Phosphorylation-dependent osterix degradation negatively
regulates osteoblast differentiation
Seira Hoshikawa1,2 | Kouhei Shimizu2 | Asami Watahiki2 | Mitsuki Chiba1,2 |
Kan Saito1 | Wenyi Wei3 | Satoshi Fukumoto1,2,4 | Hiroyuki Inuzuka2
© 2020 Federation of American Societies for Experimental Biology
Abbreviations: ALP, alkaline phosphatase; BMP, bone morphogenetic protein; CHX, cycloheximide; CRL, Cullin-RING ubiquitin ligase; Fbw7, F-box/
WD repeat-containing protein 7; MAPK, mitogen-activated protein kinase; MSCs, mesenchymal stem cells; MM, multiple myeloma; OCN, osteocalcin;
Osx, osterix; SCF, Skp1-Cullin-F-box protein; UPS, ubiquitin-proteasome system.
Division of Pediatric Dentistry,
Department of Oral Health and
Development Sciences, Tohoku University
Graduate School of Dentistry, Sendai, Japan
Center for Advanced Stem Cell and
Regenerative Research, Tohoku University
Graduate School of Dentistry, Sendai, Japan
Department of Pathology, Beth Israel
Deaconess Medical Center, Harvard
Medical School, Boston, MA, USA
Department of Pediatric Dentistry, Kyushu
University Graduate School of Dental
Science, Fukuoka, Japan
Correspondence
Satoshi Fukumoto, Division of Pediatric
Dentistry, Department of Oral Health and
Development Sciences, Tohoku University
Graduate School of Dentistry, 4-1 Seiryo,
Aoba-ku, Sendai 980-8575, Japan.
Email: [email protected]
Hiroyuki Inuzuka, Center for Advanced
Stem Cell and Regenerative Research,
Tohoku University Graduate School of
Dentistry, 4-1 Seiryo, Aoba-ku, Sendai 980-
8575, Japan.
Email: [email protected]
Funding information
Japan Society for the Promotion of Science
(JSPS), Grant/Award Number: 17H01606,
19H03834 and 19J11773
Abstract
Proteasome inhibitors exert an anabolic effect on bone formation with elevated levels
of osteoblast markers. These findings suggest the important role of the proteasomal
degradation of osteogenic regulators, while the underlying molecular mechanisms
are not fully understood. Here, we report that the proteasome inhibitors bortezomib
and ixazomib markedly increased protein levels of the osteoblastic key transcription
factor osterix/Sp7 (Osx). Furthermore, we revealed that Osx was targeted by p38 and
Fbw7 for proteasomal degradation. Mechanistically, p38-mediated Osx phosphoryl￾ation at S73/77 facilitated Fbw7 interaction to trigger subsequent Osx ubiquitination.
Consistent with these findings, p38 knockdown or pharmacological p38 inhibition
resulted in Osx protein stabilization. Treatment with p38 inhibitors following osteo￾genic stimulation efficiently induced osteoblast differentiation through Osx stabili￾zation. Conversely, pretreatment of p38 inhibitor followed by osteogenic challenge
impaired osteoblastogenesis via suppressing Osx expression, suggesting that p38
exerts dual but opposite effects in the regulation of Osx level to fine-tune its activity
during osteoblast differentiation. Furthermore, Fbw7-depleted human mesenchymal
stem cells and primary mouse calvarial cells resulted in increased osteogenic capac￾ity. Together, our findings unveil the molecular mechanisms underlying the Osx pro￾tein stability control and suggest that targeting the Osx degradation pathway could
help enhance efficient osteogenesis and bone matrix regeneration.
KEYWORDS
Fbw7, p38, ubiquitination
1 | INTRODUCTION
The ubiquitin-proteasome system (UPS) is involved in var￾ious biological processes including cell proliferation and
differentiation via protein ubiquitination and degradation.
Aberrant signaling pathways in the UPS are associated with
various diseases, and therefore, the UPS is regarded as a
potential therapeutic target.1,2 Several drugs targeting the
2 |
HOSHIKAWA et al.
UPS are currently in clinical use and are effective in treating
cancers. Bortezomib, a first-in-class proteasomal inhibitor,
was found to have significant efficacy in clinical treatment
of multiple myeloma (MM).3
Besides its antitumor effects,
bortezomib alleviates osteolytic bone lesions and increases
osteoblastic marker levels in patients.4-7 In addition, the
next-generation proteasome inhibitors carfilzomib and ixaz￾omib (bortezomib-analog) also exhibited bone anabolic effi￾cacy in a MM mouse model.8,9 These observations indicated
that preventing proteasomal degradation of major osteoblas￾tic regulators is the mechanism underlying the enhanced bone
anabolic effects.
Osteoblast differentiation is accomplished by a transcrip￾tional network program governed by osterix (Osx, also called
Sp7), a key osteogenic transcription factor.10 Osx is a mem￾ber of the Sp family of C2H2-type zinc finger transcription
factors, which play a major role in enhancing the expression
of transcripts encoding osteoblastic markers through binding
to the GC-rich region of the target gene cis-elements, as well
as serving as a co-activator for the homeobox-transcription
factor Dlx5.11-14 Accumulating evidence emphasizes the crit￾ical physiological and pathophysiological relevance of Osx
function in maintaining bone integrity.15 However, relatively
little is known about the upstream signaling and detailed mo￾lecular mechanisms that regulate Osx protein stability during
the osteogenic differentiation program.
Several lines of evidence indicate that p38 mitogen-acti￾vated protein kinase (MAPK) plays crucial roles in controlling
osteoblast differentiation, in which osteogenic induction ele￾vates p38 kinase activity to trigger the cellular differentiation
program.16-18 Mechanistically, activated p38 phosphorylates
Runx2 and Dlx5 to augment their transactivation activities,
leading to the induction of Osx transcription.18-20 In addi￾tion, p38 directly stimulates Osx transactivation activity by
phosphorylating S73/77 of Osx to recruit the histone acetyl￾ation and chromatin remodeling factors p300 and Brg-1 to
the phosphorylated Osx for the induction of downstream
target genes.13 Furthermore, this phosphorylation enhances
the interaction between Osx and Runx2, facilitating gene
expression of osteoblast markers. Together, these studies re￾vealed that p38 plays important roles in inducing Osx gene
expression and transactivation activity leading to osteoblast
differentiation.
F-box/WD repeat-containing protein 7 (Fbw7) is one of
the best characterized members of the F-box protein family
and serves as a receptor subunit of the Skp1-Cullin1-F-box￾protein (SCF)Fbw7 E3 ligase complex. Fbw7 is considered a
tumor suppressor because it selectively binds to potent on￾coproteins, including c-Myc, c-Jun, cyclin E, Notch1, and
Mcl-1 for proteasomal degradation.22,23
Besides its anticancer effects, Fbw7 reportedly contributes
to the regulation of osteoblast, chondrocyte, and osteoclast
formation through the degradation of OASIS transcription
factors and Notch2,24,25 suggesting the potential role of Fbw7
as a critical regulator of bone formation. In this study, we
focused on the mechanisms underlying cross talk between
p38/Fbw7-mediated proteolysis and the osteogenic pathway
through Osx degradation. We investigated the molecular
mechanism of phosphorylation-dependent Osx degradation
and the role of the Osx stability control in osteoblast differ￾entiation of mesenchymal stem cells (MSCs) and preosteo￾blasts. This study suggests that understanding the regulatory
mechanism of the p38/Fbw7-dependent Osx degradation
would provide insights into potential therapeutic approaches
for use in bone-regeneration medicine via inhibition of the
Osx degradation pathway.
2 | MATERIALS AND METHODS
2.1 | Cell culture
MC-3T3-E1 cells (RRID:CVCL_0409) were obtained from
RIKEN BRC (Tsukuba, Ibaraki, Japan). UCBTERT-21
(RRID:CVCL_3232) and UE7T-13 (RRID:CVCL_3238)
cells were obtained from JCRB cell bank (Ibaraki, Osaka,
Japan). 293T (RRID:CVCL_0063) and Saos-2 cells
(RRID:CVCL_0548) were obtained from ATCC (Minnesota,
VA, USA). All cell lines were cultured in Dulbecco’s modi￾fied Eagle’s medium (DMEM) supplemented with 10% of
fetal bovine serum (FBS), 100 units of penicillin, and 100 μg/
mL of streptomycin. 293T cells were used for transfection￾based assays to identify the molecular mechanism of Osx
degradation by reconstituting signaling molecules involved
in posttranslational modifications of Osx. Fbw7–/– HCT116
cells and the corresponding control Fbw7+/+ cells (kind gifts
from Dr. Bert Vogelstein)26 are widely used in studies on
Fbw7; therefore, we used these cells to investigate the effect
of Fbw7 ablation on the regulation of the Osx protein half-life.
