-β1 signaling pathway modulates Wnt signaling pathway, either through its induction or inhibition.[15–17] Using the osteo-induced AD-MSCs as a model, we first examined the expression of the Wnt and its endogenous antagonist, Dkk-1mRNA. Further, we treated the osteo-induced AD-MSCs with TGF-β1, SB, and inhibitor of Wnt production 2 (IWP2), the antagonists of the TGF-β1 and Wnt pathways. We analyzed the expression of TGF-β1 and Wnt pathways target genes as well as those of the osteogenesis markers, OCN and ALP. Our results showed firstly that both pathways are active and functional and that Wnt/β-catenin signaling is suppressed during osteogenic differentiation of AD-MSCs. Secondly, TGF-β1 requires the Wnt pathway to be active to exert its inhibitory effects on osteogenic differentiation of AD-MSCs.
METHODS
1. Isolation and treatment of mesenchymal stem cells
To isolate the AD-MSCs, adipose tissues were obtained from healthy young people, ranging 30 to 40 years of age with different genders, who had requested liposuctioning. Following the volunteers’ consent, fat tissues were obtained from different areas including the abdomen fat. Being washed three times, the tissue was digested by adding 0.1% collagenase A type I (10 mg/mL in phosphate-buffered saline [PBS]) for one hr at 37°C. Following neutralization of collagenase and its removal by addition of fetal bovine serum (FBS) (Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA), the suspension was centrifuged at 15,000 rpm. The pellet was washed with NH4Cl to remove the blood cells, centrifuged at 1,500 rpm for 5 min, and the pellet was washed again with the media and transferred to T75 flasks containing Dulbecco’s modified Eagle’s medium (DMEM) high glucose (Gibco; Thermo Fisher Scientific, Inc.), FBS 10%, Pen/Strep and fungizone, incubated at 37°C in a humidified chamber with 5% CO2. For the osteogenic differentiation of the AD-MSCs, cells were cultured in the osteogenic induction medium, which contained low glucose DMEM (Invitrogen, Carlsbad, CA, USA), 10% FBS (SPL Life science, Pocheon, Korea), 10 mM β-glycerol phosphate (Sigma-Aldrich, St. Louis, MO, USA), 50 μM ascorbic acid 2-phosphate (Sigma-Aldrich) and 10−7 M dexamethasone (Sigma-Aldrich) for 21 days. To investigate the effects of TGF-β1 on the osteogenic induction, AD-MSCs were treated with TGF-β1 (1 ng/mL; PeproTech Inc., Rocky Hill, NJ, USA) either alone or together with its antagonist SB-431542 (10 μM; Tocris Bioscience, Bristol, UK) every 3 days during the period of 21 days. To inhibit the canonical Wnt signaling pathway, IWP2 (3 μM; Tocris Bioscience) was used every 3 days for 21 days.
2. Characterization of hAD-MSCs
1) Flow cytometry
Kögler and colleagues (2004) [18] showed that mesenchymal stem cells express CD90, CD105, CD73, CD44, CD19, and CD133 Cell surface antigen markers. In this study, we also characterized the AD-MSCs, by using flow cytometry according to the following steps: Firstly, cell culture media was removed and the cells were washed with PBS, trypsinized, thoroughly washed with PBS and centrifuged at 3,000 rpm. The pellet was fixed with 4% paraformaldehyde for 15 min, washed twice with PBS and then blocked with 150 μL of 3% BSA at room temperature for 1 hr. Following incubations with different CDs antibodies and their isotype controls for 2 hr (1/1,000 diluted in 3% BSA; Sigma-Aldrich) at room temperature, the pellet was washed with PBS (once for 5 min), re-suspended in 1 mL PBS and analyzed by flow cytometer (BD Biosciences, Franklin Lakes, NJ, USA).
