Clinical Science (2004) 107, 111–123 (Printed in Great Britain)
R
E
V
I
E
W
Sensitivity of bone to glucocorticoids
Mark S. COOPER
Division of Medical Sciences, University of Birmingham, Queen Elizabeth Hospital, Edgbaston, Birmingham B15 2TH, U.K.
A
B
S
T
R
A
C
T
Glucocorticoids are used widely in a range of medical specialities, but their main limitation is an
adverse impact on bone. Although physicians are increasingly aware of these deleterious effects,
the marked variation in susceptibility between individuals makes it difficult to predict who will
develop skeletal complications with these drugs. Although the mechanisms underlying the adverse
effects on bone remain unclear, the most important effect appears to be a rapid and substantial
decrease in bone formation. This review will examine recent studies that quantify the risk of
fracture with glucocorticoids, the mechanisms that underlie this increase in risk and the potential
basis for differences in individual sensitivity. An important determinant of glucocorticoid sensitivity
appears to be the presence of glucocorticoid-metabolizing enzymes within osteoblasts and this
may enable improved estimates of risk and generate new approaches to the development of
bone-sparing anti-inflammatory drugs.
INTRODUCTION
The adverse effects of glucocorticoids on the skeleton
were first recognized by Harvey Cushing [1] 70 years
ago in the context of the disease that bears his name.
However, the problem of GIOP (glucocorticoid-induced
osteoporosis) has increased dramatically in recent decades
with the introduction of synthetic glucocorticoids to
treat a range of inflammatory diseases [2,3]. Physicians
in most medical specialities are increasingly aware of the
effects of therapeutic glucocorticoids on the skeleton,
but the marked variation in susceptibility between
individuals has made it difficult to predict individuals
who are likely to have adverse effects with these
drugs. The mechanisms underlying the effects of glucocorticoids on bone have been sought for many years, but
no single effect has emerged as being of prime importance.
In the last few years, key advances have been made in
several areas, including the epidemiology of GIOP, the
impact of glucocorticoids on bone cells in vitro and
in vivo and the basis for variation in sensitivity between
individuals. Additionally, the recent identification of
glucocorticoid-modifying enzymes in bone cells suggests
that endogenous glucocorticoids may make a greater contribution to bone metabolism than previously thought.
This article will critically review these recent advances
and will highlight areas where considerable uncertainty
still exists.
RECENT ADVANCES IN THE EPIDEMIOLOGY
OF GIOP
The epidemiology of GIOP has been difficult to define
due to the wide variability in glucocorticoid doses used
clinically, the impact of the underlying disease being
treated and the large numbers of subjects needed to
measure an effect on fracture risk. Recent epidemiological
studies have overcome some, but not all, of these
problems and have given us a more accurate estimation
of the magnitude of fracture risk.
van Staa et al. [3] used information from the
General Practice Research Database, a computerized
medical records system that covers a large number of
Key words: bone, glucocorticoids, 11β-hydroxysteroid dehydrogenase, osteoblasts, osteoporosis, prednisolone.
Abbreviations: AP-1, activator protein-1; BMD, bone mineral density; BMP, bone morphogenic protein; CBG, corticosteroid
binding globulin; GIOP, glucocorticoid-induced osteoporosis; GR, glucocorticoid receptor; 11β-HSD, 11β-hydroxysteroid
dehydrogenase; IFN-β, interferon-β; MR, mineralocorticoid receptor; NF-κB, nuclear factor κB; OPG, osteoprotegerin; PTH,
parathyroid hormone; RANK, receptor activator of NF-κB; RANKL, RANK ligand.
Correspondence: Dr Mark S. Cooper (email [email protected]).
C
2004 The Biochemical Society
111
112
M. S. Cooper
general practices in England and Wales, to estimate
glucocorticoid usage in the U.K. It was determined that
0.9 % of this population was continuously receiving
prescriptions for glucocorticoids for more than 3 months,
a figure which rose to 2.5 % in the elderly. Over 90 %
of these prescriptions were for prednisolone with only
3 % for hydrocortisone. Using this database, a group
of 244 235 oral corticosteroid users were identified and
individually matched with a similar number of control
subjects who were receiving topical, but not oral, glucocorticoids [4]. The risk of developing a fracture before,
during or after treatment was then examined in these
subjects. Using this approach it was determined that fracture risk increased substantially at all skeletal sites examined with the greatest relative risk (2.6) seen at the spine.
Fracture risk at all sites increased with steroid dose
with the largest increases again seen at the spine with
relative risks of over 5 with high doses of steroids. A
statistically significant increase in spine fracture risk was
seen even with very low doses of prednisolone (relative
risk 1.55 with doses < 2.5 mg/day), suggesting that there
may be no safe lower dose of oral glucocorticoid.
These risks appeared to be more related to current than
cumulative dose. What was a surprise was the time course
of fracture risk. Fracture risk at all sites increased rapidly
after starting glucocorticoids and then appeared to plateau
even during treatment for up to 5 years. When steroid
treatment was stopped, fracture risk fell rapidly back
towards baseline. These results suggest that changes in
BMD (bone mineral density) are unlikely to be mediating
the changes in fracture risk. Loss of BMD during glucocorticoid treatment is often rapid in the first 6–12 months
of treatment and then continues at a slower rate [5].
Additionally, although some gain in BMD might be expected on ceasing glucocorticoid treatment, it would be
unlikely that the changes would be so rapid and complete
as to account for the change in fracture risk. The origin
of this rapid and reversible effect of glucocorticoids is
likely to be the source of intense research interest in the
future [6]. From a practical point of view, these results also
suggest that attempts at prophylactic treatment should
begin early during treatment, and probably at the time of
commencing glucocorticoids. Furthermore, prophylactic
measures probably do not need to be continued for prolonged periods once steroids have been ceased. In the U.K.
these findings have been incorporated into excellent evidence-based guidelines for the management of GIOP [7].
The advantages of these studies were their large size
and relevance to patients typical of the population at
large. However, this, and related studies, do have some
significant limitations. This approach cannot separate the
impact of the underlying disease from that of glucocorticoids, although the effects seen appeared to be consistent across the various indications for glucocorticoid
use (e.g. respiratory versus rheumatological). Furthermore, the data only reflected prescriptions that had
C
2004 The Biochemical Society
Figure 1 Effects of glucocorticoids on the skeleton
Glucocorticoids can affect bone cells directly (grey arrows) or indirectly via a
range of factors (black arrows).
been issued. There was no way of determining whether
the glucocorticoids had actually been taken. More
importantly, it is likely that many individuals in the
low dose category were actually taking intermittent, but
frequent, courses of glucocorticoids. These limitations,
along with the use of a control group that were taking
inhaled steroids, would tend to weaken the relationship
and, thus, the data may actually underestimate the
magnitude of the risks associated with glucocorticoids.
IMPACT OF GLUCOCORTICOIDS ON BONE
CELLS IN VITRO
The effect of glucocorticoid excess on the skeleton may be
mediated by indirect effects on skeletal metabolism, such
as reduction in sex steroid levels, reduction in muscle mass
and strength, and secondary hyperparathyroidism due to
impairment in calcium absorption or increased calcium
excretion [8,9] (Figure 1). These effects may be very
important in various clinical settings, but the dominant
effect of glucocorticoids appears to be due to direct effects
on bone cells [10,11]. The three main cell types within
bone are osteoblasts, osteocytes and osteoclasts. Osteoblasts are cells which form bone matrix and are
essential for its mineralization. Osteocytes are terminally
differentiated osteoblasts that have been incorporated
into bone, whereas osteoclasts are derived from the
monocyte lineage and resorb bone. In various settings,
glucocorticoids have potentially important impacts on all
of these cell types.
Osteoblasts in vitro
Osteoblasts have long been thought to be the most
important target of glucocorticoids in bone. The
histological hallmark of GIOP is a rapid and profound
decrease in the capacity to form bone [12–14] and
Glucocorticoid sensitivity of bone
dramatic reductions in biochemical markers of bone
formation and osteoblastic activity are seen in a clinical
setting [15–18]. Correspondingly, glucocorticoids have
dramatic effects on osteoblasts in a range of experimental
settings. Most studies have focussed on the role of glucocorticoids in the commitment of cells to the osteoblast
lineage or on the function of mature osteoblasts. More
recently the impact of glucocorticoids on the death of
osteoblasts has been explored.
Glucocorticoids appear to be important in the
commitment of cells to the osteoblast lineage [10]. In cells
of human origin, glucocorticoids appear to be essential
for this stage. Glucocorticoids also induce the expression
of a range of genes that are characteristic of mature
osteoblasts, such as osteocalcin and type I collagen. By
contrast, glucocorticoids can also suppress features of
the osteoblast phenotype in mature cells. Various studies
have examined the impact of glucocorticoids on a range
of intracellular and extracellular signalling pathways that
may be important in bone metabolism. Glucocorticoids
decrease the stability of mRNA for the bone specific
transcription factor cbfa-1 (core-binding factor-1) [19]
and the activity of several intracellular kinases may also
be reduced [20]. PTH (parathyroid hormone) action
on osteoblasts may be reduced by glucocorticoids [21],
whereas BMP (bone morphogenic protein) generation is
increased [22]. The expression of Notch receptors (which
are important in BMP signalling) may be increased by
glucocorticoids [23]. Multiple aspects of the GH (growth
hormone)/IGF (insulin-like growth factor) system may
be affected [9,24] and glucocorticoids may reduce TGF-β
(transforming growth factor-β) activity [19]. Despite the
potential importance of all these effects, the impact on
bone physiology in vivo remain unclear.
