Osteogenic Differentiation of Human Adipose Tissue-Derived Stem Cells Is Modulated by the miR-26a Targeting of the SMAD1 Transcription Factor

Authors

  • Ettore Luzi,

    1. Metabolic Bone Unit, Laboratory of Molecular Genetics, Department of Internal Medicine, University of Florence, Florence, Italy
    Search for more papers by this author
  • Francesca Marini,

    1. Metabolic Bone Unit, Laboratory of Molecular Genetics, Department of Internal Medicine, University of Florence, Florence, Italy
    Search for more papers by this author
  • Silvia Carbonell Sala,

    1. Metabolic Bone Unit, Laboratory of Molecular Genetics, Department of Internal Medicine, University of Florence, Florence, Italy
    Search for more papers by this author
  • Isabella Tognarini,

    1. Metabolic Bone Unit, Laboratory of Molecular Genetics, Department of Internal Medicine, University of Florence, Florence, Italy
    Search for more papers by this author
  • Gianna Galli,

    1. Metabolic Bone Unit, Laboratory of Molecular Genetics, Department of Internal Medicine, University of Florence, Florence, Italy
    Search for more papers by this author
  • Maria Luisa Brandi

    Corresponding author
    1. Metabolic Bone Unit, Laboratory of Molecular Genetics, Department of Internal Medicine, University of Florence, Florence, Italy
    2. DeGene Spin-off, University of Florence, Florence, Italy
    • Maria Luisa Brandi, MD, PhD, Department of Internal Medicine, University of Florence, Viale Pieraccini 6, 50139 Florence, Italy
    Search for more papers by this author

  • The authors state that they have no conflicts of interest

  • Published online on October 8, 2007

Abstract

The molecular mechanisms that regulate hADSC differentiation toward osteogenic precursors and subsequent bone-forming osteoblasts is unknown. Using osteoblast precursors obtained from subcutaneous human adipose tissue, we observed that microRNA-26a modulated late osteoblasts differentiation by targeting the SMAD1 transcription factor.

Introduction: Elucidation of the molecular mechanisms guiding human adipose tissue-derived stem cells (hADSCs) differentiation is of extreme importance for improving the treatment of bone-related diseases such as osteoporosis. The aim of this study was to identify microRNA as a regulator of the osteogenic differentiation of hADSCs.

Materials and Methods: Osteoblast differentiation of hADSCs was induced by treatment with dexamethasone, ascorbic acid, and β-glycerol phosphate. The expression of osteoblastic phenotype was evaluated after the induction by simultaneous monitoring of alkaline phosphatase activity, the expression of genes involved in osteoblastic differentiation by real-time RT-PCR, and mineralization at the same time. MicroRNA expression was determined by Northern blot, and transfection of both antisense miR-RNA and sensor plasmids was done to validate the inhibitory role of microRNA during hADSC osteogenesis. Western blot was used to determine the expression levels of the SMAD1 protein. qRT-PCR analysis was used to compare the expression patterns of osteoblastic markers in transfected cells.

Results and Conclusions: We analyzed the role of microRNA 26a (miR-26a) during differentiation of hADSCs. Northern blot analysis of miR-26a during hADSC differentiation showed increased expression, whereas expression of the SMAD1 protein was complementary to that of miR-26a. Because the highest expression of miR-26a and the lowest expression of SMAD1 protein were reached at hADSC terminal differentiation, we carried out our study during the late stages of hADSC differentiation. The inhibition of miR-26a, by 2′-O-methyl-antisense RNA, increased protein levels of its predicted target, SMAD1 transcription factor, in treated osteoblasts, upregulating bone marker genes and thus enhancing osteoblast differentiation. Our data suggest a role for miR-26a in the differentiation induced by treatment with dexamethasone, ascorbic acid, and β-glycerol phosphate of hADSCs toward the osteogenic lineage by targeting its predicted target, the SMAD1 protein. This study contributes to a better knowledge of molecular mechanisms governing hADSC differentiation by proposing a microRNA-based control of late differentiation.