MC3T3-E1, UCBTERT-21, UE7T-13, and Saos-2 cells were
used for the osteogenic differentiation assay as they serve as a
physiologically relevant system in osteogenesis. Immortalized
human MSCs, UCBTERT-21, and UE7T-13, which maintain
multipotency and osteogenic capacity even after long-term
serial passages,27,28 were used to conduct shRNA-mediated
Fbw7 knockdown and MSC differentiation analyses. Saos-2
cells were used to generate Osx-knockout cells to evaluate the
effect of Osx ablation on osteogenic differentiation because
they maintain their osteogenic capacity after CRISPR-Cas9-
mediated Osx gene editing. Transfection was performed using
polyethylenimine (PEI).29 Briefly, 293T cells were seeded
at 3.0 × 106
cells per 100-mm plate the day before transfec￾tion. For transfection, 5 μg of plasmid DNA was diluted in
700 μL of serum-free DMEM, and 15 μg of PEI (1 μg/mL)
(Polysciences, Warrington, PA, USA) was directly added to
the diluted DNA. After vigorous vortexing for 15 seconds,
| HOSHIKAWA et al. 3
the mixture was incubated at room temperature for 15 min￾utes and added to the cells. After 48 hours of incubation, the
transfected cells were harvested for immunoblot analysis or
immunoprecipitation. Lentiviral packaging and subsequent
infection were performed according to a protocol described
previously.30 Following viral infection, the cells were selected
for 72 hours in the presence of puromycin (1 μg/mL) or hy￾gromycin (200 μg/mL), depending on the viral vectors used
for infection. The protein synthesis inhibitor cycloheximide
(CHX) was used at 100 μg/mL. The selective p38 inhibitors
SB203580 and SB239063 (Calbiochem, Merck, Darmstadt,
Germany) were used at 10 μM concentration.
2.2 | Induction of osteoblast differentiation
MC3T3-E1, UCBTERT-21, UE7T-13, and primary mouse cal￾varial cells were seeded in 96-well plates at 6.0×103
cells per
well and cultured for 48 hours until cells were confluent. For
osteoblast differentiation, MC3T3-E1 cells and primary mouse
calvarial cells were treated with bone morphogenetic protein-4
(BMP-4) (10 ng/mL) (rhBMP-4, HZ-1045, HUMANZYME,
Chicago, IL, USA) for 7 days. Human UCBTERT-21 cells
were cultured in medium supplemented with Osteoblast￾Inducer Reagent (Takara Bio Inc., MK430, Kusatsu, Shiga,
Japan) and BMP-4 (10 ng/mL) for 7 days. Differentiation of
human UE7T-13 was induced by Osteoblast-Inducer Reagent
for 1-2 weeks. For Saos-2 osteogenic stimulation, cells were
seeded in a 24-well plate at 2.0 × 104
cells per well. At 48
hours after plating, cells were treated with potassium dihydro￾gen phosphate (1.8 mM) and L-ascorbic acid (50 μg/mL) for 2
weeks and utilized for assays.
2.3 | p38 inhibitor treatments
UE7T-13 cells were seeded in 60-mm plate at a high density so
that they reached confluence the next day. For the inhibitor pre￾treatment, an inhibitor stock solution (SB203580 or SB239063,
10 mM in DMSO) was directly added to the culture medium
at 1:1000 dilution (10 μM). After 12 hours of pretreatment, the
medium was replaced with Osteoblast-Inducer Reagent contain￾ing p38 inhibitor (10 μM), and cells were harvested at the indi￾cated periods. For the inhibitor posttreatment, the cells were first
stimulated with Osteoblast-Inducer Reagent for 12 hours, and
the p38 inhibitor stock was then directly added to the medium.
The treated cells were harvested at the indicated periods.
2.4 | Antibodies and plasmids
Anti-Fbw7 antibody (A301-720A) was purchased from
Bethyl Laboratories (Montgomery, TX, USA). Anti-p-p38
(9211), anti-p38 (9212), anti-p-Smad (9516), anti-Smad1
(6944), anti-Notch1 (4380), and anti-p-MAPKAPK-2 (MK-
2) (3007) antibodies were purchased from Cell Signaling
Technology (Danvers, MA, USA). Anti-ALP (sc-137213),
anti-cyclin E (sc-247), anti-green fluorescent protein (GFP)
(sc-8334), anti-osteocalcin (sc-365797), anti-Osx (F-3) (sc-
393325), anti-Osx (A-13) (sc-22536-R), anti-Runx2 (sc-
10758), anti-β-actin (sc-47778), anti-HA (F-7) (sc-7392),
anti-HA (Y-11) (sc-805), and anti-GST (Z-5) (sc-459) an￾tibodies were purchased from Santa Cruz Biotechnology
(Dallas, TX, USA). Anti-β-catenin (C7207), anti-p-Ser/Thr￾Pro MPM-2 (05-368) antibodies, and anti-Flag M2 affinity
gel (A2220) were purchased from Sigma-Aldrich (St. Louis,
MO, USA). Anti-Flag monoclonal (018-22381) antibody,
anti-HA antibody beads (014-23081), and anti-c-Myc anti￾body beads (10D11) (016-26503) were purchased from Wako
(Osaka, Japan). Anti-Flag polyclonal (PM020) and anti-Myc￾tag (562) antibodies were purchased from MBL (Nagoya,
Japan). The sources of HA-Fbw7 (WT, R465H, R479L, and
R505C) and shRNA against GFP and Fbw7 have been de￾scribed previously.31,32 shRNAs for p38 were purchased from
Sigma-Aldrich (TRCN0000000510 and TRCN0000194820).
pcDNA-HA-GSK3β was purchased from Addgene (14753,
Watertown, MA, USA). pcDNA-Flag-Osx was a kind gift
from Dr. Takenobu Katagiri. Flag-Osx-S73/77A was gen￾erated with the QuikChange Site-Directed Mutagenesis Kit
(Agilent Technologies, Santa Clara, CA, USA) according
to the manufacturer’s instructions. GST-Osx (33-145aa)
wild-type (WT) and S73/77A expression plasmids were
constructed by subcloning the appropriate PCR fragments
into pGEX4T-3. The sources of GST-F-box proteins, Myc￾Cullin isoforms, HA-p38, HA-ERK, HA-cyclin A, and His￾ubiquitin (His-ub) were described previously.30,33,34
2.5 | Immunoblots and immunoprecipitation
Cells were lysed in NP-40 cell lysis buffer (50 mM Tris, pH
7.5, 120 mM NaCl, 0.5% NP-40) supplemented with a pro￾tease inhibitor cocktail (cOmplete, Mini Protease Inhibitor
Cocktail, Roche, Basel, Switzerland) and phosphatase inhibi￾tors (PhosSTOP, Calbiochem). The protein concentrations of
lysates were determined with the Bio-Rad Protein Assay Dye
(Bio-Rad Laboratories, Hercules, California, USA). Forty
micrograms of whole-cell lysates was resolved by sodium
dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis
(PAGE) and transferred to polyvinylidene difluoride mem￾branes (Bio-Rad). The membranes were blocked with 5% of
nonfat dry milk in Tris-buffered saline with 0.05% of Tween
20 (TBST, pH 8.0) and probed with the antibodies (at 1:1000-
1:4000 with 5% nonfat dry milk in TBST) indicated in the
Figures. For immunoprecipitation, the cells were treated with
MG132 (15 μM) for 12 hours and harvested in NP-40 cell
4 |
HOSHIKAWA et al.
lysis buffer containing protease and phosphatase inhibitors.
Six hundred microliters (1 mg) of cell lysate was incubated
with 8 μL slurry of anti-Myc- or Flag-tag antibody-conjugated
beads or 15 μL slurry of glutathione S-transferase (GST)-
tagged protein purification resin (Glutathione Sepharose 4B,
GE Healthcare, Chicago, IL, USA) for 4 hours at 4°C with
gentle rocking. The tubes were briefly microcentrifuged to
precipitate the immunoprecipitates or GST-pulldowns, and
then, the beads were washed five times with 1 mL of NP-40
washing buffer (20 mM Tris, pH 8.0, 100 mM NaCl, 1 mM
EDTA, and 0.5% NP-40), resuspended in 50 μL of 2×SDS
sample buffer, and heated at 90°C for 5 minutes. The eluted
samples (15 μL) and whole-cell lysates (input: 60 μg) were
separated by SDS-PAGE, followed by immunoblot analysis.
