| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
ARTICLES |
Department Biomedical Engineering, University of Alabama at Birmingham (A.W.E., A.Y.-J.), Birmingham, Alabama 35294; Departments of Pathology and Physiology and Cell Biology, University of Pittsburgh and Veterans Affairs Medical Center (H.C.B.), Pittsburgh, Pennsylvania 15261
Address all correspondence and requests for reprints to: Dr. Harry C. Blair, University of Pittsburgh, 705 Scaife Hall, DeSoto & Terrace Streets, Pittsburgh, Pennsylvania 15261. E-mail: hcblair{at}imap.pitt.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Glucocorticoids suppress bone formation and cause calcium loss in most species, but the mechanisms that cause severe and rapid changes in specific sites, particularly the femoral head and neck, are debated. Although glucocorticoids are commonly associated with fractures and aseptic necrosis in the femoral head, acute bone loss or aseptic necrosis occurring without glucocorticoids cannot be distinguished (2), suggesting that an underlying mechanism may be triggered by multiple stimuli.
Fat embolization was the first mechanism proposed for femoral head necrosis. This hypothesis is supported by the occurrence of aseptic necrosis in lipid disorders (3). However, whereas embolization causes necrosis and glucocorticoids increase circulating lipids, the evidence that this is a key mechanism is limited. A histological study showed fatty emboli with microfractures (4), but there is no evidence that this occurs frequently with glucocorticoid exposure, and how the mechanism would affect specific bone regions preferentially is unclear. Later work suggested that vasculitis with steroid exposure produces microemboli, suggesting that this mechanism may be involved in specific circumstances (5).
Increased ratios of osteoclast differentiation-promoting to osteoclast
differentiation-blocking cytokines occur in osteoblasts exposed to
glucocorticoids. These include the TNF
-family cytokine RANK-ligand,
which activates RANK, and a soluble decoy receptor, osteoprotegerin
(6, 7), which binds RANK-ligand. Glucocorticoids stimulate
adenylate cyclase in bone cells, a pathway that also mediates PTH
activity (8), and PTH increases the expression of
RANK-ligand. Changes in RANK-ligand expression are probably involved in
increased systemic bone resorption by glucocorticoids. However, it is
not clear that these cytokines, which regulate skeletal turnover, cause
the site-specific damage associated with glucocorticoids.
Glucocorticoid-dependent shifts in osteoblastic differentiation and glucocorticoid-dependent apoptosis also occur. Osteoblastic differentiation can be arrested by glucocorticoids (9), although it is unclear whether this is a primary factor in vivo. In mice, glucocorticoid-dependent cell death in osteoblasts and osteocytes of vertebral trabeculae and femoral cortical bone have been demonstrated (10); this same study showed cell death consistent with an apoptotic mechanism in iliac bone biopsies of humans treated with glucocorticoids for several years. Both osteoblasts and chondroblasts may be subject to glucocorticoid-mediated apoptosis (11). However, studies of isolated osteoblasts report variable findings that frequently are not consistent with glucocorticoid-mediated apoptosis (12). Osteoblastic markers such as alkaline phosphatase are often stimulated by dexamethasone (13), and glucocorticoids are often used to stimulate differentiation of osteoblasts from precursors (14).
We hypothesized that changes in bone cell activity occur with exposure to high concentrations of glucocorticoid before events such as fracture or necrosis. This hypothesis was tested by studying the femora of rabbits exposed for 28 days to 1.7 or 4 µmol/kg·day of methylprednisolone acetate. Histology, scanning electron microscopy, and tetracycline labeling were studied to evaluate bone formation and resorption. Bone quality and cell survival were examined in additional detail using tetracycline adsorption and TUNEL assays. Tetracycline adsorption occurs at the molecular interface with bone mineral; it provides a direct indication of exposed apatite surfaces. This adsorption measures small quantities of label relative to the tetracycline incorporated during bone synthesis in traditional labeling of the bone mineralizing surface, but is observable by laser scanning confocal microscopy. TUNEL assays label cellular DNA fragmentation, which occurs in apoptosis and some other forms of cell death, using terminal deoxynucleotidyl transferase to incorporate labeled nucleotides for colorimetric analysis in cross-sections.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Histology
Both femora of each animal were harvested. Femoral heads were
bisected after radiography and fixed 18 h in 10% formalin, 50
mM phosphate, 250 mM sucrose, pH 7.4. Half of
each femur was decalcified in 8% formate and embedded in paraffin.
