T-Box20 inhibits osteogenic differentiation in adipose-derived human mesenchymal stem cells: the role of T-Box20 on osteogenesis

Background Skeletal development and its cellular function are regulated by various transcription factors. The T-box (Tbx) family of transcription factors have critical roles in cellular differentiation as well as heart and limbs organogenesis. These factors possess activator and/or repressor domains to modify the expression of target genes. Despite the obvious effects of Tbx20 on heart development, its impact on bone development is still unknown. Methods To investigate the consequence by forced Tbx20 expression in the osteogenic differentiation of human mesenchymal stem cells derived from adipose tissue (Ad-MSCs), these cells were transduced with a bicistronic lentiviral vector encoding Tbx20 and an enhanced green fluorescent protein. Results Tbx20 gene delivery system suppressed the osteogenic differentiation of Ad-MSCs, as indicated by reduction in alkaline phosphatase activity and Alizarin Red S staining. Consistently, reverse transcription-polymerase chain reaction analyses showed that Tbx20 gain-of-function reduced the expression levels of osteoblast marker genes in osteo-inductive Ad-MSCs cultures. Accordingly, Tbx20 negatively affected osteogenesis through modulating expression of key factors involved in this process. Conclusion The present study suggests that Tbx20 could inhibit osteogenic differentiation in adipose-derived human mesenchymal stem cells.


Introduction
Bone is a dynamic tissue undergoing continuous remodeling, which is coordinately regulated by osteoblasts and osteoclasts [1]. A proper balance between bone formation by osteoblasts and bone resorption through osteoclasts guaranties bone maintenance, whereas an imbalance in the activity of these cells leads to bone-associated problems including osteoporosis, osteopetrosis, Padget's disease of the bone and rheumatoid arthritis [1][2][3]. Although the transcriptional regulatory networks involved in bone formation are rather well-defined, the discovery of new transcription factors involved in osteogenesis could further advance our understanding of normal and pathological bone development and homeostasis [4].
The T-box (Tbx) family of transcription factors has widespread roles in development and response pathways [5,6]. T-box proteins contain a highly conserved DNA-binding domain of 180-200 amino acids, socalled T-box domain [4,5,7]. Based on the amino acid sequence of the T-box domain, Tbx family members are divided into five subfamilies: T, Tbx1, Tbx2, Tbx6, and Tbr1 [8]. Together with Tbx1, Tbx10, Tbx15, Tbx18 and Tbx22, Tbx20 belongs to the Tbx1 subfamily. T-box genes are involved in the development of the heart, limbs, mammary glands and cancers [4,5]. Since T-box is highly conserved in T-box family [9], the members of this family may have similar DNA-binding properties. It has been demonstrated that Tbx2, Tbx3, Tbx4, and Tbx5 are extensively detectable in the limb buds and their effects on embryonic skeletogenesis have been welldocumented [10,11]. T-box transcription factors contain activator, repressor or possess both domains, which together with other transcription factors affect the transcriptional activity of target genes in a cell type-specific manner [8,12]. Studies have shown that Tbx20a, i.e. the dominant Tbx20 isoform in the heart, can act both as a transcriptional activator and repressor. Tbx20 is required for vertebrate cardiogenesis [13,14], as revealed by proliferation and chamber and valve formation [7]. Besides, it has been shown that Tbx20 is expressed in different tissues including heart, eye, ventral neural tube and limbs, thus it controls their developmental process [15]. A recent DNA microarray study suggests that skeletal development and cardiac valve maturation are accompanied with changes in the expression of overlapping sets of transcription factor genes including Twist1, Sox9, Msx1, and Msx2 [16]. Recently, it was also reported that Tbx20 plays a pivotal role in facial development [17].
Other T-box genes (e.g. Tbx2, Tbx3, Tbx18 and Tbx22) are involved in bone formation as well as cardiac development [5,18,19]. Although the impact of Tbx20 on heart development is well-investigated [14] its role in bone development has not been studied. To take into account the potential role of Tbx20 in bone progenitors, we hypothesized that Tbx20 may be an important modulator of osteoblast function. To test this idea, Tbx20 was overexpressed in adipose-derived human mesenchymal stem cells (Ad-MSCs) using a vesicular stomatitis virus G protein-pseudotyped human immunodeficiency virus type 1 (HIV1)-based lentiviral vector. These vectors are extensively used in experimental and clinical studies because of their broad tropism, high efficiency and ability to mediate stable transgene in dividing and non-dividing cell populations [20].

