OPEN
Citation: Bone Research (2016) 4, 16022; doi:10.1038/boneres.2016.22
www.nature.com/boneres
ARTICLE
miR-23a/b regulates the balance between osteoblast and
adipocyte differentiation in bone marrow mesenchymal
stem cells
Qi Guo1, Yusi Chen2, Lijuan Guo1, Tiejian Jiang1 and Zhangyuan Lin3
Age-related osteoporosis is associated with the reduced capacity of bone marrow mesenchymal stem cells
(BMSCs) to differentiate into osteoblasts instead of adipocytes. However, the molecular mechanisms that
decide the fate of BMSCs remain unclear. In our study, microRNA-23a, and microRNA-23b (miR-23a/b) were
found to be markedly downregulated in BMSCs of aged mice and humans. The overexpression of miR-23a/b
in BMSCs promoted osteogenic differentiation, whereas the inhibition of miR-23a/b increased adipogenic
differentiation. Transmembrane protein 64 (Tmem64), which has expression levels inversely related to those
of miR-23a/b in aged and young mice, was identified as a major target of miR-23a/b during BMSC
differentiation. In conclusion, our study suggests that miR-23a/b has a critical role in the regulation of
mesenchymal lineage differentiation through the suppression of Tmem64.
Bone Research (2016) 4, 16022; doi:10.1038/boneres.2016.22; published online: 24 August 2016
INTRODUCTION
Osteoblast-mediated bone formation, a key determinant
of bone mass, maintains bone homeostasis along with
osteoclast-mediated bone resorption. Bone marrow
mesenchymal stem cells (BMSCs), the progenitor cells for
osteoblasts, adipocytes, and chondrocytes, show a
decrease with age in their potential to differentiate into
osteoblasts rather than adipocytes, which results in agerelated bone loss and fat accumulation in the bone
marrow.1–4 Age-related osteoporosis, coupled with an
increase in bone marrow fat, is attributable to an
imbalance
between
osteoblast
and
adipocyte
differentiation.3 However, the molecular mechanisms that
regulate the fate of BMSCs remain unclear.
MicroRNAs (miRNAs) are small (22–24 nucleotides),
single-stranded noncoding RNAs that are involved in
diverse biological processes. Several miRNAs have been
characterized to participate in osteogenesis or adipogenesis. Li et al.5 found that miR-188 was highly expressed in
aged mice and humans and that it regulates the
1
bifurcation of differentiating BMSCs into osteoblasts and
adipocytes.5 miR-204 and miR-211 have been reported to
act as vital negative regulators of Runx2 to promote
adipogenesis and suppress osteogenesis in BMSCs.6
miR-205 has been shown to exert negative effects on
the osteogenic differentiation of BMSCs,7 whereas miR-21
promoted the osteogenic differentiation of MSCs via the
PI3K/β-catenin pathway.8 Despite these findings, the functions of miRNAs in the differentiation and lineage commitment of BMSCs requires further investigation.
In this study, we identified two novel miRNAs, miR-23a,
and miR-23b, that are downregulated in the BMSCs of
aged vs young mice and humans. We also investigated
the roles of these two miRNAs in the differentiation of
BMSCs in vitro. We demonstrate that miR-23a/b strikingly
enhanced osteoblast and attenuated adipocyte differentiation from BMSCs by targeting Tmem64. Consequently,
our study suggests that miR-23a/b acts as an age-related
‘switch’ to divert BMSCs from being adipogenic to
osteogenic.
Department of Endocrinology, The Xiangya Hospital of Central South University, Changsha 410008, China; 2Department of Gerontology, The
Xiangya Hospital of Central South University, Changsha 410008, China and 3Department of Orthopedics, The Xiangya Hospital of Central South
University, Changsha 410008, China
Correspondence: Zhangyuan Lin (zhangyuanlin1605@163.com)
Received: 15 January 2016; Revised: 29 May 2016; Accepted: 1 June 2016
Osteoblast and adipocyte differentiation
Q Guo et al
2
MATERIALS AND METHODS
BMSC culture and transfection
Following the isolation of mouse and human BMSCs, the
cells were cultured to the third-passage as previously
described.5 The miRNA blocker antagomiR-23a/b, analog
agomiR-23a/b and their negative control were synthesized
by Ribobio (Guangzhou, China). miRNA blockers and
analogs were prepared and directly mixed with cells
according to the manufacturer’s instructions.
Clinical samples
Human bone marrow samples were collected from 32 aged
patients (17 males and 15 female) who were 470 years old
and from 29 young patients (15 males and 14 female) aged
from 20 to 40 years old who underwent routine therapeutic
surgery at the Orthopedic Surgery Department of the
Xiangya Hospital of Central South University.