Six-micrometer paraffin sections were cut and deparaffinized in xylene;
these were either stained with hematoxylin and eosin for histology or
used unstained for terminal deoxynucleotidyl transferase labeling.
Undecalcified bone was dehydrated in graded alcohols, with samples
taken for scanning electron microscopy, and in the remaining part, the
alcohol was replaced with acetone and the bone was embedded in
methacrylate for undecalcified sections. After curing the plastic, 7
µm undecalcified sections were cut on a Carl Zeiss microtome and deplasticized in xylene. These were
used unstained in phase polarization and tetracycline labeling studies.
Subarticular trabecular parameters were generated by computer from
digital scans of cross-sections, using a digital pen to outline the
area to be integrated, which was trabecular bone to 3 mm distal to the
femoral head. Examples of scans used for measurements are shown in Fig. 1B
. Resorbed area was measured as percent
pitted area in 3 mm of subarticular trabecular bone from scanning
electron micrographs, as in Fig. 1C; results were similar to linear
resorbed area from cross-sections (not shown).
|
Laser scanning detection of passively adsorbed tetracycline
Laser excitation with high-gain photodetection was used to
capture the weak signals from adsorbed label at nonsynthetic sites.
This adsorption occurs at the molecular interface of the hydroxyapatite
crystal, and because of the small quantity of label it is not resolved
by standard epifluorescence microscopy. Observations of 7 µm
undecalcified deplasticized sections of bone used a Leica Corp. TCS NT instrument with an oil immersion, 1.0 NA,
40x objective. An argon laser, 488 nm, was used to excite the fluor,
with a 540–600 nm window for the photodetector. Useful amplification
was limited by background from the blue-green collagen
autofluorescence, which has broad and overlapping excitation and
emission bands. Laser power and photodetector gain were adjusted to
produce neutral background from autofluorescence in unstained controls;
then experimental sections were imaged with adsorbed tetracycline
appearing as bright bands at sites where the fluor contacts
hydroxyapatite-containing surfaces. Characteristics of Langmuir
adsorption of tetracycline are detailed in Misra (15) and
Baker et al. (16). Selected laser scanning
confocal fluorescence images were compared with phase polarized images
to illustrate changes in collagen birefringence at sites labeled with
adsorbed tetracycline.
TUNEL assays
Terminal deoxynucleotidyl transferase-mediated dUTP nick
endlabeling assays were used for qualitative analysis of DNA
fragmentation. This was performed by colorimetric apoptosis detection
incorporating biotinylated or digoxigenin-modified nucleotides
(Promega Corp., Madison, WI; Roche Molecular Biochemicals, Indianapolis, IN). For these assays, 6-µm
sections were deparaffinized by xylene washes. Sections were treated
for 5 min with 0.3% hydrogen peroxide to reduce interference from
tissue peroxidase activity if peroxidase-labeling was to be performed.
Tissue was permeabilized using 20 µg/ml proteinase K. Labeling used
biotin- or digoxigenin-coupled dUTP and 1 U/ml of terminal
deoxynucleotidyl transferase in 30 mM Tris in 140
mM Na cacodylate, 1 mM
CoCl2, pH 7.2, and was carried out 30 min at 37
C; other steps were at room temperature. In each case, controls were
produced in which terminal deoxynucleotidyl transferase was omitted,
but all other conditions were identical. After end-labeling, sections
were washed with PBS with 1% BSA. Label was reacted
anti-biotin-streptavidin horseradish peroxidase conjugate or
anti-digitonin alkaline phosphatase conjugate for 1 h using the
recommended antibody concentrations. Peroxidase detection then used 5
min incubation in 0.5 mM diaminobenzidine activated with
0.01% H2O2 to produce a
brown precipitate. Alkaline phosphatase was detected using 30 min
incubation at pH 8 in 1 mM 5-bromo-4-chloro-3-indolyl
phosphate and 0.2 mM nitro blue tetrazolium to produce a
blue precipitate.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). The trabecular
thickness and volume were decreased in all of the
methylprednisolone-treated animals relative to controls (Fig. 1B).