INTRODUCTION

The first microRNA (miRNA) was identified in Caenorhabditis elegans as a gene important for timing larval development.(1,2) miRNAs are a class of small (∼22 bp), noncoding RNAs that regulate gene expression at a post-transcriptional level. They are first transcribed as primary miRNAs by RNA polymerase II and cut by the RNase III enzyme, Drosha, into ∼70-nucleotide precursors (pre-miRNAs), which are transported to the cytoplasm by Exportin 5. Another enzyme, Dicer, processes the pre-miRNAs to mature miRNAs, which are recruited into the RNA-induced silencing complex (RISC). Finally, this complex interferes with the translation and stability of target mRNAs: these RISC complexes either bind mRNA with exact complementarity, leading to cleavage of the mRNA by the RISC, or bind with partial complementarity, leading to repression of translation. miRNAs have since been implicated in many processes in invertebrates, including cell proliferation and apoptosis,(3,4) fat metabolism,(3) neuronal patterning,(5) and tumorigenesis.(6) Because several miRNAs are conserved across species,(7–10) they are likely to be involved in developmental processes in all animals. Only a few mammalian miRNAs have been assigned a functional role, and at least five of these are involved in developmental processes. miR-181 promotes B cell development in mice,(11) and it targets the homeobox protein Hox-A11 during mammalian myoblast differentiation.(12) miR-196a regulates several Hox genes,(13) which encode for a family of transcription factors involved in various developmental processes in animals.(13) The brain-specific miR-134 regulates dendritic spine development.(14) miR-1, miR-133 and miR-206 are specifically induced during myogenesis.(15,16) Finally, miR-143 regulates adipocytic differentiation.(17) Like bone marrow, adipose tissue is a mesodermally derived organ that contains a stromal population encompassing microvascular endothelial, smooth muscle, and stem cells.(18) These cells can be enzymatically digested out of adipose tissue and separated from buoyant adipocytes by centrifugation. This population (termed adipose tissue-derived stem cells [ADSCs]) shares many of the characteristics of its counterpart in bone marrow including extensive proliferative potential and the ability of differentiating toward adipogenic, osteogenic, chondrogenic, myogenic, and neurogenic lineages.(19–21) The osteogenic potential of hADSCs has been previously described,(22) showing the in vivo bone-forming capacity of these cells.

Among the molecules capable of inducing osteoblastic differentiation, bone morphogenetic proteins (BMPs) are recognized as potent osteotropic agents, capable of inducing both osteoblast differentiation and bone formation. The signal transduction of BMPs involve SMAD proteins. SMAD1 is the downstream effector of BMP signaling, and it is phosphorylated by BMP type I receptors. The phosphorylation of SMAD1 induces its accumulation in the nucleus, where it regulates gene transcription by associating with a nuclear transcription factor or by binding directly to DNA.(23) SMAD1 protein interacts with homeodomain transcription factor HOXC-8.(24) Through this interaction, SMAD1 activates osteopontin gene expression in response to BMP stimulation by means of dislodging HOXC-8 from the promoter in response to BMP signaling.(25)

Currently, it is not known whether miRNAs play a role in regulating components of the osteoblastic differentiation. Here we show that miR-26a, a microRNA that is upregulated during osteogenic differentiation of hADSCs, modulates this process by binding to specific target sequences harbored in the 3′-untranslated region (UTR) of SMAD1 mRNAs, as predicted by bioinformatic approaches.(26)

MATERIALS AND METHODS

Cell cultures and osteogenic differentiation

ADSCs were isolated from adipose tissue obtained from the subcutaneous abdominal depot during herniotomy in accordance with a protocol approved by the Institutional Review Board for human studies. Knife biopsies of adipose tissue were immediately placed in McCOY'S 5A medium (Sigma Aldrich, St Louis, MO, USA) supplemented with sterile 22 mM HEPES (Sigma Aldrich), 100 IU/ml penicillin, and 100 μg/ml streptomycin, pH 7.4, transported to the laboratory, and processed within 30 min from excision. Using sterile technique, the tissue was cut into small pieces of ∼0.2–0.5 mm. After mincing, fragments were washed extensively with Dulbecco's PBS with calcium and magnesium (DPBS; BioWhittaker, Cambrex, Belgium) to remove contaminating debris, centrifuged at 200g for 10 min, resuspended in Ham's F12 Coon's modification medium supplemented with 20% FBS and 0.3 mg/ml collagenase type I (C-0130; Sigma Aldrich), and digested for 12 h at 37°C in plastic culture dishes. After enzymatic digestion, the contents of the dishes were mechanically dispersed and passed through a sterile 230-μm stainless steel tissue sieve into a 50-ml sterile plastic tube. Undigested stromal-vascular tissue trapped on the sieve was discarded, whereas the infranatant containing the pre-adipocyte fraction was collected, and the cells were sedimented by centrifugation at 300g for 5 min. The pellet was incubated with an erythrocyte lysis buffer (155 mM NH4Cl, 5.7 mM K2HPO4, 0.1 mM EDTA, pH 7.3) for 10 min at room temperature to eliminate red blood cells. After centrifugation, the stromal-vascular cell fraction was suspended and cultured in growth medium composed of Coon's modified Ham's F12 medium supplemented with 10% FBS, 100 IU/ml penicillin, 100 μg/ml streptomycin, and 1 ng/ml basic fibroblast growth factor (bFGF; Sigma Aldrich) in 100-mm culture plates at 37°C with 95% relative humidity in 5% CO2. Medium was refreshed twice a week, and cells were used for further subculturing or cryopreservation on reaching 3 × 105 cells/plate. All the cells used in the experiments were in three to seven passages.

