Bidirectional ephrinB2-EphB4 signaling
controls bone homeostasis
Chen Zhao,1,5Naoko Irie,1,5Yasunari Takada,1Kouji Shimoda,2Takeshi Miyamoto,3Toru Nishiwaki,4Toshio Suda,3
and Koichi Matsuo1,*
1Department of Microbiology and Immunology
2Laboratory Animal Center
3The Sakaguchi Laboratory of Developmental Biology
4Department of Orthopedic Surgery
School of Medicine, Keio University, Shinjuku-ku, Tokyo, 160-8582, Japan
5These authors contributed equally to this work.
Bone homeostasis requires a delicate balance between the activities of bone-resorbing osteoclasts and bone-forming
osteoblasts. Various molecules coordinate osteoclast function with that of osteoblasts; however, molecules that mediate
osteoclast-osteoblast interactions by simultaneous signal transduction in both cell types have not yet been identified.
Here we show that osteoclasts express the NFATc1 target gene Efnb2 (encoding ephrinB2), while osteoblasts express
the receptor EphB4, along with other ephrin-Eph family members. Using gain- and loss-of-function experiments, we dem-
onstrate that reverse signaling through ephrinB2 into osteoclast precursors suppresses osteoclast differentiation by inhib-
iting the osteoclastogenic c-Fos-NFATc1 cascade. In addition, forward signaling through EphB4 into osteoblasts enhances
osteogenic differentiation, and overexpression of EphB4 in osteoblasts increases bone mass in transgenic mice. These data
demonstrate that ephrin-Eph bidirectional signaling links two major molecular mechanisms for cell differentiation—one in
osteoclasts and the other in osteoblasts—thereby maintaining bone homeostasis.
Bone-resorbing osteoclasts are large, multinucleated cells de-
ing osteoblasts are derived from mesenchymal cells (Karsenty
and Wagner, 2002; Teitelbaum and Ross, 2003). Osteoblasts
or bone stromal cells produce macrophage-colony stimulating
factor (M-CSF, encoded by Csf1) and RANK ligand (RANKL).
When these cytokines stimulate their receptors on osteoclast
precursors, downstream signaling pathways activate the tran-
scription factors NF-kB (Franzoso et al., 1997), c-Fos (Grigoria-
dis et al., 1994), and NFATc1 (nuclear factor of activated T cells
c1) (Asagiri et al., 2005; Ishida et al., 2002; Takayanagi et al.,
2002a), all of which are required for osteoclast differentiation.
Mice lacking c-Fos (Fos2/2mice) develop osteopetrosis due to
an osteoclast differentiation block. We reported that Nfatc1 is
a major c-Fos target gene and that a constitutively active form
of NFAT restores the differentiation block in Fos2/2osteoclast
precursors (Matsuo et al., 2004). In addition, c-Fos activates
transcription of Ifnb (encoding IFN-b), which mediates a nega-
tive feedback loop that suppresses osteoclast differentiation
(Takayanagi et al., 2002b).
Bone remodeling is a complex process requiring ‘‘coupling’’
between osteoclastic and osteoblastic activities. Coordinated
function of osteoclasts and osteoblasts ensures that resorption
lacunae are filled with new bone produced by osteoblasts so as
to maintain bone integrity. Mechanical forces and certain sys-
temic factors, such as steroids or parathyroid hormone, affect
bone remodeling. For the anabolic effect of parathyroid hor-
mone to occur, bone resorption may be necessary, implying
that osteoblast formation requires signals from osteoclasts
(Martin, 2004). Local factors including TRAP secreted by osteo-
clasts (Sheu et al., 2003) and insulin-like growth factor and
Mundy, 1987), can stimulate bone formation. An imbalance be-
tweenbone resorption andformationresultsinbone remodeling
diseases such as osteoporosis and osteopetrosis. Even audi-
tory ossicles, the smallest bones in the body, can be adversely
affected by defects in bone remodeling (Kanzaki et al., 2006).
Understanding the mechanisms underlying bone remodeling
should potentially facilitate prevention and treatment of bone
Cell-surface molecules ephrinB (B1wB3) are preferential
ligands for the tyrosine kinase receptors EphB (B1wB6), while
ephrinA proteins (A1wA5) function as ligands for EphA
(A1wA10) receptors (Himanen and Nikolov, 2003; Kullander
and Klein, 2002; Murai and Pasquale, 2003; Palmer and Klein,
2003; Pasquale, 2005). The ephrinB family consists of trans-
membrane proteins with cytoplasmic domains, whereas the
ephrinA family consists of glycosyl-phosphatidyl inositol (GPI)-
anchored molecules. Interaction between ephrinB- and EphB-
expressing cells results in bidirectional signal transduction.
Activation of the EphB receptors by the ephrinB ligands is re-
ferred to as ‘‘forward signalling,’’ whereas activation of the eph-
rinB ligands by the EphB receptors is designated ‘‘reverse sig-
nalling.’’ Reverse signaling through ephrinB ligands activates
both tyrosine phosphorylation-dependent and -independent
signal transduction pathways. The intracellular domain of eph-
rinB ligands, particularly the last 33 C-terminal amino acids, is
CELL METABOLISM 4, 111–121, AUGUST 2006 ª2006 ELSEVIER INC. DOI 10.1016/j.cmet.2006.05.012 111
A R T I C L E
highly conserved and contains multiple tyrosine residues, and
the C-terminal YKV motif is a binding site for PDZ (postsynaptic
densityprotein, diskslarge, zonaoccludens) domain-containing
proteins. Similarly, forward signaling through EphB receptors
activates both tyrosine phosphorylation-dependent and -inde-
pendent pathways. The importance of bidirectional signaling
has been confirmed in angiogenesis, axon guidance, cell sort-
ing, and boundary formation (Davy and Soriano, 2005). For ex-
ample, ephrinB2 and its cognate receptor EphB4 are recipro-
cally expressed in arterial and venous endothelial cells. Mice
deficient in ephrinB2 or EphB4 exhibit similar phenotypes char-
acterized by early embryonic lethality due to disorganized arte-
riovenous formation (Adams et al., 1999; Gerety et al., 1999;
Wang et al., 1998). However, recent analysis using ephrinB2
mutants lacking the C-terminal cytoplasmic domain revealed
that ephrinB2 reverse signaling through the cytoplasmic domain
is not required for early vascular development but is required for
axon pathfinding and cardiac valve formation (Cowan et al.,
2004). Furthermore, ephrin-Eph interaction regulates stem cell
differentiation (Wang et al., 2004), the immune response (Sharfe
et al., 2002; Yu et al., 2003), intestinal epithelial cell migration
(Batlle et al., 2002), and skeletal patterning (Compagni et al.,
2003; Davy et al., 2004).
Here weidentify Efnb2(encoding ephrinB2)asan NFATtarget
gene during osteoclast differentiation. In addition we show that
EphB4 is expressed on osteoblasts. Using a combination of
in vitro and in vivo approaches we demonstrate that ephrinB2-
EphB4 bidirectional signaling links suppression of osteoclast
differentiation to stimulation of bone formation, which may reg-
ulate the transition from a bone-resorption to a bone-formation
phase in each bone remodeling cycle.