Scanning electron microscopy of control and glucocorticoid treated bone
showed that the subarticular trabecular bone was 50–80% resorptive,
whereas in controls less than 20% of the surface was pitted (Fig. 1C).
The cohort of animals treated with 1.7 µmol/kg·day showed minimal
and inconsistent changes in these tests and was not studied in detail.
Principal histomorphometric parameters (17) are summarized
in Table 1.
|
), whereas controls showed normal bone
synthesis at this site (Fig. 2B). Overall, examination of tetracycline
uptake in whole-mount specimens was decreased 70% in treated animals
(Fig. 2C). Continuity of trabecular bone in treated animals appeared to
be compromised, with fewer continuous trabeculae seen on the
cross-sections, and marked resorption across trabeculae, and pronounced
resorption was observed at intersections of trabecular plates (Fig. 2, D–F). However, derived parameters related to connectivity, such as
trabecular spacing, did not meet tests of statistical difference (Table 1). Nevertheless, the striking appearance of the sections suggests that
detailed analysis of early connectivity changes in
corticosteroid-treated subarticular bone may be useful. There was no
difference in cortical thickness of the femoral shaft. Mean thickness
of the medial cortex 2 cm distal to the articular surface was 620
± 120 µm in experimental and 640 ± 102 µm in control
animals. No specimen showed inflammation, necrosis, or vascular
changes.
|
Laser excitation with high-gain photodetection was used to capture the
weak signals from adsorbed label at nonsynthetic sites. This adsorption
occurs at the molecular interface of the hydroxyapatite crystal, and
because of the small quantity of label it is not usually resolved by
epifluorescence microscopy, although when high concentrations of
tetracycline are used faint surface labels can sometimes be detected by
standard epifluorescence at nonmetabolically labeled surfaces.
Autofluorescence of marrow cell granules and collagen also become
visible with laser stimulation and high-gain signal detection; for the
present purposes the collagen autofluorescence was adjusted to a
neutral background intensity in the bone trabeculae, and marrow
fluorescence was ignored as irrelevant. Tetracycline was bound to most
bone surfaces, as expected (arrowheads, Fig. 3A), but in addition there were numerous
areas of deep trabecular bone that adsorbed tetracycline in an odd
lamellar pattern (arrows, Fig. 3A), a pattern consistent
with the orientation of adsorbed fluor to hydroxyapatite crystals
aligned with collagen lamellae (see Discussion).
High-resolution polarized phase microscopy of these same deplasticized
undecalcified sections showed subtle defects in these same areas, where
collagen birefringence was attenuated and larger numbers of lacunae
appeared to be empty (Fig. 3B).
|
, A and C, E). No
comparable regions of deep trabecular tetracycline adsorption were seen
in cortical bone, either in treated or control animals (not
illustrated). Typically, some surface adsorbed label is observed even
in standard epifluorescence when the interval before specimen
collection is less than 3–5 days, and the 1-day interval used here was
chosen to maximize adsorbed labeling. However, the short label rendered
surface labeling visible only in the laser scanned sections. When areas
of metabolically labeled bone from control animals were observed at
exposures appropriate for this intense label (Fig. 4B), neither
adsorbed tetracycline nor matrix background were seen. Portions of
control bone lacking label, visualized at high gain in Figs. 4, D and
F, showed surface binding and background fluorescence as in the treated
animal, but none of the deep trabecular adsorption was present in
controls. The amount of surface adsorption varied regionally, in both
treated and control bone, an example of which is shown in Fig. 4F. This
suggests that some areas were protected from surface adsorption, that
some adsorbed label was removed between labeling and killing of the
animal, or that some label was lost in specimen preparation.
|
). TUNEL assays using peroxidase-labeling
showed strong nuclear staining in trabeculae of treated but not control
animals (Fig. 5, A and B). Interestingly,
whereas there was strong DNA fragmentation in the subarticular
trabecular bone, there was virtually no labeling in cortical bone (Fig. 5C). Controls without terminal nucleotidyl transferase showed no
nuclear staining (not illustrated).
