| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
-LJUBOJEVIDepartment of Biology, University of York, York, YO10 5YW, U.K.
1Correspondence: Department of Biology, University of York, PO Box 373, York, YO10 5YW, U.K. E-mail nmp2{at}york.ac.uk
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-Ljubojevi, I., Skerry, T. M. The
glutamate receptor antagonist MK801 modulates bone resorption in
vitro by a mechanism predominantly involving osteoclast
differentiation.
Key Words: osteoclast • differentiation • electrophysiology • glutamate receptor • MK801 maleate (Dizocilpine)
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
or in
individuals with aberrant PTH receptor function. (2)
. The
effects of the numerous paracrine agents on bone mass also demonstrate
the importance of local intercellular communication (3)
.
Identification by us, and subsequently by others, of glutamate
transporters (4)
and receptors (5
6
7)
in bone
suggests that paracrine signaling in bone cells may mimic synaptic
transmission between neurons.
Glutamate is the major excitatory neurotransmitter in the central
nervous system (CNS) and is released in a quantal manner by presynaptic
cells to bind to clustered specific cell surface receptors on the
postsynaptic cell (8)
. These receptors are divided into
ionotropic receptors (three subgroups of glutamate-gated ion channels)
and 7-transmembrane domain-containing metabotropic receptors that
signal via G-proteins (9)
. The ionotropic receptors are
divided according to their sensitivity to pharmacological ligands
N-methyl-D-aspartate (NMDA), AMPA, and kainate. In response
to glutamate binding, ionotropic receptors open to allow influx of ions
into the cell, regulating subsequent processes. In this report, we
focus on the function of NMDA receptors in osteoclasts, as
immunocytochemical localizations show that human and laboratory animal
osteoclasts possess NMDA receptors (6
, 7)
.
Naturally the implication of this is that the receptors play a role in
the process of bone resorption, and in this study we have used
electrophysiological techniques as well as in vitro
resorption assays to investigate that role. Patch clamping is the
method of choice for characterizing membrane ion channel activity and
allows accurate definition of specific responses to both agonists and
antagonists in single cells. This is a significant goal, since
differences between CNS and bone glutamate receptors could point toward
divergence in ligand specificity that would have profound implications
on drug targeting. Longer term in vitro assays, such as the
ones we use here, have been available for the study of
osteoclastogenesis and bone resorption for a considerable period of
time and the methods are well characterized (10
11
12)
. We
show here that although mature osteoclasts possess NMDA receptors and
respond acutely to ligand stimulation, the role of glutamate appears to
be predominantly to control osteoclastogenesis, since the effects of
NMDA antagonists are greater by far in cultures where osteoclasts
develop than in cultures where mature cell function is investigated.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Isolation of mature rabbit osteoclasts
Osteoclasts were isolated from 2- to 3-day-old neonatal rabbits
as described previously (13)
. Briefly, for each
experiment, femora and tibiae from a single rabbit were curetted into
minimal essential medium containing 10% fetal calf serum and
penicillin/streptomycin (medium). Cells were settled in medium onto 10
mm diameter glass coverslips at 37°C, 5% CO2, and
nonadherent cells were removed after 45 min by vigorous washing in
phosphate-buffered saline (PBS). The coverslips were transferred to 4
ml pre-equilibrated medium and maintained at 37°C in humidified 95%
air, 5% CO2 for 3 h prior to use.
Patch clamp studies
To assist patching, coverslips with adherent osteoclasts were
placed at 4°C for 30 min, causing the osteoclasts to round up (D.
Ypey, personal communication). The coverslips were then rinsed in
external solution before use. The cell preparation was placed in a
recording chamber mounted on an inverted microscope. Experiments were
performed in the whole-cell patch clamp configuration (14)
using an L/M-EPC7 amplifier and CED patch/voltage clamp software.
