Full text of DOI:10.1359/jbmr.2002.17.2.301

JOURNAL OF BONE AND MINERAL RESEARCH
Volume 17, Number 2, 2002
©
2002 American Society for Bone and Mineral Research
Recombinant Human Bone Morphogenetic Protein-2 Enhances
Osteotomy Healing in Glucocorticoid-Treated Rabbits*
1
1
1
1
1
1
C.A. LUPPEN, C.A. BLAKE, K.M. AMMIRATI, M.L. STEVENS, H.J. SEEHERMAN, J.M. WOZNEY,
1
,2
and M.L. BOUXSEIN
ABSTRACT
The objectives of this study were to evaluate the effect of chronic prednisolone treatment on osteotomy healing in
rabbits and to determine whether recombinant human bone morphogenetic protein-2 (rhBMP-2) would enhance
healing in the presence of chronic glucocorticoid therapy. Forty-nine skeletally mature, male rabbits were injected
with either prednisolone (n ؍ 26; 0.35 mg/kg per day, three times a week) or saline (n ؍ 23). After a 6-week
pretreatment period, bilateral ulnar osteotomies were created surgically. One osteotomy was treated with rhBMP-2
(0.2 mg/ml of rhBMP-2, 40 ␮g of rhBMP-2 total) delivered on an absorbable collage sponge (ACS), whereas the
contralateral osteotomy remained untreated. Prednisolone or saline treatment was continued until the rabbits were
killed either 6 weeks or 8 weeks after creation of the osteotomy. Osteotomy healing was evaluated by radiography,
peripheral quantitative computed tomography (pQCT), torsional biomechanics, and undecalcified histology.
Because we observed similar responses to both prednisolone and rhBMP-2/ACS treatment in the 6-week and
8-week cohorts, the results from these time points were combined. Serum osteocalcin and vertebral trabecular bone
density were lower in the prednisolone-treated rabbits. Prednisolone treatment dramatically inhibited osteotomy
healing. In the untreated ulnas, callus area and torsional strength were 25% and 55% less, respectively, in the
prednisolone-treated rabbits than in the saline group (p < 0.001 for both). rhBMP-2/ACS enhanced healing in both
the prednisolone- and the saline-treated groups, although the effect was larger in the prednisolone-treated rabbits.
In the prednisolone-treated rabbits, callus area and torsional strength were 40% and 165% greater (p < 0.001 for
both), respectively, in osteotomies treated with rhBMP-2/ACS compared with the contralateral, untreated osteot-
omies. Histological evaluation confirmed that osteotomy healing was inhibited by prednisolone and accelerated by
rhBMP-2/ACS. In summary, a single application of rhBMP-2/ACS counteracted the inhibition of osteotomy
healing caused by prednisolone exposure. These results suggest that rhBMP-2/ACS may be a useful treatment for
enhancing fracture healing in patients who are undergoing chronic glucocorticoid therapy. (J Bone Miner Res
2002;17:301–310)
Key words: fracture healing, glucocorticoid, prednisolone, recombinant human bone morphogenetic
protein-2, rabbit, biomechanics, osteotomy
INTRODUCTION
musculoskeletal diseases, pharmacologic doses of glucocor-
ticoids influence bone metabolism, leading to impaired bone
LUCOCORTICOID THERAPY is prescribed widely for its
anti-inflammatory and immunosuppressive effects. Al-
though useful in the treatment of respiratory, skin, and
(1–3)
healing and bone loss.
The majority of patients taking
G
glucocorticoids for over 6 months develop osteoporosis and,
consequently, patients on high-dose therapy have an in-
(2–4)
creased risk of fracture.
*
Presented in part at the 46th Annual Meeting of the Orthopedic
Research Society, Orlando, FL, USA, March 2000.
Dr. Cynthia Luppen, Dr. Cara Blake, Dr. Krista Ammirati, Dr.
Mary Stevens, Dr. Howard Seeherman, Dr. John Wozney, and Dr.
Several mechanisms, both indirect and direct, have been
proposed to account for the effects of glucocorticoid therapy
on bone. Glucocorticoid therapy indirectly influences bone
Mary Bouxsein have financial interests in the forms of corporate metabolism by altering calcium homeostasis and secretion
(5)
appointments and stock ownership.
of sex hormones. In particular, pharmacologic doses of
1
Bone Biology and Applications, Genetics Institute-Wyeth Ayerst Research, Cambridge, Massachusetts, USA.
2
Orthopedic Biomechanics Laboratory, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts, USA.
3
01

3
02
LUPPEN ET AL.
glucocorticoids lead to reduced calcium absorption in the
gut, secondary hyperparathyroidism, and decreased secre-
tion of anabolic hormones. The direct actions of glucocor-
ticoids on bone cells lead to inhibition of bone formation,
MATERIALS AND METHODS
Study design
Forty-nine skeletally mature, male New Zealand White
(6)
suppression of osteoblastogenesis, increased bone resorp- rabbits (Ͼ8 months old, 3.5–4.0 kg) were assigned to two
tion, and enhanced apoptosis of osteoblasts and osteo- groups: prednisolone or control (saline). Bilateral, midulnar
(
7)
cytes. In addition, glucocorticoids influence local produc- osteotomies were created surgically in all rabbits. One limb
was treated with rhBMP-2/absorbable collage sponge
ACS; total dose ϭ 40 ␮g of rhBMP-2), whereas the con-
tion of growth factors and cytokines that may affect bone
(8)
(
formation and resorption. Taking these results all to-
gether, it appears that the net result of the endocrine and
paracrine responses to glucocorticoids is inhibition of os-
tralateral limb remained untreated. To simulate chronic glu-
cocorticoid therapy, the rabbits received systemic pred-
nisolone (or saline) therapy for 6 weeks before creation of
the osteotomy as well as throughout the healing period.
Rabbits in the prednisolone group received subcutaneous
injections of prednisolone three times a week (0.35 mg/kg
per day, totaling 1.05 mg/kg per week; Steris Laboratories,
Phoenix, AZ, USA), whereas those in the saline group
received subcutaneous injections of normal saline according
to the same schedule. This prednisolone dosing regimen was
shown previously to induce bone loss and inhibit fracture
(5)
teoblasts and activation of osteoclasts.
Consistent with their detrimental effects on bone density
and fracture risk, glucocorticoids also inhibit fracture heal-
(9)
ing. In the mid-1970s, Kostenszky et al. used autoradiog-
raphy and histology to show that administration of high-
dose glucocorticoids delayed mineralization of the fracture
callus in a dog model. More recent studies have shown that
systemic prednisolone therapy impairs fracture healing in
(
10,11)
rabbits.
The proposed mechanisms of glucocorticoid
healing in rabbits and to be tolerable for long-term admin-
action in bone that were outlined previously likely contrib-
ute to this impaired fracture healing. Specifically, the reduc-
tion in bone formation and inhibition of osteoblastogenesis
seen after glucocorticoid therapy could inhibit fracture heal-
ing. Moreover, localization of the functional glucocorticoid
receptor in human chondrocytes, osteoblasts, and osteocytes
suggests that all stages of endochondral bone formation,
including those normally involved in fracture healing, may
(10,11,27)
istration.
