Article

Mechanical load regulates bone growth via

periosteal Osteocrin
Graphical abstract Authors
Haruko Watanabe-Takano, Hiroki Ochi,
Ayano Chiba, ..., Hiroki R. Ueda,
Akihiro Yasoda, Naoki Mochizuki

Correspondence
takano_h@ncvc.go.jp (H.W.-T.),
mochizuki@ncvc.go.jp (N.M.)

In brief
Watanabe-Takano et al. demonstrate that
physiological loading induces Osteocrin
(OSTN) expression in the periosteal
osteoblasts of long bones through
suppression of the Forkhead box protein
O1 transcription factor. OSTN promotes
bone growth through enhancement of
C-type natriuretic peptide signaling,
leading to elongation and appositional
growth of long bones.

Highlights
d OSTN from periosteum regulates endochondral and
intramembranous ossification

d OSTN enhances osteogenesis through potentiation of CNP

signaling

d FoxO1 suppresses periosteal expression of OSTN under

unloading condition

d OSTN is required for physiological load-induced bone gain

Watanabe-Takano et al., 2021, Cell Reports 36, 109380

July 13, 2021 ª 2021 The Author(s).
https://doi.org/10.1016/j.celrep.2021.109380 ll
 ll
OPEN ACCESS

Article
Mechanical load regulates bone growth
via periosteal Osteocrin
Haruko Watanabe-Takano,1,* Hiroki Ochi,3 Ayano Chiba,1 Ayaka Matsuo,5 Yugo Kanai,6 Shigetomo Fukuhara,7 Naoki Ito,8
Keisuke Sako,1 Takahiro Miyazaki,1 Kazuki Tainaka,9 Ichiro Harada,10 Shingo Sato,11 Yasuhiro Sawada,1,3,4
Naoto Minamino,5 Shu Takeda,12 Hiroki R. Ueda,2,13,14 Akihiro Yasoda,15 and Naoki Mochizuki1,2,16,*
1Department of Cell Biology, National Cerebral and Cardiovascular Center Research Institute, 6-1 Kishibe-shimmachi, Suita, Osaka

564-8565, Japan
2CREST, Japan Science and Technology Agency, 4-1-8 Honcho, Kawaguchi, Saitama 332-0012, Japan
3Department of Clinical Research, National Rehabilitation Center for Persons with Disabilities, 4-1 Namiki, Tokorozawa, Saitama 359-8555,

Japan
4Department of Rehabilitation for Motor Functions, National Rehabilitation Center for Persons with Disabilities, 4-1 Namiki, Tokorozawa,

Saitama 359-8555, Japan

5Omics Research Center, National Cerebral and Cardiovascular Center Research Institute, 6-1 Kishibe-shinmachi, Suita, Osaka 564-8565,

Japan
6Department of Diabetes, Endocrinology and Nutrition, Graduate School of Medicine and Faculty of Medicine, Kyoto University,

Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan

7Department of Molecular Pathophysiology, Institute of Advanced Medical Sciences, Nippon Medical School, 1-1-5 Sendagi, Bunkyo-ku,

Tokyo 113-8602, Japan

8Laboratory of Molecular Life Science, Institute of Biomedical Research and Innovation, Foundation for Biomedical Research and Innovation

at Kobe, 6-7-6 Minatojima-Minamimachi, Chuo-ku, Kobe 650-0047, Japan

9Department of System Pathology for Neurological Disorders, Center for Bioresources, Brain Research Institute, Niigata University, 1-757

Asahimachidori, Chuo-ku, Niigata 951-8585, Japan

10Medical Products Technology, Development Center, R&D headquarters, Canon Inc., 3-30-2, Shimomaruko, Ohta-ku, Tokyo 146-8501,

Japan
11Center for Innovative Cancer Treatment, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8519, Japan
12Division of Endocrinology, Toranomon Hospital Endocrine Center, 2-2-2 Toranomon, Minato-ku, Tokyo 105-8470, Japan
13Department of Systems Pharmacology, Graduate School of Medicine, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033,

Japan
14Laboratory for Synthetic Biology, RIKEN Center for Biosystems Dynamics Research, 1-3 Yamadaoka, Suita, Osaka 565-0871, Japan
15Clinical Research Center, National Hospital Organization Kyoto Medical Center, 1-1 Fukakusa-Mukaihatacho, Fushimi-ku, Kyoto 612-8555,

Japan
16Lead contact

*Correspondence: takano_h@ncvc.go.jp (H.W.-T.), mochizuki@ncvc.go.jp (N.M.)

https://doi.org/10.1016/j.celrep.2021.109380

SUMMARY

Mechanical stimuli including loading after birth promote bone growth. However, little is known about how
mechanical force triggers biochemical signals to regulate bone growth. Here, we identified a periosteal-oste-
oblast-derived secretory peptide, Osteocrin (OSTN), as a mechanotransducer involved in load-induced long
bone growth. OSTN produced by periosteal osteoblasts regulates growth plate growth by enhancing C-type
natriuretic peptide (CNP)-dependent proliferation and maturation of chondrocytes, leading to elongation of
long bones. Additionally, OSTN cooperates with CNP to regulate bone formation. CNP stimulates osteogenic
differentiation of periosteal osteoprogenitors to induce bone formation. OSTN binds to natriuretic peptide re-
ceptor 3 (NPR3) in periosteal osteoprogenitors, thereby preventing NPR3-mediated clearance of CNP and
consequently facilitating CNP-signal-mediated bone growth. Importantly, physiological loading induces
Ostn expression in periosteal osteoblasts by suppressing Forkhead box protein O1 (FoxO1) transcription fac-
tor. Thus, this study reveals a crucial role of OSTN as a mechanotransducer converting mechanical loading to
CNP-dependent bone formation.

INTRODUCTION ercise, and sleep. Long bones develop through two types of os-
sifications, namely, endochondral ossification and intramembra-
Long bone growth is a highly complicated process regulated by nous ossification, which regulate growth in the length and width
genetic factors and environmental factors, including gravity, ex- of long bones, respectively. Endochondral ossification is the

Cell Reports 36, 109380, July 13, 2021 ª 2021 The Author(s). 1
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
 ll
OPEN ACCESS
process by which chondrocytes of the cartilage template are re-
placed by bone (trabecular bone [TB]), whereas intramembra-
nous ossification is the process by which osteoblasts directly
produce mineralized bone in the inner layer of the periosteum
(PS; cambium layer) (Kozhemyakina et al., 2015; Long and Or-
nitz, 2013; Mackie et al., 2011; Rauch, 2005).
During endochondral ossification, chondrocytes are orga-
nized into the growth plate (GP) located between the epiphysis
and metaphysis and undergo a sequential maturation process
to become hypertrophic chondrocytes that are ultimately re-
placed by bone. A number of studies suggest that several secre-
tory factors such as parathyroid hormone-related protein, Wnt li-
gands, sonic hedgehog, indian hedgehog, bone morphogenetic
proteins, and fibroblast growth factor mediate endochondral
ossification in the GP by eliciting genetically programmed pro-
cesses (Kozhemyakina et al., 2015; Long and Ornitz, 2013;
Mackie et al., 2011). In addition, C-type natriuretic peptide
(CNP), a member of natriuretic peptides (NPs), regulates
endochondral ossification by promoting proliferation and matu-
ration of chondrocytes. CNP stimulates its receptor guanylate
cyclase-B (GC-B)/NP receptor-B (NPR-B)/NP receptor-2
(NPR-2) to produce cyclic GMP (cGMP), leading to activation
of cGMP-dependent protein kinases (cGKs) (Potter et al.,
2006). The mice lacking CNP, GC-B, or cGKII exhibited short
stature, whereas mice overexpressing CNP exhibited skeletal
overgrowth (Chusho et al., 2001; Kawasaki et al., 2008; Nakao
et al., 2015; Tsuji and Kunieda, 2005; Yasoda et al., 2004). These
findings suggest an essential role for a CNP/GC-B/cGKII
signaling pathway in endochondral ossification. Moreover, the
mice deficient for NP receptor 3 (NPR3)/NP receptor C
(NPR-C), a clearance receptor for NPs, show bone elongation,
suggesting that proper formation of long bone requires tight
regulation of CNP signaling by NPR3-mediated clearance of
CNP (Jaubert et al., 1999; Matsukawa et al., 1999).
During intramembranous ossification, skeletal stem cells in the
PS contribute to appositional growth by differentiating into bone-
forming osteoblasts to increase bone thickness (Debnath et al.,
2018; Duchamp de Lageneste et al., 2018; Gao et al., 2019; Or-
tinau et al., 2019). It has been reported that systemic factors,
such as estrogen and parathyroid hormone, regulate differentia-
tion and proliferation of periosteal osteoblast progenitors (Ogita
et al., 2008). In addition, local factors such as Wnt, hedgehog,
and chemokines have been known to be expressed in PS-
derived cells (PDCs) (Kim et al., 2007; Stich et al., 2008; Wang
et al., 2010). Thus, differentiation of skeletal stem cells in the
PS into osteoblasts might be regulated by genetic cues including
both systemic and local secretory factors.
In addition to those biochemical factors, mechanical force
engendered by gravitational change, physical activities, and ex-
ercise is also an important determinant that controls bone
growth. Moderate compression and tension to bones promote
chondral growth through triggering chondrocyte differentiation
(Elder et al., 2001; Ohashi et al., 2002; Rauch, 2005; Wang
and Mao, 2002). Furthermore, adding bone in width has been
thought to accommodate local mechanical demands to resist
chronic loading (Frost, 2003; Hart et al., 2017; Iolascon et al.,
2013; Rauch, 2005). However, little is known about how me-
chanical factors trigger genetically programmed biochemical
2 Cell Reports 36, 109380, July 13, 2021
Article
signaling to induce elongation and appositional growth of
bones.
Osteocrin (Ostn), also known as Musclin, encodes a secretory
protein that is expressed in bone and skeletal muscles (Nishi-
zawa et al., 2004; Thomas et al., 2003). OSTN contains two re-
gions with high homology to NPs. Although OSTN binds to
neither GC-A nor GC-B due to the lack of conserved cysteine
ring (Moffatt et al., 2007), it specifically binds to NPR3 with a
similar affinity to NPs (Kita et al., 2009). Accordingly, OSTN com-
petes with NPs for binding to NPR3, thereby increasing NP avail-
ability by protecting them from NPR3-dependent degradation
(Kanai et al., 2017; Kita et al., 2009; Miyazaki et al., 2018; Moffatt
et al., 2007; Subbotina et al., 2015). Previously, we and others
have reported that overexpression of OSTN in transgenic mice
induces an increase in bone length by enhancing CNP-signal-
mediated proliferation and maturation of chondrocytes (Kanai
et al., 2017; Moffatt et al., 2007). Our previous report on ostn
mutant fish reveals that OSTN is also involved in bone and carti-
lage development in zebrafish (Chiba et al., 2017).
In this study, we demonstrate a crucial role of OSTN as a me-
chanotransducer involved in load-induced growth of long bone.
Physiological loading induces Ostn expression in the periosteal
osteoblasts of long bones through suppression of the Forkhead
box protein O1 (FoxO1) transcription factor. Periosteal-osteo-
blast-derived OSTN promotes not only developmental bone
growth but also bone recovery from unloading, possibly through
enhancement of CNP signaling. Thus, OSTN participates in the
regulation of mechanical-load-induced long bone growth.
RESULTS
OSTN is expressed in the PS of bones
To examine the expression patterns of OSTN, we generated
knockin mice by using the targeting vector obtained from the
Knockout Mouse Project (KOMP) repository (Figure S1A). We
used heterozygous mice (Ostn+/LacZ) to detect LacZ activity as
an indicator of Ostn promoter activity. Five-bromo-4-chloro-3-
indolyl b-D-galactopyranoside (X-gal) staining in the whole
cleared body revealed that Ostn expression in some bones
started at embryonic day 15.5 (E15.5) (Figure 1A). Ostn expres-
sion became clear at 30 days after birth (postnatal day 30
[P30]). Its expression was higher in long bones (zygomatic
bones, mandible, tibias, metacarpal bones, phalanges, caudal
vertebrae, ulnae, radii, and ribs) as well as flat bones (nasal
bones, frontal bones, parietal bones, occipital bones, and scap-
ulas) than in vertebrae, femora, and humeri. It is of note that the
expression of Ostn tended to be concentrated in the distal bones
tibiae, radius, and ulna, suggesting a unique upstream signal that
induces Ostn expression.
At the age of 3 months, Ostn expression seemed to be entirely
downregulated even in the bones in which Ostn expression was
apparent at P30. To confirm whether LacZ-positive staining re-
flects the endogenous Ostn expression, we conducted qPCR
analyses by using the tissues isolated from 8-week-old mice.
Consistent with the results of the X-gal staining, Ostn expression
in tibia and femur was significantly higher than that in other tis-
sues (Figures 1B and S1B). X-gal staining images of the tibia
at the various developmental stages demonstrated that Ostn
 ll
Article OPEN ACCESS

B C D

E F G

(legend on next page)