Pipettes (2–10 M
) were pulled from unfilamented borosilicate Kovar
glass (Corning 7052) and heat polished. Experiments were performed at
room temperature (20–26°C). Seal resistance ranged from 5 to 10
G. Cells were held at 0 mV and stepped to test potentials between
-140 and 140 mV for 800 ms in 20 mV steps. To investigate the membrane
currents before and after addition of glutamate [100–300 µM] and
MK801 (Tocris) [100 µM], we used nearly symmetrical solutions on
each side of the membrane. The NMDA receptor coagonist glycine [30
µM] was used in conjunction with each application of glutamate or
NMDA. Intracellular solutions contained [mM]: 140 CsCl, 10 HEPES, 5
glucose, 5 K2ATP; pH 7.2. The extracellular
solution was Mg2+ and Na+
free and contained [mM]: 140 CsCl, 2 CaCl2, 5
HEPES, 10 glucose; pH 7.3. All solutions were filtered prior to use
(pore diameter 0.2 µM). Compounds were added to the extracellular
compartment via a pipette. In between the experiments, the bath was
perfused continuously with extracellular solution. Results are
displayed as typical traces and representative current/voltage
relationships. Statistical significance was determined by Student’s
t test.
Coculture
Osteoclasts were generated from cocultures of osteoblasts and
marrow leukocytes essentially as described previously
(10)
. To extract primary osteoblasts, neonatal mouse
calvariae were dissected free of loose tissues and washed in Hanks’
buffered salts solution with magnesium and calcium ions (Hanks’+),
then digested for 15 min with collagenase type I (crude) at 1 mg/ml in
Hanks’+ (collagenase), when the solution was aspirated and discarded.
The calvariae were then digested for 30 min in 3 ml collagenase and
washed twice with 3 ml of phosphate-buffered saline without magnesium
and calcium ions (PBS-). The collagenase suspension and the washes
were combined and the cells were collected by centrifugation at
1000 x g for 3 min, suspended in medium (
-MEM with
10% fetal calf serum and penicillin/streptomycin, and kept on ice. The
calvariae were treated with EDTA at 4 mM in PBS- for 15 min and washed
twice with Hanks’+. The cells were collected as before; the calvariae
were digested with collagenase for another 60 min and the cells were
collected. All three fractions of cells were pooled and cultured in 15
ml of medium in a 75 cm2 flask. Three hours
later, the medium was aspirated and the cells were gently washed once
with Hanks’+. The cells were then maintained in medium without
disturbance for up to a week. They were usually used for coculture
within 3 days.
Bone marrow leukocytes were extracted from adult mouse tibiae and femora. The bones were dissected and placed in Hanks’+. The muscle tissue was scraped away and the epiphyses were cut off to gain access to the marrow cavity. The marrow tissue was ejected from the shafts with Hanks’+ with 10% fetal calf serum using a syringe and a 25 gauge needle. It was then converted to a fine cell suspension by sequential passage through 19, 21 and 25 gauge needles and layered onto Ficoll. The leukocyte fraction was recovered from the top of the Ficoll after 25 min centrifugation at 600 x g and washed once in Hanks’+.
To set up the coculture, 10 mm diameter coverslips or elephant ivory wafers (6 mm diameter) were placed in 48-well plates and covered with 200 µl of medium with 10-8m 1,25 dihydroxy vitamin D3 (D3 medium). Freshly extracted leukocytes and previously prepared osteoblasts were counted and suspended in D3 medium so that addition of 100 µl of mixture yielded 2 x 104 osteoblasts and 4 x 105 leukocytes per well. Drugs or vehicle were diluted to 4x final concentration with D3 medium and added to the wells in 100 µl, bringing the total volume per well to 400 µl. Wells at the edge of each plate were not used but were filled with PBS. The D3 medium was refreshed on days 4 and 7 by preparing fresh D3 medium with a 2x final concentration of the drug or vehicle, then adding 400 µl dropwise to the wells and waiting 20 min before removing 400 µl.
To analyze osteoclast morphology and activity 11 days later, cells on
coverslips were fixed in 4% paraformaldehyde in PBS and stained for
tartrate-resistant acid phosphatase (TRAP; 15
).
Ivory wafers were placed in distilled water for 15 min and then rubbed
gently with lint-free tissue to remove cells. A Leica Quantimet image
analysis system was used to measure areas of TRAP staining and
resorption lacunae (16).
The area of TRAP-stained osteoclasts was measured in eight geometrically predetermined fields and the measured area was extrapolated to represent the whole area of the coverslip. Resorption lacunae area was measured at five geometrically predetermined fields and similarly extrapolated to the whole area of the wafer. Data from coculture studies were transformed logarithmically to normalize distributions, and differences between control and experimental cultures were analyzed by one-way analysis of variance and Dunnett’s test.
