Central control of bone remodeling by neuromedin U

Abstract

Bone remodeling, the function affected in osteoporosis, the most common of bone diseases, comprises two phases: bone formation by matrix-producing osteoblasts and bone resorption by osteoclasts. The demonstration that the anorexigenic hormone leptin inhibits bone formation through a hypothalamic relay suggests that other molecules that affect energy metabolism in the hypothalamus could also modulate bone mass. Neuromedin U (NMU) is an anorexigenic neuropeptide that acts independently of leptin through poorly defined mechanisms. Here we show that Nmu-deficient (Nmu-/-) mice have high bone mass owing to an increase in bone formation; this is more prominent in male mice than female mice. Physiological and cell-based assays indicate that NMU acts in the central nervous system, rather than directly on bone cells, to regulate bone remodeling. Notably, leptin- or sympathetic nervous system-mediated inhibition of bone formation was abolished in Nmu-/- mice, which show an altered bone expression of molecular clock genes (mediators of the inhibition of bone formation by leptin). Moreover, treatment of wild-type mice with a natural agonist for the NMU receptor decreased bone mass. Collectively, these results suggest that NMU may be the first central mediator of leptin-dependent regulation of bone mass identified to date. Given the existence of inhibitors and activators of NMU action, our results may influence the treatment of diseases involving low bone mass, such as osteoporosis.

Full text

Central control of bone remodeling by neuromedin U
Shingo Sato
1
, Reiko Hanada
2
, Ayako Kimura
1
, Tomomi Abe
3
, Takahiro Matsumoto
4,5
, Makiko Iwasaki
1
,
Hiroyuki Inose
1
, Takanori Ida
2
, Michihiro Mieda
3
, Yasuhiro Takeuchi
6
, Seiji Fukumoto
7
, Toshiro Fujita
7
,
Shigeaki Kato
4,5
, Kenji Kangawa
8
, Masayasu Kojima
2
, Ken-ichi Shinomiya
1
& Shu Takeda
1
Bone remodeling, the function affected in osteoporosis, the
most common of bone diseases, comprises two phases: bone
formation by matrix-producing osteoblasts1and bone resorption
by osteoclasts2. The demonstration that the anorexigenic
hormone leptin3–5 inhibits bone formation through a
hypothalamic relay6,7 suggests that other molecules that affect
energy metabolism in the hypothalamus could also modulate
bone mass. Neuromedin U (NMU) is an anorexigenic
neuropeptide that acts independently of leptin through poorly
defined mechanisms8,9.HereweshowthatNmu-deficient
(Nmu
–/–
) mice have high bone mass owing to an increase in
bone formation; this is more prominent in male mice than
female mice. Physiological and cell-based assays indicate that
NMU acts in the central nervous system, rather than directly
on bone cells, to regulate bone remodeling. Notably, leptin- or
sympathetic nervous system–mediated inhibition of bone
formation6,7 was abolished in Nmu
–/–
mice, which show an
altered bone expression of molecular clock genes (mediators of
the inhibition of bone formation by leptin). Moreover, treatment
of wild-type mice with a natural agonist for the NMU receptor
decreased bone mass. Collectively, these results suggest that
NMU may be the first central mediator of leptin-dependent
regulation of bone mass identified to date. Given the existence
of inhibitors and activators of NMU action10, our results may
influence the treatment of diseases involving low bone mass,
such as osteoporosis.
Bone mass is maintained at a constant level between puberty and
menopause by a succession of bone-resorption and bone-formation
phases11,12. The discovery that neuronal control of bone remodeling is
mediated by leptin6shed light on a new regulatory mechanism of bone
remodeling and also suggested that bone mass may be regulated by a
variety of neuropeptides13. In line with this observation, cannabinoids
and pituitary hormones have been shown to be intimately involved in
bone remodeling14,15. Leptin inhibits bone formation by binding to
its receptors located in hypothalamus and thereby activating the
sympathetic nervous system (SNS), which requires the adrenergic b2
receptors (Adrb2) expressed in osteoblasts7,16. Downstream of Adrb2,
leptin signaling activates molecular clock genes that regulate osteoblast
proliferation and hence bone formation17. In addition, leptin regulates
bone resorption through two distinct pathways16.
NMU is a small peptide produced by nerve cells in the submucosal
and myenteric plexuses in the small intestine, and also by structures in
the brain, including the dorsomedial nucleus of the hypothalamus9.It
is generally assumed that NMU acts as a neuropeptide to regulate
various aspects of physiology, including appetite, stress response and
SNS activation9. Indeed, NMU-deficient (Nmu
–/–
) mice develop
obesity due to increased food intake and reduced locomotor activity
that is believed, at least in part, to be leptin independent8.Inaddition,
expression of NMU is diminished in leptin-deficient (Lep
ob
) mice18,
but can be induced in these mice by leptin treatment19.Insearchof
additional neuropeptides that regulate bone remodeling, we analyzed
Nmu
–/–
mice.
When assessed at 3 and 6 months of age, both male and female
Nmu
–/–
mice showed a high bone mass phenotype as compared to the
wild type (WT), with male mice more severely affected than female
mice (Fig. 1a and data not shown). The presence of a uniform increase
in bone mineral density (BMD) along the femurs of Nmu
–/–
mice
suggested that both trabecular and cortical bone were equally affected
(Supplementary Fig. 1 online). Microcomputed tomography analysis
confirmed this observation (Fig. 1b,c). To determine whether this
phenotype was secondary to the obesity of the Nmu
–/–
mice, we
restricted their food intake for 1 month starting at 2 months of age.