Osteogenic differentiation and transfection of oligoribonucleotides

Transfection of 250 nM 2′-O-methyl RNAs in lipofectamine 2000 (Invitrogen) was performed according to manufacturer's instructions. Cells were plated in 100-mm culture plates in growth medium and grown to 60–70% confluence. Afterward, the medium was switched to Ham's F12 Coon's modification medium supplemented with 10% FBS (South America origin; BioWhittaker), 100 IU/ml penicillin, 100 μg/ml streptomycin, 10 nM dexamethasone (Sigma Aldrich), 0.2 mM ascorbic acid (Merck, Darmstadt, Germany), and 10 mM glycerol phosphate (Sigma Aldrich). Medium was refreshed twice a week. The expression of the osteoblastic phenotype was evaluated 10, 20, and 30 days after induction by the simultaneous monitoring of alkaline phosphatase (ALP) activity (histochemical staining and enzymatic assay), expression of genes involved in osteoblastic differentiation (real time RT-PCR), and mineralization (Alizarin red S, von Kossa, and calcein staining). Thirty days after osteogenic induction, ∼40% of hADSCs showed an osteoblastic-like phenotype.

Oligoribonucleotides, reporter plasmids, and luciferase assays

2′-O-methyl oligoribonucleotides (miR-26a antisense, 5′-GCCUAUCCUGGAUUACUUGAA-3′; miR-26a antisense mutated, 5′-GCCAAUCCUGGAUUACGGGAA-3′) were synthesized by IBA (Gottingen). Reporter constructs that contain a miR-26a binding site (pGL3-26a) or a mismatch sequence (pGL3-26aMUT) in the 3′-UTR of Smad1 mRNA were kindly provided by D Bartel.(27) Reporter or control plasmid (200 ng) plus 80 ng pRL-null (Promega) were transfected alone or in combination with 40 pmol of 2′-O-methyl oligoribonucleotides, using lipofectamine 2000 (Invitrogen), and hADSCs were transfected and placed under differentiation conditions as described above. Luciferase assays were performed 48 h after transfection using the Dual Luciferase Reporter Assay System (Promega). Firefly luciferase activity was normalized to Renilla luciferase activity for each transfected well. Each transfected well was assayed in triplicate.

Northern blot analysis

Total RNA was isolated with Trizol reagent (Invitrogen) according to the manufacturer's instructions. For miRNA Northern blots, 15 μg of total RNA was separated on 15% denaturing polyacrylamide gels, electrotransferred to GeneScreen Plus membranes, and hybridized using UltraHyb-Oligo buffer (Ambion) at 42°C overnight. Oligonucleotides complementary to mature miRNAs were end-labeled with T4 Kinase (Roche) and were used as probes. Probe sequences were as follows: miR-26a antisense, 5′-AGCCTATCCTGGATTACTTGAA-3′; U6 antisense, 5′-GCCATGCTAATCTTCTCTGTATC-3′.

RNase protection assay

The RNase protection assay was performed with an RPAIII kit (Ambion).

Antisense RNA probes for Smad1 mRNA and 18S RNAs were synthesized by MEGAshortscript kit (Ambion). The antisense RNA probes labeled with (α-t32P)UTP using T7 RNA polymerase were hybridized to total RNA (5 μg) from hADSC-transfected cells at 50°C overnight and digested by RNaseA/T1. The protected mRNA fragments were separated by electrophoresis through 5% polyacrilamide/1× Tris-borate EDTA (TBE)/9 M urea gels with 1× TBE as the running buffer, and mRNA was detected and quantified with a Cyclone Phosphoimager (Perkin-Elmer).

