Microphthalmia transcription factor regulates the expression of the novel
osteoclast factor GPNMB
Vera M. Ripolla,1,2, Nicholas A. Meadowsa,⁎,1,2, Liza-Jane Raggatta, Ming K. Changa,
Allison R. Pettita, Alan I. Cassadya, David A. Humeb
aInstitute for Molecular Biosciences, Co-operative Research Centre for Chronic Inflammatory Diseases,
The University of Queensland, St. Lucia, QLD 4072, Australia
bThe Roslin Institute, University of Edinburgh, Roslin, Edinburgh, EH25 9PS Scotland, United Kingdom
Received 14 November 2007; received in revised form 13 January 2008; accepted 15 January 2008
Available online 26 January 2008
Received by A.J. van Wijnen
Gpnmb is a macrophage-enriched gene that has also been shown to be expressed in osteoblasts. Here, we have shown gpnmb to be highly induced in
maturing murine osteoclasts. Microarray expression profile analysis identified gpnmb as a potential target of MITF in RAW264.7 cells, subclone C4
(RAW/C4), that overexpress this transcription factor. Electrophoretic mobility shift assays identified a MITF-binding site (M-box) in the gpnmb
promoter that is conserved in different mammalian species. Anti-MITF antibody supershifted the DNA–MITF complex for the promoter site while
MITF binding was abolished by mutation of this site. The gpnmb promoter was transactivated by co-expression of MITF in reporter gene assays while
mutation of the gpnmb M-box prevented MITF transactivation. The induction of gpnmb expression during osteoclastogenesis was shown to exhibit
similar kinetics to the known MITF targets, acp5 and clcn7. GPNMB expressed in RAW/C4 cells exhibited distinct subcellular distribution at different
LAMP-2, suggesting that GPNMB resides in the endocytic pathway of mature macrophages and is possibly targeted to the plasma membrane of bone-
resorbing osteoclasts. The inclusion of gpnmb in the MITF regulon suggests a role for GPNMB in mature osteoclast function.
Crown Copyright © 2008 Published by Elsevier B.V. All rights reserved.
Keywords: Bone; Resorption; Regulation; Osteoblast
Osteoclasts are large multinucleated cells that are derived
from the haematopoietic myeloid/monocyte lineage and are
specialised for bone resorption. Normal bone remodelling
requires tight coupling between the activity of osteoclasts and
the primary bone-synthesising cells, osteoblasts. Receptor
activator of NF-κB ligand (RANKL) and macrophage-colony
stimulating factor (CSF-1) are osteoblast products that initiate
osteoclastogenesis and functional maturation of bone-resorbing
activity (Kodama et al., 1991; Suda et al., 1999; Yasuda et al.,
1999; Boyle et al., 2003). The use of transgenic gene knockout
mice has elucidated the role of several transcription factors in
Available online at www.sciencedirect.com
Gene 413 (2008) 32–41
Abbreviations: acp5, tartrate-resistant acid phosphatase; bp, base pairs; CAGE,
stimulating factor 1; csf1r, colony stimulating factor 1 receptor; ctsk, cathepsin K;
cDNA, DNA complementary to RNA; G418, geneticin; hprt, hypoxanthine-
IL, interleukin; MEM, Minimal Essential Medium; MITF, microphthalmia
transcription factor; OCL, osteoclast; ostm1, osteopetrosis-associated transmem-
brane protein 1; NFAT, nuclear factor of activated Tcells; PBS, phosphate buffered
saline; RANKL, receptor activator of NF-κB ligand; qPCR, quantitative real time
polymerase chain reaction; RAW/C4, RAW264.7 cell line subclone C4.
⁎Corresponding author. Mammalian Genetics Unit, Medical Research
Council, Harwell, Oxfordshire OX11 0RD, England, United Kingdom.
Tel.: +44 1235 841246.
E-mail address: N.Meadows@har.mrc.ac.uk (N.A. Meadows).
1Now affiliated with Medical Research Council, Mammalian Genetics Unit,
Harwell, Oxon OX11 0RD.
2These authors contributed equally to this study.
0378-1119/$ - see front matter. Crown Copyright © 2008 Published by Elsevier B.V. All rights reserved.
regulating osteoclast differentiation and function. Null muta-
tions in PU.1 (Tondravi et al., 1997), c-Fos (Wang et al., 1992)
and NF-κB p50/p52 (Franzoso et al., 1997; Iotsova et al., 1997;
Xing et al., 2002) all result in osteopetrosis, which is defined
by a failure of bone resorption. Amongst these regulators,
the microphthalmia transcription factor (MITF) controls
the expression of genes associated with late stage osteoclast
MITF has been shown to regulate the expression of tartrate-
resistant acid phosphatase (acp5) (Luchin et al., 2000),
cathepsin K (ctsk) (Motyckova et al., 2001), chloride channel
7 (clcn7) and osteopetrosis-associated transmembrane protein 1
(ostm1) (Meadows et al., 2007) in osteoclasts. All of these genes
are required for osteoclast function and bone resorption.
Together with the Ets transcription factor PU.1, MITF acts
downstream of CSF-1 and RANKL to activate target genes.
This complex is also involved with recruiting another key
osteoclast transcription factor NFATc1, to maintain target gene
expression in differentiated cells (Sharma et al., 2007).
Similarly the zinc finger protein Eos, an Ikaros family member,
forms a complex with MITF and PU.1 to repress target genes
(Hu et al., 2007). MITF is a member of the basic helix-loop-
helix-leucine zipper (b-HLH-ZIP) family of transcription
factors and together with TFEB, TFEC and TFE3, MITF is
part of the MiT sub-group of the b-HLH-ZIP family. All of the
MiT factors are expressed in cells of the mononuclear phago-
cyte lineage with TFE3 and TFEB being widely expressed in
other cell types but TFEC being restricted to myeloid cells
(Rehli et al., 1999). We have employed the identification of
novel genes regulated by factors like MITF to elucidate the
mechanism of osteoclast function.
