Evidence for the role of matrix metalloproteinase-13 in
bone resorption by giant cell tumor of bone☆,☆☆
Isabella W. Y. Mak MSca,b, Eric P. Seidlitz PhDa,c, Robert W. Cowan MSca,c,
Robert E. Turcotte MD, FRCSCd, Snezana Popovic MD, PhD, MSc, FRCPCa,
William C. H. Wu PhDe, Gurmit Singh PhDa,c, Michelle Ghert MD, FRCSCa,b,⁎
aMcMaster University, Hamilton, Ontario, Canada L8V 5C2
bDepartment of Surgery, Juravinski Cancer Centre, Hamilton Health Sciences, Hamilton, Ontario, Canada
cResearch Department, Juravinski Cancer Centre, Hamilton Health Sciences, Hamilton, Ontario, Canada
dDepartment of Orthopedic Surgery, McGill University Health Centre, Montreal General Hospital, Quebec, Canada
eHenry Ford Hospital, Detroit, MI, USA
Received 2 February 2010; revised 11 March 2010; accepted 17 March 2010
Giant cell tumor of
Summary Giant cell tumor of bone (GCT) is an aggressively osteolytic primary bone tumor that is
characterized by the presence of abundant multinucleated osteoclast-like giant cells, hematopoietic
monocytes, and a distinct mesenchymal stromal cell component. Previous work in our laboratory has
shown that matrix metalloproteinase (MMP)-13 is the principal proteinase expressed by the stromal cells
of GCT. The release of cytokines, particularly interleukin-1β, by the giant cells of GCT acts on stromal
cells to stimulate a surge in MMP-13 secretion. The purpose of this study was to determine the bone
resorption capabilities of the cellular elements of GCT and the significance of the MMP-13 expression
involved in GCT bone resorption. We present a 3-dimensional histomorphometric technique developed
to analyze resorption pit depth and yield an accurate measurement of bone resorption with a direct
physical view of lacunae on bone slices. In this study, we demonstrate that the mesenchymal stromal
cells and the multinucleated giant cells of GCT are independently capable of bone resorption. However,
coculture of these 2 cell fractions shows a synergistic increase in bone resorption. In addition, inhibition
of MMP-13 reduces resorptive activity of the cells indicating that MMP-13 likely plays an important
role in this tumor. This cell-cell cooperation involves giant cell-derived cytokine up-regulation of MMP-
13 in the stromal cells, which in turn assists the giant cells in bone resorption. Future research will
involve elucidation of the role of cell-cell/matrix communication pathways in bone resorption and
tumorigenesis in GCT.
© 2010 Elsevier Inc. All rights reserved.
☆Funding sources: Hamilton Health Science New Investigator Fund, Juravinski Cancer Centre Foundation Grant, McMaster University Surgical
Associates Grant, and Canadian Institutes of Health Research (CIHR) Grant.
☆☆Conflict of Interest: All authors have no conflict of interest.
⁎Corresponding author. Hamilton, Ontario, Canada L8V 5C2.
E-mail addresses: email@example.com (I. W. Y. Mak), firstname.lastname@example.org (E. P. Seidlitz), email@example.com (R. W. Cowan),
firstname.lastname@example.org (R. E. Turcotte), email@example.com (S. Popovic), firstname.lastname@example.org (W. C. H. Wu), email@example.com (G. Singh),
firstname.lastname@example.org (M. Ghert).
0046-8177/$ – see front matter © 2010 Elsevier Inc. All rights reserved.
Human Pathology (2010) xx, xxx–xxx
ARTICLE IN PRESS
Giant cell tumor of bone (GCT) is the most common
benign aggressive skeletal neoplasm . It arises preferen-
tially in the meta-epiphyseal regions of long bones including
the femur and humerus . Although the lesion is benign
and possesses very limited metastatic and malignant
potential, it does produce substantial local osteolysis,
regional pain, and the predisposition to pathologic fractures
. Current preferred treatment of GCT consists of limb-
sparing surgery by the means of extended curettage with the
addition of local adjuvant therapies . Albeit anatomy and
function are preserved with such an approach, local
recurrence rates remain high , thus, emphasizing the
importance of developing an understanding of the biology
of the tumor and subsequent creation of more effective
On the microscopic level, GCT is characterized by 2 cell
types: the stromal cell and the giant cell. The first is the
mesenchymal stromal cell, which is of osteoblastic lineage
as demonstrated by the expression of the osteoblastic
markers osteopontin, osteocalcin, and osteonectin .