Quantitative real-time PCR

Real-time RT-PCR was carried out on bone marker genes to check mRNA levels. Two independent experiments using two separate cell samples were conducted. Ten microgrmas of total RNA from each sample was DNase treated with the DNA-free kit (Ambion, Austin, TX, USA). The purity and concentration of the RNA was checked with an ND-1000 Spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA), and a 10-ng/ml dilution was prepared for successive qRT-PCR reactions. Two milliliters of diluted RNA was amplified using a one-step Brilliant SYBR Green QRT-PCR Master mix kit (Stratagene) and 400 nM of forward and reverse primers (Table 1). Each RNA sample was evaluated for transcript levels in triplicate (including the normalization control β-actin) and quantified with an MX3000P multiplex quantitative PCR instrument (Stratagene) following a unique three-step protocol: 1 cycle of initial incubation at 50°C for 30 min for the reverse reaction, 1 cycle at 95°C for 10 min to fully activate the DNA polymerase, and 40 amplification cycles (95°C for 1.5 min at the appropriate annealing temperature and 72°C for 30 s). Sample fluorescence was detected during the annealing step. Finally, the amplified product was incubated for 1 min at 95°C to denature the PCR product, ramped down to 55°C, and finally ramped up from 55°C to 95°C for a dissociation curve, at a rate of 0.2°C/s. By collecting fluorescence data continuously, we obtained a dissociation curve. Fluorescence was plotted versus the Ct (threshold cycle) based on dRn (baseline-corrected, reference dye-normalized fluorescence) to obtain the standard curve and to measure the initial template quantity. Gene expression was normalized to β-actin. Fluorescence data were analyzed using MXPro Software.

Table Table 1.. Primers for Quantitative RT-PCR of Genes Involved in Osteogenic Differentiation of hADSCs
original image

Western blotting

Total protein extracts were prepared in a cell disruption buffer (Ambion). The purified protein concentrations were determined by the BCA protein assay (Pierce, Rockford, IL, USA) using BSA as the standard protein. SMAD1 protein expression was evaluated by Western blotting analysis. Twenty micrograms of protein was denatured for 10 min at 95°C in a volume of 95% Laemmli sample buffer (Bio-Rad, Hercules, CA, USA) and 5% β-mercaptoethanol solution. Proteins were separated by SDS-12% PAGE. The separated protein bands were transferred electrophoretically to a nitrocellulose membrane (Optitran BA-S 83; Schleicher & Schuell, Dassel, Germany) for 1 h in a Mini Transblot electrophoretic transfer cell (Bio-Rad) at 100 V. Filters were blocked for 1 h at room temperature in a blocking solution of PBS, 0.1% Tween 20, and 2% ECL Advance blocking agent (Amersham Biosciences). To evaluate SMAD1 expressions, the blocked membranes were incubated with mouse anti-Smad1 (A-4) polyclonal antibody (Santa Cruz Biotechnologies) for 1 h at room temperature (1:500 dilution of stock in blocking solution), washed four times for 5 min with a solution of PBS and 1% Tween 20, incubated with an anti-mouse horseradish peroxidase-conjugated secondary antibody (Bethyl, Montgomery, TX, USA) for 1 h at room temperature (1:2000 dilution of stock in blocking solution), washed four times for 5 min with a solution of PBS and 1% Tween 20, and finally washed once with PBS. Detection of membrane-bound antibodies was performed by chemiluminescence. β-actin was used as the standard protein.

Statistical analysis

Statistical analyses were performed using a Student's t-test. p < 0.05 was considered statistically significant.

RESULTS

miR-26a interacts with the 3′-UTR of the Smad1 mRNA

Bioinformatic approaches have been used to identify potential miRNA targets.(26) TargetScan (http://genes.mit.edu/targetscan) software predicted two target sites for miR-26a in orthologous Smad1 3′-UTR sequences from humans (Hs), rats (Rn), and Fugu (Fr).(27)