Cell Reports 36, 109380, July 13, 2021 3

 ll
OPEN ACCESS Article

expression increased until P30 and decreased by aging, sug-
gesting its biological significance in bone morphogenesis (Fig-
ure 1C). The expression of Ostn was site specific and likely to
be higher on the convex side of tibia than on the concave side.
Detailed examination revealed the characteristic expression of
Ostn in the anteromedial part but not in the posterolateral part
of the tibia (Figures 1D and S1C).
To define where Ostn-expressing cells reside in the tibia, we
histologically examined the expression in X-gal-stained sec-
tions (Figure 1E). It is of note that the expression of Ostn was
high in the PS of the developing tibia from E15.5 to P30 (Figures
1E and S1D). The periosteal expression of Ostn was high in
other bones such as interparietal bone and caudal vertebra
(Figure S1D). On the other hand, any X-gal staining was not de-
tected in the GP or endosteum (Figure 1E). To validate whether
X-gal staining faithfully reflects the endogenous Ostn expres-
sion, we fractionated the PS, bone marrow (BM), and trabec-
ular/cortical bone (TB/CB) from the tibia, which was verified
by the expression of specific marker genes, and analyzed
Ostn expression in each fraction by qPCR (Figure S1E). The
expression of Ostn mRNA was more in PS than that in BM
and TB/CB (Figure S1E). We further confirmed the periosteal
expression of Ostn by qPCR using laser-microdissected PS,
BM, and GP from the tibia (Figures 1F and S1F). The qPCR
data revealed a dominant expression of Ostn in the PS over
the BM and GP (Figure 1F).
To further characterize the Ostn-expressing cells, we isolated
the cells exhibiting high LacZ activity by using fluorescence
activated cell sorting (FACS) and analyzed the expression of
marker genes specific for individual cell types by qPCR (Figures
1G and S1G). We found that Ostn-expressing cells belong to a
certain population of PDCs and that specific markers for oste-
oblasts were highly expressed in Ostn-expressing cells. In
contrast, the expression of markers for other cell types,
including fibroblasts, endothelial cells, chondrocytes, adipo-
cytes, osteoclasts, and mesenchymal cells, were low or com-
parable in LacZ-positive cells compared with that in LacZ-
negative cells. Furthermore, immunostaining with the antibody
for Osterix, a specific marker for osteoblasts, revealed that
the LacZ-positive cells partially overlapped with Osterix-posi-
tive cells (Figure S1H). These results suggest that Ostn is likely
to be expressed in certain populations of periosteal osteoblasts
of distal bones.
OSTN regulates endochondral ossification
Previously, we and others have reported that overexpression of
OSTN in transgenic mice promotes CNP-signal-mediated prolif-
eration and maturation of chondrocytes to elongate bone length
(Kanai et al., 2017; Moffatt et al., 2007). Thus, we investigated
whether endogenous OSTN is required for physiological bone
growth by generating homozygous mutant mice (OstnLacZ/LacZ
mice). Analyses using qPCR and chemiluminescent enzyme
immunoassay (CLEIA) revealed the significant reduction of
OSTN in the OstnLacZ/LacZ mice as compared with that in the
wild-type mice (Ostn+/+ mice) (Figures 2A and 2B). Deficiency
of Ostn in OstnLacZ/LacZ mice was confirmed by the lack of
Ostn mRNA revealed by in situ hybridization (ISH) (Figures S2A
and S2B). Consistent with the results of X-gal staining of
Ostn+/LacZ mice and qPCR analyses, Ostn mRNA was detected
in the PS of Ostn+/LacZ mice but not in that of OstnLacZ/LacZ
mice. No or subtle Ostn mRNA expression was detected in the
GP, BM, and the region of TB of the Ostn+/LacZ mice, consistent
with Figures 1F and S1F. These results indicate that Ostn is
depleted in the OstnLacZ/LacZ mice and that X-gal staining in the
Ostn+/LacZ mice recapitulates the expression of endogenous
Ostn.
We assessed whole-body imaging of skeletal systems
scanned by computed tomography (CT) (Figures 2C and 2D).
The 8-week-old OstnLacZ/LacZ mice exhibited slight but signifi-
cant shortening of some long bones. The length of femora, tibiae,
and caudal vertebrae except for carvarie of the OstnLacZ/LacZ
mice was significantly shorter than those of the Ostn+/+ mice,
suggesting that OSTN is required for bone elongation.
To elucidate whether OSTN is involved in chondral growth, we
examined the GP of 3-, 5-, or 8-week-old OstnLacZ/LacZ mice by
histological analyses. As shown in Figure 2E, the OstnLacZ/LacZ
mice had thinner GPs than the control mice. Quantitative ana-
lyses indicated a significant reduction of GP thickness of the
OstnLacZ/LacZ mice as compared with the Ostn+/+ mice at the
ages of both 3 and 8 weeks (Figure 2F). Furthermore, we exam-
ined the proliferation of chondrocytes in vivo by the incorporation
of 5-ethynyl-20 -deoxyuridine (EdU) (Figure S2C). Notably, the
number of EdU-positive cells of the OstnLacZ/LacZ mice was lower
than that of the Ostn+/+ mice (Figure 2G). The number of EdU-
positive cells of both mice decreased gradually along with the
progression of skeletal development. Furthermore, to evaluate
chondrocyte maturation in the GP, the expression of Col10a1

Figure 1. OSTN is expressed in the periosteum (PS) of bones

(A) X-gal staining of the cleared whole body of Ostn+/LacZ mice at the stage indicated at the top. E, embryonic day; P, postnatal day.
(B) Ostn mRNA expression in various tissues of 8-week-old Institute of Cancer Research (ICR) male mice. The gene expression value of each sample was
normalized by each Gapdh mRNA value. The relative expression to the value of tibia is indicated as mean with ± standard error of mean (SEM) (n = 4).
(C) X-gal staining of the cleared tibiae of Ostn+/LacZ mice at the stage indicated on the left.
(D) X-gal staining of tibia in Ostn+/LacZ mice at the age of 8 weeks. Note the expression of Ostn in the anteromedial surface of the tibia. Arrowheads show the
convex side of tibia.
(E) Eosin and X-gal staining of the tibia of the Ostn+/LacZ mouse at the age of P30. Note that Ostn is expressed in the PS (red arrowheads) but not in the growth plate
(GP) and endosteum (black arrowheads).
(F) qPCR analyses of Ostn by using laser-microdissected PS, bone marrow (BM), and GP of the tibia of 8-week-old mice. The relative expression is indicated as
mean with ± SEM similarly to (B) (n = 3, each case contains 2 mice).
(G) Marker analysis of LacZ-expressing cells by qPCR. Values represent log10 fold change of the expression levels in LacZ-positive cells relative to those in LacZ-
negative cells as mean with ± SEM (n = 3).
**p < 0.01, *p < 0.05; by one-way ANOVA followed by Turkey-Kramer test in (B) and (F) or by Student’s t test in (G). Scale bars, 0.5 cm in (A) and (C), 2.5 mm in (D),
1 mm in (E, top image), and 250 mm in (E, bottom images). See also Figure S1 and Table S1.

4 Cell Reports 36, 109380, July 13, 2021

 ll
Article OPEN ACCESS

A B C D

F G H
E

Figure 2. OSTN regulates endochondral ossification

(A) qPCR analysis of Ostn expression in the hindlimbs of OstnLacZ/LacZ and Ostn+/+. Relative expression to the value obtained from Ostn+/+ as mean with ± SEM
similarly to Figure 1B (n = 6).
(B) The concentration of OSTN in tibia measured by CLEIA as mean with ± SEM (n = 6).
(C) Computed tomography (CT) images of tibiae (top) and caudal vertebrae (bottom) in 8-week-old OstnLacZ/LacZ and Ostn+/+ mice. Red arrowheads indicate end
of each bone.
(D) Bone (as indicated at the bottom) lengths of OstnLacZ/LacZ and Ostn+/+ mice as mean with ± SEM (n = 15–16). Note the shortened bone length of OstnLacZ/LacZ
mice.
(E) Representative images of hematoxylin and eosin with Alcian blue staining of GPs of the 3-week-old OstnLacZ/LacZ and Ostn+/+ mice. Note the decreased width
of GP in OstnLacZ/LacZ mice (two-way arrows).
(F) The thickness of the GPs of the 3- and 8-week-old OstnLacZ/LacZ and Ostn+/+ mice as mean with ± SEM (n = 6–9).
(G) Percentages of EdU-positive cells per total cells in the GPs of 3- and 5-week-old OstnLacZ/LacZ and Ostn +/+ mice by using sections in Figure S2C as mean with
± SEM (n = 4–6).
(H) Quantitative analyses of the Col10a1(+) area assessed by ISH using the sections shown in Figure S2D. Values represent mean with ± SEM (n = 5).
Scale bars; 100 mm in (E). **p < 0.01, *p < 0.05; by Student’s t test. See also Figure S2 and Table S1.

was assessed by ISH (Figure S2D). We measured thicknesses of
the area where Col10a1-expressing cells were detected and
found a significant reduction of the thickness in the OstnLacZ/LacZ
mice as compared with Ostn+/+ mice at the age of 5 weeks (Fig-
ure 2H). These results indicate that OSTN regulates chondrocyte
proliferation and maturation in the GP, thereby contributing to
chondral growth.
OSTN enhances bone growth by stimulating osteoblast
activity
Because chondrogenesis in the GP eventually leads to ossifica-
tion, we investigated whether OSTN regulates endochondral
ossification. To this end, we quantified the bone microstructure
of tibia using microCT. We observed a marked deficiency of mi-
croarchitecture of TBs at proximal metaphysis of the tibia in the
OstnLacZ/LacZ mice at the age of 8 weeks (Figures 3A and 3B).
Although total bone volume (TV) of the OstnLacZ/LacZ mice was
comparable to that of the Ostn+/+ mice, bone volume (BV),
bone volume fraction (BV/TV), bone surface (BS), TB thickness
(Tb.Th.), and TB number (Tb.N.) were significantly decreased in
the OstnLacZ/LacZ mice, indicating the failure of TB growth in the
OstnLacZ/LacZ mice. To understand the cause of the impaired
TB growth, we immunohistochemically analyzed Osterix expres-
sion and found that staining of Osterix beneath the GP was
reduced in the OstnLacZ/LacZ mice (Figures 3C and 3D). To
analyze the activity of osteoblasts, we further performed bone
morphometric analyses. Although the number of osteoblasts
(N.Ob/BS) was unchanged in the OstnLacZ/LacZ mice, we found

Cell Reports 36, 109380, July 13, 2021 5

 ll
OPEN ACCESS Article

A C D

F
E

Figure 3. OSTN enhances bone growth by stimulating osteoblast activity

(A) Representative microCT images of trabecular bone at the proximal-metaphyseal side of the tibia in the 8-week-old OstnLacZ/LacZ and Ostn+/+ mice.
(B) Measurement of trabecular bone parameters: total bone volume (TV), bone volume (BV), bone volume fraction (BV/TV), bone surface (BS), trabecular bone
thickness (Tb.Th.), and trabecular bone number (Tb.N.) assessed by microCT. All results are expressed as mean with ± SEM (n = 12–14).
(C) Representative images of immunostaining of GPs of 5-week-old OstnLacZ/LacZ and Ostn+/+ mice by using anti-Osterix antibody.
(D) The area of Osterix-positive cells in the GP of 3- and 5-week-old OstnLacZ/LacZ and Ostn+/+ mice as mean with ± SEM (n = 5).
(E) Representative microCT images of cortical bone at the mid-diaphyseal side of the tibia in 8-week-old OstnLacZ/LacZ and Ostn+/+ mice.
(F) Measurement of the following cortical bone parameters: cortical area (Ct.Ar.), total area (Tt.Ar.), and cortical thickness (Ct.Th.). All results are expressed as
mean with ± SEM (n = 12–14).
Scale bar, 1,000 mm in (A) and (E) and 100 mm in (C). **p < 0.01, *p < 0.05; by Student’s t test. See also Figure S3.

the reduced parameters of osteoblast activity including bone-
forming rate (BFR/BS) and mineral apposition rate (MAR) in the
OstnLacZ/LacZ mice (Figure S3A). Consistently, mineralizing (MS/
BS) and osteoid (OS/BS) surfaces in the OstnLacZ/LacZ mice
were decreased and increased, respectively, indicating the
defect of osteoblast. On the other hand, osteoclast-related pa-
rameters, namely, osteoclast number (N.Oc/BS) and eroded sur-
face (ES/BS) in the OstnLacZ/LacZ mice, were comparable to those
of the Ostn+/+ mice. These findings suggest that OSTN potenti-
ates TB growth by stimulating osteoblast activity.
We further investigated the role of OSTN in short axis growth
of bones by imaging a cross-sectional area of the tibia of the
OstnLacZ/LacZ mice. Morphological microCT analyses of tibia re-
vealed that the cortical thickness (Ct.Th.), cortical area (Ct.Ar.),
and other parameters (total area [Tt.Ar.] and Ct.Ar/Tt.Ar.) of
CBs were attenuated in the OstnLacZ/LacZ mice (Figures 3E and
3F). There was a significant reduction in external line length
(ELL) but not in internal line length (ILL), suggesting the functional
deterioration of osteoblast-dependent CB formation in the PS
(Figure S3B). Bone morphometric analyses of CB also showed
a tendency of decreased bone formation rate in the PS (Ps.BFR)
of the OstnLacZ/LacZ mice (Figure S3C). The results of microCT
and bone morphometric analyses imply that OSTN might regu-
late osteoblast-dependent CB formation in the PS.