Mature osteoclast resorption assay
Rat and rabbit osteoclasts were studied to determine the role of
glutamate in the bone resorption by mature cells. Osteoclasts from
neonatal rats (3–5 days old) were prepared and cultured according to
the method of Arnett and Spowage (11)
. Briefly, for each
experiment, femora and tibiae pooled from four 3-day-old Wistar rats
were curetted into minimal essential medium containing 10% fetal
bovine serum and penicillin/streptomycin, acidified with 82 µl
concentrated HCl per 100 ml (MEM). Rabbit osteoclasts were prepared as
described in the first section of Materials and Methods. Cells were
settled in MEM onto 6 mm diameter wafers of elephant ivory at 37°C,
5% CO2 and nonadherent cells were removed after
45 min by vigorous washing in PBS. Experimental media containing MK801
were prepared from 100x aqueous stock solutions and control
experimental medium included a 1/100th volume of sterile water. The
ivory wafers were transferred to 4 ml pre-equilibrated experimental
media, with or without MK801, in 12-well tissue culture plates (four
wafers per well) and cultured for 26 h at 37°C, 5%
CO2. At the end of the incubation period, the
cells and pits were visualized and measured as before.
Calvarial explant assay
The effect of MK801 on bone resorption in calvarial explants was
assessed by measuring the release of 45Ca from
prelabeled neonatal mouse calvariae as described previously
(12)
. Briefly, neonatal (1 to 2 days old) mice were
injected subcutaneously with 1 µCi 45Ca
chloride (Amersham, Little Chalfont, U.K.) in 50 µl of physiological
saline. After 4 days, the animals were killed and the calvarial bones
were excised and divided into equal halves along the sagittal suture.
Half-calvariae were preincubated at 37°C, 5%
CO2 on stainless steel grids in 12-well tissue
culture plates in 1 ml of phenol red free BGJ culture medium (ICN Flow,
High Wycombe, U.K.) containing 5% fetal bovine serum and
penicillin/streptomycin. After an 18 h preincubation period, the
medium was replaced with experimental medium to which MK801 was added
from concentrated 100x aqueous stock solutions. Control experiments
included a 1/100th volume of sterile water. After a 48 h culture
period, the experiment was terminated and bones were dissolved in 50%
trichloroacetic acid. An aliquot of the medium and an aliquot of the
dissolved bone were analyzed for 45Ca by liquid
scintillation counting and the percentage of 45Ca
released from each bone was calculated. Differences between control and
experimental cultures were assessed by one-way analysis of variance.
Visualization of actin rings in osteoclasts
Rabbit osteoclasts were prepared as described above and settled
onto 10 mm glass coverslips for 45 min. After rinsing vigorously in
PBS, the coverslips were transferred to MEM and preincubated for 3 h at 37°C, 95% air, 5% CO2. The coverslips
were then transferred to fresh MEM with or without MK801 (at 30 or 100
µM) and incubated for an additional 3 h at 37°C, 95% air, 5%
CO2. Cells were fixed in 4% formaldehyde in PBS
for 5 min, then washed three times with PBS. F-actin was stained at
room temperature in the dark with 10-7 M
rhodamine-conjugated phalloidin (Sigma, St. Louis, Mo.) and visualized
by fluorescence microscopy.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). The NMDA receptor antagonist MK801 did not completely
block glutamate-induced currents, but the reduction was significant
(P<0.01). Glutamate-induced currents were also blocked to a
similar extent by 1 mM magnesium (data not shown). Application of MK801
alone did not cause significant changes in control currents (not
shown).
|
Coculture
MK801 produced consistent reductions in osteoclast size,
nuclearity, TRAP expression, and resorptive activity. Specifically, in
control cultures, cells were large, with up to 20 nuclei visible in
each cell (Fig. 2a, b
), whereas after exposure to MK801, most cells that
stained positive for TRAP were small and either mono- or binucleate
(Fig. 2c, d
). The effect of MK801 on the TRAP-positive cell
area was investigated at nine different concentrations between 3 and
100 µM and shown to be concentration dependent and significant above
50 µM (P<0.05 at 60 µM increasing to
P<0.001 at 80 µM and above) (Fig. 3
).
|
|
Pits in ivory wafers from control cultures were linked in trails and
overlapped (Fig. 4a
), sometimes covering up to 80% of the surface of the
dentine, whereas those exposed to MK801 were smaller, less overlapping,
and less continuous, forming isolated patches (Fig. 4b
). The
addition of MK801 for short periods at different times revealed that
addition early during the culture (days 2–4 or 2–7) was as effective
in inhibiting osteoclast formation and resorption of dentine as was
treatment with MK801 for the entire experiment. However, addition at
any time after day 7 did not have a significant effect in reducing
resorbing activity (Fig. 5
).