Because the SMAD1 protein plays a pivotal role in osteoblastic differentiation, we studied the expression pattern of both miR-26a and SMAD1 protein during hADSC osteogenic differentiation. Northern blot for mature miR-26a (Figs. 1A and 1B) at various times after induction showed that expression of miR-26a increased with time, reaching maximum levels when osteogenic maturation occurred. The overexpression of this and other microRNAs in various differentiated tissue implicated microRNAs in the determination or maintenance of the differentiated state. Osteoblastic differentiation was confirmed by the increased expression of genes associated with osteoblastic differentiation, such as bone ALP (Alp), type I collagen (ColI), osteopontin (Opn), runx2/cbfa1 (Runx2), and osteocalcin (Ocn) (Fig. 1D). The pattern of expression of the SMAD1 protein was complementary to that of miR-26a: SMAD1 was highly expressed in cells with low miR-26a levels, whereas SMAD1 downregulation coincided with miRNA induction during terminal differentiation (Fig. 1C). Because hADSCs reached terminal osteogenic differentiation 4 wk after induction with osteogenic medium ((as showed by calcified extracellular matrix; Fig. 1E), when miR-26a expression was the highest and SMAD1 expression lowest, this study was carried out during the late stages of hADSC differentiation.

Figure Figure 1.

Kinetics of the induction of the endogenous miR-26a in differentiating osteoblasts. (A) Northern blotting for mature miR-26a in differentiating hADSCs at various times after induction. The blot was reprobed for U6 snRNA as a loading control. (B) Quantitation of Northern blot. RNA was quantified after standardization with U6 small RNA. Bars indicate SD. *,**,***p < 0.05 (paired Student's t-test, n = 3) with respect to the control sample, the undifferentiated cells. (C) Quantitation of Western blot for SMAD1 protein in differentiating hADSCs at various times after induction. Western blot was quantified after standardization with β-actin. Bars indicate SD. *,**,***p < 0.05 (paired Student's t-test, n = 3) with respect to the control sample, the undifferentiated cells. (D) Osteoblastic marker genes expression (I collagen [ColI], ALP [Alp], osteopontin [Opn], runx2/cbfa1 [Runx2], and osteocalcin [Ocn]) was measured by absolute qRT-PCR in undifferentiated (dashed columns) and differentiated (solid columns) cells, respectively, at time 0 and after 30 days of exposure to the differentiation medium. Bars indicate SD. (E) The expression of the osteoblastic phenotype was evaluated 30 days after the induction by the monitoring of the mineralization (Alizarin red S).

2′-O-Me RNA inhibition of miRNA activity

2′-O-methyl oligoribonucleotides were characterized as sequence-specific inhibitors of miRNA function and miRNA-direct RNA-induced silencing complex (RISC) activity.(28) These molecules stoichiometrically bind and irreversibly inactivate miRNAs, providing a valuable tool for inhibiting the function of a single miRNA in vitro and in vivo. To monitor the degree of miRNA inhibition, the inhibition of miRNA 26a activity was verified using pGL3/miR-26a, a luciferase expression plasmid containing two complementary miR-26a binding sites.(27) Both the luciferase expression plasmid containing two complementary miR-26a binding sites ((pGL3/26a) and an analogous reporter with point substitutions disrupting the pairing to the miRNA seed of miR-26a as an negative control (pGL3/26MUT) were transfected in differentiating osteoblastic cells. The insertion of wildtype sequences rendered the reporter sensitive to endogenous miR-26a, with a reduction in luciferase activity compared with the analogous mutated reporter (Fig. 2A). Thus, the effect of cellular endogenous miR-26a on translation of the luciferase mRNA was dependent on the presence of the miR-26a cognate binding site within the 3′-UTR, because expression of the luciferase reporter containing the mutant miR-26a binding site within the 3′-UTR was unaffected by the presence of endogenous cellular miR-26a.

Figure Figure 2.

Luciferase assay and effects of miRNA inhibition on differentiating osteoblasts. (A) WT reporter or mutated control luciferase plasmids were transfected into hADSCs alone or with 250 nM of 2′-O-methyl miR-26a antisense RNA or 2′-O-methyl miR26a mutated antisense RNA. Cells were placed in osteogenic differentiation media. Reporter activities were measured after 2 days in differentiation media and were normalized to Renilla luciferase activity. Bars indicate SD. *p < 0.05 (paired Student's t-test, n = 3) with respect to the control sample. (B) Northern blot experiment showing selective inhibition of the endogenous miRNA 26a with 2′-O-methyl antisense RNA (WT) and with mutated control 2′-O-methyl RNA (mut). U6 was used as a loading control.

Co-transfection of the reporter plasmid with 2′-O-methyl RNA antisense miR-26a (pGL3/26a wt + 2′-O-Me) enhanced the expression of the reporter construct, indicating inhibition of endogenous miRNAs. In contrast, co-transfection of the reporter plasmid with control antisense RNA (pGL3/26a +mut 2′-O-Me) did not have the same effect (Fig. 2A). This antisense-based loss of function assay showed inhibition of endogenous miR-26a at a functional level.