2) Adipogenic and chondrogenic differentiation potentials of the AD-MSCs
Mesenchymal characteristics of these cells were also analyzed through their differentiation potentials toward chondro-, adipo- and osteogenic lineages.[19]
3. Analysis of bone matrix production
1) Calcium content assay
Total calcium contents of the cells were measured using a Calcium Content Assay kit (Pars Azmoon, Tehran, Iran) and alizarin red staining on days 14 and 21, respectively. Briefly, 20,000 cells were plated in each well of a 96-well plate followed by addition of 0.6 N hydrochloride acid to each well. At the end of the treatment periods, the cells were homogenized for 1 min, centrifuged for 10 min at 10,000 g, and the supernatant was used to measure the calcium content. Adding 10 μL of the kit reagent to each well of a 96-well plate containing 90 μL of supernatant and incubating for 5 min at 37°C, the calcium quantity was measured at 630 nm using the enzyme-linked immunoassay (ELISA) reader.
2) Alizarin red staining
Evaluation of bone matrix production was performed as follows: Cell culture media was removed, cells were washed with PBS, fixed for 10 min with methanol (Merck & CO, Inc., West Port, PA, USA) at room temperature and stained for 10 min with alizarin red (2%; Sigma-Aldrich) which is able to bind the precipitated calcium in the osteoblasts and generate a red color. The excess dye was removed by further washes in PBS, followed by 30 min incubation in 10% acetic acid (Merck & CO, Inc.). Cell supernatant was collected and heated 10 min in 85°C water bath, immediately placed on ice for 5 min, centrifuged (Sigma-Aldrich) for 15 min at 20,000 g, diluted with 10% ammonium hydroxide (Merck & CO, Inc.) 1:3 to neutralize the acidic pH. The optical density (OD) was read at 405 nm by the ELISA reader.
4. Gene expression analysis
Total RNAs were extracted from the control and treated cells using easy-BLUETM total RNA extraction solution (iN-tRON Biotechnology, Seoul, Korea) according to the manufacturer’s protocol. RNAs were then reverse transcribed to the complementary DNAs using the reverse transcriptase enzyme (Yekta Tajhiz Azma, Tehran, Iran) according to manufacturer’s protocol. Finally, real-time polymerase chain reaction (PCR) was performed using SYBR green master mix and sequence-specific primers (Yekta Tajhiz Azma) in Corbett Q PCR machine. The reaction was done at 95°C for 3 min as a holding step, followed by 40 amplification cycles of 5 sec at 95°C (denaturation), 20 sec at 60°C (annealing), and 10 sec at 72°C (extension). All reactions were done in duplicates using β2-microglobulinmRNA as a housekeeping gene. Fold changes of gene expressions were quantified using the comparative delta δCT method through the freely available relative expression software tool 2009. Primer sequences are shown in Table 1.
Primers used for real-time polymerase chain reaction
5. Statistical analyses
All experiments were performed in three replicates and analyzed by SPSS version 16 (SPSS Inc., Chicago, IL, USA). Results were expressed as mean±standard error of the mean. A value of P value less than 0.05 was considered statistically significant.
RESULTS
1. AD-MSCs express cell surface antigens of the MSCs
The isolated AD-MSCs were characterized by using flow cytometry. Our results showed that these cells were positive for mesenchymal CD markers such as CD73, CD90, and CD105, but negative for those of the hematopoietic stem cell markers, CD34 and CD45 (Fig. 1).
Flow cytometric analysis of the isolated adipose-derived mesenchymal stem cells for CD markers. They were positive for mesenchymal CD markers such as CD73, CD90, and CD105, but negative for those of the hematopoietic stem cell markers, CD34 and CD45. Blue lines show the labelled cells whereas the red lines show the unstained cells. Percentages of the unstained cells are shown on the top left hand side and those of the labelled cells on the top right hand side.