One gene that was identified was gpnmb, (also referred to as
dc.hil and osteoactivin) which encodes for a transmembrane
glycoprotein, first described in human melanoma cell lines
(Weterman et al., 1995). Murine GPNMB was identified as a
candidate adhesion molecule in myeloid dendritic cells
(Shikano et al., 2001). The rat GPNMB orthologue, termed
osteoactivin, has been attributed a role in osteoblast differ-
entiation and function (Selim et al., 2003; Abdelmagid et al.,
2007) while murine osteoactivin was identified in primary
calvarial osteoblasts cultures (Bachner et al., 2002). Since
GPNMB also shares significant sequence homology to
melanosomal proteins including quail neuroretina clone 71
(QNR-71) (Turque et al., 1996), it was proposed that GPNMB
may have a role in melanin biosynthesis and the development
Correlated with this, mutation of the gpnmb gene has been
found to be responsible for iris pigment dispersion disorder in
DBA/2J mice (Anderson et al., 2002). The pathogenesis of the
pigmentary glaucoma inDBA/2J mice has also been associated
to immune dysfunction and chronic inflammatory responses
involving IL-18 (Mo et al., 2003; Libby et al., 2005; Zhou
et al., 2005). In spite of these observations the mechanism of
GPNMB function and its role in different cell types remains
We have previously shown that gpnmb expression is highly
macrophage-enriched, elevated in inflammatory macrophages
and functions as a feedback repressor of inflammation (Ripoll
et al., 2007). In the current study, we have identified gpnmb as a
target of MITF and provide evidence for a possible function in
2. Materials and methods
2.1. General reagents
Recombinant human RANKL (Peprotech), human colony-
pGL2-Basic (Promega). Mouse monoclonal anti-MITF antibody
(NeoMarkers), rat anti-mouse MAC-3 monoclonal antibody
(BD Pharmingen), mouse anti-V5 tag monoclonal antibody
(Invitrogen). C57BL/6 mice were obtained from a specific
pathogen free (SPF) colony at the Animal Breeding House
accordance with local animal ethics guidelines.
2.2. Cell lines and cell culture
RAW/C4 cells: a subclone of the RAW264.7 macrophage-
like cell line (ATCC) (Cassady et al., 2003). RAW/C4 cells were
cultured in Dulbecco's modified Eagle medium (DMEM)
(Invitrogen) containing 5% heat-inactivated fetal calf serum
(FCS) (Biowhittaker). For osteoclast differentiation, 1 × 104
RAW/C4 or bone marrow cells were cultured on 0.1% gelatin
pre-coated plates for 7 days in the above medium supplemented
with 50 μg/mL ascorbic acid (Sigma), 40 ng/mL RANKL and
104units/mL CSF-1. Medium was changed every 2–3 days.
Osteoclasts were observed after 5–7 days in culture. Primary
osteoblasts were isolated from the calvarial of 2-day-old
neonatal mice. Calvarial was dissected, rinsed with PBS and
digested six times for 10 min at 37 °C in Hanks solution
containing 0.25 U/mL collagenase D (Roche) and 2.2 U/mL
dispase (Invitrogen). 6.9 × 104cells were cultured in Minimal
Essential Medium (MEM, Invitrogen) containing 10% heat-
inactivated FBS (Invitrogen). Cells were cultured at 37 °C and
5% CO2. Cells were differentiated in BGJb medium (Invitro-
gen) supplemented with 50 μg/mL ascorbic acid (Sigma) and 10
β-glycerolphosphate (Sigma) from day 7. MC3T3-E1-S14 cells
were maintained in MEM with 10% heat-inactivated FBS. After
confluence (day 2) the cells were cultured in differentiation
medium, with medium changes every 2–3 days.
2.3. Enrichment of primary osteoblasts and depletion of
macrophages from calvarial preparations
Osteoblasts were purified from freshly digested primary
preparations using the mouse lineage cell depletion kit
supplemented with mouse CD11b microbeads (Miltenyi,
Biotec.). Approximately 1 × 107calvarial digested cells were
incubated with lineage cell depletion antibody cocktail
supplemented with CD11b beads. Cells were sorted using a
MACS separation column where lineage positive haematopoie-
tic cells were retained and enriched osteoblasts collected. Cells
were counted and seeded for differentiation experiments.
33 V.M. Ripoll et al. / Gene 413 (2008) 32–41
2.4. Generation of RAW/C4 cell lines stably overexpressing
Mouse gpnmb was PCR amplified from cDNA (forward
primer: 5′-TCGGAGTCAGCATGGAAAGT-3′; reverse pri-
mer: 5′-GAGTGTCCTTGGCTTGTCCT-3′) and cloned into
the pEF6/V5-His TOPO TAvector (Invitrogen). RAW/C4 cells
before transfection. RAW/C4 cells (5 × 106cells/transfection)
were electroporated at0.28kVand 1000 μFusing a Gene-Pulser
(Bio-Rad). Cells were co-transfected with the selection plasmid,
pNT-Neo, at a 3:1 ratio, plated into tissue culture plates and
grown for 48 h before selection with G418 (Invitrogen) at 450
μg/mL. The stable transfectant cell lines, RAW/C4-gpnmb-
G418 selective agent during routine culture.
2.5. RNA isolation and quantitative PCR
Total cellular RNAwas extracted and purified to make cDNA.