Taken together, these results indicated that antisense 2′-O-methyl RNA can be used to analyze miRNA function in human osteoblastic cells along the pathway of differentiation from hADSCs. The effect of miR-26a antisense 2′-O-methyl RNA treatment on endogenous miR-26a levels in differentiating osteoblasts was also determined. These RNA oligonucleotides were transfected into osteoblasts at the time of onset of the miRNA expression, and the endogenous expression levels of miR-26a were analyzed by Northern blot. In these cells, the target miR-26a in 2′-O-methyl RNA-treated osteoblasts were markedly reduced in a sequence-specific manner, because a mutated negative control 2′-O-mehyl RNA did not cause a reduction in miR-26a expression (Fig. 2B).

Expression of miR-26a target SMAD1 during osteoblastic differentiation

One of the targets predicted for miR-26a is the SMAD1 transcription factor protein, the downstream effector of the BMP involved in osteoblastic differentiation.(27) Whether miR-26a could affect SMAD1 cellular protein levels by modulating osteoblastic differentiation was examined firstly by Western blot of cellular extracts from differentiating osteoblasts transfected with miR-26a antisense 2′-O-methyl RNA. SMAD1 protein was upregulated by the inhibition of miR-26a, because absolute levels of SMAD1 protein were higher in cells treated with the 26a antisense 2′-O-methyl RNA than in control cells (Fig. 3A). Twofold upregulation of the SMAD1 protein was estimated by semiquantitative analysis of the Western blots, similarly to previously published data on miR-181 and its target, Hox-A11.(12) The levels of Smad1 mRNA were not affected by the inhibition of miR-26a with antisense 2′-O-methyl RNA (Fig. 3B), even though the protein levels increased, supporting a post-transcriptional mechanism of action. These data are consistent with the hypothesis that SMAD1 is a direct target of miR-26a in osteoblasts differentiating from hADSCs.

Figure Figure 3.

Predicted miR-26a target SMAD1 is upregulated in miR-26a 2′-O-methyl antisense RNA-treated osteoblasts. (A) Western blotting for SMAD1 protein in differentiating osteoblasts treated with 250 nM 2′-O-methyl antisense RNA targeting miR-26a (WT) and with control 2′-O-methyl miR26a mutated antisense RNA (mut) at 30 days after induction of differentiation. The blot was reprobed for β-actin as a loading control. (B) RNase mapping analysis of expression of endogenous Smad1 mRNA in differentiating osteoblasts treated with 250 nM 2′-O-methyl antisense RNA targeting miR-26a (WT) and control (mut) at 30 days after induction of differentiation.

Inhibition of endogenous miR-26a promotes osteogenic differentiation of hADSCs

It was reported that the antisense inhibition of miR-181 dramatically affected expression of markers of myoblast differentiation C2C12 cells.(12) Moreover, inhibition of miR-143 regulated adipogenesis.(17) Because the Smad1 gene transfection promoted the commitment of murine mesenchymal progenitors to the osteoblastic lineage,(29) we analyzed mRNA expression of osteogenic differentiation markers in hADSCs.

Markers for osteogenic differentiation were assessed by comparative real-time qRT-PCR (Fig. 4). Significant upregulation of the ALP, COLAI, OPN, and OCN genes were shown in miR-26a antisense RNA-treated cells compared with cells treated with mutated antisense RNA (Fig. 4). The induction of genes associated with osteoblastic differentiation was observed in hADSCs where the SMAD1 protein expression was upregulated by RNA antisense inhibition of miR-26a binding. The upregulation affected specifically the genes associated with osteoblastic differentiation as shown by the unchanged levels of runx2/cbfa1, a transcription factor that binds to the promoters of several osteogenic genes. The expression of this gene was observed at all time-points in osteo-induced hADSCs and human mesenchymal stem cells (hMSCs).(18–20) Furthermore, runx2/cbfa1 gene expression was not specific to osteo-induced cells, because basal expression was observed in noninduced hADSCs and hMSCs.(18–20) Interestingly, OPN gene expression (an intermediate marker of differentiation) is upregulated in miR-26a antisense RNA-treated cells compared with the cells treated with mutated antisense RNA (Fig. 4).

Figure Figure 4.