Rabbits were assigned randomly to be
killed at either 6 weeks (13 rabbits, prednisolone; 12 rabbits,
saline) or 8 weeks (13 rabbits, prednisolone; 11 rabbits,
saline) after creation of the osteotomies. These time points
were selected based on a previous study using the same
(23)
model in which the osteotomies were healed at 6 weeks.
Thus, the 8-week time point was chosen to account for the
potential of delayed healing in the prednisolone group. The
rabbits were monitored daily to assess general health and
(12)
be influenced by glucocorticoid therapy.
Bone morphogenetic protein-2 (BMP-2) is a potent neurological function and were weighed weekly to deter-
differentiation factor that is essential for embryonic skel- mine prednisolone (and saline) injection volumes. The rab-
etal development and for osteoblastogenesis in the bits were housed individually and fed standard rabbit chow
(
13,14)
adult.
In addition, BMP-2 is up-regulated during and water ad libitum. The protocol was approved by the
(
15,16)
fracture healing.
The primary action of BMP-2 is to Institutional Animal Care and Use Committee (IACUC) of
induce the differentiation of mesenchymal precursor cells the Genetics Institute and all procedures were carried out
into osteoblasts, although it also may play an indirect role according to Association for Assessment and Accreditation
of Laboratory Animal Care (AAALAC) guidelines.
in the chemotactic and proliferative behavior of these
(
17)
cells.
In several animal models, exogenous application
of recombinant human BMP-2 (rhBMP-2) has been Surgical procedure and general animal care
shown to promote healing of critical-sized, long bone
(
18–20)
Bilateral, midulnar osteotomies were created surgically in
all rabbits. The rabbits received injections of analgesics
defects 21–23a) nd to accelerate healing of diaphyseal
(
fractures.
(
Buprenex, 0.3 mg/kg twice a day [b.i.d.]; Recitt & Colman
Pharmaceuticals, Inc., Richmond, VA, USA) and antibiotics
Monocid, 25 mg/kg once a day [s.i.d.]; SmithKline Beed-
Despite the important clinical ramifications, few studies
have investigated the potential interaction between inhibi-
tion of bone formation by glucocorticoid therapy and stim-
ulation of bone formation by rhBMP-2 in vivo. Although in
vitro studies have shown that physiological levels of glu-
cocorticoids potentiate the proliferative effects of rhBMP-
(
nam, Philadelphia, PA, USA) before surgery and for 2 days
after surgery. Before surgery, rabbits were anesthetized by
intramuscular injections of ketamine (35 mg/kg) and xyla-
zine (5 mg/kg). Both forelimbs were shaved and prepped
with chlorhexidine scrub, 70% ethanol, and chlorhexidine
solution. After sedation and surgical preparation, the rabbit
was placed in a supine position and general anesthesia by
halothane/oxygen was administered. A posterior longitudi-
nal incision was made over the ulna and the underlying
muscles were retracted, exposing the middiaphyseal sur-
(
24,25)
2
and that the inhibition of osteoblastogenesis induced
by pharmacologic levels of glucocorticoids may be coun-
(26)
teracted by rhBMP-2,
it is not known whether glucocor-
ticoids will inhibit the stimulatory effects of rhBMP-2 on
bone formation in vivo. Therefore, this study was designed
to investigate this potential interaction between glucocorti-
coid therapy and rhBMP-2 in vivo. We used a rabbit ulnar faces of both the ulna and the radius. Care was taken to
osteotomy model to evaluate whether fracture healing is minimize damage to the periosteum. Osteotomies were cre-
inhibited by chronic glucocorticoid therapy and if so, ated approximately 45–50 mm distal to the olecranon pro-
whether rhBMP-2 can overcome this inhibition and enhance cess using a high-speed oscillating saw (0.5- to 1-mm blade
healing.
width defect). During the osteotomy procedure, the site was

rhBMP-2, PREDNISOLONE, AND FRACTURE HEALING
303
irrigated continuously with saline containing 0.2 mg/ml histomorphometric analysis or left unstained for dynamic
gentamicin sulfate to minimize thermal damage. In each histomorphometric analyses. Dynamic histomorphometry
rabbit, one ulna was treated with rhBMP-2 and the con- measurements were performed for the rabbits killed at the
tralateral ulna was left untreated. Previously, we have 6-week time point only. Trabecular perimeter, trabecular
shown that healing of osteotomies treated with an ACS that area, and total area were measured using KSS stereology
has been soak-loaded with buffer is equivalent radiograph- software (KSS, Salt Lake City, UT, USA). Static parameters
ically, histologically, and biomechanically to untreated os- (trabecular bone volume [BV/TV, %], trabecular thickness
(
23)
teotomies.
Thus, in this study, we chose to use an un- [Tb.Th, ␮m], trabecular separation [Tb.Sp, ␮m], and tra-
treated osteotomy for the control limb.
The rhBMP-2 (200 ␮l of 0.2 mg/ml of rhBMP-2) was model.
becular number [Tb.N]) were calculated based on a plate
(29)
Label length and double label width were mea-
applied to an 8 mm ϫ 20 mm strip of ACS (Integra Life sured using Osteomeasure software (Scimeasure Analytical
Sciences Corp., Plainsboro, NJ, USA) at least 30 minutes Systems, Inc., Atlanta, GA, USA) and dynamic indices
before implantation. Then, the rhBMP-2/ACS implant was (mineral apposition rate [MAR, ␮m/day], mineralizing
placed on the palmar aspect of the ulna and wrapped cir- surface [MS, %], and bone formation rate [BFR, %/year])
cumferentially around the lateral and part of the dorsal were calculated according to methods described previous-
(29)
aspect of the ulna. The medial aspect of the ulna could not ly. The individual performing the histological evaluation
be reached because of the position of the radius. Therefore, was blinded to the study design and treatment groups.
the rhBMP-2/ACS implant surrounded approximately two-
thirds of the ulna. After ensuring adequate hemostasis, the
wound was closed in layers. The rabbit was taken into
Evaluation of osteotomy healing
recovery, placed in lateral recumbency, and monitored until
Osteotomy healing was evaluated in vivo using antero-
the righting reflex returned. All rabbits were ambulatory posterior and lateral radiographs. Postmortem assessments
after surgery. Because the osteotomy was stabilized by the included peripheral quantitative computed tomography
radius, no casts or splints were used. Fluorochrome labels (pQCT) to assess the geometry and density of the mineral-
were administered 10 days after surgery (tetracycline, 25 ized callus, torsional biomechanics to assess the mechanical
mg/kg) and 13 days and 3 days before death (calcein, 5 properties of the healing osteotomy, and undecalcified his-
mg/kg). Ulnas from 10 rabbits per group were designated tology.
for biomechanical testing, and 2 representative ulnas from
Anteroposterior and lateral radiographs (40 kVp, 200
each group were selected for histological evaluation. To mA, 0.2 s; Kodak Ektascan M film; Kodak, Rochester, NY,
choose the ulna for histological evaluation, we reviewed the USA) were taken immediately postoperatively and weekly
final set of in vivo radiographs and ranked the rabbits in thereafter. The radiographs were evaluated by an indepen-
each group according to the extent of healing. The two dent reader who was blinded to the study design and treat-
rabbits that ranked in the middle of each group with respect ment groups. We recorded the onset of callus mineralization
to healing were chosen for histological evaluation. This and the week that each of the four cortices first appeared
approach ensured that, for our qualitative histological eval- bridged with bone. When three of four cortices were
uation, we chose ulnas that represented the average healing bridged, the ulna was designated as “radiographically
of the group. After euthanasia, the rabbit forelimbs were bridged.” An ulna was designated as having achieved “ra-
cleared carefully of excess tissue and stored frozen at diographic union” when all four cortices were bridged, there
Ϫ20°C for biomechanical testing or fixed in 70% ethanol was a radiopaque callus, and the osteotomy line was no
for histology.
longer evident.