2.6 | In vitro ubiquitination assays
293T cells were transfected with constructs encoding Flag-Osx
(WT or S73/77A) and HA-Fbw7. At 48 hours after transfec￾tion, the cells were treated with MG132 (15 μM) alone or to￾gether with the p38 inhibitors (10 μM) for 12 hours and lysed
with NP-40 cell lysis buffer for HA-immunoprecipitation to
purify the HA-Fbw7/Flag-Osx complex. The complex was
mixed with recombinant E1, E2, and ubiquitin (R&D systems,
Minneapolis, MN, USA), and the in vitro ubiquitin conjuga￾tion reaction was performed according to the manufacturer’s
instructions. The reactions were separated by SDS-PAGE, and
Osx polyubiquitination was detected by immunoblot analysis.
2.7 | In vitro kinase reaction and
binding assay
Recombinant GST-Osx proteins were purified from trans￾formed E. coli DH5α. Briefly, 20 mL of LB medium was
inoculated with 1 mL of fresh overnight culture, and the cells
were grown at 37°C with shaking until an O.D.600 of 0.8. The
expression of the GST-fusion protein was then induced with
0.4 mM of IPTG. After culturing at 30°C for 3 hours with
vigorous shaking, cells were collected by centrifugation, and
pellets were resuspended in 1 mL of Tris-buffered saline
(TBS) for sonication. Cell debris was discarded following
centrifugation, and the supernatant was incubated with 50
µl of 50% Glutathione Sepharose 4B slurry (GE Healthcare)
with gentle rocking for 3 hours at 4°C. The GST-protein￾bound beads were washed three times with 1 mL of TBS and
stored at 4°C. The amount of the GST-fusion protein was es￾timated with the GelCode Blue Stain Reagent (Thermo Fisher
Scientific, Waltham, MA, USA). For the in vitro kinase reac￾tion, 5 μg of GST-Osx-bound beads (WT and S73/77A) was
incubated with purified enzymatically active p38α (R&D
systems) in the presence of 20 μM ATP in kinase reaction
buffer (New England Biolabs, Ipswich, MA, USA) for 30
minutes at 30°C. The reaction was stopped by the addition
of SDS sample buffer and resolved by SDS-PAGE for im￾munoblot analysis, and the phosphorylation was detected by
anti-p-Ser/Thr-Pro antibody. For the in vitro binding assay,
the in vitro kinase reaction was stopped by washing the GST￾protein-bound beads three times with 1 mL of NP-40 wash￾ing buffer. The beads were then incubated with whole-cell
lysates derived from 293T cells transfected with HA-Fbw7
for 3 hours at 4°C with gentle rocking, washed five times
with 1 mL of NP-40 washing buffer, and resolved by SDS￾PAGE for immunoblot analysis. The binding of HA-Fbw7
was detected using anti-HA antibody.
2.8 | Real-time reverse transcription (RT)-
PCR analysis
Total RNA was extracted with TRIzol Reagent (Thermo
Fisher Scientific). The reverse transcription (RT) reaction
was performed with ReverTra Ace qPCR RT Master Mix
(TOYOBO, Osaka, Japan). The real-time RT-PCR reaction
was performed with SYBR Select Master Mix, or TaqMan
Gene Expression Assay and TaqMan Gene Expression
Master Mix using a 7500 Real-Time PCR System (Applied
Biosystems, Thermo Fisher Scientific). Relative gene ex￾pression was calculated using the ddCt method, and Gapdh
was used for the normalization of transcript levels. All pro￾cedures were performed according to the manufacturers’ in￾structions. The primers used for the PCR reactions are listed
in Supplementary Table 1.
2.9 | Osx-knockout Saos-2 and Fbw7-
knockout primary mouse calvarial cells
For generating Osx-knockout Saos-2 cells, the sgRNA plasmid
was constructed by inserting the sgRNA oligos (human-Osx￾sgOligo-Fw: 5′ CACCGCACAAAGAAGCCGTACTCTG
3′, Rv: 5′ AAACCAGAGTACGGCTTCTTTGTGC 3′)
into pLenti-CRISPRv2-Blast and transfected into Saos-2
by PEI. At 48 hours after transfection, cells were selected
with blasticidin (1.5 μg/mL) for 3 days. The resulting cells
were subjected to clonal isolation by the single cell dilution
method in a 96-well plate. Knockout of Osx was validated
by immunoblot analysis. For the culture of Fbw7-knockout
primary mouse calvarial cells, osteoblast-specific Fbw7-
knockout (Col1a1-Cre;Fbw7fl/fl) mice were generated by
crossing floxed Fbw7 (Fbw7fl/fl) mice (kindly provided by
Dr. Keiko Nakayama) with Collagen1a1-promoter-Cre re￾combinase transgenic mice (Col1a1-Cre) (RIKEN BRC,
05524). Genomic DNA was extracted from tail samples or
primary mouse calvarial cells and used as a template for PCR
| HOSHIKAWA et al. 5
genotyping. The PCR primers used for genotyping are listed in
Supplementary Table 1. Primary mouse calvarial cells were iso￾lated from neonatal mice for the evaluation of osteoblast differ￾entiation capacity. All care and experimental procedures were
conducted according to the Tohoku University Institutional
Animal Care and Use Committee protocol (#2019DnA-059).
2.10 | Alkaline phosphatase (ALP) and
matrix mineralization assays
For ALP staining, the cells were washed twice with saline
and fixed with 4% of paraformaldehyde for 5 minutes. The
cells were again washed with saline, refixed with ethanol/ac￾etone (50:50 v/v), washed with saline once, and stained with
Sigmafast BCIP/NBT (Sigma-Aldrich). ALP activity was de￾termined with the TRACP and ALP Assay Kit (Takara Bio).
For activity measurement, the reaction was terminated by add￾ing 5 M of NaOH, and the absorbance was measured at 405
nm by using a multimode reader (TriStar2
LB942; Berthold
Technologies GmbH, Bad Wildbad, Germany). For the min￾eralization assay, the cells were washed twice with saline,
fixed with 4% of paraformaldehyde for 5 minutes, washed with
saline once, and stained with Alizarin Red S (PG Research,
Kodaira, Tokyo, Japan) for 30 minutes at room temperature.
The stained cells were washed with water and dehydrated in
the air. To quantitate calcium deposition, extraction solution
(5% formic acid) was added to the dehydrated stained cells, and
absorbance was measured at 405 nm using a multimode reader.
2.11 | Quantification and statistical analysis
All quantitative data are presented as mean ± standard error
of the mean (SEM) values indicative of at least three inde￾pendent experiments or biological replicates. Between-group
differences were analyzed by one-way ANOVA with multi￾ple-comparison tests or Student’s t test. Statistical analyses
were performed using GraphPad Prism (GraphPad Software,
San Diego, CA, USA), and P < .05 was considered a statisti￾cally significant difference.
3 | RESULTS
3.1 | Proteasome inhibitors enhance
osteoblast differentiation and synergistic
induction of Osx protein in preosteoblastic and
mesenchymal stem cells
The clinical usage of bortezomib in the treatment of multi￾ple myeloma improved osteolytic bone lesions and increased
osteoblastic markers in patients.4-7 Besides, the first oral pro￾teasome inhibitor ixazomib exhibited bone anabolic effects
in multiple myeloma mouse models.8
To investigate the ef￾ficacy of these proteasome inhibitors in osteoblast differen￾tiation, we treated two osteogenic cell lines, mouse calvarial
preosteoblasts MC3T3-E1 and human umbilical cord blood￾derived MSCs UCBTERT-21, with BMP and the proteasome
inhibitors simultaneously. For evaluating drug effectiveness
and efficacy, we utilized these compounds with a range of
concentrations as low as 0.1-10 pM, which was 1/1000th that
used in previous studies.5,7,8 After a 1-week treatment, we
evaluated cellular differentiation with ALP activity obtained
as readout. Proteasome inhibitor treatments significantly en￾hanced BMP-elicited osteoblast differentiation in MC3T3-E1
(Figures 1A and S1A) and UCBTERT-21 (Figures 1B and
S1B) cells. On osteogenic stimulation, Osx protein levels re￾markably increased between 12 and 48 hours in response to
simultaneous treatments involving proteasome inhibitors and
osteogenic stimulation (Figure 1C,D), whereas Runx2, osteo￾calcin (OCN), and phosphorylated Smad levels did not show
any significant differences. In contrast, proteasome inhibi￾tor treatments did not affect Osx mRNA expression levels
(Figure 1E). Together, these data indicate that the treatments
with proteasome inhibitors efficiently promoted osteoblast
differentiation and markedly enhanced Osx protein levels in
the osteogenic cell lines MC3T3-E1 and UCBTERT-21.