Adipogenic differentiation assay and Oil Red staining
To induce adipogenic differentiation, BMSCs were grown in
6-well plates at 2.5 × 106 cells per well in adipogenic-inducing
medium α-MEM (Gibco, Waltham, MA, USA) containing 10%
FBS (Gibco), 5 μg·mL − 1 insulin, 0.5 mmol·L − 1 3-isobutyl-1methylxanthine, and 1 μmol·L − 1 dexamethasone for 14 days.
Culture medium was changed every 2 or 3 days. After
14 days of adipogenic induction, Oil Red O staining was performed to detect the lipid droplets as previously described.9
Osteogenic differentiation and Alizarin Red staining
To induce osteogenic differentiation, BMSCs were grown in
24-well plates at 5 × 105 cells per well in osteogenic-inducing
medium (300 ng·mL − 1 BMP-2, 50 μg·mL − 1 ascorbic acid,
and 5 mmol·L − 1 β-glycerolphosphate) for 48 h. An alkaline
phosphatase (ALP) activity assay and the assessment of
secreted osteocalcin levels were performed using an
enzymatic colorimetric ALP kit (Roche Diagnostics, Minneapolis, MN, USA) and a specific immunoassay kit (DiaSorin,
Stillwater, MN, USA), respectively, as previously described.5
To induce osteoblastic mineralization, BMSCs were grown in
mineralization-inducing medium as described above in 6well plates at 2.5 × 106 cells per well for 21 days. Alizarin Red
staining was performed as previously described.5 Spectrophotometry was used to quantify Alizarin Red S at 540 nm.
qRT-PCR analysis
Quantitative reverse transcription PCR (qRT-PCR) was
performed using a Roche Molecular Light Cycler (Basel,
Basel-Stadt, Switzerland) as previously described.10–12 We
used TRIzol reagent (Invitrogen, Carlsbad, CA, USA) to
isolate total RNA from cultured cells or tissues, and 1.0 μg
total RNA and SuperScript II (Invitrogen) were used to
perform reverse transcription. The amplification reactions,
Bone Research (2016) 16022
Table 1.
mRNAs
Primer sequences used for qRT-PCR detection for
Gene
Acc. No
Primer sequence (5′ to 3′)
Pparg (mouse)
NM_001127330
Fabp4 (mouse)
NM_024406
Runx2 (mouse)
NM_001146038
Osterix (mouse)
NM_130458
Tmem64 (mouse)
NM_181401
β-actin (mouse)
NM_007393
F: GACCACTCGCATTCCTTT
R: CCACAGACTCGGCACTCA
F: AAATCACCGCAGACGACA
R: CACATTCCACCACCAGCT
F: ACTTCCTGTGCTCCGTGCTG
R: TCGTTGAACCTGGCTACTTGG
F: ACCAGGTCCAGGCAACAC
R: GCAAAGTCAGATGGGTAAGTAG
F: AGGAAGCGGCCTGAAGGT
R: GAAGGAAGAGCCACTGGGAT
F: CTGTCCCTGTATGCCTCTG
R: TGATGTCACGCACGATTT
Acc. No, Genbank accession numbers; F, forward primer; mRNA, messenger RNA;
R, reverse primer; qRT-PCR, quantitative reverse transcription PCR.
which contained amplification primers and SYBR Green
PCR Master Mix (Perkin-Elmer Corporation, Applied Biosystems, Foster City, CA, USA), were set up in 25-μL reaction
volumes, and a 1 μL of complementary DNA was added to
each amplification reaction. The nucleotide sequences of
primers for β-actin, Tmem64, Pparg, Fabp4, Runx2, Osterix,
miR-23a/b, and U6 are listed in Tables 1 and 2.
Western blot
Western blotting was performed as previously described.13–15
Total cell lysates were separated by SDS-polyacrylamide
gel electrophoresis and then transferred to PVDF
membranes (Millipore, Bedford, MA, USA). Tmem64 levels
were detected using an anti-Tmem64 antibody (sc-87460;
Santa Cruz, Dallas, TX, USA) and were normalized to β-actin
(ab3280; Abcam, Cambridge, MA, USA). The membranes
were incubated with appropriate HRP-conjugated secondary antibodies, and the blots were visualized using an ECL kit
(Santa Cruz) and exposed to X-ray films.