Expression of ephrinB and EphB families in bone cells
Overexpression of an active form of NFAT in Fos2/2osteoclast
precursors induces expression of Efnb2 in addition to upregula-
tion of osteoclast markers such as Acp5 (encoding tartrate-re-
sistant acid phosphatase, TRAP) and Calcr (encoding the calci-
tonin receptor) (Matsuo et al., 2004). Therefore, we examined
Efnb2expression byquantitative reversetranscription-polymer-
ase chain reaction (RT-PCR) analysis using M-CSF-dependent
as osteolast precursors. Expression of the macrophage marker
in the presence of RANKL for 3 days along with Acp5 and Calcr
in wild-type but not in Fos2/2cultures by RANKL (Figure 1A).
Next we identified an NFAT binding site (22863 relative to the
ATG initiation codon) in the mouse Efnb2 promoter by electro-
phoretic mobility shift assays (EMSA) and showed that an
Efnb2 promoter-reporter construct is activated by RANKL or
NFAT (Figure S1). At the protein level, ephrinB2 was induced
during osteoclast differentiation (Figure 1B). Using immunofluo-
rescence microscopy, we detected ephrinB2 protein in multinu-
cleated and differentiating mononuclear osteoclasts formed in
the presence of RANKL and intact Fos (Figure 1C, Wt). When
cells were immunostained before fixation and permeabilization,
ephrinB2 was detected on the membrane of wild-type osteo-
clasts(Figure 1C,membrane). Finally, wedetectedephrinB2ex-
pression in TRAP-positive multinucleated osteoclasts in bone
sections (Figure 1D). These data demonstrate that Efnb2 ex-
pression is induced during osteoclast differentiation through
the c-Fos-NFATc1 transcriptional cascade in wild-type cells.
Because ephrinB2 is known to interact with the receptors
EphB1, EphB2, EphB3, EphB4, EphB6, and EphA4, we exam-
ined expression of Efnb and Ephb family members and Epha4
in osteoclasts and osteoblasts differentiated in vitro. In differen-
but not Ephb encoding receptors were detected (Figure 1E). By
contrast, calvarial osteoblasts constitutively expressed several
Efnb ligands, Ephb receptors, and Epha4 (Figure 1E).
Reverse signaling suppresses osteoclast formation
The indiscriminate ephrinB2 interacts with multiple Ephs, while
EphB4 only interacts with ephrinB2 (Gale and Yancopoulos,
1999; Myshkin and Wang, 2003). Therefore, to analyze the ef-
fects of reverse signaling in osteoclast differentiation we used
EphB4 to stimulate ephrinB2, and to analyze the effects of for-
ward signaling in osteoblast differentiation we used ephrinB2
to stimulate Eph receptors. Addition of clustered EphB4 to os-
teoclastogenic cultures suppressed osteoclast differentiation
and hence bone resorption in a dose-dependent manner (Fig-
ure 2A). To investigate whether membrane bound EphB4 also
suppresses osteoclast differentiation through ephrinB2, we
established subclones of ST2 stromal cells stably expressing
EphB4. Quantitative RT-PCR analysis showed that forced ex-
pression of EphB4 did not reduce RANKL or M-CSF or increase
osteoprotegerin (OPG) expression levels (Figure S2). We then
cocultured bonemarrow cells with EphB4-expressing ST2 cells.
Consistent with the soluble EphB4 experiment, membrane
In addition to soluble EphB4, we treated osteoclast precursors
with soluble EphA4 and EphB2, which are known to interact
with ephrinB2. As a negative control, we used ephrinB2 itself
since ephrinB2 does not interact homophilically. The soluble re-
ceptors EphB4, EphA4, and EphB2 but not the soluble ligand
expression (Figure 2C). These data collectively indicate that
Overexpression of ephrinB2 inhibits osteoclastogenesis
We further examined the role of ephrinB2 in osteoclast differen-
tiation by gain-of-function experiments. Infection of bone mar-
row cells with a retroviral vector expressing wild-type ephrinB2
suppressed formation of large multinucleated osteoclasts when
cocultured with ST2 cells, indicating that overexpressed eph-
rinB2 transduced inhibitory signals into osteoclast precursors
(Figure 3A). To determine whether the negative signal was in-
deed transduced through the ephrinB2 cytoplasmic domain,
we generated ephrinB2 mutants lacking the entire cytoplasmic
domain (DC) or only the highly conserved C-terminal 33 amino
acids (D33). In addition, we generated a deletion mutant lacking
only the C-terminal YKV (DYKV), which interacts with PDZ do-
mains, andamutantwith sixtyrosine-to-phenylalanine substitu-
tions in the cytoplasmic domain (Y6F) (Figure 3B). The DC and
D33 constructs were synthesized and properly expressed on
the cellsurface, basedonobservations thattheextracellulardo-
mains of both were recognized by an anti-ephrinB2 antibody
(Figure3CandFigure S3A) andthatDCandD33overexpression
stimulated osteoblast differentiation in vitro (see Figure 5B).
As expected, neither DC nor D33 suppressed formation of
A R T I C L E
CELL METABOLISM : AUGUST 2006
multinucleated giant cells when infected bone marrow cells
were cocultured with ST2 cells, demonstrating that the intracel-
lular domain of ephrinB2 is essential for reverse signaling. Im-
portantly, the DYKV but not the Y6F mutant lacked inhibitory
activity, strongly suggesting that interaction with PDZ domain
proteins, but not tyrosine phosphorylation, mediates the inhibi-
tory signals (Figure 3B). Similar results were obtained in stromal
cell-free cultures (data not shown). These data indicate that the
cytoplasmic domain of ephrinB2 transduces the inhibitory effect
and that the C-terminal YKV is critical for signal transduction.
Loss of reverse signaling enhances osteoclastogenesis
We next examined the role of ephrinB2 during osteoclastogen-
esis by loss-of-function experiments. First, we knocked down
endogenous Efnb2 transcripts by small interfering RNA (siRNA)
using the retroviral vector RVH1 (Barton and Medzhitov, 2002).
Retroviral gene transfer and immunoblot analysis confirmed
thatthe knockdownvectorefficientlyinhibited Efnb2expression
in M-CSF-dependent bone marrow-derived macrophages
cent protein (GFP)-positive bone marrow cells, which had been
Figure 1. Expression of ephrins and Ephs in osteo-
clasts and osteoblasts
A) Quantitative RT-PCR analysis of gene expression
in wild-type (Wt) and Fos2/2osteoclast precursors
(MDMs; see Experimental Procedures) in the pres-
ence of RANKL. Spl, spleen; BM, bone marrow.
Bars represent means 6 SEM.
B) Immunoblot analysis of ephrinB2 during osteo-
clast differentiation of wild-type MDMs. Actin serves
as a loading control. C, control expressing exoge-
C) Immunofluorescence of ephrinB2 in wild-type
MDMs treated with RANKL for 3 days. Negative con-
trols are wild-type MDMs in the absence of RANKL,
and Fos2/2MDMs in the presence of RANKL. ‘‘Wt
(membrane)’’ indicates that immunostaining was
performed before cell fixation and permeabilization.
Nuclei were stained with DAPI (blue) as shown in
D) Expression of ephrinB2 in bone. Upper panel,
TRAP-activity staining (TRAP). Lower panel, immu-
nofluorescence of ephrinB2 in tibia.