To determine the density (mg/cm ), bone mineral content
3
2
(
g), and area (mm ) of the mineralized callus, the excised
Evaluation of the systemic effect of prednisolone
treatment
limbs were scanned using pQCT (XCT3000; Stratec, Pfor-
zheim, Germany). For each forelimb, we acquired five con-
To assess the systemic effect of prednisolone treatment, tiguous slices through the fracture callus, perpendicular to
we monitored bone formation via serum osteocalcin mea- the diaphyseal axis. The slices were 2.2 mm thick, 2 mm
surements. Serum samples were obtained at 0, 3, 6 (osteot- apart, and had an in-plane pixel size of 0.2 ϫ 0.2 mm. The
omy surgery), and 12 weeks after the initiation of pred- three slices surrounding the osteotomy, including a middle
nisolone or saline treatment. Serum osteocalcin was slice that was closest to the osteotomy, were analyzed. To
determined using a competitive immunoassay that binds analyze each slice, a region of interest was drawn around the
osteocalcin in serum using a monoclonal antiosteocalcin ulna and its callus, excluding the radius. The first analysis
3
antibody (NovoCalcin; Metrabiosystems, Mountain View, used a low bone density threshold (375 mg/cm ) to measure
CA, USA).
the properties of the ulna and callus. Then, a second analysis
3
In addition, we performed static and dynamic histomor- was performed using a high threshold (750 mg/cm ) to
phometry of vertebral trabecular bone. At death, the sixth measure the properties of the ulna only. Combining the data
lumbar vertebral body was dissected free from surrounding from these two analyses yielded the area, mineral content,
soft tissues and placed in 70% ethanol for fixation. Unde- and density of the fracture callus itself. Average values from
calcified specimens were processed through alcohol gradi- the three slices are reported.
ents, embedded in methylmethacrylate (MMA), and sec-
Ulnas from 10 rabbits per group were tested biomechani-
(28)
tioned in the midsagittal plane.
Thin sections (8 ␮m) cally using a servohydraulics materials testing system
(28)
were either stained using a modified von Kossa for static (model 8500; Instron Corp., Canton, MA, USA). The fore-

3
04
LUPPEN ET AL.
limbs were cleared carefully of soft tissue and then embed-
ded for testing in square molds using poly-MMA (PMMA).
Once embedded, a high-speed saw was used to cut through
the radius so that it could be gently pried away from the
ulna. Care was taken to not disturb the callus or ulna. After
embedding, the ulnas were tested to failure in torsion at a
constant loading rate (1.5°/s) and torque and rotation data
were recorded at 50 Hz. Failure torque (N-m), torsional
stiffness (N-m/deg), and energy absorbed to failure (N-m/
deg) were computed from the torque versus rotation data.
For those specimens with no clear failure, the maximum
rotation and energy-to-failure were computed for a displace-
ment up to 35°. Faxitron radiographs of the ulnas were
performed after mechanical testing and used to determine
the location of the fractures according to a modification of
(30)
the guidelines established by White et al. Thus, the ulnas
were classified as having fractured either through the orig-
inal osteotomy site, through a combination of the original
osteotomy site and intact host bone, or through the intact
host bone only.
FIG. 1. Serum osteocalcin (ng/ml; mean Ϯ SEM) in prednisolone-
and saline-treated rabbits over 12 weeks of treatment. ϩPrednisolone Ͻ
saline (p Ͻ 0.0001); *significant change from previous time point (p Ͻ
Based on radiographic observations, ulnas from two rep-
resentative rabbits per group were excluded from biome-
chanical testing and were instead evaluated for callus com- 0.05).
position, defect bridging, and cortical continuity using
qualitative undecalcified histology. After fixation in 70%
ethanol, the specimens were processed through alcohol gra-
dients, embedded in MMA, and sectioned in a sagittal plane
through the center of the diaphysis and region of interest.
Sections (5 ␮m thick) were placed on slides and stained
RESULTS
The surgical procedure was well tolerated. The rabbits
were fully weight bearing immediately after surgery and
were freely ambulating shortly thereafter. Throughout the
study, the rabbits displayed normal eating and behavior
patterns. Despite this, all rabbits lost weight during the
course of the study, with the majority of the weight loss
with Goldner’s trichrome.
Data analysis
Standard descriptive statistics were computed for all out- (approximately 200 g) occurring the week after surgery.
come variables. Analysis of variance (ANOVA) was used to Over the course of the study, the average weight loss in the
determine the effect of endpoint (6 weeks or 8 weeks). prednisolone-treated rabbits was greater than that in the
Because this analysis suggested that the effects of pred- saline-treated rabbits (390 Ϯ 200 g vs. 220 Ϯ 160 g; p ϭ
nisolone and rhBMP-2/ACS treatments were similar in the 0.01). Because rabbits are expected to lose weight after
6
-week and 8-week cohorts, results from the two time points surgery and there was large intragroup variation in the
were combined for all subsequent analyses, except for the weights, we consider the difference in weight loss between
qualitative histology evaluations. The systemic effect of the prednisolone and saline groups to be a minor finding.
prednisolone exposure was evaluated by comparing histo-
morphometry of vertebral trabecular bone using a Student’s
t-test and serum osteocalcin levels using an ANOVA with
Effects of prednisolone on bone metabolism and
vertebral trabecular bone histomorphometry
repeated measures in the prednisolone and saline groups.
Osteocalcin levels in the prednisolone group were 42–
The effect of prednisolone treatment on osteotomy healing 72% lower than those in the saline group as early as 3 weeks
was evaluated by comparing properties of the untreated after initiation of prednisolone treatment and continuing
limbs in the prednisolone and saline groups by Student’s throughout the study (p Ͻ 0.0001 for all; Fig. 1). In the
2
t-tests (for pQCT and biomechanics) or ␹ analyses (for prednisolone group, osteocalcin levels were 62–72% below
radiographic union and fracture location). The effect of baseline levels at all time points after initiation of pred-
rhBMP-2/ACS on fracture healing was evaluated using an nisolone treatment (p Ͻ 0.0001 for all). In the saline group,
ANOVA with repeated measures and paired t-tests (for osteocalcin levels were 22% and 38% lower than baseline at
2
pQCT and biomechanics) as well as ␹ analyses for assess- 6 weeks and 12 weeks, respectively, after initiation of saline
ment of radiographic union. Data are presented as means Ϯ injections (p Ͻ 0.0005 for both).