More recently, a significant impact of glucocorticoids
on osteoblast apoptosis has been noted. In some systems,
glucocorticoids induce apoptosis [25], whereas in
others glucocorticoids protect against apoptosis [26,27].
However, in primary osteoblasts and osteoblasts in vivo,
it is likely that high doses of glucocorticoids have primarily an apoptotic action [25,28].
In vitro experiments continue to be plagued by
problems with interpretation. There are clearly contrasts
between transformed cells and primary cultures, e.g. it is
difficult to interpret studies examining proliferation and
apoptosis in transformed cells. There is also difficulty in
defining the impact of glucocorticoids at various stages
of differentiation especially when trying to extrapolate
to the in vivo situation. Additionally, the presence of
multiple effects on a range of genes potentially important
in bone cell function suggests that no one pathway
is likely to explain the clinically relevant actions of
glucocorticoids. In vivo, bone is remodelled at distinct
sites with controlled coordination of groups of osteoclasts
and osteoblasts as part of the ‘bone remodelling unit’.
In vitro models are currently unable to reproduce this
level of complexity. An example of where in vitro experiments may give insight into the pathogenesis of GIOP
is the capacity of intermittent PTH to protect against
apoptosis in osteoblast cell lines that are normally sensitive to glucocorticoid-induced apoptosis [29]. These
studies have the capacity to stimulate in vivo and clinical
studies.
Osteocytes in vitro
Osteoblasts have a number of fates. They can die via
apoptosis, become quiescent osteoblasts lining the bone
surface or become incorporated into bone matrix and
undergo differentiation into osteocytes. The function of
osteocytes remains poorly defined, but roles in mechanosensing, the detection of microfractures and the regulation of bone repair have been proposed. This has
recently raised the possibility that osteocyte damage
may be important in the pathogenesis of glucocorticoidinduced osteonecrosis [25,30]. Progress in the understanding of osteocyte function has been limited by the
inability to study viable human osteocytes in vitro. An
important advance has been the development of a rodent
cell line that demonstrates many features typical of osteocytes, such as high osteocalcin levels, expression of
connexin43 and a dendritic morphology [31]. Although
the function of osteocytes remains poorly defined, it is
difficult to assess the functional impact of glucocorticoids
on osteocytes except in the extreme example of osteocyte
death. More evidence of the role of glucocorticoids on
osteocyte function has come from in vivo, rather than
in vitro, experiments.
The role of glucocorticoids on osteoblasts that line
the bone surface is also obscure. These cells may be
important in the initiation of bone remodelling via the expression of enzymes capable of resorbing the thin layer
of unmineralized matrix which has to be removed to
allow access of osteoclasts to the bone surface [32,33].
Glucocorticoids may increase expression of these
enzymes [34], and this effect could account for some of
the increased resorption of bone seen during treatment
with glucocorticoids.
Osteoclasts in vitro
The other cell type within bone that may be a target
for glucocorticoids is the osteoclast. Osteoclasts are
multinucleated cells derived from the mononuclear cell
lineage. The formation and function of osteoclasts is
regulated by a range of factors, many of which are derived
from osteoblasts. A critical step in the differentiation
of mononuclear cells to osteoclasts is signalling via the
RANK (receptor activator of NF-κB, where NF-κB is
nuclear factor κB) pathway by RANKL (RANK ligand)
[35]. RANKL is a cell-surface receptor expressed on a
range of osteoblasts and some immune cells. In addition
to being required for osteoclast formation, RANKL
also stimulates the activity of mature osteoclasts. A
C
2004 The Biochemical Society
113
114
M. S. Cooper
naturally occurring inhibitor of RANKL signalling, OPG
(osteoprotegerin) is also secreted by osteoblasts and, thus,
the local RANKL/OPG ratio is likely to determine the
degree of osteoclast formation.
Several protocols are now available for the in vitro
generation of osteoclasts. In all of these in vitro systems,
high doses of synthetic glucocorticoids, primarily dexamethasone, are required. In contrast with osteoblasts and
osteocytes, high doses of glucocorticoids do not appear
to induce osteoclast apoptosis. Glucocorticoids appear to
have an important role in the formation and function of
osteoclasts via effects on osteoblasts [36]. Glucocorticoids
up-regulate the expression of RANKL on the surface
of a range of osteoblasts. Additionally, glucocorticoids
reduce the production of OPG by osteoblasts. This
increase in the RANKL/OPG ratio would be expected
to increase the generation and activity of osteoclasts and
this may account for some of the early bone resorption
seen with glucocorticoids. Glucocorticoids may also
have direct effects on osteoclasts. In osteoclasts cultured
in vitro, glucocorticoids stimulate intracellular signalling
pathways which are associated with inflammation. In
this setting, glucocorticoids exert their stimulatory effect
on osteoclast formation not by directly stimulating the
activity of the RANK signalling pathway, but rather
by inhibiting the production of IFN-β (interferon-β)
[37]. IFN-β exerts an important inhibitory action on the
RANKL pathway and, thus, this effect of glucocorticoids
may be due to their relief of IFN-β-mediated inhibition
of RANK signalling. The importance of these effects
in vivo during glucocorticoid treatment or in inflammatory disease has yet to be explored.
Factors regulating sensitivity of bone
cells to glucocorticoids in vitro
When performing in vitro studies it is usual to try to
create a reproducible and uniform model system. Such an
approach may mask variations in the sensitivity of bone
cells to glucocorticoids. Sensitivity to glucocorticoids
could be regulated at a number of levels (Figure 2).
Individuals could differ in the absorption, distribution
or metabolism of steroid, the number of and affinity of
GRs (glucocorticoid receptors), or amount and binding
of nuclear cofactors. Additionally, all of these factors
could change in a clinical setting during inflammation
or prolonged glucocorticoid exposure.
For glucocorticoids to act, steroids need to diffuse into
a target cell, bind to specific receptors and alter gene
transcription either by binding to DNA or interfering
with other signalling pathways, such as NF-κB or AP-1
(activator protein-1). GRs are expressed in a range of
bone cells, including osteoblasts and osteocytes [38,39],
and this topic has previously been reviewed in Clinical
Science [40]. The number of GRs varies between tissues,
but whether this has a role in determining differences in
glucocorticoid sensitivity is unclear. Some GR variants
C
2004 The Biochemical Society
Figure 2 Schematic illustration of the factors influencing
the action of therapeutic glucocorticoid on bone cells
Glucocorticoid action is affected by changes in the circulating level of free glucocorticoid, the local metabolism of glucocorticoids, the number and affinity of
steroid receptors and the level of nuclear cofactors.
have been described. An asparagine to serine change
at codon 363 results in increased suppression of serum
cortisol levels in response to low dose dexamethasone.
Heterozygous individuals have a trend to decreased BMD
at the lumbar spine, but the overall effect appears small
[41]. GR phosphorylation status has also been implicated
in regulation of GR function [42], but its physiological
significance remains unclear. Glucocorticoids can also
bind to MRs (mineralocorticoid receptors) in tissues
that do not express the 11β-HSD2 (11β-hydroxysteroid
dehydrogenase 2) enzyme (see below). MRs are expressed
within bone cells and may be involved in glucocorticoid
signalling [39]. Nuclear coactivators and corepressors
are likely to be important in tissue-specific responses to
glucocorticoids. Whether expression of these cofactors is
likely to explain any differences in sensitivity between
individuals is currently unclear.
Recently, the potential importance of glucocorticoidmodifying enzymes in a range of tissues has become apparent. 11β-HSDs are intracellular enzymes that catalyse
the interconversion of the inactive hormone cortisone and
hormonally active cortisol [43]. Even though cortisone
has been used for over 50 years as an orally active glucocorticoid, it requires conversion (predominantly on firstpass metabolism in the liver) into cortisol (referred to
as hydrocortisone when given as a pharmaceutical)
to have bioactivity [44]. Two distinct 11β-HSD enzymes
have been described in man [43]. 11β-HSD1 is bidirectional, but in most tissues examined to date is primarily
a glucocorticoid activator converting cortisone into
cortisol. This enzyme is found in tissues which express
high levels of GR, such as liver, adipose and gonad. 11βHSD2 is unidirectional, solely inactivating cortisol to
cortisone, and is found in tissues which express MR, such
as the kidney. Cortisol and aldosterone can both bind
Glucocorticoid sensitivity of bone
the MR with similar affinity, but the circulating level of
cortisol is substantially higher than that of aldosterone.