6 Cell Reports 36, 109380, July 13, 2021

 ll
Article OPEN ACCESS

A B C

D E

Figure 4. OSTN may regulate osteogenic differentiation of PS-derived cells by NPR3

(A) Representative histograms of FACS analysis under the indicated conditions. The x axis and y axis indicate Cy3 intensity and cell number, respectively.
Numbers indicate the percentages of OSTN-binding cells per parent population.
(B and C) Representative alizarin red S (ARS) staining images (B) and quantitative analysis (mean with ± SEM) (C) of the sorted PDCs cultured in growth medium or
differentiation medium for 8 days (n = 3). BD, OSTN bound; NB, OSTN unbound.
(D) qPCR analyses of the expression of receptors for natriuretic peptides (Npr1, Npr2, and Npr3) in fractionated bones: PS, BM, and trabecular bone/cortical bone
(TB/CB). The gene expression value of each sample was normalized by each Gapdh mRNA level. The relative expression to the value of PS is indicated as mean
with ± SEM (n = 4).
(E) qPCR analyses of the expression of Npr2, Npr3, and Ostn in total PDCs and BD and NB PDC fractions. Data are normalized similarly to (D) and indicated as
mean with ± SEM (n = 3).
**p < 0.01, *p < 0.05; by one-way ANOVA followed by Turkey-Kramer test. See also Figure S4 and Table S1.

OSTN may regulate osteogenic differentiation of PDCs
by NPR3
To investigate the mechanism of OSTN-mediated bone forma-
tion, we isolated the PDCs that are expected to contain skeletal
stem cells having the potential to differentiate into osteoblasts,
chondrocytes, and adipocytes (Duchamp de Lageneste et al.,
2018; van Gastel et al., 2012). Biotin-tagged OSTN, but not biotin
alone, increased Cy3 intensity in about 30% of total PDCs (Fig-
ures S4A and 4A). Furthermore, untagged OSTN decreased
the number of PDCs that bound to labeled OSTN, confirming
the binding specificity of biotin-tagged OSTN. Then, we investi-
gated whether OSTN-bound (BD) PDCs have the capability to
differentiate into osteoblast lineage. When PDCs were cultured
under osteogenic differentiation media, Ca2+ deposition as-
sessed by alizarin red S (ARS) staining of the cells originated
from BD PDCs was greater than that of OSTN-unbound (NB)
PDCs (Figures 4B and 4C). These results suggest that BD
PDCs include osteoprogenitors.
Because OSTN binds to NPR3 to regulate NP signaling (Kanai
et al., 2017; Kita et al., 2009; Miyazaki et al., 2018; Moffatt et al.,
2007; Subbotina et al., 2015), we hypothesized that OSTN might
promote osteoblast differentiation by affecting the NP signaling.
To address this hypothesis, we first examined expression of
GC-A (Npr1), GC-B (Npr2), and Npr3 in the fractionated bone
samples (PS, BM, and TB/CB). The cells derived from BM highly
expressed Npr1, whereas those derived from TB/CB and the PS

Cell Reports 36, 109380, July 13, 2021 7

 ll
OPEN ACCESS Article

(legend on next page)

8 Cell Reports 36, 109380, July 13, 2021

 ll
Article OPEN ACCESS

expressed much less Npr1. In contrast, Npr2 and Npr3 were
mainly expressed in the PS (Figure 4D). The biotin-labeled
OSTN failed to bind the Npr3-deficient PDCs (Figure 4A). Consis-
tently, Npr3 was predominantly expressed in BD PDCs, whereas
Ostn was exclusively expressed in NB PDCs (Figure 4E). On the
other hand, the expression of Npr2 was comparable between BD
and NB PDCs. Moreover, we examined the mutually exclusive
expression of Ostn and Npr3 in PDCs by re-analyzing the bulk
RNA sequencing (RNA-seq) data reported in the previous study
(Debnath et al., 2018). Periosteal stem cells marked by
Cathepsin K (Ctsk)-Cre predominantly expressed Ostn, whereas
mesenchymal stem cells unmarked by Ctsk-Cre expressed Npr3
(Figure S4B). These findings suggest that OSTN binds to PDCs
by NPR3 and might regulate the GC-B signaling to promote oste-
ogenic differentiation in a paracrine manner.
OSTN promotes osteogenic differentiation through
potentiation of CNP signaling
CNP regulates chondrogenesis to induce endochondral ossifi-
cation through GC-B (Nakao et al., 2015). Therefore, we asked
whether the CNP-GC-B signaling axis regulates osteogenic dif-
ferentiation in the PS. First, we analyzed the expression of Nppc
and its receptors during osteogenic differentiation of PDCs.
When PDCs were cultured in the osteogenic differentiation me-
dia, we detected the upregulation of osteoblast markers Osterix
and Collagen type I alpha1, encoded by Sp7 and Col1a1, indi-
cating that PDCs undergo osteoblastic differentiation (Fig-
ure S5A). The expression of Nppc and Npr2 was significantly
increased during osteogenic differentiation, whereas that of
Npr3 was unchanged (Figure S5A), implying a positive endoge-
nous regulation of the CNP-GC-B signaling for osteogenic
differentiation. We next investigated whether CNP promotes
osteogenic differentiation of PDCs. For that investigation, we
induced osteogenic differentiation of PDCs in the presence
and absence of exogenous CNP (CNP-22). CNP significantly
augmented alkaline phosphatase (ALP) activity and Ca2+ depo-
sition in the PDCs (Figures 5A–5C). Consistently, the expression
of osteoblast marker genes such as Sp7 and Col1a1 in the PDCs
was increased by the treatment with CNP (Figures 5D and 5E).
CNP neither affected adipogenic nor chondrogenic differentia-
tion of PDCs when cultured under the corresponding differentia-
tion media (Figures S5B–S5E). These results suggest that the
CNP-GC-B signaling potentiates osteogenic differentiation of
PDCs.
Next, we examined the effect of OSTN on the osteogenic dif-
ferentiation of PDCs in the presence and absence of CNP by
examining Ca2+ deposition, ALP activity, and the expression
of osteogenic markers (Figures 5F, 5G, and S5F–S5I). In the
absence of exogenous CNP, OSTN significantly enhanced
Ca2+ deposition and ALP activity (Figures 5F, 5G, S5F, and
S5G). OSTN further augmented exogenous CNP-dependent
differentiation of the PDCs into osteoblasts, as evidenced by
increased ALP activity and Ca2+ deposition. To investigate
the synergistic effect of the combination of CNP and OSTN
on the differentiation, we measured cGMP in the PDCs stimu-
lated with CNP. Although OSTN alone did not increase
cGMP, the combination of OSTN and CNP significantly
increased cGMP (Figure 5H). These results suggest that
OSTN promotes osteogenic differentiation through the
enhancement of CNP signaling.
To investigate whether impaired bone formation in the
OstnLacZ/LacZ mice is ascribed to the defects of osteogenic
differentiation by CNP signaling, we isolated PDCs from the
OstnLacZ/LacZ mice and analyzed their osteogenic differentiation.
The Ostn-depleted PDCs showed weaker osteogenic differenti-
ation than the control (Figures 5I, 5J, and S5J–S5L). However,
exogenous CNP rescued decreased osteogenic differentiation
of the PDCs derived from the OstnLacZ/LacZ mice. To confirm
that the reduced CNP signaling is responsible for the defect of
osteoblast differentiation in vivo, we calculated Osterix-positive
cells in the freshly isolated tibial periosteal cells at P11 days by
flow cytometry (Figure S5M). The percentage of periosteal Os-
terix-positive cells was lower in the OstnLacZ/LacZ mice than
that in the Ostn+/+ mice. However, overexpression of CNP in vivo
by crossing the OstnLacZ/LacZ mice with the transgenic mice ex-
pressing the human Nppc gene under serum amyloid P compo-
nent promoter (SAP-CNP) resulted in increased Osterix-positive
periosteal cells. Thus, these results indicate that the decreased
CNP signaling in the OstnLacZ/LacZ mice might be the cause of
impaired osteogenic differentiation in vivo.

Figure 5. OSTN promotes osteogenic differentiation through potentiation of CNP signaling

(A) Representative images of alkaline phosphatase (ALP) activity (top) and ARS staining (bottom) images at 1 day and 8 days of osteogenic differentiation,
respectively, with or without CNP-22 (10 nM).
(B and C) Quantified results of ALP- and ARS-positive areas used in (A) with or without indicated concentrations of CNP-22 as mean with ± SEM (B, n = 5; C, n = 7).
(D and E) qPCR analyses of the expression of osteoblast markers Sp7 (D) and Col1a1 (E) at 4 days of osteogenic differentiation with or without CNP-22 (1,000 nM).
The expression value of each sample was normalized by Gapdh mRNA. The relative expression to the value of the cells cultured in the growth medium is indicated
as mean with ± SEM (n = 3).
(F) Representative ARS images of the staining at 5 days of osteogenic differentiation in PDCs with 5 nM (top) or without (bottom) CNP-22 in the combination with
OSTN (1 nM).
(G) The fold increase of the quantified result of ARS staining area used in (F) relative to the cells cultured in differentiation medium as mean with ± SEM (n = 4).
(H) Effect of the indicated concentration of OSTN on cGMP production of PDCs at 15 min after the stimulation with or without CNP-22 (5 nM) as mean with ± SEM
(n = 3).
(I) Representative ARS staining at 8 days of osteogenic differentiation of PDCs isolated from 8-week-old OstnLacZ/LacZ and Ostn+/+ mice.
(J) Quantitative analysis of the ARS-positive area in the images used in (I) as mean with ± SEM (n = 3).
(K and L) Representative ARS staining (K) and quantified results (mean with ± SEM) (L) at 10 days of osteogenic differentiation of PDCs transfected with the control
siRNA and siRNA against Npr3 (bottom) with or without OSTN (1 nM) (n = 8).
**p < 0.01, *p < 0.05; by one-way ANOVA followed by Turkey-Kramer test in (B), (C), (D), (E), (G), (H), and (L) or by Student’s t test in (J). See also Figure S5 and Table
S1.

Cell Reports 36, 109380, July 13, 2021 9

 ll
OPEN ACCESS Article

A B C D

E F G I J

H K

Figure 6. Mechanical-load-induced expression of OSTN regulates bone formation

(A) Ostn expression in tibial PS and in tibial TB/CB/BM of the control mice kept on the ground (GC), under the condition of tail-suspension (TS) for 14 days and
subsequent reloading (RL) for 3 days. Gene expression value of each sample was normalized by each Gapdh mRNA level. The relative expression to the value of
PS in GC is indicated as mean with ± SEM (n = 4).
(B) X-gal staining of the cleared tibia of Ostn+/LacZ mouse under the indicated condition similarly to (A). Note the significant reduction of X-gal staining at the convex
and distal region of tibia applied to TS (arrowheads).