This manipulation normalized the body weight and serum insulin
level of the Nmu
–/–
mice but did not affect their high bone mass
phenotype (Fig. 1d and data not shown). Of note, when Nmu
–/–
mice
were backcrossed to the C57BL/6J genetic background, their body
weight became similar to that of their WT littermates; however,
their BMD remained high (data not shown). These results suggest
that NMU regulates bone mass independently of its regulation of
energy metabolism, just as leptin does7.Tobettercharacterizethe
cellular nature of the bone phenotype in the Nmu
–/–
mice, we
© 2007 Nature Publishing Group http://www.nature.com/naturemedicine
Received 4 June; accepted 8 August; published online 16 September 2007; doi:10.1038/nm1640
1
Department of Orthopaedic Surgery, Graduate School, 21
st
Century Center of Excellence Program, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku,
Tokyo 113-8519, Japan.
2
Division of Molecular Genetics, Institute of Life Science, Kurume University, 1-1 Hyakunen-kohen, Kurume, Fukuoka 839-0842, Japan.
3
Department of Molecular Neuroscience, Tokyo Medical and Dental University 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8519, Japan.
4
Institute of Molecular and
Cellular Biosciences, University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo 113-0032, Japan.
5
Exploratory Research for Advanced Technology, Japan Science and
Technology Agency, 4-1-8 Honcho, Kawaguchi, Saitama 332-0012, Japan.
6
Toranomon Hospital Endocrine Center, 2-2-2 Toranomon, Minato-ku, Tokyo 105-8470,
Japan.
7
Division of Nephrology and Endocrinology, Department of Internal Medicine, University of Tokyo Hospital, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8655, Japan.
8
Department of Biochemistry, National Cardiovascular Center Research Institute, 5-7-1 Fujishiro-dai, Suita-shi, Osaka 565-8565, Japan. Correspondence should be
addressed to S.T. (shu-tky@umin.ac.jp).
NATURE MEDICINE ADVANCE ONLINE PUBLICATION 1
LETTERS

performed histological and histomorphometric analyses of vertebrae
and tibiae in both male and female animals (Fig. 1e and Supplemen-
tary Fig. 1).At3and6monthsofage,Nmu
–/–
mice showed greater
bone volume in both vertebrae and tibiae than did WT littermates,
with male mice having a more pronounced phenotype (Fig. 1e and
Supplementary Fig. 1). At the present time we do not have a clear
explanation of the difference in phenotype severity between male and
female mice. Bone formation rates (WT mice, 146.9 ± 12.3, Nmu
–/–
mice, 183.7 ± 10.3, Po0.05) and osteoblast numbers were both
significantly greater in the vertebrae and tibiae of Nmu
–/–
mice (Fig. 1f
and Supplementary Fig. 1). The higher osteoblast numbers in the
presence of a normal mineral apposition rate (Fig. 1f and Supple-
mentary Fig. 1), which reflects the function of individual osteo-
blasts20, suggested that osteoblast proliferation may be increased in
Nmu
–/–
mice. Indeed, 5-bromo-2-deoxyuridine (BrdU)-positive pro-
liferative osteoblast counts were significantly increased in Nmu
–/–
mice
in vivo (Fig. 1g), demonstrating that NMU affects osteoblast prolif-
eration. In contrast, Nmu
–/–
and WT mice showed comparable
osteoclast numbers and osteoclast surface areas (Fig. 1f and Supple-
mentary Fig. 1), suggesting that NMU does not affect bone resorp-
tion. This observation was further supported by the normal level of
urinary elimination of deoxypyridinoline in Nmu
–/–
mice (Fig. 1h).
Taken together, these results demonstrate
that NMU deficiency results in an isolated
increase in bone formation leading to high
bone mass. Nmu-heterozygote mice did not
have an overt bone abnormality at any age
analyzed (Fig. 1e).
Two cognate G protein–coupled receptors
have been reported to be NMU receptors:
NMUR1, which is expressed in various
tissues, including the small intestine and
lung (data not shown), and NMUR2, which
is predominantly expressed in the hypothala-
mus and the small intestine (Fig. 2a)18. Both
receptors and NMU itself were barely detectable in bone (Fig. 2a). To
further exclude the possibility of a direct action of NMU on osteo-
blasts, we treated mouse primary osteoblasts with varying concentra-
tions of NMU. Alkaline phosphatase activity, mineralization and
expression of osteoblastic genes were all unaffected by this treatment
(Fig. 2b,c). In addition, there were no differences between WT mice
and Nmu
–/–
mice in the expression of osteoblastic genes in vivo
(Fig. 2d). Moreover, both WT and Nmu
–/–
osteoblasts proliferated
normally in vitro in response to NMU treatment (Fig. 2e), though
Nmu
–/–
osteoblasts proliferated more than WT osteoblasts
in vivo (Fig. 1g). Osteoclastic differentiation from bone marrow
macrophages was unchanged by NMU treatment (Fig. 2f), as expected
from the absence of a bone resorption defect in vivo (Fig. 1f,h). Taken
together, these results strongly suggest that NMU’s effect on bone may
not come from its direct action on osteoblasts, but rather through
another relay.