E) RT-PCR analysis of ephrins and Ephs in MDMs
and calvarial osteoblasts during differentiation. C,
control adult mouse brain.
ephrin-Eph signaling in bone homeostasis
CELL METABOLISM : AUGUST 2006113
infected with the siRNA virus, were sorted and cocultured with
ST2 cells. After 5 days, ephrinB2 knockdown cultures displayed
lower). Next, we isolated bone marrow cells from conditional
ephrinB2 knockout mice (Gerety and Anderson, 2002) carrying
the LysMcre allele (Clausen et al., 1999), which deletes the
floxed Efnb2 loci (Efnb2f/f) in a myeloid lineage. Deletion of
Efnb2 was confirmed by genomic PCR analysis in M-CSF-
dependent bone marrow-derived macrophages (Figure 3E and
Figure S3B). Furthermore, Efnb2 induction was abolished in
Efnb2f/fcells in the presence of the LysMcre allele (Figure 3F),
and these cells showed more efficient osteoclast formation
compared to control cells without the deletion (Figure 3G).
Therefore, both gain- and loss-of-function experiments demon-
To evaluate in vivo effect of Efnb2 deletion, we also analyzed
femurs and tibias of mice lacking ephrinB2 in the macrophage-
and the bone mineral density (BMD) of conditional ephrinB2
knockout mice (Efnb2f/f, Cre+) were comparable to that of con-
rinB1 or other compensatory factors can maintain bone homeo-
Reverse signaling inhibits Fos and Nfatc1 expression
We next determined whether inhibitory signaling downstream of
ephrinB2 affects expression of two pivotal osteoclastogenic
transcription factors, c-Fos and NFATc1. Immunoblot analysis
showed that treatment with EphB4 suppressed induction of
c-Fos on day 2 and NFATc1 on day 4 (Figure 4A). Levels of
Fos and Nfatc1 transcripts, as well as the numbers of TRAP-
positive multinucleated cells, were reduced in bone marrow
cells in the presence of EphB4 under osteoclastogenic condi-
tions (Figure 4B). Furthermore, expression of Fcgr3, which was
high in Fos2/2osteoclast precursors (Figure 1A), was elevated
in the presence of EphB4 (Figure 4B). These data suggest that
ephrinB2 signaling inhibits osteoclastogenesis by blocking in-
duction of Fos and its transcriptional target Nfatc1. To confirm
these findings, we retrovirally introduced c-Fos and NFATc1
into bone marrow cells and cultured them with EphB4 in the
presence of M-CSF and RANKL. EphB4 suppressed formation
of large multinucleated cells from bone marrow cells infected
with empty virus. In contrast, bone marrow cells infected with
c-Fos virus were insensitive to suppression by EphB4 and
formed well-differentiated large osteoclasts (Figure 4C). In-
fection with NFATc1 virus also conferred resistance to EphB4
treatment (Figure 4D). Therefore, inhibition of Fos and Nfatc1
transcription is one mechanism by which reverse signaling
through ephrinB2 suppresses osteoclast differentiation.
Forward signaling enhances osteoblast formation
Apart from the reverse signaling into osteoclasts through eph-
rinB2,ephrinB2-EphB4 interactionbetweenosteoclasts andos-
teoblasts must simultaneously transmit signals in the opposite
oblasts. The addition of clustered ephrinB2 to osteoblastogenic
cultures significantly stimulated differentiation of calvarial oste-
oblasts, as judged from alkaline phosphatase (ALP) staining,
and clustered soluble EphB4 neutralized this stimulatory effect
(Figure 5A). Overexpression of ephrinB2 in osteoblasts by retro-
viral gene transfer also stimulated osteoblast differentiation
(Figure 5B). This stimulatory effect was independent of the eph-
rinB2 cytoplasmic domain, given that wild-type, DC, and D33
mutants all enhanced ALP staining (Figure 5B), and supports
the idea that it is the ephrinB2 extracellular domain that stimu-
lates Eph receptors. Quantitative RT-PCR analysis showed
that when stimulated with clustered ephrinB2, expression of
Figure 2. Reverse signaling through ephrinB2 inhibits osteoclast differentiation
A) Treatment of MDMs with clustered EphB4 or control Fc at indicated concentra-
tions in the presence of RANKL (10 ng/ml). Upper panels, TRAP staining after 5
days. Values represent relative TRAP activities in cell lysates. a.u., arbitrary units.
Lower panels, resorption pits on bovine bone slices after culturing in the presence
of RANKL for 7 days; values represent bone surface resorbed (%).
B) Overexpression of EphB4 in ST2 cells suppresses osteoclast differentiation.
Receptor body staining using ephrinB2-Fc, and fluorescein-isothiocyanate
(FITC)-conjugated anti-Fc antibody confirmed higher levels of EphB4 expression
in transfected ST2 cells (FITC). After coculturing with wild-type bone marrow cells,
TRAP-positive multinucleated cells (MNCs) were counted. Values represent
means 6 SD.
C) Quantitative RT-PCR analysisof Calcr expressionduring RANKL-induced oste-
oclast differentiation of wild-type MDMs on coverglasses coated with EphA4-Fc,
EphB2-Fc, or EphB4-Fc (day 6). Fc and ephrinB2-Fc were included as negative
controls. Bars represent means 6 SEM.
A R T I C L E
CELL METABOLISM : AUGUST 2006
osteoblast differentiation markers, such as Alp, Col1 (Col1a1,
collagen Ia), and Og2 (osteocalcin), was increased (Figure 5C).
Consistently, mRNAs encoding transcription factors critical for
osteoblast differentiation, such as Dlx5 (distal-less homeobox
5), Osx (osterix, also known as Sp7) (Nakashima et al., 2002),
and Runx2 (runt-related transcription factor 2, also known as
Cbfa1) (Ducy et al., 1997; Komori et al., 1997), were significantly
induced by ephrinB2 (Figure 5C). We knocked down Ephb4 ex-
pression by introducing siRNA vectors into calvarial osteoblasts
Figure 4. Reverse signaling inhibits expression of c-Fos and NFATc1
A) Immunoblot analysis of c-Fos and NFATc1 during osteoclast formation in the
presence of clustered Fc or EphB4-Fc (2 mg/ml).
B) Quantitative RT-PCR analysis of the expression of Fos (day 2), Nfatc1 (day 4),
and Fcgr3 (day 4), and number of TRAP-positive MNCs (day 4) in the presence
of clustered Fc or EphB4-Fc in stromal cell-free cultures. Bars represent means
C) EphB4 treatment of MDMs overexpressing c-Fos. MDMs were infected with
pMX-IRES-GFP (empty) or pMX-Fos-IRES-GFP (Fos) and cultured without or
with EphB4-Fc (2 mg/ml).
D) Number of TRAP-positive MNCs in 4 day osteoclastogenic cultures in the pres-
ence of Fc or EphB4-Fc. MDMs were infected with pMX-IRES-GFP (empty), pMX-
Fos-IRES-GFP (Fos), or pMSCV-NFATc1-IRES-GFP (NFATc1) prior to osteoclas-
togenesis. Bars represent means 6 SD.
Figure 3. Gain- and loss-of-function of ephrinB2 in osteoclasts
A) MDMs were infected with the expression viruses pMX-IRES-GFP (empty) or
pMX-ephrinB2-IRES-GFP (Wt ephrinB2) and cocultured with ST2 cells for 6
days. Arrowheads, GFP-positive multinucleated cells (MNC).
B) Upper, number of large (>65 mm), TRAP-positive MNCs (means 6 SEM) deter-
mined after overexpression of Wt ephrinB2 and mutants. Lower, schematic repre-
sentation of Wt and mutant constructs.
C) Upper panels, immunofluorescence of extracellular domains of retrovirally
expressed ephrinB2 mutants (DC and D33) in MDMs. Immunostaining was per-
formed before cell fixation and permeabilization. Lower panels, DAPI staining of
the same culture. Experiment was performed as in Figure 1C.