Cell culture and markers
Ad-MSCs were obtained from healthy individuals with the approval of the Ethics Committee of Mashhad University of Medical Sciences and with written informed consent of the donors. Ad-MSCs were extracted and cultured by a previously published protocol [21,22]. Briefly, Ad-MSCs from lipoaspirate wastes of healthy donors undergoing aesthetic surgery were isolated by digestion with collagenase type I (0.075%) (Invitrogen, Massachusetts, USA) for 45 min at 37 °C in a shaker water bath followed by the addition of Dulbecco's modified Eagle's medium-low glucose (DMEM-LG) (Gibco BRL, Paisley, Scotland) supplemented with 20% fetal bovine serum (FBS; Gibco BRL, Paisley, Scotland) to inactivate the enzyme. The resulting cell suspension was subsequently centrifuged at 1000g for 15 min, and the cell pellet was then suspended in DMEM-LG containing 15% FBS, and 1× penicillin/streptomycin (Pen/ Strep; Gibco). The cell suspension was maintained at 37 °C and 5% CO 2 in T75 culture flasks. After 3 days, the cells were washed with phosphate-buffered saline (PBS) and cultured in fresh medium, which was subsequently replaced twice a week. The identity of the Ad-MSCs was verified by flow cytometry (BD FACSCalibur System, Biosciences, San Jose, CA) following immunostaining for cell surface antigens at passage three. Mouse monoclonal antibodies directed against CD11b, CD34, CD44, CD45, CD90, and CD105 were purchased from Acris Antibodies (Herford, Germany). FITC-conjugated goat anti-mouse antibodies were obtained from Biolegend (San Diego, CA).
To induce osteogenic differentiation, cells were kept in basal medium containing 0.1 µM dexamethasone, 50 μg/ml ascorbic acid and 10 mM β-glycerophosphate (all from Sigma-Aldrich Chemie, Taufkirchen, Germany). Adipogenesis was induced in the same medium but with 100 µM indomethacin (Sigma-Aldrich Chemie, Taufkirchen, Germany) instead of the ascorbic acid.

Plasmid construction
DNA modifying enzymes used in this experiment were purchased from New England Biolabs (Bioke, Leiden, the Netherlands) and Fermentas (Fisher Scientific, Fermentas, the Netherlands) and applied as specified or by using standard protocols.
The murine Tbx20-coding sequence was excised from plasmid pCRII Topo-Tbx20 (Addgene, Cambridge, MA; plasmid number: 24715) with EcoRV and SpeI and inserted into the polylinker of the lentiviral vector shuttle construct pLV.hCMV-IE.IRES.EGFP.hHBVPRE that was previously known as pLV-CMV-IRES-EGFP [23] digested with the same enzymes. The resulting plasmid, which codes for Tbx20 and the enhanced green fluorescent protein (EGFP), was designated pLV-Tbx20. To verify the correctness of pLV-Tbx20, restriction endonuclease digestions and nucleotide sequence analysis were performed. Sequencing was carried out by Leiden Genome Technology Center (http://www.lgtc.nl/) using a BigDye Terminator v3.1 Cycle Sequencing Kit and a 3730xl DNA Analyzer (both from Applied Biosystems). A schematic genetic maps of pLV-Tbx20 and pLV-Ctrl constructs are provided in Fig. 1.