Because the anorexigenic effect of NMU requires a hypothalamic
relay8,19 and because hypothalamic neurons have been shown to
regulate bone mass, we tested whether NMU’s regulation of bone
formation could involve a central relay. Continuous intracerebroven-
tricular (i.c.v.) infusion of NMU into Nmu
–/–
mice decreased their fat
mass and fat pad weight significantly, although body weight was not
© 2007 Nature Publishing Group http://www.nature.com/naturemedicine
0
2
4
6
8
10
Oc.S/BS (%)
0
5
10
15
20
Ob.S/BS (%)
*
0
2
4
6
8
10
12
DPD
(nmol/mmol creatinine)
abc
f
g
0
200
400
600
800
MAR (µm/year)
0
50
100
150
200
*
BFR/BS (µm3/µm2/year)
0
10
20
30
40
50
60
Tb.Th (µm)
*
h
0
2
4
6
8
10
12
Osteoblast mitotic
index (%)
*
0
0.2
0.4
0.6
0.8
1.0
1.2
Cortical cross-sectional
area (mm2)
**
ed
6 months
**
Cortical thickness (µm)
0
50
100
150
200
250
300
**
3 months
*
0
10
20
30
40
50
60
Nmu–/–
WT
Nmu–/–
WT
Nmu–/–
WT
Nmu–/–
WT
Nmu–/–
WT
Nmu–/–
Nmu+/–
WT
Nmu–/–
WT
Nmu–/–
Nmu–/–
WT
BMD (mg/cm2)
BMD (mg/cm2)
0
10
20
30
40
50
60
BMD (mg/cm2)
Pair-fed
**
0
10
20
30
40
50
60
0
5
10
15
20
25
30
35
Body weight (g)
Pair-fed
VertebraTibia
15.5 ± 0.5 14.7 ± 1.0 19.2 ± 0.7**
17.0 ± 0.5 17.5 ± 0.4 21.9 ± 0.9**
BV/TV (%)
BV/TV (%)
WT
Nmu–/–
WT Nmu–/–
WT
Nmu–/–
WT
Nmu–/–
WT Nmu–/–
WT
Nmu–/–
WT Nmu–/–
WT
Figure 1 High bone mass in Nmu
–/–
mice due
to increased bone formation. (a) Bone mineral
density (BMD) of the femurs of 3 (left)- and
6 (right)-month-old male wild-type (WT) and
Nmu
–/–
mice. (b) Micro-computed tomography
(mCT) analysis of the distal femurs of male mice
at 3 months. Scale bars, 500 mm. (c) Cortical
thickness and cross-sectional area of the femurs
of 3-month-old male mice. (d) Body weight and
BMD of 3-month-old male mice with restricted
food intake. (e) Histological analysis of the
vertebrae and tibiae of 3-month-old male WT,
Nmu
+/–
and Nmu
–/–
mice. Bone volume per
tissue volume (BV/TV). Scale bars, 1 mm.
(f) Histomorphometric analysis of the vertebrae of
3-month-old male mice. Mineral apposition rate
(MAR), bone formation rate over bone surface
area (BFR/BS), osteoblast surface area over bone
surface area (Ob.S/BS), trabecular thickness
(Tb. Th) and osteoclast surface area over bone
surface area (Oc.S/BS). (g) Increased osteoblast
proliferation in newborn Nmu
–/–
mice. Immuno-
localization of BrdU incorporation (arrows) in
the calvariae of WT and Nmu
–/–
mice (left).
Osteoblast mitotic index (right). Scale bar,
20 mm. (h) Urinary elimination of deoxy-
pyridinoline (DPD) in WT and Nmu
–/–
mice.
**, Po0.01; *, Po0.05.
LETTERS
2ADVANCE ONLINE PUBLICATION NATURE MEDICINE

affected (Fig. 2g and Supplementary Fig. 2 online). In addition,
NMU i.c.v. infusion eliminated the high bone mass phenotype in
Nmu
–/–
mice (Fig. 2g and Supplementar y Fig. 2), suggesting that
NMU inhibits bone formation through the central nervous system.
The central nature of bone remodeling regulation by NMU, along
with the notion that the anorexigenic effect of NMU may be
independent of leptin8, prompted us to examine whether leptin
could be involved in the regulation of bone formation by NMU. To
address this question, we performed i.c.v. infusion of NMU or leptin
in Lep
ob
mice. NMU decreased fat pad weight significantly, albeit to a
milder extent than that achieved by leptin (Fig. 3a and Supplemen-
tary Fig. 3 online). Body weight was not significantly changed by the
NMU infusion, indicating that this treatment had only a mild effect
on energy metabolism (data not shown). In contrast, NMU decreased
bone mass in Lep
ob
mice as efficiently as leptin did (Fig. 3a). These
results indicate that NMU inhibits bone formation in a leptin-
independent manner. Next, we asked whether leptin could correct
the high bone mass phenotype of Nmu
–/–
mice. Leptin i.c.v. infusion
decreased bone volume and bone formation in WTmice, as previously
reported (Fig. 3b and Supplementary Fig. 3)6. However, the leptin
paradoxically increased bone volume and osteoblast number in
Nmu
–/–
mice (Fig. 3b,cand Supplementary Fig. 3). The fact that
leptin decreased fat mass and fat pad weight in Nmu
–/–
mice and
increased urinary elimination of normetanephrine, a metabolite of
noradrenaline17, verified that the administration of leptin was properly
performed (Fig. 3b,dand Supplementary Fig. 3). Therefore, taken
together, these results suggest that NMU acts downstream of leptin to
regulate bone formation.