Furthermore, these cells are known to be the neoplastic
element of the tumor, and they serve to regulate the tumor
environment through the expression and secretion of the
cytokine receptor activator of nuclear factor κB ligand
(RANKL) and macrophage colony-stimulating factor (M-
CSF) . In bone, RANKL acts in concert with M-CSF to
promote the proliferation and differentiation of monocytes
(of macrophage lineage) into osteoclast-like multinucleated
giant cells . In contrast to the stromal cells, the giant cells
are the main effector cell type of GCT. The secretion of
cathepsin K by these cells promotes cleavage of type I
collagen , the major structural protein in bone. This
secretion results in characteristic bone resorption observed
However, there is evidence to suggest considerable
reciprocity between these 2 cell types in both the regulation
of the tumor environment and mediation of bone resorption
[8-10]. Previous experiments in our laboratory have
demonstrated that the giant cells secrete interleukin-1β (IL-
1β) and tumor necrosis factor-α . These cytokines signal
through their corresponding receptors on the stromal cells to
up-regulate the expression of matrix metalloproteinases
(MMPs)-2, 9, and 13 [5,9,10]. We have found the up-
regulation of MMP-13 in the stromal cells of GCT to be at
least partially controlled via the osteoblastic transcription
factor Runx2 and the ERK and JNK signaling pathways. The
secretion of these MMPs, particularly the very high
expression of MMP-13, suggests that there may be an
important role for the mesenchymal stromal cells in the
process of bone resorption .
The aims of this study were to analyze the bone
resorption capabilities of the cellular elements of GCT to
determine the significance of the mesenchymal stromal
cell expression of MMP-13 in GCT and to highlight the
importance of cellular reciprocity in the bone resorption
process. We also present a technique for analyzing resorp-
tion pit volume with an accurate 3-dimensional histomor-
2. Materials and methods
2.1. GCT sample collection
The use of all patient-derived material was approved by
our institution's Research Ethics Board, and patient informed
consent was obtained individually. The diagnosis of GCT of
bone was established by biopsy before surgical excision.
Tissues were obtained at the time of surgery from patients
undergoing tumor resection, and the diagnosis of GCT was
verified postoperatively by a bone pathologist. Tissue
samples from 4 cases of GCT of bone were used in this
study, and all experiments were performed in triplicate or as
otherwise stated for all 4 bone tumors.
2.2. Primary cell lines and cultures
We established primary cell cultures of GCT stromal
tumor cells from fresh GCT tissue. The specimens were
freshly minced in Dulbecco's modified Eagle medium (D-
MEM; Gibco, Burlington, Ontario, Canada) producing a cell
suspension with small fragments of tissue. The resultant
suspension was passed through a 20-gauge needle before
seeding in cell culture flasks with D-MEM supplemented
with 10% fetal bovine serum (FBS), 2 mmol/L L-glutamine,
100 U/mL penicillin, and 100 μg/mL streptomycin (Gibco).
The resulting cell suspension, together with macerated tissue,
was cultured in 37°C humidified air with 5% CO2. Culture
medium was changed every 2 to 3 days until approximately
80% confluence. Confluent cells (approximately 80%) were
subcultured after dissociating with trypsin and EDTA. After
several successive passages, the mesenchymal stromal cells
became the homogeneous cell type, whereas the multinu-
cleated giant cells were eliminated from the culture. Primary
cultures obtained after the fifth or sixth passage (without
any hematopoietic markers) and up to the 10th passage,
which represents the proliferating homogenous stromal
tumor cell population, were used for experiments and
Human fetal osteoblast (hFOB) 1.19 cells (American
Type Culture Collection, ATCC no. CRL-11372) and
mouse RAW 264.7 cells (ATCC no. TIB-71) were used as
control cell lines and tested in triplicate as well. Similarly,
these cell lines were maintained in supplemented D-MEM as
described for the GCT cells in optimized conditions. RAW
264.7 cells were cultured and differentiated into mature
osteoclasts in minimum essential medium-α modification
containing 10% FBS, 100 U/mL penicillin, and 100 μg/mL
2I. W. Y. Mak et al.
ARTICLE IN PRESS
streptomycin, and 50 μmol/L RANKL (R&D Systems,
Minneapolis, MN) before applying to the bone resorption
assay. At the end of the culture period, tartrate-resistant acid
phosphatase (TRAP)-positive cells with 3 or more nuclei per
cell in RAW 264.7 cells were counted.
2.3. Subfractionation of GCT populations from
Fresh GCT samples were cultured for 1 day in tissue
culture flasks containing supplemented D-MEM. After
aspirating the nonadherent blood cells and fat tissue, most
of the mononuclear stromal cells were detached by a brief
trypsin digestion, leaving giant cells and monocytes
attached to the flask. Stromal cell-enriched and giant cell/
monocyte-enriched fractions were further separated by size
using sterile 40 μm cell strainers (Fisherbrand; Fisher
Scientific International Inc., Pittsburgh, PA, USA). Fur-
thermore, the stromal cells-enriched fractions were refined
using the MACS Technology (Miltenyi Biotec, Bergisch
Gladbach, Germany), a magnetic cell-surface antigen
labeling system that we have optimized for this application.