Lentiviral vector production
To produce lentiviral vector particles, human embryonic kidney (HEK) 293T cells were co-transfected at a molar ratio of 1:1:2 with packaging plasmids psPAX2 (Addgene; plasmid number; 12260) and pLP/VSVG (Life Technologies Europe) and with lentiviral vector shuttle plasmid pLV-Tbx20 (to generate LV-Tbx20 particles) or pLV.hCMV-IE.IRES.EGFP.hHBVPRE (to produce control vector [LV-Ctrl] particles). Human embryonic kidney 293T cells were grown in DMEM-HG supplemented with 10% FBS. After the cells reached 70-80% confluency, the transfection medium was added containing the packaging plasmids supplemented with NaCl 150 mM and 25 kDa linear polyethylenimine (PEI; Polysciences Europe, Hirschberg an der Bergstrasse, Germany). The DNA-PEI complexes were left on the cells overnight. The next day, the transfection medium was replaced by DMEM-LG supplemented with 5% FBS and 25 mM HEPES-NaOH (pH 7.4). Producer cell supernatants were collected 48 h after the start of the transfection procedure and following low speed centrifuge, they were filtered through 0.45-μm pore-sized, 33-mm diameter polyethersulfone Millex-HP syringe filters (Millipore, Amsterdam, The Netherlands). To concentrate the vector particles, the cleared supernatants were underlayed with a cushion of 20% (wt/vol) sucrose followed by ultracentrifugation at 15,000 revolutions per minute and 4 °C for 2 h in a SW32 swinging bucket rotor (Beckman Coulter Nederland, Woerden, The Netherlands). The supernatant was removed completely and viral vectors were then suspended in PBS containing 1% bovine serum albumin (Sigma-Aldrich Chemie, Taufkirchen, Germany). Finally, viral vectors were aliquoted in 1.5 ml microtubes and stocked at − 80 °C for subsequent investigations.

Lentiviral transduction
Ad-MSCs were plated at density of 5 × 10 3 cells/cm 2 . The next day, the cells were infected with LV-Tbx20 or LV-Ctrl particles in culture medium containing 2 µg/ ml DEAE-dextran sulfate (Sigma-Aldrich Chemie, Taufkirchen, Germany) under antibiotic-free conditions. After 24 h, the cells were washed with PBS and the transfection medium was replaced with osteogenic differentiation medium. The transduction efficiency was checked by direct EGFP fluorescence microscopy. All experiments were performed in triplicate.

Mineralization and alkaline phosphatase (ALP) activity assays
To stain calcific deposits as previously described [24], after 4th week of differentiation, cells were fixed with 10% formalin (Sigma-Aldrich Chemie, Taufkirchen, Germany) for 45 min at room temperature (RT). After washing away the fixative with water, the fixed cells were incubated with Alizarin Red S (ARS, Sigma-Aldrich Chemie, Taufkirchen, Germany) staining solution of pH 4.1-4.3 for 45 min at RT. Next, unbound ARS was aspirated out and washed away with water. Then, mineralized/red bone nodules were visualized by light microscopy (Nikon Eclipse TE2000-U, Tokyo, Japan). Stained cells were subjected to ARS quantification using 10% acetic acid (Sigma-Aldrich Chemie, Taufkirchen, Germany). The resulting slurry was transferred to microcentrifuge tube followed by heating at 85 °C and cooling on ice. Then, Fig. 1 Map of LV DNA in the LV shuttle plasmids. pLV.Tbx20 (a), pLV.Ctrl (b). 5′ LTR chimeric 5′ long terminal repeat including the Rous sarcoma virus U3 region as well as the HIV1 R and U5 regions. Ψ HIV1 packaging signal, RRE HIV1 Rev-responsive element, cPPT HIV1 central polypurine tract and termination site, hCMV-IE human cytomegalovirus immediate early gene promoter, Tbx20 coding sequence of desired gene, EMCV IRES encephalomyocarditis virus internal ribosomal entry site, EGFP enhanced green fluorescent coding sequence, hHBVPRE human hepatitis B virus posttranscriptional regulatory element, 3′ LTR 3′ HIV1 long terminal repeat involving a deletion in the U3 region to render the LV self-inactivating the tubes were centrifuged at 20,000g for 15 min and 10% ammonium hydroxide was added to the supernatant. In the final step, absorbance was read at 405 nm using a BioTek plate reader (Bad Friedrichshall, Germany). To detect alkaline phosphatase (ALP) activity, cells were cultured in differentiation medium for about 21 days. Then, they were fixed with 10% formalin, washed with PBS, and finally stained with BCIP/NBT (5-bromo-4-chloro-3-indolyl phosphate/nitroblue tetrazolium; Sigma, Germany). The ALP activity was visualized as a dark purple color. To measure ALP activity, differentiated cells were detached with lysis buffer (50 mM Tris-HCl [pH 7.4]), 1% Triton X-100 (Merck, Germany) followed by incubating with p-nitro-phenyl phosphate (pNPP; Sigma-Aldrich Chemie, Taufkirchen, Germany) for 30 min at RT. Finally, the chromogenic reaction product (i.e.; p-nitro-phenol [pNP]) was detected spectrophotometrically at 405 nm.

