Full text of DOI:10.1359/jbmr.2002.17.8.1486

JOURNAL OF BONE AND MINERAL RESEARCH
Volume 17, Number 8, 2002
©
2002 American Society for Bone and Mineral Research
Colocalization of Intracellular Osteopontin With CD44 Is
Associated With Migration, Cell Fusion, and Resorption
in Osteoclasts
1
1
2
2
3
K. SUZUKI, B. ZHU, S.R. RITTLING, D.T. DENHARDT, H.A. GOLDBERG,
1
1
C.A.G. MCCULLOCH, and J. SODEK
ABSTRACT
Although osteopontin (OPN) is recognized generally as a secreted protein, an intracellular form of osteopontin
iOPN), associated with the CD44 complex, has been identified in migrating fibroblastic cells. Because both
(
OPN and CD44 are expressed at high levels in osteoclasts, we have used double immunofluorescence analysis
and confocal microscopy to determine whether colocalization of these proteins has functional significance in
the formation and activity of osteoclasts. Analysis of rat bone marrow–derived osteoclasts revealed strong
surface staining for CD44 and ␤1- and ␤3-integrins, whereas little or no staining for OPN or bone sialoprotein
(
BSP) was observed in nonpermeabilized cells. In permeabilized perfusion osteoclasts and multinucleated
osteoclasts, staining for OPN and CD44 was prominent in cell processes, including filopodia and pseudopodia.
Confocal microscopy revealed a high degree of colocalization of OPN with CD44 in motile osteoclasts. In cells
treated with cycloheximide (CHX), perinuclear staining for OPN and BSP was lost, but iOPN staining was
retained within cell processes. In osteoclasts generated from the OPN-null and CD44-null mice, cell spreading
and protrusion of pseudopodia were reduced and cell fusion was impaired. Moreover, osteoclast motility and
resorptive activity were significantly compromised. Although the area resorbed by OPN-null osteoclasts could
be rescued partially by exogenous OPN, the resorption depth was not affected. These studies have identified
an intracellular form of OPN, colocalizing with CD44 in cell processes, that appears to function in the
formation and activity of osteoclasts. (J Bone Miner Res 2002;17:1486–1497)
Key words: osteopontin, CD44, osteoclasts, osteoclastogenesis, cell migration
INTRODUCTION
tion in the adult, while maintaining bone homeostasis. Con-
sequently, regulation of osteoclast formation and activity,
which is linked closely to bone-forming osteoblasts through
hormonal and cytokine control, is critical for maintaining
bone mass. Disruption of the balance between osteoclast
and osteoblast activities is evident in several metabolic bone
diseases such as osteopetrosis and osteoporosis, in which
the activity of osteoclasts relative to osteoblasts is reduced
or increased, respectively. Osteoclasts, like macrophages,
are derived from the granulocyte/monocyte myeloid cell
STEOCLASTS ARE specialized cells that are responsible
for the physiological and pathological dissolution of
mineral and the breakdown of the organic matrix of miner-
alized connective tissues. As such, they have a critical role
in the remodeling of bone, allowing structural changes to
occur during growth and development and functional adap-
O
All authors have no conflict of interest.
1
CIHR Group in Matrix Dynamics, Faculty of Dentistry, University of Toronto, Toronto, Canada.
Department of Cell Biology, Rutgers University, Piscataway, New Jersey, USA.
2
3
CIHR Group in Skeletal Development and Remodeling, School of Dentistry, University of Western Ontario, London, Ontario, Canada.
1
486

OPN AND CD44 IN OSTEOCLASTOGENESIS
1487
lineage. Osteopontin (OPN) and CD44 are expressed by F129 mice (wild-type [WT] and OPN) C57BL/6 (WT and
promyelocytic cells early in the development of osteoclasts CD44-null) were treated with 0.83% NH Cl in 10 mM of
4
(1,2)
and macrophages
and coexpression of these proteins has Tris-HCl buffer (pH 7.4) for 20 minutes on ice to hemolyze
been reported in differentiating HL-60 monocytes. More- red blood cells. The remaining cells were suspended in a
(3)
over, these proteins are highly expressed in “multinu- small amount of ␣-minimal essential medium (␣-MEM) and
(4,5)
cleated” osteoclasts.
Although OPN-null mice show applied onto a Sephadex G-10 (Amersham-Pharmacia Bio-
anomalies in both the formation and the activity of tech, Baie d’Urfe, Quebec, Canada) column and incubated
(
6–8)
osteoclasts,
the precise role of OPN in bone resorption for 45 minutes at 37°C to remove macrophages and stromal
(30)
has not been elucidated.
cells.
Nonadherent cells were collected by eluting with
5
OPN is a phosphorylated glycoprotein, expressed as a the culture media and plated at a density of 5 ϫ 10
secreted protein in a variety of tissues, that can regulate cells/cm and then cultured in ␣-MEM with 10% fetal calf
2
mineral crystal growth through an aspartate-rich sequence serum (FCS; adjusted to pH 7.0 after adding FCS). Lab-Tek
and mediate cell attachment and signaling through an RGD chamber slides (8-well, glass slide; Nunc, Roskilde, Den-
(9–14)
motif that recognizes several integrins.
Recent studies mark) were coated with the mixture of 20 ng/ml of either
also have shown that CD44 is a receptor for OPN that recombinant murine (R & D Systems, Minneapolis, MN,
mediates both chemotactic and signaling activities in mac- USA) or human (Pepro Tech, Rocky Hill, NJ, USA) colony-
(
15,16)
rophages.
Various forms of OPN have been recog- stimulating factor-1 (CSF-1) and 60 ng/ml of recombinant
nized that are generated primarily by differences in post- human receptor activator of nuclear factor ␬B (NF-␬B)
translational modifications that influence their ability to ligand (RANKL; Pepro Tech) and air-dried for 1 h. In some
(
17–19)
bind to other matrix proteins,
growth,
regulate mineral crystal cultures, the chamber slides were precoated with bovine
(
20)
(21)
and mediate osteoclast attachment.
serum albumin (BSA; 5 ␮g/ml), OPN (5 ␮g/ml), or bone
CD44, which is the major cell-surface receptor for hyal- sialoprotein (BSP; 5 ␮g/ml) by incubating overnight at
(
22,23)
uronate,
is a transmembrane glycoprotein that also is 37°C. Both the OPN and the BSP were purified to homo-
expressed in a number of variant forms. The heterogeneity geneity from E-extracts of porcine bone as described pre-
(31)
of CD44 is generated by alternative splicing of at least 10 viously,
and the concentrations were determined from
exons (exons 6–15) that form the proximal extracellular amino acid analysis. After removing the coating solution
domain with additional variation introduced through post- and washing with phosphate-buffered saline (PBS), bone
(24,25)
translational modifications.
These variant receptor marrow monocytes were plated as described previously.
forms mediate cell adhesion, migration, and signaling ac- The mouse marrow cells were grown for 3–8 days in the
tivities in a variety of cells including immune cells and presence of 20 ng/ml of CSF-1 and 60 ng/ml of RANKL
(25)
transformed cells.
However, the most prevalent form of and for some experiments with 50 ng/ml of recombinant
CD44 is an 85- to 95-kDa protein in which the amino murine tumor necrosis factor-␣ (TNF-␣; Pepro Tech), as
terminus contains two binding sites for hyaluronan and the detailed in following section and in the Results section.
carboxy-terminal cytoplasmic domain includes two putative
sites for ankyrin binding and four potential sites for serine
phosphorylation. The CD44 receptor complex is linked to
Immunostaining
the cytoskeleton through ezrin, radixin, and moesin (ERM)
proteins, which interact with the cytoplasmic domain.