Based on immunohistochemistry studies done by ourselves
and others, CD-14 MACS microbeads were used to
magnetically label the giant cells/monocytes of GCT. After
incubating the less than 40 μm of GCT stromal cell-enriched
fraction with CD-14 microbeads for 15 minutes at 4°C, the
cell suspension was then passed through a magnetic column
and the labeled giant cells/monocytes of GCT were retained,
whereas the unlabeled mesenchymal stromal cells passed
through. The column was removed from the magnetic
separator, and the labeled cells were eluted from the column.
The resulting mesenchymal stromal cell fraction and
hematopoietic giant cell fraction (mixed with monocyte
precursors) were subcultured accordingly in specific condi-
tions for the bone resorption assay.
2.4. Bone resorption assay with cytokine
stimulation and MMP-13 inhibition
Fresh bovine femoral cortical bone was obtained
commercially, the marrow removed, and the bones were
sterilized in ethanol. The bone was cut into 200 μm
rectangular slices using an Isomet low-speed diamond-
edged saw (Buehler, Evanston, IL). The slices were then
punched into 6-mm-diameter circles. Next, the slices were
subjected to ultrasonic cleaning and sterilized with 70%
ethanol, dried, and left under ultraviolet light for 5 minutes
before use. Before each experiment, the slices were placed
into wells of a 48-well culture plate and incubated for 1 hour
in serum-free medium. To evaluate the effects of cytokine
stimulation and MMP-13 inhibition, the same number of
cells in each GCT fraction were seeded accordingly and
treated in serum-free medium containing dimethyl sulfoxide
vehicle as the control or 1.0 ng/mL of IL-1β (R&D Systems)
with or without 200 nmol/L of selective MMP-13 inhibitors
CP-471474  and PD-331179 . IL-1β was used to
recreate the tumor environment as shown in the model in our
bovine bone slices. (A) GCT cells directly from fresh tumor sample,
(B) homogeneous stromal cells, and (C) giant cells in the giant cell-
enriched fraction. Giant cells are marked with asterisks, monocytes
with the arrow, and stromal cells with the number sign.
Representative pictures of each cell type were taken with light
microscope at magnification ×400. Bars represent 25 μm.
Cell morphology of GCT cells after 6 days of culture on
3Bone resorption in GCT of Bone
ARTICLE IN PRESS
4I. W. Y. Mak et al.
ARTICLE IN PRESS
previous study . The chemical name of CP-471474 is 2-
methylpropanamide and PD-331179 is 4-((1-methyl-2,4-
(4H)-yl)methyl)benzoic acid. For CP-471474, the published
IC50 is 0.90 nmol/L for MMP-13 and 1200 nmol/L for
MMP-1. For PD-331179, the IC50 is 0.67 nmol/L for MMP-
13, but it is more than 30 000 nmol/L for MMP-1, MMP-2,
MMP-3, MMP-7, MMP-8, MMP-9, MMP-12, MMP-14,
and MMP-17. Therefore, only MMP-13 would be inhibited
at a final concentration of 200 nmol/L used in this study.
Bone slices were cultured in duplicate for each experimental
condition for 6 days, and the medium was changed and
collected every other day. The collected conditioned media
samples were evaluated using the type I collagen fragment
enzyme-linked immunosorbent assay (ELISA) as described
in following sections.
2.5. Type I collagen fragment ELISA assay on fresh
GCT tissue media
The release of the C-terminal type I collagen fragments
(CTX) from cortical bovine bone slices was determined
using the CrossLaps for Culture kit (Nordic Bioscience
Diagnostics, Herlev, Denmark) according to the manufac-
turer's directions. Briefly, diluted conditioned culture media
and control culture medium from wells containing bone
slices without cells, were incubated for 2 hours at room
temperature in the wells of a streptavidin-coated microplate
with a biotinylated monoclonal antibody specific for
degradation products of C-terminal telopeptides of type I
collagen. After washing, a chromogenic substrate solution
was added and the color product developed in proportion to
the amount of CTX bound in the initial step. The color
development was stopped and the intensity of the color was
measured using the CytoFluor Multi-Well Plate reader series
4000 (PerSeptive Biosystems, Framingham, MA, USA), set
to 450 ± 20 nm, with a reference at 650 ± 20 nm.