3.2 | Fbw7 controls Osx protein stability
Given the significant effects of proteasome inhibitors on en￾hancing Osx levels (Figure 1C,D), we attempted to identify
the E3 ligase responsible for the regulation of Osx protein
stability. Among several hundred E3s, cullin-RING ubiqui￾tin ligases (CRLs) compose the largest E3 superfamily and
mediate approximately 20% in total of proteasomal degrada￾tion.35,36 To this end, we looked at the possible involvement
of CRLs in Osx ubiquitination and subsequent degradation.
Using co-immunoprecipitation (co-IP), we investigated the
interaction between Osx and cullin family proteins, the cen￾tral scaffold subunit in the CRL complexes, and found that
Osx selectively bound to Cullin1, a component of the CRL1
generally termed SCF E3 ubiquitin ligase (Figure 2A). Since
the SCF includes a substrate subunit F-box protein that de￾termines its substrate specificity, we next tested which F-box
protein interacts with Osx. By co-expression of a set of GST￾F-box proteins with Flag-Osx in 293T cells, we identified
that Osx specifically interacted with Fbw7, but not with the
other F-box proteins we tested (Figure 2B). Furthermore, a
co-IP experiment revealed that Fbw7 mutants deficient in
substrate recognition37 abolished the interaction with Osx
(Figure 2C). We also conducted a co-IP experiment using
6 |
HOSHIKAWA et al.
Fbw7+/+ and Fbw7–/– HCT116 cells transfected with Myc￾Cullin1 and Flag-Osx and observed that the interaction be￾tween Osx and Cullin1 was sustained in Fbw7+/+ but not in
Fbw7–/– cells, suggesting that this interaction is likely medi￾ated by Fbw7 (Figure 2D). Next, to investigate the effects of
Fbw7 on Osx protein half-life, we performed cycloheximide
(CHX) chase assay with HCT-116 cells in which Fbw7 is
depleted and demonstrated that the Osx protein half-life was
greater in Fbw7–/– HCT-116 cells than in control Fbw7+/+
cells (Figure 2E,F). Together, these data indicate that the
SCFFbw7 complex plays a role in controlling Osx protein
stability.
3.3 | Osx is destabilized via p38-mediated
phosphorylation
Fbw7 selectively binds to its target proteins via the consensus
Fbw7 degron motif.22 We found the conserved Fbw7 degron￾like motif in the Osx amino sequence (Figure 3A). Consistent
with this observation, our co-IP experiment showed that mu￾tating the sequence in Osx resulted in a marked decrease in
its interaction with Fbw7 (Figure 3B). Given that prior phos￾phorylation of the central Ser/Thr residues in this motif is
required for the interaction between Fbw7 and substrates,22
we attempted to identify Fbw7 degron-priming kinases by
FIGURE 1 Proteasome inhibitors enhance osteoblast differentiation and induced Osx accumulation in MC3T3-E1 and UCBTERT-21 cells
(See also Figure S1). A-B, MC3T3-E1 (A) and UCBTERT-21 (B) were treated for 7 days with or without BMP-4 (10 ng/mL) and proteasome
inhibitors at the indicated concentrations. The cells were fixed and stained to assess ALP activity. C-D, Immunoblot (IB) analysis of whole-cell
lysates (WCLs) derived from MC3T3-E1 (C) and UCBTERT-21 (D) treated with BMP-4 (10 ng/mL) and proteasome inhibitors at the indicated
concentrations. The cells were harvested at the indicated time points for IB analysis. E, Real-time RT-PCR analysis to determine the relative
mRNA expression levels of Osx in MC3T3-E1 cells presented in (C). Gapdh was utilized for normalization. Data are presented as mean ± SEM
values (n = 3). n.s., not significant. Comparison was made with one-way ANOVA followed by multiple-comparison test at the same time points
between two groups
| HOSHIKAWA et al. 7
using a set of kinases previously reported to phosphorylate
this motif.38,39 Ectopic expression of p38, but not of the oth￾ers we tested, facilitated a decrease in Osx protein levels in
293T cells (Figure 3C). The Osx-S73/77A mutant was pro￾tected from p38-dependent Osx downregulation (Figure 3D).
To confirm that the Osx residues S73/77 represent the phos￾phorylation sites for p38, we conducted an in vitro kinase
reaction using bacteria-purified GST-Osx (33-145aa) fusion
proteins (WT or S73/77A) and a recombinant enzymatically
active p38. We found that p38 phosphorylated Osx mainly
at the S73/77 sites (Figure 3E). Besides, we conducted an
in vitro binding assay by incubating p38-treated GST-Osx
proteins in whole-cell lysates derived from HA-Fbw7-
expressing 293T cells, followed by a GST-pulldown assay.
We observed specific interaction of Fbw7 with p38-treated
GST-Osx-WT but not with S73/77A (Figure 3F). Consistent
with these results, the Fbw7 degron mutant exhibited a de￾creased polyubiquitination level in the in vitro ubiquitination
assay (Figure 3G), suggesting that the p38-mediated Osx
phosphorylation at S73/77 is critical in the control of Osx
protein stability. Furthermore, pharmacological inhibition of
p38 resulted in impaired ubiquitination of Osx (Figure 3H).
Next, to investigate the effects of p38-mediated phosphoryla￾tion of Osx S73/77 on Osx protein turnover, we conducted a
CHX chase experiment using 293T cells ectopically express￾ing Osx, Fbw7, and p38. Forced expression of p38 resulted in
a shortened protein half-life of Osx-WT, but not of S73/77A
(Figure 3I,J). Moreover, we evaluated the endogenous Osx
half-life in cells with Fbw7 ablation in combination with
shRNA-mediated p38 depletion. The knockdown of p38
led to marked Osx stabilization in Fbw7+/+ cells; this effect
was refractory to Fbw7 ablation as evident in Fbw7–/– cells
FIGURE 2 SCFFbw7 controls Osx protein stability. A, Immunoblot (IB) analysis of whole-cell lysates (WCLs) and anti-Myc
immunoprecipitates (IPs) derived from 293T cells transfected with Flag-Osx along with empty vector (EV) or Myc-Cullins (Cullin1, 2, 3, 4A, or
5), as indicated. At 36 hours after transfection, the cells were treated with the proteasome inhibitor MG132 (15 μM) for 12 hours before harvesting.
B, IB analysis of WCLs and GST-pulldowns derived from 293T cells transfected with EV or the indicated GST-F-box plasmids along with Flag￾Osx, as shown. At 36 hours after transfection, cells were treated with the proteasome inhibitor MG132 (15 μM) for 12 hours before harvesting. C,
IB analysis of WCLs and anti-Flag IPs derived from 293T cells transfected with Flag-Osx along with EV and HA-Fbw7 (WT, R465H, R479L, or
R505C), as shown. At 36 hours after transfection, cells were treated with the proteasome inhibitor MG132 (15 μM) for 12 hours before harvesting.
D, IB analysis of WCLs and anti-Myc IPs derived from wild-type (WT) (Fbw7+/+) and Fbw7-knockout (Fbw7–/–) HCT116 cells transfected with
Myc-Cullin1 and Flag-Osx, as shown. At 36 hours after transfection, cells were treated with the proteasome inhibitor MG132 (15 μM) for 12
hours before harvesting. E, IB analysis of WCLs derived from WT (Fbw7+/+) and Fbw7-knockout (Fbw7–/–) HCT116 cells treated with the protein
synthesis inhibitor cycloheximide (CHX; 100 μg/mL). The cells were harvested at the indicated time points for IB analysis. F, Quantification of the
relative Osx band intensities in (C), which were normalized to those of the loading control β-actin and to the time point t = 0. Data are presented as
mean ± SEM values (n = 3). *P < .05; Student’s t test
8 |
HOSHIKAWA et al.
(Figure 3K,L), confirming the fact that p38 is required for
Fbw7-directed degradation of Osx. Together, these data
show that p38-mediated phosphorylation at S73 and S77 is
required for Osx ubiquitination and subsequent degradation
by Fbw7.