11β-HSD2 is thus required in mineralocorticoid target
tissues to reduce the intracellular level of cortisol to enable
specific binding of aldosterone to the MR.
The expression of glucocorticoid-modifying enzymes
has recently been examined in bone cells. Enzyme activity
and mRNA expression was first studied in a range of
rat and human osteosarcoma cells [45,46]. Only 11βHSD2 activity was detected, and the direction of activity
was glucocorticoid inactivation. By contrast, 11β-HSD1,
but not 11β-HSD2, activity was seen in primary cultures of human, mouse or rat osteoblasts, and these
cells generate active from inactive glucocorticoids [46].
Osteoblasts from the CD-1 mouse strain have exclusive
11β-HSD1 expression and activity [47], as do osteoblasts
derived from rat vertebrae [48]. Interestingly, osteoblasts
from the C57/Bl6 mouse strain exhibit 11β-HSD1
activity, but also express 11β-HSD2 mRNA [49]. The
presence of functional 11β-HSD1 activity (cortisone to
cortisol activation) within osteoblasts is also indicated
by the ability of cortisone to influence the formation of
bone nodules by bone marrow derived from CD-1 mice
and Wistar rats [50], an effect that could be blocked by
the 11β-HSD1 inhibitor carbenoxolone. The functional
consequences of 11β-HSD enzyme expression have
been explored in osteosarcoma cells stably transfected
with cDNAs for the human enzymes [51,52]. In cells
expressing 11β-HSD2, cortisol was unable to induce
alkaline phosphatase activity in contrast with control
empty vector cells. In cells expressing 11β-HSD1, alkaline phosphatase activity was induced normally by
cortisol, but these cells were rendered sensitive to cortisone despite this steroid having no effect in control
cells. These experiments suggest that circulating cortisone
can behave as an active glucocorticoid in osteoblasts
expressing 11β-HSD1.
Although the metabolism of cortisol (hydrocortisone)
is likely to be of relevance in normal physiology,
the predominant glucocorticoids used therapeutically
are prednisolone (used in the U.K.) and prednisone
(the most frequently used oral glucocorticoid in the
U.S.A.) [3]. Prednisolone and prednisone are synthetic
derivatives of cortisol and cortisone respectively and, like
cortisone, prednisone requires activation by 11β-HSD1
for bioactivity. The metabolism of these synthetic steroids
has been examined in primary osteoblasts and other
cells expressing 11β-HSD1 [53]. The enzyme kinetics
of the conversion of prednisone into prednisolone were
indistinguishable from those of cortisone into cortisol.
This suggests that local glucocorticoid metabolism will
also be relevant to these clinically important synthetic
glucocorticoids.
The expression of 11β-HSD1 and the ability to
generate cortisol from cortisone in human osteoblasts
increases substantially when cells are exposed to the
pro-inflammatory cytokines TNF-α (tumour necrosis
factor-α) or IL-1β (interleukin-1β) [54]. Glucocorticoid
levels are thus likely to be higher near sites of inflammation and may contribute to the periarticular bone
loss seen in patients with long-standing rheumatoid
arthritis [55]. Additionally, glucocorticoids themselves
can increase 11β-HSD1 activity and mRNA expression
in these cells [53]. These effects are likely to have
implications for the regulation of local cortisol generation
in various clinical settings.
Other levels of potential interest will be variation in
enzyme expression with skeletal site, especially in view
of the marked sensitivity of the bones of the axial skeleton
(e.g. vertebrae and ribs) compared with that of the peripheral skeleton. Although bone found throughout the
body has similar physical properties, there are considerable regional differences in developmental origin and
mechanisms by which various bones are formed [56].
During the earliest phases of development, templates for
bones are formed from condensations of mesenchymal
cells. These cells have distinct origins with the craniofacial
skeleton formed by cranial neural crest cells [57], the
axial skeletal elements from paraxial (somatic) mesoderm
[58] and the limb elements from lateral plate mesoderm [59]. There is a suggestion that expression of 11βHSD1 might vary with skeletal site, with variability of
activity being less in osteoblasts derived from the calcaneus compared with other sites in the lower limb
[53], but further work needs to be done in this area.
Along similar lines, it is also clear that osteoblasts in the
periosteum (the outer layer of the bone that is important in apposition growth) differ in some ways from
their endosteal (the inner part of the bone) counterparts.
Changes in the periosteum are likely to be important in
determining skeletal structure during aging [60]. GRs are
expressed in osteoblasts at both periosteal and endosteal
sites [61], but functional differences in the response of
these osteoblasts to glucocorticoids have not been examined. An additional under-appreciated area is the potential differences in the behaviour of bone cells involved
in either bone modelling (the establishment of the basic
skeletal structure) or remodelling (the ongoing repair and
renewal of the skeleton) [62]. These differences are extremely difficult to examine in vitro and new approaches
will be needed.
There are also species differences in the response to
glucocorticoids. Many transformed cell lines and primary
osteoblasts derived from mice can spontaneously mineralize without the need for additional glucocorticoids [63–
65]. Human cell lines and primary cultures of osteoblasts
in contrast appear to need supplemental glucocorticoids
to induce mineralization [66–68]. This raises important
questions about the applicability of murine cell lines to
potentially important features of osteoblast behaviour
such as matrix composition and the ability to mineralize
this matrix.
C
2004 The Biochemical Society
115
116
M. S. Cooper
In contrast with osteoblasts, little is known about the
variation in glucocorticoid sensitivity of osteoclasts or
osteocytes. In osteoclasts, the capacity of cells to differentiate in vitro in response to the synthetic glucocorticoid dexamethasone does not appear to differ between males and females [69]. Furthermore, sensitivity
to dexamethasone does not change with age [69]. Dexamethasone is not metabolized to any great extent by
glucocorticoid-modifying enzymes, thus differences
might still be seen with endogenous glucocorticoids
if enzyme expression was to change in these settings.
As a consequence of the inability to culture human
osteocytes in vitro, any differences in sensitivity due
to age, sex, skeletal site or disease status have not been
examined, but would have relevance to glucocorticoidinduced apoptosis and osteonecrosis.
IMPACT OF GLUCOCORTICOIDS IN VIVO
Animal models
Many animal models of GIOP have been suggested, but
most have important limitations. The response of various
bone cells to glucocorticoids can differ considerably
between species. In the rat, bone density increases with
glucocorticoid treatment, an effect that appears to be
due to glucocorticoid-induced osteoclast apoptosis [70].
Additionally, the organization of cortical bone structure
in rodents is different from that in humans and other
primates. Once cortical bone is formed in rodents, it does
not undergo further remodelling with age. By contrast,
human cortical bone has an osteonal (Haversian) organization that undergoes remodelling throughout life, a
process that removes bone that has sustained microfracture or become hypermineralized [71]. This limits
rodent models in their ability to mimic changes that
occur with aging or that focus primarily on cortical bone.
Osteonal modelling also occurs in sheep and this animal
has thus been proposed as a model of GIOP [72]. The
changes in biochemical bone markers and bone histology
in sheep exposed to glucocorticoids are similar to those
seen in humans. The sheep model is likely to be of particular value in studying the impact of glucocorticoids on
biomechanical properties of bone and the interaction with
oestrogen deficiency. Despite the limitations of rodent
models, the most important animal model to emerge
recently is the mouse. Mice appear to develop changes in
the skeleton which are broadly similar to those seen
in humans [25]. The ease of use, short breeding time
and the capacity to generate genetically modified animals
gives the mouse model a number of strengths, and recent
studies have generated data that are now testable in
humans and of potential relevance to clinical bone disease
in man.
Several studies have now been reported from the
Weinstein’s group using mice treated with subcutaneous
C
2004 The Biochemical Society
prednisolone pellets [25,73–76]. These animals have
changes in bone histomorphology analogous with those
seen in man [25]. These studies have shed light on the importance of the rates of birth and death of bone cells.
Glucocorticoids appear to increase the rate of osteoblast
and osteocyte apoptosis. Additionally, glucocorticoids
reduce the generation of early osteoblasts as assessed by
the capacity of bone marrow to generate colony forming
units containing mineralized matrix. The capacity of
bone marrow to generate osteoclasts was also reduced.
The survival of mature osteoclasts was, however,
maintained by glucocorticoid treatment [74]. Intermittent
PTH injections were able to reduce the level of osteoblast
apoptosis in these studies [29]. More recently, evidence
has been presented suggesting that glucocorticoids can
reduce vertebral bone compression strength independently of bone density [75].
The expression of glucocorticoid-metabolizing enzymes has also been defined in mice. An 11β-HSD1
gene-ablation mouse model has been produced [77],
but it has not been examined whether there are differences in the response of these animals to therapeutic
glucocorticoids [49]. Murine osteoblasts, as in humans,
express 11β-HSD1 and can generate active from inactive
glucocorticoids. An interesting approach to the study
of the effects of glucocorticoids on bone has been the
development of mice that have targeted overexpression
of the glucocorticoid-inactivating enzyme 11β-HSD2
within osteoblasts [47]. These animals were generated
using the well characterized Col2.3 type 1 collagen promoter [78]. Hydrocortisone inhibited collagen synthesis
in calvaria from wild-type mice, but this effect was
blunted in mice with osteoblastic 11β-HSD2 expression.
