J. Cell Sd. 76, 155-165 (1985)
Printed in Great Britain © Company of Biologists Limited 1985
155
BONE CELLS PREDISPOSE BONE SURFACES TO
RESORPTION BY EXPOSURE OF MINERAL TO
OSTEOCLASTIC CONTACT
T. J. CHAMBERS AND K. FULLER
Department of Experimental Pathology, St Bartholomew's Hospital, West Smithfield,
London EC1, U.K.
SUMMARY
The cell-free endocranial surface of young adult rat parietal bones was used as a substrate for
osteoclastic bone resorption, either without prior treatment, or after incubation of the parietal bones
with collagenase or neonatal rat calvarial cells. Untreated, the endocranial surface consisted of
unmineralized organic fibres; incubation with calvarial cells or collagenase caused disruption and
removal of these fibres, with extensive exposure of bone mineral on the endocranial surface, without
morphologically detectable mineral dissolution. Neonatal rabbit osteoclasts resorbed bone to a
greater extent from parietal bones pre-incubated with calvarial cells or collagenase than from
untreated bones; mineral exposure and subsequent osteoclastic resorption were both increased if
calvarial cells were incubated with parathyroid hormone; removal of bone mineral after incubation
with calvarial cells removed the predisposition to osteoclastic resorption. These experiments
demonstrate that calvarial cells are capable of osteoid destruction, and indicate that one mechanism
by which osteoblasts induce osteoclastic bone resorption may be through digestion of the unmineralized organic material that covers bone surfaces, to expose the underlying resorptionstimulating bone mineral to osteoclastic contact.
INTRODUCTION
The shape and structure of bones depend upon intricate and dynamic patterns of
osteoclastic bone resorption during morphogenesis and restructuring. It has been
established that osteoclasts are derived from circulating mononuclear precursors
(Loutit & Nisbet, 1982; Marks, 1983); non-random bone resorption by
haematogenous cells implies that special properties develop in bone (Chambers,
1980) or in bone-lining cells (Rodan & Martin, 1981) at sites of resorption, which
induce osteoclasts to resorption activity. We have recently found that osteoclasts do
not resorb bone from native bone surfaces, in which mineral is covered by a layer of
unmineralized osteoid, but do so if bone mineral is artefactually exposed (Chambers,
Thompson & Fuller, 19846). This raises the possibility that bone-lining cells may
initiate osteoclastic resorption by removal of unmineralized organic material from
bone surfaces at sites appropriate for bone resorption, to expose underlying
mineralized bone to osteoclastic contact.
Although unmineralized osteoid is present on all bone surfaces (Raina, 1972;
Parfitt, 1976; Fornasier, 1977, 1980; Vanderwiel, 1980) except in areas of osteoclastic
Key words: osteoclast, osteoblast, bone resorption, parathyroid hormone, bone mineral.
156
T. J. Chambers and K. Fuller
resorption, it is generally too shallow to be readily identified in the scanning electron
microscope (SEM) on surfaces on which bone formation has ceased (Boyde & Hobdell, 1969) (resting surfaces); where bone formation is in progress, however,
mineralization lags behind osteoid deposition, and the relatively deeper mineral front
has a characteristic appearance that allows easy distinction in the SEM between the
organic and mineral surfaces in formative bone (Boyde & Hobdell, 1969). For this
reason we used the formative surface of rat parietal bone to test the capacity of bone
cells to remove this organic lining, and to assess the relationship between bone-cellmediated osteoid destruction and the susceptibility of bone to osteoclastic resorption.