(legend continued on next page)

10 Cell Reports 36, 109380, July 13, 2021

 ll
Article OPEN ACCESS

Moreover, we investigated whether OSTN augments CNP-
dependent osteogenic differentiation through NPR3 (Figures
5K, 5L, and S5N). Knockdown of Npr3 augmented osteogenic
medium-induced Ca2+ deposition even without OSTN, implying
positive regulation of the endogenous CNP signaling. Osteo-
genic medium-induced Ca2+ deposition was enhanced by
OSTN stimulation in control cells but not in Npr3-depleted cells,
suggesting that OSTN promotes osteogenic differentiation in
PDCs though NPR3. Collectively, these findings suggest that
OSTN augments CNP signaling through NPR3 to induce osteo-
genic differentiation of PDCs.
Physiological-load-induced expression of OSTN
regulates bone formation
The characteristic expression pattern of Ostn in the convex and
medial surface of distal bones prompted us to hypothesize that
OSTN might act as a mechanotransducer to regulate bone
growth because OSTN was observed where tensile force gener-
ated by axial overload was applied (Razi et al., 2015a; Sugiyama
et al., 2012). To address this hypothesis, we performed a tail-
suspension (TS) test to eliminate physiological load to hindlimb
in mice. Consistent with the previous studies (Gerbaix et al.,
2015), the expression of Postn in periosteal cells was decreased
after TS for 14 days and was completely restored by reloading
(RL) for 3 days when compared to that in the mice kept on the
ground (Figure S6A). Conversely, the expression of Sost, an
osteocyte-derived mechanoresponsive factor, in the mixture of
TB/CB and BM (TB/CB/BM) cells was significantly increased af-
ter TS and decreased by subsequent RL, indicating that our TS
and RL were correctly applied to the bones (Figure S6B).
Therefore, we investigated the effect of TS and RL on the
expression of Ostn in two bone fractions, namely, PS and TB/
CB/BM. The Ostn expression in PS was significantly decreased
by TS for 14 days, whereas that in the TB/CB/BM cells was un-
changed (Figure 6A). However, subsequent RL for 3 days
reversed the TS-induced decrease in Ostn expression in the
periosteal cells. We also analyzed the effect of TS and RL on
the distribution of Ostn-expressing cells on the tibial surface
similarly to Figure 1A (Figure 6B). The expression of Ostn was
decreased on the convex and anteromedial surface of the tibia
by TS and restored by subsequent RL (arrowheads in Figure 6B).
To confirm mechanical-load-induced Ostn expression by elimi-
nating systemic side effects, we performed an additional expres-
sion analysis by using the tibiae obtained from the mice applied
for the unilateral denervation of sciatic nerve (Figure S6C; Baek
et al., 2020). The expression of Ostn was decreased in the dener-
vated tibia but not in the contralateral tibia. These results suggest
that periosteal Ostn expression is induced by physiological me-
chanical loading.
To elucidate the molecular mechanism underlying mechani-
cal-load-regulated Ostn expression, we focused on Forkhead
box protein O1 (FoxO1) because FoxO1 negatively regulates
Ostn expression in C2C12 myoblasts (Yasui et al., 2007). First,
we tested whether FoxO1 is modulated by mechanical stimuli
by using periosteal extracts from the tibiae of mice subjected
to TS and subsequent RL (Figures 6C and 6D). Expression of
FoxO1 in the tibial PS was increased by TS and decreased by
subsequent RL. Furthermore, nuclear localization of FoxO1 in
the PS was more frequently detected under an unloaded condi-
tion than under normal condition (Figure 6E). The mRNA of peri-
osteal glutathione peroxidase 1 (Gpx1), one of the target genes
for FoxO1 (Akasaki et al., 2014), was significantly higher in the
mice subjected to TS than that in the controls (Figure 6F). These
results indicate that the physiological loading in the bones in-
hibits FoxO1 activity, suggesting the role of FoxO1 as a me-
chano-responsive transcription factor.
We next investigated if FoxO1 suppresses Ostn expression in
the PS. Treatment with FoxO1 inhibitor and small interfering RNA
(siRNA)-mediated knockdown of FoxO1 resulted in an increased
expression of Ostn in the isolated PDCs, suggesting that FoxO1
negatively regulates Ostn expression (Figures 6G, S6D, and
S6E). Then, we tested whether FoxO1 directly regulates Ostn
gene expression in PDCs by performing a chromatin immuno-
precipitation (ChIP)-qPCR assay. As previously reported (Chief-
ari et al., 2018), the anti-FoxO1 antibody we used in ChIP-qPCR
was validated by the results, showing the successful immune-
precipitation of the promoter region of the Igfbp1 gene but not
that of the Rpl30 gene (Figure S6F). There are about 19 putative

(C and D) Western blot analysis for FoxO1 (top) and b-actin (bottom) by using the extract of tibial PS of the mice in GC, applied to TS for 5 days, and subsequent RL
for 5 days. Each lane represents a single animal. The quantitative result normalized to total protein is shown in (D) as mean with ± SEM (n = 3).
(E) Images of tibial PS stained with anti-FoxO1 (white) and 4’,6-diamidino-2-phenylindole (DAPI) (blue) by using mice of GC and TS for 5 days. Note the increased
nuclear staining of FoxO1 after TS (red arrowheads).
(F) qPCR analysis of the expression of Gpx1 in tibial PS applied to TS for 5 days. Data are indicated similarly to (A) (n = 3).
(G) qPCR analysis of the expression of Ostn in tibial PDCs after the treatment with FoxO1 inhibitor (FoxO1i) for 48 h. Data are normalized similarly to (A) as mean
with ± SEM (n = 3).
(H) Schematic representation of the potential FoxO1-binding sites predicted by JASPAR database within the 1.8 kb upstream of the Ostn transcription start site
(top) and percent inputs calculated by ChIP-qPCR analysis with the indicated primer sets and antibodies. Red and blue triangles indicate the potential FoxO1-
binding sites of plus (+) and minus (–) strands, respectively. Region 1 (R1) to region 5 (R5) show the area covered by qPCR. Data were shown as mean with ± SEM
(n = 4–6).
(I) Representative X-gal staining images of the cleared tibia of Ostn+/LacZ mouse and Ostn+/LacZ;Ctsk-Cre;CAG-FoxO1AAA for 1 h at P21. Note the significant
reduction of X-gal staining at distal region of tibia in Ostn+/LacZ;Ctsk-Cre;CAG-FoxO1AAA (arrowhead).
(J) qPCR analysis of the expression of Ostn mRNA in tibial PS isolated from Ostn+/LacZ and Ostn+/LacZ;Ctsk-Cre;CAG-FoxO1AAA mice at P21. Data are indicated
similarly to (A) as mean with ± SEM (n = 6–8).
(K) microCT parameters of the tibial trabecular bone from the mice of GC, applied to TS for 14 days, and subsequent RL for 7 days. All results are expressed as
mean with ± SEM (n = 5–10).
Scale bars, 2.5 mm in (B) and (I) and 50 mm in (E). **p < 0.01, *p < 0.05; by one-way ANOVA followed by Turkey-Kramer test (A, D, and K) or by Student’s t test (F, G,
H, and J). See also Figure S6 and Table S1.

Cell Reports 36, 109380, July 13, 2021 11

 ll
OPEN ACCESS
FoxO1 binding sites between the 1.8 kb upstream and the Ostn
transcription start site (Figure 6H). Therefore, we investigated
whether FoxO1 binds to these predicted sites by performing
ChIP-qPCR. Among them, the region of 229 to +53 bp from
the transcription start site was identified as a FoxO1 binding site.
To further investigate whether FoxO1 suppresses Ostn
expression in vivo, we examined the effect of ectopic expression
of constitutively active mutant of FoxO1 (FoxO1AAA) on Ostn
gene expression in the PS. To this end, Ctsk-Cre mice in which
Cre expression is induced in the periosteal skeletal stem cell line-
age and in osteoclast lineage within BM were crossed with the
CAG-FoxO1AAA transgenic mice to induce FoxO1AAA expres-
sion. Before that step, we examined where Cre expression is
induced in vivo by crossing the Ctsk-Cre mice with ROSA-(yellow
fluorescent protein)YFP reporter mice and found YFP expressed
in the periosteal cells that expressed Ostn (Figure S6H). We then
examined Ostn expression in the PS of Ctsk-Cre;CAG-Fox-
O1AAA;Ostn+/LacZ mice by performing X-gal staining and qPCR
analysis and found that forced expression of FoxO1AAA resulted
in the suppression of periosteal Ostn expression (Figures 6I and
6J). In addition, we observed a severe osteopetrosis-like pheno-
type in the Ctsk-Cre;CAG-FoxO1AAA mice (Figures S6I, S6J, and
S6K), which was likely caused by the expression of FoxO1AAA in
osteoclasts because it has been reported that FoxO1 restrains
osteoclastogenesis (Bartell et al., 2014). Taken together with
the data obtained from the studies using PDCs and ChIP-
qPCR analyses, these results show the role of FoxO1 as a nega-
tive regulator of periosteal Ostn expression.
We further investigated whether FoxO1 is involved in TS-
induced downregulation of Ostn expression. As expected, TS
for 1 week dramatically decreased the number of OSTN/Osterix
double-positive cells in the PS. However, the FoxO1 inhibitor
partially blocked TS-induced reduction of the number of
OSTN/Osterix double-positive cells (Figure S6G). Collectively,
these findings suggest that mechanical loading induces Ostn
expression in the PS by inhibiting FoxO1-dependent repression
of Ostn expression.
Finally, we tested whether mechanical-load-induced expres-
sion of OSTN regulates bone formation. The Ostn+/+ mice sub-
jected to TS for 14 days exhibited a dramatic loss of TB, as evi-
denced by decreased bone parameters including BV, BV/TV,
BS, Tb.Th., and Tb.N. (Figures 6K, S6L, and S6M). However,
subsequent RL for 7 days partially but significantly restored the
loss of TB, indicating that RL leads to physiological-load-medi-
ated restoration of bone mass. TS also caused a loss of TB in
the OstnLacZ/LacZ mice to the level observed in the Ostn+/+
mice. However, RL-induced restoration of TB loss was signifi-
cantly impaired in the OstnLacZ/LacZ mice compared with that in
Ostn+/+ mice. These results suggest that OSTN plays a signifi-
cant role in physiological-load-induced bone formation.
DISCUSSION
In this study, we clearly demonstrated defective formation of
long bone in our OstnLacZ/LacZ mice. However, Subbotina et al.
(2015) have previously reported no abnormalities of tibiae in
the Ostn-deficient mice. Currently, the reason for this discrep-
ancy remains unclear, but it might be attributed to different
12 Cell Reports 36, 109380, July 13, 2021
Article
experimental procedures used to analyze skeletal structures.
They analyzed thickness and length of hindlimb bone in only a
limited number of Ostn-deficient mice, whereas we assessed
skeletal systems in more than 12 OstnLacZ/LacZ mice. In addition,
spatial resolution in microCT used in this study is higher than that
used in their study. Furthermore, different sites of tibial CB were
measured in this and their studies. Another possibility is that
skeletal phenotypes of Ostn-deficient mice might depend on
the strains of the mice used for the analyses.
OSTN promotes growth of long bone possibly by facilitating
the CNP signaling pathway. We have shown that overexpression
of OSTN leads to the elongation of bones in both zebrafish and
mouse in a CNP-dependent manner (Chiba et al., 2017; Kanai
et al., 2017). Furthermore, several studies have indicated that
CNP plays a pivotal role in proliferation/differentiation of chon-
drocytes in GP during endochondral ossification (Chusho
et al., 2001; Kawasaki et al., 2008; Nakao et al., 2015; Tsuji
and Kunieda, 2005; Yasoda et al., 2004). Consistently, the pre-
sent study further demonstrated the crucial role of endogenous
OSTN as a physiological regulator of endochondral ossification.
In addition, our study showed the role of OSTN in osteoblast-
dependent bone formation. Indeed, Ostn-deficient mice ex-
hibited reduced osteoblast activity in tibial TB. Furthermore,
OSTN potentiated CNP-induced osteogenic differentiation of
PDCs in vitro. Thus, OSTN may directly regulate bone formation
by promoting CNP-dependent osteogenic differentiation.
Consistently, the roles of CNP and cGKII, a downstream effector
of CNP, in osteogenic differentiation and bone formation has
been suggested in vitro and in vitro (Hagiwara et al., 1996; Kondo
et al., 2015; Ramdani et al., 2018). Thus, OSTN may regulate
bone formation by promoting CNP-signal-mediated formation
of GP and by potentiating CNP-dependent osteogenic
differentiation.
How does periosteal-derived OSTN regulate endochondral
ossification? Our careful examination revealed the predominant
expression of Ostn in periosteal osteoblasts but not in GP, BM,
and CB of tibia. Furthermore, we previously reported that over-
expression of OSTN under the liver-specific promoter in mice in-
duces an increase in bone length, suggesting that OSTN can be
delivered through blood vessels (Kanai et al., 2017; Moffatt et al.,
2007). Thus, periosteal OSTN might be delivered to the GP and
TB by passing across the CB by blood vessels. Although there
are no penetrating blood vessels into the layers of GPs, epiphy-
seal and metaphyseal ends of GP are supplied by epiphyseal ar-
teries and metaphyseal arteries, respectively (Ramasamy, 2017).
In addition, periosteal blood vessels have been shown to pass
thorough CB and anastomose with arterioles orvenules in BM
(Gru€neboom et al., 2019). Therefore, OSTN released from the
periosteal osteoblasts might be transported into the GP and
TB by the blood stream to promote CNP-signal-mediated long
bone growth, although this hypothesis needs to be verified by
tissue-specific gene deletions of Ostn, Npr3, and Npr2.
OSTN responds to physiological loading to regulate bone
growth. The quadruped animals support their body weights by
their forearms and hindlimbs, especially through distal bones in
which Ostn is highly expressed. Our analyses using the TS model
showed that the mechanical load induces the expression of
Ostn. Moreover, Ostn is expressed in the antero-medial surface