© 2007 Nature Publishing Group http://www.nature.com/naturemedicine
ac
de
g
b
Atf4Sp7Runx2BglapAkp2
1.5
1.0
0.5
0
Nmu–/–
WT Nmu–/–
Nmu–/–
WT
Fold change mRNA
f
0
5
10
15
20
25
0
TRAP-positive cells
(per mm2)
0
5
10
15
20
25
30
35
40
NMUPBS
Nmu–/– Nmu–/–
NMUPBS
Body weight (g)
0
0.1
0.2
0.3
0.4
0.5
0.6
Perigonadal fat (g)
*
0
0.2
0.4
0.6
0.8
1.0
1.2
Atf4Sp7BglapAkp2
Fold change mRNA
0
Runx2
0
0.2
0.4
0.6
0.8
1.0
1.2
Cell proliferation,
fold induction
0
0.1
0.2
0.3
0.4
0.5
10–12M
10–12M
10–10M
10–10M
10–10M NMU
10–8M
10–8M
10–8M NMU
0
AKP2
(µmol per mg protein per min)
0
10–12M10
–10M10
–8M NMU
10–10M10
–8M10
–6M NMU
0
0
5
10
15
20
25
NMUPBS
*NMUPBS
Tibia Vertebra
18.4 ± 0.6 14.9 ± 0.9*
20.6 ± 0.7 16.7 ± 0.4*
BV/TV
(
%
)
BV/TV (%)
Nmu Nmur1 Nmur2
Hypothalamus
Nmu Nmur1 Nmur2
Femur
NMU
NMU
Fat (%)
Figure 2 Absence of NMU’s direct effect on osteoblasts; decrease in bone mass by NMU i.c.v. infusion. (a) Expression of Nmu,Nmur1 and Nmur2 in
the hypothalamus at the atlas-levels of 38 (top) and 43 (bottom) and in the femur as shown by in situ hybridization. Note the expression of Nmu in the
dorsomedial nucleus of the hypothalamus (DMH) (bottom) and Nmur2 in paraventricular nucleus (top), arcuate nucleus and DMH (bottom). Scale bars,
500 mm. (b–d) Effect of NMU on osteoblast differentiation. (b,c) WT osteoblasts treated with NMU. (b) Alkaline phosphatase (AKP2) activity (top),
mineralized nodule formation (bottom). (c) Expression of osteoblastic genes (Akp2,Bglap,Runx2,Sp7 and Atf4), depicted as fold change over WT
expression. (d) Expression of osteoblastic genes in WT and Nmu
–/–
femurs. (e) Effect of NMU on osteoblast proliferation. WT or Nmu
–/–
osteoblasts treated
with NMU. (f) Effect of NMU on osteoclast differentiation. Bone marrow–derived osteoclasts treated with NMU. (g) Effect of NMU i.c.v. infusion on body
weight, fat pad weight (perigonadal fat) and fat mass (left). Histological analysis of the vertebrae (top right) and tibiae (bottom right). Male mice at 3 months
of age were used. Scale bars, 1 mm. *, Po0.05.
LETTERS
NATURE MEDICINE ADVANCE ONLINE PUBLICATION 3

The SNS is a major mediator of leptin’s antiosteogenic action7.
NMUR2 is expressed in paraventricular nuclei, whose neurons directly
project to the sympathetic preganglionic neurons, and NMU stimu-
lates sympathetic outflow9,21. These observations, along with the fact
that Nmu
–/–
mice have osteoblastic defects similar to the one observed
in Adrb2-deficient mice16, prompted us to explore whether NMU and
sympathetic tone are in the same pathway regulating bone formation.
Indeed, Nmu/Adrb2 double heterozygote mice had higher bone mass
than Adrb2 single heterozygote mice (Fig. 4a), although Nmu single
heterozygote mice had normal bone mass (Fig. 1e and Supplemen-
tary Fig. 1). Given that Nmu expression in the hypothalamus was
reduced in Nmu single heterozygote mice (data not shown), com-
pound heterozygosity of Nmu and Adrb2 may have resulted in higher
bone mass. Furthermore, this result suggests that these two pathways
share a common molecule. Of note, Nmu
–/–
mice had a higher degree
of urinary elimination of normetanephrine than WT littermates
(Fig. 4b), which would decrease bone mass, yet they had high bone
mass. This suggests that their high bone mass phenotype is not caused
by decreased SNS activity, but is instead the result of resistance to the
antiosteogenic activity of the SNS. This is in agreement with the
observation that i.c.v. infusion of leptin, a potent stimulator of SNS
activity, did not decrease bone mass in Nmu
–/–
mice (Fig. 3b and
Supplementary Fig. 3). Furthermore, injection of isoproterenol, a
sympathomimetic, reduced bone mass in WTmice7but not in Nmu
–/–
mice (Fig. 4c and Supplementar y Fig. 4 online). Thus, Nmu
–/–
mice
are resistant to the antiosteogenic effects of both leptin and the SNS.
We present six experimental arguments to strongly suggest that the
failure of leptin or isoproterenol to decrease bone mass in Nmu
–/–
mice is not due to leptin-SNS signaling defects. First, leptin infusion
decreased fat pad weight equally well in WT and in Nmu
–/–
mice and
could increase normetanephrine abundance in Nmu
–/–
mice (Fig. 3b,d
and Supplementary Fig. 3). Second, the expression of Adrb2 was not
different in WT and Nmu
–/–
bones (Fig. 4d). Third, treatment with
NMU did not affect Adrb2 expression in osteoblasts (Supplementary
Fig. 5 online). Fourth, isoproterenol induced expression of Tnfsf11
(encoding tumor necrosis factor superfamily, member 11) and
decreased expression of Tnfrsf11b (encoding tumor necrosis factor
superfamily, member 11b, also known as osteoprotegerin), Runx2
(encoding runt-related transcription factor-2) and Col1a1 (encoding
collagen type I), molecular markers for the effect of SNS activation on
osteoblasts, in both WT and Nmu
–/–
osteoblasts (Fig. 4d). Fifth,
isoproterenol induced cAMP production equally well in WT and
Nmu
–/–
osteoblasts (Fig. 4e). Sixth, and most notably, leptin increased
bone resorption to a similar extent in WT and Nmu
–/–
mice (Fig. 3c
and Supplementary Fig. 3).