2. Active TGF-β1 signaling in the osteogenic AD-MSCs
To determine whether TGF-β1 signaling is active in ADMSCs, we examined if TGF-β1 activates downstream signaling components such as its target gene plasminogen activator inhibitor type PAI-1mRNA. As shown in Figure 2, the expression of PAI-1mRNA was significantly increased in the osteo-induced AD-MSCs treated with the recombinant protein TGF-β1 for 21 days compared to that of the control. Accordingly, treatment with the TGF-β1 antagonist (SB431542) either alone or combined with the TGF-β1 strongly inhibited the PAI-1mRNA expression in the osteo-induced AD-MSCs for 21 days. Altogether, these results showed that TGF-β1 is active and functional in the osteo-induced AD-MSCs.
Real-time polymerase chain reaction analysis of expression of transforming growth factor-β1 (TGF-β1) target gene (plasminogen activator inhibitor-1 [PAI-1]) in the osteo-induced adipose-derived mesenchymal stem cells treated with the recombinant protein TGF-β1 alone or together with the TGF-β1 antagonist (SB431542). While TGF-β1 treatment significantly increased the expression of PAI-1 compared to that of the control, that in combination with SB431542 strongly inhibited the expression. a)P<0.001.
3. Active TGF-β1 signaling inhibits the osteogenic differentiation of AD-MSCs
To evaluate the role of TGF-β1 signaling on osteogenic differentiation of AD-MSCs, the osteo-induced AD-MSCs were treated with the TGF-β1 recombinant protein, SB431542, or both followed by the analysis of the calcium content on day 14 and alizarin red staining on day 21. Quantitation of the calcium content level showed a significant decrease after the TGF-β1 treatment, but an increase following the treatment with SB either alone or together with TGF-β1 (Fig. 3A). Consistently, the result of alizarin red staining showed that the amount of cellular calcium precipitate was significantly lower following the TGF-β1 treatment compared to those after the SB431542 treatment alone or combined with the TGF-β1 (Fig. 3B, C), suggesting that inhibition of osteogenesis and bone matrix production by TGF-β1 was reversed following antagonizing TGF-β1.
Analysis of the osteogenic differentiation of adipose-derived mesenchymal stem cells, using calcium content assay on day 14, alizarin red staining on day 21 and real-time polymerase chain reaction for the osteogenic markers on day 3. Cells were treated with the osteogenic media (OS; control), transforming growth factor-β1 (TGF-β1) recombinant protein, SB431542 or both. The results of calcium content level (A) and alizarin red staining (B, C) showed that cellular calcium precipitant was decreased significantly by TGF-β1 treatment, but increased significantly following the treatment with SB alone or together with TGF-β1 (arrows). (D, E) The analysis of the mRNA expressions of two specific osteogenic gene markers, the alkaline phosphatase (D) and osteocalcin (E) showed that they were significantly decreased following the TGF-β1 treatment, but increased after SB431542 treatment. a)P<0.001.
The inhibitory effects of TGF-β1 were further examined by the analysis of the expressions of two specific gene markers, the ALPmRNA and OCNmRNA after day 21. The results of real-time PCR analysis showed that while the expressions of both genes were significantly decreased by TGF-β1 treatment, they were significantly enhanced after SB431542 treatment (Fig. 3D, E), further confirming the inhibitory action of TGF-β1 on the osteogenic differentiation of ADMSCs.