2.6. Histologic condition of bone slices and
corresponding image analysis
After 6 days of bone resorption, cells were fixed in 10%
buffered formaldehyde. To confirm the consistent presence
of mature osteoclast-like multinucleated giant cells, TRAP
immunohistochemical characterization using TRAP mouse
monoclonal antibodies (26E5, Novocastra Laboratories Ltd.,
Newcastle Upon Tyne, UK) were performed on independent
experiments by identifying TRAP-positive multinucleated
cells containing 3 or more nuclei. At the end of the bone
resorption culture period, bone slices were washed in
phosphate-buffered saline and subjected to ultrasonic testing
for 5 minutes to remove adherent cells. Remaining cells were
gently removed with a cotton swab, and the bone slices were
washed in distilled water. Bone slices were mounted on glass
slides with Vectamount (Vector Laboratories, Burlingame,
CA), and resorption pits were visualized after staining with
hematoxylin (Gill's no. 3). The excavation areas on the
surface of the bone slices were observed under reflected light
with a Leitz Diaplan light microscope (Leica Microsystems,
Richmond Hill, Ontario) and a CCD camera (QImaging
Micropublisher 3.3 RTV; Surrey, BC). Photos of the stained
resorption pits from 10 random fields for each bone slice
surface were analyzed using Adobe Photoshop 7.0 (Adobe
Systems, San Jose, CA) with the Image Processing Toolkit
(John C. Russ, Reindeer Games, Inc, Raleigh, NC). Pits were
counted and expressed as a percentage of image area. A valid
bone resorption pit, evaluated by a bone pathologist, can be
recognized by its features: a small unnatural cavity or groove
on the bone slice surface, usually irregular in outline, formed
by a bone cell.
Histologic examination on the cross sections (200 μm) of
these bone slices embedded vertically in paraffin blocks was
used to detect the depth of the resorption pits from the bone
surface. Bone cross sections (5 μm) were deparaffinized in
xylene and rehydrated in ethanol before rinsing in water,
stained with hematoxylin and eosin, and examined under a
reflected Leitz Diaplan light microscope equipped with a
CCD camera. The average depth of pits was measured using
multiple evenly distributed perpendicular readings at 2 μm
apart across the pit with the Northern Eclipse 6.0 software
(Empix Imaging Inc., Mississauga, Ontario). The resorbed
volume on bone slices were the product of the average
resorbed areas and the average depth of pits.
2.7. Statistical analysis
GraphPad Prism software (GraphPad Software, Inc., La
Jolla, CA, USA) was used for statistical analysis. All data
are presented as mean ± SEM and are representative of
measurements that were performed on 4 different GCT
patient samples (n = 4). To assess variations in pit area/
volume and ELISA measurement versus treatments and cell
types, 1-way or multiway analysis of variance and the post
plated at 2 × 104cells/well in serum-free media into a 48-well culture dish in the presence of IL-1β. After 6 days of culture, cells were brushed
off and the bone slices were stained with hematoxylin. (A) Homogeneous GCT stromal cells, (B) multinucleated giant cell-enriched fraction,
(C) coculture of the homogeneous GCT stromal cell and the multinucleated giant cell-enriched fractions, (D) differentiated RAW 264.7-
positive control cells, (E) hFOB-negative control cells, and (F) untreated bone slice control. Lateral migration track of giant cells are marked
with arrows. Representative pictures of each cell type were taken with light microscope at magnification ×200. Bars represent 50 μm.
Corresponding black and white analytical images for pit area quantification are shown on the right of each bone resorption image.
Two-dimensional imaging of resorbed lacunae on the surface of bovine bone slices by various cell fractions of GCT. Cells were
5Bone resorption in GCT of Bone
ARTICLE IN PRESS
the pit (top). Representative cross-sectional images of bone slices containing GCT stromal (middle, white arrows) and giant cell (bottom, black
arrow) resorptionlacunae. Images were taken with a light microscope at ×200 (on the left) and ×400 (on the right) magnification. Bars represent
50 μm. (B) Mean resorption pit depth of GCT stromal cell fractions, giant cell fractions, coculture of both fractions, differentiated RAW 264.7
cells (positive control), and hFOB cells (negative control) for 6 days of culture on bovine bone slices in standard media in the presence of IL-1β.
Values represent the means ± SEM of 4 individual triplicate experiments after normalized to the bone slice control. Coculture of both GCT cell
fractions resulted in significantly increased average pit depth, both relative to stromal cells alone and to giant cells alone. *P b .05, **P b .01
versus stromal cells; #P b .05 versus giant cells. Statistical comparison by analysis of variance with post hoc Tukey tests.
6 I. W. Y. Mak et al.
ARTICLE IN PRESS
hoc multiple comparison Tukey test (P b .05) were applied.
Measurements were normalized to the bone slices in
cell-free culture media. Each experiment was performed at
least 3 times. P values b .05 were considered to be
3.1. Morphology of GCT cell fractions
Mesenchymal stromal cells and multinucleated osteo-
clast-like giant cells were isolated from fresh GCT patient
samples using both size and biomarker separation with cell
strainers and the CD14-labeled magnetic column, respec-
tively. The fresh GCT sample (Fig. 1A) features all 3 cell
types: the mesenchymal stromal cells, the round monocytes
of macrophage lineage, and the multinucleated giant cells.