Immunostaining
Two weeks after osteogenic induction, cells were fixed with 4% paraformaldehyde (in 1× PBS) at 4 °C for 30 min, permeabilized by incubation with 0.1% Triton X-100 in PBS and incubated overnight at 4 °C with mouse antialkaline phosphatase (ALP; BD Biosciences, San Diego, CA) and anti-collagen type I (ColI; Sigma-Aldrich Chemie, Taufkirchen, Germany), diluted as 1:200 and 1:1000 in PBS and 0.1% donkey serum, respectively (Santa Cruz Biotechnology, Santa Cruz, CA). The next day, cells were incubated at RT for 4 h with Alexa Fluor 568-conjugated donkey anti-mouse IgG (H + L) secondary antibody (Life Technologies Europe, Bleiswijk, The Netherlands) diluted as 1:400 in PBS and 0.1% donkey serum. Then, nuclei were incubated with Hoechst 33258 (Sigma-Aldrich Chemie, Taufkirchen, Germany) for 10 min at RT. After each incubation step, cells were thoroughly rinsed with PBS. Pictures were taken with an inverse fluorescence microscope (Nikon eclipse TE2000-U) attached to digital color camera. ImageJ, version 4.1 (National Institutes of Health, Bethesda, MA) was used for mean fluorescence signal intensity.

Reverse transcription-quantitative polymerase chain reaction (RT-qPCR) analyses
To measure transcript levels of various genes, total RNA was isolated from cells at 7 and 14 days of osteogenic differentiation using Hybrid-R ™ RNA purification kit (GeneAll, Seoul, Republic of Korea). Next, first strand-cDNA was synthesized by reverse transcription with GeneAll's HyperScript kit (Seoul, Republic of Korea). Expression levels of desired genes was accessed using SYBR Green (Amplicon, Odense, Denmark)-based real time PCR (StepOne detection system; Applied Biosystems, CA, USA). Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was applied as reference gene and gene expression changes were calculated using 2 −ΔΔCt method. Forward and reverse primer sequences (Takapouzist, Tehran, Iran) used are shown in Table 1.

Statistical analysis
All experiments were performed in triplicate and all data were expressed as the mean ± standard deviation (SD). The statistical analysis was performed using Prism 6 (Graphpad Software, San Diego, CA). Significant associations between various groups were analyzed by ANOVA and defined as p values.

Identification of human Ad-MSCs
The mesenchymal stem cells isolated from human adipose tissue had a spindle-shaped fibroblastic appearance and could form single-cell. As previously demonstrated, immunostaining and flow cytometric analysis revealed  that Ad-MSCs expressed the mesenchymal stem cell markers CD44, CD90, CD105 at their surface while they were negative for CD11b, CD45, CD34 [21]. Culturing of Ad-MSCs in osteogenic differentiation medium stimulated calcification as evidenced by the bright-red mineralized nodules observed following ARS staining. Moreover, when the Ad-MSCs were cultured in adipogenic differentiation medium, the cells accumulated lipid droplets as indicated by Oil Red O. The results of the osteogenesis and adipogenesis assays showed that the Ad-MSCs possess multilineage differentiation potential capacities (Additional file 1: Fig. S1A and B).