mRNA 3′-UTR cloning and luciferase reporter assay
A segment of the mouse Tmem64 3′-untranslated region
(UTR) containing the predicted miR-23a/b binding site was
amplified by PCR using the forward primer 5′-CTAGAGGAA
TTCTGAAATGTGAAATTGTCTCAAGGCCGG - 3′ and the
reverse primer 5′-CCTTGAGACAATTTCACATTTCAGAATTCC
T-3′. The PCR products were purified and inserted into the
XbaI–FseI site downstream of the stop codon in the pGL3
control luciferase reporter vector (Promega, Madison, WI,
USA), resulting in WT-pGL3-Tmem64. A QuikChange sitedirected mutagenesis kit (Stratagene, La Jolla, CA, USA)
was used to insert mutations into the miR-23a/b seed
region to obtain MUT-pGL3-Tmem64. The primers for
Tmem64 3′-UTR mutagenesis were 5′-CTAGAGGAATTCT
GAACACTGAAATTGTCTCAAGGCCGG-3′ (forward) and
5′-CCTTGAGACAATTTCAGTGTTCAGAATTCCT-3′ (reverse).
Osteoblast and adipocyte differentiation
Q Guo et al
3
Table 2. Primer sequences used for qRT-PCR detection for
microRNA
microRNA
Primer
Primer sequence (5′ to 3′)
miR-23a
RT primer
miR-23b
Forward
Reverese
RT primer
Forward
Reverese
GTCGTATCCAGTGCAGGGTCCGAGG
TATTCGCACTGGATACGAC GGTAATC
GAGTGATCACATTGCCAGG
GCAGGGTCCGAGGTATTC
GAACGCTTCACGAATTTGCGTGTCAT
CTCGCTTCGGCAGCACA
AACGCTTCACGAATTTGCGT
U6
qRT-PCR, quantitative reverse transcription PCR.
BMSCs were transfected with wild type (WT) or MUTpGL3-Tmem64 constructs (200 ng) and either agomiR-23a/b
or agomiR-NC for 48 h using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. The
modified pGL3 control vector without a 3′-UTR insert was
used as a positive control. BMSCs treated with Lipofectamine only served as negative controls. The luminescence
signal was quantified by a dual luciferase reporter assay
system (Promega) using a luminometer (Glomax, Promega). Values from the firefly luciferase assay were
normalized to the Renilla luciferase assay value from the
transfected phRL-null vector (Promega).
The 3′-UTR of mouse Tmem64 was amplified by RT-PCR
from total RNA extracted from BMSCs and using primers
designed based on the mouse Tmem64 complementary
DNA sequence. The forward primer was 5′-GCTCTAGATTG
TTGAGAGCCTAGCGTGC-3′, and the reverse primer was
5′-GCGGTACCCAGCTCAGACGTACCAGGTC-3′. A QuikChange site-directed mutagenesis kit (Stratagene) was
used to generate the two mutations in Tmem64 (mutant
Tmem64) by PCR using the WT Tmem64 construct as the
template. The introduced mutations did not result in aminoacid changes in the Tmem64 protein. Finally, WT and mutant
Tmem64 were cloned into the pCDNA3.1 (+) expression
vector (Invitrogen) at XbaI/KpnI sites. Then, we co-transfected the WT or the mutant Tmem64 3′-UTR construct with
agomiR-23a/b into mouse BMSCs as described above.
Statistical analyses
Statistics were analyzed using SPSS 16.0 (Polar Engineering
and Consulting, http://www.winwrap.com/). Data are
presented as the mean ± s.d. A Student’s t-test was used
for comparing the differences between two groups.
Comparisons of multiple groups were made using oneway ANOVA. All experiments were repeated at least three
times, and representative experiments are shown. Po0.05
was considered statistically significant.
Study approval
All animal care protocols and experiments were
reviewed and approved by the Animal Care and Use
Committee of the Laboratory Animal Research Center at
Xiangya Medical School of Central South University. All
female WT C57BL/6 mice of different ages used in
experiments were housed under specific pathogen-free
conditions (22 °C, 12-h light/12-h dark cycles, and
50%–55% humidity) with free access to food pellets and
tap water.
The clinical study was approved by the Ethics Committee
of Central South University, and informed consent was
obtained from each participant before the collection of
clinical samples.