|
|
Mature osteoclast assay
Routinely, the area resorbed in control cultures of disaggregated
neonatal rat osteoclasts was greater than 25 x
103 µm2. In three independent
experiments, we observed no significant effect of MK801 (10–100 µM)
on resorption. The area of resorption in control cultures of one
representative study was 45 x 103
µm2) (±SD 12, n=4), or 0.03% of
the total area of dentine, compared with 38 x
103 µm2 (±SD
10 n=4, NS) in cultures containing 100 µM MK801.
Rabbit osteoclasts were similarly unresponsive to the effects of MK801, and there was no significant difference between the resorption area or TRAP activity in control and MK801-treated cultures at any of the concentrations tested.
Calvarial explant assay
Between 12 and 22% of the calcium label in mouse calvarial bones
in control cultures was released into the medium over 48 h in
culture. In three replicate experiments, MK801 (10–100 µM) had no
significant effect on calvarial resorption. In controls, 17.6% (±
SE 3.2%) of the incorporated label was released into the
medium compared with 17.7% (± SE 3%) in cultures treated
with 10 µM MK801 and 14.5% (± SE 3%) after treatment
with 100 µM MK801 (Fig. 6
).
|
Actin rings in osteoclasts
There was no detectable difference between F-actin ring formation
in osteoclasts treated with MK801 at 30 µM or 100 µM compared with
vehicle-treated controls (Fig. 7
).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Furthermore, in
studies some years ago, it was shown that glutamate induced chemotactic
responses in cells of the macrophage/monocyte lineage, including
osteoclasts, although at that time the mechanism for such an effect was
unknown (17)
.
Our electrophysiological studies provide direct evidence that receptors
expressed by osteoclasts are functional. However, an interesting
feature of this report is that the effect of the antagonist MK801
occurs predominantly during the earlier phase of osteoclastogenesis and
is less or nonexistent in mature rat, mouse, or rabbit osteoclasts in
cocultures, explant models, and disaggregated osteoclasts. This result
contrasts with other studies (7)
that showed inhibition of
bone resorption by preparations of isolated rabbit osteoclasts similar
to the ones we used. To clarify this point, we investigated the effect
of MK801 on actin ring formation, since it is well known that this
process is pivotal in polarization, attachment, and resorption activity
of osteoclasts (18
19
20
21)
. The lack of effect of MK801 in
disrupting actin rings in mature rabbit osteoclasts seems to support
the lack of significant effect of the agent on bone resorption except
in the coculture model, in which we show regulation of osteoclast
differentiation.
It is hard to reconcile the presence of receptors on osteoclasts and the effects of ligands and antagonists on membrane currents with a lack of effect on resorption of antagonists on isolated mature cells in culture. Our data suggest, therefore, that glutamate mediated changes in membrane currents are not involved in regulation of resorption in mature osteoclasts, but may be involved in other intracellular processes.
Whether or not MK801 affects mature osteoclast function, from our
studies it is clear that its effects on osteoclastogenesis are of
greater magnitude than those on mature cells. This raises a further
question, namely, whether the effects under those circumstances are
mediated by receptors on osteoclast precursors or on the osteoblasts
that are necessary to induce their fusion and activity. We have
demonstrated the expression of glutamate receptors on osteoblasts and
characterized their properties extensively (5
, 22)
. It is
therefore possible that effects of glutamate and/or antagonists on
osteoclastogenesis are mediated indirectly through osteoblasts. Direct
contact between hematopoietic progenitor cells and osteoblasts/stromal
cells is required for osteoclastogenesis, and an osteoblastic
membrane-associated factor has recently been described that fulfills
the role of osteoclast differentiation factor [ODF, also known as
tumor necrosis factor-related, activation-induced cytokine (TRANCE) and
receptor activator of nuclear factor 
B ligand (RANKL)].
ODF/TRANCE/RANKL interacts with the RANK receptor on osteoclast
precursors to drive osteoclastogenesis (23)
. Studies of
the effects of MK801 on osteoclastogenesis driven by RANKL in cultures
in the presence or absence of osteoblasts would discriminate between
direct and/or indirect effects of glutamate signaling on osteoclasts.