The PCR amplicon was quantitated using SYBR Green (Applied
Biosystem) using an ABI Prism 7000 sequence detection system
(Applied Biosystem). Sample amplicon levels during the
linear phase of amplification were normalised against hypox-
anthine phosphoribosyl transferase (hprt) control PCR product.
Assays were performed in triplicate and the means ± SD were
determined. The specific primers used for qPCR were as
follows: mouse TRAP1C (GB: NM_007388) forward primer
(5′-ACCTGTGCTTCCTCCAGGAT-3′), reverse primer (5′-
TCTCAGGGTGGGAGTGGG-3′); mouse Clcn7 (GB:
NM_011930) forward primer (5′-GACTGGCTGTGGGAAAG-
GAA-3′), reverse primer (5′-TCTCGCTTGAGTGATGTT-
GACC-3′); mouse Gpnmb (GB: NM_053110) forward primer
(5′-AGCACAACCAATTACGTGGC-3′), reverse primer (5′-
CTTCCCAGGAGTCCTTCCA-3′); mouse csf1r (GB:
NM_007779) forward primer (5′-CCACCATCCACTTG-
TATGTCAAAGAT-3′), reverse primer (5′-CTCAACCACTGT-
CACCTCCTGT-3′); mouse Runx2 (GB: NM_009820.3)
forward primer (5′-ACAAACAACCACAGAACCACAAGT-
3′), reverse primer (5′-GTCTCGGTGGCTGGTAGTGA-3′).
Relative expression levels were calculated using experimentally
determined primer efficiency and the ΔCTmethod (Pfaffl, 2001).
2.6. Analysis of promoter activity
The murine promoter region forgpnmb was amplified by PCR
from mouse genomic DNA (forward primer: 5′-GGTAAA-
GAATGTCAGAACAGGA-3′; reverse primer: 5′-GTCAG-
GGCTTGACTCTGACT-3′) and cloned into a promoter-less
luciferase reporter vector (pGL2-Basic, Promega). Site directed
was mutated to CTCGAG (5′-GGAGATCtcGaGATGTTT-
Sequence analysis of these promoter constructs was used to
electroporation (280 V/1000 μF, Gene-Pulser, Bio-Rad). 5 × 106
assay were performed according to Meadows et al. (2007).
Luciferase activity was normalised to the total protein concentra-
tion of the cell lysate to give relative light units (RLU). The
standard error of the mean was calculated within, and between
2.7. Electrophoretic mobility gel shift assay (EMSA)
Nuclear extracts were prepared according to the protocol
described by Meadows et al. (2007). Complementary oligonu-
cleotides used for double-stranded probe preparation were as
follows with mutated residues shown in lower case: Gpnmb
oligo (5′-CTGCTTAAA ACATCACATGATCTCCC-3′) and
Oligonucleotides were 5′ end-labelled with γ-32P-ATP and
T4 polynucleotide kinase for 30 min at 37 °C. Nuclear extract
proteins were bound to the DNA probe in a 10 μL reaction
containing 20 mM HEPES pH 7.9, 500 nM DTT, 2 mM EDTA,
40 mM KCl, 12% glycerol, 1 μg salmon sperm DNA, 0.04 pmol
purified probe and 2 μg nuclear extract. Reactions in which
excess of unlabelled competitor probe (0.4, 2 or 4 pmol, respec-
tively). The supershift reaction included the anti-Mitf antibody
(NeoMarkers). The Tris–glycine–EDTA gel system was used for
RAW/C4 cells overexpressing GPNMB were grown on
coverslips and differentiated in the presence of RANKL and
CSF-1 for 7 days, as described above. At indicated time points,
cells were washed twice with PBS and fixed with 4% para-
formaldehyde. Cells were permeabilized with 0.1% Triton X-100
and anti-MAC-3 for 1 h. Primary antibodies were detected with
secondary Alexa 488-conjugated (for V5) and Alexa 594-con-
coverslipsusingDAKO Cytomation fluorescent mountingmedia
(Dako Corporation). Slides were photographed using a Zeiss
LSM 510 META confocal microscope (Carl Zeiss).
Microarray data that wehave published as part ofthe Novartis
symatlas project (symatlas.gnf.org) supports published evidence
that gpnmb mRNA is detectable in primary mouse calvarial
osteoblasts, and increases, as they are stimulated towards matrix
calcification in vitro. Nevertheless, such primary cultures may be
contaminated with myeloid cells, and known macrophage-
specific genes (emr1, csf1r) can be detected in the osteoblasts.
34V.M. Ripoll et al. / Gene 413 (2008) 32–41
To assess whether the osteoblasts themselves express gpnmb,
magnetic assisted cell sorting (MACS) was used to deplete
macrophages and other haematopoietic cells from the calvarial
cell preparation and enrich osteoblasts. Cell populations were
cultured over a standard 21-day time course, RNAwas harvested
Fig. 1. Gpnmb expression in macrophages and osteoclasts is dominant over gpnmb expression in osteoblasts. qPCR for gpnmb, csf1r and runx2 in primary osteoblasts
before and after cellular sorting to remove contaminating macrophages (A). While gpnmb is up-regulated over the time course in both populations, both csf1r and
gpnmb expression is markedly lower in the sorted osteoblast population. Runx2 expression is increased in the sorted population. qPCR for gpnmb in primary
osteoclasts and enriched osteoblasts (B), and in osteoclast and osteoblast cell lines over a time course of differentiation (C). Gpnmb expression in osteoclasts is at least
3-fold above the expression detected in osteoblasts in both primary and cell line derived cells. Experiments were performed in triplicate and the bars represent the
standard deviation for one experiment. Statistically significant difference (⁎pb0.05,⁎⁎pb0.01,⁎⁎⁎pb0.001) was analysed by paired t-test. Statistical significance
across a time course was performed using one way ANOVA (#pb0.05).