|
), a pattern
consistent with the areas of passive tetracycline adsorption (Figs. 3 and 4). TUNEL assays using alkaline phosphatase were compared because
of concern regarding background seen when using the peroxidase method
(Fig. 6B); this showed the same nuclear pattern but nonnuclear marrow
background, attributed to residual peroxidase activity in erythrocytes,
was eliminated. Sections from control bone were not negative for TUNEL
staining, but rather showed a different pattern, with a fraction of 1%
of osteoblasts and occasional marrow cells labeled by the terminal
deoxynucleotidyl transferase. This is shown at lower magnification in
Fig. 6C to indicate the small proportion of nuclei labeled in the
control bone. No labeling of osteocytes in the matrix was observed in
control sections, indicating that apoptosis in matrix cells is, at
most, a rare event in control bone.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This model was used to examine changes in bone formation and
degradation before the occurrence of fracture or collapse of the
femoral head. This avoids potential confounding secondary changes, such
as thrombosis, which can either cause skeletal damage or occur as a
consequence of it. Embolization, fat-related thromboemboli, and other
occlusive vascular events have been hypothesized to be the major cause
of femoral head necrosis related to glucocorticoid exposure. Under the
conditions studied, no vascular changes, thrombi, or necrosis of marrow
cells were observed (Fig. 1B). On the other hand, detailed analysis of
the bone indicated that patches of the trabecular bone matrix had
become damaged and permeable (Figs. 3 and 4), and contained
osteoblastic and osteocytic nuclei with fragmented DNA (Figs. 5 and 6).
This suggests that apoptosis is an early event in
glucocorticoid-mediated bone loss. At some sites, cells with fragmented
DNA and damaged matrix constituted a large proportion of the trabecular
bone (Fig. 4E). This suggests that aseptic necrosis may also occur,
independently of vascular changes, when glucocorticoid-dependent
trabecular-cell death is extensive.
The lamellar pattern of tetracycline adsorption in the damaged bone
(Figs. 3 and 4) is very interesting and has not been reported. Bone,
even when devitalized, normally retains a very tight structure and can
be used as an impervious substrate. The novel lamellar pattern
indicates that, when osteocytic death occurs, the permeability of the
matrix increases rapidly, for which no mechanism or precedent is known.
The findings are, however, consistent with previous studies of
tetracycline adsorption. Adsorption is oriented to hydroxyapatite
deposition, which is in turn aligned with collagen lamellae that occur
in repeating rows of collagen bundles oriented with and orthogonal to
the major axis of stress in a section of bone (21, 22). In
the cortex, cross-sections of lamellae produce generally alternating
bands of light and dark, whereas in the trabecular bone, due to the
complexity of the surface, on a random cross-section the pattern shows
concentric curved layers (see Fig. 3).
TUNEL-positivity may reflect apoptosis or other causes of cell death,
but the coordinated pattern in groups of osteoblasts and association
with matrix changes suggests that a programmed, apoptotic process is
indeed involved. The matrix permeability demonstrated by lamellar
tetracycline adsorption suggests further that osteoblasts, before or
during their coordinated death, participate in matrix degradation. The
subtle matrix changes, which might relate to enzyme secretion, are
reminiscent of those in hypertrophic chondrocytes in the transition to
bone, and are worthy of further investigation. Hypertrophic
chondrocytes produce collagenases (23), and osteoblasts
in vitro are known to have the potential to do the same
(24, 25), but the physiological importance of this is
unknown. The changes in collagen birefringence (Fig. 3B) indicate that
the organic component of the matrix at labeled sites was altered.
However, further studies will be required to determine the biochemical
mechanisms responsible for the observed permeabilization of the
osteocytic matrix.