D) Knockdown of ephrinB2. Upper, immunoblot analysis of ephrinB2 after retrovi-
ralgene-transfer ofRVH1-GFP(empty)andRVH1-siRNA-GFP (siRNA)vectorsinto
wild-type MDMs, followed by RANKL treatment for 3 days. Lower, TRAP staining
of cocultures containing sorted GFP-positive MDMs infected with empty or siRNA
viruses (TRAP). Numbers represent TRAP-positive giant MNCs in triplicate 24-well
samples (means 6 SEM).
E) Genomic PCR analysis of Efnb2 deletion in MDMs isolated from Efnb2f/fmice
without (2) or with (+) the LysMCre allele. The 636 bp (flox) and 309 bp (D) bands
represent floxed and deleted Efnb2 loci, respectively.
F) Quantitative RT-PCR analysis of Efnb2 during osteoclast differentiation of
MDMs described in (E). Bars represent means 6 SD.
G) Number of TRAP-positive MNCs (means 6 SEM) in 5 day cocultures containing
MDMs described in (E) and ST2 cells.
ephrin-Eph signaling in bone homeostasis
CELL METABOLISM : AUGUST 2006 115
and differentiated them into mature osteoblasts in the presence
of ephrinB2. The siRNA vector reduced Ephb4 transcripts and
lowered ALP activity, whereas irrelevant Epha4 and Ephb6 tran-
scripts were not affected, indicating that signaling from EphB4
specifically stimulates osteoblast differentiation (Figure 5D).
We examined potential downstream effectors of forward sig-
naling in osteoblasts stimulated with ephrinB2. During in vitro
differentiation of osteoblasts, phosphorylation of ERK1/2 was
enhanced (Figure S4), which may lead to stimulation of osteo-
blast differentiation. We also analyzed the active guanosine
triphosphate (GTP) bound form of the small GTPase RhoA (Eti-
enne-Manneville and Hall, 2002) since inhibition of Rho-Rho
kinase signaling stimulates differentiation of mouse calvarial
osteoblasts (Harmey et al., 2004). GTP-RhoA was reduced by
ephrinB2 treatment (Figure 5E). To ask whether RhoA activity
negatively regulates osteoblast differentiation, we introduced
by retroviral gene transfer a constitutively active (V14) or a dom-
inant-negative (N19) form of RhoA (Gebbink et al., 1997) into os-
teoblasts in culture. After 6 days, RhoA (V14) suppressed ALP
activity and RhoA (N19) enhanced it (Figure 5F), indicating that
RhoA inactivation is a mechanism by which forward signaling
through EphB4 enhances osteoblast formation.
Enhanced bone formation in EphB4 transgenic mice
To examine the effects of forward signaling through EphB4 on
bone homeostasis in vivo, we generated EphB4 transgenic
mice, in which EphB4 is expressed in osteoblasts under the
control of the mouse Col1 promoter (Liu et al., 2001). Overex-
pression of EphB4 in transgenic osteoblasts was confirmed by
immunohistochemical staining of sections of femur (Figure 6A)
and tibia (Figure 6B). To compare expression levels of Ephb4
between wild-type and EphB4 transgenic mice, we performed
quantitative RT-PCR using RNA prepared from calvaria, femur,
and tibia. On average, EphB4 transgenic mice expressed 6.7-
fold higher levels of Ephb4 transcripts in calvaria and 4.8-fold
higher in the long bones (Figure 6C). We next compared
Ephb4 transcripts using cultured calvarial osteoblasts and
bone marrow-derived osteoclasts. EphB4 transgenic osteo-
blasts expressed a 5.2-fold higher level of Ephb4 than wild-
type controls, whereas osteoclasts did not express detectable
Ephb4 from the transgene (Figure 6D). These calvarial osteo-
blasts isolated from EphB4 transgenic mice differentiated
more efficiently than those from wild-type controls, as judged
from both ALP and Alizarin Red staining (Figure 6E).
that EphB4 transgenic mice exhibited increased bone mass by
10 weeks of age (Figure 7A). Dual energy X-ray absorptiometry
(DEXA) showed that the BMD of femurs from transgenic mice
was consistently greater than that of wild-type controls (Fig-
ure 7B). Quantitative bone histomorphometry further showed
that bone volume, osteoid thickness, mineralizing surface, and
bone formation rate were significantly increased in transgenic
mice versus wild-type controls (Figure 7C). Serum levels of
osteocalcin, a marker of bone formation, were also slightly ele-
vated in EphB4 transgenic mice (Figure 7D). Strikingly, osteo-
clast number, osteoclast surface, and eroded surface were
Figure 5. Forward signaling through EphB4 en-
hances osteoblast differentiation
A) ALP staining of osteoblasts in osteoblastogenic
cultures containing indicated concentrations of clus-
tered ephrinB2 and EphB4 for 6 days. Two indepen-
dent preparations of osteoblasts showed similar re-
sults. Staining intensity was quantified with Image
J. *p < 0.05, **p < 0.01, versus no treatment. Bars
represent means 6 SEM.
B) ALP staining of osteoblasts 6 days after infection
with wild-type (Wt) and mutant ephrinB2 viruses
(DC and D33). Bars represent means 6 SEM.
C) Quantitative RT-PCR analysis of osteoblasts cul-
tured under osteoblastogenic conditions in the pres-
ence of ephrinB2-Fc. RNAs were prepared on day 6
(for Alp, Dlx5, Osx, and Runx2) and on day 9 (for Col1
and Og2). Bars represent means 6 SD.
D) EphB4 knockdown experiment. Quantitative RT-
PCR analysis of Ephb4 in osteoblasts differentiated
in the presence of ephrinB2 for 9 days after infecting
were evaluated to examine the specificity of Ephb4
siRNA. ALP activity in lysates was also quantified.
yp = 0.063, **p < 0.01 versus GFP virus. Bars repre-
sent means 6 SEM.
E) Immunoblot analysis of RhoA activity in osteo-
blasts stimulated with Fc or ephrinB2-Fc for the indi-
F) Effect of RhoA activity on osteoblast differentia-
tion. V14, constitutively active RhoA; N19, domi-
nant-negative RhoA. Upper, ALP staining (day 6) of
calvarial osteoblasts. Lower, ALP activity in lysates.
Bars represent means 6 SEM.
A R T I C L E
CELL METABOLISM : AUGUST 2006
significantly reduced in EphB4 transgenic mice compared to
wild-type controls, indicating that osteoclast function was in-
hibited in these mice (Figure 7C). Consistently, EphB4 trans-
genic mice showed a decrease in urinary deoxypyridinoline
(DPD) crosslinks, a marker for osteoclastic bone resorption, in-
dicating that osteoclast function was suppressed (Figure 7E).
Such reduced osteoclastic activity is not due to reduction in
RANKL or M-CSF or to an increase in OPG since there was no
significant difference in Rankl, Csf1, and Opg mRNA levels in
bone or RANKL and OPG levels in serum between wild-type
and EphB4 transgenic mice (data not shown). Since EphB4
binds only to ephrinB2 (Gale and Yancopoulos, 1999; Myshkin
and Wang, 2003), we conclude that the results of the bone his-
tomorphometric analysis are a consequence of direct interac-
tion between EphB4 and ephrinB2. Collectively, elevated levels
of EphB4 in osteoblasts not only enhance bone formation but
also inhibit bone resorption in vivo.