Fig. 2. Microarray and qPCR data for gpnmb expression in cells overexpressing MITF. Expression of gpnmb mRNA is induced by RANKL. Treatment of RAW/C4
cells for 5 days and superinduced by MITF overexpression. The qPCR was consistent with the microarray data. Vector refers to RAW/C4 cells stably transfected with
the empty pEF6 control. Statistically significant difference (⁎⁎pb0.01,⁎⁎⁎pb0.001) was analysed by paired t-test.
35 V.M. Ripoll et al. / Gene 413 (2008) 32–41
Fig. 3. Promoter annotation and Clustal Walignment of the gpnmb promoter with other mammalian species. CAGE analysis has been used to identify the transcription
start sites of gpnmb in both mouse (red bars) and human (blue bars) (A). There is a clear dominant CA initiator site (3206 tags) for mouse that is relatively conserved
with the human gpnmb transcription start site. The conserved gpnmb promoter region for mouse and human has been aligned (B). Aligned and conserved consensus
binding sites for MITF and AP-1 have been boxed. A conserved TATA-like element proximal to the transcription start sites has also been represented. Transcription
start sites have been marked by an arrow and the ATG start codon is marked by⁎. The M-box in the gpnmb promoter is conserved between at least 11 different
mammalian species (C). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
36V.M. Ripoll et al. / Gene 413 (2008) 32–41
time PCR (qPCR). The expression of gpnmb increased during
differentiation of both the unsorted and enriched osteoblast
cultures but was at least 6-fold higher in the unsorted population
compared to enriched osteoblasts (Fig. 1A). Similarly, the
expression of the macrophage gene, csf1r was induced during
unsorted osteoblasts differentiation and expressed at minimal
levels in enriched osteoblast cultures (Fig. 1A). Runx2 mRNA
levels increased in the enriched osteoblast cells over the time
course of differentiation, confirming that the sorting process
enriched for osteoblasts (Fig. 1A). These data show that
osteoblasts only express low levels of gpnmb and that the
macrophage contamination of these cultures.
The microarray data in the Novartis symatlas project also
revealed that gpnmb mRNA is expressed in differentiated
osteoclasts. Considering the lineage relationship between macro-
phages and osteoclasts we sought to compare the expression of
gpnmb in osteoclasts and osteoblasts. A time course for gpnmb
mRNA levels was performed in osteoclast and osteoblast cells
using both primary cells and cell lines. While gpnmb is up-
regulated during differentiation of both cell types, maximal
expression of gpnmb in primary osteoclasts is at least 4-fold
above the maximal expression of gpnmb in enriched primary
osteoblasts (Fig. 1B). Maximal expression of gpnmb in
osteoclast-like cells differentiated from the RAW/C4 cell line is
at least 3-fold above the maximal expression of gpnmb in the
osteoblast cell line MC3T3 (Fig. 1C). This is the first evidence
that suggests thatgpnmb expression in osteoclasts is significantly
higher than osteoblasts. In fact gpnmb could be described
predominantly as a mature myeloid cell expressed gene.
3.2. Overexpression of MITF increases basal gpnmb expression
and superinduces gpnmb during osteoclastogenesis
We showed previously that overexpression of MITF in the
precursor RAW/C4 line increases OCL number and the en-
dogenous expression of typical osteoclast markers following
treatment with RANKL and CSF-1 (Meadows et al., 2007).
Within the list of potential MITF transcriptional targets, a number
ofgenes with relatively unknownroles in osteoclast biology were
also identified, and by association, these genes are likely can-
gpnmb, which increased in expression in RAW/C4-GFP-mitf/
pEF6 cells compared to RAW/C4-GFP-pEF6 cells in the
microarray analysis (Fig. 2). This result was validated using
qPCR for gpnmb with RNA prepared from the same time
course. Given that gpnmb is also expressed in melanocytes
(Weterman et al., 1995), which require MITF for differentiation
be a direct MITF target and a contributor to osteoclast function.
3.3. The gpnmb proximal promoter region is conserved
between mouse and human
The transcription start site of gpnmb was identified from an
extensive genome-wide promoter analysis using CAGE (cap
analysis gene expression) data for both mouse and human
(Fig. 3A) (Carninci et al., 2006). In both mouse and human
there is a major peak of transcription initiation associated with a
conserved CA initiator (Carninci et al., 2006). Of the 3206 tags
mapped in mice, two thirds were derived from bone marrow and
myeloid specific tissues (primary data can be accessed at www.
macrophages.com). The promoter for gpnmb was annotated
using a Clustal Walignment of the mouse and human conserved
regions (Fig. 3B). Several features are relevant to the regulation
of gpnmb. Unlike many myeloid specific promoters, gpnmb has
a single dominant start site around 30 bp downstream of a
TATA-like element in both mouse and human. A conserved AP-
1 (TGAGTCA) site was identified in the gpnmb promoter and
the most conserved region contains a consensus M-box element
(TCACATGA) for binding of MITF. This MITF-binding site is
aligned and conserved between at least 11 different species
(Fig. 3C). The clear conservation of these elements suggests that
gpnmb has similar regulation in all mammals.
Fig. 4. EMSA and luciferase transfection assays confirm that MITF binds and transactivates the gpnmb promoter. (A) Nuclear extracts from RAW/C4 cells treated with
RANKL and CSF-1 (5 days) incubated with probes containing the M-box from the gpnmb promoter region. MITF-binding specificity was demonstrated with cold
competition with both WTand M-box mutation oligonucleotides (arrow). Incubation with an anti-MITF antibody produces a supershifted band (⁎). Selective loss of
the MITF complex by wild type, and not mutant, competitors suggests that MITF has specificity for the M-box probe. (B) Co-transfections with mitf/pEF6 and a
gpnmb promoter–reporter construct produce at least a 6-fold induction in relative luciferase activity compared to basal levels and transfections with mi/pEF6.