SMAD1 upregulation promotes hADSC osteogenesis. Comparative qRT-PCR analysis of expression of osteoblastic markers in differentiating osteoblasts treated with 250 nM 2′-O-methyl antisense RNA targeting miR-26a (WT) and control (mut) at 30 days after induction of differentiation. qRT-PCR analysis revealed a statistically significant increase in I collagen (ColI), ALP (Alp), osteopontin (Opn), and osteocalcin (Ocn) when hADSCs were treated with complementary miR-26a RNA. **Enhancement was not seen in runx2/cbfa1 (Runx2). Bars indicate SD. *p < 0.05 (paired Student's t-test, n = 3).

DISCUSSION

Adipose tissue, like bone marrow, is a mesodermally derived organ that contains stromal cellular components with different phenotypes, such as endothelial, smooth muscle, fibroblastic, and stem cells. Under conditions supportive for stromal cells growth, a homogeneous cell population emerges in culture.(18,19) This population, named hADSCs, exhibited reproducible growth and differentiation kinetics in culture, when driven with osteogenic, adipogenic, myogenic, or chondrogenic lineage-specific culture media.(18,19) Several studies suggested a high osteogenic potential for hADSCs,(18–22) but the molecular mechanisms that underlie hADSCs differentiation toward the osteoblastic phenotype are still unknown. Cell differentiation involves complex pathways regulated at both transcriptional and post-transcriptional levels. Recently, it has been shown that miRNAs, small noncoding RNAs, have an influence on the complexity of the stemness state, through negative regulation of gene expression at the post-transcriptional level.(30) In this study, we investigated the possibility that miR-26a could be involved in the osteogenic differentiation of hADSCs by modulating the expression of its predicted target, the SMAD1 protein.(27) Identification of genes targeted by miRNAs is widely believed to be an important step toward understanding the role of miRNAs in gene regulatory networks. As part of the effort to understand interactions between miRNAs and their targets, computational algorithms have been developed based on observed rules for features such as the degree of hybridization between the two RNA molecules. These in silico approaches provide important tools for miRNA target detection, and together with experimental validation, help to reveal regulated targets of miRNAs.

Indeed, TargetScan (http://genes.mit.edu/targetscan) software predicted two target sites for miR-26a in the orthologous Smad1 3′ UTR sequences from humans (Hs), rats (Rn), and Fugu (Fr),(27) as confirmed by miRanda prediction software.(31) TargetScan combines thermodynamics-based modeling of RNA:RNA duplex interactions with comparative sequence analysis to predict miRNA targets conserved across multiple genomes. Given an miRNA that is conserved in multiple organisms and a set of orthologous 3′-UTR sequences from these organisms, TargetScan (1) searches the UTRs in the first organism for segments of perfect Watson-Crick complementarity to bases 2–8 of the miRNA (numbered from the 5′ end)—we refer to this 7 nt segment of the miRNA as the “miRNA seed” and UTR heptamers with perfect Watson-Crick complementarity to the seed as “seed matches”; (2) extends each seed match with additional base pairs to the miRNA as far as possible in each direction, allowing G:U pairs, but stopping at mismatches; (3) optimizes basepairing of the remaining 3′ portion of the miRNA to the 35 bases of the UTR immediately 5′ of each seed match using the RNAfold program, thus extending each seed match to a longer “target site”; and (4) assigns a folding free energy G to each such miRNA:target site interaction (ignoring initiation free energy). Lewis et al.(27) cloned a 106 nt fragment of the Smad1 3′ UTR segment that included miR-26a target sites within the firefly luciferase ORF. The luciferase activity was measured in HeLa cells transfected with pGL3/26a reporter plasmid and with an analogous reporter with point substitutions disrupting the target sites.(27) These point substitutions, by disrupting the pairing to the miRNA seed, significantly enhanced the expression of the control reporter gene, as expected if the endogenous miR-26a in the HeLa cells was specifying the repression of the reporter gene expression by pairing to the predicted target site.(27)

The role of the BMP signaling mediator SMAD1 protein in osteogenic differentiation was studied in murine parental mesenchymal progenitors—the pluripotential mesenchymal cell line C3H10T that represents a relatively early stage of mesenchymal cell determination with the ability to differentiate into osteoblasts, chondrocytes, myoblasts, and adipocytes.(29) Expression of recombinant SMAD1 protein in these cells enhanced the Smad1 mRNA levels 20- to 50-fold in relation to endogenous Smad1 mRNA levels, thus resulting in enhanced osteogenic development.(29) To test the hypothesis that SMAD1 is a direct target of miR-26a during osteogenic differentiation in hADSCs, we transfected the reporter plasmid with a tandem repeat of the Smad1-predicted target gene sequences inserted within the firefly luciferase gene(27) into hADCSs before osteoblastic induction and induced differentiation. Insertion of the wildtype tandem repeat of Smad1-predicted target sequences rendered the reporter sensitive to endogenous miR-26a, whereas mutation of the target sequences abolished this effect, thus indicating that SMAD1 is a direct target of the endogenous miR-26a during hADSCs differentiation.