The mesenchymal stromal cells (Fig. 1B) have been
confirmed in our previous work to be a homogeneous
culture after 5 passages based on molecular expression and
immunohistochemical marker characterization [5,9], that is,
the round monocytes of macrophage lineage and the
multinucleated giant cells were completely eliminated. The
isolated giant cell-enriched fraction contains not only mainly
giant cells and monocytes but also includes about 10% to
20% of residual stromal cells (Fig. 1C). An increase in
percentage and size of the giant cells in the giant
cell-enriched fraction was observed when compared to the
with or without 200 nmol/L of MMP-13 inhibitors CP-471474 or
PD-331179 on pit formation in GCT. Surface area resorbed was
quantified using image analysis software. The volume of resorbed
bone was calculated by multiplying the average pit depth by the 2-
dimensional pit area. (A) IL-1β significantly increased pit volume in
the giant cell fraction and in the coculture fraction but not in the
stromal cell fraction as compared to the dimethyl sulfoxide vehicle
control. *P b .05, **P b .01 versus control. (B) MMP-13 inhibition
significantly decreased pit volume in all cell fractions. *P b .05,
**P b .01 versus IL-1β. Values represent the means ± SEM of 4
individual triplicate experiments using statistical comparison by
analysis of variance with post hoc Tukey tests.
The effect of IL-1β stimulation and MMP-13 inhibition
fragment release in IL-1β-stimulated GCT cells. Quantification of
pit formation was determined by measuring CTX release in the
medium. (A) The CTX level significantly increased when the
homogeneous stromal cells were cocultured with the giant cell-
enriched fraction in the dimethyl sulfoxide vehicle control. **P b
.01 versus stromal cells; ##P b .01 versus giant cells. (B) CP-
471474 showed a significant inhibitory effect on CTX release in
IL-1β-stimulated GCT cells. *P b .05; **P b .01 versus IL-1β.
Values represent the means ± SEM of triplicate from 4 individual
experiments using statistical comparison by analysis of variance
with post hoc Tukey tests.
The effect of MMP-13 inhibition on type I collagen
7 Bone resorption in GCT of Bone
ARTICLE IN PRESS
GCT cells obtained directly from fresh tumor samples.
Giant cells and monocytes were positively stained with
TRAP to confirm this characteristic osteoclast feature (data
3.2. Two-dimensional bone resorption images
Functional assessment of the role and interaction of each
cell type in GCT was determined using a bone resorption
assay. After 6 days of culture on the bovine bone slices, cells
were removed and the bone slices were stained with
hematoxylin to visualize resorption pits. Ten random images
from each bone slice were taken and analyzed using image
analysis software to measure the contrast and size of stained
pits. Pits created by the homogeneous GCT stromal cells
were smaller and denser (Fig. 2A) as compared to the bone
slice control without cells (Fig. 2F). The number of large
excavations formed by the giant cells in the multinucleated
giant cell-enriched fraction (Fig. 2B) was slightly decreased
compared to the number of pits formed when mixing both
the homogeneous GCT stromal cells and the multinucleated
giant cell-enriched fraction (Fig. 2C). As a positive control,
the differentiated RAW 264.7 cells showed characteristic
osteoclastic lacunae formation (Fig. 2D). Note that giant
cells leave not only classic round pits like the ones of RAW
264.7 cells but also long and narrow pits. As a negative
control, hFOB cells (Fig. 2E) did not appear different from
the untreated bone slice control (Fig. 2F). Corresponding
black and white analytical images for pit area quantification
are shown on the right of each bone resorption image and the
extent of resorption was determined by calculating the
average percentage of “lacunar resorption area” (black
portion) from each image.
3.3. Cross-sectional bone resorption images
Cross sections (200 μm thick) of bone slices embedded
vertically in paraffin blocks were cut into 5-μm-thick tissue
slides. Multiple evenly distributed perpendicular lines were
drawn and measured across each pit to determine the
average depth (Fig. 3A top). Pits formed by the
mesenchymal stromal cells were shallow and less distinct
(Fig. 3A middle). However, these pits were evaluated by a
bone pathologist and determined to be valid bone
resorption pits despite their different morphology. Pits
created by the multinucleated giant cells were deep and
extended to larger coverage areas (Fig. 3A bottom);
however, those small but distinctive pits created by the
stromal cells should not be ignored as well (Fig. 3A
middle). The mean depth of resorption pits in GCT cells,
differentiated RAW 264.7 cells (positive control), and
hFOB cells (negative control) after 6 days is shown in
Fig. 3B. The giant cells resorbed significantly deeper
lacunae than the mesenchymal stromal cells. When the
giant cells were cocultured with the stromal cells, the depth
of resorption significantly exceeded the giant cell baseline
resorptive activity. Note that the bone chip surface exposed
to cell culture was relatively scalloped, comparing to the
smooth surface on the underside.