There was no transactivation of the reporter construct in which the M-box site was mutated. Bars represent means±standard errors of the mean (n=9).
37 V.M. Ripoll et al. / Gene 413 (2008) 32–41
3.4. MITF binds and transactivates the gpnmb promoter
To establish whether MITF directly transcriptionally reg-
ulates gpnmb expression in osteoclasts, the ability of MITF to
bind and transactivate the gpnmb promoter in RAW/C4-derived
OCLs was determined. Radiolabelled oligonucleotide probes
containing the gpnmb promoter M-box sequence were designed
for EMSA, under conditions previously optimized with the
acp5 promoter M-box (Meadows et al., 2007). Endogenous
MITF from nuclear extracts of RAW/C4 cells cultured with
RANKL and CSF-1 for 5 days was found to bind specifically to
the gpnmb probe, shown by the supershift that occurs in the
presence of a mouse anti-MITF antibody (Fig. 4A). Binding
specificity was demonstrated using cold competition assays
performed with the wild type gpnmb M-box and a mutated M-
box oligonucleotide in which the CACATG core was mutated to
Transient transfections of RAW/C4 cells with a gpnmb
promoter (− 250 to + 73) luciferase reporter construct and a
MITF expression plasmid were performed to assess the activity
and responsiveness of the proximal gpnmb promoter to MITF.
The gpnmb promoter was transactivated by the wild type mitf
expression plasmid (Fig. 4B) and mutations of the conserved
M-box not only ablated this transactivation but also reduced
basal promoter activity. These data indicate that MITF regulates
the gpnmb promoter in vitro via the conserved M-box.
Consequently these experiments identify a specific MITF-
binding site within the gpnmb promoter and suggest that the
gpnmb promoter is responsive to wild type MITF promoter
3.5. Gpnmb expression is induced with similar kinetics to clcn7
and acp5 during osteoclastogenesis
The regulation of gpnmb expression during osteoclastogen-
esis of primary cells was compared to known MITF targets,
acp5 and clcn7 (Fig. 5). Over a 7-day time course with
RANKL and CSF-1, qPCR showed that gpnmb was up-
regulated with similar kinetics to clcn7 and acp5 which sug-
gests a common pattern of regulation.
3.6. RANKL induces a change in GPNMB localisation during
To characterize the role of GPNMB in osteoclasts, its
subcellular localisation in different stages of osteoclast
differentiation was examined by immunofluorescence using
confocal microscopy. RAW/C4-gpnmb-pEF6 cells overexpres-
sing GPNMB, were treated with RANKL over 7 days, and were
stained with anti-V5 antibody to detect V5-tagged GPNMB. As
we previously described in macrophages (Ripoll et al., 2007), at
day 0 and day 3 osteoclasts, GPNMB was confined to the
membrane compartments around the nuclei typical of the Golgi
network. GPNMB staining in both undifferentiated and day 3
did not coincide with that of the endosomal marker, MAC-3/
LAMP-2 as shown in Fig. 6. This indicates that GPNMB is not
present in lysosomes or late endosomal compartments at these
time points. After further stimulation with RANKL, GPNMB
cellular distribution dramatically changed from a single peri-
nuclear Golgi compartment to dot-like organelles throughout
the cytoplasm in multinuclear osteoclast-like cells. At 5 and 7
days after stimulation, MAC-3/LAMP-2 overlapped with
Fig. 5. Clcn7, acp5 and gpnmb are up-regulated with similar kinetics during
osteoclastogenesis. qPCR for clcn7, acp5 and gpnmb in primary osteoclasts
cultured with RANKL and CSF-1 over a time course of 7 days. Expression of
clcn7, acp5 and gpnmb was induced after 3 days reaching maximal expression
between 5 and 7 days. All three genes display similar kinetics of expression
during the time course. Experiments were performed in triplicate and the bars
represent the standard deviation for one experiment.
38 V.M. Ripoll et al. / Gene 413 (2008) 32–41
GPNMB, which indicated it was localised to late endosomes
and lysosomes (Fig. 6). Intracellular membrane trafficking and
endocytic pathways are essential for osteoclast function and are
regulated by factors such as RANKL (Sakai et al., 2001). Many
other proteins that are enriched in osteoclasts such as TRAP,
CTSK and CLCN7 are closely associated with these pathways
and some of them are known to be transported from late endo-
somes/lysosomes or recycling compartments to the cell peri-
phery, extracellular space and plasma membrane (ruffled
border) (Sahara, 2001; Hollberg et al., 2002; Lange et al.,
2006). Our findings suggest that GPNMB resides in the endo-
cytic pathway of mature osteoclast-like cells and is possibly
targeted to the plasma membrane or extracellular space upon
osteoclast terminal differentiation.
Genes regulated by MITF have a strong association with
osteoclast function. As late markers of osteoclast differentiation,
the direct MITF targets TRAP, Cathepsin K, CLCN7, OSTM1,
E-cadherin and OSCAR have all been implicated with either
osteoclast resorption or late stages of osteoclast maturation
(Luchin et al., 2000; Motyckova et al., 2001; Mansky et al.,
2002; So et al., 2003; Meadows et al., 2007). In light of this, the
identification of novel targets of master regulators of osteoclast
function like MITF and NFATc1 may offer further insights into
the mechanism of osteoclast function. This study has identified
GPNMB as a novel candidate for osteoclast function by direct
determination of MITF regulation of gpnmb transcription.