A similar mouse has been generated with 11β-HSD2
expression under an osteocalcin promoter [75]. These
animals were protected against many of the skeletal
changes induced by short-term prednisolone treatment.
The authors conclude that the effects of glucocorticoids
on bone are thus direct on osteoblasts in this model rather
than being mediated by changes in factors such as sex
steroid levels or calcium homoeostasis. The applicability
of enzyme expression to mice models of GIOP will
depend on the levels of circulating glucocorticoids during
glucocorticoid treatment. Prednisolone and prednisone
levels have not yet been reported in these models.
Studies in mice have contributed a great deal to our
understanding of the potential mechanisms underlying
GIOP. Future models should allow the impact of inflammatory conditions on the response to glucocorticoids
to be defined and thus be better able to mimic clinical
scenarios.
Human studies
Studies in humans have also been performed. The
hallmark of GIOP is a rapid reduction in bone formation
markers, such as osteocalcin, alkaline phosphatase and
Glucocorticoid sensitivity of bone
the propeptides of type 1 collagen [15]. By contrast,
there is no change in markers of bone resorption. Bone
density decreases and this can occur very rapidly
[5,79,80]. This loss of bone appears to be due to sustained
resorption in the face of reduced formation. The ability
of bisphosphonates, which inhibit bone resorption, to
reduce fracture risk at the spine in patients taking
glucocorticoids [79,81,82] also suggests that resorption
plays an important role in GIOP. At a histological
level, studies have again supported a dominant role
of reduced bone formation, but, in post-menopausal
women, increased resorption also appears to be a feature
[13]. In addition to the features described above, a
decrease in the viability of osteocytes has been observed
in trans-iliac biopsies from patients with rheumatoid
arthritis treated with prednisolone [28].
In humans, in contrast with animals, mechanistically
based approaches to treatment can be evaluated in genuine
clinical settings. The potential of intermittent PTH to
protect against osteoblast apoptosis in isolated cells
and in animal models was discussed earlier. The use
of PTH in a clinical setting has been examined in a
trial of women taking prednisone and HRT (hormonereplacement therapy) [83]. Women randomized to PTH
showed an increase in spine BMD compared with
women who did not receive the drug. There was no
early change in femoral neck BMD, but a follow-up
study demonstrated a delayed increase [84]. Using QCT
(quantitative computerized tomography) it was apparent
that these increases were predominantly in trabecular
bone. The study was not powered to examine changes
in fracture risk, but the changes in BMD seem likely to
be protective. Very few studies have been able to look
at the material properties of glucocorticoid-treated bone,
but the profound changes in matrix protein expression
are unlikely to be without consequence. The recent
increased focus on bone quality [85,86] should allow a
fuller examination of this area.
Although the commonest skeletal problems associated
with glucocorticoid use is fracture, the complication
perhaps most feared by physicians is osteonecrosis.
Osteonecrosis (also termed avascular necrosis) involves
the rapid and focal deterioration in bone quality and
primarily affects the femoral head [87]. It can lead to pain
and ultimately collapse of the bone often requiring hip
replacement. It can affect individuals of all ages and can
occur with relatively low doses of glucocorticoids (e.g.
during corticosteroid-replacement therapy for adrenal
failure) [88]. The pathogenesis and basis for susceptibility
between individuals remains unknown, but recent data
implicate glucocorticoid-induced osteocyte death in the
condition. These conclusions are based on the relative
frequency of osteocyte apoptosis observed in animal
models and the increased rate of local osteocyte apoptosis
in femoral heads removed for this reason [25,30]. The
lack of a direct role for an interrupted blood supply
suggests that the term osteonecrosis is preferable in
this context to avascular necrosis. Even if osteocyte
apoptosis is the mechanism underlying glucocorticoidinduced osteonecrosis, there is still no explanation for
why this occurs in only a few individuals and there is no
way of predicting the risk of the complication occurring
in an individual.
Local and systemic metabolism of
glucocorticoids in vivo
Until recently, very little attention was paid to the pharmacology of commonly used glucocorticoids. The effects
of glucocorticoids were thought to be broadly similar
and effects on bone were an inevitable consequence of
anti-inflammatory potency. Pharmaceutical companies
have recently driven attempts to develop dissociated
glucocorticoids [89]. These compounds would ideally
maintain their anti-inflammatory action (thought to be
via transrepression, GR binding to transcription factors,
such as NF-κB and AP-1) whilst lacking their negative
impact on bone (thought to be via transactivation, binding
of GR to DNA). Although promising candidates have
been developed, none have demonstrated a steroidsparing effect on bone in vivo [90,91].
What has been a surprise is the realization that the
pharmacology of the most commonly used oral glucocorticoids may be more complex than previously thought.
Furthermore, there may be simple ways in which the
negative impact on bone could be reduced. As discussed
above, cortisone and prednisone are biologically inactive
and thus little attention has been paid to these steroids
compared with their active counterparts cortisol and
prednisolone. The recent identification of glucocorticoidmodifying enzymes within bone cells suggests that these
inactive steroids may have a functional role in bone. Even
in normal individuals there is interconversion of cortisol
and cortisone with shuttling of these steroids between
liver and kidney (Figure 3) [43]. Circulating cortisone
derives primarily from the kidneys where the enzyme
11β-HSD2 is expressed [92]. Cortisone can be converted
into cortisol in the liver or any other tissue expressing
11β-HSD1. Both cortisone and prednisone can be given
orally since there is almost complete first-pass hepatic
conversion into their active counterparts. However, in
a similar manner to cortisone, prednisolone can be
metabolized to prednisone by renal 11β-HSD2 [93]. The
levels of prednisone generated in normal individuals after
a 5 mg dose are of the order of 60 nmol/l, which are
similar to those of cortisone [53]. At these concentrations,
sufficient substrate should be available for conversion
into prednisolone in tissues expressing 11β-HSD1. This
suggests that the effects of glucocorticoids may be
direct via circulating active glucocorticoids or by local
conversion of inactive glucocorticoids into their active
counterparts.
C
2004 The Biochemical Society
117
118
M. S. Cooper
Figure 3 Systemic and local metabolism of glucocorticoids
by 11β-HSD enzymes
Endogenous and exogenous glucocorticoids undergo extensive interconversion
between active and inactive forms in vivo with inactivation occurring primarily
in the kidney and reactivation in the liver and other tissues such as bone that
express 11β-HSD1.
The relative contributions of these two pathways has
been examined in a clinical trial of healthy subjects taking
prednisolone [18]. All subjects taking 5 mg of prednisolone twice daily demonstrated decreases in the bone formation markers osteocalcin and N-terminal propeptide
of type 1 collagen. The extent of the fall was highly
correlated with baseline 11β-HSD1 activity measured
by the urinary ratio of cortisol/cortisone metabolites.
This study suggests that local glucocorticoid generation
makes a considerable contribution to the effects of glucocorticoids on bone formation markers. However, the
urinary ratios of steroid metabolites are likely to reflect
enzyme activity occurring in several tissues, including
liver, fat and kidney, in addition to bone. Although
predictive of changes in bone metabolism in healthy
subjects, it remains to be determined if the same relationships apply in subjects during illness, since inflammation
and glucocorticoid treatment may lead to tissue-specific
dysregulation of steroid metabolism in this setting [44].
This area of research suggests new ways in which
tissue selectivity of glucocorticoids can be achieved. If
a steroid can maintain its effects on immune tissue, but
is unable to be reactivated locally via 11β-HSD1 in
osteoblasts, it is likely to have a reduced impact on bone.
It is interesting that local glucocorticoid metabolism
may explain some of the findings with deflazacort, an
oxazoline derivative of prednisolone. This steroid was
promoted as a relatively bone-sparing glucocorticoid
when compared with prednisone/prednisolone. Several
studies suggested a reduced effect on bone compared
with prednisolone when used in equivalent anti-inflammatory doses, whereas others found no difference or
that higher doses of deflazacort were needed to maintain
an anti-inflammatory effect equivalent to that of prednisolone. These studies were done at a time when
C
2004 The Biochemical Society
the potential significance of tissue-specific metabolism
of glucocorticoids was not appreciated. Prednisolone
is inactivated to prednisone in the kidney by 11βHSD2 and thus the intrarenal levels of prednisolone
are selectively reduced. In mice given an intraperitoneal
injection of prednisolone the prednisolone/prednisone
ratio in the kidney was 0.8 after 60 min but 7 in the
plasma [93]. Tissues that express 11β-HSD1 can convert
prednisone back into prednisolone, thus the corresponding prednisolone/prednisone ratio in the liver was
43. In contrast with prednisolone, deacetyl-deflazacort,
the active metabolite of deflazacort formed by firstpass hepatic metabolism, is a relatively poor substrate
for 11β-HSD2 [94] and, thus, renal levels would not be
expected to be significantly lower than in other tissues.
Additionally, the metabolites of deflazacort are poor
substrates for reactivation by 11β-HSD1 [95].