RESULTS
miR-23a/b is markedly downregulated in BMSCs during
the aging process
Our previous findings5 revealed that BMSCs tended to
differentiate into adipocytes rather than osteoblasts over
the course of aging. To investigate the differences in the
miRNA expression profiles of BMSCs from young and aged
mice, we had previously performed microarray analysis to
identify dysregulated miRNAs.5 We identified miR-23a/b to
be the most significantly downregulated miRNAs in aged vs
young mice, and in this study, we chose to study miR-23a/b
further and investigated its function in the regulation of
BMSC differentiation. We confirmed the decreased level of
miR-23a/b qRT-PCR (Figure 1a), and we then measured the
levels of miR-23a/b in human BMSCs that were isolated
from the bone marrow cells of patients aged either 20 to 40
or 470 years. Consistently, the expression of miR-23a/b was
notably decreased in elderly samples compared with that
in young samples (Figure 1b and c). This result suggests that
miR-23a/b is involved in age-related effects on BMSCs in
mouse and human.
miR-23a/b inhibits the adipogenic differentiation of BMSCs
miR-23a/b expression was revealed by qRT-PCR analysis
to gradually decrease during adipogenic differentiation
in the BMSCs of 6- to 8-week-old mice (Figure 2a). To
overexpress or silence miR-23a/b in BMSCs for functional
investigation, we transfected BMSCs with agomiR-23a/b,
antagomiR-23a/b or their negative control and subsequently induced adipogenic differentiation (Figure 2b). The
overexpression of miR-23a/b attenuated lipid droplet
formation in adipogenesis-induced BMSCs (Figure 2c and
d). Likewise, the messenger RNA (mRNA) levels of two
important adipocyte markers, peroxisome proliferatoractivated receptor-g (Pparg) and fatty acid binding
protein 4 (Fabp4), were also reduced compared with
controls (Figure 2e and f). Conversely, silencing miR-23a/b
promoted lipid droplet formation (Figure 2c and d) and
increased the levels of Pparg and Fabp4 mRNA during
the adipogenic differentiation of BMSCs (Figure 2e and f).
Bone Research (2016) 16022
Osteoblast and adipocyte differentiation
Q Guo et al
4
Figure 1. miR-23a/b is gradually downregulated in BMSCs throughout the aging process. (a) qRT-PCR was used to analyze the relative levels of
miR-23a/b in BMSCs isolated from C57BL/6 mice of different ages. n = 5 per group. (b and c) Comparison of miR-23a/b levels in young and old
human BMSCs as determined by qRT-PCR of human male (b) and female samples (c). Male: ny = 15, no = 17. Female: ny = 14, no = 15. Data are shown
as the mean ± s.d. **P o0.01 (ANOVA or Student’s t-test). ANOVA, analysis of variance; BMSCs, bone marrow mesenchymal stem cells; qRT-PCR,
quantitative reverse transcription PCR.
Taken together, these observations suggest that miR-23a/b
negatively regulates the adipogenic differentiation of
BMSCs.
miR-23a/b promotes the osteogenic differentiation of
BMSCs
Our results showed that miR-23a/b expression gradually
increased in BMSCs from 6- to 8-week-old mice during
the process of osteoblastic differentiation (Figure 3a).
We next determined the role of miR-23a/b during the
osteogenic differentiation of BMSCs by overexpressing or
silencing miR-23a/b in BMSCs. After being transfected
with agomiR-23a/b, antagomiR-23a/b, or their negative
controls, BMSCs were cultured in osteogenic-inducing
medium. Alizarin Red staining indicated that the overexpression of miR-23a/b facilitated the osteogenic differentiation of BMSCs, whereas the silencing of miR-23a/b
inhibited osteogenic differentiation (Figure 3b and c).
In addition, ALP activity and osteocalcin secretion, both
markers of osteoblast differentiation, were evaluated in
agomiR-23a/b-transfected BMSCs and compared with
control BMSCs (Figure 3d). Furthermore, the mRNA levels of
two critical osteoblast transcription factors, Osterix, and
Runx2, were also increased following transfection with
agomiR-23a/b (Figure 3e). In contrast, the transfection of
antagomiR-23a/b attenuated ALP activity and osteocalcin
secretion, in addition to Osterix and Runx2 expression
(Figure 3d and e). Altogether, all of these data indicate
that miR-23a/b enhances the osteogenic differentiation
of BMSCs.
Bone Research (2016) 16022
miR-23a/b directly targets Tmem64
miRNAs have been shown to regulate the expression
of mRNAs by binding to coding sequences or the
3’-untranslated regions (3′-UTRs) of target genes.16 We
used Starbase v2.0 (http://starbase.sysu.edu.cn/contact.
php/; Li et al. Nucleic Acids Res. 2014 & Yang et al. Nucleic
Acids Res. 2011) to predict the possible target genes
of miR-23a/b, considering the predicted intersections
of miRanda, PicTar, and TargetScan and using medium
stringency. Among the 26 possible target genes predicted,
we chose Transmembrane protein 64 (Tmem64), which
had been reported to participate in the regulation of
mesenchymal lineage allocation,9 for further analysis.