Fig. 6. GPNMB localisation in RAW/C4 cells treated with RANKL over 7 days. GPNMB/V5 (green) localisation shifts from the Golgi apparatus to a diffuse vesicular
pattern following culture with RANKL over a time course of 7 days. MAC-3 (red) has been included as a lysosomal and late endosome marker. The 2 proteins occupy
different cellular compartments at days 0 and 3. By days 5 and 7 the 2 proteins co-localise to similar cellular compartments (yellow). (For interpretation of the
references to colour in this figure legend, the reader is referred to the web version of this article.)
39 V.M. Ripoll et al. / Gene 413 (2008) 32–41
In a previous study, we showed that gpnmb expression is
highly macrophage-enriched and further elevated in inflamma-
tory macrophages (Ripoll et al., 2007). In bone, gpnmb has been
observed to be inducibly expressed specifically in primary
osteoblast differentiation cultures (Safadi et al., 2001; Bachner
et al., 2002; Abdelmagid et al., 2007). Our cell-specific analyses
reveal that gpnmb is highly expressed in both macrophages and
detection of inducible expression of many macrophage-specific
genes within heterogeneous populations of primary osteoblast
cultures. The macrophage marker gene, csf1r, is one gene we
have observed to have disproportionately high expression in
RNA harvested from primary osteoblast populations. Similarly,
we find that at least some of the gpnmb expression in primary
osteoblast cultures can be attributed to macrophage contamina-
tion,andthat macrophagesandosteoclasts expressthemRNAat
from Nomiyama et al. (2005) which also identified gpnmb as a
gene differentially expressed in osteoclast-like cells. This find-
ing is not surprising as macrophages and osteoclasts share the
same lineage and a requirement for CSF-1.
Gpnmb has been identified in low-metastatic melanoma cell
lines and its expression pattern is reminiscent of migrating
neural cell populations identified with c-kit and trp-2 markers.
MITF has been shown to regulate the expression of both c-kit
and trp-2 (Tsujimura et al., 1996; Ferguson and Kidson, 1997)
in melanocytes. MITF is a master regulator of melanocyte gene
expression and its dysregulation has been linked to skin cancer
melanomas (Garraway et al., 2005). Furthermore, in a
homology search using the protein–protein BLAST database,
the amino acid sequence of GPNMB shares identity to the
melanoma markers QNR-71 (quail) SILV (mouse) and
PMEL17 (human homologue of SILV). QNR-71, SILV and
PMEL17 have all been shown to be regulated by MITF (Turque
et al., 1996; Du et al., 2003).
Microarray and qPCR data from RAW/C4 cells that
constitutively overexpress MITF revealed that gpnmb is up-
regulated both before and during osteoclast differentiation
(Meadows et al., 2007). Examination of the gpnmb promoter
showed a strong conservation in the proximal region amongst at
least 11 different mammalian species. This suggests that a
common regulator of gpnmb is necessary for its expression in
humans as in other mammals. The presence of a conserved
consensus MITF-binding site (M-box) within this promoter
region proposes that regulation by MITF plays a critical role in
directing gpnmb expression in osteoclasts. We have shown a
clear association of MITF with the gpnmb promoter. Further-
more, the conserved gpnmb promoter region was activated by
co-transfection of MITF.
GPNMB localisation in mature osteoclasts differs from its
distribution in macrophages, suggesting a cell type specific
function. We showed that during the differentiation of multi-
nuclear osteoclast-like cells, GPNMB translocates from a
perinuclear location to lysosomes/late endosomes with MAC-
3/LAMP-2. The localisation of GPNMB to these compartments
is consistent with the presence of a predicted endosomal/
lysosomal-sorting signal located immediately after the trans-
membrane domain on the C terminal region of GPNMB
(Anderson et al., 2002), (smart.embl-heidelberg.de). Intracel-
lular membrane trafficking and endocytic pathways are
regulated by the critical osteoclastogenic factor RANKL
and are crucial for osteoclast function (Sakai et al., 2001). An
important number of osteoclast molecules such as CTSK,
CLCN7 and v-ATPase are transported from late endosomes/
lysosomes and recycling compartments to the extracellular
space and ruffled border to carry out their function (Sahara,
2001; Toyomura et al., 2003; Lange et al., 2006). There is an
exciting possibility that like these molecules, GPNMB is also
transported to the vicinity of the cell surface to undertake its
role. It is possible that as has been shown for GTP-binding RAB
proteins, the localisation of some proteins to the ruffled border
can only be observed in resorbing osteoclasts cultured on bone
slices and not in non-resorbing osteoclast cultured on glass
coverslips (Zhao et al., 2002). In addition to the structural and
regulatory relationship between GPNMB and PMEL17,
GPNMB subcellular location in osteoclasts is also reminiscent
of PMEL17. Just as GPNMB moves from a perinuclear location
in immature osteoclast-like cells to endosome-related structures,
PMEL17 is a major component of melanosomes in melano-
cytes, and traffics from a perinuclear location to the intralume-
nal vesicles of multivesicular endosomes in differentiating
melanocytes. A mutation in PMEL17 (silver) disrupts melano-
some formation (Theos et al., 2006). By extension, we predict
that GPNMB will have a function in formation of a population
of endosome-like structure involved in bone resorption and is
probably also involved in melanosome formation in melano-
cytes. Further functional studies are required to understand
GPNMB intracellular sorting.