Figure 6. Expression of EphB4 in transgenic mice and in vitro osteoblast differen-
A) Immunohistochemical staining of EphB4 (red) in distal femurs from 8-week-old,
wild-type (Wt) and transgenic (Tg) mice. Counterstaining, methyl-green. Orienta-
tion is as in Figure 7A.
B) Immunofluorescence detection of EphB4 (red) in tibia from 8-week-old trans-
genic mice. Counterstaining, DAPI (blue). BM, bone marrow.
C) Quantitative RT-PCR analysis of Ephb4 expression. Total RNA was isolated di-
Tg mice (n = 3 for each genotype) were separately measured. Bars represent
means 6 SEM.
D) Quantitative RT-PCR analysis using calvarial osteoblasts and bone marrow-de-
rived osteoclasts prepared from Wt and EphB4 Tg mice. Bars represent means 6
E) In vitro osteoblastogenesis of EphB4-expressing osteoblasts. Duplicate cul-
tures were terminated on the days indicated and stained for ALP activity or with
Figure 7. EphB4 transgenic mice show increased bone mass
A)Radiographicanalysis(leftpanels,radiograph; rightpanels, mCT)offemursfrom
B) Quantification of BMD of femurs isolated from 10-week-old female mice by
DEXA (Wt n = 6, Tg n = 10, data shown as means 6 SEM, *p < 0.05, **p < 0.01).
C)Histomorphometrical analysis oftibia from 10-week-old mice(eachgroupn=8,
volume; Ob.S/BS, osteoblast surface per bone surface; O.Th, osteoid thickness;
MS/BS, mineralizing surface per bone surface; BFR/BS, bone formation rate per
bone surface; N.Oc/B.Pm, osteoclast number per bone perimeter; Oc.S/BS, oste-
oclast surface per bone surface; ES/BS, eroded surface per bone surface.
E) Urinary deoxypyridinoline (DPD) measurement in WtandTg mice. (Wtn = 22, Tg
n = 16, *p < 0.05). Bars represent means 6 SEM.
ephrin-Eph signaling in bone homeostasis
CELL METABOLISM : AUGUST 2006 117
The interaction between ephrins and Ephs has been extensively
investigated in neural development, angiogenesis, skeletal pat-
terning, and other contexts (Compagni et al., 2003; Davy et al.,
2004; Himanen and Nikolov, 2003; Kullander and Klein, 2002;
Murai and Pasquale, 2003; Palmer and Klein, 2003). However,
roles for ephrin-Eph bidirectional signaling in bone remodeling
have not been demonstrated. In this study, we show that eph-
rinB2 and EphB4 regulate differentiation of osteoclasts and
osteoblasts, suppressing bone resorption and enhancing bone
formation. Our results indicate that ephrinB2 expression is de-
pendent on a RANKL-induced c-Fos-NFATc1 transcriptional
cascade during osteoclast differentiation. On the other hand,
EphB4 expression is constitutive and the forward signaling
through EphB4 induces osteogenic regulatory factors, such as
Dlx5, Osx, and Runx2, in calvarial osteoblasts, suggesting that
EphB4 is at the top of the regulatory cascade during osteoblast
We observed that reverse signaling through ephrinB2 into he-
matopoietic precursors suppresses osteoclast differentiation.
Although ephrin-Eph signaling frequently decreases cell adhe-
sion in various cell types, we did not observe decreased adhe-
sion (increased detachment) of osteoclast precursors stimu-
lated with EphB4 (data not shown). Therefore, the observed
inhibitory effects of ephrinB2 are not due to a loss of osteoclast
precursors from the culture. How then does ephrinB2 reverse
signaling suppress osteoclast differentiation? Our data indicate
that such suppression requires the ephrinB2 C-terminal YKV
motif, suggesting the involvement of as yet unidentified PDZ do-
main proteins in mediating the signal. Although how signals
downstream of M-CSF or RANKL are abrogated in these cells
remains unknown, transcription of Fos and its target Nfatc1 is
inhibited by the ephrinB2 reverse signaling, and overexpression
of c-Fos or NFATc1 relieves the inhibitory effect. Therefore, the
ative feedback loop through ephrinB2.
Bidirectional ephrin-Eph signaling links the negative feedback
loop in osteoclast differentiation to positive regulation of osteo-
blast differentiation, thereby maintaining bone homeostasis.
By overexpressing EphB4, we showed that forward signaling
through EphB4 enhances osteoblast differentiation both in cul-
tured cells and in mice. Consistently, knockdown of EphB4 sig-
naling by siRNA resulted in reduced osteoblast differentiation.
EphB receptors are known to interact with various signaling
of Src, PI3K-Akt (Steinle et al., 2002), and the Ras/Rho families
ofsmallmolecular weight GTPases(NorenandPasquale,2004).
We showed that EphB4 signaling activated ERK1/2 pathways,
which may explain enhanced osteoblast differentiation (Jaiswal
et al., 2000; Wang et al., 2002; Xiao et al., 2000). We also ob-
served that EphB4 signaling attenuates RhoA activity. Although
this is in agreement with the notion that higher RhoA activity in-
hibits osteoblast differentiation (Harmey et al., 2004), McBeath
et al. (2004) using human mesenchymal stem cells suggest
that active RhoA enhances osteoblast differentiation. In mouse
primary calvarial osteoblasts, a constitutively active form of
RhoA inhibited and a dominant-negative form of RhoA stimu-
lated osteoblast differentiation, supporting the notion that
EphB4-forward signaling enhances osteoblast differentiation
by lowering RhoA activity. Whether and how EphB4-forward
signaling critically affects ERKs, RhoA, or other heterologous
signaling pathways should be determined by future pharmaco-
logical or genetic studies.
Transgenic mice overexpressing EphB4 showed increased
femoral bone density. Although the bone formation rate was el-
tion may be enhanced in transgenic mice. On the other hand,
EphB4 transgenic mice showed a reduction in osteoclast num-
ber and osteoclast surface, suggesting that reverse signaling
through ephrinB2 into osteoclasts was increased by forced ex-
pression of EphB4 in vivo. Furthermore, the decrease in urinary
DPD crosslinks provides additional support for the idea that re-
EphB4 transgenic mice. Since a reduced number of osteoclasts
should lead to overall reduction in forward signaling in osteo-
blasts, the net result of EphB4 overexpression on bone metabo-
lism in vivo would be complex compared to EphB4 overexpres-
sion in cultured osteoblasts stimulated with soluble ephrinB2.
Why do conditional knockout mice lacking ephrinB2 in the
myeloid lineage not exhibit a strong bone phenotype? The na-
ture of reverse signaling of ephrinB1 and ephrinB2 is likely to
be similar because the amino acid sequence of their intracellular
domains is highly conserved. This implies that upon Eph bind-
ing, ephrinB1, which is also expressed in differentiating osteo-
clasts, signals in a manner similar ephrinB2. Since EphB4
does not interact with ephrinB1, other Eph receptors such as
EphB2 or EphB3, which are also expressed in osteoblasts,
may interact with ephrinB1. Eph receptors other than EphB4
likely compensate for forward signaling mediated by EphB4 in
the absence of ephrinB2.