For immunofluorescent staining, cells were either incu-
bated with primary antibodies 1 h before fixation (nonper-
(26,27)
Recently, CD44 variants have been shown to cooperate with meabilized cell staining) or first fixed with 2% paraformal-
␤
1-integrins in binding OPN as an important step in the dehyde for 15 minutes at 4°C (permeabilized cell staining).
(28)
stimulation of cell motility and chemotaxis.
To minimize nonspecific staining of permeabilized cells, the
Although OPN generally is recognized as a secreted fixed cells were incubated with 2% BSA in PBS overnight
protein, previously, we have identified an intracellular form at 4°C. For double staining, cells were incubated with a
of OPN (iOPN) that is associated with the CD44 complex in mixture of two kinds of primary antibodies diluted in 2%
migrating fibroblastic cells, activated macrophages, and BSA-0.01% Triton-X in PBS (BSA buffer) for 1 h at room
(29)
metastatic cells. Because osteoclasts also are motile cells temperature (RT, 21°C). After three washes of 10 minutes
and share with macrophages a common myeloid lineage, we each with 2% BSA in PBS, cells were incubated with the
have used double immunofluorescence and confocal mi- appropriate secondary antibodies diluted with BSA buffer
croscopy to determine whether OPN associates with the for 1 h at RT. After three washes of 10 minutes each with
CD44 receptor complex in osteoclasts. Further, to determine 2% BSA in PBS, cells were incubated with 4,6-diamidino-
the functional significance of this putative association, we 2-phenylindole (DAPI) for nuclear staining.
have studied the formation and activity of osteoclasts de-
rived from OPNϪ/Ϫ and CD44Ϫ/Ϫ mice.
The following primary antibodies were used at the dilu-
tions given in brackets: rabbit anti-human OPN LF123,
which reacts with rat and mouse OPN (used at a 1:200
dilution), was a kind gift from Dr. L.W. Fisher (National
Institutes of Health [NIH], Bethesda, MD, USA); mouse
anti-rat (1:200; MPIIIB1O1; Developmental Studies Hy-
bridoma Bank, John Hopkins University, Baltimore, MD,
MATERIALS AND METHODS
Preparation of osteoclasts
Bone marrow cells obtained from femurs of 5-week-old USA; under contract from National Institute of Child Health
Wister rats or from tibias and femurs of 5- to 7-week-old and Human Development [NICHD]); biotin-labeled mouse

1
488
SUZUKI ET AL.
anti-rat CD44 (1:200; clone OX-49; PharMingen, Missis- and replated onto the membrane and allowed to attach for
sauga, Ontario, Canada), biotin-labeled rat anti-mouse 2 h. CSF-1 (10 ng/ml) was added to the lower well as a
CD44 (1:200; clone KM201; Southern Biotechnology As- chemoattractant and the cells were allowed to migrate for
sociates, Birmingham, AL, USA), mouse anti-rat integrin 2 h. Membranes were fixed with 2% paraformaldehyde and
␤3-chain (1:50; clone F11; PharMingen), mouse anti- removed carefully from the insert. The membranes then
human ␤1 (clone 18; Transduction Laboratories, Lexington, were washed with PBS and cut in half and the cells were
KY, USA), and rabbit anti-rat BSP (1:200). Secondary scraped from either side using a razor blade. The remaining
antibodies used were fluorescein isothiocyanate (FITC)– cells were stained for TRAP activity or for actin filaments
conjugated goat anti-rabbit immunoglobulin G (IgG; 1:200, using rhodamine-labeled phalloidin, and the cell number
whole molecule; Sigma-Aldrich, Oakville, Ontario, Can- was determined.
ada), TX red (TR)–conjugated goat anti-rabbit IgG (1:200,
H ϩ L chains; Vector Laboratories, Burlington, Ontario,
Canada), FITC-conjugated streptavidin (1:200; Cedarlane
Laboratories, Hornby, Ontario, Canada), TR-conjugated
streptavidin (1:200; Cedarlane Laboratories), FITC-
conjugated goat anti-mouse IgG (1:200, H ϩ L; Sigma),
TR-conjugated goat anti-mouse IgG (1:200, H ϩ L chains;
Cedarlane Laboratories), and Alexa Fluor 488 rabbit anti-
mouse IgG F(abЈ)2 (A-21204; Molecular Probes, Eugene,
OR, USA). For actin staining, either FITC- or rhodamine-
labeled phalloidin (Cedarlane Laboratories) was used at a
dilution of 1:500. Immunofluorescent images were acquired
using an air-cooled charge-coupled device (CCD) camera
Resorption assays
ϩ
Dentine slices: TRAP osteoclasts generated from bone
marrow monocytes in the presence of CSF-1 and RANKL
for 3 days were trypsinized and replated onto dentine slices
(
4 mm in diameter) placed in individual wells of a 96-well
plate. Cells were allowed to attach for 2 h. The dentine
slices with osteoclasts on top were then transferred to a
2
4-well plate and cultured for 48 h in the presence of CSF-1
and RANKL. The cells were removed by soaking the den-
tine slice in 1 M of NH OH solution overnight and soni-
4
cating for 1 minute. Resorption pits were examined by
backscattered electron microscopy after carbon coating and
the area was quantitated.
Osteologic slides: Individual wells of Osteologic chamber
slides (BD Biosciences, Billerica, MA, USA) were precoated
with BSA (100 ␮g/well), rat OPN (100 ng/well), or porcine
BSP (100 ng/well) by applying 30 ␮l of protein solution to the
(Quantix; Photometrics, Ltd., Tucson, AZ, USA) mounted
on an inverted fluorescent microscope (Axiovert 135; Zeiss,
Oberkochen, Germany). For FITC- or TR-labeled probes,
excitation was set at 490 nm or 555 nm and emission was
collected at 528 nm or 617 nm, respectively.
ϩ
well and drying the solution at 37°C. TRAP osteoclasts
Confocal microscopy
generated from bone marrow monocytes in the presence of
CSF-1 and RANKL for 3 days were trypsinized and replated
onto the Osteologic chamber slide and cultured for 48 h in the
presence of CSF-1 and RANKL. Then, cells were fixed with
Osteoclasts grown on Lab-Tek chamber slides and double-
stained for OPN and CD44 were examined and photographed
using a 63ϫ, 1.3 oil immersion objective, using epifluores-
cent optics and confocal imaging on a confocal laser scan-
ning microscope (CLSM; Leica, Heidelberg, Germany).
Transverse sections were imaged from the dorsal surface
down to the cell-substratum interface at 2-␮m intervals.
2% paraformaldehyde, permeabilized with 0.1% Triton-X for
5
minutes, and stained for actin filaments using rhodamine-
conjugated phalloidin for 30 minutes. After photographing the
actin staining, the cells were removed by treatment with a 20%
wt/vol solution of sodium hypochlorite for 1–2 minutes. After
washing with PBS, the discs were stained with von Kossa’s
reagent. In some cultures, 50 ng/ml of TNF-␣ was added to
stimulate the mineral resorption activity. The area of resorption
was determined using MetaMorph 4.5 (Universal Imaging
Corp., West Chester, PA, USA).