3.4 | Inhibition of p38 leads to Osx protein
stabilization
To evaluate the physiological significance of p38-dependent
Osx degradation, we treated the human bone marrow-derived
| HOSHIKAWA et al. 9
MSC cell line, UE7T-13, with the p38-selective inhibitors
SB203580 and SB239063 and assessed Osx protein abun￾dance. The inhibitor treatment resulted in an increase in Osx
protein but not mRNA levels (Figure 4A,B). Next, to confirm
that the Osx upregulation is due to Osx protein stabilization,
we conducted a CHX chase experiment and demonstrated that
p38 inhibitor treatments led to extended Osx protein half-life
(Figure 4C,D). In accordance with the negative role of p38 ki￾nase activity in the regulation of Osx protein stability, the level
of phosphorylated p38 was negatively associated with Osx
protein abundance in osteogenically stimulated MC3T3-E1
cells, most likely through posttranslational and transcriptional
regulations (Figure S2A,B). These results suggest that p38
kinase activity negatively contributes to the regulation of pro￾tein stability and abundance of Osx in osteoblastogenesis.
3.5 | p38 exerts dual effects in both
induction and downregulation of Osx activity
Osteogenic extracellular stimulation elevates p38 activity to
trigger the cellular differentiation program, in which p38 plays
an important role in the induction of Osx expression and trans￾activation activity.13,16,18-21 Thus, we postulated that p38 is
necessary to fine-tune Osx activity through both inducing and
decreasing Osx protein levels. To this end, UE7T-13 cells were
treated with p38 inhibitors for two different time courses, 12
hours before osteogenic stimulation to inhibit Osx mRNA ex￾pression, and 12 hours after osteogenic stimulation to evaluate
whether this treatment inhibits Osx degradation leading to Osx
accumulation (Figure 5A). In accordance with previous stud￾ies20, inhibitor pretreatment resulted in suppression of Osx lev￾els compared to those of untreated cells by impairing the Osx
mRNA level (Figure 5B,C). On the contrary, the p38 inhibitor
treatment following osteogenic induction resulted in enhanced
Osx protein accumulation (Figure 5D). Consistent with these
results, inhibitor pretreatments resulted in a decrease in osteo￾blast differentiation compared with that noted for control cells,
as shown by the staining with ALP (Figure 5E,F) and Alizarin
Red S (Figure 5G,H) and the expression of the osteoblast
marker gene OCN (Figure 5I). In contrast, posttreatment with
inhibitors resulted in significantly increased osteoblastogen￾esis compared to that of control cells, as shown in the staining
with ALP (Figure 5J,K) and Alizarin Red S (Figure 5L,M),
FIGURE 3 p38-mediated Osx phosphorylation facilitates Fbw7 interaction and subsequent Osx degradation. A, Consensus Fbw7 recognition
sequence in human and mouse Osx. The Fbw7 degron sequence is defined as T/S-P-X-X-S/T/D/E-P, where Thr and Ser residues need to be
phosphorylated for Fbw7 to recognize this motif. B, Immunoblot (IB) analysis of whole-cell lysates (WCLs) and Flag immunoprecipitates (IPs)
derived from 293T cells transfected with empty vector (EV) or Flag-Osx (WT or S73/77A) along with HA-Fbw7, as indicated. At 36 hours after
transfection, cells were treated with the proteasome inhibitor MG132 (15 μM) for 12 hours before harvesting. C, IB analysis of WCLs derived from
293T cells transfected with the expression plasmids encoding the indicated protein kinases along with Flag-Osx and HA-Fbw7, as shown. GFP
was used as an internal control for transfection efficiency. At 48 hours after transfection, the cells were harvested for IB analysis. D, IB analysis
of WCLs derived from 293T cells transfected with Flag-Osx (WT or S73/77A) along with HA-p38 and HA-Fbw7, as indicated. GFP was used as
an internal control for transfection efficiency. At 48 hours after transfection, the cells were harvested for IB analysis. E, p38 in vitro kinase assay
showing that the residues S73/77 of Osx are the phosphorylation sites for p38. Bacteria-purified GST-Osx (33-145aa) proteins (WT or S73/77A)
bound to beads were treated with recombinant enzymatically active p38, as shown. The kinase reactions were then subjected to IB analysis with
the indicated antibodies. GST was used as a negative control. F, In vitro binding assay showing that phosphorylation of Osx at S73/77 is required
for the binding of Fbw7. Following the in vitro kinase reaction in (E), the p38-treated or nontreated beads-bound GST-Osx proteins were incubated
with WCLs derived from 293T cells transfected with HA-Fbw7 and subjected to GST-pulldown. The beads-bound proteins were analyzed by
IB analysis with the indicated antibodies. G, The complex of HA-Fbw7/Flag-Osx (WT or S73/77A) was purified with IP from WCLs derived
from 293T cells transfected with HA-Fbw7/Flag-Osx and treated with MG132 (15 μM) for 12 hours before harvesting. The complex was mixed
with recombinant E1, E2, and ubiquitin (ub) proteins to conduct an in vitro ubiquitination reaction. The reactions were separated by SDS-PAGE,
and polyubiquitination of Osx was analyzed by IB analysis. H, The complex of HA-Fbw7/Flag-Osx was purified with IP from WCLs derived
from 293T cells transfected with HA-Fbw7/Flag-Osx and treated with the p38-selective inhibitors (10 μM SB239063 or SB203580) and MG132
(15 μM) for 12 hours before harvesting. DMSO was used as the control. The complex was mixed with recombinant E1, E2, and ubiquitin (ub)
proteins to conduct an in vitro ubiquitination reaction. The reactions were separated by SDS-PAGE, and polyubiquitination of Osx was analyzed
by IB analysis. I, IB analysis of WCLs derived from 293T cells transfected with Flag-Osx (WT or S73/77A), HA-Fbw7, or HA-p38, as indicated.
GFP was used as the internal control for transfection efficiency. At 48 hours after transfection, the cells were treated with the protein synthesis
inhibitor CHX (100 μg/mL) and harvested at the indicated time points. J, Quantification of the relative Osx band intensities in (I), which were
normalized to those of GFP and to the time point t = 0. Data are presented as mean ± SEM values (n = 3). *P < .05, Osx-WT+p38 versus Osx￾WT; ## P < .01, Osx-WT+p38 versus Osx-S73/77A or S73/77A+p38; n.s., not significant; one-way ANOVA with multiple-comparison test. K,
IB analysis of WCLs derived from WT (Fbw7+/+) and FBW7-knockout (Fbw7–/–) HCT116 cells infected with shRNA lentiviral vectors specific
for p38 (two independent constructs: A and B). shRNA for GFP was used for control cells. The cells were selected for 7 days in the presence of
puromycin (1 μg/mL) to remove noninfected cells, and then, treated with CHX (100 μg/mL) during the period of time indicated before harvesting.
L, Quantification of the relative Osx band intensities in (K), which were normalized to those of the loading control β-actin and to the time point t =
0. Data are presented as mean ± SEM values (n = 3). *P < .05, **P < .01, Fbw7+/+ GFP versus Fbw7+/+ shp38; ##P < .01, Fbw7+/+ GFP versus
Fbw7–/– GFP or shp38; n.s., not significant; one-way ANOVA with multiple-comparison test
10 |
HOSHIKAWA et al.
and the expression of OCN (Figure 5N). Taken together, these
data demonstrated that blockade of p38 activity following Osx
induction promotes Osx protein accumulation and facilitates
osteoblast differentiation.
3.6 | Fbw7 ablation results in enhanced
osteogenic capacity
Having obtained mechanistic insights into Fbw7-mediated
Osx degradation (Figures 2 and 3), we next investigated
whether Fbw7 knockdown promotes osteoblast differen￾tiation in UE7T-13 cells. We found that Fbw7 depletion
remarkably and persistently increased the protein levels of
Osx as well as the Fbw7 substrate cyclin E in UE7T-13 cells
(Figure 6A). While Osx mRNA levels were decreased in
Fbw7-knockdown cells (Figure S3A), the protein half-life
markedly increased following Fbw7 depletion (Figure S3B),
suggesting that elevation of Osx levels was mainly through
Osx protein stabilization. Next, we evaluated osteoblast dif￾ferentiation capacity using the control and Fbw7-depleted
UE7T-13 cells. ALP and Alizarin Red S staining indicated
that Fbw7 depletion noticeably facilitated osteoblastogenesis
(Figure 6B-E). To further investigate the physiological role
of Fbw7-mediated Osx degradation in osteogenesis, primary
mouse calvarial cells were isolated from the neonatal calvarias
of control and Fbw7-knockout mice (Col1a1Cre(+);Fbw7fl/fl),
cultured under osteogenic condition for 7 days, and subjected
to differentiation analysis. ALP and Alizarin Red S staining
indicated that Fbw7 ablation led to significant enhancement
of osteoblast differentiation (Figure 6F-I). Furthermore, ex￾pression of the osteoblast marker genes ALP and OCN in￾creased in the differentiated Fbw7-knockout osteoblasts
(Figure 6J). Immunoblot analysis demonstrated that Fbw7
ablation accumulated Osx protein as well as the Fbw7 sub￾strate Notch140 in primary osteoblasts (Figure 6K), while the
Osx mRNA level was impaired by Fbw7 ablation (Figure 6J).