SD unless otherwise noted. All tests were two-tailed and
Both static and dynamic histomorphometric variables
differences were considered significant at p Ͻ 0.05. Post from vertebral trabecular bone were consistently lower in
hoc testing was done using Fisher’s protected least squares the prednisolone group, although the differences did not
difference test. Statistical analyses were done using Stat- always reach statistical significance (Table 1). Trabecular
view 5.0 (SAS Institute, Inc., Cary, NC, USA) and Super- BV/TV was 11.5% lower (p ϭ 0.07) and Tb.Th was 7.6%
ANOVA (SAS Institute, Inc.).
lower (p ϭ 0.03) in the prednisolone-treated rabbits com-

rhBMP-2, PREDNISOLONE, AND FRACTURE HEALING
305
TABLE 1. STATIC AND DYNAMIC HISTOMORPHOMETRY OF
weeks, the osteotomies in the untreated ulnas of the saline
group were bridged with bone and isolated regions of car-
tilage. In comparison, at this time point, the osteotomies in
the prednisolone group were bridged with a combination of
cartilage, fibrocartilage, and loose fibrous tissue. The over-
all callus size, including both the mineralized and the un-
mineralized portions of the callus, did not appear to be
THE LUMBAR VERTEBRAE (MEAN Ϯ SD)
Variable
Prednisolone
Saline
BV/TV (%)
Tb.N (mm
Tb.Th (␮)
Tb.Sp (␮)
17.7 Ϯ 3.4
1.38 Ϯ 0.19
127.8 Ϯ 15.3
609.9 Ϯ 118.7
1.24 Ϯ 0.20
22.9 Ϯ 10.4
139.1 Ϯ 70.8
20.0 Ϯ 5.1*
1.43 Ϯ 0.24
138.4 Ϯ 16.8
Ϫ1
†
)
577.6 Ϯ 125.7 affected by prednisolone treatment. Taken together, these
1.34 Ϯ 0.22
28.8 Ϯ 9.5
a
MAR (␮m/day)
histological observations suggest that the primary effect of
prednisolone treatment was to inhibit the overall maturation
of the callus and to delay endochondral bone formation.
a
MS (%)
a
BFR (%/year)
185.9 Ϯ 77.7
Note that results from the 6- and 8-week time points have been
combined.
Effects of rhBMP-2/ACS on osteotomy healing
a
Dynamic histomorphometry only available for animals treated
The effect of rhBMP-2/ACS on osteotomy healing was
determined by comparing the ulnas treated with rhBMP-2/
ACS with their contralateral untreated ulnas in both the
prednisolone and the saline groups. Radiographic, densito-
metric, biomechanical, and histological results showed that
rhBMP-2/ACS enhanced osteotomy healing in both groups,
with prednisolone for 12 weeks.
*
Greater than prednisolone-treated rabbits ( p
Ͻ
0.10);
†
Greater than prednisolone-treated rabbits ( p Ͻ 0.05).
pared with those treated with saline. In addition, the MAR, although the effect was larger in the prednisolone group.
percent MS, and BFR were 7.7, 25.2, and 20.5% lower, Poor radiographic healing was observed in the untreated
respectively, in the prednisolone group, although these dif- ulnas in both the prednisolone and the saline groups, in
ferences did not reach statistical significance (Table 1).
which only 16% (8/49) of the limbs achieved radiographic
union. In contrast, 80% (39/49) of the rhBMP-2/ACS-
treated ulnas achieved radiographic union (␹ ϭ39.3; p Ͻ
2
Effects of prednisolone on osteotomy healing
0
.0001 for prednisolone and saline groups combined, Table
Radiographic, densitometric, biomechanical, and histo- 2). The number of limbs achieving radiographic union was
logical comparisons of the untreated ulnas in the pred- similar in the rhBMP-2–treated limbs in the prednisolone-
nisolone and saline groups showed that prednisolone treat- and saline-treated groups. Healing was accelerated with
ment dramatically inhibited osteotomy healing. Assessment rhBMP-2/ACS treatment, because a majority of the ulnas
of the in vivo radiographs showed that none of the untreated treated with rhBMP-2/ACS showed mineralized callus as
ulnas in the prednisolone group achieved radiographic early as 2 weeks (81% prednisolone and 91% saline) and
union, whereas 8/23 (35%) of untreated ulnas in the saline showed callus bridging at 3 weeks (58% prednisolone and
2
group achieved radiographic union (␹ ϭ 10.8; p ϭ 0.001; 65% saline). In comparison, few of the untreated ulnas
Table 2). The area, mineral content, and density of the showed callus formation at 2 weeks (12% prednisolone and
mineralized callus in the untreated ulnas of rabbits treated 22% saline) and none were bridged with bone at 3 weeks
with prednisolone were 25, 28, and 3.5% lower, respec- (Table 2).
tively, than corresponding values from untreated ulnas in
rabbits treated with saline (p Ͻ 0.001 for all; Table 3; Fig. behavior of the healing osteotomy were enhanced by
). Moreover, the torsional strength, stiffness, and energy rhBMP-2/ACS treatment in both the saline and the pred-
Properties of the mineralized callus and biomechanical
2
absorbed to failure were 55–57% lower in the untreated nisolone groups; however, the magnitude of improvement
ulnas of the prednisolone-treated rabbits compared with was greater in the prednisolone group (Table 3). In the
those receiving saline (p Ͻ 0.0001 for all; Table 3; Fig. 3). prednisolone-treated rabbits, the area, mineral content, and
During postmortem torsion testing, only 13.6% (3/22) of the density of the mineralized callus were 40, 43, and 2.7%
untreated ulnas in the prednisolone group fractured outside greater, respectively, in ulnas treated with rhBMP-2/ACS
the original osteotomy in the intact host bone, whereas the than the contralateral untreated ulnas (p Ͻ 0.0005 for all;
remainder of ulnas fractured through the original osteotomy Table 3; Fig. 2). In the prednisolone-treated rabbits, ulnas
site. In comparison, 55% (11/20) of untreated ulnas in the treated with rhBMP-2/ACS had greater torsional strength
saline group fractured outside the original osteotomy (ϩ165%), stiffness (ϩ156%) and energy-to-failure (ϩ199%)
2
(
␹ ϭ8.1; p ϭ 0.005).
compared with the untreated ulnas (p Ͻ 0.0001 for all;
The histological evaluation, although limited to two spec- Table 3; Fig. 4). Moreover, 82% (18/22) of limbs treated
imens per group, confirmed the inhibition of healing that with rhBMP-2/ACS fractured outside the original osteot-
was observed radiographically and biomechanically in the omy site compared with only 14% (3/22) of the contralateral
2
prednisolone-treated rabbits. Compared with the untreated untreated limbs (␹ ϭ 20.5; p Ͻ 0.0001).
ulnas in the saline group, those in the prednisolone group
In comparison, in the saline-treated rabbits, ulnas treated
exhibited less extensive bridging of the osteotomy and a less with rhBMP-2/ACS had greater callus area and bone min-
mature callus (Fig. 4). For example, at 6 weeks, the osteot- eral content (ϩ17% for both; p Ͻ 0.02), but similar callus
omies in the untreated ulnas of the saline group were density as the contralateral untreated ulnas (Table 2, Fig. 2).
bridged primarily with cartilage, whereas those in the pred- In the saline group, compared with untreated ulnas, those
nisolone group were bridged with fibrocartilage. At 8 treated with rhBMP-2/ACS had greater torsional strength

3
06
LUPPEN ET AL.