MATERIALS
AND
METHODS
The effect of resident and neonatal rat calvarial cells on parietal endocranium
Wistar rats (220—250 g) were killed by decapitation after ether anaesthesia. Calvariae were
removed, cleaned of adherent soft tissues and divided sagitally. Each of the parietal bones was cut
into three specimens, 3 mm X 2 mm. The three specimens from one side were each incubated in 1 ml
RPMI 1640 (Gibco U.K.) supplemented with 10 % (v/v) heat-inactivated foetal calf serum (Flow,
Irvine, U.K.; FCS) and antibiotics (benzylpenicillin lOOi.u./ml and streptomycin 100(jg/m\)
(Glaxo, Middlesex, U.K.) in the wells (16 mm diameter) of Linbro (Sterilin, Teddington, U.K.)
tissue culture plates for 3—16 days. After incubation one specimen was fixed in 5 % glutaraldehyde
and prepared for scanning electron microscopy as previously described (Chambers et al, 19846) to
confirm the presence of cells, the remainder were immersed in ammonium hydroxide (0-25 M,
60min) to remove cells to allow visualization of endocranium, before glutaraldehyde fixation and
preparation for the SEM as above. The three specimens from the contralateral calvaria were
immersed in ammonium hydroxide to remove cells, washed in glass distilled water and either fixed
immediately or fixed after incubation for 3-16 days in culture medium.
The effect of neonatal rat calvarial cells was assessed using cells digested into suspension by
collagenase digestion of neonatal rat calvariae. Calvariae were removed from rats within 24 h of
birth, dissected free of soft tissues and periosteum, and incubated (2 h) in collagenase (Sigma type 11;
1 mg/ml in Hepes-buffered medium 199; Flow). The cell suspension was washed three times in
medium 199 containing 10% FCS and resuspended at 3 X 105 cells/ml in RPMI 1640 with FCS
and antibiotics. The suspension was divided and one half centrifuged (1000^, 10 min) for addition
of supernatant to (cell-free) control cultures. Calvariae were removed from 220-250 g rats as before,
immersed in NH4OH (60min), washed in distilled water, immersed in ethanol (2h) and dried.
Three specimens of parietal bone from one half were placed in 16-mm culture wells, and to these
were added 1 ml of calvarial cell suspension. These specimens were incubated overnight, washed
and placed in fresh culture medium with parathyroid hormone (PTH; Dr Zanelli, National Institute
for Biological Standards, Hampstead, U.K.) (1 i.u./ml) or vehicle, and incubation was continued
for 3, 6, 10 and 16 days (at least 5 animals at each time). Medium and hormone or vehicle were
replaced every 4—5 days. One specimen from the contralateral calvaria was fixed without incubation,
the remainder were treated as above but with substitution of the osteoblast suspension by supernatant. After incubation parietal bones were either fixed to confirm the presence of cells, or cells
were removed by immersion in NH4OH before fixation and preparation for SEM as before.
Quantitation of mineral exposure was performed on 15 specimens incubated with, and 15 specimens
incubated without, PTH for 10 days, by point counts at a screen magnification of X400. Alkaline
phosphatase cytochemistry was performed on samples of neonatal calvarial cell suspensions, as
previously described, after incubation overnight; although these cell digests are a heterogeneous
population, 70—85 % of cells were positive for alkaline phosphatase activity by the naphthol A5-B1
method (Bancroft & Stevens, 1977; Chambers, 1982).
To test the ability of neonatal rat calvarial cells to remove mineral, similar experiments were
performed on parietal bones pre-incubated in collagenase (5 mg/ml, 2h) before incubation in
culture medium, or with neonatal rat calvarial cells for 14 days before removal of cells and inspection
in the SEM.
Induction of osteoclastic resorption by osteoblasts
157
Incubation of osteoclasts on endocranial surfaces
Parietal bones from 220-250 g rats were rendered acellular with NH»OH and incubated for 10
days in RPMI 1640 with FCS and antibiotics, either with or without neonatal rat calvarial cells,
in the presence or absence of PTH (1 i.u./ml), as above. Parietal bones were also incubated in
bacterial collagenase (Sigma type II, 5 mg/ml, 2h). All parietal bones were washed, immersed in
NH4CI (to remove cells where present), washed in medium 199 and placed in 16mm diam.