Sequence analysis showed one miR-23a/b binding site in
the 3′-UTR of the Tmem64 gene (position 1069-1076;
Figure 4a). To clarify whether miR-23a/b could directly target
the Tmem64 gene, a luciferase reporter construct including
the putative binding site of the Tmem64 3′-UTR (WT-pGL3Tmem64) was generated, and three mutant nucleotides
were introduced into the predicted target sequences
(MUT-pGL3-Tmem64) and used as a control. We transfected
WT-pGL3-Tmem64 or MUT-pGL3-Tmem64 along with
agomiR-23a/b or agomiR-NC into BMSCs and assessed the
effects of miR-23a/b on luciferase translation by luciferase
enzyme activity. Transfection with agomiR-23a/b was able
to repress the luciferase activity of the Tmem64 3′-UTR
reporter gene, whereas MUT-pGL3-Tmem64 prevented this
inhibition (Figure 4b). This finding confirmed that miR-23a/b
can specifically bind to the predicted 3′-UTR of Tmem64.
To further determine whether this conserved site was
the actual binding region, we transfected BMSCs with
Osteoblast and adipocyte differentiation
Q Guo et al
5
Figure 2. miR-23a/b inhibits the adipogenic differentiation of BMSCs. (a) qRT-PCR analysis of the relative levels of miR-23a/b in BMSCs induced
to differentiate into adipocytes for 14 days. (b) The relative levels of miR-23a/b in BMSCs transfected with 10 μmol·L − 1 agomiR-23a/b,
antagomiR-23a/b or their NC were analyzed by qRT-PCR. (c and d) Representative images of Oil Red staining of lipid droplets (c), and the
quantitative analysis of the number of Oil Red spots (d) in BMSCs induced to differentiate into adipocytes for 14 days. (e and f) The relative mRNA
expression levels of adipogenic markers, Pparg (e) and Fabp4 (f), were measured by qRT-PCR in BMSCs induced to differentiate into adipocytes for
48 h. Scale bars: 120 μm. n = 5 per group. Data are shown as the mean ± s.d. *P o0.05, **Po 0.01 (ANOVA). ANOVA, analysis of variance; BMSCs,
bone marrow mesenchymal stem cells; NC, negative controls; qRT-PCR, quantitative reverse transcription PCR.
Bone Research (2016) 16022
Osteoblast and adipocyte differentiation
Q Guo et al
6
Figure 3. miR-23a/b promotes the osteogenic differentiation of BMSCs. (a) qRT-PCR analysis showed the relative levels of miR-23a/b in BMSCs
induced to differentiate into osteoblasts for 14 days. (b and c) Representative images of Alizarin Red staining (b) and the quantitative analysis of
matrix mineralization (c) in BMSCs induced to differentiate into osteoblasts for 21 days after transfection. (d) ALP activity and osteocalcin secretion
were measured in BMSCs induced to generate osteoblasts for 48 h. (e) qRT-PCR was used to analyze the relative expression levels of Osterix and
Runx2 in BMSCs induced to differentiate into osteoblasts for 48 h. Scale bars: 100 μm. n = 5 per group. Data are shown as the mean ± s.d. *P o0.05,
**P o0.01 (ANOVA). ANOVA, analysis of variance; BMSCs, bone marrow mesenchymal stem cells; qRT-PCR, quantitative reverse
transcription PCR.
agomiR-23a/b or antagomiR-23a/b and detected the
mRNA and protein levels of Tmem64. The overexpression
of miR-23a/b decreased endogenous levels of Tmem64
protein, whereas the inhibition of miR-23a/b elevated
Bone Research (2016) 16022
Tmem64 protein levels (Figure 4c); however, Tmem64
mRNA levels remained stable (Figure 4d). We also measured the levels of Tmem64 mRNA and protein in BMSCs
from of 3- and 18-month-old mice, and we found
Osteoblast and adipocyte differentiation
Q Guo et al
7
Figure 4. miR-23a/b directly targets Tmem64. (a) Schematic representation of the predicted miR-23a/b target site in the 3′-UTR of mouse Tmem64.
The alignment of miR-23a/b with WT and MUT 3′-UTR region is shown by complementary pairing, and three mutated nucleotides are underlined.