Article
but not in the posterior surface of the tibia, suggesting the rele-
vance of Ostn expression to mechanical loading. Indeed, axial-
loading-induced tensile and compressed stress toward anterior
and posterior sides of tibia elicit site-specific response of bone
(re)modeling (Razi et al., 2015a; Sugiyama et al., 2012). However,
because daily physical activities transmit complex mechanical
loads including tension, compression, torsion, and shear to the
skeleton (Hart et al., 2017), more simplified models such as axial
loading need to be used to identify the physical cues that induce
Ostn expression in future analysis.
OSTN regulates physiological-load-induced bone growth
independently of osteocyte-mediated mechanical response. It
has been shown that ablation of osteocytes in mice results in
the resistance to mechanical unloading as in the TS model (Qin
et al., 2020; Tatsumi et al., 2007). Thus, osteocytes are generally
accepted as potent negative regulators that sense multiple me-
chanical stresses. Osteocyte-derived Sclerostin (SOST) and re-
ceptor activator of nuclear factor-kappa B ligand (RANKL)
reduce bone mass by preventing the osteoblast proliferation/dif-
ferentiation/apoptosis and activating osteoclasts, respectively
(Cardoso et al., 2009; Lin et al., 2009; Sutherland et al., 2004).
After unloading, the bone mass in Ostn-deficient mice was
severely decreased to the level observed in wild-type mice, sug-
gesting OSTN is dispensable for unloading-induced bone loss.
On the other hand, osteocytes and SOST have been shown to
be dispensable for loading-induced bone gain, suggesting the
primary role of other cells in mechanical loading (Morse et al.,
2014; Tatsumi et al., 2007). Consistently, Ostn-deficient mice ex-
hibited an impairment of RL-induced growth of TB. Therefore,
OSTN is required for physiological-load-induced bone growth
besides osteocyte-mediated mechanical responses.
Limitations of study
The Ctsk-Cre;CAG-FoxO1AAA mice show that FoxO1 is a nega-
tive regulator for Ostn expression (Figures 6I and 6J). However,
the Ctsk-Cre;CAG-FoxO1AAA mice exhibited a severe osteopet-
rosis-like phenotype (Figures S6I, S6J, and S6K), which was
likely to be caused by the expression of FoxO1AAA in osteoclasts
because it has been reported that FoxO1 restrains osteoclasto-
genesis (Bartell et al., 2014). Therefore, this osteopetrosis-like
phenotype is likely ascribed to the impairment of osteoclasto-
genesis because the Ctsk promoter is active not only in perios-
teal skeletal stem cells but also in osteoclasts (Debnath et al.,
2018; Nakamura et al., 2007). Thus, we cannot totally exclude
the possibility that downregulation of periosteal Ostn is second-
ary to the osteopetrosis-like phenotype. Currently, in spite of ad-
vances in periosteal research, there are no Cre driver mice that
express Cre recombinase specifically in the periosteal stem cells
(Debnath et al., 2018; Duchamp de Lageneste et al., 2018; Gao
et al., 2019; Ortinau et al., 2019). Our conclusion will be
confirmed by the future studies after development of the
methods that specifically modify the lineage of periosteal stem
cells/osteoblasts.
STAR+METHODS
Detailed methods are provided in the online version of this paper
and include the following:
ll
OPEN ACCESS
d KEY RESOURCES TABLE
d RESOURCE AVAILABILITY
B Lead contact
B Materials availability
B Data and code availability
d EXPERIMENTAL MODEL AND SUBJECT DETAILS
B Animals
B Primary cell cultures
d METHOD DETAILS
B Whole mount X-gal staining and clearing
B Whole-mount periosteum staining
B Laser microdissection
B MicroCT
B Bone length measurement
B In situ hybridization for Col10a1
B In situ hybridization for Ostn
B Tissue preparation and immunohistochemistry
B Sorting of LacZ-positive or YFP-positive cells
B Estimation of Osterix (+) cells by FACS
B Determination of lineage-specific cell differentiation
B OSTN binding assay
B RNA preparation and qPCR
B Morphometric analysis
B Analysis of processed RNAseq data
B Knockdown of Npr3 and Foxo1 in PDCs
B Analysis of proliferating chondrocytes
B Measurement of cGMP activity
B Western blotting
B Measurement of OSTN concentration in plasma
B ChIP-qPCR assay
d QUANTIFICATION AND STATISTICAL ANALYSIS
SUPPLEMENTAL INFORMATION
Supplemental information can be found online at https://doi.org/10.1016/j.
celrep.2021.109380.
ACKNOWLEDGMENTS
We appreciate Dr. Takahashi, Dr. Nakamura, and Dr. Watanabe for providing
Npr3-deficinet, Ctsk-Cre, and Rosa-Lox-STOP-Lox-YFP mice, respectively. In
addition, we thank Ms. Takatsu for her technical assistance in laser microdis-
section. This work was supported by Grants-in-Aid for Young Scientists (A)
(15H05646 to H.W.-T.), for Scientific Research on Innovative Areas
(18H04994 to H.W.-T.), and for Scientific Research (C) (18K09050 to H.W.-
T.) from Japan Society for the Promotion of Science (JSPS), AMED-CREST
(Japan Agency for Medical Research and Development, Core Research for
Evolutional Science and Technology) (JP17gm0610010 to N. Mochizuki);
and grants from the Uehara Memorial Foundation, Japan (to H.W.-T.), the Mo-
chida Memorial Foundation, Japan (to H.W.-T.), Meiji Yasuda Life Foundation
of Health and Welfare, Japan (to H.W.-T.), the Okinaka Memorial Institute for
Medical Research, Japan (to H.W.-T.), and Takeda Science Foundation, Japan
(to N. Mochizuki).
AUTHOR CONTRIBUTIONS
H.W.-T. and N. Mochizuki designed the experiments and wrote the paper.
H.W.-T. performed the almost analysis. H.O. helped with microCT analysis.
Other coauthors contributed to considering plans of experiments, developed
materials, and discussed the results.
Cell Reports 36, 109380, July 13, 2021 13
 ll
OPEN ACCESS
DECLARATION OF INTERESTS
The authors declare no competing interests.
INCLUSION AND DIVERSITY
The author list of this paper includes contributors from the location where the
research was conducted who participated in the data collection, design, anal-
ysis, and/or interpretation of the work.
Received: October 28, 2019
Revised: April 15, 2021
Accepted: June 21, 2021
Published: July 13, 2021
REFERENCES
Akasaki, Y., Alvarez-Garcia, O., Saito, M., Caramés, B., Iwamoto, Y., and Lotz,
M.K. (2014). FoxO transcription factors support oxidative stress resistance in
human chondrocytes. Arthritis Rheumatol. 66, 3349–3358.
Baek, K.-W., Jung, Y.-K., Kim, J.-S., Park, J.S., Hah, Y.-S., Kim, S.-J., and
Yoo, J.-I. (2020). Rodent Model of Muscular Atrophy for Sarcopenia Study.
J. Bone Metab. 27, 97–110.
Bartell, S.M., Kim, H.-N., Ambrogini, E., Han, L., Iyer, S., Serra Ucer, S., Rabi-
novitch, P., Jilka, R.L., Weinstein, R.S., Zhao, H., et al. (2014). FoxO proteins
restrain osteoclastogenesis and bone resorption by attenuating H2O2 accu-
mulation. Nat. Commun. 5, 3773.
Cardoso, L., Herman, B.C., Verborgt, O., Laudier, D., Majeska, R.J., and
Schaffler, M.B. (2009). Osteocyte apoptosis controls activation of intracortical
resorption in response to bone fatigue. J. Bone Miner. Res. 24, 597–605.
Chiba, A., Watanabe-Takano, H., Terai, K., Fukui, H., Miyazaki, T., Uemura, M.,
Hashimoto, H., Hibi, M., Fukuhara, S., and Mochizuki, N. (2017). Osteocrin, a
peptide secreted from the heart and other tissues, contributes to cranial osteo-
genesis and chondrogenesis in zebrafish. Development 144, 334–344.
Chiefari, E., Arcidiacono, B., Palmieri, C., Corigliano, D.M., Morittu, V.M., Britti,
D., Armoni, M., Foti, D.P., and Brunetti, A. (2018). Cross-talk among HMGA1
and FoxO1 in control of nuclear insulin signaling. Sci. Rep. 8, 8540.
Chusho, H., Tamura, N., Ogawa, Y., Yasoda, A., Suda, M., Miyazawa, T., Na-
kamura, K., Nakao, K., Kurihara, T., Komatsu, Y., et al. (2001). Dwarfism and
early death in mice lacking C-type natriuretic peptide. Proc. Natl. Acad. Sci.
USA 98, 4016–4021.
Debnath, S., Yallowitz, A.R., McCormick, J., Lalani, S., Zhang, T., Xu, R., Li, N.,
Liu, Y., Yang, Y.S., Eiseman, M., et al. (2018). Discovery of a periosteal stem
cell mediating intramembranous bone formation. Nature 562, 133–139.
Duchamp de Lageneste, O., Julien, A., Abou-Khalil, R., Frangi, G., Carvalho,
C., Cagnard, N., Cordier, C., Conway, S.J., and Colnot, C. (2018). Periosteum
contains skeletal stem cells with high bone regenerative potential controlled by
Periostin. Nat. Commun. 9, 773.
Elder, S.H., Goldstein, S.A., Kimura, J.H., Soslowsky, L.J., and Spengler, D.M.
(2001). Chondrocyte differentiation is modulated by frequency and duration of
cyclic compressive loading. Ann. Biomed. Eng. 29, 476–482.
Frost, H.M. (2003). Bone’s mechanostat: a 2003 update. Anat. Rec. A Discov.
Mol. Cell. Evol. Biol. 275, 1081–1101.
Gao, B., Deng, R., Chai, Y., Chen, H., Hu, B., Wang, X., Zhu, S., Cao, Y., Ni, S.,
Wan, M., et al. (2019). Macrophage-lineage TRAP+ cells recruit periosteum-
derived cells for periosteal osteogenesis and regeneration. J. Clin. Invest.
129, 2578–2594.
Gerbaix, M., Vico, L., Ferrari, S.L., and Bonnet, N. (2015). Periostin expression
contributes to cortical bone loss during unloading. Bone 71, 94–100.
Gru€neboom, A., Hawwari, I., Weidner, D., Culemann, S., Mu €ller, S., Henne-
berg, S., Brenzel, A., Merz, S., Bornemann, L., Zec, K., et al. (2019). A network
of trans-cortical capillaries as mainstay for blood circulation in long bones. Nat.
Metab. 1, 236–250.
14 Cell Reports 36, 109380, July 13, 2021
Article
Hagiwara, H., Inoue, A., Yamaguchi, A., Yokose, S., Furuya, M., Tanaka, S.,
and Hirose, S. (1996). cGMP produced in response to ANP and CNP regulates
proliferation and differentiation of osteoblastic cells. Am. J. Physiol. 270,
C1311–C1318.
Hart, N.H., Nimphius, S., Rantalainen, T., Ireland, A., Siafarikas, A., and
Newton, R.U. (2017). Mechanical basis of bone strength: influence of bone ma-
terial, bone structure and muscle action. J. Musculoskelet. Neuronal Interact.
17, 114–139.
Iolascon, G., Resmini, G., and Tarantino, U. (2013). Mechanobiology of bone.
Aging Clin. Exp. Res. 25, S3–S7.
Jaubert, J., Jaubert, F., Martin, N., Washburn, L.L., Lee, B.K., Eicher, E.M., and
Guénet, J.L. (1999). Three new allelic mouse mutations that cause skeletal
overgrowth involve the natriuretic peptide receptor C gene (Npr3). Proc.
Natl. Acad. Sci. USA 96, 10278–10283.
Kanai, Y., Yasoda, A., Mori, K.P., Watanabe-Takano, H., Nagai-Okatani, C.,
Yamashita, Y., Hirota, K., Ueda, Y., Yamauchi, I., Kondo, E., et al. (2017).
Circulating osteocrin stimulates bone growth by limiting C-type natriuretic
peptide clearance. J. Clin. Invest. 127, 4136–4147.
Kawasaki, Y., Kugimiya, F., Chikuda, H., Kamekura, S., Ikeda, T., Kawamura,
N., Saito, T., Shinoda, Y., Higashikawa, A., Yano, F., et al. (2008). Phosphory-
lation of GSK-3beta by cGMP-dependent protein kinase II promotes hypertro-
phic differentiation of murine chondrocytes. J. Clin. Invest. 118, 2506–2515.
Kim, J.-B., Leucht, P., Lam, K., Luppen, C., Ten Berge, D., Nusse, R., and
Helms, J.A. (2007). Bone regeneration is regulated by wnt signaling. J. Bone
Miner. Res. 22, 1913–1923.
Kita, S., Nishizawa, H., Okuno, Y., Tanaka, M., Yasui, A., Matsuda, M., Ya-
mada, Y., and Shimomura, I. (2009). Competitive binding of musclin to natri-
uretic peptide receptor 3 with atrial natriuretic peptide. J. Endocrinol. 201,
287–295.
Kondo, E., Yasoda, A., Fujii, T., Nakao, K., Yamashita, Y., Ueda-Sakane, Y.,
Kanamoto, N., Miura, M., Arai, H., Mukoyama, M., et al. (2015). Increased
Bone Turnover and Possible Accelerated Fracture Healing in a Murine Model
With an Increased Circulating C-Type Natriuretic Peptide. Endocrinology
156, 2518–2529.
Kozhemyakina, E., Lassar, A.B., and Zelzer, E. (2015). A pathway to bone:
signaling molecules and transcription factors involved in chondrocyte devel-
opment and maturation. Development 142, 817–831.
Lin, C., Jiang, X., Dai, Z., Guo, X., Weng, T., Wang, J., Li, Y., Feng, G., Gao, X.,
and He, L. (2009). Sclerostin mediates bone response to mechanical unloading
through antagonizing Wnt/beta-catenin signaling. J. Bone Miner. Res. 24,
1651–1661.
Long, F., and Ornitz, D.M. (2013). Development of the endochondral skeleton.
Cold Spring Harb. Perspect. Biol. 5, a008334.
Mackie, E.J., Tatarczuch, L., and Mirams, M. (2011). The skeleton: a multi-
functional complex organ: the growth plate chondrocyte and endochondral
ossification. J. Endocrinol. 211, 109–121.
Matsukawa, N., Grzesik, W.J., Takahashi, N., Pandey, K.N., Pang, S., Yamau-
chi, M., and Smithies, O. (1999). The natriuretic peptide clearance receptor
locally modulates the physiological effects of the natriuretic peptide system.
Proc. Natl. Acad. Sci. USA 96, 7403–7408.
Miyazaki, T., Otani, K., Chiba, A., Nishimura, H., Tokudome, T., Takano-Wata-
nabe, H., Matsuo, A., Ishikawa, H., Shimamoto, K., Fukui, H., et al. (2018). A
New Secretory Peptide of Natriuretic Peptide Family, Osteocrin, Suppresses
the Progression of Congestive Heart Failure After Myocardial Infarction. Circ.
Res. 122, 742–751.
Moffatt, P., Thomas, G., Sellin, K., Bessette, M.-C., Lafrenière, F., Akhouayri,
O., St-Arnaud, R., and Lancto ^ t, C. (2007). Osteocrin is a specific ligand of
the natriuretic Peptide clearance receptor that modulates bone growth.
J. Biol. Chem. 282, 36454–36462.
Morse, A., McDonald, M.M., Kelly, N.H., Melville, K.M., Schindeler, A., Kramer,
I., Kneissel, M., van der Meulen, M.C.H., and Little, D.G. (2014). Mechanical
load increases in bone formation via a sclerostin-independent pathway.
J. Bone Miner. Res. 29, 2456–2467.