The fact that the leptin-SNS pathway is intact in Nmu
–/–
mice,
together with the paradoxical increase in osteoblast number induced
by leptin i.c.v. infusion in Nmu
–/–
mice (Fig. 3c), suggests that NMU
affects only the negative regulator of bone remodeling by leptin, that
is, the molecular clock. Indeed, the expression of Per1 and Per2
(encoding period homolog-1 and -2, respectively) was downregulated
in Nmu
–/–
bones as compared to WT bones (Fig. 4f and Supplemen-
tary Fig. 6 online). Thus, NMU, acting through the central nervous
system, affects the molecular clock in bone.
Because bone resorption in Nmu
–/–
mice was comparable to that in
the wild type, despite the high SNS activity in these mice, we also
© 2007 Nature Publishing Group http://www.nature.com/naturemedicine
ab
Perigonadal fat (g)
0
1
2
3
4
5
6
7
8
NMULeptinPBS
**
*
Lepob Lepob
Lepob
Lepob
c
d
0
20
40
60
80
100
*
Normetanephrine
(ng/mmol creatinine)
0
200
400
600
800
LeptinPBS
MAR (µm/year)
0
50
100
150
200
BFR/BS (µm3/µm2/year)
**
0
2
4
6
8
10
N.Ob/B.Pm (/mm)
**
0
3
6
9
12
15
LeptinPBS
Ob.S/BS (%)
*
Perigonadal fat (g)
0
0.1
0.2
0.3
0.4
0.5
0.6
LeptinPBSLeptinPBS
*
**
**
0
5
10
15
20
25
LeptinPBSLeptinPBS
BV/TV (%)
*
*
*
0
5
10
15
20
25
LeptinPBSLeptinPBS
Fat (%)
*
*
*
BV/TV (%)
0
5
10
15
20
NMU
Leptin
PBS NMU
Leptin
PBS
*
*
Cortical thickness (µm)
0
50
100
150
200
250
**
*
Oc.S/BS (%)
0
2
4
6
8
10
*
LeptinPBS LeptinPBSNMULeptinPBS
Nmu–/–
WT
Nmu–/–
WT
Nmu–/–
WT
Nmu–/–
Nmu –/– Nmu –/– Nmu –/– Nmu –/– Nmu –/– Nmu –/–
LeptinPBS LeptinPBS
LeptinPBS LeptinPBS
WT
Figure 3 Leptin does not eliminate high bone mass in Nmu
–/–
mice. (a) Effect of NMU or leptin i.c.v. infusion in Lep
ob
mice (3-month-old males). Fat pad
weight and bone mass were determined by histology and cortical thickness by mCT analysis. (b–d) Effect of leptin i.c.v. infusion on Nmu
–/–
mice (3-month-old
males). (b) Fat pad weight, fat mass and bone mass shown by histology. (c) Histomorphometric analysis. N. Ob/B.Pm indicates the number of osteoblasts per
bone perimeter. (d) Urinary elimination of normetanephrine. Scale bars, 1 mm. **, Po0.01; *, Po0.05.
LETTERS
4ADVANCE ONLINE PUBLICATION NATURE MEDICINE

© 2007 Nature Publishing Group http://www.nature.com/naturemedicine
a
b
c
d
0
20
40
60
80
100
WT Nmu –/–
Nmu –/–
Nmu–/–
Nmu–/–
*
Normetanephrine
(ng/mmol creatinine)
Fold change Tnfrsf11b mRNA
0
0.2
0.4
0.6
0.8
1.0
1.2 **
––
Fold change Tnfsf11 mRNA
ISO ISO
0
10
20
30
40
50 **
**
e
Fold change Runx2 mRNA
0
0.2
0.4
0.6
0.8
1.0
1.2 **
Fold change Col1a1 mRNA
0
0.2
0.4
0.6
0.8
1.0
1.2 **
cAMP (pmol/ml)
0
50
100
150
200
250
PTH 10 nM
PTH 1 nM
ISO 1 µMControl
f
0
0.2
0.4
0.6
0.8
1.0
1.2
Fold change Adrb2 mRNA
WT
Nmu –/–
WT
Nmu –/–
WT Nmu –/–
WT
––
ISO ISO
Nmu –/–
WT
––
ISO ISO
Nmu –/–
WT
––
ISO ISO
Nmu –/–
Nmu –/–
WT
g
h
i
Rutin
Control
Fold change Cartpt mRNA
0
0.5
1.0
1.5
WT
Nmu–/–
WT
*
0
5
10
15
20
RutinControl
BV/TV (%)
*
*
**
0
5
10
15
20
ISOPBSISOPBS
BV/TV (%)
WT
0
5
10
15
20
BV/TV (%)
*
0
1
2
3
4
5
6
7
8
2218141062ZT
Fold change Per1 mRNA
**
ISOPBS
ISOPBS
WT
Leptin Leptin
Osteoblast
proliferation
SNS
Clock
Adrb2
NMU
receptor
Hypothalamus
NMU
Leptin
receptor
Osteoblast
Leptin
receptor
Osteoblast
proliferation
Clock
Adrb2
NMU
receptor
NMU
SNS
Nmu+/+
Adrb2+/–
Nmu+/–
Adrb2+/–
Figure 4 Sympathetic activation does not rescue high bone mass in Nmu
–/–
mice. (a)BonemassinAdrb2
+/–
/Nmu
+/–
and Adrb2
+/–
/Nmu
+/+
mice as determined
by histology (3-month-old males). (b) Increased urinary elimination of normetanephrine in Nmu
–/–
mice. (c) Effect of sympathetic activation by isoproterenol
(ISO) injection in Nmu
–/–
mice (3-month-old males). Shown is the bone mass of vertebrae as determined by histology. (d)ExpressionofAdrb2 in the femurs
of WT and Nmu
–/–
mice (left). Gene expression changes induced by isoproterenol (ISO) treatment of WT and Nmu
–/–
osteoblasts (four rightmost graphs).