These results indicate that Fbw7 negatively regulates Osx
protein stability and osteoblast differentiation.
3.7 | Nondegradable Osx-S73/77A mutant
enhances osteoblast differentiation
Given the possible pleiotropic effect of Fbw7 on bone forma￾tion through degradation of multiple substrates, we focused
FIGURE 4 p38 activity negatively regulates abundance and stability of Osx protein (See also Figure S2). A, Immunoblot (IB) analysis of
whole-cell lysates (WCLs) derived from UE7T-13. Cells were cultured in osteoblast differentiation medium for 12 hours and serially treated with
SB239063 or SB203580 (10 μM) for 24 hours before harvesting. DMSO was used for control cells. B, Real-time RT-PCR analysis to determine
the relative mRNA expression levels of Osx in p38 inhibitor-treated UE7T-13 cells presented in (A). Data are presented as mean ± SEM values (n
= 3). n.s., not significant; one-way ANOVA with multiple-comparison test. C, IB analysis of WCLs derived from UE7T-13. Cells were cultured
in osteoblast differentiation medium for 12 hours, serially treated with SB239063 or SB203580 (10 μM) for 24 hours, and treated with the protein
synthesis inhibitor cycloheximide (CHX) (100 μg/mL). The cells were harvested at the indicated time points for IB analysis. Phospho-MK2 is a
readout of p38 activity. DMSO was used for control cells. D, Quantification of the relative Osx band intensities in (C), which were normalized
to those of the loading control β-actin and to the time point t = 0. Data are presented as mean ± SEM values (n = 3). *P < .05, control versus
inhibitor-treated groups; one-way ANOVA with multiple-comparison test
| HOSHIKAWA et al. 11
on the relevance of Fbw7-mediated Osx degradation. We
first generated CRISPR/Cas9-mediated Osx-knockout cells
by using the human osteoblastic osteosarcoma cell line Saos-
2. To characterize the significance of Fbw7-directed Osx
degradation in osteoblast differentiation, the Osx-knockout
cells were further transferred with Osx-WT and nondegrada￾ble Osx-S73/77A. The levels of reintroduced Osx were con￾firmed by immunoblot analysis (Figure 7A). The resulting
cells were challenged with osteogenic stimulation, and we
confirmed that Osx-knockout cells had decreased potential
to differentiate to osteoblasts, and Osx-WT restored the dif￾ferentiation capacity of Osx-knockout cells (Figure 7B,C).
Furthermore, the expression of the nondegradable Osx￾S73/77A mutant exhibited greater osteogenic potential as
evident in the comparison of differentiation status between
Osx-WT and Osx-S73/77A expressing cells (Figure 7B,C).
Moreover, transcript levels of the osteoblast marker gene
OCN were consistent with the results of calcium deposition
12 |
HOSHIKAWA et al.
(Figure 7C,D). Together, these results showed that Osx
stabilization is critical for efficiently inducing osteoblast
differentiation.
4 | DISCUSSION
Osx is an essential transcription factor governing the os￾teogenic program and the process of bone formation.10,15
Therefore, understanding the upstream signaling that regu￾lates Osx protein stability is necessary to provide a potential
avenue for bone-regeneration therapies. In this study, we an￾alyzed molecular mechanisms through which p38 and Fbw7
cooperatively promote Osx ubiquitination and degradation.
We found that p38 is the priming kinase responsible for the
interaction between Fbw7 binding and subsequent Osx deg￾radation (Figure 3). Pharmacological inhibition of p38 fol￾lowing osteogenic stimulation resulted in Osx stabilization
and enhanced osteoblast differentiation (Figures 4 and 5).
Furthermore, Fbw7-knockdown MSCs and Fbw7-knockout
primary mouse calvarial cells displayed enhanced osteoblast
formation (Figure 6). Our findings demonstrate that there is
precise coordination between p38 and Fbw7 in the regulation
of osteoblast differentiation. In accordance with these find￾ings, ectopic expression of the nondegradable Osx-S73/77A
mutant augmented osteoblast differentiation (Figure 7), sug￾gesting the use of MSCs expressing nondegradable Osx as a
potential therapeutic approach for efficient bone regeneration.
The current study shows that p38-mediated phosphoryla￾tion negatively regulates Osx protein stability; however, p38
plays a positive role in promoting osteoblast gene expression
during the early stage of differentiation. Our data show that
the distinct timing of p38 inhibitor treatment results in oppo￾site effects on Osx protein levels and osteoblast differentia￾tion, unveiling the previously unidentified p38 function. One
possible reason for this coupling is the necessity for a safe￾guard switch for preventing Osx overactivation. Given the
crucial roles of Osx in osteoblast differentiation, increased
Osx activity needs to be quickly downregulated to maintain
the integrity of osteoblastic gene expression. Another con￾ceivable reason is for the induction of pulsatile gene expres￾sion of Osx downstream targets. Regulatory repetition of Osx
activation and degradation might lead to fine-tuning of the
proper extent and duration of target gene expression and acti￾vation of the downstream transcriptional cascade.
Previous studies indicated that Osx is subjected to prote￾asome-dependent degradation.41-43 The carboxyl terminus of
Hsp70-interacting protein (CHIP), an E3 ligase, is reported
to negatively regulate osteoblast differentiation through deg￾radation of several osteogenic transcription factors, includ￾ing Osx, Runx2, and Smad1/542. Besides, Cbl-b/c-Cbl also
promotes Osx degradation.43 Cbl interacts with its binding
partners via a conserved phospho-Tyr binding domain,44 im￾plying that the Cbl-mediated Osx ubiquitination may also be
regulated in a phosphorylation-dependent manner. Although
the precise regulatory mechanisms of CHIP and Cbl during
osteogenesis remain to be defined, it is possible that the CHIP
and Cbl signaling pathways interplay with the p38/Fbw7
pathway for efficient Osx degradation in a coordinated and
orchestrated manner. Our data demonstrated that treatments
FIGURE 5 p38 regulates Osx through inducing and downregulating its protein abundance following osteogenic stimulation. A, Time course
of p38 inhibitor pretreatment/posttreatment and osteogenic induction for the experiments. B, IB analysis of WCLs derived from UE7T-13 cells.
Cells were pretreated with the selective p38 inhibitor SB239063 or SB203580 (10 μM) for 12 hours, then, exposed to osteogenic stimulation by
the replacement with osteoblast differentiation medium containing the inhibitor, and cultured for the time periods indicated before harvesting. C,
Real-time RT-PCR analysis to determine the relative mRNA expression levels of Osx in p38 inhibitor-treated UE7T-13 cells presented in (B).
Data are presented as mean ± SEM values (n = 3). **P < .01; one-way ANOVA with multiple-comparison test. D, IB analysis of WCLs derived
from UE7T-13 cells. Cells were cultured in osteoblast differentiation medium for 12 hours, serially treated with SB239063 or SB203580 (10 μM),
and cultured for the time periods indicated before harvesting. E, Alkaline phosphatase (ALP) staining of UE7T-13. Cells were pretreated with
SB239063 or SB203580 (10 μM) for 12 hours, and then, cultured in osteoblast differentiation medium containing the inhibitors for 7 days. DMSO
was used as the control. The cells were fixed and stained for ALP. F, Quantification of ALP activity in the UE7T-13 cells presented in (E). Data
are presented as mean ± SEM values (n = 3). **P < .01; one-way ANOVA with multiple-comparison test. G, Alizarin Red S staining of UE7T-13.
Cells were pretreated with SB239063 or SB203580 (10 μM) for 12 hours, and then, cultured in osteoblast differentiation medium containing the
inhibitors for 2 weeks. DMSO was used as the control. Cells were fixed and stained with Alizarin Red S. H, Quantification of calcium deposition
in the UE7T-13 cells presented in (G). Data are presented as mean ± SEM (n = 3). **P < .01: one-way ANOVA with multiple-comparison test. I,
Real-time RT-PCR analysis to determine the relative mRNA expression levels of OCN in the UE7T-13 cells presented in (E). Data are presented
as mean ± SEM values (n = 3). *P < .05; one-way ANOVA with multiple-comparison test. J, Alkaline phosphatase (ALP) staining of UE7T-13.