3.4. Volumetric bone resorption analysis
With a focus on the homogeneous stromal cells, the
giant cell-enriched fraction, and the combination of both in
a direct cell-cell coculture, the giant cell-containing groups
resorbed consistently more in resorbed volume than the
homogeneous stromal cells (Fig. 4A). The resorbed
volumes of IL-1β-stimulated GCT cells increased signifi-
cantly in both giant cell-containing groups, comparing to
the untreated control. There was no significant change in
volume observed on the hFOB cell-incubated bone slices
when normalized with the bone slice control (data not
shown). RAW 264.7 cells were graded a similar level of
resorbed volume as of the giant cells (data not shown).
Next, 2 newly developed MMP-13 inhibitors were tested to
isolate the role of MMP-13 in resorbing bone (Fig. 4B).
CP-471474 significantly inhibited volumetric bone resorp-
tion in all 3 groups of IL-1β-stimulated GCT cells. In
contrast, PD-331179 suppressed pit formation in the giant
cell-enriched fraction only.
3.5. Type I collagen fragments as the medium
To validate the volumetric bone resorption measure-
ment, quantification of pit formation was determined by an
ELISA that measures CTX release, the C-terminal type I
collagen fragments, in the medium after 6 days of culture.
The CTX level was highest in the medium when the
homogeneous stromal cells were cocultured with the giant
cell-enriched fraction (Fig. 5A). The capability of the
mesenchymal stromal cells in resorbing bone at a
consistently low level has also been reconfirmed here.
hFOB cells did not show any significant CTX release (data
not shown). Likewise, the effects of MMP-13 inhibitors on
CTX release in IL-1β-stimulated GCT cells demonstrated a
similar trend to that of the volumetric bone resorption
analysis. Again, CP-471474 showed a significant inhibitory
effect on CTX release in all 3 groups of IL-1β-stimulated
GCT cells (Fig. 5B).
Understanding the relationship between GCT and the
bone microenvironment is important for developing better
therapeutic strategies for this tumor. Bone is a complex
8 I. W. Y. Mak et al.
ARTICLE IN PRESS
and dynamic connective tissue composed primarily of
mineral and type I collagen . The mechanisms that
regulate the bone resorption, formation, and remodeling
are part of a complex process that involves multiple
cellular functions and interaction. In GCT and other bone
diseases, the fine balance between bone formation and
bone resorption is disrupted . We have recently
identified that large amounts of MMP-13 were secreted
into the extracellular environment by the mesenchymal
stromal cells after stimulation with cytokines, which
originated from the giant cells . We believe that this
factor may play an important role in specifically disrupting
the balance in bone formation and resorption in GCT. The
current study was designed to assess the impact of
cytokine stimulation and MMP-13 inhibition in an in
vitro model of GCT bone resorption.
In this study, we found that homogeneous GCT stromal
cells are modestly capable of resorbing bone as reflected by
resorbed bone volume and type I collagen fragment release.
Although the giant cells have been recognized as the key
cellular effectors of pathologic bone erosion [14-16], the
neoplastic stromal cells should not be overlooked. In the
study by Wen et al  in 1999, stromal cells isolated from
GCT samples with a simple 50-μm wire mesh size separation
did not show any significant bone resorption using a classical
calvariae calcium release approach, which may yield biased
results due to nonstandardized starting material. We have
demonstrated that stromal cells isolated from fresh GCT
patient samples separated by both size and biomarker
features exhibit a small but significant stromal cell initiated
excavation. This effect was validated with both physical and
biochemical measurement techniques.
Isolated multinucleated giant cells of GCT formed larger
lacunae than the RANKL-differentiated RAW 264.7 cells.
Pits excavated by mature RAW 264.7 cells in this study
showed a classic circular pattern as described in the literature
. However, in the case of giant cells, not only round pits
but also long and narrow pits were observed, suggesting that
giant cells may migrate on the surface of the bone slice to
enlarge the pit area laterally. Fusion of monocytes with
multinucleated giant cells may be occurring, as suggested by
the appearance of plasmalemmal continuity of the giant cells
during formation . The functional cycle of osteoclasts
consists of migration toward bone, followed by adherence to
the bone surface where the cell polarizes and initiates the
resorptive process . Most studies on multinucleated giant
cells focus on chemotactic attraction using migration assays
[21,22], and here, we visually demonstrate the migration of
these resorbing cells based on the morphological character-
istics of the resorption pits.
The surface area data extracted from the bone resorption
assay could not truly reflect the level of bone excavation by
itself without consideration of the depth of the pits. After
incorporating 3 dimensions into the analysis, a comprehen-
sive volumetric assessment of bone resorption would
approximate the actual bone excavation. Measuring the pit
depth on the cross section of resorbed bone slices is a
straightforward but original method, yielding the resorbed
volume rather than the more commonly used resorbed area
. The deep and extensive lacunae formed by giant cells
contrasted to stromal cells' shallow but frequent pits. Parikka
et al  showed a similar concave shape from a cross
section of a lacuna excavated by osteoclasts using an indirect
technique. It would be interesting to compare the calculated
bone resorption volume with estimated results from scanning
electron microscopy stereophotogrammetry or confocal
microscopy (both technologies based on reflection) [25,26]
in future studies. However, shortcomings of the reflection-
based microscopy such as distortions from magnification and
sample alignment could yield spurious results . The
objectively measured type-I collagen fragment release data
presented herein support similar findings in our 3-dimen-
sional pit analysis.