(e) cAMP concentration in the culture medium of WT and Nmu
–/–
osteoblasts after ISO treatment. Parathyroid hormone (PTH) was used as a control.
(f)ExpressionofPer1 in the femurs of WTand Nmu
–/–
mice. Zeitberger time (ZT) is indicated on the x-axis. (g)ExpressionofCartpt in the hypothalamus
of WT and Nmu
–/–
mice. (h) Rutin decreases bone mass in WT mice as determined by histological analysis of vertebrae (left) and quantitative
histomorphometric analysis (right) (3-month-old males). Scale bar, 1 mm. **, Po0.01; *, Po0.05. (i) Model of leptin, sympathetic nervous system
(SNS) and NMU signaling for the regulation of bone formation in WT mice (left) and Nmu
–/–
mice (right).
LETTERS
NATURE MEDICINE ADVANCE ONLINE PUBLICATION 5

tested whether the expression of Cartpt (encoding cocaine- and
amphetamine-regulated transcript propeptide), a central mediator of
leptin’s action on bone resorption16, was altered in these mice. Indeed,
Cartpt expression was increased in Nmu
–/–
mice as compared to WT
littermates (Fig. 4g and Supplementary Fig. 7 online). These results
suggest that the protective activity of Cart on bone resorption
compensates for the bone-resorbing activity induced by the SNS in
Nmu
–/–
mice. The effect of other leptin-regulated neuropeptides, such
as NPY (neuropeptide Y), AgRP (agouti-related protein) and a-MSH
(a-melanotropin), will be limited, because the expression of Npy and
Agrp was unchanged in Nmu
–/–
mice8and melanocortin 4 receptor, a
major receptor for a-MSH, has been shown to have little effect on
bone remodeling by itself 22.
Lastly, we treated WT mice with rutin, a natural NMUR2 agonist
found in daily foods such as buckwheat23. Consistent with the high
bone mass phenotype of the Nmu
–/–
mice, rutin decreased bone mass
significantly in WT mice (Fig. 4h). This result, together with the
predominant expression of Nmur2 in the hypothalamus (Fig. 2a),
suggests that NMU regulates bone remodeling through NMUR2.
Collectively, these results suggest that NMU, through a central relay
and via an unidentified pathway, acts as a modulator of leptin-
SNS-Adrb2 regulation of bone formation (Fig. 4i). However, one
concern still remains: because leptin affects several pathways originat-
ing in the hypothalamus and elsewhere in the brain, i.c.v. infusion of
leptin may have resulted in an uncoordinated change in leptin-
regulated bone remodeling that does not reflect a physiological role
of leptin. To rigorously address that question, an analysis of a mouse
model in which a specific nucleus of the hypothalamus is activated by
leptin will be necessary. From a therapeutic point of view, given the
lack of an obesity phenotype in Nmur2-deficient mice24,anNMU
antagonist may be a candidate for the treatment of bone-loss disorders
without inducing unwanted body weight gain.
METHODS
Animals. Nmu
–/–
and Adrb2
–/–
mice were previously described8,16.Wepur-
chased C57BL/6J mice and C57BL/6J Lep
ob
from the Jackson Laboratory. We
maintained all of the mice under a 12 hr light-dark cycle with ad libitum access
to regular food and water, unless specified. For pair-fed experiments, we caged
Nmu
–/–
and WT mice individually for 12 weeks as described8. In brief, Nmu
–/–
mice were given access to water ad libitum and fed the amount of chow eaten
on the previous day by a WT littermate. We determined mouse genotypes by
PCR as previously described8,16. We injected isoproterenol (10 mg/kg, Sigma)
intraperitoneally (i.p.) once daily for 4 weeks. Rutin (Sigma) was administered
orally 300 mg per kg body weight per day for 4 weeks. All animal experiments
were performed with the approval of the Animal Study Committee of Tokyo
Medical and Dental University and conformed to relevant guidelines and laws.
Dual X-ray absorptiometry and microcomputed tomography analysis. We
measured bone mineral density (BMD) of the femurs and fat pad composition
by DCS–600 (Aloka). We obtained two-dimensional images of the distal femurs
by microcomputed tomography (mCT, Comscan). We measured cortical
thickness and cross-sectional area at the center of the femur. We examined at
least eight mice for each group.
Histological and histomorphometric analysis. We injected calcein (25 mg/kg,
Sigma) i.p. 5 and 2 d before sacrifice. We stained undecalcified sections of
the third and fourth lumbar vertebrae and tibiae with von Kossa staining.
We performed static and dynamic histomorphometric analyses using the
Osteomeasure Analysis System (Osteometrics). We analyzed 8–10 mice for
each group.