Analysis of osteoclast fusion
Osteoclasts prepared from bone marrow cells of WT,
OPN-null, and CD44-null mice were grown in the presence
of CSF-1 (20 ng/ml) and RANKL (60 ng/ml) for 8 days. To
increase fusion, a second series of cells were cultured in the
additional presence of tumor necrosis factor (TNF) ␣ (50
ng/ml). The cells were stained with DAPI and the number of
Statistical analysis
Quantitative data obtained for cell, fusion, cell migration,
and osteoclast resorption was averaged from a minimum of
ϩ
nuclei in tartrate-resistant acid phosphate–positive (TRAP )
cells was determined from the analysis of eight randomly
selected fields.
3
values in replicate experiments from which mean val-
ues Ϯ SD were calculated. Leven’s test was used to deter-
mine homogeneity of variance and two sample t-tests were
used to determine the difference between two means.
Cell migration assays
Transwell membranes (polycarbonate; pore size, 8 ␮m;
Corning Costar, Corning, NY, USA) were precoated on
both sides with BSA (100 ␮g/membrane) or on either the
upper or the lower surface with rat OPN (100 ng/membrane)
RESULTS
Immunocytochemical analysis of OPN and CD44 in rat
osteoclasts
ϩ
or porcine BSP (100 ng/membrane). TRAP prefusion os-
teoclasts generated from bone marrow monocytes in the
Previous studies have shown that OPN colocalizes with
presence of CSF-1 and RANKL for 3 days were trypsinized CD44 inside embryonic fibroblasts and influences their abil-

OPN AND CD44 IN OSTEOCLASTOGENESIS
1489
FIG. 1. Immunofluorescence analysis of permeabilized and nonpermeabilized rat osteoclasts. The localization of OPN relative to the (A and B)
CD44, (E) ␤3-integrin, and (F) ␤1-integrin are shown and compared with the relative distribution of (C) BSP and CD44 and (G) CD44 and
␤3-integrin in rat osteoclasts permeabilized by treatment with Triton-X 100. The OPN, like BSP, is stained strongly around the nucleus but unlike
BSP, also has a perimembranous distribution in lamellipodia (long arrows in panels A, B, E, and F) and in filopodia (short arrows in panels A
and B). In the cell processes the OPN has a similar distribution to the CD44 and ␤3- and ␤1-integrin receptors (arrows in panels A, B, E, and
F). The immunofluorescent staining generally was stronger with TR (CD44 in panels A, C, and D; OPN in panel B; ␤3 in panels E and H) than
FITC. CD44 and ␤3-integrin, which have a comparable distribution in pseudopodia (arrowheads) and filopodia (arrows) shown in panel G are
stained also strongly in (D and H) nonpermeabilized cells, whereas staining for OPN in the same cells was at background (minus primary antibody)
levels. The OPN was stained with a polyclonal antibody, LF 123, except in panel B in which a mouse anti-rat monoclonal antibody MPIIIB1O1
was used.
(29)
ity to migrate and bind hyaluronan.
Because OPN is a the fluorescent labels. As shown in Figs. 1A and 1B, when
ligand for ␣v␤3 and the CD44 receptor, both of which are TR was used for the CD44 and OPN, respectively, a stron-
(32–34)
enriched on the surface of osteoclasts,
we first ana- ger fluorescence was obtained. To determine whether the
lyzed the distribution of OPN relative to these receptors in pericellular staining for OPN was intracellular or extracel-
permeabilized multinucleated osteoclasts generated from rat lular, immunostaining was performed without cell perme-
bone marrow monocytes. Multinucleated osteoclasts were abilization (Figs. 1D and 1H). Although the CD44 and
isolated from rat femurs and stimulated with CSF-1 for 3 ␤3-integrin receptors stained strongly without cell perme-
minutes, double-stained using TR- and FITC-labeled sec- abilization, OPN immunostaining was negligible, reflecting
ond antibodies. OPN was localized primarily around the the intracellular location of the OPN and the extracellular
nucleus but also was present in lamellar structures and cell location of epitopes for the CD44 and ␤3.
processes, which also stained for CD44, and the ␤3- and
1-integrins (Figs. 1A, 1B, 1E, and 1F). In comparison, actively synthesized and secreted and intracellular OPN
BSP, a secreted protein with similar properties to OPN that (iOPN), putatively associated with the receptors bone marrow
To discriminate between the intracellular OPN that is being
␤
(35)
is also a ligand for ␣v␤3, was concentrated around nuclei monocytes grown for 3 days in the presence of CSF-1 and
but was not present in cell processes or at the cell surface RANKL were treated with 25 ␮g/ml of cycloheximide (CHX)
(Fig. 1C). Because the immunofluorescence generated with for various times before fixing and immunostaining. At this
FITC appeared to be weaker than with TR, the staining stage, most of the cells were still mononuclear. As shown by
intensity of OPN and CD44 was compared by alternating phase-contrast microscopy, cell processes in the developing

1
490
SUZUKI ET AL.
FIG. 2. Distribution of OPN and CD44 relative to actin. Rat osteoclasts were grown on chamber slides for 3 days and, before fixing and
permeabilization with Triton-X100, incubated with 25 ␮g/ml of CHX for 0–3 h. Cells were either double-stained for OPN and CD44 using FITC-
and TR-labeled second antibodies, respectively, or double-stained for either OPN and CD44, using FITC-labeled second antibodies, and actin,
using phalloidin-rhodamine. At this early stage of osteoclastogenesis, the osteoclastic cells form extensive cell processes, as shown in the
phase-contrast photograph in the top left panel. Cells immunostained for OPN and CD44 reveal staining for both proteins in cell processes and
in lamellae at the end of the processes (arrows). Although OPN and actin also are codistributed in the cell processes (arrows), the relationship
between the CD44 and actin is more readily apparent and they appear to be codistributed in the periphery of the cell body (arrows). Colocalization
of OPN and CD44 was still evident in some multinucleated osteoclasts 1 h after CHX treatment (CHX1h, arrows). However, with increasing time,
the cell processes were lost, as seen in the phase-contrast photograph (CHX2h), as was the colocalization of the OPN and CD44 (CHX2–3h).
Although staining for BSP and ␤3-integrin was lost within 1 h of CHX treatment (not shown), staining for OPN and CD44 was still quite strong
3
h after CHX treatment.
osteoclasts were lost within 2 h of CHX treatment (Fig. 2). At
To discriminate between the immunolocalization of OPN
the onset of CHX treatment, actin, used to evaluate changes in and CD44 more accurately, permeabilized rat osteoclasts were
cytoskeletal architecture that are important for migration and immunostained for both proteins and optical sections were
cell fusion, was prominent in long cell processes, which also examined using CLSM (Fig. 3A, a–l). In multinucleated oste-
stained for OPN but more strongly for CD44. Although the oclasts with extensive filopodia and lamellipodia, the CD44
staining for both BSP and ␤3 was lost within 1 h (results not and OPN were colocalized at the cell periphery and at the base
shown), the perimembranous location of the OPN and CD44 of filopodia, as indicated by arrows in merged images from the
was still evident, while their codistribution was less clear and optical sections shown in Fig. 3. Notably, the colocalization of
the intensity of staining was reduced (Fig. 2). Notably, after 2 h OPN and CD44 in the area of the nuclei likely represents
of CHX treatment, colocalization of OPN and CD44 was lost newly synthesized proteins, which would include secreted
with the loss of cell processes, although cortical staining for OPN. In larger osteoclasts with numerous nuclei, OPN expres-
both proteins was evident even after 3 h of CHX treatment. To sion generally was much weaker and codistribution with CD44
determine whether the OPN associated with the CD44 had was more restricted (not shown).
been internalized, biotinylated porcine bone OPN was added to
the culture medium or coated on the surface of the glass slide,
and its presence inside the osteoclasts was analyzed using
FITC- or TR-labeled streptavidin. However, no fluorescent
Immunocytochemical analysis of OPN and CD44 in
mouse osteoclasts
label was detected after incubating cells with the OPN for 2 h
results not shown).