Among the potential pitfalls in a coculture system is the
difficulty of establishing the secreted factors responsible for
the bone resorption. To further our understanding of the
relationship between the stromal cell-derived MMP-13 and
the giant cells in bone resorption, 2 newly developed MMP-
13 inhibitors were tested. Specific MMP-13 inhibition is
achieved by both inhibitors at a final concentration of 200
nmol/L, which is well below that of other MMP inhibition.
MMP-13, the highest expressed interstitial collagenase in
GCT stromal cells [9,10], initiates bone resorption by
generating collagen fragments that activate osteoclasts at
the ruffled membranes [28-31]. MMP-13 has also been
shown to activate other latent MMPs such as progelatinase
B/pro-MMP-9 [32,33]. CP-471474 demonstrated a signifi-
cant inhibitory effect on excavation in both resorbed bone
volume and CTX release in all 3 cell groups, whereas
PD-331179 did not appear to be effective. The divergence of
inhibitory capabilities of these 2 MMP-13 inhibitors may
arise from differences in inhibitory mechanisms. Whereas
CP-471474 binds to the catalytic zinc ion, PD-331179
In summary, our analysis of the bone resorption
capabilities of the cellular elements of GCT has revealed
several interesting characteristics of the cells of GCT. The
alternative technique developed for analyzing resorption pit
depth allows a direct physical view of lacunae on bone
slices. We have shown that although the mesenchymal
stromal cells of GCT are able to resorb bone, they did so
in an inefficient way when cultured alone. The multinu-
cleated giant cells in coculture with the stromal cells
surpassed the bone resorbing capacity of giant cells alone,
suggesting synergism between the 2 cell types. We also
confirmed that MMP-13 plays an important role in bone
resorption in this tumor. Therefore, this synergism between
the giant cells and the stromal cells may involve the giant
cell-derived cytokine-regulation of MMP-13 secretion by
the stromal cells, which in turn assists the giant cells in
bone resorption. More work is required to elucidate the
mechanism of MMP-13 in GCT cell reciprocity and to
9 Bone resorption in GCT of Bone
ARTICLE IN PRESS
determine other potential players in bone resorption and
tumorigenesis in GCT.
We would like to thank Pfizer for the courtesy in
providing the MMP-13 inhibitors. We would also like to
acknowledge Cindy Wong and Konrad Salata for their help
in this study.
 Gupta R, Seethalakshmi V, Jambhekar NA, et al. Clinicopathologic
profile of 470 giant cell tumors of bone from a cancer hospital in
western India. Ann Diagn Pathol 2008;12:239-48.
 Turcotte RE. Giant cell tumor of bone. Orthop Clin North Am 2006;
 Balke M, Schremper L, Gebert C, et al. Giant cell tumor of bone:
treatment and outcome of 214 cases. J Cancer Res Clin Oncol 2008;
 Becker WT, Dohle J, Bernd L, et al. Local recurrence of giant cell
tumor of bone after intralesional treatment with and without adjuvant
therapy. J Bone Joint Surg Am 2008;90:1060-7.
 Ghert M, Simunovic N, Cowan RW, Colterjohn N, Singh G. Properties
of the stromal cell in giant cell tumor of bone. Clin Orthop Relat Res
 Nishimura M, Yuasa K, Mori K, et al. Cytological properties of
stromal cells derived from giant cell tumor of bone (GCTSC) which
can induce osteoclast formation of human blood monocytes without
cell to cell contact. J Orthop Res 2005;23:979-87.
 Lindeman JH, Hanemaaijer R, Mulder A, et al. Cathepsin K is the
principal protease in giant cell tumor of bone. Am J Pathol 2004;165:
 Sims NA, Gooi JH. Bone remodeling: multiple cellular interactions
required for coupling of bone formation and resorption. Semin Cell
Dev Biol 2008;19:444-51.
 Mak IW, Cowan RW, Popovic S, Colterjohn N, Singh G, Ghert M.
Upregulation of MMP-13 via Runx2 in the stromal cell of giant cell
tumor of bone. Bone 2009;45:377-86.
 Cowan RW, Mak IW, Colterjohn N, Singh G, Ghert M. Collagenase
expression and activity in the stromal cells from giant cell tumour of
bone. Bone 2009;44:865-71.