Article
Nagai, C., and Minamino, N. (2014). Direct chemiluminescent enzyme immu-
noassay for atrial natriuretic peptide in mammalian plasma using a PEGylated
antibody. Anal. Biochem. 461, 10–16.
Nakamura, T., Imai, Y., Matsumoto, T., Sato, S., Takeuchi, K., Igarashi, K., Har-
ada, Y., Azuma, Y., Krust, A., Yamamoto, Y., et al. (2007). Estrogen prevents
bone loss via estrogen receptor alpha and induction of Fas ligand in osteo-
clasts. Cell 130, 811–823.
Nakao, K., Osawa, K., Yasoda, A., Yamanaka, S., Fujii, T., Kondo, E., Koyama,
N., Kanamoto, N., Miura, M., Kuwahara, K., et al. (2015). The Local CNP/GC-B
system in growth plate is responsible for physiological endochondral bone
growth. Sci. Rep. 5, 10554.
Nishizawa, H., Matsuda, M., Yamada, Y., Kawai, K., Suzuki, E., Makishima, M.,
Kitamura, T., and Shimomura, I. (2004). Musclin, a novel skeletal muscle-
derived secretory factor. J. Biol. Chem. 279, 19391–19395.
Ogita, M., Rached, M.T., Dworakowski, E., Bilezikian, J.P., and Kousteni, S.
(2008). Differentiation and proliferation of periosteal osteoblast progenitors
are differentially regulated by estrogens and intermittent parathyroid hormone
administration. Endocrinology 149, 5713–5723.
Ohashi, N., Robling, A.G., Burr, D.B., and Turner, C.H. (2002). The effects of
dynamic axial loading on the rat growth plate. J. Bone Miner. Res. 17, 284–292.
Ortinau, L.C., Wang, H., Lei, K., Deveza, L., Jeong, Y., Hara, Y., Grafe, I., Rose-
nfeld, S.B., Lee, D., Lee, B., et al. (2019). Identification of Functionally Distinct
Mx1+aSMA+ Periosteal Skeletal Stem Cells. Cell Stem Cell 25, 784–796.e5.
Ouyang, W., Liao, W., Luo, C.T., Yin, N., Huse, M., Kim, M.V., Peng, M., Chan,
P., Ma, Q., Mo, Y., et al. (2012). Novel Foxo1-dependent transcriptional pro-
grams control T(reg) cell function. Nature 491, 554–559.
Potter, L.R., Abbey-Hosch, S., and Dickey, D.M. (2006). Natriuretic peptides,
their receptors, and cyclic guanosine monophosphate-dependent signaling
functions. Endocr. Rev. 27, 47–72.
Qin, L., Liu, W., Cao, H., and Xiao, G. (2020). Molecular mechanosensors in os-
teocytes. Bone Res. 8, 23–24.
Ramasamy, S.K. (2017). Structure and Functions of Blood Vessels and
Vascular Niches in Bone. Stem Cells Int. 2017, 5046953.
Ramdani, G., Schall, N., Kalyanaraman, H., Wahwah, N., Moheize, S., Lee,
J.J., Sah, R.L., Pfeifer, A., Casteel, D.E., and Pilz, R.B. (2018). cGMP-depen-
dent protein kinase-2 regulates bone mass and prevents diabetic bone loss.
J. Endocrinol. 238, 203–219.
Rauch, F. (2005). Bone growth in length and width: the Yin and Yang of bone
stability. J. Musculoskelet. Neuronal Interact. 5, 194–201.
Razi, H., Birkhold, A.I., Zaslansky, P., Weinkamer, R., Duda, G.N., Willie, B.M.,
and Checa, S. (2015a). Skeletal maturity leads to a reduction in the strain mag-
nitudes induced within the bone: a murine tibia study. Acta Biomater. 13,
301–310.
Srinivas, S., Watanabe, T., Lin, C.S., William, C.M., Tanabe, Y., Jessell, T.M.,
and Costantini, F. (2001). Cre reporter strains produced by targeted insertion
of EYFP and ECFP into the ROSA26 locus. BMC Dev. Biol. 1, 4.
Stich, S., Loch, A., Leinhase, I., Neumann, K., Kaps, C., Sittinger, M., and
Ringe, J. (2008). Human periosteum-derived progenitor cells express distinct
ll
OPEN ACCESS
chemokine receptors and migrate upon stimulation with CCL2, CCL25,
CXCL8, CXCL12, and CXCL13. Eur. J. Cell Biol. 87, 365–376.
Subbotina, E., Sierra, A., Zhu, Z., Gao, Z., Koganti, S.R.K., Reyes, S., Stepniak,
E., Walsh, S.A., Acevedo, M.R., Perez-Terzic, C.M., et al. (2015). Musclin is an
activity-stimulated myokine that enhances physical endurance. Proc. Natl.
Acad. Sci. USA 112, 16042–16047.
Sugiyama, T., Meakin, L.B., Browne, W.J., Galea, G.L., Price, J.S., and Lan-
yon, L.E. (2012). Bones’ adaptive response to mechanical loading is essentially
linear between the low strains associated with disuse and the high strains
associated with the lamellar/woven bone transition. J. Bone Miner. Res. 27,
1784–1793.
Sutherland, M.K., Geoghegan, J.C., Yu, C., Turcott, E., Skonier, J.E., Winkler,
D.G., and Latham, J.A. (2004). Sclerostin promotes the apoptosis of human
osteoblastic cells: a novel regulation of bone formation. Bone 35, 828–835.
Tainaka, K., Kubota, S.I., Suyama, T.Q., Susaki, E.A., Perrin, D., Ukai-Tade-
numa, M., Ukai, H., and Ueda, H.R. (2014). Whole-body imaging with single-
cell resolution by tissue decolorization. Cell 159, 911–924.
Tatsumi, S., Ishii, K., Amizuka, N., Li, M., Kobayashi, T., Kohno, K., Ito, M.,
Takeshita, S., and Ikeda, K. (2007). Targeted ablation of osteocytes induces
osteoporosis with defective mechanotransduction. Cell Metab. 5, 464–475.
Thomas, G., Moffatt, P., Salois, P., Gaumond, M.-H., Gingras, R., Godin, E.,
Miao, D., Goltzman, D., and Lancto ^ t, C. (2003). Osteocrin, a novel bone-spe-
cific secreted protein that modulates the osteoblast phenotype. J. Biol. Chem.
278, 50563–50571.
Tsuji, T., and Kunieda, T. (2005). A loss-of-function mutation in natriuretic pep-
tide receptor 2 (Npr2) gene is responsible for disproportionate dwarfism in cn/
cn mouse. J. Biol. Chem. 280, 14288–14292.
van Gastel, N., Torrekens, S., Roberts, S.J., Moermans, K., Schrooten, J., Car-
meliet, P., Luttun, A., Luyten, F.P., and Carmeliet, G. (2012). Engineering vas-
cularized bone: osteogenic and proangiogenic potential of murine periosteal
cells. Stem Cells 30, 2460–2471.
Wang, X., and Mao, J.J. (2002). Accelerated chondrogenesis of the rabbit cra-
nial base growth plate by oscillatory mechanical stimuli. J. Bone Miner. Res.
17, 1843–1850.
Wang, Q., Huang, C., Zeng, F., Xue, M., and Zhang, X. (2010). Activation of the
Hh pathway in periosteum-derived mesenchymal stem cells induces bone for-
mation in vivo: implication for postnatal bone repair. Am. J. Pathol. 177, 3100–
3111.
Xu, C., Ochi, H., Fukuda, T., Sato, S., Sunamura, S., Takarada, T., Hinoi, E.,
Okawa, A., and Takeda, S. (2016). Circadian Clock Regulates Bone Resorption
in Mice. J. Bone Miner. Res. 31, 1344–1355.
Yasoda, A., Komatsu, Y., Chusho, H., Miyazawa, T., Ozasa, A., Miura, M., Kur-
ihara, T., Rogi, T., Tanaka, S., Suda, M., et al. (2004). Overexpression of CNP in
chondrocytes rescues achondroplasia through a MAPK-dependent pathway.
Nat. Med. 10, 80–86.
Yasui, A., Nishizawa, H., Okuno, Y., Morita, K., Kobayashi, H., Kawai, K., Mat-
suda, M., Kishida, K., Kihara, S., Kamei, Y., et al. (2007). Foxo1 represses
expression of musclin, a skeletal muscle-derived secretory factor. Biochem.
Biophys. Res. Commun. 364, 358–365.
Cell Reports 36, 109380, July 13, 2021 15
 ll
OPEN ACCESS Article

STAR+METHODS

KEY RESOURCES TABLE

REAGENT or RESOURCE SOURCE IDENTIFIER

Antibodies
Anti-b-actin Sigma-Aldrich Cat#A5316; RRID:AB_476743
Anti-Digoxigenin-AP, Fab fragments Sigma-Aldrich Cat#11093274910; RRID:AB_2734716
Anti-FoxO1(C29H4) Cell Signaling Technology Cat#2880; RRID:AB_2106495
Anti-Osterix Abcam Cat#ab22552; RRID:AB_2194492
Anti-Osteocrin R&D Systems Cat#MAB2620; RRID:AB_2156721
Dako EnVision+ System-HRP Labeled DAKO Cat#K4002; RRID:AB_2630375
Polymer anti-Rabbit
Goat anti-Rabbit IgG (H+L) Highly Thermo Fisher Scientific Cat#A11034; RRID:AB_2576217
Cross-Adsorbed Secondary Antibody,
Alexa Fluor 488
Goat anti-Rabbit IgG (H+L) Highly Thermo Fisher Scientific Cat#A11035; RRID:AB_2534093
Cross-Adsorbed Secondary Antibody,
Alexa Fluor 546
Chemicals, peptides, and recombinant proteins
Alcian Blue Sigma-Aldrich Cat#A3157; CAS:33864-99-2
Biotin-tagged Osteocrin Toray Research Center, Inc Special order; CAS:N/A
BM-Purple Sigma-Aldrich Cat#11442074001; CAS:N/A
Calcein Dojindo Cat#340-00433; CAS:1461-15-0
CNP-22 Peptide Institute. Inc. Cat#4229-v; CAS:127869-51-6
Collagenase type II Worthington Cat#LS004176; CAS:9001-12-1
DDAO galactoside (9H-(1,3-Dichloro-9,9- Thermo Fisher Scientific Cat# D6488; CAS: 503178-95-8
Dimethylacridin-2-One-7-yl)
b-D-Galactopyranoside)
Dispase II, powder GIBCO Cat#17105-041; CAS:42613-33- 2
FoxO1 inhibitor AS1842856 Sigma-Aldrich Cat#344355; CAS:836620-48-5
KOD SYBR qPCR Mix TOYOBO Cat#QKD-201; CAS:N/A
LipofectamineTM RNAiMAX Thermo Fisher Scientific Cat#13778075; CAS:N/A
Transfection Reagent
Oil Red O Sigma-Aldrich Cat#O0625; CAS:1320-06-5
Osteocrin Toray Research Center, Inc Special order; CAS:N/A
Proteinase K QIAGEN Cat#19131; CAS:39450-01-6
Random primers Thermo Fisher Scientific Cat#48190-011; CAS:N/A
RNase OUT Thermo Fisher Scientific Cat#10777-019; CAS:N/A
SsoAdvancedTM Universal BIO-RAD Cat#172-5270; CAS:N/A
SYBR Green Super mix
StreptavidinCy3 from Sigma-Aldrich Cat#S6402; CAS:N/A
Streptomyces avidinii buffered
aqueous solution
TRIzol RNA Isolation Reagent Thermo Fisher Scientific Cat#15596018; CAS:N/A
1% Alizarin Red S solution Muto Pure. Chemicals Cat#17972; CAS:130-22-3
3-Isobutyl-1-methylxanthine Sigma-Aldrich Cat#I5879; CAS:28822-58-4
4% Paraformaldehyde WAKO Cat#161-20141; CAS:30525-89-4
Phosphate Buffer Solution
40 ,6-Diamidine-20 -phenylindole Sigma-Aldrich Cat#D9542; CAS:28718-90-3
dihydrochloride
(Continued on next page)