To further investigate the relationship between the CD44
and OPN, we prepared osteoclasts from the bone marrow of
(

OPN AND CD44 IN OSTEOCLASTOGENESIS
1491
FIG. 3. Double-immunofluores-
cence analysis of OPN and CD44
in rat and mouse osteoclasts. (A)
Optical sections at 2-␮m intervals
from the dorsal to ventral surface
of a rat osteoclasts immunostained
for both OPN (green; a–d) and
CD44 (red; f–h) of a rat osteoclast
examined by confocal microscopy.
The merged images (i–l) reveal
perimembranous
arrows) of the OPN and CD44
yellow) in the podosomes, which
colocalization
(
(
are shown in a phase contrast im-
age of the multinucleated oste-
oclast (g) bar ϭ 20 ␮m. (B) Single
images of WT mouse osteoclasts
immunostained for both OPN (red)
and CD44 (green) revealed by con-
focal microscopy. Staining of WT
(
a–c), OPN-null (d–f) and CD44-
null (g–i) cells are shown for OPN
upper) and CD44 (middle) with
(
the lower panels showing the
merged images. Colocalization of
OPN and CD44 is evident in a per-
imembranous location (arrows) in
the merged images of the WT cells.
In contrast, the OPN- and CD44-
null cells do not stain for OPN or
CD44, respectively, and the stain-
ing for CD44 and OPN essentially
is restricted to a perinuclear loca-
tion, with no appreciable perimem-
branous staining. Osteoclasts were
fixed, permeabilized, and double-
stained for OPN and CD44 using
combinations of mouse anti-rat
OPN monoclonal antibody (MAb)
and biotin-labeled mouse anti-
rat CD44 antibodies with Alexa-
Fluor 488-labeled second anti-
bodies and TR-streptavidin,
respectively, for the rat oste-
oclasts. For the mouse osteoclasts
a combination of LF 123 and
biotin-labeled rat anti-mouse
CD44 antibodies were used with
TR-labeled second antibodies
and FITC-labeled streptavidin,
respectively, bar ϭ 10 ␮m.
OPN-null, CD44-null, and WT mice and examined these perimembranous staining of CD44 was markedly reduced in
cells by immunofluorescence microscopy. As observed with the OPN-null cells and absent in the CD44-null cells, re-
the rat osteoclasts, OPN and CD44, immunostained using spectively (Fig. 3B, d–i). Further, the OPN staining (Fig.
different primary and secondary antibodies, were colocal- 3B, a and g) was clearly reduced in the CD44-null cells
ized at the cell periphery in WT cells as shown by arrows in compared with the WT cells.
merged images (Fig. 3B, a–c). The OPN-null and CD44-
null osteoclasts, which were smaller and more rounded than
the WT cells, showed only background staining for the null
Osteoclastogenesis in mouse bone marrow cells
proteins (Fig. 3B, d and h). CD44 staining of OPN-null cells
To study the generation of osteoclasts, mouse bone mar-
and OPN staining of the CD44-null cells was limited to the row monocytes were cultured in the presence of CSF-1 and
nuclear area and vesicular structures in the cytoplasm. The RANKL for 5 days. Morphologically, the cells derived from

1
492
SUZUKI ET AL.
motes cell process formation in the WT cells. Moreover,
when OPN-null osteoclasts treated with TNF-␣ were
stained for actin filaments, cortical actin ring structures,
typically seen in multinucleated osteoclasts, were much
smaller than those seen in the WT cells, which extended
long pseudopodia toward adjacent cells (Fig. 4, f and g).
ϩ
The small size of the TRAP cells from OPN- and CD44-
null mice appeared to result from reduced cell fusion be-
cause the average number of nuclei in the TRAP-stained
osteoclasts from OPN-null and CD44-null mice was signif-
icantly (p Ͻ 0.001) lower than in WT osteoclasts (Fig. 5A).
Notably, coating the chamber slides with porcine or rat bone
OPN or bovine milk OPN did not improve cell fusion in
either the OPN-null or the CD44-null cells, and the exoge-
nous OPN did not appear to influence cell spreading or the
formation of pseudopodia. Similarly, BSP from porcine
bone had no effect on osteoclastogenesis (results not
shown). When WT marrow cells were cultured in the pres-
ence of TNF-␣, the formation of multinuclear cells with
characteristic podosomes and cortical actin ring structures
was increased markedly. However, the effects of TNF-␣ on
the OPN-null cells were more modest, as indicated by the
analysis of nuclei number in osteoclasts with actin ring
structures (Fig. 5B).
Functional analysis of mouse bone marrow osteoclasts
Because reduced fusion in the OPN-null and CD44-null
cells could, in part, result from impaired migration, which
also has been observed in embryonic fibroblasts from the
(
29)
FIG. 4. Morphological characteristics of osteoclasts generated from OPN-null mice,
the ability of mouse osteoclasts to mi-
WT, OPN-, and CD44-null mice. Comparison of osteoclasts prepared
grate in response to a chemotactic gradient of CSF-1 was
determined using a modified Boyden chamber system. Os-
teoclasts were plated on a BSA-coated polycarbonate mem-
brane of an insert placed in a culture well containing the
same medium and serum but with 1 ng/ml of CSF-1 added
to the lower chamber. Although 47.20 Ϯ 11.56% (mean Ϯ
SD) of the WT cells migrated onto the lower membrane
from the marrow of WT (control) and OPN-null and CD44-null mice
and treated with TNF-␣ revealed a reduced formation of cell processes
and pseudopodia in the cells from the OPN-null mice, as observed by
phase-contrast microscopy (a and b) and in both OPN-null and CD44-
null cells after TRAP staining (c–e). Although the osteoclasts gener-
ated from the OPN-null and CD44-null cells stained as strongly for
TRAP as the WT cells, they were smaller with fewer nuclei (Fig. 6).
The smaller size of the OPN-null osteoclasts, even after treatment with after 1 h only 18.64 Ϯ 8.77% of the OPN-null osteoclasts
TNF-␣, was reflected also in much smaller cortical actin rings, shown and 23.50 Ϯ 9.08% of the CD44-null osteoclasts, despite
by staining with phalloidin-rhodamine (f and g). The arrows show the their smaller size, had migrated over the same time period
actin-ring structure, which encompassed more than 10 nuclei in the WT
(
Figs. 6A and 6B). Because OPN has been reported to
cells and two to three nuclei in the OPN-null cells, as determined by
nuclear staining (not shown). The osteoclasts were generated from
mouse bone marrow cells grown in the presence of CSF-1 (20 ng/ml)
and RANKL (60 ng/ml) for 3 days (a and b) or 6 days (c–g).
stimulate cell migration of osteoclasts through activation of
the ␣v␤3-integrin, we coated the upper or lower surfaces
of the polycarbonate membrane with porcine or rat bone
OPN. However, neither the OPN nor BSP had a significant
effect on the migration of either the WT or OPN-null cells
(37)
(
Fig. 6B). Because the pore size restricted the migration of
the OPN-null and CD44-null mice were smaller, did not the large multinucleated cells, the migrant cells were rela-
ϩ
appear to spread as well as the WT cells, and extended tively enriched in TRAP prefusion osteoclasts and smaller
fewer cell processes, as shown by phase-contrast micros- multinucleated osteoclasts.