It is thus not a surprise that the most dramatic
studies supporting a bone-sparing effect of deflazacort
were seen in patients with renal disease [96,97]. Two
double-blind studies randomized patients with nephrotic
syndrome or renal transplantation to prednisone or
deflazacort assuming a 1:1.2 relative potency. Bone loss
was significantly reduced in patients taking deflazacort
compared with prednisone in both of these trials with
equivalent levels of immunosuppression. However, when
the relative potencies of deflazacort and prednisolone
were assessed in patients with polymyalgia rheumatica,
higher doses of deflazacort were required for disease
control (equivalent to a ratio of 1:1.4) [98]. The equivalent
doses of prednisolone and deflazacort needed for effects
in various tissues in healthy subjects has been examined.
For suppression of eosinophils the dose ratio was 1.12, for
osteocalcin the ratio was 1.54 and for cortisol suppression
the ratio was 2.27 [99]. On this basis it would appear
that deflazacort is superior to prednisolone for treating
renal inflammation. It is also possible that deflazacort
may have reduced effects on bone due to a lack of
reactivation by 11β-HSD1 in osteoblasts. This is likely to
be the case, but the clinical significance will be dependent
on whether 11β-HSD1 is expressed in the tissue being
targeted therapeutically. 11β-HSD1 can be induced in
a range of tissues during inflammation [100–102] and
this has implications for the design of anti-inflammatory
medications that have selective enzymic reactivation. It
is possible that glucocorticoid therapy in the future may
be optimized on the basis of the underlying disease being
treated and the capacity of various tissues to undergo
enzymic modification of the particular glucocorticoid.
SENSITIVITY TO ENDOGENOUS
GLUCOCORTICOIDS
Endogenous corticosteroids may play a role in normal
bone physiology. Previous studies have examined natural
Glucocorticoid sensitivity of bone
variation in serum cortisol levels and found weak
correlations with BMD. In healthy men, integrated
24 h cortisol levels correlated with lumbar spine BMD
(r = − 0.37, P < 0.05) and femoral neck BMD (r = − 0.31,
P = 0.06), although the relationships were removed by adjustment for BMI (body mass index) [103]. Peak, trough
or integrated serum cortisol levels in men do not correlate
with biochemical bone markers, but trough cortisol levels
predicted bone loss at the spine and hip over a 4 year
period after adjustment for adiposity [103]. No associations were seen between these variables in women. Other
investigators have found weak associations of serum
cortisol and lumbar spine BMD in men [104], but not
women [105]. A further study examined the association
of salivary cortisol with lumbar spine BMD. Evening
salivary cortisol levels were found to be negatively
correlated with spine BMD in elderly women, whereas
morning salivary cortisol was negatively correlated with
spine BMD in elderly men [106]. In the MacArthur
Study of Successful Aging [107], the relationship between
baseline urinary free cortisol excretion, measured on an
overnight urine sample, and fracture risk was examined
in 684 community-dwelling men and women aged 70–79
over a 7 year period. Increasing quartiles of urinary free
cortisol were associated with increasing risk of fracture,
an effect that was seen primarily in men. These results
support a relationship between free cortisol and bone
metabolism, but other factors are possible, such as an
impact of cortisol on muscle mass and propensity to fall.
There are technical difficulties with examining any
relationships between endogenous cortisol and bone
health such as problems in measuring free cortisol levels
rather than total [primarily CBG (corticosteroid binding
globulin)-bound] cortisol, and the complicating factors
of stress responses and diurnal variation. Additionally,
urinary free cortisol measurements are heavily influenced
by individual differences in cortisol metabolism with only
approx. 1 % of the total cortisol synthesized excreted as
‘free’ cortisol [108].
The presence of 11β-HSD1 within bone suggests that
endogenous ‘inactive’ steroids may play a role in normal
bone physiology via local reactivation to their active
counterparts. Such an effect would not be reflected in circulating levels of cortisol. The circulating concentration
of cortisone in healthy individuals is approx. 50 nmol/l
[109]. Since cortisone has much less binding to CBG
than cortisol, the free concentrations of cortisol and cortisone are likely to be very similar and the levels of
cortisone would be sufficient to have an impact in tissues
if they express sufficient 11β-HSD1. An independent
impact of cortisone on bone has yet to be shown and
it is unclear whether this impact would be anabolic
(by virtue of increased commitment of immature cells
to the osteoblasts lineage) or catabolic (by suppressing
the function of mature osteoblasts). Enzyme expression
is likely to vary across differentiation and, thus, the
impact is likely to depend on the stage of osteoblast
differentiation at which 11β-HSD1 is expressed.
Changes in the capacity of human osteoblasts to generate active glucocorticoids locally may also occur with age
[53,110]. There is considerable variation in the activity
of 11β-HSD1 in osteoblasts cultured from different
individuals [53]. Additionally, significant correlations
were seen between cortisol generation and donor age
with osteoblasts from older individuals able to generate
2–3 times more cortisol than osteoblasts from younger
donors. Were such an effect to occur in vivo then
osteoblasts within older subjects would be exposed to
much higher levels of active glucocorticoids than those
of younger individuals. It is thus possible that part of
the dramatic increase in fracture risk that occurs with
age [111] is due to a localized form of GIOP without
changes in the circulating levels of glucocorticoids. This
hypothesis requires further testing in vivo.
Insights into the role of endogenous steroids have
been derived from mice which overexpress 11β-HSD2
in osteoblasts and osteocytes. Reduced vertebral bone
density was seen in female mice expressing 11β-HSD2
under a type 1 collagen promoter, suggesting an anabolic
effect of glucocorticoids at this site [47]. Currently, it
is unclear whether this is due to blocking circulating
or locally generated corticosteroids. An 11β-HSD1
knockout model has been examined and displays subtle
abnormalities in trabecular structure, suggesting that
local glucocorticoid generation may have a role in bone
development [49]. Analysis of the bones of this animal
is complicated by the alterations in HPA (hypothalamo–
pituitary–adrenal) axis function that occur in these mice
[112] and the selection of a mouse strain (C57/Bl6)
which expresses 11β-HSD2, as well as 11β-HSD1, in
osteoblasts. The animal models described above are not
ideal when examining the changes that occur in the human
skeleton with aging and there is a need for further clinical
studies in this area.
OUTSTANDING QUESTIONS
The reasons for the rapid and transient nature of the
increase in fracture risk during glucocorticoid treatment
remains an important area of research. Proposed explanations such as osteocyte viability cannot explain
the rapid reduction in fracture risk that occurs with
glucocorticoid withdrawal. A more plausible explanation
is that glucocorticoids have a subtle impact on the material
properties of bone. No study has yet examined in depth
the material properties of glucocorticoid-treated bone,
but the profound reductions in alkaline phosphatase,
osteocalcin and other enzymes during glucocorticoid
treatment are unlikely to be without consequence. New
techniques such as NMR and infrared spectroscopy are
becoming more accessible and will be used to examine
this issue. An area of considerable clinical importance
C
2004 The Biochemical Society
119
120
M. S. Cooper
is the impact of recurrent short-term high dose glucocorticoid exposure on the skeleton. The impact on fracture risk has yet to be determined in this setting, although
large numbers of individuals, primarily with respiratory
disease, are treated in this manner. Further clarification of
the basis for differences in sensitivity between individuals
is also needed. If this is due primarily to differences in
local glucocorticoid metabolism, the clinical utility of
measures of glucocorticoid metabolism needs to be determined. There is also the prospect that some individuals
will have aberrant expression of glucocorticoid-metabolizing enzymes and are thus exposed to abnormally
high glucocorticoid levels within bone and thus may
develop a localized form of GIOP. Blocking local glucocorticoid production is likely to have a beneficial effect
in these individuals.
REFERENCES
1 Cushing, H. W. (1932) The basophil adenomas of the
pituitary body and their clinical manifestations
(pituitary basophilism). Bull. Johns Hopkins Hosp. 50,
137–195
2 Hench, P. S., Slocumb, C. H., Polley, H. F. and
Kendall, E. C. (1951) Effect of cortisone and pituitary
adrenocorticotropic hormone (ACTH) on rheumatic
diseases. JAMA, J. Am. Med. Assoc. 144, 1327–1335
3 van Staa, T. P., Leufkens, H. G., Abenhaim, L., Begaud,
B., Zhang, B. and Cooper, C. (2000) Use of oral
corticosteroids in the United Kingdom. Q. J. Med. 93,
105–111
4 van Staa, T. P., Leufkens, H. G., Abenhaim, L., Zhang, B.
and Cooper, C. (2000) Use of oral corticosteroids and risk
of fractures. J. Bone Miner. Res. 15, 993–1000
5 Gennari, C. and Imbimbo, B. (1985) Effects of prednisone
and deflazacort on vertebral bone mass. Calcif. Tissue Int.
37, 592–593
6 Manolagas, S. C. (2000) Corticosteroids and fractures:
a close encounter of the third cell kind. J. Bone Miner.