4. Inhibition of the osteogenic differentiation of ADSCs by TGF-β1 requires the active Wnt/β-catenin
Evidence on the interaction between TGF-β1 and the canonical Wnt/β-catenin signaling pathways prompted us to examine the existence/nature of this interaction during osteogenic differentiation. Co treatment of the osteo-induced ADSCs with the Wnt/β-catenin signaling inhibitor, IWP2 (small molecule antagonizing the Wnt/β-catenin pathway), and TGF-β1 for three days was followed by analysis of themRNA expression of Wnt/β-catenin target genes, c-Myc and Cyclin D1, using real-time PCR. Our results showed that treatment with TGF-β1 alone upregulated the expressions of two Wnt target genes, c-MycmRNA and Cyclin DmRNA, significantly compared to those of the untreated control, whereas treatment with TGF-β1 and IWP2 significantly reduced their expressions returning almost to those of the control levels (Fig. 4C, D). Following the combined TGF-β1 and IWP2 treatment there was no reduced bone matrix production but instead remarkably increased level compared to the reduced level after treatment with the TGF-β1 alone (Fig. 4A, B), indicating that the inhibitory effect of TGF-β1 on the osteogenic differentiation is attenuated if the Wnt/β-catenin pathway is antagonized.
Analysis of bone matrix production by alizarin red staining (A, B) and mRNA expression of Wnt/β-catenin target genes, c-Myc and Cyclin D (C, D) after treatment with transforming growth factor-β1 (TGF-β1) alone or combined with the Wnt inhibitor (IWP2). (A, B) Bone matrix production was remarkably increased in cells cultured in the presence of both TGF-β1 and IWP2 compared to cells that cultured in the presence of only TGF-β1. (C, D) Real-time polymerase chain reaction analysis of the expression of Wnt/β-catenin target genes, c-Myc (C) and Cyclin D1 (D) showed that treatment with TGF-β1 alone upregulated the expressions of the two Wnt target genes significantly compared to those of the untreated control, whereas that with both TGF-β1 and IWP2, reduced their expressions significantly returning them to almost the control levels. a)P<0.001.
5. Constantly low expression of Wnt3a and its target gene Cyclin D1 but progressively increasing expression of Dkk-1 during the osteogenic differentiation of AD-MSCs
From the above results, it can be implied that Wnt/β-catenin pathway must be inhibited during osteogenic differentiation of AD-MSCs. To address this question, the status of WntmRNA expression and its signaling activity were examined during the osteogenic differentiation of these cells. Our results showed that there were constantly low levels of Wnt3amRNA expression (Wnt/β-catenin pathway agonist; 0.2 fold) and Cyclin D1mRNA (the Wnt target gene; 0.2–0.5 fold) on days 7, 14 and 21 (Fig. 5). The expression of Dkk-1mRNA (Wnt/β-catenin pathway antagonist) however was progressively increasing (2.5 folds increase on day 14 and almost 3 folds on day 21; Fig. 5). These results indicate that Wnt/β-catenin signaling is being suppressed throughout the osteogenic differentiation of AD-MSCs.
Analysis of the mRNA expressions of Wnt endogenous antagonist (Dkk-1), Wnt3a and its target gene (Cyclin D1) by real-time polymerase chain reaction. There were constantly low levels of Wnt3a expression (0.2 fold) and Cyclin D1 (0.2–0.5 fold) on days 7, 14, and 21, whereas that of Dkk-1mRNA was progressively increasing, 2.5 folds on day 14 and almost 3 folds on day 21. a)P<0.01, b)P<0.001.
DISCUSSION
TGF-β1 and the Wnt signaling pathways have widely been recognized to have roles in maintenance, expansion, early differentiation and commitment of the mesenchymal stem/progenitor cells towards different lineages including the osteoblastic lineage. Their exact roles however are controversial and not well understood. Using AD-MSCs as a standard model, we sought to partially unravel the mechanism by which TGF-β1 and Wnt signaling pathways modulate the osteogenic differentiation of these cells.