These observations were reinforced by both SMAD1 protein and miR-26a expression profiles. miR-26a was barely expressed in undifferentiated hADSCs, but under differentiation, its expression level increased, reaching the maximum when osteogenic maturation occurred. Conversely, SMAD1 was highly expressed in cells with low miR-26a levels, whereas SMAD1 downregulation coincided with miRNA induction during terminal differentiation.

To determine whether the interaction between miR-26 and Smad1 3′-UTR was involved in osteogenic differentiation of hADSCs through the modulation of SMAD1 expression, hADSCs were treated with 2′-O-methyl antisense oligonucleotides against the miRNA 26a(28) at the time of onset of miRNA expression. Inhibition of miR-26a, as shown by Northern blot, resulted in upregulation of SMAD1 protein without affecting the level of Smad-1 mRNA, thus supporting the postulated post-transcriptional mechanism for miR-RNA-regulated gene expression.(32,33) The upregulation of SMAD1 protein induced an increased osteogenic activity on cells treated with 2′-O-methyl RNA complementary to endogenous miR-26a, as stressed by the enhanced expression of the Ocn bone-specific gene, the highly specific late marker in the osteoblastic differentiation process.

SMAD1 interacts with the transcription repressor HOXC8,(24) disrupting HOX DNA binding, and thus resulting in activation of OPN gene transcription. HOX proteins act as downstream DNA-binding proteins in the BMP signaling pathway, because their transcriptional activities are regulated by SMADs through physical interactions.(24) A domain of SMAD1, termed SMAD1C, interacts with HOXC8.(34) Ectopic expression of Smad1C was able to induce osteoblast differentiation and bone formation in vitro.(34) We showed that OPN gene expression (an intermediate marker of differentiation) is upregulated in miR-26a antisense RNA-treated cells compared with the cells treated with mutated antisense RNA, whereas both HOXC8 mRNA and protein levels remained unchanged (data not shown).

These data suggested a negative effect or control of miR-26a over differentiation. miR-26a could participate in the differentiation process of hADSCs by diminishing the availability of the active SMAD1 transcription factor. Gain of function caused by 2′-O-methyl antisense oligonucleotides against miRNA 26a confirmed this interpretation. Typically, thousands of mammalian genes contain a sequence complementary to the 7-nt seed of a given miRNA, but only 1/10 of these target sites are conserved across species: large-scale microarray analyses that compare changes in gene expression in response to the presence of specific miRNAs show that many targets are subject to downregulation at the mRNA level.(28) Thus, we point out that the miR-26a is necessary for a negative control of hADSC terminal differentiation, although more miRNAs should be necessary for early and late bone formation. Transcription factors such as OCT4, SOX2, and NANOG,(34) and other regulators, such as NOTCH, IDS, and PTEN,(30) important for stem cell self-renewal, are potential targets of miRNAs. These targets should be experimentally confirmed in futures studies. In a recent study, Cui et al.(35) systematically analyzed the relationship between transcription factors (TFs) and miRNAs in gene regulation. They found that the genes with more TF-binding sites have a higher probability of being targeted by miRNAs and have more miRNA-binding sites on average. This observation indicates that the genes with higher cis-regulation complexity are more coordinately regulated by TFs at the transcriptional level and by miRNAs at the post-transcriptional level. Altogether, these findings open new avenues in the understanding of the mechanism that control the coordinated regulation of gene expression. Our demonstration that miR26a-modulated SMAD1 transcription factor could provide mechanistic insights both into the molecular processes of stem cell differentatiation and into the network involving transcription factors and miRNAs in the differentiation processes, making this cellular model appealing as a useful target for cellular therapy, tissue engineering, and gene transfer through miRNAs.

Acknowledgements

This work was supported by a grant of the Fondazione Ente Cassa di Risparmio di Firenze to MLB.

Ancillary