copy (Fig. 4, a and b) and in cells stained for TRAP (Fig. 4,
To determine whether the defective migration and cell
c–e). Notably, almost all of the mononuclear cells also fusion in the OPN- and CD44-null osteoclasts affected their
stained for cathepsin K (results not shown), using rabbit functional ability to resorb bone, prefusion osteoclasts were
anti-mouse cathepsin K antibodies (kindly provided by Dr. grown for 48 h in the presence of CSF-1 and RANKL on
J. Mort, McGill University, Montreal, Canada), indicating dentine slices. Extensive resorption pits of variable size
that these are prefusion or nonfused osteoclasts. To empha- were formed by the WT cells, whereas the number, size, and
size the differences between WT and null cells (Fig. 4, a and depth of the resorption pits, measured by backscattered
b), the cells were incubated with TNF-␣ (50 ng/ml), which electron microscopy, were reduced (WT, 0.0797 Ϯ 0.011;
(36)
stimulates RANKL-induced osteoclastogenesis
and pro- OPNϪ/Ϫ, 0.0093 Ϯ 0.0016; CD44Ϫ/Ϫ, 0.0138 Ϯ 0.0020

OPN AND CD44 IN OSTEOCLASTOGENESIS
1493
FIG. 6. Cell migration of osteoclasts from WT, OPN-null, and
CD44-null mice. Using a modified Boyden chamber, the proportion of
ϩ
TRAP cells able to migrate in response to CSF-1 (1 ng/ml) was
determined. (A) After a 2-h migration period, the cells from WT,
OPN-null (OPNϪ/Ϫ), and CD44-null (CD44Ϫ/Ϫ) on the upper (non-
migrant) and lower (migrant) surfaces of the membrane were stained
for TRAP. (B) Quantitation of the TRAP-stained cells (left side).
Significantly fewer TRAP-stained cells obtained from the OPN-null
(
0
18.64 Ϯ 8.77%; p Ͼ 0.001) and CD44-null (23.50 Ϯ 9.08%; p Ͼ
.001) mice migrated through the 8-␮m pores compared with the WT
FIG. 5. Analysis of nuclei in osteoclasts from WT, OPN-null, and
CD44-null mice. Osteoclasts were prepared from bone marrow cells
cells (47.20 Ϯ 11.56%) after 2 h, despite their smaller size. Coating the
lower or upper surface of the polycarbonate filter with OPN or BSP did
not affect the migration of either the WT or the OPN-null cells when
compared with BSA coated on both sides (B, right side).
grown in the presence of CSF-1 (20 ng/ml) and RANKL (60 ng/ml) for
ϩ
8
days. (A) The cells were stained and the number of nuclei in TRAP
cells was determined. Although most TRAP-stained cells contained
single nuclei, significantly more (p Ͻ 0.001) WT cells (white bars)
contained multiple nuclei (average, two nuclei per cell); almost all of
the OPN-null (OPNϪ/Ϫ, gray bars) and CD44-null osteoclasts
when the cells were treated with TNF-␣ (Figs. 7B and 7C).
In the OPN-null osteoclast cultures, the resorbed areas ap-
peared gray rather than white after Von Kossa staining, as
seen in the WT cultures, indicating that the mineral was
removed incompletely. Although coating the mineral sur-
face with OPN, but not BSP, increased resorption by the
WT and the OPN-null cells, the resorption depth achieved
by the OPN-null cells was still less than was observed in the
WT cells, as indicated by the gray color of the resorption
pits excavated by the OPN-null cells. Coating with OPN
also enhanced the TNF-␣ affects, particularly in the OPN-
null cells (Fig. 7B).
(
CD44Ϫ/Ϫ, black bars) contained single nuclei. (B) To increase cell
fusion, cells from WT and OPN-null mice were grown in the additional
presence of TNF-␣ and the number of nuclei was analyzed in cells
displaying cortical actin ring structures when stained with phalloidin-
rhodamine. Although the nuclei number was increased markedly in the
WT cells (white bars), only a modest increase was seen in OPN-null
cells (OPNϪ/Ϫ, hatched bars).
2
2
mm /4 mm dentine slice) when comparable numbers of
osteoclasts, generated from the OPN-null and CD44-null
mice, were analyzed (Fig. 7A). To evaluate the effects of
protein coating on resorption, osteoclasts were first cultured
for 48 h on uncoated Osteologic slides, on which WT
osteoclasts resorbed larger areas of mineral than the OPN-
DISCUSSION
In this study, OPN has been shown to codistribute with
null osteoclasts; these differences being further emphasized CD44 inside motile rat and mouse osteoclasts, with the

1
494
SUZUKI ET AL.
colocalization of the proteins being especially prominent at
the periphery of cell processes extended by migrating cells.
That the association between the intracellular OPN and
CD44 may have functional significance is suggested by an
impairment of cell process formation, cell spreading, mi-
gration, and cell fusion during osteoclastogenesis of both
OPN- and CD44-null cells. Although resorption was also
impaired in OPN- and CD44-null osteoclasts, this could be
partially rescued in the OPN-null cells by coating the sub-
stratum with exogenous OPN, consistent with the require-
(21)
ment of OPN for mediating osteoclast attachment.
In recent studies, we have identified iOPN, colocalizing
with the CD44 receptor and ERM proteins, which appears to
be required for CD44-mediated hyaluronan binding and for
(29,38)
normal cell migration in fibroblasts.
In osteoclasts,
CD44 has been localized to the basolateral plasma mem-
branes with significantly less immunoreactivity in the bone-
(4)
contact area. Electron microscopy has been used to show
that the CD44 is present in microvilli of the basolateral
plasma membrane and colocalizes with moesin but not
(
39)
radixin or ezrin.
Notably, CD44 is enriched also in cell
(5)
(40)
processes of osteoblasts and especially in osteocytes in
which CD44 is a phenotypic marker. Here, we show that
OPN in osteoclasts codistributes with CD44 and with both
␤
1- and ␤3-receptors. However, in contrast to the cell-
surface receptors, immunostaining for OPN required cell
permeabilization, indicative of an intracellular location. To
differentiate between the pericellular iOPN and OPN in
secretory vesicles, we first showed that immunostaining for
BSP was entirely perinuclear with no perimembranous
staining (Fig. 1), whereas in motile osteoclasts optical sec-
tions revealed colocalization of OPN with CD44 throughout
some of the osteoclasts (Fig. 3) with more variable colocal-
ization in other cells. In addition, when protein synthesis
was blocked with CHX, staining for BSP disappeared,
whereas OPN could still be localized with CD44 within the
cells. Although the codistribution of OPN with CD44 was
subsequently lost, this coincided with the loss of cell pro-
cesses in which CD44 and OPN are enriched. The loss of
colocalization may be indicative of a more labile protein
that links the OPN and CD44.