The concentrations of MK801 we have used [3–100 µM] are well
within the range considered to be selective in neuroscience studies.
However, it is not completely clear that the channel characteristics of
osteoclast NMDA receptors are identical to those expressed in the CNS.
One reason for this may be connected with the heteromeric nature of the
receptors. NMDA receptors are made up of four or five NMDAR1 and NMDAR2
subunits (24)
. The NMDAR2 subunits are divided into 2A, B,
C, and D. CNS type receptors are predominantly composed of 1/2A
heteromers, whereas in bone we have shown that NMDA receptors are the
much less well-characterized 1/2C or 1/2D type, which are known to have
gating characteristics that are different from the 1/2A receptors
(9)
.
Although the effects of NMDA receptor antagonists on osteoclastic bone
resorption are therefore clear in vitro, they are not well
characterized in vivo. Several knockout mice exist with
deletions of single glutamate receptors, but the NMDAR1 knockout does
not exhibit a profound skeletal phenotype and is not osteopetrotic
(25)
(Fig. 8
). Perhaps it is not surprising given the expression of numerous
different receptor subtypes by all the bone cells we have studied,
which would suggest redundancy in the system. Such redundancy would be
consistent with many other paracrine systems in bone. In a different
glutamate receptor transgenic animal, in which editing of the GluRB
subunit of an AMPA receptor abrogates its function, there is a mild
skeletal phenotype with reduced bone growth and kyphosis
(26)
. Since we have detected expression of this receptor
in bone, the phenotype may be directly linked to glutamate signaling in
bone, in which there may be less redundancy.
|
Although it is not surprising that the glutamate receptors expressed in
bone would be functional, there are only a small number of tissues in
which glutamate signaling has been characterized outside of the CNS. In
addition to earlier demonstrations of expression of functional
glutamate receptors in the adrenal gland, pancreas, lung, and some
regions of the gut (27
28
29
30)
, we have recently shown that
glutamate receptors are present in megakaryocytes and basal layer
keratinocytes (31
, 32)
, implicating excitatory amino acid
signaling in platelet formation and wound healing. There is now an
increasing body of evidence for neurotransmitter actions on bone
(33
34
35
36)
. These data therefore add to the possibility that
excitatory amino acids in particular, and neurotransmitters in general,
are probably more ubiquitously functional than previously thought. A
parallel may be drawn between this possibility and the growth in
understanding of the multiple functions of cytokines, which were
initially thought to be involved solely in signaling within the immune
system where they were discovered.
The ultimate implication of studies of neurotransmitter function in the musculoskeletal system is the possibility of manipulating bone mass by drug targeting those signaling systems in vivo. The blood–brain barrier may provide a useful way to engineer compounds that will act on the skeleton without affecting brain function.
|
ACKNOWLEDGMENTS |
|---|
from bones generously donated by Drs. Tom Curran and
Michi Yuzaki from the St. Jude Children’s Research Hospital in
Memphis. |
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  E. Hinoi, T. Takarada, K. Uno, M. Inoue, Y. Murafuji, and Y. Yoneda Glutamate Suppresses Osteoclastogenesis through the Cystine/Glutamate Antiporter Am. J. Pathol., April 1, 2007; 170(4): 1277 - 1290. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. J. Spencer, J. C. Utting, S. L. Etheridge, T. R. Arnett, and P. G. Genever Wnt signalling in osteoblasts regulates expression of the receptor activator of NF{kappa}B ligand and inhibits osteoclastogenesis in vitro J. Cell Sci., April 1, 2006; 119(7): 1283 - 1296. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Burt-Pichat, M. H. Lafage-Proust, F. Duboeuf, N. Laroche, C. Itzstein, L. Vico, P. D. Delmas, and C. Chenu Dramatic Decrease of Innervation Density in Bone after Ovariectomy Endocrinology, January 1, 2005; 146(1): 503 - 510. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Gu and S. J. Publicover Expression of Functional Metabotropic Glutamate Receptors in Primary Cultured Rat Osteoblasts. CROSS-TALK WITH N-METHYL-D-ASPARTATE RECEPTORS J. Biol. Chem., October 27, 2000; 275(44): 34252 - 34259. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. Rzeski, L. Turski, and C. Ikonomidou From the Cover: Glutamate antagonists limit tumor growth PNAS, May 22, 2001; 98(11): 6372 - 6377. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
P<0.01;
P<0.001.