Incontrast towhatwe showin this study,the word ‘‘coupling’’
traditionally refers to the positive correlation between systemic
bone resorption and bone formation (Harris and Heaney,
1969). Our finding is consistent with such ‘‘coupling’’ when
each bone remodeling cycle is considered at a microscopic
level. To achieve coupling, termination of osteoclastic bone re-
sorption has to be followed by osteoblastic bone formation,
which refills resorption lacunae with new bone (Hattner et al.,
1965; Parfitt, 1994). We suggest that a bidirectional interaction
between ephrinB ligands expressed on osteoclasts and Eph re-
ceptors on osteoblasts can dynamically enhance the transition
from resorption to a reversal phase occurring at each resorption
cycle (Parfitt, 1994), when osteoclast activity is attenuated and
osteoblasts initiate bone formation at resorption lacunae. Then,
an osteoclast-free area might be maintained by ephrinB reverse
signaling to osteoclast precursors. Furthermore, continuous
bone formation by osteoblasts over the several months required
to refill resorption lacunae might be self-maintained by ephrin-
Eph interactions on osteoblasts. This study establishes the con-
Mouse Ephb4 cDNA was subcloned into the NotI site of modified pNASSb,
which contains the 2.3 kb osteoblast-specific promoter region for the mouse
pro-a1 (I) collagen gene (Liu et al., 2001). The NarI-SalI fragment containing
the 2.3 kb pro-a1 (I) promoter-Ephb4 coding region-poly (A) signal was iso-
lated and microinjected into pronuclei of fertilized eggs from B6C3F1
(C57BL/6 3 C3H/He) females. Fos2/2mice were produced by mating
Fos+/2mice on a 129 3 C57BL/6 mixed background (Wang et al., 1992).
A R T I C L E
CELL METABOLISM : AUGUST 2006
To specifically delete ephrinB2 in myeloid cells, conditional ephrinB2 knock-
out mice (Gerety and Anderson, 2002) and LysMcre mice (Clausen et al.,
1999) were crossed. All animals were treated in accordance with the guide-
lines of Keio University for animal and recombinant DNA experiments.
Bone phenotype analysis
urinary DPD measurement were performed as described (Nishiwaki et al.,
2006). Osteocalcin was quantified with an ELISA kit (Biomedical Technolo-
gies). Bone tissues and cultured cells were immunostained using polyclonal
anti-EphB4 goat (R&D), polyclonal anti-ephrinB2 goat (R&D), and Alexa647-
conjugated anti-goat chicken (Invitrogen) antibodies. Fluorescent images
were acquired using a Fluoview1000 (Olympus) confocal microscope.
Bone resorption assay
Bone slices were prepared from diaphysis of bovine long bones cut out with
a band saw. The diaphysis tube was cut longitudinally into pieces, and bone
marrow and softtissue were removed under running water. Blocksof cortical
bone were immersed in xylene for 30 min, transferred to alcohol for 30 min,
dried overnight in oven at 45?C, and sliced with a diamond saw (Isomet,
Buehler). Slices were washed three times in water, sonicated for 10 min,
washed three times in water, immersed in acetone for 10 min, washed in al-
cohol twice, drained, and dried in air. The bone slices were placed in the bot-
tom of wells and cells were cultured on top of bone slices. Cells on bone sli-
ces were removed in 10% sodium hypochlorite. Air dried bone slices were
stained with hematoxylin. The entire surface of each bone slice was exam-
ined and the total area resorbed per bone slice was quantified blind using
ImageJ (National Institutes of Health).
DC (deletion of entire cytoplasmic domain), D33 (deletion of the last C-termi-
nal 33 amino acids), and DYKV were generated by PCR-mediated mutagen-
esis based on mouse Efnb2 cDNA and cloned into pMX-IRES-GFP and
were substituted with phenylalanines) was generated using a QuickChange
site-directed mutagenesis kit (Stratagene). Expression of mutants was con-
firmed in COS7 cells by transient transfection and immunoblotting. Ephb4
cDNA was cloned between the BamHI and XhoI sites of pCMV-Tag4A (Stra-
tagene). Constitutively active RhoA (V14) and dominant-negative RhoA (N19)
cDNAs (Gebbink et al., 1997) were cloned into the EcoRI and NotI sites of
pMX-IRES-GFP. Inserts were verified by DNA sequencing. Retroviral Efnb2
siRNA vectors were constructed in RVH1 (Barton and Medzhitov, 2002).
The CD4 cDNA of RVH1 was replaced by GFP to generate RVH1-GFP.
Efnb2 siRNA sequences were:
Ephb4 siRNA 1 sequences were:
Ephb4 siRNA 2 sequences were:
as osteoclast precursors. Briefly, total bone marrow cells were flushed out
from wild-type femurs or tibias and were cultured overnight in tissue culture
dishesin a-minimalessentialmedium(a-MEM) containing 10%FCS. Nonad-
or 5 3 104cells per 48-well plate in a-MEM containing 10 ng/ml of M-CSF
saline and adherent cells were used as MDMs. Fos2/2MDMs were prepared
from splenocyotes. With or without retroviral infection, MDMs were induced
to differentiate into osteoclasts in the presence of 10 ng/ml each of M-CSF
and RANKL (R&D) for 3 days. Adherent cells were fixed with 4% paraformal-
dehyde, treated with ethanol-acetone (50:50), and stained for TRAP activity
using a kit (Sigma). TRAP activity in cell lysates was quantified by measuring
absorbance at 405 nm using a kit (Sigma). For cocultures, bone marrow cells
(nonadherent cells obtained after culturing total bone marrow cells for 18 hr)
were seeded at a density of 5 3 105cells per 48-well plate containing 5 3
104ST2 cells in 0.35 ml a-MEM supplemented with 10% FCS, 1028M 1,
25-dihydroxyvitamin D3, and 1027M dexamethasone. The medium was
changed every 2 days. Mouse primary calvarial osteoblasts were prepared,
differentiated, and stained as described (Nishiwaki et al., 2006). Images of
wells stained for ALP activity were processed using ImageJ. To establish
EphB4-overexpressing stromal cells, ST2 cells (2 3 105) were lipofected
with 4 mg of pCMV-EphB4-Tag4A using FuGENE 6 (Roche). Cells were se-
lected in medium containing 500 mg/ml of geneticin (G418). Receptor body
staining for EphB4 was as described (Fu ¨ller et al., 2003). Twelve EphB4
expression clones were obtained and tested for osteoclastogenesis in the
coculture system. To stimulate forward or reverse signaling, ephrinB2-Fc,
with anti-Fc antibody or coated on poly-L-lysine coverglasses at 2 mg/ml
unless otherwise indicated.
Retroviral gene transduction
Retroviruses were produced as described (Nishiwaki et al., 2006). Recombi-
nant retrovirus was used to infect primary bone marrow cells or calvarial os-
teoblasts in the presence of 8 mg/ml polybrene (Sigma).