Glucocorticoid-dependent cell death or apoptosis with matrix damage is
one of several mechanisms that may underlie femoral bone damage and
lead to similar outcomes. For example, traumatic damage or embolic
events cause necrosis, and in disorders of lipid metabolism embolic
events are frequent (3). Glucocorticoids increase serum
lipids and the fat content of the bones (26), and fatty
emboli with microfractures occur when experimental vascular damage is
combined with glucocorticoid exposure (5). It has also
been noted that glucocorticoids stimulate in bone the formation of cAMP
(8), the principal second messenger mediating PTH
activity. In recent work, it has been shown that a key pathway in
osteoclast development, osteoblast expression of RANK-ligand, is
promoted by glucocorticoid exposure (6, 7). RANK-ligand is
a cell membrane-factor whose activity is mediated by contact
(27). The present study does not measure the activity of
the RANK-ligand on trabecular osteoblasts. However, in trabecular bone
the matrix is damaged and permeable (Fig. 4) and the osteoblasts and
osteocytes have fragmented DNA (Fig. 6). Thus, increased expression of
RANK-ligand by these cells is highly unlikely. The source of
differentiation factors for the obviously accelerated osteoclastic
activity (Fig. 1C) may be stromal cells, which were viable despite the
apoptosis of the trabecular cells. Marrow stromal cells produce
RANK-ligand and other important osteoclast-modifying cytokines, and
respond to stimuli for bone turnover including PTH
(28, 29).
Glucocorticoids modify osteoblastic differentiation and survival. Our
study suggests that this is a primary factor in response to high
concentrations of glucocorticoids in vivo. Apoptosis
following glucocorticoid treatment has been observed in vertebral
trabecular bone and femoral cortical bone in mice, in iliac biopsies of
humans following years of corticosteroid therapy (10), in
chondroblasts (11), and even in osteoclasts
(30). In this study, more specific patterns of
osteoblastic and osteocytic death were observed. Rare osteoblasts, less
than 1% of trabecular lining cells, were TUNEL positive in normal
bone. In bone from methylprednisolone-treated animals, the appearance
of the cortex resembled control bone, but in trabecular bone undergoing
rapid resorption, a significant proportion of the trabecular
osteoblasts and osteocytes were TUNEL-reactive (Figs. 5 and 6).
Although the pattern of osteocytic permeability and TUNEL-positivity
(
Figs. 3–6) suggest apoptosis of coordinated groups of cells, the
TUNEL assay measures only DNA fragmentation. This is a late change in
apoptosis but may also reflect other mechanisms of cell death than
apoptosis, and further investigation of this point with specific assays
for mechanisms, such as caspase or BCL-2 activity, will be important.
TUNEL-positive osteoclasts were not observed, but we cannot exclude the
presence of apoptotic osteoclasts because such cells are degraded
rapidly and difficult to detect in bone sections.
The presence of contiguous regions of damaged matrix (Figs. 4 and 5)
suggests that apoptosis does probably occur in cohort of connected
osteoblasts and osteocytes. Osteoblasts produce bone in patches of
cells approximately 50–100 µm in diameter, which are connected by
gap junctions both on the surface (31) and which are
contiguous with buried layers of osteocytes (32, 33),
which are probably related to units of bone formation in trabecular
bone, similar to osteons in lamellar bone.
Glucocorticoids function in normal bone turnover, and the apoptosis in
rabbit bone thus probably reflects extreme stimulation of a
physiologically important receptor. Tissue culture studies show that
osteoblast maturation is stimulated by glucocorticoids at appropriate
concentrations (13, 14). Glucocorticoids often have
variable effects on cell differentiation and survival that are
concentration dependent. Curiously, glucocorticoid-related osteocytic
death occurred selectively in the subarticular trabecular bone (Fig. 5). It is likely that the sensitivity of the trabecular bone reflects
differences in the microenvironment of the osteocytes and osteoblasts.
Trabecular bone is composed of thin plates that flex under load to a
greater extent than in cortical bone. Further, mechanical strain
stimulates production of osteoblast signals including nitric oxide
(34), which are proapoptotic and may affect the
glucocorticoid response. An interaction of glucocorticoid and stretch
responses would be consistent with the observed osteocyte and
osteoblast apoptosis, but such an interaction has not been verified
experimentally.