Res. 15, 1001–1005
7 Compston, J. on behalf of Guidelines Writing Group.
(2002) Glucocorticoid-induced osteoporosis: Guidelines
for prevention and treatment. Royal College of
Physicians of London, UK
8 Lukert, B. P. and Raisz, L. G. (1990) Glucocorticoidinduced osteoporosis: pathogenesis and management.
Ann. Intern. Med. 112, 352–364
9 Canalis, E. (1996) Clinical review 83: mechanisms of
glucocorticoid action in bone: implications to
glucocorticoid-induced osteoporosis. J. Clin. Endocrinol.
Metab 81, 3441–3447
10 Cooper, M. S., Hewison, M. and Stewart, P. M. (1999)
Glucocorticoid activity, inactivity and the osteoblast.
J. Endocrinol. 163, 159–164
11 Manolagas, S. C. (2000) Birth and death of bone cells:
basic regulatory mechanisms and implications for the
pathogenesis and treatment of osteoporosis.
Endocr. Rev. 21, 115–137
12 Dempster, D. W. (1989) Bone histomorphometry in
glucocorticoid-induced osteoporosis. J. Bone Miner. Res.
4, 137–141
13 Carbonare, L. D., Arlot, M. E., Chavassieux, P. M., Roux,
J. P., Portero, N. R. and Meunier, P. J. (2001) Comparison
of trabecular bone microarchitecture and remodeling in
glucocorticoid-induced and postmenopausal
osteoporosis. J. Bone Miner. Res. 16, 97–103
14 Chavassieux, P. M., Arlot, M. E., Roux, J. P. et al. (2000)
Effects of alendronate on bone quality and remodeling in
glucocorticoid-induced osteoporosis: a histomorphometric analysis of transiliac biopsies. J. Bone Miner. Res.
15, 754–762
C
2004 The Biochemical Society
15 Godschalk, M. F. and Downs, R. W. (1988) Effect of
short-term glucocorticoids on serum osteocalcin in
healthy young men. J. Bone Miner. Res. 3, 113–115
16 Prummel, M. F., Wiersinga, W. M., Lips, P., Sanders, G. T.
and Sauerwein, H. P. (1991) The course of biochemical
parameters of bone turnover during treatment with
corticosteroids. J. Clin. Endocrinol. Metab. 72, 382–386
17 Pearce, G., Tabensky, D. A., Delmas, P. D., Baker, H. W.
and Seeman, E. (1998) Corticosteroid-induced bone loss
in men. J. Clin. Endocrinol. Metab. 83, 801–806
18 Cooper, M. S., Blumsohn, A., Goddard, P. E. et al. (2003)
11β-hydroxysteroid dehydrogenase type 1 activity
predicts the effects of glucocorticoids on bone. J. Clin.
Endocrinol. Metab. 88, 3874–3877
19 Chang, D. J., Ji, C., Kim, K. K., Casinghino, S., McCarthy,
T. L. and Centrella, M. (1998) Reduction in transforming
growth factor β receptor I expression and transcription
factor CBFa1 on bone cells by glucocorticoid.
J. Biol. Chem. 273, 4892–4896
20 Engelbrecht, Y., de Wet, H., Horsch, K., Langeveldt,
C. R., Hough, F. S. and Hulley, P. A. (2003)
Glucocorticoids induce rapid up-regulation of mitogenactivated protein kinase phosphatase-1 and dephosphorylation of extracellular signal-regulated kinase and
impair proliferation in human and mouse osteoblast cell
lines. Endocrinology 144, 412–422
21 Rubin, M. R. and Bilezikian, J. P. (2002) Clinical review
151: The role of parathyroid hormone in the pathogenesis
of glucocorticoid-induced osteoporosis: a re-examination
of the evidence. J. Clin. Endocrinol. Metab 87, 4033–4041
22 Boden, S. D., Hair, G., Titus, L. et al. (1997)
Glucocorticoid-induced differentiation of fetal rat
calvarial osteoblasts is mediated by bone morphogenetic
protein-6. Endocrinology 138, 2820–2828
23 Pereira, R. M., Delany, A. M., Durant, D. and Canalis, E.
(2002) Cortisol regulates the expression of Notch in
osteoblasts. J. Cell Biochem. 85, 252–258
24 McCarthy, T. L., Ji, C., Chen, Y., Kim, K. and Centrella,
M. (2000) Time- and dose-related interactions between
glucocorticoid and cyclic adenosine 3 , 5 -monophosphate
on CCAAT/enhancer-binding protein-dependent
insulin-like growth factor I expression by osteoblasts.
Endocrinology 141, 127–137
25 Weinstein, R. S., Jilka, R. L., Parfitt, A. M. and Manolagas,
S. C. (1998) Inhibition of osteoblastogenesis and
promotion of apoptosis of osteoblasts and osteocytes by
glucocorticoids. Potential mechanisms of their deleterious
effects on bone. J. Clin. Invest 102, 274–282
26 Nakashima, T., Sasaki, H., Tsuboi, M. et al. (1998)
Inhibitory effect of glucocorticoid for osteoblast
apoptosis induced by activated peripheral blood
mononuclear cells. Endocrinology 139, 2032–2040
27 Pereira, R. M., Delany, A. M. and Canalis, E. (2001)
Cortisol inhibits the differentiation and apoptosis of
osteoblasts in culture. Bone 28, 484–490
28 Sambrook, P. N., Hughes, D. R., Nelson, A. E.,
Robinson, B. G. and Mason, R. S. (2003) Osteocyte
viability with glucocorticoid treatment: relation to
histomorphometry. Ann. Rheum. Dis. 62, 1215–1217
29 Jilka, R. L., Weinstein, R. S., Bellido, T., Roberson, P.,
Parfitt, A. M. and Manolagas, S. C. (1999) Increased bone
formation by prevention of osteoblast apoptosis with
parathyroid hormone. J. Clin. Invest. 104, 439–446
30 Weinstein, R. S., Nicholas, R. W. and Manolagas, S. C.
(2000) Apoptosis of osteocytes in glucocorticoid-induced
osteonecrosis of the hip. J. Clin. Endocrinol. Metab. 85,
2907–2912
31 Kato, Y., Windle, J. J., Koop, B. A., Mundy, G. R. and
Bonewald, L. F. (1997) Establishment of an osteocyte-like
cell line, MLO-Y4. J. Bone Miner. Res. 12, 2014–2023
32 Chambers, T. J. and Fuller, K. (1985) Bone cells
predispose bone surfaces to resorption by exposure of
mineral to osteoclastic contact. J.Cell Sci. 76, 155–165
33 Bord, S., Horner, A., Hembry, R. M., Reynolds, J. J. and
Compston, J. E. (1996) Production of collagenase by
human osteoblasts and osteoclasts in vivo. Bone 19, 35–40
34 Delany, A. M., Jeffrey, J. J., Rydziel, S. and Canalis, E.
(1995) Cortisol increases interstitial collagenase
expression in osteoblasts by post-transcriptional
mechanisms. J. Biol. Chem. 270, 26607–26612
Glucocorticoid sensitivity of bone
35 Hofbauer, L. C., Khosla, S., Dunstan, C. R., Lacey, D. L.,
Boyle, W. J. and Riggs, B. L. (2000) The roles of
osteoprotegerin and osteoprotegerin ligand in the
paracrine regulation of bone resorption.
J. Bone Miner. Res. 15, 2–12
36 Hofbauer, L. C., Gori, F., Riggs, B. L. et al. (1999)
Stimulation of osteoprotegerin ligand and inhibition of
osteoprotegerin production by glucocorticoids in
human osteoblastic lineage cells: potential paracrine
mechanisms of glucocorticoid-induced osteoporosis.
Endocrinology 140, 4382–4389
37 Takuma, A., Kaneda, T., Sato, T., Ninomiya, S.,
Kumegawa, M. and Hakeda, Y. (2003) Dexamethasone
enhances osteoclast formation synergistically with
transforming growth factor-β by stimulating the priming
of osteoclast progenitors for differentiation into
osteoclasts. J. Biol. Chem. 278, 44667–44674
38 Manolagas, S. C. and Anderson, D. C. (1978) Detection
of high-affinity glucocorticoid binding in rat bone.
J. Endocrinol. 76, 379–380
39 Beavan, S., Horner, A., Bord, S., Ireland, D. and
Compston, J. (2001) Colocalization of glucocorticoid and
mineralocorticoid receptors in human bone. J. Bone
Miner. Res. 16, 1496–1504
40 Bland, R. (2000) Steroid hormone receptor expression and
action in bone. Clin. Sci. 98, 217–240
41 Huizenga, N. A., Koper, J. W., De Lange, P. et al. (1998)
A polymorphism in the glucocorticoid receptor gene
may be associated with and increased sensitivity to
glucocorticoids in vivo. J. Clin. Endocrinol. Metab. 83,
144–151
42 Webster, J. C., Jewell, C. M., Bodwell, J. E., Munck, A.,
Sar, M. and Cidlowski, J. A. (1997) Mouse glucocorticoid
receptor phosphorylation status influences multiple
functions of the receptor protein. J. Biol. Chem. 272,
9287–9293
43 Tomlinson, J. W., Walker, E. A., Bujalska, I. J. et al. (2004)
11β-hydroxysteroid dehydrogenase type 1: A tissuespecific regulator of glucocorticoid response. Endocr.