Production and functional testing of LV-Tbx20 and LV-Ctrl particles
Twenty-four hours after the addition of the lentiviral vector shuttle and packaging plasmids, essentially all 293T producer cells were EGFP positive indicative of a high transfection efficiency. To test the gene transfer activity of the LV-Tbx20 and LV-Ctrl preparations, 5 × 10 4 / cm 2 293T indicator cells were incubated with 2 μl of a 1000-fold dilution of both vector stocks. This resulted in the expression of EGFP in virtually all cells confirming that both vectors were functional and had a similar titer (Additional file 2: Fig. S2A, B).

The effect of Tbx20 overexpression on the osteogenic differentiation ability of Ad-MSCs
To investigate the effects of Tbx20 expression on osteogenesis, Ad-MSCs were transduced with LV-Tbx20, which expresses both Tbx20 and EGFP, or with LV-Ctrl, which only codes for EGFP. The transduction efficiency of the Ad-MSCs at 72 h post-transduction was about 90% as judged by direct EGFP fluorescence microscopy (Additional file 2: Fig. S2A, B). The transduced cells were subsequently cultured in osteogenic differentiation medium for up to 21 days (Fig. 2a, b), after which they were subjected to ARS staining and ALP activity measurement. Besides, the expression of murine Tbx20 after 14 days showed 20 times more expression in LV-Tbx20 transduced cells in comparison with LV-Ctrl cells (Fig. 2c). The ARS staining (Fig. 3a) revealed that Tbx20 overexpression reduced the formation of mineralized nodules in Ad-MSCs. Consistent with these findings, ARS quantification confirmed less calcification in LV-Tbx20-treated cells than in LV-Ctrl-transduced Ad-MSCs (Fig. 3b). Also, staining the cells with BCIP/NBT revealed that in LV-Tbx20 group the blue cells were less prominent than LV-Ctrl group (Fig. 3c). ALP activity in Tbx20-overexpressing Ad-MSCs at 14 days after initiating osteogenesis was about 1.2-fold lower than in LV-Ctrl-transduced Ad-MSCs (Fig. 3d). This was accompanied by a reduction in immunostaining of ALP and ColI in LV-Tbx20-expressing cells in comparison with LV-Ctrl-transduced Ad-MSCs (Fig. 4). Collectively, these results demonstrate an inhibitory effects of Tbx20 overexpression on the osteogenic differentiation ability of Ad-MSCs.

The effect of Tbx20 overexpression on osteoblastic gene expression in Ad-MSCs
To study the negative effects of Tbx20 overexpression on osteogenesis in more detail, the mRNA levels of osteoblast-related genes in LV-Tbx20-and LV-Ctrl-transduced Ad-MSCs at different times after osteoinduction were evaluated by RT-qPCR. Overexpression of Tbx20 reduced the expression of the osteoblast differentiation markers ALPL (Fig. 5a), COL1A1 (Fig. 5b), RUNX2 (Fig. 5c), and OPN (Fig. 5d) in Ad-MSCs in a time-dependent manner. In comparison, human Tbx20 expression has been demonstrated in Fig. 5e. At 14 days after the initiation of osteogenesis, ALPL, COL1A1, RUNX2, and OPN mRNA levels were respectively, 4-, 4.6-, 1.6-and 3-fold lower in LV-Tbx20-treated Ad-MSCs than in LV-Ctrl-transduced Ad-MSCs (p < 0.05). Enhanced expression of Tbx20 in Ad-MSCs reduced osteogenesis. Cells were infected with control viral vector (LV-Ctrl) or LV-Tbx20 and allowed to differentiate in osteogenic medium. a After 21 days, mineral nodules deposition was stained with Alizarin Red S (ARS). b The stained cells were extracted using 10% acetic acid and its absorbance was read at 405 nm to quantify the mineralization level. c After 21 days, induced cells were stained with BCIP/NBT to detect secreted alkaline phosphatase. d After 7 and 14 days ALP activity was conducted using pNPP liquid substrate system. All data are presented as mean ± SD. ****Shows significant differences (p < 0.0001) between negative control (cells grown in normal medium) and infected groups; ***Indicates significant differences (p < 0.001) between LV-Ctrl and LV-Tbx20. Experiments were performed in triplicate and repeated three times