Despite the contradiction on the role of TGF-β1 in osteogenesis, evidence from in vitro studies favors its inhibitory role. Using TGF-β1 in a physiological range, we first showed that TGF-β1 was able to activate its target gene, PAI-1, indicating that this signaling pathway was active and functional in the osteo-induced AD-MSCs. Our analysis of the TGF-β1 function alone showed that the recombinant TGF-β1 protein strongly attenuated the calcium content level in 14 days and bone matrix production as well as the osteogenic specific gene expression in 21 days of culture. These were reversed when TGF-β1 was antagonized by using SB431542, that is, upregulation of calcium content and osteogenic markers expression, suggesting that the inhibition of TGF-β1 specifically increased the osteogenic differentiation of the osteo-induced AD-MSCs. These results are consistent with some previously published articles using human unrestricted somatic stem cells,[20] rat bone marrow mesenchymal stem cells [21] and human bone marrow mesenchymal stem cells.[21,22] Contrastingly, the stimulatory effects of TGF-β1 signaling on osteogenic differentiation have also been shown in different types of stem cells such as pluripotent mesenchymal precursor cell line C2C12,[23] murine bone marrow stromal cells [24] and bone mesenchymal stem cells.[10,11] These controversies could either be related to different applied concentrations, sources and types of stem cell or to the interactions with other signaling pathways and downstream effectors. As an example, Li and colleagues (2012) [25] reported that TGF-β1 at low concentration induced the alkaline phosphatase activity synergistically in BMP9-transduced C3H10T1/2 cells, whereas at higher concentrations inhibited the BMP9-induced osteogenic activity.
The interaction between the TGF-β1 and Wnt signaling has been reported on specification of cell fate during development [26] as well as in the differentiation of adult stem cells and cell lines.[27,28] They reported that the cross-talk between TGF-β1 and Wnt signaling pathways occurs through transcription factors TCF/LEF1 and SMAD3 [27] and that Axin, a negative regulator of the Wnt/β-catenin signaling, can regulate the activity of TGF-β1 signaling.[28] If TGF-β1 has inhibitory on osteogenesis mediated by the active Wnt/β-catenin pathway, then one would expect Wnt/β-catenin pathway to be down-regulated during osteogenic differentiation. Indeed, our further investigation showed that this was the case. Our real-time PCR analysis revealed that the expressions of Wnt3amRNA and Cyclin D1mRNA (Wnt target genes) were at quite low levels throughout the osteogenesis. Contrastingly, that of the Wnt antagonist Dkk-1mRNA was progressively increasing, suggesting that the Wnt pathway is antagonized during osteogenesis.
The inhibitory role of the Wnt pathway in bone differentiation has been highlighted by some researchers. For example, De Boer and colleagues (2004) [29] reported the inhibitory effect of canonical Wnt pathway on bone differentiation in human MSC cells induced by dexamethasone. In cultured USSCs, we have previously shown that treatment with BIO, a GSK3 inhibitor and Wnt activator, for 21 days led to the highest inhibition of osteo matrix production in first 3 days.[30] Accordingly, treatment with Dkk-1, the Wnt antagonist, enhanced the expression of the two osteogenic genes, alkaline phosphatase and osteopontin.[31] There are also contradictory evidence indicating the stimulatory effect of Wnt on osteogenesis for example in bone marrow MSCs [29,32,33] and murine pluripotent stem cells.[34,35] The discrepancies between the Wnt effects on osteogenesis might be due to differences in cellular context (development vs. adult), species, the specific used doses, timelines and the interaction between Wnt and other pathways such as BMP and TGF-β1 reported by many studies.[4,29,36,37]
Altogether, in the context of AD-MSCs, our study has provided the first evidence on the interplay between TGF-β1 and Wnt/β-catenin signaling pathways which favors the inhibition of osteogenesis. This would therefore suggest reconsidering a proper niche for future application of stem cell therapy in bone tissue regeneration.
Notes
Acknowledgments
The authors wish to thank Professor Elahe Elahi and her team at the University of Tehran who supported us using the Corbett real time PCR machine.
Funding
The authors received no financial support for this article.
Ethics approval and consent to participate
All studies were accomplished in accordance with the guidelines approved by the Tehran University of Medical Sciences Research Ethics Committee. Samples were taken from patients following their consent.
Conflict of interest
No potential conflict of interest relevant to this article was reported.