(b) BMSCs were co-transfected with the luciferase reporter carrying WT-pGL3-Tmem64 or MUT-pGL3-Tmem64 along with agomiR-23a/b or
agomiR-NC. The effects of miR-23a/b on the luciferase reporter constructs were determined 48 h after transfection. The firefly luciferase values were
normalized to Renilla luciferase; n = 5. (c) After BMSCs were transfected with agomiR-23a/b or antagomiR-23a/b, the relative levels of Tmem64
protein expression were determined by western blot; β-actin was used as loading control; n = 5. (d) The relative levels of Tmem64 mRNA were
determined using qRT-PCR and normalized to β-actin; n = 5. (e) Tmem64 protein levels in BMSCs from 3- and 18-month-old mice were measured by
western blot and expressed as the densitometry of Tmem64/β-actin. Tmem64 mRNA levels were determined by qRT-PCR and are shown as the
fold-induction relative to β-actin; n = 3. (f) The increase in ALP activity induced by agomiR-23a/b was blocked by the transfection of MUT Tmem64
3′-UTR into osteogenic-induced-BMSCs. *P o0.05 vs. agomiR-23a/b+WT Tmem64 3′-UTR. n = 3. Data are shown as the mean ± s.d. *P o0.05
(Student’s t-test or ANOVA). ALP, alkaline phosphatase; ANOVA, analysis of variance; BMSCs, bone marrow mesenchymal stem cells; CDS, coding
sequence; qRT-PCR, quantitative reverse transcription PCR; WT, wild type; 3′-UTR, 3′-untranslated region.
Bone Research (2016) 16022
Osteoblast and adipocyte differentiation
Q Guo et al
8
increased levels of Tmem64 protein in aged mice;
however, the increase of Tmem64 mRNA in aged mice
was not statistically significant (Figure 4e).
We next co-transfected the WT or mutant Tmem64
3′-UTR construct along with agomiR-23a/b or miR-NC
into mouse BMSCs cultured in osteogenic-inducing
medium; miR-NC and agomiR-23a/b were used as
negative and positive controls, respectively. We observed
an increase in ALP activity in those cells transfected
with agomiR-23a/b and agomiR-23a/b+WT Tmem64
3′-UTR. The increase in ALP activity induced by agomiR-23a/b was blocked by the mutant Tmem64 3′-UTR
construct (Figure 4f). These results together show that
Tmem64 shows increased expression with age and is the
major target of miR-23a/b during BMSC differentiation
and that miR-23a/b affects Tmem64 expression at the
post-transcriptional level.
DISCUSSION
The maintenance of bone homeostasis primarily depends
on osteoblast-mediated bone formation and osteoclastmediated bone resorption. During the aging process,
BMSCs show a gradual decline in their capacity to
differentiate into osteoblasts vs adipocytes, which results
in progressive bone loss and fat accumulation and leads to
age-related osteoporosis.3,17–22 However, the mechanism
behind this switch in differentiation potential requires further
investigation.
In the present study, we observed that miR-23a/b is
prominently downregulated in BMSCs of aged mice and
humans. The overexpression of miR-23a/b promoted the
osteogenic differentiation of BMSCs, whereas the inhibition
of miR-23a/b intensified adipogenic differentiation from
BMSCs in vitro. Furthermore, we determined that miR-23a/b
regulated BMSCs differentiation by directly targeting
Teme64. These results suggest that miR-23a/b has a critical
role in BMSC differentiation.
Previously, we performed miRNA microarray analysis to
determine that miR-188 becomes remarkably elevated in
BMSCs with age, and we identified its vital function in
determining the differentiation potential of BMSCs.
However, the regulation of BMSC differentiation involves
multiple miRNAs. To further investigate the age-related
switch in differentiation potential of BMSCs, we observed
and identified two important downregulated miRNAs,
miR-23a and miR-23b, in the BMSCs of aged mice. In
addition, we confirmed that the level of miR-23a/b
expression in human BMSCs also showed significant agerelated differences. Taken together, these findings indicate
that miR-23a/b has a crucial effect on the aging process of
BMSCs in both mouse and human.
Bone Research (2016) 16022
miR-23a and miR-23b belong to the same family and
have strong similarities in their nucleotide sequences, and
importantly, they function as synergistic regulators of BMSC
functions. Previously, it had been reported that miR-23a/b
reinforces the expression of glutaminase in mitochondria
and participates in glutamine metabolism.23 Several
studies have shown that the activation of miR-23a by
NFATc3 regulates cardiac hypertrophy24 and that miR-23b
inhibits autoimmune inflammation.25 However, there had
been no studies of the action of miR-23a/b on the
regulation of BMSC differentiation.