Discussion
Tissue development is spatiotemporally regulated by transcription factors and noncoding RNAs including microRNAs, which either stimulate or inhibit the expression of target genes in a cell type-specific manner [25]. Several factors negatively regulate bone formation including microRNAs such as miR-378, miR-138, and miR-182 [26,27] and transcription factors such as NFE2 [2], GATA4 [28], and HOXC8 [29] as well as p53 as a tumor suppressor [30]. Previously, we have demonstrated that Tbx20 has cardiogenic effects [22]. In the present study, we introduced Tbx20 in Ad-MSCs using a HIV1-based lentiviral vector to test its modulatory effects on osteogenesis. Our results demonstrated that Tbx20 reduced the osteoblastic differentiation of Ad-MSCs. ARS staining revealed decreased calcified nodule formation and ALP assays showed a reduction in ALP activity in Tbx20overexpressed cells compared to control cells. Moreover, Fig. 4 Increased expression of Tbx20 decreased ALP and ColI expression. Representative micrograph of 14 days incubation of Ad-MSCs exposed to LV-Ctrl or LV-Tbx20 with EGFP-positive, ALP and ColI expressions (a-d). Ad-MSCs overexpressed Tbx20 demonstrated two-and threefolds reduction in ALP and ColI expression respectively compared to LV-Ctrl (e, f). ****Shows significant differences (p < 0.0001) between LV-Ctrl and LV-Tbx20. Experiments were performed in triplicate and repeated three times the proteins expression of ALP and ColI significantly decreased in LV-Tbx20-treated Ad-MSCs in comparison to LV-Ctrl-transduced Ad-MSCs. Similar results were obtained by RT-qPCR in which it has been demonstrated that Tbx20 overexpression significantly down-regulated ALPL, COL1A1, RUNX2, and OPN expressions on Ad-MSCs in a time dependent manner. Possible mechanisms involved in the osteosuppressive effects of Tbx20 could be explained by its binding to important osteoblast transcription factors and changing their function and/ or expression at key stages of differentiation. It has been well-documented that Tbx20 can act as a transcriptional repressor [31], raising the possibility that Tbx20 could down-regulate osteoblastic gene expression in Ad-MSCs. RUNX2 is a key transcription factor regulating expression of various osteoblast marker genes including those encoding type Iα1 collagen, osteocalcin, ALP, and osteopontin [32]. As it was assumed, overexpression of Tbx20 in Ad-MSCs indeed down-regulated RUNX2, ALPL and COL1A1 mRNA levels. It has been reported that Tbx3 blocks osteoblast differentiation through inhibition of Osterix and RUNX2 [4]. As mentioned, the members of T-box family have similar DNA-binding properties [10,11]. Therefore, a more likely explanation for our findings could be that Tbx20 similar to Tbx3 regulated osteogenesis by binding to master transcription factor i.e. RUNX2. Consistently, it was shown that increased expression of Tbx3 could mislocalize RUNX2 [33]. Similarly, Tbx15 deficiency in mutant mice not only leads to reduction in bone size and length but also delays endochondral bone formation. Interestingly, overexpression of Tbx2 in NIH3T3 cell line up-regulates osteoblastogenesis and chondroblastogenesis [34]. Accordingly, we could not ignore the potential role of Tbx20 in regulation of bone formation although some probable functions of recombinant Tbx20 could be hindered by spatial interference , and human Tbx20 (e) in LV-Ctrl-and LV-Tbx20-transduced Ad-MSCs that were maintained in osteogenic differentiation medium for 7 or 14 days. Expression levels were normalized to those in untreated cells (i.e. cells cultured in normal growth medium) as mean ± SD. The GAPDH gene served as internal control. ****p< 0.0001. ***p< 0.001. **p< 0.01. ns not significant. Experiments were performed in triplicate and repeated three times