Rev., in the press
44 Cooper, M. S. and Stewart, P. M. (2003) Corticosteroid
insufficiency in acutely ill patients. N. Engl. J. Med. 348,
727–734
45 Eyre, L. J., Rabbitt, E. H., Bland, R. et al. (2001)
Expression of 11β-hydroxysteroid dehydrogenase
in rat osteoblastic cells: Pre-receptor regulation of
glucocorticoid responses in bone. J. Cell Biochem. 81,
453–462
46 Bland, R., Worker, C. A., Noble, B. S. et al. (1999)
Characterization of 11β-hydroxysteroid dehydrogenase
activity and corticosteroid receptor expression in human
osteosarcoma cell lines. J. Endocrinol. 161, 455–464
47 Sher, L. B., Woitge, H. W., Adams, D. J. et al. (2004)
Transgenic expression of 11β-hydroxysteroid
dehydrogenase type 2 in osteoblasts reveals an anabolic
role for endogenous glucocorticoids in bone.
Endocrinology 145, 922–929
48 Ishida, Y., Killinger, D. W., Khalil, M. W., Yang, K.,
Strutt, B. and Heersche, J. N. (2002) Expression of
steroid-converting enzymes in osteoblasts derived from
rat vertebrae. Osteoporos. Int. 13, 235–240
49 Justesen, J., Mosekilde, L., Holmes, M. et al. (2004) Mice
deficient in 11β-hydroxysteroid dehydrogenase type 1
lack bone marrow adipocytes, but maintain normal bone
formation. Endocrinology 145, 1916–1925
50 Bellows, C. G., Ciaccia, A. and Heersche, J. N. (1998)
Osteoprogenitor cells in cell populations derived from
mouse and rat calvaria differ in their response to
corticosterone, cortisol, and cortisone. Bone 23, 119–125
51 Rabbitt, E., Lavery, G. G., Walker, E. A., Cooper, M. S.,
Stewart, P. M. and Hewison, M. (2002) Pre-receptor
regulation of glucocorticoid action by 11βhydroxysteroid dehydrogenase: a novel determinant of
cell proliferation. FASEB J. 16, 36–44
52 Woitge, H., Harrison, J., Ivkosic, A., Krozowski, Z. and
Kream, B. (2001) Cloning and in vitro characterization of
α 1(I)-collagen 11β-hydroxysteroid dehydrogenase type 2
transgenes as models for osteoblast-selective inactivation
of natural glucocorticoids. Endocrinology 142, 1341–1348
53 Cooper, M. S., Rabbitt, E. H., Goddard, P. E., Bartlett,
W. A., Hewison, M. and Stewart, P. M. (2002)
Osteoblastic 11β-hydroxysteroid dehydrogenase type 1
activity increases with age and glucocorticoid exposure.
J. Bone Miner. Res. 17, 979–986
54 Cooper, M. S., Bujalska, I., Rabbitt, E. et al. (2001)
Modulation of 11β-hydroxysteroid dehydrogenase
isozymes by proinflammatory cytokines in osteoblasts:
an autocrine switch from glucocorticoid inactivation to
activation. J. Bone Miner. Res. 16, 1037–1044
55 Cooper, M. S., Emery, P., Hewison, M. and Stewart, P. M.
(2003) The role of local corticosteroid generation in
inflammation-associated bone loss. J. Bone Miner. Res. 18
(Suppl.), S44
56 Olsen, B. R., Reginato, A. M. and Wang, W. (2000) Bone
development. Annu. Rev. Cell Dev. Biol. 16, 191–220
57 Bronner-Fraser, M. (1994) Neural crest cell formation and
migration in the developing embryo. FASEB J. 8, 699–706
58 Tam, P. P. and Trainor, P. A. (1994) Specification and
segmentation of the paraxial mesoderm. Anat. Embryol.
(Berlin) 189, 275–305
59 Cohn, M. J. and Tickle, C. (1996) Limbs: a model for
pattern formation within the vertebrate body plan.
Trends Genet. 12, 253–257
60 Ahlborg, H. G., Johnell, O., Turner, C. H., Rannevik, G.
and Karlsson, M. K. (2003) Bone loss and bone size after
menopause. N. Engl. J. Med 349, 327–334
61 Abu, E. O., Horner, A., Kusec, V., Triffitt, J. T. and
Compston, J. E. (2000) The localization of the functional
glucocorticoid receptor α in human bone. J. Clin.
Endocrinol. Metab. 85, 883–889
62 Duan, Y., Beck, T. J., Wang, X. F. and Seeman, E. (2003)
Structural and biomechanical basis of sexual dimorphism
in femoral neck fragility has its origins in growth and
aging. J. Bone Miner. Res 18, 1766–1774
63 Ecarot-Charrier, B., Glorieux, F. H., van der, R. M. and
Pereira, G. (1983) Osteoblasts isolated from mouse
calvaria initiate matrix mineralization in culture.
J. Cell Biol. 96, 639–643
64 Ikeda, K., Matsumoto, T., Morita, K., Kurokawa, K. and
Ogata, E. (1986) Inhibition of in vitro mineralization by
aluminum in a clonal osteoblastlike cell line, MC3T3-E1.
Calcif. Tissue Int. 39, 319–323
65 zur Nieden, N. I., Kempka, G. and Ahr, H. J. (2003)
In vitro differentiation of embryonic stem cells into
mineralized osteoblasts. Differentiation 71, 18–27
66 Cheng, S. L., Yang, J. W., Rifas, L., Zhang, S. F. and
Avioli, L. V. (1994) Differentiation of human bone
marrow osteogenic stromal cells in vitro: induction
of the osteoblast phenotype by dexamethasone.
Endocrinology 134, 277–286
67 Iba, K., Sawada, N., Nuka, S. et al. (1996)
Phase-dependent effects of transforming growth factor β
1 on osteoblastic markers of human osteoblastic cell line
sV-HFO during mineralization. Bone 19, 363–369
68 Jonsson, K. B., Frost, A., Nilsson, O., Ljunghall, S. and
Ljunggren, O. (1999) Three isolation techniques for
primary culture of human osteoblast-like cells:
a comparison. Acta Orthop. Scand. 70, 365–373
69 Jevon, M., Sabokbar, A., Fujikawa, Y. et al. (2002) Genderand age-related differences in osteoclast formation from
circulating precursors. J. Endocrinol. 172, 673–681
70 Dempster, D. W., Moonga, B. S., Stein, L. S., Horbert,
W. R. and Antakly, T. (1997) Glucocorticoids inhibit bone
resorption by isolated rat osteoclasts by enhancing
apoptosis. J. Endocrinol. 154, 397–406
71 Parfitt, A. M. (1994) Osteonal and hemi-osteonal
remodeling: the spatial and temporal framework for signal
traffic in adult human bone. J. Cell Biochem. 55, 273–286
C
2004 The Biochemical Society
121
122
M. S. Cooper
72 Chavassieux, P., Pastoureau, P., Chapuy, M. C., Delmas,
P. D. and Meunier, P. J. (1993) Glucocorticoid-induced
inhibition of osteoblastic bone formation in ewes:
a biochemical and histomorphometric study.
Osteoporos. Int. 3, 97–102
73 Plotkin, L. I., Weinstein, R. S., Parfitt, A. M., Roberson,
P. K., Manolagas, S. C. and Bellido, T. (1999) Prevention
of osteocyte and osteoblast apoptosis by bisphosphonates
and calcitonin. J. Clin. Invest. 104, 1363–1374
74 Weinstein, R. S., Chen, J. R., Powers, C. C. et al. (2002)
Promotion of osteoclast survival and antagonism of
bisphosphonate-induced osteoclast apoptosis by
glucocorticoids. J. Clin. Invest. 109, 1041–1048
75 O’Brien, C. A., Jia, D., Plotkin, L. I. et al. (2003)
Glucocorticoids act directly on osteoblasts and osteocytes
to induce their apoptosis and reduce bone formation and
strength. Endocrinology 145, 1835–1841
76 Weinstein, R. S., Jia, D., Powers, C. C. et al. (2004) The
skeletal effects of glucocorticoid excess override those of
orchidectomy in mice. Endocrinology 145, 1980–1987
77 Kotelevtsev, Y., Holmes, M. C., Burchell, A. et al. (1997)
11β-hydroxysteroid dehydrogenase type 1 knockout
mice show attenuated glucocorticoid-inducible responses
and resist hyperglycemia on obesity or stress. Proc. Natl.