Cells were cultured in osteoblast differentiation medium for 12 hours and serially treated with SB239063 or SB203580 (10 μM) for 7 days. DMSO
was used as the control. The cells were fixed and stained for ALP. K, Quantification of ALP activity in the UE7T-13 cells presented in (J). Data
are presented as mean ± SEM values (n = 3). *P < .05, **P < .01; one-way ANOVA with multiple-comparison test. L, Alizarin Red S staining of
UE7T-13. Cells were cultured in osteoblast differentiation medium for 12 hours and serially treated with SB239063 or SB203580 (10 μM) for 2
weeks. DMSO was used as the control. Cells were fixed and stained with Alizarin Red S. M, Quantification of calcium deposition in the UE7T-13
cells presented in (L). Data are presented as mean ± SEM (n = 3). **P < .01; one-way ANOVA with multiple-comparison test. N, Real-time RT￾PCR analysis to determine the relative mRNA expression levels of OCN in the UE7T-13 cells presented in (J). Data are presented as mean ± SEM
values (n = 3). *P < .05; one-way ANOVA with multiple-comparison test
| HOSHIKAWA et al. 13
FIGURE 6 Fbw7 depletion promotes osteoblast differentiation (See also Figure S3). A, Immunoblot (IB) analysis of whole-cell lysates
(WCLs) derived from UE7T-13 cells infected with shRNA lentiviral vectors specific for Fbw7. The empty vector was used as the control. The
cells were selected for 7 days in the presence of puromycin (1 μg/mL) to remove noninfected cells, and then, cultured in osteoblast differentiation
medium for the indicated time points and harvested for IB analysis. B, Alkaline phosphatase (ALP) staining of UE7T-13 cells presented in (A).
C, Quantification of ALP activity in UE7T-13 cells presented in (B). Data are presented as mean ± SEM values (n = 3). **P < .01; Student’s t
test. D, Alizarin Red S staining of UE7T-13 cells presented in (A). E, Quantification of calcium deposition in UE7T-13 cells presented in (D).
Data are presented as mean ± SEM values (n = 3). **P < .01; Student’s t test. F, Alkaline phosphatase (ALP) staining of primary mouse calvarial
cells isolated from control and Fbw7-knockout mice. Cells were cultured in osteoblast differentiation medium for 7 days and fixed for staining.
G, Quantification of ALP activity in primary mouse calvarial cells presented in (F). Data are presented as mean ± SEM values (n = 3). **P < .01;
Student’s t test. H, Alizarin Red S staining of primary mouse calvarial cells isolated from control and Fbw7-knockout mice. Cells were cultured
in osteoblast differentiation medium for 2 weeks and fixed for staining. I, Quantification of calcium deposition in primary mouse calvarial cells
presented in (H). Data are presented as mean ± SEM values (n = 3). **P < .01; Student’s t test. J, Real-time RT-PCR analysis to determine the
relative mRNA expression levels of ALP, OCN, and Osx in primary mouse calvarial cells isolated from control and Fbw7-knockout mice. Cells
were cultured in osteoblast differentiation medium for 7 days and harvested for RNA extraction. Data are presented as mean ± SEM values (n =
3). *P < .05; Student’s t test. K, IB analysis of WCLs derived from primary mouse calvarial cells isolated from control and Fbw7-knockout mice.
Cells were cultured in osteoblast differentiation medium for 7 days and harvested for IB analysis
14 |
HOSHIKAWA et al.
with bortezomib and ixazomib at minimal concentrations re￾sulted in a marked increase in the Osx protein level. At the
same time, other osteogenic transcription factors, including
Runx2, Smad1, and β-catenin, were not affected under our
experimental conditions (Figure 1). This finding implies that
the regulation of Osx activity may highly depend on prote￾asome-dependent signaling, like p53, the protein stability
of which is tightly controlled by multiple E3s.45 Since the
biological significance of multiple E3-mediated Osx deg￾radation processes is currently unknown, future studies are
warranted to elucidate how the osteogenic E3 signaling net￾work regulates Osx activity in a concerted, context-depen￾dent, and stage-specific manner.
In this study, we discovered that treatments with a minimal
concentration of bortezomib and ixazomib led to efficient os￾teoblast differentiation in osteogenic cell lines accompanied
by a remarkable enhancement of Osx protein levels (Figure
1). The results suggest that the synergistic induction of Osx
protein was likely due to its posttranslational regulation.
Mechanistically, p38-mediated Osx phosphorylation trig￾gered Fbw7-directed ubiquitination and proteasomal degra￾dation of Osx. In agreement with these results, p38 inhibitor
treatments following osteogenic induction or Fbw7 ablation
promoted Osx stabilization and elevated osteoblast differen￾tiation. These findings suggest that targeting the p38/Fbw7
signaling pathway would provide a promising approach for
tissue engineering that would take advantage of MSC im￾plantation and bone matrix regeneration for various bone-re￾lated disorders.
ACKNOWLEDGMENTS
We thank Dr. Keiko Nakayama for providing floxed Fbw7
mice. We thank Drs. Masahiro Saito and Hiroshi Egusa for
helpful discussion. This study was supported by JSPS Grants￾in-Aid for Scientific Research (17H01606 and 19H03834)
and Grant-in-Aid for JSPS Fellows (19J11773).
FIGURE 7 Nondegradable Osx-S73/77A mutant enhances osteoblast differentiation in Saos-2 cells. A, Immunoblot (IB) analysis of
whole-cell lysates (WCLs) derived from CRISPR/Cas9-mediated Osx-knockout Saos-2 cells reexpressing Flag-Osx (WT or S73/77A) using
lentiviral vectors. Empty vectors (EV) were used as control. The cells were selected for 2 weeks in the presence of hygromycin (100 μg/mL)
to remove noninfected cells, and then, treated with potassium dihydrogen phosphate (1.8 mM) and L-ascorbic acid (50 μg/mL) for 2 weeks to
induce differentiation. B, Alizarin Red S staining of the Osx-knockout Saos-2 cells reexpressing Flag-Osx (WT or S73/77A) presented in (A). C,
Quantification of calcium deposition in Osx-knockout Saos-2 cells reexpressing Flag-Osx (WT or S73/77A) presented in (B). Data are presented as
mean ± SEM values (n = 3). **P < .01; one-way ANOVA with multiple-comparison test. D, Real-time RT-PCR analysis to determine the relative
mRNA expression levels of OCN in the Osx-knockout Saos-2 cells reexpressing Flag-Osx (WT or S73/77A) presented in (A). The cells were
cultured with potassium dihydrogen phosphate (1.8 mM) and L-ascorbic acid (50 μg/mL) for 2 weeks and harvested for RNA extraction. Data are
presented as mean ± SEM values (n = 3). **P < .01; one-way ANOVA with multiple-comparison test
| HOSHIKAWA et al. 15
CONFLICT OF INTEREST
The authors declare no competing interests.
AUTHOR CONTRIBUTIONS
SH performed the experiments with assistance from KSh,
AW, MC, KSa, and HI; SH and HI designed the experi￾ments; WW, SF, and HI conceived and supervised the study;
SF and HI wrote the manuscript; and all authors discussed
the manuscript.
REFERENCES
1. Komander D, Rape M. The ubiquitin code. Annu Rev Biochem.
2012;81:203-229.
2. Huang X, Dixit VM. Drugging the undruggables: exploring the
ubiquitin system for drug development. Cell Res. 2016;26:484-498.
3. Park JE, Miller Z, Jun Y, Lee W, Kim KB. Next-generation protea￾some inhibitors for cancer therapy. Transl Res. 2018;198:1-16.
4. Pennisi A, Li X, Ling W, Khan S, Zangari M, Yaccoby S. The pro￾teasome inhibitor, bortezomib suppresses primary myeloma and
stimulates bone formation in myelomatous and nonmyelomatous
bones in vivo. Am J Hematol. 2009;84:6-14.
5. Mukherjee S, Raje N, Schoonmaker JA, et al. Pharmacologic
targeting of a stem/progenitor population in vivo is associ￾ated with enhanced bone regeneration in mice. J Clin Invest.
2008;118:491-504.
6. Oyajobi BO, Garrett IR, Gupta A, et al. Stimulation of new bone
formation by the proteasome inhibitor, bortezomib: implications
for myeloma bone disease. Br J Haematol. 2007;139:434-438.
7. Giuliani N, Morandi F, Tagliaferri S, et al. The proteasome inhibi￾tor bortezomib affects osteoblast differentiation in vitro and in vivo
in multiple myeloma patients. Blood. 2007;110:334-338.