TABLE 2. EVALUATION OF IN VIVO RADIOGRAPHS AT 2, 3, 6, AND 8 WEEKS AFTER OSTEOTOMY
Saline
Prednisolone
Untreated
rhBMP-2/ACS
Untreated
rhBMP-2/ACS
‡
‡
Callus Formation at 2 weeks
Callus Bridging at 3 weeks
22% (5/23)
0% (0/23)
52%* (12/23)
35%* (8/23)
91% (21/23)
12% (3/26)
0% (0/26)
12% (3/26)
0% (0/26)
81% (21/26)
‡
‡
65% (15/23)
58% (15/26)
a
†
‡
Radiographic bridging
100% (23/23)
92% (24/26)
a
‡
‡
Radiographic union
83% (19/23)
77% (20/26)
a
Refers to whether the osteotomy was identified to be bridged or united at any time.
2
†
‡
*
Greater than prednisolone untreated ( p Ͻ 0.005 by ␹ analysis); greater than untreated contralateral p Ͻ 0.005 and p Ͻ
2
0
.0001 by ␹ analysis.
TABLE 3. GEOMETRY AND DENSITY OF THE MINERALIZED CALLUS AND BIOMECHANICAL PROPERTIES OF THE HEALING
OSTEOTOMY (MEAN Ϯ SD)
Saline
Prednisolone
rhBMP-2/ACS
Untreated
rhBMP-2/ACS
Untreated
pQCT evaluation
Area (mm )
Mineral content (mg)
Density (mg/cm )
2
‡
§
14.4 Ϯ 4.2*
8.32 Ϯ 2.57*
575 Ϯ 18*
16.9 Ϯ 5.3
10.8 Ϯ 2.9
5.98 Ϯ 1.59
553 Ϯ 9
15.1 Ϯ 4.4
§
‡
9.76 Ϯ 3.05
8.56 Ϯ 2.30
3
§
577 Ϯ 15
568 Ϯ 18
Torsion testing
Failure torque (N-m)
Torsional stiffness (N-m/deg)
Energy-to-failure (N-m/deg)
†
†
‡
§
§
0.465 Ϯ 0.166
0.021 Ϯ 0.009
6.24 Ϯ 3.11
0.584 Ϯ 0.162
0.024 Ϯ 0.008
8.38 Ϯ 2.55
0.209 Ϯ 0.156
0.009 Ϯ 0.008
2.81 Ϯ 2.20
0.554 Ϯ 0.189
0.023 Ϯ 0.007
8.41 Ϯ 3.68
†
‡
§
Note that results from the 6- and 8-week time points have been combined.
Greater than prednisolone untreated * p Ͻ 0.005 and p Ͻ 0.0005; greater than untreated contralateral p Ͻ 0.05 and p Ͻ
†
‡
§
0
.0001.
FIG. 2. Area (left), bone mineral content (middle), and density (right) of the mineralized callus in the untreated and rhBMP-2–treated ulnas of
prednisolone and saline groups. Error bars represent 1 SD. Results from the 6-week and 8-week cohorts were combined. ϩprednisolone Ͻ saline
(
p Ͻ 0.001); *rhBMP-2/ACS Ͼ untreated (p Ͻ 0.05); **rhBMP-2 Ͼ untreated (p Ͻ 0.0005).
(ϩ26%; p Ͻ 0.05), energy-to-failure (ϩ34%; p Ͻ 0.05), ACS-treated ulnas relative to the untreated ulnas, as the
and torsional stiffness (ϩ14%, NS; Table 2; Fig. 3). Finally, periosteal calluses were undergoing remodeling and the
in these saline-treated rabbits, 84% (16/19) of limbs treated endosteal calluses were being resorbed by osteoclasts. Sim-
with rhBMP-2/ACS fractured outside the original osteot- ilar remodeling and resorption of the callus was not ob-
omy site, compared with 55% (11/20) of the contralateral served in the untreated osteotomies.
2
untreated limbs (␹ ϭ 3.9; p Ͻ 0.05).
Histological evaluation confirmed that rhBMP-2/ACS
treatment dramatically enhanced healing of the osteotomies
in both the prednisolone and the saline groups (Fig. 4). At
both the 6-week and the 8-week time points, the ulnas
DISCUSSION
In this study we confirmed previous findings that chronic
treated with rhBMP-2/ACS were fully bridged with bone. administration of prednisolone inhibits osteotomy healing in
(10,11)
At 6 weeks, the periosteal calluses of the rhBMP-2/ACS rabbits.
In addition, we showed that a single, local
treated ulnas were larger than those left untreated. By 8 application of rhBMP-2/ACS was able to overcome this
weeks, the healing process was advanced in the rhBMP-2/ inhibition and enhance osteotomy healing. For example, in

rhBMP-2, PREDNISOLONE, AND FRACTURE HEALING
307
FIG. 3. Failure torque (left), torsional stiffness (middle), and energy absorbed-to-failure (right) of the untreated and rhBMP-2–treated ulnas of
prednisolone and saline groups. Error bars represent 1 SD. Results from the 6-week and 8-week cohorts were combined. ϩPrednisolone Ͻ saline
(
p Ͻ 0.001); *rhBMP-2/ACS Ͼ untreated (p Ͻ 0.05); **rhBMP-2 Ͼ untreated (p Ͻ 0.0001).
FIG. 4. Representative histo-
logical sections of the healing os-
teotomy in (A and B) saline and
(
C and D) prednisolone-treated
rabbits 6 weeks after creation of
the osteotomy (Goldner’s
trichrome stain, 1ϫ). Note the
presence of cartilage in the os-
teotomy gap in the untreated
limbs of the (A) saline and (C)
prednisolone rabbit, whereas (B
and D) the limbs treated with
rhBMP-2/ACS were completely
bridged with bone.
the prednisolone-treated rabbits, 65% of the osteotomies in which the healing environment is compromised such as
treated with rhBMP-2/ACS achieved radiographic union, nicotine exposure, alcohol abuse, and diabetes.
whereas none of the contralateral untreated osteotomies
Based strictly on a per body weight basis, the dose of
achieved radiographic union after 6 weeks of healing. More- prednisolone administered to the rabbits is equivalent to 10
over, in rabbits that received prednisolone, the biomechani- mg/day of prednisone (assuming a body weight of 68 kg or
cal properties of osteotomies treated with rhBMP-2/ACS, 150 lbs). Ebling et al. reported that rabbits and humans have
(31)
were at least 2-fold greater than those of the untreated similar plasma binding of methylprednisolone.