e1 Cell Reports 36, 109380, July 13, 2021

 ll
Article OPEN ACCESS

Continued
REAGENT or RESOURCE SOURCE IDENTIFIER
5-Bromo-4-chloro-3-indolyl WAKO Cat#021-07852; CAS:7240-90-6
b-D-galactopyranoside
10 mM dNTPs Thermo Fisher Scientific Cat#18427-013; CAS:56-38-2
Critical commercial assays
Click-iT EdU Cell Proliferation Kit for Imaging, Thermo Fisher Scientific Cat#C10337
Alexa Fluor 488 dye
Cyclic GMP ELISA Kit Cayman chemical Cat#581021
DAB Substrate Kit, Peroxidase Vector Labs Cat#SK-4001
(HRP), with Nickel,
(3,30 -diaminobenzidine)
DIG RNA Labeling Mix Sigma-Aldrich Cat#11277073910
FluoReporterTM lacZ Flow Cytometry Kit Thermo Fisher Scientific Cat#F1930
Mesenchymal Stem Cell Chondrogenic PromoCell Cat#C-28012
Differentiation Medium (Ready-to-use)
PicopureTM RNA isolation Kit Thermo Fisher Scientific Cat#KIT0204
RNAscope Multiplex ACDbio Cat#323100
Fluorescent Reagent Kit v2
RNAscope Probe-Mm-Ostn ACDbio Cat#440791
RNeasy Plus Mini Kit QIAGEN Cat#74136
SimpleChIP Plus Enzymatic Chromatin IP kit Cell Signaling Technology Cat#9004
SuperScriptTM III Reverse Transcriptase Thermo Fisher Scientific Cat#18080-044
SuperScriptTM IV VILOTM Master Mix Thermo Fisher Scientific Cat#11756050
Deposited data
Reanalyze processed RNA-seq data from Debnath et al., 2018 GEO: GSE106235
periosteal stem cell
Experimental models: Organisms/strains
B6.129S-OSTNtm1a(KOMP)ncvc Laboratory of Naoki Mochizuki N/A
Ctsktm1(cre)Ska Laboratory of Takashi Nakamura N/A
Gt(ROSA)26Sortm1(CAG-FOXO1,GFP)Moli The Jackson Laboratory JAX:029316; RRID:IMSR_JAX:029316
Gt(ROSA)26Sortm1(EYFP)Cos/J The Jackson Laboratory JAX:006148; RRID:IMSR_JAX:006148
Npr3tm1Unc Laboratory of Nobuyuki Takahashi N/A
Tg(hSAP-Nppc) Laboratory of Akihiro Yasoda N/A
Oligonucleotides
ChIP primers See Table S1 N/A
Cloning primers See Table S1 N/A
Genotyping primers See Table S1 N/A
ON-TARGETplus SMARTpool for FoxO1 DharmaconTM Cat# L-041127-00-0005
qPCR primers See Table S1 N/A
StealthTM RNAi See Table S1 N/A
Software and algorithms
Graph Pad prism 5 GraphPad Software, Inc RRID: SCR_002798; https://scicrunch.org/
resources/Any/search?q=SCR_002798&l=
SCR_002798SCR_002798
ImageJ 1.51 https://imagej.nih.gov/ij/ RRID: SCR 003070; https://scicrunch.org/
resources/Any/search?q=SCR_003070&l=
SCR_003070SCR_003070; https://imagej.nih.gov/ij/
Imaris 8 Bitplane RRID: SCR_007370; https://scicrunch.org/
resources/Any/search?q=SCR_007370&l=
SCR_007370SCR_007370
RStudio Version 1.4.1103 https://www.rstudio.com/ RRID: SCR_000432; https://scicrunch.org/
resources/Any/search?q=SCR_000432&l=
SCR_000432SCR_000432; https://www.rstudio.com/

Cell Reports 36, 109380, July 13, 2021 e2

 ll
OPEN ACCESS Article

RESOURCE AVAILABILITY

Lead contact
Further information and requests for resources and reagents should be directed to lead contact, Naoki Mochizuki (mochizuki@ncvc.
go.jp).

Materials availability
All unique reagents generated in this study are available from the lead contact without restriction.

Data and code availability

This study did not generate any unique datasets or code.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

All protocols for mice were approved by the Institutional Animal Care and Use Committee of National Cerebral and Cardiovascular
Center Research Institute.

Animals
To generate OstnLacZ/LacZ mouse, a targeting vector that harbors LacZ and neomycin cassette (PG00061_Z_2_D09) in Ostn locus
was obtained from KOMP. The targeting vector was linearized with NotI and transfected into R1 ES cells by electroporation (Bio-
Rad) followed by the culture in the medium containing 250 mg/ml G418 (Wako) for 1 week. G418-resistant clones were selected
by genomic PCR and homologous recombination in ES clone was confirmed by genomic PCR or Southern blotting. Chimeric
mice were backcrossed into C57BL/6J background (F9). Animals were able to move freely and access food or water in the cages.
All cages were always kept clean and sanitized. The temperature, humidity, and light in rooms were monitored and maintained appro-
priately. To standardize the size of pups, we used the offspring from heterozygous parent and restricted the number of pups to 5 per
one mother from P7. We routinely determined the genotypes of offspring by genomic PCR using KOD FX Neo DNA polymerase
(TOYOBO) with primers listed in Table S1. All animals used here were males to avoid gender differences in bone metabolism. To
confirm the effect of CNP on impaired osteogenic differentiation of OstnLacZ/LacZ mice in vivo, we used the transgenic mouse over-
expressing human Nppc gene driven by serum amyloid P promoter (Kanai et al., 2017; Miyazaki et al., 2018). To overexpress
FoxO1AAA, a mutant that is constitutively localized to the nucleus, in periosteal osteoblasts, we crossed OstnLacZ/LacZ lines to
Rosa-Lox-STOP-Lox-YFP (The Jackson laboratory), CAG-Lox-STOP-Lox-FoxO1AAA (The Jackson laboratory, stock No.029316),
and Ctsk-Cre (gifted from T. Nakamura) (Nakamura et al., 2007; Ouyang et al., 2012; Srinivas et al., 2001). In TS model, 5-week-
old C57BL/6J mice (purchased from Japan SLC) were applied to hind limb unloading through TS hanging device for 5 to 14 days
as described previously (Gerbaix et al., 2015). Animals can freely access to water and pelleted food during TS. GC mice were raised
in normal condition for 2 weeks and RL mice were kept in normal condition for 3 days or 5 days following TS. When mice were treated
with FoxO1 inhibitor, the mice were intraperitoneally administered with 200 ng/g/day FoxO1 inhibitor (AS1842856) for 7 days. For the
model of unilateral sciatic nerve transection, eight-week-old C57BL/6J mice (purchased from Japan SLC) were used and performed
the surgery according to the method described previously (Baek et al., 2020). Briefly, the mice were anesthetized using constant-flow
isoflurane inhalation, the right side of sciatic nerve in each mouse was exposed, and then removed 5 mm section of the nerve. The left
side of nerve was preserved as a contralateral control.

Primary cell cultures

PDCs were collected according to the modified protocol described previously (van Gastel et al., 2012). Tibiae of 8-week-old male
mice were dissected and immediately soaked into PBS. To eliminate the contamination from perichondrium, both ends of tibiae
were coated with 4% low-melting agarose. Tibiae were incubated with isolating medium consist of aMEM (GIBCO), 3 mg/ml, colla-
genase type II (Worthington), Dispase (GIBCO), GlutaMAXTM (GIBCO), and antibiotics (FUJIFILM) for 10 min at 37 C. The samples
were vortex for 10 s and this first supernatant was discarded. Again, the samples were incubated with new isolating buffer for
60 min at 37 C. The reaction was stopped with aMEM containing 10% FBS, antibiotics, GlutaMAXTM and filtered with 70 mm cell
strainer (FALCON). The cells were collected by centrifuged, suspended with growth medium, and incubated at 37 C in 5% CO2
for 4 days. After the culture for 4 days, the cells were passaged and grown for 3 days. To induce osteogenic and adipogenic differ-
entiation, cells were plated in 24 well at 2.5 3 104 cells per well and grown in normal medium for 48 hours. The osteogenic medium
contains 10 mM b-glycerophosphate, 50 mg/ml ascorbic acid, and the adipogenic medium is supplemented with 500 mM isobutyl-
methylxanthine (IBMX), 2.5 mM dexamethasone (Dex), and 10 mg/ml insulin. For chondrogenic differentiation, 2.5 3 105 cells were
plated in a 96-well U-bottom suspension culture plate. After the spontaneous formation of spheroid for 48 hours, the medium was
exchanged with differentiation medium (PromoCell) with or without CNP-22 (Peptide Institute. Inc.). The differentiation medium
was changed every other day for 14 days.

e3 Cell Reports 36, 109380, July 13, 2021

 ll
Article OPEN ACCESS

METHOD DETAILS

Whole mount X-gal staining and clearing

Detection of LacZ activity in whole mount was performed using X-gal. Animals were anesthetized, perfused with PBS, and fixed with
fixative solution containing [0.1 M Na-PO4 buffer (pH 7.3), 1% paraformaldehyde (PFA), 0.05% glutaraldehyde, 5 mM EGTA, 2 mM
MgCl2, and 0.1% NP-40]. The deskinned animals were incubated in staining solution [0.1 M Na-PO4 buffer (pH 7.3), 2 mM MgCl2,
0.01% sodium deoxycholate (DOC), 0.02% NP-40, 5 mM K-ferricyanide, 5 mM K-ferrocyanide, 20 mM Tris-HCl(pH7.3), and 1 mg/ml
X-gal] for overnight or 1 hour at RT. They were washed with PBS and fixed with 4% PFA/PBS NaN3 for 1 hour at 4 C. Clearing was
carried out with CUBIC according to the modified method described previously (Tainaka et al., 2014). Briefly, animals were soaked in
1/2 CUBIC overnight at 37 C with gentle shaking and transferred to CUBIC for 1 week at 37 C or 28 C.

Whole-mount periosteum staining

For fluorescent detection of periosteum, animals were killed by cervical dislocation and their tibiae were removed. Anteromedial sur-
face of tibiae was exposed by the removal of muscles that were connected to the tibial surface. Tibiae were fixed with 4% PFA/PBS
NaN3 (Wako) at 4 C for 10 min. After washing with PBS NaN3 several times, tibiae were blocked with PBS containing 5% normal goat
serum and 0.3% Triton X-100 at room temperature for 6 hours and immunostained with antibodies for FoxO1 at 4 C for overnight.
Tibiae were washed with PBS NaN3 several times, stained with 2nd antibodies, and washed with PBS NaN3 again. The staining of
nuclei was performed by incubation in PBS containing 4’6-diamidino-2-phenylidole (Sigma-Aldrich) for 30 min at room temperature.
Tibiae were soaked into PBS NaN3 and observed by confocal microscopy (Olympus FV3000).

Laser microdissection
Tibiae from 8-week-old mice were immediately fixed with 4% PFA/PBS NaN3 (Wako) for overnight and then soaked into 20% EDTA
(pH 7.4) solution to decalcification for 2-3 weeks. Decalcifierd-tibiae were soaked in 30% sucrose/PBS and followed by freezing in
Tissue-Tek OCT compound (Sakura). 10 mm thick sections were cut from blocks, placed on the slides for laser microdissection (Carl
Zeiss), wash with EtOH, and dried for 1 minute by a circulator. Each ROI was excised by Laser Microdissection System (Leica) and
corrected. Extraction of mRNA was conducted by PicoPureTM RNA Isolation Kit (Thermo Fisher Scientific) following to the manufac-
ture’s instruction. cDNA was synthesized using Superscript IV VILO Master Mix.

MicroCT
Tibiae of mice at the age of 8 weeks were collected, fixed with 4% PFA/PBS NaN3 (Wako) overnight, and stored in 70% EtOH at 4 C
until analysis. For calculation of bone volume, the 3D microCT data was obtained by ScanX-mate-E90 (Comscantecno Co., Ltd.) ac-
cording to the method described previously (Xu et al., 2016). For trabecular bone analysis, 300 mm section below the tibial proximal
growth plate was used to calculate bone volume. In the case of cortical bone analysis, section of 1 mm at the half of whole tibia was
analyzed. Three-dimensional morphometric parameters were determined by TRI/3D-BON software (RATOC).

Bone length measurement

Bone lengths of calvaria, femora, tibia, and caudal vertebra were measured by Latheta LCT-200 (Hitachi). The lengths of femora and
tibia were calculated by the average length of right and left bones. The lengths of caudal vertebrae (C) were calculated by measuring
the total lengths from C1 to C3.