Although the immunostaining of rat osteoclasts provided
evidence of an iOPN associated with the CD44 receptor
complex in osteoclasts, we sought to further strengthen the
significance of this observation by studying osteoclasts from
(16)
(41)
mice with targeted disruptions of the OPN
or CD44
genes. Although bones appear normal in CD44-null mice, a
reduction in mineral content and density has been observed
FIG. 7. Analysis of resorption activity by osteoclasts from WT,
OPN-null, and CD44-null mice. (A) The ability of osteoclasts from
WT, OPN-null (OPNϪ/Ϫ), and CD44-null (CD44Ϫ/Ϫ) to resorb a
mineralized substratum was first determined on dentine slices and the
formation of resorption pits examined by backscattered electron mi-
croscopy. As shown, fewer and smaller resorption pits were formed by
area in the OPN-null cultures (light gray bar) was not accompanied by
the OPN-null and CD44-null cells. (B) Osteoclastic resorption by the an increase in resorption depth, as indicated by the “gray” color of the
WT and OPN-null cells was examined also on Osteologic discs, coated
with BSA, OPN, or BSP. After 48 h, the cells were removed by soaking
in 20% sodium hypochlorite solution overnight and the discs were
stained with Von Kossa reagent to reveal areas of mineral resorption.
resorption pits in panel B. In the additional presence of TNF-␣, resorp-
tion was markedly increased in WT cells under all conditions and OPN
coating increased resorption in the OPN-null cells. Osteoclasts used in
these studies were generated from mouse bone marrow monocytes
(
C) Resorption by WT cells (white bars) was extensive in comparison grown in the presence of CSF-1 (20 ng/ml) and RANKL (60 ng/ml) for
to the OPN-null cells (dark gray bars) on all surfaces (p Ͼ 0.001). 3 days. The cells were then trypsinized and replated onto the dentine
Coating the surface with OPN increased the total resorption area in
both WT and OPN-null cultures. However, the increased resorption
slices (4 mm in diameter) or Osteologic discs and cultured in the
presence of CSF-1 and RANKL.

OPN AND CD44 IN OSTEOCLASTOGENESIS
1495
(Sodek and Grynpas, unpublished data, 2001). In the OPN- OPN to restore the impaired cell migration and cell fusion,
null mice, an increased capacity of spleen cells to generate observed in both OPNϪ/Ϫ and CD44Ϫ/Ϫ cells, suggests
(6)
osteoclasts has been shown, which also may reflect a that iOPN may be required for the functionality of the CD44
(7,42)
compensation for a deficiency in osteoclastic resorption
receptor.
The functional relationship between CD44 and OPN,
The de- evident in osteoclasts derived from CD44- and OPN-null
associated with increased expression of the osteoprotegerin
(43)
ligand and its receptor in the OPN-null mice.
creased osteoclast activity in the OPN-null mice can be mice, also has been observed in other cell types.
attributed to the lack of extracellular OPN, which can recruit onic fibroblasts and macrophages derived from OPN-null
(29)
Embry-
(29,38,50)
osteoclast precursors through chemotactic activity mediated and CD44-null mice have impaired migration,
(
15)
(44)
by the CD44
and ␣v␤3-receptors,
and promote oste- which can be rescued by transfection of an OPN expression
(33)
(38)
oclast attachment through the ␣v␤3-receptor. Indeed, the vector.
The impaired migration of CD44-null cells is
ability of OPN to mediate osteoclast attachment would be consistent with the suggestion that the CD44 receptor may
consistent with the increased depth of resorption pits pro- be more important in cell migration than in cell attach-
(51)
duced by OPN-null osteoclasts when the mineralized sub- ment
and with the reduced metastatic potential of tumor
stratum is coated with exogenous OPN. That OPN secreted cells transfected with antisense CD44s complementary
(52)
by osteoclasts is involved in an autocrine regulation of DNA (cDNA).
osteoclast activity is indicated also by the inhibition of
osteoclast activity by OPN antisense oligonucleotides
The reduced resorption by OPN- and CD44-null cells
may relate to the impaired fusion because larger osteoclasts
(45)
and the requirement of OPN for parathyroid hormone are known to be more active in bone resorption. In contrast,
(8)
(
PTH)–induced bone resorption. However, our studies the reduced depth of resorption pits produced by the OPN-
indicate that the impaired activity of OPN-null osteoclasts null osteoclasts appears to reflect the absence of extracellu-
also may involve the iOPN and its association with CD44. lar OPN, because an increased depth of resorption could be
Thus, immunostaining for iOPN is reduced in CD44-null achieved by coating the mineralized surface with exogenous
osteoclasts, and the perimembranous location of OPN and OPN. Notably, as a ligand for ␣v␤3, OPN is particularly
CD44 is lost in CD44Ϫ/Ϫ and OPNϪ/Ϫ osteoclasts, re- effective in stimulating the production of resorptive en-
spectively (Fig. 3), suggesting that the formation of a cell zymes including membrane type 1-matrix metalloproteinase
(
53)
surface CD44 receptor complex in these cells is dependent (MT1-MMP), MMP-9, and TRAP
on the iOPN. Notably, this contrasts ERM proteins, which mote osteoclastic resorption.
and is used to pro-
Although the requirement
bind to the cytoplasmic tail of CD44 and retain a perimem- of CD44 in resorptive activity of osteoclasts has not been
(54)
(46)
branous distribution when the binding site is ablated.
reported previously, a recent study of nonresorptive oste-
The relationship between CD44 and OPN also is evident oclasts from 4 patients with malignant osteopetrosis has
in shared functional anomalies apparent in the CD44-null revealed that CD44 expression was absent in the osteoclasts
and OPN-null osteoclasts. Morphological examination of from 2 patients and expressed weakly in the other 2 patients,
osteoclastogenesis has revealed a reduced ability of OPN- whereas CD44 was expressed strongly in osteoclasts from
(55)
and CD44-null cells to form pseudopodia and cell pro- control patients.
In addition, the osteoclasts from the
cesses, in which CD44 and iOPN are characteristically osteopetrotic patients appeared to lack the dendritic pro-
concentrated. This also may form the basis for detrimental cesses observed in normal osteoclasts.
effects on subsequent cell fusion and impaired mobility of
In summary, these studies have shown that an intracellu-
osteoclasts because it has been shown that the ␣v␤3- lar form of OPN exists in osteoclasts and osteoclast precur-
disintegrin echistatin also blocks osteoclast migration and sors, which codistributes with CD44 in cell processes. Al-
(
32,47)
fusion in vitro,
the ␣v␤3 being important for osteoclast though a physical relationship between iOPN and CD44 or
attachment and spreading. Although the formation of its receptor components has not been established in these
multinucleated osteoclasts does not appear to be adversely cells, similarities in functional defects observed in oste-
(48)
affected by the lack of the ␤3-integrin in vivo,
it is oclasts derived from OPNϪ/Ϫ and CD44Ϫ/Ϫ mice indi-
conceivable that a compensatory mechanism may mask the cate that interaction between the iOPN and the CD44 re-
role of the ␣v␤3-integrin. In this regard, no obvious effects ceptor complex may be required for the formation of cell
of CD44 deficiency are evident in the monocyte chemotac- processes and for cell normal cell migration. These defi-
tic protein (MCP)–induced accumulation of peritoneal mac- ciencies are reflected in reduced fusion of osteoclastic pre-
rophages in CD44-null mice, despite an 80% reduction in cursors and impaired resorptive activity of osteoclasts. The
their migration toward this cytokine in vitro (our unpub- absence of OPN resulting in impaired function suggests that
lished observations, 2001). Although the mechanism of cell it serves a regulatory rather than a structural role inside the
fusion in osteoclasts is poorly understood, impaired fusion cell.
of osteoclast precursors from CD44Ϫ/Ϫ mice is consistent
with the studies showing that formation of osteoclasts can
be inhibited by an anti-CD44 monoclonal antibody, but not
ACKNOWLEDGMENTS
in a monomeric form, indicating that cross-linking of the
(49)
receptor is required for inhibition. Although it is difficult
We thank Dr. Tak W. Mak (Amgen Institute, Toronto,
to discriminate between the effects of secreted extracellular Canada) for providing the CD44-null mice for these studies.