8. Garcia-Gomez A, Quwaider D, Canavese M, et al. Preclinical ac￾tivity of the oral proteasome inhibitor MLN9708 in Myeloma bone
disease. Clin Cancer Res. 2014;20:1542-1554.
9. Hurchla MA, Garcia-Gomez A, Hornick MC, et al. The epoxyke￾tone-based proteasome inhibitors carfilzomib and orally bioavail￾able oprozomib have anti-resorptive and bone-anabolic activity in
addition to anti-myeloma effects. Leukemia. 2013;27:430-440.
10. Karsenty G, Kronenberg HM, Settembre C. Genetic control of
bone formation. Annu Rev Cell Dev Biol. 2009;25:629-648.
11. Nakashima K, Zhou X, Kunkel G, et al. The novel zinc finger-con￾taining transcription factor osterix is required for osteoblast differ￾entiation and bone formation. Cell. 2002;108:17-29.
12. Koga T, Matsui Y, Asagiri M, et al. NFAT and Osterix coopera￾tively regulate bone formation. Nat Med. 2005;11:880-885.
13. Ortuno MJ, Ruiz-Gaspa S, Rodriguez-Carballo E, et al. p38 regu￾lates expression of osteoblast-specific genes by phosphorylation of
osterix. J Biol Chem. 2010;285:31985-31994.
14. Hojo H, Ohba S, He X, Lai LP, McMahon AP. Sp7/osterix is re￾stricted to bone-forming vertebrates where it acts as a Dlx co-factor
in osteoblast specification. Dev Cell. 2016;37:238-253.
15. Sinha KM, Zhou X. Genetic and molecular control of osterix in
skeletal formation. J Cell Biochem. 2013;114:975-984.
16. Rodriguez-Carballo E, Gamez B, Ventura F. p38 MAPK signaling
in osteoblast differentiation. Front Cell Dev Biol. 2016;4:40.
17. Thouverey C, Caverzasio J. Focus on the p38 MAPK signaling
pathway in bone development and maintenance. Bonekey Rep.
2015;4:711.
18. Greenblatt MB, Shim JH, Zou W, et al. The p38 MAPK pathway is
essential for skeletogenesis and bone homeostasis in mice. J Clin
Invest. 2010;120:2457-2473.
19. Ge C, Yang Q, Zhao G, Yu H, Kirkwood KL, Franceschi RT.
Interactions between extracellular signal-regulated kinase 1/2
and p38 MAP kinase pathways in the control of RUNX2 phos￾phorylation and transcriptional activity. J Bone Miner Res.
2012;27:538-551.
20. Ulsamer A, Ortuno MJ, Ruiz S, et al. BMP-2 induces Osterix ex￾pression through up-regulation of Dlx5 and its phosphorylation by
p38. J Biol Chem. 2008;283:3816-3826.
21. Artigas N, Urena C, Rodriguez-Carballo E, Rosa JL, Ventura F.
Mitogen-activated protein kinase (MAPK)-regulated interactions
between Osterix and Runx2 are critical for the transcriptional os￾teogenic program. J Biol Chem. 2014;289:27105-27117.
22. Davis RJ, Welcker M, Clurman BE. Tumor suppression by the
Fbw7 ubiquitin ligase: mechanisms and opportunities. Cancer
Cell. 2014;26:455-464.
23. Shimizu K, Nihira NT, Inuzuka H, Wei W. Physiological functions
of FBW7 in cancer and metabolism. Cell Signal. 2018;46:15-22.
24. Yumimoto K, Matsumoto M, Onoyama I, Imaizumi K, Nakayama
KI. F-box and WD repeat domain-containing-7 (Fbxw7) protein
targets endoplasmic reticulum-anchored osteogenic and chon￾drogenic transcriptional factors for degradation. J Biol Chem.
2013;288:28488-28502.
25. Fukushima H, Shimizu K, Watahiki A, et al. NOTCH2 Hajdu￾Cheney mutations escape SCF(FBW7)-dependent proteolysis to
promote osteoporosis. Mol Cell. 2017;68:645-658 e645.
26. Rajagopalan H, Jallepalli PV, Rago C, et al. Inactivation of hCDC4
can cause chromosomal instability. Nature. 2004;428:77-81.
27. Terai M, Uyama T, Sugiki T, Li XK, Umezawa A, Kiyono T.
Immortalization of human fetal cells: the life span of umbili￾cal cord blood-derived cells can be prolonged without ma￾nipulating p16INK4a/RB braking pathway. Mol Biol Cell.
2005;16:1491-1499.
28. Aomatsu E, Takahashi N, Sawada S, et al. Novel SCRG1/BST1
axis regulates self-renewal, migration, and osteogenic differentia￾tion potential in mesenchymal stem cells. Sci Rep. 2014;4:3652.
29. Longo PA, Kavran JM, Kim MS, Leahy DJ. Transient mammalian
cell transfection with polyethylenimine (PEI). Methods Enzymol.
2013;529:227-240.
30. Gao D, Inuzuka H, Tan MK, et al. mTOR drives its own activation
via SCF(betaTrCP)-dependent degradation of the mTOR inhibitor
DEPTOR. Mol Cell. 2011;44:290-303.
31. Inuzuka H, Shaik S, Onoyama I, et al. SCF(FBW7) regulates cellu￾lar apoptosis by targeting MCL1 for ubiquitylation and destruction.
Nature. 2011;471:104-109.
32. Gasser JA, Inuzuka H, Lau AW, Wei W, Beroukhim R, Toker A.
SGK3 mediates INPP4B-dependent PI3K signaling in breast can￾cer. Mol Cell. 2014;56:595-607.
33. Inuzuka H, Tseng A, Gao D, et al. Phosphorylation by casein ki￾nase I promotes the turnover of the Mdm2 oncoprotein via the
SCF(beta-TRCP) ubiquitin ligase. Cancer Cell. 2010;18:147-159.
34. Shimizu K, Fukushima H, Ogura K, et al. The SCFbeta-TRCP
E3 ubiquitin ligase complex targets Lipin1 for ubiquitination and
degradation to promote hepatic lipogenesis. Sci Signal. 2017;10.
eaah4117
35. Deshaies RJ, Joazeiro CA. RING domain E3 ubiquitin ligases.
Annu Rev Biochem. 2009;78:399-434.
16 |
HOSHIKAWA et al.
36. Soucy TA, Smith PG, Milhollen MA, et al. An inhibitor of
NEDD8-activating enzyme as a new approach to treat cancer.
Nature. 2009;458:732-736.
37. Maser RS, Choudhury B, Campbell PJ, et al. Chromosomally un￾stable mouse tumours have genomic alterations similar to diverse
human cancers. Nature. 2007;447:966-971.
38. Wang Z, Liu P, Inuzuka H, Wei W. Roles of F-box proteins in can￾cer. Nat Rev Cancer. 2014;14:233-247.
39. Kitagawa K, Shibata K, Matsumoto A, et al. Fbw7 targets GATA3
through cyclin-dependent kinase 2-dependent proteolysis and
contributes to regulation of T-cell development. Mol Cell Biol.
2014;34:2732-2744.
40. O’Neil J, Grim J, Strack P, et al. FBW7 mutations in leukemic
cells mediate NOTCH pathway activation and resistance to gam￾ma-secretase inhibitors. J Exp Med. 2007;204:1813-1824.
41. Peng Y, Shi K, Wang L, et al. Characterization of Osterix protein
stability and physiological role in osteoblast differentiation. PLoS
One. 2013;8:e56451.
42. Xie J, Gu J. Identification of C-terminal Hsp70-interacting protein
as a mediator of tumour necrosis factor action in osteoblast dif￾ferentiation by targeting osterix for degradation. J Cell Mol Med.
2015;19:1814-1824.
43. Choi YH, Han Y, Lee SH, et al. Cbl-b and c-Cbl negatively regu￾late osteoblast differentiation by enhancing ubiquitination and deg￾radation of Osterix. Bone. 2015;75:201-209.
44. Meng W, Sawasdikosol S, Burakoff SJ, Eck MJ. Structure of the
amino-terminal domain of Cbl complexed to its binding site on
ZAP-70 kinase. Nature. 1999;398:84-90.
45. Lee JT, Gu W. The multiple levels of regulation by p53 ubiquitina￾tion. Cell Death Differ. 2010;17:86-92.
SUPPORTING INFORMATION
Additional supporting information may be found online in
the Supporting Information section.
How to cite this article: Hoshikawa S, Shimizu K,
Watahiki A, et al. Phosphorylation-dependent osterix
degradation negatively regulates osteoblast
differentiation. The FASEB Journal. 2020;00:1–16.

https://doi.org/10.1096/fj.202001340R