In situ hybridization analysis. We performed in situ hybridization analysis
according to the established protocol25. Antisense cRNA probe for Cartpt was
previously described26. We used fragments of cDNA for Nmu (105 base pairs
upstream to 647 base pairs downstream of the initiation codon), Nmur1
(13–1242 base pairs downstream of the initiation codon) and Nmur2 (16–1252
base pairs downstream of the initiation codon) to generate antisense probes.
We stained sections hybridized with
35
S-labeled probes with Hoechst 33528
and quantitatively analyzed the expression of Cartpt with a phosphorimager
(Bass–2500, Fuji). The atlas-level of designations corresponds to those
described previously27. We analyzed six mice for each group.
Measurement of deoxypyridinoline cross-links and normetanephrine. We
measured urinary deoxypyridinoline cross-links (DPD) and normetanephrine
with the METRA DPD-EIA kit (Quidel) and the Normetanephrine-ELISA kit
(ALPCO), respectively, according to the manufacturer’s instructions. We used
creatinine values to standardize between samples (Creatinine Assay Kit,
Cayman). We examined eight samples for each group.
Cell culture. In vitro primary osteoblast cultures were established as previously
described6. Briefly, we cultured primary osteoblasts from calvariae of 4-d-old
mice in a-MEM (Sigma) containing ascorbic acid (0.1 mg/ml, Sigma). We
added NMU to the medium twice daily. After 14 d, we measured alkaline
phosphatase activity with the ALP kit (Wako). For the mineralization assay, we
supplemented the medium with b-glycerophosphate (5 mM, Sigma). We
assessed mineralized nodule formation by von Kossa staining. We performed
the cell proliferation and cAMP assays with the Cell Proliferation Assay
(Promega) and cAMP EIA kit (Cayman Chemical), respectively. In vitro
osteoclast differentiation has been described previously16. Briefly, bone marrow
cells of 2-month-old mice were cultured in the presence of human macrophage
colony-stimulating factor (10 ng/ml, R&D Systems) for 2 d and then differ-
entiated into osteoclasts with human RANKL (50 ng/ml, Peprotech) and
human macrophage colony-stimulating factor (10 ng/ml) for 3 d. We counted
tartrate-resistant acid phosphatase (TRAP)-positive multinucleated cells (more
than 3 nuclei). We performed all the cell cultures in triplicate or quadruplicate
wells and repeated more than 3 times.
BrdU immunohistochemistry. For BrdU labeling, we injected 100 mgBrdUi.p.
into 3-d-old mice 1 h before sacrifice. We embedded calvariae in paraffin and
cut coronally. We detected BrdU-incorporated osteoblasts with the BrdU
Immunohistochemistry Kit (Exalpha Biologicals). We calculated the number
of BrdU-positive osteoblasts over the total number of osteoblasts (osteoblast
mitotic index) at three different locations (+3.0, 3.5 and 4.0 AP (0 point:
bregma)) per mouse. We analyzed six mice per group.
Intracerebroventricular infusion. Intracerebroventricular infusion was per-
formed as previously described6. Briefly, we exposed the calvaria of an
anesthetized mouse, implanted a 28-gauge cannula (Plastics ONE) into the
third ventricle and then connected the cannula to an osmotic pump (Durect)
placed in the dorsal subcutaneous space of the mouse. We infused rat
Neuromedin U-23 (Peptide Institute) or human leptin (Sigma) at
0.125 nmol/hr or 8 ng/hr, respectively, for 28 d.
Quantitative RT–PCR analysis. After flushing mouse bone marrow out of the
bone with PBS, we extracted bone RNAwith Trizol (Invitrogen) and performed
reverse transcription for cDNA synthesis. We performed quantitative analysis of
gene expression with the Mx3000P real-time PCR system (Stratagene). Primer
sequences are available upon request. We used GAPDH expression as an
internal control.
Statistical analysis. All data are represented as mean ± s.d. (n¼8 or more). We
performed statistical analysis by Student’s t-test. Values were considered
statistically significant at Po0.05. Results are representative of more than
four individual experiments.
Note: Supplementary information is available on the Nature Medicine website.
ACKNOWLEDGMENTS
We thank G. Karsenty, M. Patel and P. Ducy for critical review of the manuscript
and for helpful discussions; K. Nakao, M. Noda, T. Matsumoto and S. Ito for
insightful suggestions; P. Barrett (Rowett Research Institute, UK) for providing
a plasmid for the Cartpt probe; and J. Chen, M. Starbuck, S. Sunamura,
H. Murayama, H. Yamato, and M. Kajiwara for technical assistance. This work
was supported by grant-in-aid for scientific research from the Japan Society for
© 2007 Nature Publishing Group http://www.nature.com/naturemedicine
LETTERS
6ADVANCE ONLINE PUBLICATION NATURE MEDICINE

the Promotion of Science, a grant for the 21st Century Center of Excellence
program from the Ministry of Education, Culture, Sports, Science, and
Technology of Japan, Ono Medical Research Foundation, Yamanouchi Foundation
for Research on Metabolic Disorders, Kanae Foundation for the Promotion of the
Medical Science and the Program for Promotion of Fundamental Studies in
Health Sciences of the National Institute of Biomedical Innovation of Japan.
AUTHOR CONTRIBUTIONS
S. Sato conducted most of the experiments. K. Kangawa and M. Kojima
generated Nmu
–/–
mice. R. Hanada and T. Ida conducted in vitro experiments.