In conclusion, our data indicate that high concentrations of
glucocorticoids damage bone by a mechanism that includes osteocyte and
osteoblast death, most likely by apoptosis of contiguous groups of
cells in osteons. In addition, this study shows that adsorption of
tetracycline can be used to detect very early lesions in bone that
create only subtle morphological changes (Fig. 5). We hypothesize that
the anatomy of the femur, which has a large active surface area and
high stress load, makes the subarticular trabecular bone particularly
sensitive to glucocorticoid-induced apoptosis.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received September 8, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  F.-S. Wang, J.-Y. Ko, D.-W. Yeh, H.-C. Ke, and H.-L. Wu Modulation of Dickkopf-1 Attenuates Glucocorticoid Induction of Osteoblast Apoptosis, Adipocytic Differentiation, and Bone Mass Loss Endocrinology, April 1, 2008; 149(4): 1793 - 1801. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. J. ROBINSON, C. W. BORYSENKO, and H. C. BLAIR Tumor Necrosis Factor Family Receptors Regulating Bone Turnover: New Observations in Osteoblastic and Osteoclastic Cell Lines Ann. N.Y. Acad. Sci., November 1, 2007; 1116(1): 432 - 443. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Chavassieux, E. Seeman, and P. D. Delmas Insights into Material and Structural Basis of Bone Fragility from Diseases Associated with Fractures: How Determinants of the Biomechanical Properties of Bone Are Compromised by Disease Endocr. Rev., April 1, 2007; 28(2): 151 - 164. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Tasic, Z. Trajkovski, G. Zafirovski, Z. Gucev, and R. S. Trompeter Aseptic necrosis of both tali in a child with steroid-dependent nephrotic syndrome Nephrol. Dial. Transplant., June 1, 2006; 21(6): 1702 - 1704. [Full Text] [PDF]
|
|
|
|
 |
|
 C. Kristo, R. Jemtland, T. Ueland, K. Godang, and J. Bollerslev Restoration of the coupling process and normalization of bone mass following successful treatment of endogenous Cushing's syndrome: A prospective, long-term study Eur. J. Endocrinol., January 1, 2006; 154(1): 109 - 118. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Sun, H. C. Blair, Y. Peng, N. Zaidi, O. A. Adebanjo, X. B. Wu, X. Y. Wu, J. Iqbal, S. Epstein, E. Abe, et al. Calcineurin regulates bone formation by the osteoblast PNAS, November 22, 2005; 102(47): 17130 - 17135. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Talamo, E. Angtuaco, R. C. Walker, L. Dong, M. H. Miceli, M. Zangari, G. Tricot, B. Barlogie, and E. Anaissie Avascular Necrosis of Femoral and/or Humeral Heads in Multiple Myeloma: Results of a Prospective Study of Patients Treated With Dexamethasone-Based Regimens and High-Dose Chemotherapy J. Clin. Oncol., August 1, 2005; 23(22): 5217 - 5223. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F.-S. Wang, C.-L. Lin, Y.-J. Chen, C.-J. Wang, K. D. Yang, Y.-T. Huang, Y.-C. Sun, and H.-C. Huang Secreted Frizzled-Related Protein 1 Modulates Glucocorticoid Attenuation of Osteogenic Activities and Bone Mass Endocrinology, May 1, 2005; 146(5): 2415 - 2423. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R Okazaki, A Sakai, A Ootsuyama, T Sakata, T Nakamura, and T Norimura Trabecular bone mass and bone formation are preserved after limb immobilisation in p53 null mice Ann Rheum Dis, April 1, 2004; 63(4): 453 - 456. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Takuma, T. Kaneda, T. Sato, S. Ninomiya, M. Kumegawa, and Y. Hakeda Dexamethasone Enhances Osteoclast Formation Synergistically with Transforming Growth Factor-{beta} by Stimulating the Priming of Osteoclast Progenitors for Differentiation into Osteoclasts J. Biol. Chem., November 7, 2003; 278(45): 44667 - 44674. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. H. Boss and I. Misselevich Osteonecrosis of the Femoral Head of Laboratory Animals: The Lessons Learned from a Comparative Study of Osteonecrosis in Man and Experimental Animals Vet. Pathol., July 1, 2003; 40(4): 345 - 354. [Abstract] [Full Text] [PDF]
|
|