The mutation of the gpnmb gene has been found to be
responsible for an iris pigment dispersion disorder in DBA/2J
mice (Anderson et al., 2002). Because of the melanocyte
involvement in iris pigment development and the critical role of
MITF in melanocyte biology, this phenotype supports our
observation of MITF regulation of gpnmb expression. There is
no evident bone phenotype in DBA/2J mice. This does not
necessarily imply that there is no function in osteoclasts; the
TRAP and cathepsin K knockouts have relatively mild
phenotypes, partly due to the presence of compensatory phos-
phatases and protease activities from other loci, and partly
because bone turnover is balanced by competing actions of
osteoblasts and osteoclasts. The DBA/2J mouse strain already
has a high bone mineral density, and has been used in mapping
of genetic determinants of bone mineral density (Klein et al.,
2001). It will be of interest to study the gpnmb mutation in a
different mouse background such as C57/Bl6 or BALB/c.
Abdelmagid, S.M., Barbe, M.F., Arango-Hisijara, I., Owen, T.A., Popoff, S.N.,
Safadi, F.F., 2007. Osteoactivin acts as downstream mediator of BMP-2
effects on osteoblast function. J. Cell. Physiol. 210 (1), 26–37.
Anderson, M.G., et al., 2002. Mutations in genes encoding melanosomal proteins
cause pigmentary glaucoma in DBA/2J mice. Nat. Genet. 30 (1), 81–85.
Bachner, D., Schroder, D., Gross, G., 2002. mRNA expression of the murine
glycoprotein (transmembrane) nmb (Gpnmb) gene is linked to the
40 V.M. Ripoll et al. / Gene 413 (2008) 32–41
developing retinal pigment epithelium and iris. Brain Res. Gene Expr. Download full-text
Patterns 1 (3–4), 159–165.
Boyle, W.J., Simonet, W.S., Lacey, D.L., 2003. Osteoclast differentiation and
activation. Nature 423 (6937), 337–342.
Carninci, P., et al., 2006. Genome-wide analysis of mammalian promoter
architecture and evolution. Nat. Genet. 38 (6), 626–635.
Cassady, A.I., Luchin, A., Ostrowski, M.C., Hume, D.A., 2003. Regulation of
the murine TRACP gene promoter. J. Bone Miner. Res. 18 (10), 1901–1904.
2003. MLANA/MART1 and SILV/PMEL17/GP100 are transcriptionally
regulated by MITF in melanocytes and melanoma. Am. J. Pathol. 163 (1),
Ferguson, C.A., Kidson, S.H., 1997. The regulation of tyrosinase gene
transcription. Pigment Cell Res. 10 (3), 127–138.
Franzoso, G., et al., 1997. Requirement for NF-kB in osteoclast and B-cell
development. Genes Dev. 11, 3482–3496.
Garraway, L.A., et al., 2005. Integrative genomic analyses identify MITF as
a lineage survival oncogene amplified in malignant melanoma. Nature
436 (7047), 117–122.
Ho, S.N., Hunt, H.D., Horton, R.M., Pullen, J.K., Pease, L.R., 1989. Site-
directed mutagenesis by overlap extension using the polymerase chain
reaction. Gene 77 (1), 51–59.
Hollberg, K., Hultenby, K., Hayman, A., Cox, T., Andersson, G., 2002.
Osteoclasts from mice deficient in tartrate-resistant acid phosphatase have
altered ruffled borders and disturbed intracellular vesicular transport. Exp
Cell Res. 279 (2), 227–238.
Hu, R., Sharma, S.M., Bronisz, A., Srinivasan, R., Sankar, U., Ostrowski, M.C.,
2007. Eos, MITF, and PU.1 recruit corepressors to osteoclast-specific genes
in committed myeloid progenitors. Mol Cell. Biol. 27 (11), 4018–4027.
Iotsova, V., Caamano, J., Loy, J., Yang, Y., Lewin, A., Bravo, R., 1997.
Klein, R.F., Shea, M., Gunness, M.E., Pelz, G.B., Belknap, J.K., Orwoll, E.S.,
2001. Phenotypic characterization of mice bred for high and low peak bone
mass. J. Bone Miner. Res. 16 (1), 63–71.
Kodama, H., Nose, M., Shumpei, N., Yamasaki, A., 1991. Essential role of
macrophage colony-stimulating factor in the osteoclast differentiation
supported by stromal cells. J. Exp. Med. 173, 1291–1294.
Lange, P.F., Wartosch, L., Jentsch, T.J., Fuhrmann, J.C., 2006. ClC-7 requires
Ostm1 as a beta-subunit to support bone resorption and lysosomal function.
Nature 440 (7081), 220–223.
Libby, R.T., Gould, D.B., Anderson, M.G., John, S.W., 2005. Complex genetics
of glaucoma susceptibility. Annu. Rev. Genomics Hum. Genet. 6, 15–44.
Luchin, A., et al., 2000. The microphthalmia transcription factor regulates
expression of the tartrate-resistant acid phosphatase gene during terminal
differentiation of osteoclasts. J. Bone Miner. Res. 15 (3), 451–460.
Mansky, K.C., Marfatia, K., Purdom, G.H., Luchin, A., Hume, D.A., Ostrowski,
M.C., 2002. The microphthalmia transcription factor (MITF) contains two
N-terminal domains required for transactivation of osteoclast target
promoters and rescue of mi mutant osteoclasts. J. Leukoc. Biol. 71 (2),
Meadows, N.A., Sharma, S.M., Faulkner, G.J., Ostrowski, M.C., Hume, D.A.,
Cassady, A.I., 2007. The expression of clcn7 and ostm1 in osteoclasts is
coregulated by microphthalmia transcription factor. J. Biol. Chem. 282 (3),
Mo, J.S., et al., 2003. By altering ocular immune privilege, bone marrow-
derived cells pathogenically contribute to DBA/2J pigmentary glaucoma.