 Rohde LE, Ducharme A, Arroyo LH, et al. Matrix metalloproteinase
inhibition attenuates early left ventricular enlargement after experi-
mental myocardial infarction in mice. Circulation 1999;99:3063-70.
 Johnson AR, Pavlovsky AG, Ortwine DF, et al. Discovery and
characterization of a novel inhibitor of matrix metalloprotease-13 that
reduces cartilage damage in vivo without joint fibroplasia side effects.
J Biol Chem 2007;282:27781-91.
 Werner M. Giant cell tumour of bone: morphological, biological and
histogenetical aspects. Int Orthop 2006;30:484-9.
 Kanehisa J, Izumo T, Takeuchi M, Yamanaka T, Fujii T, Takeuchi H.
In vitro bone resorption by isolated multinucleated giant cells from
giant cell tumour of bone: light and electron microscopic study.
Virchows Arch A Pathol Anat Histopathol 1991;419:327-38.
 Ohsaki Y, Takahashi S, Scarcez T, et al. Evidence for an autocrine/
paracrine role for interleukin-6 in bone resorption by giant cells from
giant cell tumors of bone. Endocrinology 1992;131:2229-34.
 Oreffo RO, Marshall GJ, Kirchen M, et al. Characterization of a cell
line derived from a human giant cell tumor that stimulates osteoclastic
bone resorption. Clin Orthop Relat Res 1993:229-41.
 Wen J, Xie D, Yao J, Zhang M, Bi J. Effect of cytokines on in vitro
bone resorption by cells isolated from giant cell tumor of bone. Chin
Med J (Engl) 1999;112:443-7.
 Battaglino R, Kim D, Fu J, Vaage B, Fu XY, Stashenko P. c-myc is
required for osteoclast differentiation. J Bone Miner Res 2002;17:
 Bird MC, Garside D, Jones HB. Multinucleated giant cells in primary
cultures derived from canine bone marrow—evidence for formation of
putative osteoclasts. Cell Tissue Res 1992;268:17-30.
 Blair HC. How the osteoclast degrades bone. Bioessays 1998;20:
 Liao TS, Yurgelun MB, Chang SS, et al. Recruitment of osteoclast
precursors by stromal cell derived factor-1 (SDF-1) in giant cell tumor
of bone. J Orthop Res 2005;23:203-9.
 Zheng MH, Fan Y, Smith A, Wysocki S, Papadimitriou JM, Wood DJ.
Gene expression of monocyte chemoattractant protein-1 in giant cell
tumors of bone osteoclastoma: possible involvement in CD68+
macrophage-like cell migration. J Cell Biochem 1998;70:121-9.
 Sugawara K, Hamada M, Hosoi S, Tamaoki T. A useful method to
evaluate bone resorption inhibitors, using osteoclast-like multinucle-
ated cells. Anal Biochem 1998;255:204-10.
 Parikka V, Lehenkari P, Sassi ML, et al. Estrogen reduces the depth of
resorption pits by disturbing the organic bone matrix degradation
activity of mature osteoclasts. Endocrinology 2001;142:5371-8.
 Boyde A, Dillon CE, Jones SJ. Measurement of osteoclastic resorption
pits with a tandem scanning microscope. J Microsc 1990;158:261-5.
 Boyde A, Howell PG, Franc F. Simple SEM stereophotogrammetric
method for three-dimensional evaluation of features on flat substrates.
J Microsc 1986;143:257-64.
 Boyde A, Jones SJ. Pitfalls in pit measurement. Calcif Tissue Int
 Fuller K, Chambers TJ. Localisation of mRNA for collagenase in
osteocytic, bone surface and chondrocytic cells but not osteoclasts. J
Cell Sci 1995;108(Pt 6):2221-30.
 Holliday LS, Welgus HG, Fliszar CJ, Veith GM, Jeffrey JJ, Gluck SL.
Initiation of osteoclast bone resorption by interstitial collagenase. J
Biol Chem 1997;272:22053-8.
 Holliday LS, Welgus HG, Hanna J, et al. Interstitial collagenase
activity stimulates the formation of actin rings and ruffled membranes
in mouse marrow osteoclasts. Calcif Tissue Int 2003;72:206-14.
 Nakamura H, Sato G, Hirata A, Yamamoto T. Immunolocalization of
matrix metalloproteinase-13 on bone surface under osteoclasts in rat
tibia. Bone 2004;34:48-56.
 Knauper V, Cowell S, Smith B, et al. The role of the C-terminal
domain of human collagenase-3 (MMP-13) in the activation of
procollagenase-3, substrate specificity, and tissue inhibitor of
metalloproteinase interaction. J Biol Chem 1997;272:7608-16.
 Knauper V, Smith B, Lopez-Otin C, Murphy G. Activation of
progelatinase B (proMMP-9) by active collagenase-3 (MMP-13). Eur J
10I. W. Y. Mak et al.
ARTICLE IN PRESS