Our study confirmed that miR-23a/b has promoting
effects on the osteogenic differentiation of mouse BMSCs
in vitro. However, Hassan and colleagues have reported
that miR-23a had an inhibitory role in the maturation of
primary rat osteoblasts and mouse MC3T3-E1 cells through
the targeting of SATB2.26 These two results are not likely to
be contradictory because different cell types show
specificity and different mechanisms of action, which
could explain these inconsistent results. Moreover, a
microRNA can regulate the expression of multiple target
genes; therefore, the miR-23 target genes that are relevant
to osteoblast maturation and BMSC differentiation might
be different. In the present study, we demonstrated that
Tmem64 was the major target of miR-23a/b during mouse
BMSC differentiation.
miRNAs have been reported to downregulate gene
expression by inhibiting mRNA translation or reducing
mRNA stability through binding to sites in the CDS or 3′UTR of the target gene.27 Studies have indicated that Fas,28
Runx2,29 CXCL12,29 Has2,30 and Hes1 (ref. 31) are potential
target genes of miR-23a or miR-23b. In this study, we
demonstrated that Tmem64 was directly targeted by
miR-23a/b and was responsible for regulating BMSC
differentiation. Tmem64 has been found to positively
modulate osteoclast differentiation via RANKL-mediated
Ca2+ signaling pathway.32 Recently, it was shown that mice
in which the Tmem64 gene was silenced presented
increased osteoblast and decreased adipocyte differentiation from BMSCs. Conversely, the overexpression of Tmem64
accelerated adipogenesis and inhibited osteogenesis.
Tmem64 regulates the switch in the lineage commitment
of MSCs to adipogenesis rather than to osteogenesis by
suppressing β-catenin, the key Wnt signaling molecule.9
Our study revealed that miR-23a/b mediates BMSC
differentiation by post-transcriptionally repressing Tmem64
expression. The decline in miR-23a/b expression in BMSCs
with age results in an attenuation of the suppression of
Tmem64 and consequently the increased expression of
Tmem64 protein, which inhibits the Wnt/β-catenin signaling pathway. Consequently, the regulation of Tmem64
causes BMSCs to have a tendency towards favoring
differentiation into adipocytes rather than osteoblasts.
Osteoblast and adipocyte differentiation
Q Guo et al
9
A previous study5 revealed that the level of miR-188
expression is markedly higher in BMSCs from aged compared with young mice and humans, and the
BMSC-specific inhibition of miR-188 stimulated new bone
formation. In the present study, our results showed that
miR-23a/b levels are decreased in BMSCs from aged
compared with young mice and humans, and the
activation of miR-23a/b in BMSCs promoted osteogenic
differentiation. These findings suggest that the upregulation
of miR-23a/b in BMSCs could be a potential therapeutic
target for osteoporosis.
Competing interests
The authors declare no conflicts of interest.
References
1 Pittenger MF, Mackay AM, Beck SC et al. Multilineage potential of adult
human mesenchymal stem cells. Science 1999; 284: 143–147.
2 Sekiya I, Larson BL, Vuoristo JT et al. Adipogenic differentiation of
human adult stem cells from bone marrow stroma (MSCs). J Bone Miner
Res 2004; 19: 256–264.
3 Yeung DK, Griffith JF, Antonio GE et al. Osteoporosis is associated with
increased marrow fat content and decreased marrow fat unsaturation: a
proton MR spectroscopy study. J Magn Reson Imaging 2005; 22: 279–285.
4 Crane JL, Zhao L, Frye JS et al. IGF-1 signaling is essential for differentiation of mesenchymal stem cells for peak bone mass. Bone Res 2013;
1: 186–194.
5 Li CJ, Cheng P, Liang MK et al. MicroRNA-188 regulates age-related
switch between osteoblast and adipocyte differentiation. J Clin Invest
2015; 125: 1509–1522.
6 Huang J, Zhao L, Xing L et al. MicroRNA-204 regulates Runx2 protein
expression and mesenchymal progenitor cell differentiation. Stem Cells
2010; 28: 357–364.
7 Hu N, Feng C, Jiang Y et al. Regulative effect of Mir-205 on osteogenic
differentiation of bone mesenchymal stem cells (BMSCs): possible role of
SATB2/Runx2 and ERK/MAPK pathway. Int J Mol Sci 2015; 16:
10491–10506.
8 Meng YB, Li X, Li ZY et al. microRNA-21 promotes osteogenic differentiation of mesenchymal stem cells by the PI3K/beta-catenin pathway.
J Orthop Res 2015; 33: 957–964.
9 Jeong BC, Kim TS, Kim HS et al. Transmembrane protein 64 reciprocally
regulates osteoblast and adipocyte differentiation by modulating
Wnt/beta-catenin signaling. Bone 2015; 78: 165–173.
10 Li H, Xie H, Liu W et al. A novel microRNA targeting HDAC5 regulates
osteoblast differentiation in mice and contributes to primary osteoporosis in humans. J Clin Invest 2009; 119: 3666–3677.
11 Ye L, Fan Z, Yu B et al. Histone demethylases KDM4B and KDM6B
promotes osteogenic differentiation of human MSCs. Cell Stem Cell 2012;
11: 50–61.