Linbro culture wells. Osteoclasts were disaggregated from neonatal rabbit long bones as previously
described (Chambers, Revell, Fuller & Athanasou, 1984a). A sample (lml) of the suspension
containing osteoclasts released by curettage (2 ml from each long bone) was placed in each parietal
bone-containing Linbro well and incubated for 30min at 37°C. The parietal bones were then
washed in medium 199 and placed in fresh Linbro wells containing 1 ml of phosphate-free MEM
(Flow) supplemented with 10% FCS and antibiotics. After 18 h of incubation specimens were
either fixed in glutaraldehyde for assessment of osteoclast number (Chambers et al. 1984a) or
immersed in NH4CI to remove cells before glutaraldehyde fixation. All specimens were then
prepared for the SEM. The outline of all the excavations, each defined as an area of osteoclastic
resorption within a continuous border of unresorbed bone, was drawn on polythene on the SEM
screen and the image transferred to a graphic tablet input into an Apple He microcomputer
programmed for surface area measurements. Mean numbers of osteoclasts per specimen and pits
per specimen, and the mean surface area of endosteum resorbed per specimen, were calculated for
each group. The level of significance of differences between groups was calculated by Student's
<-test. Additional controls for these cultures consisted of bones incubated as above but inspected
without addition of osteoclasts; and bones incubated with neonatal calvarial cells and PTH for 10
days, then demineralized with 1M-HC1 (8 h) or EDTA (saturated solution, 3 days) before addition
of osteoclasts.
RESULTS
Effect of calvarial cells on endocranial surfaces
The endocranial surface of bones incubated in culture medium without cells was
indistinguishable from that of bones fixed immediately after removal of lining cells.
The surface consisted of sheets of parallel fibres punctuated by canalicular openings
(Fig. 1) and with occasional osteocytic lacunae at varying stages of development.
Parietal bones incubated with bone cells, whether resident or from neonatal calvariae,
showed disruption of this surface appearance by 3 days. The earliest change was focal
splitting of the continuous layer of surface fibrils to reveal deeper, not necessarily
parallel, layers of subjacent fibres (Fig. 2). Unlike the most superficial fibres, these
deeper fibres were often associated with nodules of varying size (Fig. 3), either
adjacent or along their length, embedded within them. These presumably represent
the most superficial deposits of mineral in osteoid: they were not seen in specimens
demineralized after culture. After further incubation (Fig. 4) these nodule-associated
fibres were also removed and the uneven, non-fibrillary surface of bone mineral was
revealed between frayed, partly digested bundles of fibrillary organic material. By 10
days the uneven, nodular appearance of the mineral front was observed over extensive
tracts of the endocranial surface. In specimens incubated for 10-16 days fibrillary
material had been removed from between these most superficial promotories of the
mineral front, to expose the coarse, raised, intermittent, irregular, approximately
parallel ridges of the mineralization front (Fig. 5), devoid of fibrillary organic
material: an appearance indistinguishable from that observed after treatment with
158
T. J. Chambers and K. Fuller
Figs 1-2. For legend see p. 160
Induction of osteoclastic resorption by osteoblasts
Figs 3-4. For legend see p. 160
159
160
T. J. Chambers and K. Fuller
collagenase or sodium hypochlorite (Chambers et al. 19846), and never seen on the
surface of untreated bone. Demineralization of such specimens in HCl or EDTA was
associated with loss of these coarse nodular protrusions and a return of fibrillary
material to the bone surface.
Mineral exposure could not be explained on the basis of residual collagenase from
the calvarial cell digests: the cells were extensively washed, and incubated in serum;
similar appearances were never seen in control parietal bones incubated with supernatant from the calvarial cell suspensions; although Sodek & Heersche (1981) demonstrated uptake of collagenase during bone cell extraction, they found that this
exogenous enzyme became undetectable within the first 2 days of incubation, while
osteoid destruction in our cultures continued throughout the 10-14-day incubation
period; furthermore, identical appearances were observed on parietal bones incubated with bone cells left in situ (Fig. 6), without enzyme treatment. We presume
that bone cells have removed the unmineralized osteoid from the parietal bone surface
and have exposed the mineral front.