S. Fukumoto, Y. Takeuchi and T. Fujita contributed by conducting dual X-ray
absorptiometry analyses and providing suggestions on the project. M. Iwasaki
prepared the constructs. A. Kimura performed i.c.v. infusion experiments.
H. Inose conducted mCT analyses. T. Matsumoto and S. Kato conducted
histological analyses for brain tissue. T. Abe and M. Mieda performed in situ
hybridization analysis. S. Takeda and K. Shinomiya designed the project.
S. Takeda supervised the project and wrote most of the manuscript.
Published online at http://www.nature.com/naturemedicine
Reprints and permissions information is available online at http://npg.nature.com/
reprintsandpermissions
1. Rodan, G.A. & Martin, T.J. Therapeutic approaches to bone diseases. Science 289,
1508–1514 (2000).
2. Teitelbaum, S.L. & Ross, F.P. Genetic regulation of osteoclast development and
function. Nat. Rev. Genet. 4, 638–649 (2003).
3. Saper, C.B., Chou, T.C. & Elmquist, J.K. The need to feed: homeostatic and hedonic
control of eating. Neuron 36, 199–211 (2002).
4. Ahima, R.S. & Flier, J.S. Leptin. Annu. Rev. Physiol. 62, 413–437 (2000).
5. Spiegelman, B.M. & Flier, J.S. Obesity and the regulation of energy balance. Cell 104,
531–543 (2001).
6. Ducy, P. et al. Leptin inhibits bone formation through a hypothalamic relay: a central
control of bone mass. Cell 100, 197–207 (2000).
7. Takeda, S. et al. Leptin regulates bone formation via the sympathetic nervous system.
Cell 111, 305–317 (2002).
8. Hanada, R. et al. Neuromedin U has a novel anorexigenic effect independent of the
leptin signaling pathway. Nat. Med. 10, 1067–1073 (2004).
9. Brighton, P.J., Szekeres, P.G. & Willars, G.B. Neuromedin U and its receptors:
structure, function, and physiological roles. Pharmacol. Rev. 56, 231–248 (2004).
10. Fang, L., Zhang, M., Li, C., Dong, S. & Hu, Y. Chemical genetic analysis reveals
the effects of NMU2R on the expression of peptide hormones. Neurosci. Lett. 404,
148–153 (2006).
11. Riggs, B.L., Khosla, S. & Melton, L.J. 3rd. A unitary model for involutional osteo-
porosis: estrogen deficiency causes both type I and type II osteoporosis in postmeno-
pausal women and contributes to bone loss in aging men. J. Bone Miner. Res. 13,
763–773 (1998).
12. Karsenty, G. & Wagner, E.F. Reaching a genetic and molecular understanding of
skeletal development. Dev. Cell 2, 389 –406 (2002).
13. Harada, S. & Rodan, G.A. Control of osteoblast function and regulation of bone mass.
Nature 423, 349–355 (2003).
14. Idris, A.I. et al. Regulation of bone mass, bone loss and osteoclast activity by
cannabinoid receptors. Nat. Med. 11, 774–779 (2005 ).
15. Abe, E. et al. TSH is a negative regulator of skeletal remodeling. Cell 115, 151–162
(2003).
16. Elefteriou, F. et al. Leptin regulation of bone resorption by the sympathetic nervous
system and CART. Nature 434, 514–520 (20 05).
17. Fu, L., Patel, M.S., Bradley, A., Wagner, E.F. & Karsenty, G. The molecular clock
mediates leptin-regulated bone formation. Cell 122, 803–815 (2005).
18. Howard, A.D. et al. Identification of receptors for neuromedin U and its role in feeding.
Nature 406, 70–74 (2000).
19. Wren, A.M. et al. Hypothalamic actions of neuromedin U. Endocrinology 143,
4227–4234 (2002).
20. Parfitt, A.M. The physiological and clinical significance of bone histomorphometric
data. in Bone Histomorphometry (ed. Recker, R.R.) 143–223 (CRC Press, Boca Raton,
FL, 1983).
21. Chu, C. et al. Cardiovascular actions of central neuromedin U in conscious rats. Regul.
Pept. 105, 29–34 (2002).
22. Ahn, J.D., Dubern, B., Lubrano-Berthelier, C., Clement, K. & Karsenty, G. Cart
overexpression is the only identifiable cause of high bone mass in melanocortin 4
receptor deficiency. Endocrinology 147, 3196–3202 (2006).
23. Kalinova, J., Triska, J. & Vrchotova, N. Distribution of vitamin E, squalene, epicatechin,
and rutin in common buckwheat plants (Fagopyrum esculentum Moench).
J. Agric. Food Chem. 54, 5330–5335 (2006).
24. Zeng, H. et al. Neuromedin U receptor 2–deficient mice display differential responses
in sensory perception, stress, and feeding. Mol. Cell. Biol. 26, 9352–9363 (2006).
25. Elias, C.F. et al. Leptin differentially regulates NPY and POMC neurons projecting to
the lateral hypothalamic area. Neuron 23, 775–786 (1999).
26. Graham, E.S. et al. Neuromedin U and Neuromedin U receptor-2 expression in
the mouse and rat hypothalamus: effects of nutritional status. J. Neurochem. 87,
1165–1173 (2003).
27. Paxinos, G. & Franklin, K. The Mouse Brain in Stereotaxic Coordinates 2nd edn.
(Academic Press, San Diego, 2001).
© 2007 Nature Publishing Group http://www.nature.com/naturemedicine
LETTERS
NATURE MEDICINE ADVANCE ONLINE PUBLICATION 7