In situ hybridization for Col10a1

Tibiae were immediately fixed with 4% PFA/PBS NaN3 (Wako) for overnight and then soaked into 20% EDTA (pH 7.4) solution to
decalcification for 2-3 weeks. Tibiae were infused with 30% sucrose and followed by freezing in Tissue-Tek OCT compound (Sakura).
10 mm thick sections were cut from blocks and dried at 37 C for 2 hours. After washing with PBS, slides were treated with 2 mg/ml
Proteinase K (QIAGEN) for 5 min at room temperature. After washing with PBS twice, slides were treated with 2 mg/ml Proteinase K
(Roche) for 15 min at room temperature. After washing with PBS, the slides were fixed again with 4% PFA in 0.1 M Na-PO4 buffer (pH
7.4) for 20 min at RT. The slides were incubated with 0.1 M triethanolamine-HCl (pH 8.0) for 5 min and 0.625 mL acetic anhydride was
added. Slides were pre-hybridized at 60 C for 1 hour and incubated with DIG-labeled probes at 60 C for 16 hours. The slides were
washed with pre-warmed washing buffer (4 3 SSC, 2 3 SSC/50% formamide, 1 3 SSC/50% formamide, 0.2 3 SSC) at 50 C for
30 min. After washing with TNT buffer [0.1 M Tris-HCl (pH 7.5), 0.15 M NaCl, 0.05% Tween-20], the slides were blocked with blocking
solution containing 0.1 M Tris-HCl (pH 7.5), 0.15 M NaCl, 0.5% Blocking Reagent (Roche) for 1 hour, and anti-DIG-AP pAb (Roche)
was applied for overnight at 4 C. The slides were extensively washed with TNT buffer and incubated BM-purple solution (Roche) in
the dark until a signal appears. Primers for the construction of probes are listed in Table S1. DIG-labeled probes were synthesized
using DIG RNA Labeling Mix (Roche). The staining area detected by ALP activity was measured by ImageJ and calculated the height
of the hypertrophic chondrocyte layer.

Cell Reports 36, 109380, July 13, 2021 e4

 ll
OPEN ACCESS Article

In situ hybridization for Ostn

To detect Ostn mRNA, we adopted RNAscope technology and performed ISH on 10 mm of the frozen section of 5-week-old
Ostn+/LacZ and OstnLacZ/LacZ mice according to the manufacture’s protocol of RNAscope Multiplex Fluorescent Reagent Kit v2 (ACD-
bio). The probes used here is RNAscope Probe-Mm-Ostn (#440791).

Tissue preparation and immunohistochemistry

Tibiae were fixed with 4% paraformaldehyde/PBS NaN3 (Wako) overnight. After decalcification with 20% EDTA solution (pH7.4), tis-
sues were embedded in paraffin wax. For histological examination, tissue sections were cut from paraffin-embedded blocks of the
tibia in 4 mm thick and stained with hematoxylin and eosin. For histological analysis of cartilage, Alcian Blue staining was carried out
according to the method described previously (Xu et al., 2016). For immunostaining with anti-Osterix antibody, proteolysis-depen-
dent antigen retrieval was needed. The sections were incubated in 1 mg/ml proteinase K in PBS at room temperature for 5 min.
The sections were permeabilized with 0.3% Triton X-100 in PBS NaN3. After blocking with blocking buffer (5% normal goat serum
in 1 3 PBS NaN3), sections were immunostained with anti-Osterix antibody diluted (1/600) in the buffer containing 3% BSA/ PBS
NaN3 overnight at 4 C. Detection was carried out with HRP-conjugated secondary antibody followed by DAB staining (Vector
Labs). The sections immunostained with anti-Osterix antibody were quantitatively analyzed by ImageJ. The region of 655 mm 3
300 mm was extracted and converted image to 8-bit grayscale.

Sorting of LacZ-positive or YFP-positive cells

Sorting of the cells expressing LacZ was conducted using FluoReporterTM lacZ Flow Cytometry Kit (InvitrogenTM). PDCs were iso-
lated according to the above-mentioned method and reacted with fluorescein di-V-galactoside (FDG). The cells showing high FITC
intensity were sorted by FACS Aria III (BD). For the isolation of YFP-positive cells, cells were separated based on their YFP intensity by
FACS Aria III. SYTOXTMBlue (Thermo Fisher Scientific) was used for eliminating dead cells. Sorted cells were centrifuged and lysed
with TRIzol RNA Isolation Reagents (Thermo Fisher Scientific).

Estimation of Osterix (+) cells by FACS

PDCs were prepared from mice at the age of P11 according to the above-mentioned method. To eliminate red blood cells, cells were
treated with hemolytic reagent and then fixed with 4% paraformaldehyde/PBS NaN3 for 15 min at 4 C. Cells were permeabilized with
0.1% Triton X-100//PBS NaN3 for 5 min and incubated for 30 min on ice in 1% BSA/ PBS NaN3. Cells were stained with anti-Osterix
antibody (1/1200) for 30 min on ice and then detected by A488-conjugated secondary antibody (Thermo Fisher Scientific). Since the
periosteal cells are very small, we used SYTOXTMBlue (Thermo Fisher Scientific) to get out cellular debris.

Determination of lineage-specific cell differentiation

Cell differentiation was determined by the method described previously (Kanai et al., 2017). Mineralization was detected by Alizarin
Red S staining (Muto Pure. Chemicals). Briefly, cells were fixed with 4% PFA/PBS NaN3 and washed with DDW for several times.
Then, the staining solution was added into dishes, and the cells were incubated for 5 min at RT. Samples were washed three times
with DDW and dried. Adipocytes were assessed by Oil Red O staining (Sigma). Formalin-fixed cells were washed with DDW and incu-
bated in 0.3% Oil Red O/60% isopropanol for 30 min at RT. Stained samples were washed with DDW and observed by stereomicro-
scopy (OLYMPUS). To quantify the differentiation states of samples, the stained area of images was measured by ImageJ. For the
detection of chondrogenic differentiation, cells were fixed with a fixative solution containing 30% ethanol, 0.4% PFA, and 4% acetic
acid for 15 min at RT. Then, the cells were stained with Alcian Blue solution (Sigma) overnight at 37 C and extensively washed with
DDW. For quantification, Alcian Blue was extracted by 6 M guanidine hydrochloride solution and measured by optical density at
600 nm (O.D. 600) using a spectrophotometer (Bio-Rad).

OSTN binding assay

Biotin-tagged OSTN was labeled by Cy3-streptavidin (Sigma). Equal volume of 25 mM biotin-tagged OSTN and 0.1 mg/ml Cy3-strep-
tavidin were incubated with on ice for 15 min at 4 C. Cell suspension was incubated with labeled OSTN for 30 min at 37 C. To detect
specific binding between OSTN and bound cells, an excess of biotin free OSTN was added before the addition of labeled OSTN. To
eliminate the nonspecific binding, the cells were extensively washed with PBS and suspended with 5% FBS/PBS. Analysis was con-
ducted by BD FACSAriaTM III (BD Bioscience).

RNA preparation and qPCR

Bones were homogenized by physcotron (nition) and isolated total RNA using TRIzol RNA Isolation Reagents (Thermo Fisher Scien-
tific). When we use periosteum extract, the tibia was first soaked into TRIzol RNA Isolation Reagents for 15 min on ice, followed by
vortex-mixing for a few minutes and then the supernatant was transferred into a new tube (PS). The remaining bone was cut at both
ends and bone marrow cells were flushed out using a syringe filled with cold PBS (BM). The remaining bone was used as a bone
sample (TB/CB). Purification was carried out RNeasy Plus Kit (QIAGEN) and reverse transcription of mRNA was performed using
the SuperScriptTM III Reverse Transcriptase, 10 mM dNTPs, RNase OUT, and Random primers (Thermo Fisher Scientific) according
to the company’s protocol. Quantitative PCR was performed on Mastercycler realplex4 (Eppendorf) or CFX Connect Real-Time Sys-

e5 Cell Reports 36, 109380, July 13, 2021

 ll
Article OPEN ACCESS

tem (BIO-RAD) using KOD SYBR qPCR Mix (TOYOBO) or SSOAdvanced Universal SYBR Green Supermix (BIO-RAD) and specific
primers. Primers used for real-time PCR were listed in Table S1.

Morphometric analysis
For morphometric analysis of bones, 10 ml/g calcein (Dojindo) solution that contains 2.5 mg/ml calcein, 0.15 M NaCl, 2% NaHCO3
was administered intraperitoneally at 2 and 5 days before sacrifice. Tibiae were isolated, immediately soaked in 4% PFA/PBS NaN3
overnight and stored in 70% EtOH at 4 C until analysis. Morphometric parameters of the osteoblasts were calculated (Niigata Bone
Science Institute).

Analysis of processed RNAseq data

Processed RNAseq data deposited previously was reanalyzed. The heatmap was illustrated using heatmap.2 in the R gplots package
according to the previous report (Debnath et al., 2018) .

Knockdown of Npr3 and Foxo1 in PDCs

Synthesized StealthTM RNAi against Npr3 was transfected into PDCs using LipofectamineTM RNAiMAX transfection reagent (Thermo
Fisher Scientific). Two different StealthTM RNAis were mixed in equal amount and 5 pmol mixed StealthTM RNAi was used per well in a
24-well plate. StealthTM RNAi siRNA negative control medium GC (Thermo Fisher Scientific) was used as a negative control. Se-
quences of the two synthesized oligonucleotides are listed in Table S1. For the knockdown of Foxo1, ON-TARGETplus Mouse
Foxo1 siRNA (DharmacomTM) was transfected into PDCs using LipofectamineTM RNAiMAX transfection reagent according to the
manufacture’s protocol.

Analysis of proliferating chondrocytes

For labeling of proliferating chondrocytes, 10 mM EdU diluted with 10% DMSO/PBS was injected intraperitoneally (10 ml/g). The mice
injected with EdU were incubated for 4 hours and dissected. They were immediately fixed with 4% PFA/PBS NaN3, transferred into
20% EDTA (pH 7.4) about 3 weeks, and embedded in paraffin. EdU labeled cells were visualized with Click-iT EdU Alexa Fluor 488
Imaging Kit (Thermo Fisher Scientific) according to the manufacturer’s protocol.

Measurement of cGMP activity

Confluent PDCs in 12 well plates were incubated with serum-free aMEM containing 250 mM IBMX for 10 min before the stimulation.
Stimulation with CNP-22 (Peptide Institute. Inc.) and OSTN (Toray Research Center, Inc) was performed for 15 min at 37 C. cGMP
concentration was determined using Cyclic GMP ELISA Kit (Cayman chemical) according to the manufacture’s protocol.

Western blotting
Isolated tibia was soaked for 15 min into SDS sample buffer containing 50 mM Tris-HCl(pH6.8), 2% SDS, 10% glycerol, 1% b-mer-
captoethanol, 12.5 mM EDTA, and 0.02% bromophenol blue. The samples were boiled for 3 min and separated by SDS-polyacryl-
amide gel electrophoresis using TGX Stain-Free Precast Gel (Bio-Rad). The SDS polyacrylamide gel was transferred to PVDF mem-
brane and immunoblotted with anti-FoxO1 or b-actin. Total protein concentration per each lane was calculated according to the
manufacture’s protocol.

Measurement of OSTN concentration in plasma

OSTN levels in the plasma were measured by a sandwich direct chemiluminescent enzyme immunoassay (CLEIA), which was newly
developed with a slight modification of the previously reported method for ANP CLEIA. The OSTN CLEIA system was comprised of
two anti-mouse OSTN antibodies, one directed to the C terminus and another to the middle region of OSTN (R&D Systems), and
showed no cross-reactivity with mouse ANP, BNP, CNP, and other related peptides. This system also had a high sensitivity of a
detection limit of 0.16 pmol/l (Nagai and Minamino, 2014).

ChIP-qPCR assay
ChIP assay was performed using SimpleChIP Plus Enzymatic Chromatin IP kit (Cell Signaling Technology) according to the man-
ufacture’s protocol. Confluent of PDCs in 15 cm dish was cross-linked with DMEM containing 0.5% formaldehyde, extract DNA, and
digested DNA with micrococcal nuclease for 20 min. The fixed DNA-protein complex was immunoprecipitated with anti-FoxO1 (Cell
Signaling Technology), control anti-Rabbit IgG (negative control), or anti-HistonH3 (positive control). After the extensive washing,
precipitated DNA was eluted and purified. Quantitative PCR was performed on CFX Connect Real-Time System (BIO-RAD) using
KODSYBR qPCR Mix (TOYOBO) with the primers listed in Table S1 or simpleChIP Mouse RPL30 Intron2 primers (CST).

QUANTIFICATION AND STATISTICAL ANALYSIS

All results represent mean ± SEM unless otherwise noted. Statistical analysis was performed by either 2-tailed Student’s t test or
1-way ANOVA followed by Tukey-Kramer test using GraphPad Prism. Statistical significance was defined as p < 0.05.

Cell Reports 36, 109380, July 13, 2021 e6