OPN and iOPN on osteoclastogenesis by bone marrow cells The development of the OPN-deficient mice at Rutgers
derived from OPNϪ/Ϫ mice, the inability of exogenous University was supported by the U.S. NIH. These studies

1
496
SUZUKI ET AL.
were supported by grants from the Canadian Institutes of
Health Research (MOP-36333) and the Canadian Arthritis
Network.
tein, osteopontin) synthesized by normal and transformed rat
bone cell populations: Regulation by TGF-beta. Biochem Bio-
phys Res Commun 162:1453–1459.
1
8. Singh K, DeVouge MW, Mukherjee BB 1990 Physiological
properties and differential glycosylation of phosphorylated and
nonphosphorylated forms of osteopontin secreted by normal
rat kidney cells. J Biol Chem 265:18696–18701.
9. Sodek J, Chen J, Nagata T, Kasugai S, Todescan R, Li IW,
Kim RH 1995 Regulation of osteopontin expression in osteo-
blasts. Ann NY Acad Sci 760:223–241.
0. Goldberg HA, Hunter GK 1995 The inhibitory activity of
osteopontin on hydroxyapatite formation in vitro. Ann NY
Acad Sci 760:305–308.
1. Ek-Rylander B, Flores M, Wendel M, Heinegard D, Anders-
son G 1994 Dephosphorylation of osteopontin and bone sia-
loprotein by osteoclastic tartrate-resistant acid phosphatase.
Modulation of osteoclast adhesion in vitro. J Biol Chem 269:
REFERENCES
1
2
2
1
. Yamate T, Mocharla H, Taguchi Y, Igietseme JU, Manolagas
SC, Abe E 1997 Osteopontin expression by osteoclast and
osteoblast progenitors in the murine bone marrow: Demonstra-
tion of its requirement for osteoclastogenesis and its increase
after ovariectomy. Endocrinology 138:3047–3055.
2
3
4
. Krause SW, Rehli M, Kreutz M, Schwarzfischer L, Paulauskis
JD, Andreesen R 1996 Differential screening identifies genetic
markers of monocyte to macrophage maturation. J Leukoc
Biol 60:540–545.
. Atkins K, Berry JE, Zhang WZ, Harris JF, Chambers AF,
Simpson RU, Somerman MJ 1998 Coordinate expression of
OPN and associated receptors during monocyte/macrophage
differentiation of HL-60 cells. J Cell Physiol 175:229–237.
. Nakamura H, Kenmotsu S, Sakai H, Ozawa H 1995 Localiza-
tion of CD44, the hyaluronate receptor, on the plasma mem-
brane of osteocytes and osteoclasts in rat tibiae. Cell Tissue
Res 280:225–233.
1
4853–14856.
2. Aruffo A, Stamenkovic I, Melnick M, Underhill CB, Seed B
990 CD44 is the principal cell surface receptor for hyaluro-
2
2
2
2
2
2
1
nate. Cell 6:1303–1313.
3. Miyake K, Underhill CB, Lesley J, Kincade PW 1990 Hyal-
uronate can function as a cell adhesion molecule and CD44
participates in hyaluronate recognition. J Exp Med 172:69–75.
4. Tolg C, Hofmann M, Herrlich P, Ponta H 1993 Splicing choice
from ten variant exons establishes CD44 variability. Nucleic
Acids Res 21:1225–1229.
5. Naor D, Sionov RV, Ish-Shalom D 1997 CD44: Structure,
function, and association with the malignant process. Adv
Cancer Res 71:241–319.
6. Tsukita S, Yonemura S, Tsukita S 1997 ERM proteins: Head-
to-tail regulation of actin-plasma membrane interaction.
Trends Biochem Sci 22:53–58.
7. Yonemura S, Tsukita S, Tsukita S 1999 Direct involvement of
ezrin/radixin/moesin (ERM)-binding membrane proteins in the
organization of microvilli in collaboration with activated ERM
proteins. J Cell Biol 145:1497–1509.
5
. Merry K, Dodds R, Littlewood A, Gowen M 1993 Expression
of osteopontin mRNA by osteoclasts and osteoblasts in mod-
elling adult human bone. J Cell Sci 104:1013–1020.
. Rittling SR, Matsumoto HN, McKee MD, Nanci A, An XR,
Novick KE, Kowalski AJ, Noda M, Denhardt DT 1998 Mice
lacking osteopontin show normal development and bone struc-
ture but display altered osteoclast formation in vitro. J Bone
Miner Res 13:1101–1111.
6
7
. Yoshitake H, Rittling SR, Denhardt DT, Noda M 1999
Osteopontin-deficient mice are resistant to ovariectomy-
induced bone resorption. Proc Natl Acad Sci USA 96:8156–
6
810.
8
. Ihara H, Denhardt DT, Furuya K, Yamashita T, Muguruma Y, 28. Katagiri YU, Sleeman J, Fujii H, Herrlich P, Hotta H, Tanaka
Tsuji K, Hruska KA, Higashio K, Enomoto S, Nifuji A,
Rittling SR, Noda M 2001 Parathyroid hormone-induced bone
resorption does not occur in the absence of osteopontin. J Biol
Chem 276:13065–13071.
. Denhardt DT, Guo X 1993 Osteopontin: A protein with di-
verse functions. FASEB J 7:1475–1482.
K, Chikuma S, Yagita H, Okumura K, Murakami M, Saiki I,
Chambers AF, Uede T 1999 CD44 variants but not CD44s
cooperate with beta1-containing integrins to permit cells to
bind to osteopontin independently of arginine-glycine-aspartic
acid, thereby stimulating cell motility and chemotaxis. Cancer
Res 59:219–226.
9
1
0. Butler WT, Ridall A, McKee MD 1996 Osteopontin. In: Bilez- 29. Zohar R, Suzuki N, Suzuki K, Arora P, Glogauer M, McCul-
ikian JP, Raisz LG, Rodan GA (eds.) Principles of Bone
Biology. Academic Press, San Diego, CA, USA, pp. 167–181.
1. Giachelli CM, Lombardi D, Johnson RJ, Murry CE, Almeida
loch CAG, Sodek J 2000 Intracellular OPN is an integral
component of the CD44-ERM complex involved in cell mi-
gration. J Cell Physiol 184:118–130.
1
M 1998 Evidence for a role of osteopontin in macrophage 30. Ly IA, Mishell RI 1974 Separation of mouse spleen cells by
infiltration in response to pathological stimuli in vivo. Am J
Pathol 152:353–358.
passage through column of Sephadex G-10. J Immunol Meth-
ods 5:239–247.
1
1
2. Sodek J, Ganss B, McKee MD 2000 Osteopontin. Crit Rev 31. Zhang Q, Domenicucci C, Goldberg HA, Wrana JL, J Sodek
Oral Biol Med 11:279–303.
1990 Characterization of fetal porcine bone sialoproteins: Se-
creted phosphoprotein 1 (SPP1, osteopontin), bone sialopro-
tein, and a 23 kDa glycoprotein. Demonstration that the 23
kDa glycoprotein is derived from the carboxy-terminus of
SPP1. J Biol Chem 265:7583–7589.
3. Rittling SR, Denhardt DT 1999 Osteopontin function in pa-
thology: Lessons from osteopontin-deficient mice. Exp Neph-
rol 7:103–113.
1
4. Denhardt DT, Noda M, O’Regan AW, Pavlin D, Berman JS
2
001 Osteopontin as a means to cope with environmental 32. Nakamura I, Tanaka H, Rodan GA, Duong LT 1998 Echistatin
insults: Regulation of inflammation, tissue remodeling and cell
survival. J Clin Invest 107:1055–1062.