More-
osteotomies. These results show the dramatic positive effect over, after adjustment for body weight, the mean residence
of rhBMP-2/ACS on fracture healing. Although previous time and half-life of prednisolone was similar in rabbits and
studies also have shown the ability of rhBMP-2/ACS to humans. Thus, given the high bioavailability of oral pred-
(22,23)
(32,33)
enhance fracture healing,
to our knowledge, this is the nisone administration in humans,
it appears that the
first study that shows the ability of rhBMP-2/ACS to over- prednisolone dose administered to the rabbits in this study is
come the inhibition of healing associated with glucocorti- approximately comparable with a 10-mg/day oral pred-
coid therapy and, furthermore, to enhance healing in this nisone dose in humans. This dose is used commonly for
compromised environment. These data have important clin- treatment of respiratory, immune, skin, and musculoskeletal
(34)
ical ramifications because individuals undergoing chronic disorders.
glucocorticoid therapy are both at increased risk of fracture Our results showing negative effects of glucocorticoid
and increased risk of poor healing of their fractures. Fur- administration on fracture healing concur with previously
thermore, these data suggest that rhBMP-2/ACS also may published findings. In the 1970s, studies in rabbits provided
be able to enhance bone healing in other clinical situations histological evidence of retardation at all stages of fracture

3
08
LUPPEN ET AL.
healing after administration of high doses of cortisone
The role of BMP-2 in fracture healing has been described
(35,36)
(47,48)
(10–25 mg/kg per day).
More recently, an inhibition previously.
It is believed that during normal fracture
of osteotomy healing was observed in rabbits that received healing, BMP-2 is released from the bone matrix and peri-
prednisolone for 2 months before creation of the osteotomy osteum and that it stimulates the differentiation of mesen-
(
10,11)
and also throughout the healing period.
were similar to those of Waters et al.,
Our results chymal cells into osteoblasts and chondrocytes. The in-
who reported that creased osteoblast differentiation attributed to BMPs has
radiographic union was achieved in only 3 of 20 limbs in been found to act partly, although not completely, through
(10)
(49)
prednisolone-treated rabbits, whereas 13 of 16 limbs from the up-regulation of Cbfa1.
Application of exogenous
saline-treated rabbits achieved radiographic union at 6 rhBMP-2 likely enhances fracture healing via the same
weeks. Moreover, our observation of a 55–60% decrease in mechanisms (i.e., increased differentiation of preosteoblasts
the biomechanical properties of the healing osteotomies in and up-regulation of Cbfa1) and induction of bone forma-
prednisolone-treated rabbits compared with the saline- tion. Because expression of Cbfa1 is inhibited after glu-
(
40)
treated rabbits was nearly identical to the 60–65% decrease cocorticoid exposure,
either rhBMP-2 overcomes this
(10)
previously reported.
In addition to effects on fracture healing, we observed are independent of Cbfa1.
negative effects of prednisolone on bone metabolism in the Our study cannot explain the molecular mechanisms un-
first months of treatment. Serum osteocalcin levels were derlying the effects of glucocorticoids on fracture healing or
2% lower than baseline levels in the prednisolone-treated the potential interaction between glucocorticoids and
inhibition or it acts at least partially through pathways that
(26,50)
7
rabbits at the time of osteotomy (i.e., after the 6-week rhBMP-2. However, it provides an interesting framework
pretreatment period). Serum osteocalcin was reduced also for future studies. Among the most important considerations
(Ϫ38%) in the saline-treated rabbits at the time of osteot- in examining these results are the timing and duration of
omy. It is not clear why osteocalcin levels declined in the prednisolone exposure and the profound positive effects of
control rabbits. However, we believe that this gradual de- the rhBMP-2/ACS treatment. The prednisolone treatment
cline may be attributable to a stress response associated with began 6 weeks before the osteotomy surgery and continued
the frequent handling for weighing and injections. Never- throughout the healing period. In comparison, the rhBMP-
theless, the decrease in osteocalcin was greater in the 2/ACS was applied only once, at the time the osteotomy was
prednisolone-treated rabbits than in the saline-treated rab- created. It should be noted that although the rhBMP-2/ACS
bits at all time points. In addition to the reduction in serum was applied at a single time point, Bouxsein and colleagues
osteocalcin, we also observed a reduction in vertebral tra- have shown that small amounts of rhBMP-2 may remain at
(23)
becular bone indices. Fourteen weeks after the initiation of the osteotomy site up to 2 weeks after implantation.
prednisolone treatment, there was a significant reduction in Nonetheless, the prednisolone therapy had several weeks to
vertebral trabecular bone density and trabecular thickness in establish its deleterious effects before BMP-2 was intro-
the prednisolone-treated rabbits. In addition, dynamic indi- duced, and therefore rhBMP-2 was acting in an environment
ces of bone formation were lower in the rabbits receiving with altered bone metabolism, as indicated by the dramatic
prednisolone, although the differences did not reach statis- reduction in osteocalcin levels at the time of osteotomy.
tical significance. This decrease in trabecular bone integrity Despite this, the rhBMP-2/ACS restored the osteotomies in
is consistent with reports of the rapid bone loss seen during the prednisolone-treated rabbits to the same degree of en-
the first few months of glucocorticoid treatment in hu- hanced radiographic and biomechanical healing as was seen
(
3,8)
mans.
in the saline-treated rabbits that received rhBMP-2/ACS.
These in vivo studies confirm that glucocorticoid therapy Based on this study, we can only suggest possible explana-
inhibits bone healing. However, the mechanisms by which tions for this enhancement of fracture healing after rhBMP-
glucocorticoid therapy inhibits fracture healing are not yet 2/ACS treatment in rabbits exposed to prednisolone.
defined. Indications of these mechanisms may be gained
One possible explanation for the positive effect of
from previous studies on the deleterious bone-related effects rhBMP-2 in this compromised healing environment may be
of glucocorticoids given at supraphysiological concentra- that the creation of the osteotomy initiated an injury re-
tions. Both in vitro and in vivo observations suggest that the sponse with recruitment of sufficient mesenchymal cells to
primary effects of supraphysiological doses of glucocorti- overwhelm the glucocorticoid inhibition of osteoblast activ-
coids on bone are to decrease osteoblastogenesis and osteo- ity. The differentiation of these recruited mesenchymal cells
blast differentiation and enhance apoptosis of existing os- would have to have been initiated by the pharmacologic
(7,37–39)
teoblasts and osteocytes.