Acad. Sci. U.S.A. 94, 14924–14929
78 Kalajzic, I., Kalajzic, Z., Kaliterna, M. et al. (2002) Use of
type I collagen green fluorescent protein transgenes to
identify subpopulations of cells at different stages of the
osteoblast lineage. J. Bone Miner. Res. 17, 15–25
79 Adachi, J. D., Bensen, W. G., Brown, J. et al. (1997)
Intermittent etidronate therapy to prevent corticosteroidinduced osteoporosis. N. Engl. J. Med. 337, 382–387
80 Cohen, S., Levy, R. M., Keller, M. et al. (1999)
Risedronate therapy prevents corticosteroid-induced
bone loss: a twelve-month, multicenter, randomized,
double-blind, placebo-controlled, parallel-group study.
Arthritis Rheum. 42, 2309–2318
81 Saag, K. G., Emkey, R., Schnitzer, T. J. et al. (1998)
Alendronate for the prevention and treatment of
glucocorticoid-induced osteoporosis. GlucocorticoidInduced Osteoporosis Intervention Study Group.
N. Engl. J. Med. 339, 292–299
82 Wallach, S., Cohen, S., Reid, D.M. et al. (2000) Effects of
risedronate treatment on bone density and vertebral
fracture in patients on corticosteroid therapy.
Calcif. Tissue Int. 67, 277–285
83 Lane, N. E., Sanchez, S., Modin, G. W., Genant, H. K.,
ini, E. and Arnaud, C. D. (1998) Parathyroid hormone
treatment can reverse corticosteroid-induced
osteoporosis. Results of a randomized controlled clinical
trial. J. Clin. Invest. 102, 1627–1633
84 Lane, N. E., Sanchez, S., Modin, G. W., Genant, H. K.,
Pierini, E. and Arnaud, C. D. (2000) Bone mass continues
to increase at the hip after parathyroid hormone treatment
is discontinued in glucocorticoid-induced osteoporosis:
results of a randomized controlled clinical trial.
J. Bone Miner. Res. 15, 944–951
85 Borah, B., Dufresne, T. E., Chmielewski, P. A., Gross,
G. J., Prenger, M. C. and Phipps, R. J. (2002) Risedronate
preserves trabecular architecture and increases bone
strength in vertebra of ovariectomized minipigs as
measured by three-dimensional microcomputed
tomography. J. Bone Miner. Res. 17, 1139–1147
86 Watts, N. B. (2002) Bone quality: getting closer to a
definition. J. Bone Miner. Res. 17, 1148–1150
87 Mankin, H. J. (1992) Nontraumatic necrosis of bone
(osteonecrosis). N. Engl. J. Med 326, 1473–1479
88 Williams, P. L. and Corbett, M. (1983) Avascular necrosis
of bone complicating corticosteroid replacement therapy.
Ann. Rheum. Dis. 42, 276–279
89 Belvisi, M. G., Brown, T. J., Wicks, S. and Foster, M. L.
(2001) New glucocorticosteroids with an improved
therapeutic ratio? Pulm. Pharmacol. Ther. 14, 221–227
90 Belvisi, M. G., Wicks, S. L., Battram, C. H. et al. (2001)
Therapeutic benefit of a dissociated glucocorticoid
and the relevance of in vitro separation of transrepression from transactivation activity. J. Immunol. 166,
1975–1982
C
2004 The Biochemical Society
91 Schacke, H., Schottelius, A., Docke, W. D. et al. (2004)
Dissociation of transactivation from transrepression by a
selective glucocorticoid receptor agonist leads to
separation of therapeutic effects from side effects. Proc.
Natl. Acad. Sci.U.S.A. 101, 227–232
92 Walker, B. R., Campbell, J. C., Fraser, R., Stewart, P. M.
and Edwards, C. R. (1992) Mineralocorticoid excess and
inhibition of 11β-hydroxysteroid dehydrogenase in
patients with ectopic ACTH syndrome.
Clin. Endocrinol. (Oxford) 37, 483–492
93 Escher, G., Frey, F. J. and Frey, B. M. (1994)
11β-Hydroxysteroid dehydrogenase accounts for low
prednisolone/prednisone ratios in the kidney.
Endocrinology 135, 101–106
94 Diederich, S., Eigendorff, E., Burkhardt, P. et al. (2002)
11β-hydroxysteroid dehydrogenase types 1 and 2:
an important pharmacokinetic determinant for the activity
of synthetic mine. J. Clin. Endocrinol. Metab. 87,
5695–5701
95 Hult, M., Jornvall, H. and Oppermann, U. C. (1998)
Selective inhibition of human type 1 11β-hydroxysteroid
dehydrogenase by synthetic steroids and xenobiotics.
FEBS Lett. 441, 25–28
96 Olgaard, K., Storm, T., van Wowern, N. et al. (1992)
Glucocorticoid-induced osteoporosis in the lumbar spine,
forearm, and mandible of nephrotic patients: a doubleblind study on the high-dose, long-term effects of
prednisone versus deflazacort. Calcif. Tissue Int. 50,
490–497
97 Lippuner, K., Casez, J. P., Horber, F. F. and Jaeger, P.
(1998) Effects of deflazacort versus prednisone on bone
mass, body composition, and lipid profile: a randomized,
double blind study in kidney transplant patients. J. Clin.
Endocrinol. Metab. 83, 3795–3802
98 Krogsgaard, M. R., Lund, B. and Johnsson, B. (1995) A
longterm prospective study of the equipotency between
deflazacort and prednisolone in the treatment of patients
with polymyalgia rheumatica. J. Rheumatol. 22,
1660–1662
99 Babadjanova, G., Allolio, B., Vollmer, M., Reincke, M.
and Schulte, H. M. (1996) Comparison of the
pharmacodynamic effects of deflazacort and
prednisolone in healthy subjects. Eur. J. Clin.
Pharmacol. 51, 53–57
100 Hennebold, J. D., Ryu, S. Y., Mu, H. H., Galbraith, A.
and Daynes, R. A. (1996) 11β-Hydroxysteroid
dehydrogenase modulation of glucocorticoid activities in
lymphoid organs. Am. J. Physiol 270, R1296–R1306
101 Hennebold, J. D., Mu, H. H., Poynter, M. E., Chen, X. P.
and Daynes, R. A. (1997) Active catabolism of
glucocorticoids by 11β-hydroxysteroid dehydrogenase
in vivo is a necessary requirement for natural resistance
to infection with Listeria monocytogenes. Int. Immunol.
9, 105–115
102 Thieringer, R., Le Grand, C. B., Carbin, L. et al. (2001)
11β-Hydroxysteroid dehydrogenase type 1 is induced in
human monocytes upon differentiation to macrophages.
J. Immunol. 167, 30–35
103 Dennison, E., Hindmarsh, P., Fall, C. et al. (1999) Profiles
of endogenous circulating cortisol and bone mineral
density in healthy elderly men. J. Clin. Endocrinol.
Metab. 84, 3058–3063
104 Cetin, A., Gokce-Kutsal, Y. and Celiker, R. (2001)
Predictors of bone mineral density in healthy males.
Rheumatol. Int. 21, 85–88
105 Acar, B., Uslu, T., Topuz, A. et al. (1998) Relation
between bone mineral content and clinical, hormonal and
biochemical parameters in postmenopausal women.
Arch. Gynecol. Obstet. 261, 121–128
106 Raff, H., Raff, J. L., Duthie, E. H. et al. (1999) Elevated
salivary cortisol in the evening in healthy elderly men and
women: correlation with bone mineral density.
J. Gerontol. A Biol. Sci. Med. Sci. 54, M479–M483
107 Greendale, G. A., Unger, J. B., Rowe, J. W. and Seeman,
T. E. (1999) The relation between cortisol excretion
and fractures in healthy older people: results from the
MacArthur studies-Mac. J. Am. Geriatr. Soc. 47,
799–803
Glucocorticoid sensitivity of bone
108 Palermo, M., Shackleton, C. H., Mantero, F. and Stewart,
P. M. (1996) Urinary free cortisone and the assessment of
11β-hydroxysteroid dehydrogenase activity in man.
Clin. Endocrinol. (Oxford) 45, 605–611
109 Andrews, R. C. and Walker, B. R. (1999) Glucocorticoids
and insulin resistance: old hormones, new targets. Clin.
Sci. 96, 513–523
110 Canalis, E. and Delany, A. M. (2002) 11β-hydroxysteroid
dehydrogenase, an amplifier of glucocorticoid action in
osteoblasts. J. Bone Miner. Res. 17, 987–990
111 Cooper, C. and Joseph Melton III, L. (1996) Magnitude
and impact of osteoporosis and fractures. In Osteoporosis
(Marcus, R., Feldman, D. and Kelsey, J., eds), Academic
Press, San Diego
112 Harris, H. J., Kotelevtsev, Y., Mullins, J. J., Seckl, J. R. and
Holmes, M. C. (2001) Intracellular regeneration of
glucocorticoids by 11β-hydroxysteroid dehydrogenase
(11β-HSD)-1 plays a key role in regulation of the
hypothalamic–pituitary–adrenal axis: analysis of
11β-HSD-1-deficient mice. Endocrinology 142, 114–120
Received 8 March 2004/14 April 2004; accepted 28 April 2004
Published as Immediate Publication 28 April 2004, DOI 10.1042/CS20040070
C
2004 The Biochemical Society
123