5. Weber GF, Ashkar S, Glimcher MJ, Cantor H 1996 Receptor-
inhibits the migration of murine prefusion osteoclasts and the
formation of multinucleated osteoclast-like cells. Endocrinol-
ogy 139:5182–5193.
1
1
ligand interaction between CD44 and osteopontin (Eta-1). 33. Duong LT, Lakkakorpi P, Nakamura I, Rodan GA 2000 Inte-
Science 271:509–512.
6. Ashkar S, Weber GF, Panoutsakopoulou V, Sanchirico ME,
grins and signaling in osteoclast function. Matrix Biol 19:97–
105.
Jansson M, Zawaideh S, Rittling SR, Denhardt DT, Glimcher 34. Duong LT, Rodan GA 1998 Integrin-mediated signaling in the
MJ, Cantor H 2000 Eta-1 (osteopontin): An early component
regulation of osteoclast adhesion and activation. Front Biosci
of type-1 (cell-mediated) immunity. Science 287:860–864.
3:D757–D768.
1
7. Kubota T, Zhang Q, Wrana JL, Ber R, Aubin JE, Butler WT, 35. Ganss B, Kim RH, Sodek J 1999 Bone sialoprotein. Crit Rev
Sodek J 1989 Multiple forms of SppI (secreted phosphopro- Oral Biol Med 10:79–98.

OPN AND CD44 IN OSTEOCLASTOGENESIS
1497
3
3
3
6. Zhang YH, Heulsmann A, Tondravi MM, Mukherjee A, Abu-
beta(3) integrin in osteoclast migration and formation of the
sealing zone. J Cell Sci 112:3985–3993.
Amer Y 2001 Tumor necrosis factor-alpha (TNF␣) stimulates
RANKL-induced osteoclastogenesis via coupling of TNF type 48. McHugh KP, Hodivala-Dilke K, Zheng M-H, Namba N, Lam
1
5
receptor and RANK signaling pathways. J Biol Chem 276:
63–568.
J, Novack D, Feng X, Ross FP, Hynes RO, Teitelbaum SL
2000 Mice lacking ␤ integrins are osteosclerotic due to dys-
3
7. Faccio R, Grano M, Colucci S, Zallone AZ, Quaranta V,
functional osteoclasts. J Clin Invest 104:433–440.
Pelletier AJ 1998 Activation of alphav beta3 integrin on hu- 49. Kania JR, Kehat-Stadler T, Kupfer SR 1997 CD44 antibodies
man osteoclast-like cells stimulates adhesion and migration in inhibit osteoclast formation. J Bone Miner Res 12:1155–1164.
response to osteopontin. Biochem Biophys Res Commun 249: 50. Sodek J, Zhu B, Huynh M-H, Brown TJ, Ringuette M 2002
5
22–555.
Novel functions of the matricellular proteins osteopontin and
osteonectin/SPARC. Connect Tissue Res (in press).
8. Zhu J, Suzuki K, Zohar R Suzuki S. Rittling SR, Denhardt DT,
Goldberg HA, McCulloch CAG, Sodek J 2000 Functional 51. Katayama Y, House CM, Udagawa N, Kazama JJ, McFarland
significance of intracellular osteopontin associated with the
CD44 receptor complex. J Bone Miner Res 15:S345.
9. Nakamura H, Ozawa H 1996 Immunolocalization of CD44
and the ERM family in bone cells of mouse tibiae. J Bone 52. Harada N, Mizoi T, Kinouchi M, Hoshi K, Ishii S, Shiiba K,
Miner Res 11:1715–1722.
0. Kato Y, Windle JJ, Koop BA, Mundy GR, Bonewald LF 1997
Establishment of an osteocyte-like cell line, MLO-Y4. J Bone
Miner Res 12:2014–2023.
RJ, Martin TJ, Findlay DM 1998 Casein kinase 2 phosphory-
lation of recombinant rat osteopontin enhances adhesion of
osteoclasts but not osteoblasts. J Cell Physiol 176:179–187.
3
4
4
Sasaki I, Matsuno S 2001 Introduction of antisense CD44S
cDNA down-regulates expression of overall CD44 isoforms
and inhibits tumor growth and metastasis in highly metastatic
colon carcinoma cells. Int J Cancer 91:67–75.
1. Schmits R, Filmus J, Gerwin N, Senaldi G, Kiefer F, Kundig 53. Uemura T, Liu YK, Kuboki Y 2000 mRNA expression of
T, Wakeham A, Shahinian A, Catzavelos C, Rak J, Furlonger
C, Zakarian A, Simard JJ, Ohashi PS, Paige CJ, Gutierrez-
Ramos JC, Mak TW 1997 CD44 regulates hematopoietic
MT1-MMP, MMP-9, cathepsin K, and TRAP in highly en-
riched osteoclasts cultured on several matrix proteins and
ivory surfaces. Biosci Biotechnol Biochem 64:1771–1773.
progenitor distribution, granuloma formation, and tumorige- 54. Chellaiah M, Kizer N, Silva M, Alvarez U, Kwiatkowski D,
nicity. Blood 90:2217–2233.
2. Noda M, Rittling SR, Mugurama Y, Asou Y, Yoshitake H,
Tsuji K, Denhardt D 1998 Osteopontin-deficient mice (OD-)
Hruska KA 2000 Gelsolin deficiency blocks podosome assem-
bly and produces increased bone mass and strength. J Cell Biol
148:665–678.
4
cells exhibit enhanced actin fibre formation and spreading 55. Flanagan AM, Sarma U, Steward CG, Vellodi A, Horton MA
while OD-osteoclasts are less efficient in removing hydroxy-
apatite coated on glass and in recruitment to ectopically im-
planted bone matrix. J Bone Miner Res 23:S220.
2000 Study of the nonresorptive phenotype of osteoclast-like
cells from patients with malignant osteopetrosis: A new ap-
proach to investigating pathogenesis. J Bone Miner Res 15:
352–360.
4
3. Yoshitake H, Rittling SR, Tsuji K, Yamashita T, Wong B,
Choi D, Denhardt D, Noda M 1998 Osteopontin-deficient mice
reveal increase in osteoclast number, enhanced TRANCE/
RANKL/ODF expression in bone but are resistant to bone
resorption that follows bone formation after bone marrow
ablation. J Bone Miner Res 23:S200.
Address reprint requests to:
J. Sodek, Ph.D.
CIHR Group in Periodontal Physiology
4
4
4
4
4. Giachelli CM, Steitz S 2000 Osteopontin: A versatile regulator
2
34, FitzGerald Building
150 College Street
of inflammation and biomineralization. Matrix Biol 19:615–
6
22.
5. Tani-Ishii N, Tsunoda A, Umemoto T 1997 Osteopontin anti-
sense deoxyoligonucleotides inhibit bone resorption by mouse
osteoclasts in vitro. J Periodontal Res 32:480–486.
6. Legg JW, Isacke CM 1998 Identification and functional anal-
ysis of the ezrin-binding site in the hyaluronan receptor, CD44.
Curr Biol 8:705–708.
Faculty of Dentistry
University of Toronto
Toronto, Canada M5S 3E2
7. Nakamura I, Pilkington MF, Lakkakorpi PT, Lipfert L, Sims Received in original form May 27, 2001; in revised form October
SM, Dixon SJ, Rodan GA, Duong LT 1999 Role of alpha(v)- 23, 2001; accepted November 9, 2001.