The inhibitory effects of doses of rhBMP-2, because endogenous BMP-2 in the limbs
glucocorticoids on fracture healing also may be mediated by of the prednisolone-treated rabbits that did not receive
(40)
suppression of the osteoblast transcription factor Cbfa1,
rhBMP-2/ACS was not sufficient to overcome the inhibition
attenuation of type I collagen synthesis and messenger RNA associated with prednisolone exposure. Because the pred-
(41–43)
(
mRNA) levels,
insulin-like growth factor 1 (IGF-1),
attenuation of the osteoblast cell cycle.
decreased synthesis and activity of nisolone treatment continued throughout the healing pro-
(41,44,45)
and premature cess, previous studies would indicate that bone formation by
Thus, it is likely these newly formed osteoblasts would be inhibited. Thus,
(46)
that glucocorticoids negatively influence fracture healing by either the newly differentiated osteoblasts were resistant to
affecting various cells involved in the healing process, in- subsequent glucocorticoid exposure or they were present in
cluding osteoblast precursors, preosteoblasts, mature osteo- sufficient number to enhance healing, despite an inhibition
blasts, and osteocytes.
in activity. The precise role of rhBMP-2 in overcoming the

rhBMP-2, PREDNISOLONE, AND FRACTURE HEALING
309
glucocorticoid-mediated inhibition of specific osteoblast
functions in fracture healing requires additional study.
In summary, we have shown that a single application of
rhBMP-2/ACS not only enhances osteotomy healing, but
also overcomes inhibition of healing in rabbits exposed
chronically to prednisolone. Therefore, rhBMP-2/ACS may
prove useful for patients undergoing chronic glucocorticoid
Essential requirement of BMPs-2/4 for both osteoblast and
osteoclast formation in murine bone marrow cultures from
adult mice: Antagonism by noggin. J Bone Miner Res 15:663–
6
73.
1
5. Bostrom MP, Lane JM, Berberian WS, Missri AA, Tomin E,
Weiland A, Doty SB, Glaser D, Rosen VM 1995 Immunolo-
calization and expression of bone morphogenetic proteins 2
and 4 in fracture healing. J Orthop Res 13:357–367.
therapy, because they are not only at increased risk of 16. Onishi T, Ishidou Y, Nagamine T, Yone K, Imamura T, Kato
M, Sampath TK, ten Dijke P, Sakou T 1998 Distinct and
fracture, but also are at increased risk of delayed healing or
overlapping patterns of localization of bone morphogenetic
nonunion.
protein (BMP) family members and a BMP type II receptor
during fracture healing in rats. Bone 22:605–612.
1
1
1
7. Rosen V, Wozney J 2000 Bone morphogenetic proteins and
the adult skeleton. In: Canalis E (ed.) Skeletal Growth Factors.
Lippincott Williams & Wilkins, Philadelphia, PA, USA, pp.
ACKNOWLEDGMENTS
We would like to thank the Laboratory Animal Resources
Group at the Genetics Institute for assistance with the study.
We also thank Dr. Rod Riedel and Dr. Baruch Frenkel for
reviewing the manuscript.
2
99–310.
8. Gerhart TN, Kirker-Head CA, Kriz MJ, Holtrop ME, Hennig
GE, Hipp J, Schelling SH, Wang E 1993 Healing segmental
femoral defects in sheep using recombinant human bone mor-
phogenetic protein. Clin Orthop:317–326.
9. Yasko AW, Lane JM, Fellinger EJ, Rosen V, Wozney JM,
Wang EA 1992 The healing of segmental bone defects, in-
duced by recombinant human bone morphogenetic protein
REFERENCES
(
rhBMP-2). A radiographic, histological, and biomechanical
1
2
. Lukert B 1996 Glucocorticoid-induced osteoporosis. In: Mar-
cus R (ed.) Osteoporosis. Academic Press, San Diego, CA,
USA, pp. 801–820.
study in rats. J Bone Joint Surg Am 74:659–670.
0. Bostrom M, Lane J, Tomin E, Browne M, Berberian W, Turek
T, Smith J, Wozney J, Schildhauer T 1996 Use of bone
morphogenetic protein-2 in the rabbit ulnar nonunion model.
Clin Orthop 327:272–282.
1. Bax BE, Wozney JM, Ashhurst DE 1999 Bone morphogenetic
protein-2 increases the rate of callus formation after fracture of
the rabbit tibia. Calcif Tissue Int 65:83–89.
2. Welch RD, Jones AL, Bucholz RW, Reinert CM, Tjia JS,
Pierce WA, Wozney JM, Li XJ 1998 Effect of recombinant
human bone morphogenetic protein-2 on fracture healing in a
goat tibial fracture model. J Bone Miner Res 13:1483–1490.
3. Bouxsein M, Turek T, Blake C, Seeherman H, Wozney J 2001
Recombinant human bone morphogenetic protein-2 enhances
healing in a rabbit ulnar osteotomy model. J Bone Joint Surg
Am 83:1219–1230.
2
. Dykman TR, Gluck OS, Murphy WA, Hahn TJ, Hahn BH
1
985 Evaluation of factors associated with glucocorticoid-
induced osteopenia in patients with rheumatic diseases. Arthri-
tis Rheum 28:361–368.
2
2
3
. Eastell R, Reid D, Compston J, Cooper C, Fogelman I, Francis
R, Hosking D, Purdie D, Ralston S, Reeve J, Russell R,
Stevenson J, Torgerson D 1998 A UK consensus group on
management of glucocorticoid-induced osteoporosis: An up-
date. J Intern Med 244:271–292.
4
. Van Staa TP, Leufkens HG, Abenhaim L, Zhang B, Cooper C
2
2
2
2
000 Use of oral corticosteroids and risk of fractures. J Bone
Miner Res 15:993–1000.
5
6
. Tannirandorn P, Epstein S 2000 Drug-induced bone loss. Os-
teoporos Int 11:637–659.
. Lukert B, Kream B 1996 Clinical and basic aspects of glu-
cocorticoid action. In: Bilezikian JP, Raisz LG, Rodan CA
4. Rickard DJ, Sullivan TA, Shenker BJ, Leboy PS, Kazhdan I
1
994 Induction of rapid osteoblast differentiation in rat bone
marrow stromal cell cultures by dexamethasone and BMP-2.
Dev Biol 161:218–228.
(eds.) Bone Principles of Bone Biology. Academic Press, San
Diego, CA, USA, pp. 533–548.
5. Boden SD, McCuaig K, Hair G, Racine M, Titus L, Wozney
JM, Nanes MS 1996 Differential effects and glucocorticoid
potentiation of bone morphogenetic protein action during rat
osteoblast differentiation in vitro. Endocrinology 137:3401–
3407.
26. Centrella M, Rosen V, Wozney JM, Casinghino SR, McCarthy
TL 1997 Opposing effects by glucocorticoid and bone mor-
phogenetic protein-2 in fetal rat bone cell cultures. J Cell
Biochem 67:528–540.
7
8
9
. Manolagas SC, Weinstein RS 1999 New developments in the
pathogenesis and treatment of steroid-induced osteoporosis.
J Bone Miner Res 14:1061–1066 (editorial).
. Lane NE, Lukert B 1998 The science and therapy of
glucocorticoid-induced bone loss. Endocrinol Metab Clin
North Am 27:465–483.
. Kostenszky K, Olah E 1974 Effect of prednisolone on callus
formation. Acta Biol Hung 25:49–60.