We could detect no further change in the appearance of the mineral ridges after
incubation with cells. Beneath the surface at which mineralization is in progress the
mineral is solid, but we observed no areas in which the ridged appearance of the
mineralization front was replaced by the compact amorphous appearance of subjacent
completely mineralized bone: there was no evidence of simplification of the intricate
architecture of the mineral front to suggest that this too was resorbed by bone cells.
This was also the case when bone cells were incubated for 2 weeks on parietal bones
in which the mineral front had already been exposed by collagenase. These observations suggest that, while calvarial cells remove osteoid in culture, they seem incapable
of mineral dissolution. The resorptive pattern induced by calvarial cells was thus
quite distinct from the effect of osteoclasts, which not only cause sharply defined
Fig. 1. SEM of the endocranial surface of a young adult rat after 10 days incubation in
culture medium, showing an even, fibrillary surface punctuated by canaliculi. X 700.
Fig. 2. Endocranium after incubation with neonatal calvarial cells (6 days) showing more
extensive destruction of fibrillarly material. The superficial layer is disrupted and subjacent layers of fibrils, sometimes running in a different direction, have been exposed.
X700.
Fig. 3. Endocranial surface of rat parietal bone after 10 days incubation with neonatal
calvarial cells. There is extensive disruption of surface fibrillarly material. The deeper
layers of fibrils so exposed show pale nodules of mineral. X 2200.
Fig. 4. An area of endocranium from which all the superficial, fibrillary unmineralized
osteoid has been removed. Mineral is visible on the surface as coarse, intermittent raised
ridges, but unmineralized segments of osteoid remain between mineral projections.
X2700.
Fig. 5. SEM of endocranium after 16 days incubation with bone cells. The appearance
of this surface is indistinguishable from an anorganic preparation. The appearance is that
of the mineralization front; similar features were never seen on endocranial surfaces
incubated without calvarial cells or collagenase. X 1040.
Fig. 6. Endocranium after incubation (14 days) without removal of bone-lining cells. The
native bone cells have largely removed the osteoid and have exposed the nodular surfa.ce
of the mineral front. X 1350.
Induction of osteoclastic resorption by osteoblasts
Figs 5-6
Ml
162
T. J. Chambers and K. Fuller
excavations of both mineral and organic components, but also leave demineralized organic fibres, not mineral, in the base of their excavations (Chambers et al. 1984a, b).
Mineral exposure, defined as the appearance of raised, intermittent, non-fibrillary
nodules, was quantitated by random point counting in the SEM. It was not seen in
bones incubated without cells, or after demineralization, but covered the entire surface after treatment with collagenase. In parietal bones incubated with neonatal
calvarial cells they were observed over 33-8(±3 - 8)%of the bone surface, and in the
presence of PTH 45-7 (± 4-4) % in 15 specimens incubated for 10 days (P< 0-05).
Incubation of osteoclasts on endocranial surfaces
In preliminary experiments we found that in older rats (>400g) pre-existing
osteoclastic bone resorption lacunae were commonly present in unpredictable areas
of the calvarial endocranium. Young adult rats were chosen because pre-existing bone
resorption generally appears as distinctive confluent tracts of resorption in areas
around cranial sutures, areas not included in the central portions of the parietal bones
used for these experiments. Osteoclastic pits, easily identified in the SEM as welldefined excavations with a fibrillary base and blind-ending fibres in the pit walls
(Chambers et al. 1984a, b) were observed only very rarely in parietal bones incubated
without either neonatal calvarial cells or osteoclasts (mean 0-05 excavation per
specimen, ± 0-04 standard error; 62 specimens). Excavations were also noted very
rarely after incubation of calvarial cells without subsequent addition of osteoclast
(without PTH, 0-18 ± 0-17 (« = 11); with PTH, 0-38 ± 0-25 (n = 8)). This suggests
that the calvarial cell suspensions were substantially free of osteoclasts, a view reinforced by morphological examination of these cell populations, in which it is very
unusual to observe multinucleate cells. The slight increase in the number of excava-'
tions present, compared to parietal bones incubated without cells, however, presumably reflects the occasional presence of osteoclasts in calvarial bone cell preparations.