RT-PCR and quantitative PCR analysis
synthesized using the SuperScript First-Strand Synthesis System for an
RT-PCR kit (Invitrogen). RT-PCRprimersforAcp5andCalcrareasdescribed
(Matsuo et al., 2004) and others are listed in Table S1. TaqMan RT-PCR
primers for Efnb2, Ephb4, Ephb6, Epha4, Fcgr3, Fos, Gapdh, and Nfatc1
Proteins prepared from osteoclastic or osteoblastic cultures were separated
on 4%–12% SDS-polyacrylamide gels (Novex) and transferred onto Hybond
nitrocellulose membranes (Amersham). The following primary antibodies
were used: polyclonal anti-c-Fos rabbit antibody (Ab-1, Oncogene), a mono-
clonal anti-NFATc1 mouse antibody (7A6, Santa Cruz), and polyclonal anti-
ERK1/2rabbit andpolyclonalanti-ephrinB2 goat antibodies (R&D).Where in-
dicated, blots were stripped and reprobed with a polyclonal anti-actin goat
antibody (Santa Cruz) to monitor protein loading. RhoA activity was analyzed
by an affinity precipitation assay using a GST-tagged fusion protein of the
mouse Rhotekin Rho binding domain (Upstate). GTP-RhoA and total RhoA
were detected using a monoclonal anti-RhoA mouse antibody (26C4, Santa
Supplemental data include four figures and one table and can be found with
this article online at http://www.cellmetabolism.org/cgi/content/full/4/2/111/
We thank R. Medzhitov for siRNA vectors, D. Anderson for conditional eph-
rinB2 knockout mice, I. Fo ¨rster for LysMcre mice, N. Clipstone for NFATc1
retroviral vectors, H. van Dam for RhoA mutants, H. Murayama, M. Tsuri-
Jinno, H. Suzuki, K. Maruyama, and E. Kobayashi for technical assistance,
for encouragement, and N. Ray, H. Takayanagi, S. Koyasu, L. Bakiri, D. Gal-
son, E.F. Wagner, L. DiMascio, and E. Lamar for critical reading of the man-
uscript. This work was supported by a Grant-in-Aid for Scientific Research B
(17390420) and by Keio Gijuku Academic Development Funds.
Received: December 29, 2005
Revised: April 5, 2006
Accepted: May 12, 2006
Published: August 8, 2006
ephrin-Eph signaling in bone homeostasis
CELL METABOLISM : AUGUST 2006119
Adams, R.H., Wilkinson, G.A., Weiss, C., Diella, F., Gale, N.W., Deutsch, U.,
Risau, W., and Klein, R. (1999). Roles of ephrinB ligands and EphB receptors
in cardiovascular development: demarcation of arterial/venous domains,
vascular morphogenesis, and sprouting angiogenesis. Genes Dev. 13,
Asagiri, M., Sato, K., Usami, T., Ochi, S., Nishina, H., Yoshida, H., Morita, I.,
Wagner, E.F., Mak, T.W., Serfling, E., and Takayanagi, H. (2005). Autoampli-
fication of NFATc1 expression determines its essential role in bone homeo-
stasis. J. Exp. Med. 202, 1261–1269.
RNA into primary cells. Proc. Natl. Acad. Sci. USA 99, 14943–14945.
Batlle, E., Henderson, J.T., Beghtel, H., van den Born, M.M., Sancho, E.,
Huls, G., Meeldijk, J., Robertson, J., van de Wetering, M., Pawson, T., and
epithelium by controlling theexpressionofEphB/ephrinB. Cell111,251–263.
Clausen, B.E., Burkhardt, C., Reith, W., Renkawitz, R., and Fo ¨rster, I. (1999).
Conditional gene targeting in macrophages and granulocytes using LysMcre
mice. Transgenic Res. 8, 265–277.
Compagni, A., Logan, M., Klein, R., and Adams, R.H. (2003). Control of skel-
etal patterning by ephrinB1-EphB interactions. Dev. Cell 5, 217–230.
L.A., Srivastava, D., and Henkemeyer, M. (2004). Ephrin-B2reverse signaling
is required for axon pathfinding and cardiac valve formation but not early
vascular development. Dev. Biol. 271, 263–271.
Davy, A., and Soriano, P. (2005). Ephrin signaling in vivo: look both ways.
Dev. Dyn. 232, 1–10.
Davy, A., Aubin, J., and Soriano, P. (2004). Ephrin-B1 forward and reverse
signaling are required during mouse development. Genes Dev. 18, 572–583.
Ducy, P., Zhang, R., Geoffroy, V., Ridall, A.L., and Karsenty, G. (1997). Osf2/
Cbfa1: a transcriptional activator of osteoblast differentiation. Cell 89, 747–
Etienne-Manneville, S., and Hall, A. (2002). Rho GTPases in cell biology.
Nature 420, 629–635.
nardi, A., Tran, T., Boyce, B.F., and Siebenlist, U. (1997). Requirement for
NF-kB in osteoclast and B-cell development. Genes Dev. 11, 3482–3496.
Fu ¨ller, T., Korff, T., Kilian, A., Dandekar, G., and Augustin, H.G. (2003). For-
ward EphB4 signaling in endothelial cells controls cellular repulsion and seg-
regation from ephrinB2 positive cells. J. Cell Sci. 116, 2461–2470.
Gale, N.W., and Yancopoulos, G.D. (1999). Growth factors acting via endo-
thelial cell-specific receptor tyrosine kinases: VEGFs, angiopoietins, and
ephrins in vascular development. Genes Dev. 13, 1055–1066.
Gebbink, M.F., Kranenburg, O., Poland, M., van Horck, F.P., Houssa, B., and
Moolenaar, W.H.(1997). Identification of a novel, putative Rho-specific GDP/
GTP exchange factor and a RhoA-binding protein: control of neuronal mor-
phology. J. Cell Biol. 137, 1603–1613.
Gerety, S.S., and Anderson, D.J. (2002). Cardiovascular ephrinB2 function is
essential for embryonic angiogenesis. Development 129, 1397–1410.
Gerety, S.S., Wang, H.U., Chen, Z.F., and Anderson, D.J. (1999). Symmetri-
cal mutant phenotypes of the receptor EphB4 and its specific transmem-
brane ligand ephrin-B2 in cardiovascular development. Mol. Cell 4, 403–414.
Grigoriadis, A.E., Wang, Z.Q., Cecchini, M.G., Hofstetter, W., Felix, R.,
Fleisch, H.A., and Wagner, E.F. (1994). c-Fos: a key regulator of osteo-
clast-macrophage lineage determination and bone remodeling. Science
Harmey, D., Stenbeck, G., Nobes, C.D., Lax, A.J., and Grigoriadis, A.E.
(2004). Regulation ofosteoblast differentiation by Pasteurella multocida toxin
(PMT): a role for Rho GTPase in bone formation. J. Bone Miner. Res. 19, 661–
Harris, W.H., and Heaney, R.P. (1969). Skeletal renewal and metabolic bone
disease. N. Engl. J. Med. 280, 253–259.
Hattner, R., Epker, B.N., and Frost, H.M. (1965). Suggested sequential mode
of control of changes in cell behaviour in adult bone remodelling. Nature 206,
Hayden, J.M., Mohan, S., and Baylink, D.J. (1995). The insulin-like growth
Himanen, J.P., and Nikolov, D.B. (2003). Eph receptors and ephrins. Int. J.
Biochem. Cell Biol. 35, 130–134.
Ishida, N., Hayashi, K., Hoshijima, M., Ogawa, T., Koga, S., Miyatake, Y., Ku-
megawa, M., Kimura, T., and Takeya, T. (2002). Large scale gene expression
analysis of osteoclastogenesis in vitro and elucidation of NFAT2 as a key
regulator. J. Biol. Chem. 277, 41147–41156.
Jaiswal, R.K., Jaiswal, N., Bruder, S.P., Mbalaviele, G., Marshak, D.R., and
Pittenger, M.F. (2000). Adult human mesenchymal stem cell differentiation
to the osteogenic or adipogenic lineage is regulated by mitogen-activated
protein kinase. J. Biol. Chem. 275, 9645–9652.
Kanzaki, S., Ito, M., Takada, Y., Ogawa, K., and Matsuo, K. (2006). Resorp-
tion of auditory ossicles and hearing loss in mice lacking osteoprotegerin.