1
0. Waters RV, Gamradt SC, Asnis P, Vickery BH, Avnur Z, Hill
E, Bostrom M 2000 Systemic corticosteroids inhibit bone 27. Waters R, Vickery B, Avnur Z, Hirano T, Burr D 1997 Bone
healing in a rabbit ulnar osteotomy model. Acta Orthop Scand
1:316–321.
1. Bostrom MP, Gamradt SC, Asnis P, Vickery BH, Hill E,
structure and turnover following treatment of glucocorticoid
osteopenia in rabbits with RS-66271 abd bPTH(1–34). J Bone
Miner Res 12:S512.
7
1
Avnur Z, Waters RV 2000 Parathyroid hormone-related pro- 28. Schenk R, Olah A, Herrman W 1984 Preparation of calcified
tein analog RS-66271 is an effective therapy for impaired bone
healing in rabbits on corticosteroid therapy. Bone 26:437–442.
2. Abu EO, Horner A, Kusec V, Triffitt JT, Compston JE 2000
tissues for light microscopy. In: Dickson G (ed.) Methods of
Calcified Tissue Preparation. Elsevier Science Publication,
New York, NY, USA, pp. 1–42.
1
1
The localization of the functional glucocorticoid receptor al- 29. Parfitt A, Drezner M, Glorieux F, Kanis J, Recker R 1987
pha in human bone. J Clin Endocrinol Metab 85:883–889.
3. Rosen V, Cox K, Hattersly G 1996 Bone morphogenetic
Bone histomorphometry: Standardization of nomenclature,
symbols and units. J Bone Miner Res 2:595–610.
proteins. In: Bilezikian JP, Raisz LG, Rodan CA (eds.) Prin- 30. White AA, Panjabi MM, Southwick WO 1977 The four bio-
ciples of Bone Biology. Academic Press, San Diego, CA,
mechanical stages of fracture repair. J Bone Joint Surg Am
USA, pp. 661–671.
59:188–192.
1
4. Abe E, Yamamoto M, Taguchi Y, Lecka-Czernik B, O’Brien 31. Ebling W, Szefler S, Jusko W 1985 Methylprednisolone dis-
CA, Economides AN, Stahl N, Jilka RL, Manolagas SC 2000 position in rabbits: Analysis, prodrug conversion, reversible

3
10
LUPPEN ET AL.
metabolism, and comparison with man. Drug Metab Dispos 43. Kasperk C, Schneider U, Sommer U, Niethard F, Ziegler R
1
3:296–304.
1995 Differential effects of glucocorticoids on human osteo-
blastic cell metabolism in vitro. Calcif Tissue Int 57:120–126.
3
2. Ferry JJ, Horvath AM, Bekersky I, Heath EC, Ryan CF,
Colburn WA 1988 Relative and absolute bioavailability of 44. McCarthy TL, Centrella M, Canalis E 1990 Cortisol inhibits
prednisone and prednisolone after separate oral and intrave-
nous doses. J Clin Pharmacol 28:81–87.
the synthesis of insulin-like growth factor-I in skeletal cells.
Endocrinology 126:1569–1575.
3
3
3
3
3
3. Georgitis JW, Flesher KA, Szefler SJ 1982 Bioavailability 45. Swolin D, Brantsing C, Matejka G, Ohlsson C 1996 Cortisol
assessment of a liquid prednisone preparation. J Allergy Clin
decreases IGF-I mRNA levels in human osteoblast-like cells. J
Immunol 70:243–247.
Endocrinol 149:397–403.
4. Nishimura J, Ikuyama S 2000 Glucocorticoid-induced osteo- 46. Smith E, Redman RA, Logg CR, Coetzee GA, Kasahara N,
porosis: Pathogenesis and management. J Bone Miner Metab
8:350–352.
5. Blunt J, Plotz C, Lattes R, Howes E, Meyer K, Ragan C 1950
Effect of cortisone on experimental fractures in the rabbit. Proc
Soc Exp Biol 73:678–681.
Frenkel B 2000 Glucocorticoids inhibit developmental stage-
specific osteoblast cell cycle. Dissociation of cyclin A-cyclin-
dependent kinase 2 from E2F4–p130 complexes. J Biol Chem
275:19992–20001.
1
47. Einhorn TA 1998 The cell and molecular biology of fracture
healing. Clin Orthop S7–S21.
6. Sissons H, Hadfield G 1951 The influence of cortisone on the
repair of experimental fractures in the rabbit. Br J Surg 39: 48. Bostrom MP 1998 Expression of bone morphogenetic proteins
72–178. in fracture healing. Clin Orthop S116–S123.
7. Ishida Y, Heersche JN 1998 Glucocorticoid-induced osteopo- 49. Ducy P, Zhang R, Geoffroy V, Ridall AL, Karsenty G 1997
1
rosis: Both in vivo and in vitro concentrations of glucocorti-
coids higher than physiological levels attenuate osteoblast
differentiation. J Bone Miner Res 13:1822–1826.
Osf2/Cbfa1: A transcriptional activator of osteoblast differen-
tiation. Cell 89:747–754.
50. Komori T, Yagi H, Nomura S, Yamaguchi A, Sasaki K,
Deguchi K, Shimizu Y, Bronson RT, Gao YH, Inada M, Sato
M, Okamoto R, Kitamura Y, Yoshiki S, Kishimoto T 1997
Targeted disruption of Cbfa1 results in a complete lack of bone
formation owing to maturational arrest of osteoblasts. Cell
89:755–764.
3
8. Weinstein RS, Jilka RL, Parfitt AM, Manolagas SC 1998
Inhibition of osteoblastogenesis and promotion of apoptosis of
osteoblasts and osteocytes by glucocorticoids. Potential mech-
anisms of their deleterious effects on bone. J Clin Invest
1
02:274–282.
3
4
9. Weinstein RS, Nicholas RW, Manolagas SC 2000 Apoptosis
of osteocytes in glucocorticoid-induced osteonecrosis of the
hip. J Clin Endocrinol Metab 85:2907–2912.
0. Chang DJ, Ji C, Kim KK, Casinghino S, McCarthy TL, Cen-
trella M 1998 Reduction in transforming growth factor beta
receptor I expression and transcription factor CBFa1 on bone
cells by glucocorticoid. J Biol Chem 273:4892–4896.
1. Delany AM, Canalis E 1995 Transcriptional repression of
insulin-like growth factor I by glucocorticoids in rat bone cells.
Endocrinology 136:4776–4781.
Address reprint requests to:
Mary L. Bouxsein, Ph.D.
Orthopedic Biomechanics Laboratory
Beth Israel Deaconess Medical Center
330 Brookline Avenue
4
4
Boston, MA 02215, USA
2. Delany AM, Gabbitas BY, Canalis E 1995 Cortisol downregu-
lates osteoblast alpha 1 (I) procollagen mRNA by transcrip-
tional and posttranscriptional mechanisms. J Cell Biochem Received in original form March 14, 2001; in revised form July 19,
7:488–494. 2001; accepted August 9, 2001.
5