12 Hu R, Liu W, Li H et al. A Runx2/miR-3960/miR-2861 regulatory
feedback loop during mouse osteoblast differentiation. J Biol Chem 2011;
286: 12328–12339.
13 Chen Q, Xu J, Li L et al. MicroRNA-23a/b and microRNA-27a/b suppress Apaf-1 protein and alleviate hypoxia-induced neuronal apoptosis.
Cell Death Dis 2014; 5: e1132.
14 Cheng P, Chen C, He HB et al. miR-148a regulates osteoclastogenesis by
targeting
V-maf
musculoaponeurotic
fibrosarcoma
oncogene
homolog B. J Bone Miner Res 2013; 28: 1180–1190.
15 Yang L, Cheng P, Chen C et al. miR-93/Sp7 function loop mediates
osteoblast mineralization. J Bone Miner Res 2012; 27: 1598–1606.
16 Tay Y, Zhang J, Thomson AM et al. MicroRNAs to Nanog, Oct4 and
Sox2 coding regions modulate embryonic stem cell differentiation.
Nature 2008; 455: 1124–1128.
17 Gimble JM, Zvonic S, Floyd ZE et al. Playing with bone and fat. J Cell
Biochem 2006; 98: 251–266.
18 Manolagas SC. Birth and death of bone cells: basic regulatory mechanisms and implications for the pathogenesis and treatment of osteoporosis. Endocr Rev 2000; 21: 115–137.
19 Moerman EJ, Teng K, Lipschitz DA et al. Aging activates adipogenic and
suppresses
osteogenic
programs
in
mesenchymal
marrow
stroma/stem cells: the role of PPAR-gamma2 transcription factor and
TGF-beta/BMP signaling pathways. Aging Cell 2004; 3: 379–389.
20 Shen W, Chen J, Gantz M et al. MRI-measured pelvic bone marrow
adipose tissue is inversely related to DXA-measured bone mineral in
younger and older adults. Eur J Clin Nutr 2012; 66: 983–988.
21 Kim VN, Han J, Siomi MC. Biogenesis of small RNAs in animals. Nat
Rev Mol Cell Biol 2009; 10: 126–139.
22 Bartel DP. MicroRNAs: genomics, biogenesis, mechanism, and function.
Cell 2004; 116: 281–297.
23 Gao P, Tchernyshyov I, Chang TC et al. c-Myc suppression of miR-23a/b
enhances mitochondrial glutaminase expression and glutamine
metabolism. Nature 2009; 458: 762–765.
24 Lin Z, Murtaza I, Wang K et al. miR-23a functions downstream of
NFATc3 to regulate cardiac hypertrophy. Proc Natl Acad Sci USA 2009;
106: 12103–12108.
25 Zhu S, Pan W, Song X et al. The microRNA miR-23b suppresses
IL-17-associated autoimmune inflammation by targeting TAB2, TAB3
and IKK-alpha. Nat Med 2012; 18: 1077–1086.
26 Hassan MQ, Gordon JA, Beloti MM et al. A network connecting Runx2,
SATB2, and the miR-23a ~ 27a ~ 24-2 cluster regulates the osteoblast
differentiation program. Proc Natl Acad Sci USA 2010; 107: 19879–19884.
27 Carthew RW, Sontheimer EJ. Origins and mechanisms of miRNAs
and siRNAs. Cell 2009; 136: 642–655.
28 Li B, Sun M, Gao F et al. Up-regulated expression of miR-23a/b targeted
the pro-apoptotic Fas in radiation-induced thymic lymphoma. Cell
Physiol Biochem 2013; 32: 1729–1740.
29 He Y, Meng C, Shao Z et al. MiR-23a functions as a tumor suppressor in
osteosarcoma. Cell Physiol Biochem 2014; 34: 1485–1496.
30 Lagendijk AK, Goumans MJ, Burkhard SB et al. MicroRNA-23 restricts
cardiac valve formation by inhibiting Has2 and extracellular hyaluronic
acid production. Circ Res 2011; 109: 649–657.
31 Kawasaki H, Taira K. Hes1 is a target of microRNA-23 during retinoic-acidinduced neuronal differentiation of NT2 cells. Nature 2003; 423: 838–842.
32 Kim H, Kim T, Jeong BC et al. Tmem64 modulates calcium signaling during
RANKL-mediated osteoclast differentiation. Cell Metab 2013; 17: 249–260.
This work is licensed under a Creative Commons Attribution 4.0
International License. The images or other third party material in
this article are included in the article’s Creative Commons license, unless indicated
otherwise in the credit line; if the material is not included under the Creative
Commons license, users will need to obtain permission from the license holder to
reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/
© The Author(s) 2016
Bone Research (2016) 16022