Bone 39, 414–419.
Karsenty, G., and Wagner, E.F. (2002). Reaching a genetic and molecular un-
derstanding of skeletal development. Dev. Cell 2, 389–406.
Komori,T.,Yagi,H.,Nomura, S.,Yamaguchi, A.,Sasaki,K.,Deguchi, K.,Shi-
mizu, Y., Bronson, R.T., Gao, Y.H., Inada, M., et al. (1997). Targeted disrup-
tion of Cbfa1 results in a complete lack of bone formation owing to matura-
tional arrest of osteoblasts. Cell 89, 755–764.
Kullander, K., and Klein, R. (2002). Mechanisms and functions of Eph and
ephrin signalling. Nat. Rev. Mol. Cell Biol. 3, 475–486.
M., Narai,S., Yamaguchi, A.,andKomori,T.(2001).Overexpression ofCbfa1
in osteoblasts inhibits osteoblast maturation and causes osteopenia with
multiple fractures. J. Cell Biol. 155, 157–166.
Martin, T.J. (2004). Does bone resorption inhibition affect the anabolic re-
sponse to parathyroid hormone? Trends Endocrinol. Metab. 15, 49–50.
Matsuo, K., Galson, D.L., Zhao, C., Peng, L., Laplace, C., Wang, K.Z., Bach-
ler, M.A., Amano, H., Aburatani, H., Ishikawa, H., and Wagner, E.F. (2004).
Nuclear factor of activated T-cells (NFAT) rescues osteoclastogenesis in pre-
cursors lacking c-Fos. J. Biol. Chem. 279, 26475–26480.
McBeath, R., Pirone, D.M., Nelson, C.M., Bhadriraju, K., and Chen, C.S.
(2004). Cell shape, cytoskeletal tension, and RhoA regulate stem cell lineage
commitment. Dev. Cell 6, 483–495.
Murai, K.K., and Pasquale, E.B. (2003). ‘Eph’ective signaling: forward, re-
verse and crosstalk. J. Cell Sci. 116, 2823–2832.
Myshkin, E., and Wang, B. (2003). Chemometrical classification of ephrin li-
gands and Eph kinases using GRID/CPCA approach. J. Chem. Inf. Comput.
Sci. 43, 1004–1010.
Nakashima, K., Zhou, X., Kunkel, G., Zhang, Z., Deng, J.M., Behringer, R.R.,
and de Crombrugghe, B. (2002). The novel zinc finger-containing transcrip-
tion factor osterix is required for osteoblast differentiation and bone forma-
tion. Cell 108, 17–29.
Nishiwaki, T., Yamaguchi, T., Zhao, C., Amno, H., Hankenson, K.D., Born-
stein, P., Toyama, Y., and Matsuo, K. (2006). Reduced expression of throm-
bospondins and craniofacial dysmorphism in mice overexpressing Fra1.
J. Bone Miner. Res. 21, 596–604.
Noren, N.K., and Pasquale, E.B. (2004). Eph receptor-ephrin bidirectional
signals that target Ras and Rho proteins. Cell. Signal. 16, 655–666.
Palmer, A., and Klein, R. (2003). Multiple roles of ephrins in morphogenesis,
neuronal networking, and brain function. Genes Dev. 17, 1429–1450.
Parfitt, A.M. (1994). Osteonal and hemi-osteonal remodeling: the spatial and
temporal framework for signal traffic in adult human bone. J. Cell. Biochem.
A R T I C L E
CELL METABOLISM : AUGUST 2006
Pasquale, E.B. (2005). Ephreceptor signallingcasts awide net oncell behav- Download full-text
iour. Nat. Rev. Mol. Cell Biol. 6, 462–475.
Pfeilschifter, J., and Mundy, G.R. (1987). Modulation of type beta transform-
ing growth factor activity in bone cultures by osteotropic hormones. Proc.
Natl. Acad. Sci. USA 84, 2024–2028.
Ray, N., Kuwahara, M., Takada, Y., Maruyama, K., Kawaguchi, T., Tsubone,
H., Ishikawa, H., and Matsuo, K. (2006). c-Fos suppresses systemic inflam-
matory response to endotoxin. Int. Immunol. 18, 671–677.
Sharfe, N., Freywald, A., Toro, A., Dadi, H., and Roifman, C. (2002). Ephrin
stimulation modulates T cell chemotaxis. Eur. J. Immunol. 32, 3745–3755.
Sheu, T.-J., Schwarz, E.M., Martinez, D.A., O’Keefe, R.J., Rosier, R.N., Zus-
cik, M.J., and Puzas, J.E. (2003). A phage display technique identifies a novel
regulator of cell differentiation. J. Biol. Chem. 278, 438–443.
Steinle, J.J., Meininger, C.J., Forough, R., Wu, G., Wu, M.H., and Granger,
H.J. (2002). Eph B4 receptor signaling mediates endothelial cell migration
and proliferation via the phosphatidylinositol 3-kinase pathway. J. Biol.
Chem. 277, 43830–43835.
A., Isobe, M., Yokochi, T., Inoue, J., et al. (2002a). Induction and activation of
the transcription factor NFATc1 (NFAT2) integrate RANKL signaling in termi-
nal differentiation of osteoclasts. Dev. Cell 3, 889–901.
Takayanagi, H.,Kim, S., Matsuo, K., Suzuki, H.,Suzuki, T., Sato, K.,Yokochi,
T., Oda, H., Nakamura, K., Ida, N., et al. (2002b). RANKL maintains bone
homeostasis through c-Fos-dependent induction of interferon-b. Nature
Teitelbaum, S.L., and Ross, F.P. (2003). Genetic regulation of osteoclast de-
velopment and function. Nat. Rev. Genet. 4, 638–649.
Wang, F.-S., Wang, C.-J., Sheen-Chen, S.-M., Kuo, Y.-R., Chen, R.-F., and
Yang, K.D. (2002). Superoxide mediates shock wave induction of ERK-
dependent osteogenic transcription factor (CBFA1) and mesenchymal cell
differentiation toward osteoprogenitors. J. Biol. Chem. 277, 10931–10937.
Wang,H.U., Chen, Z.F., and Anderson,D.J.(1998).Molecular distinction and
angiogenic interaction between embryonic arteries and veins revealed by
ephrin-B2 and its receptor Eph-B4. Cell 93, 741–753.
Wang, Z.Q., Ovitt, C., Grigoriadis, A.E., Mo ¨hle-Steinlein, U., Ru ¨ther, U., and
Wagner, E.F. (1992). Bone and haematopoietic defects in mice lacking
c-fos. Nature 360, 741–744.
Wang,Z.,Cohen,K.,Shao, Y.,Mole,P., Dombkowski,D., and Scadden, D.T.
(2004). Ephrin receptor, EphB4, regulates ES cell differentiation of primitive
mammalian hemangioblasts, blood, cardiomyocytes, and blood vessels.
Blood 103, 100–109.
Xiao, G., Jiang, D., Thomas, P., Benson, M.D., Guan, K., Karsenty, G., and
Franceschi, R.T. (2000). MAPK pathways activate and phosphorylate the os-
teoblast-specific transcription factor, Cbfa1. J. Biol. Chem. 275, 4453–4459.
Yu, G., Luo, H., Wu, Y., and Wu, J. (2003). Ephrin B2 induces T cell costimu-
lation. J. Immunol. 171, 106–114.
ephrin-Eph signaling in bone homeostasis
CELL METABOLISM : AUGUST 2006121