Parietal bones in all the groups to which osteoclasts were added showed increased
numbers of excavations (see Table 1). These were relatively uncommon on bones preincubated with culture medium alone and, when seen, were in areas of endocranium
Table 1. The number of excavations induced by rabbit osteoclasts in parietal bones
after various pre-treatments
Incubation
conditions before
osteoclast addition
No cells
Calvarial cells alone
Calvarial cells + PTH
Collagenase
Osteoclasts per
Surface area of
No. of
Excavations
specimens
per specimen excavations (fim2)
specimen
(mean ± S.E.M.) (mean ± S.E.M.) (mean ± S.E.M.) inspected
2-0 ±0-6
51 ±0-9
10-1 ±2.2
19-2 ±2-8
1340 ±210
1260 ±200
1430 ±180
2031 ±170
13 ±2
17 ±2
12 ±3
14 ± 2
41
32
30
28
All the differences between numbers of pits per specimen were significant (P<0 - 05). The
number of excavations on specimens pre-incubated without cells but with PTH was not significantly
different from PTH-free control (mean 3-2 ± standard error, 0-8 excavation per slice).
Induction of osteoclastic resorption by osteoblasts
163
torn or scratched by forceps during preparation, accidents associated with disruption
of surface osteoid. Pre-incubation with bone cells or collagenase increased the susceptibility of calvariae to osteoclastic resorption. This predisposition correlated with the
extent of mineral exposure; PTH thus induced an increased susceptibility to bone
resorption in the substrate that persisted in the absence of hormone.
There was no significant difference between the number of osteoclasts identified on
the different substrates, and the excavations, when present, were of similar surface
area. This suggests that the differing predispositions of the various substrates to
osteoclastic bone resorption operated through the presence or absence of a local
stimulus in the substrate to resorption, acting independently on each individual
osteoclast. A causal role for mineral exposure was reinforced further by the results of
experiments in which parietal bones were demineralized after incubation with neonatal calvarial cells. Sample bones from the same batch showed mineral exposure;
both demineralization techniques (HC1, 7 specimens; EDTA, 8 specimens) removed
the exposed mineral, and there was no evidence of osteoclastic bone resorption after
subsequent incubation with osteoclasts.
DISCUSSION
We have found that both resident bone cells and neonatal calvarial cells are able to
digest osteoid in vitro. Amongst the cell types present in our cultures, cells of the
osteoblastic lineage may be responsible: they are known to produce collagenase and
plasminogen activator, and produce increased amounts of both in response to agents
that stimulate bone resorption (Sakamoto & Sakamoto, 1982; Heath, Atkinson,
Meikle & Reynolds, 1984; Hamilton, Lingelback, Partridge & Martin, 1984). A
causal role for cells of the osteoblastic lineage in osteoid destruction in our cultures
is further suggested by the hormone-dependent nature of the process, since
osteoblasts appear to be the cell type in bone with PTH receptors (Rodan & Martin,
1981; O'Grady & Cameron, 1971; Silve, Hradek, Jones & Arnaud, 1982; Chambers
& Dunn, 1983). One report (Rao, Murray & Heerche, 1983) additionally describes
PTH receptors on osteoclasts, but the osteoid destruction we have described is unlikely
to be due to osteoclastic cells in the calvarial cell digests since we detected no evidence
of mineral resorption typical of osteoclasts (Chambers et al. 1984a).