J. Exp. Med. 197 (10), 1335–1344.
Motyckova, G., Weilbaecher, K.N., Horstmann, M., Rieman, D.J., Fisher, D.Z.,
Fisher, D.E., 2001. Linking osteopetrosis and pycnodysostosis: regulation of
cathepsin K expression by the microphthalmia transcription factor family.
Proc. Natl. Acad. Sci. U.S.A. 98 (10), 5798–5803.
Nomiyama, H., et al., 2005. Identification of genes differentially expressed in
osteoclast-like cells. J. Interferon Cytokine Res. 25 (4), 227–231.
Pfaffl, M.W., 2001. A new mathematical model for relative quantification in
real-time RT-PCR. Nucleic Acids Res. 29 (9), e45.
Rehli, M., Lichanska, A., Cassady, A.I., Ostrowski, M.C., Hume, D.A., 1999.
TFEC is a macrophage-restricted member of the microphthalmia-TFE
subfamily of basic helix-loop-helix leucine zipper transcription factors.
J. Immunol. 162 (3), 1559–1565.
Ripoll, V.M., Irvine, K.M., Ravasi, T., Sweet, M.J., Hume, D.A., 2007. Gpnmb
is induced in macrophages by IFN-gamma and lipopolysaccharide and acts
as a feedback regulator of proinflammatory responses. J. Immunol. 178 (10),
Safadi, F.F., Xu, J., Smock, S.L., Rico, M.C., Owen, T.A., Popoff, S.N., 2001.
Cloning and characterization of osteoactivin, a novel cDNA expressed in
osteoblasts. J. Cell. Biochem. 84 (1), 12–26.
Sahara, N., 2001. Cellular events at the onset of physiological root resorption in
rabbit deciduous teeth. Anat. Rec. 264 (4), 387–396.
Sakai, E., Miyamoto, H., Okamoto, K., Kato, Y., Yamamoto, K., Sakai, H., 2001.
Characterization of phagosomal subpopulations along endocytic routes in
osteoclasts and macrophages. J. Biochem. (Tokyo) 130 (6), 823–831.
Selim, A.A., et al., 2003. Anti-osteoactivin antibody inhibits osteoblast differentia-
tion and function in vitro. Crit. Rev. Eukaryot. Gene Expr. 13 (2–4), 265–275.
Sharma, S.M., et al., 2007. MITF and PU.1 recruit p38 MAPK and NFATc1
to target genes during osteoclast differentiation. J. Biol. Chem. 282 (21),
Shikano,S., Bonkobara, M., Zukas, P.K., Ariizumi, K., 2001. Molecular cloning
of a dendritic cell-associated transmembrane protein, DC-HIL, that
promotes RGD-dependent adhesion of endothelial cells through recognition
of heparan sulfate proteoglycans. J. Biol. Chem. 276 (11), 8125–8134.
So,H.,et al., 2003.Microphthalmia transcriptionfactor andPU.1synergistically
induce the leukocyte receptor OSCAR gene expression. J. Biol. Chem. 278
Steingrímsson, E., Copeland, N.G., Jenkins, N.A., 2004. Melanocytes and the
microphthalmia transcription factor network. Annu. Rev. Genet. 38, 365–411.
Suda, T., Takahashi, N., Udagawa, N., Jimi, E., Gillespie, M.T., Martin, T.J.,
1999. Modulation of osteoclast differentiation and function by the new
members of the tumor necrosis factor receptor and ligand families. Endocr.
Rev. 20 (3), 345–357.
Theos, A.C., et al., 2006. A lumenal domain-dependent pathway for sorting to
intralumenal vesicles of multivesicular endosomes involved in organelle
morphogenesis. Dev Cell 10 (3), 343–354.
Tondravi, M.M., et al., 1997. Osteopetrosis in mice lacking haematopoietic
transcription factor PU.1. Nature 386, 81–84.
Toyomura, T., et al., 2003. From lysosomes to the plasma membrane:
localization of vacuolar-type H+-ATPase with the a3 isoform during
osteoclast differentiation. J. Biol. Chem. 278 (24), 22023–22030.
Tsujimura, T., et al., 1996.Involvement of transcription factor encodedby the mi
locus in the expression of c-kit receptor tyrosine kinase in cultured mast cells
of mice. Blood 88 (4), 1225–1233.
Turque, N., et al., 1996. Characterization of a new melanocyte-specific gene
(QNR-71) expressed in v-myc-transformed quail neuroretina. Embo J.
15 (13), 3338–3350.
1992. Bone and haematopoietic defects in mice lacking c-fos. Nature 360,
Weterman, M.A., et al., 1995. nmb, a novel gene, is expressed in low-metastatic
human melanoma cell lines and xenografts. Int. J. Cancer 60 (1), 73–81.
Xing, L., et al., 2002. NF-kappaB p50 and p52 expression is not required for
RANK-expressing osteoclast progenitor formation but is essential for RANK-
and cytokine-mediated osteoclastogenesis. J. Bone Miner. Res. 17 (7),
Yasuda, H., et al., 1999. A novel molecular mechanism modulating osteoclast
differentiation and function. Bone 25 (1), 109–113.
Zhao, H., Ettala, O., Vaananen, H.K., 2002. Intracellular membrane trafficking
pathways in bone-resorbing osteoclasts revealed by cloning and subcellular
localization studies of small GTP-binding rab proteins. Biochem. Biophys.
Res. Commun. 293 (3), 1060–1065.
Zhou, X., Li, F., Kong, L., Tomita, H., Li, C., Cao, W., 2005. Involvement of
inflammation, degradation, and apoptosis in a mouse model of glaucoma.
J. Biol. Chem. 280 (35), 31240–31248.
41 V.M. Ripoll et al. / Gene 413 (2008) 32–41