Osteoid destruction by neonatal calvarial cells raises the possibility that they, like
osteoclasts, resorb bone. Against this, there was no evidence that such cells were
capable of resorption of bone mineral; unlike osteoclasts, which resorb both mineral
and organic components of mineralized bone (Chambers et al. 1984a, b), neonatal
calvarial cells appear to be capable of removing only osteoid. An alternative possibility
is that the hormonally responsive osteoid destruction may represent a phase in the
process by which osteoblasts, which have PTH receptors, induce osteoclasts, which
probably lack them, to increased bone resorption in response to PTH. We have
previously found that osteoclasts are stimulated to resorptive activity by mineralized
but not by demineralized or unmineralized bone (Chambers et al. 1984a). In the
present experiments the susceptibility of the variously pre-incubated substrates was
164
T.J. Chambers and K. Fuller
similarly proportional to the degree of mineral exposure induced by the treatment.
Correspondingly, PTH increased both the extent of mineral exposure by bone cells
and the susceptibility of the substrate to subsequent bone resorption. A causal role for
mineral exposure as the mechanism through which calvarial cells digest predisposed
bone to osteoclastic resorption is also supported by the results of the experiments
in which parietal bones were demineralized after incubation with bone cells; this
treatment completely removed the bone-cell-induced substrate predisposition to
osteoclastic resorption.
Young adult rat parietal bones were chosen for this study because they enable an
easy distinction to be made between mineral and organic surfaces in the SEM, and are
relatively free of pre-existing osteoclastic-resorptive activity. The substrate does,
however, represent a relatively rigorous test for the hypothesis, since the very thickness of the osteoid (approximately 10 ^m), which makes an SEM study possible,
necessitates the removal of a greater depth of unmineralized osteoid before the mineral
layer is revealed. If osteoid on resting surfaces, where unmineralized osteoid forms
a layer a few hundred nanometres in depth, is removed in vivo at a rate similar to that
we have observed in vitro, then cells of the osteoblastic lineage could effect mineral
exposure on resting surfaces (which comprise 80% of bone surfaces; Parfitt, 1976)
within a relatively short time.
It seems likely that, in vivo, mineral exposure would be followed rapidly by
osteoclastic resorption, and thus that such mineral-exposed surfaces would be rarely
observed. However, it is of interest that in a model of bone resorption in which
osteoclasts can be induced to resorb bone simultaneously in a well-defined temporospatial sequence in vivo, the earliest observation, before multinucleate cells
appear, is obliteration of unmineralized osteoid from the bone surface (Tran Van,
Vignery & Baron, 1982). Also consistent with a requirement for mineral exposure
before bone can be resorbed is the well-recognized preference of osteoclasts for resorption of mineralized compared to unmineralized bone in osteomalacia (Parfitt, 1976),
and the failure of osteoclasts to resorb demineralized, but not mineralized, bone
implants (Glowacki, Altobelli & Mulliken, 1981).
Many of the agents that stimulate osteoclastic bone resorption appear not to
stimulate osteoclasts directly (Rodan & Martin, 1981; Chambers & Dunn, 1983), but
do exert morphological and functional effects on osteoblasts (see Rodan & Martin,
1981); this suggests that osteoblasts play a major role in the control of osteoclastic
bone resorption. One of the effects that stimulators of bone resorption have on
osteoblasts is enhancement of secretion of collagenase and plasminogen activator
(Sakamoto & Sakamoto, 1982; Heath et al. 1984). A final common pathway for
osteoblastic stimulation of osteoclastic bone resorption in response to local and
systemic influences may be through proteolytic digestion of surface osteoid to expose
bone mineral to osteoclastic contact.
This work was supported by the Medical Research Council, the Wellcome Trust and the North
East Thames Regional Health Authority. We are grateful to Miss Caroline Judge for typing the
manuscript.
Induction of osteoclastic resorption by osteoblasts
165
REFERENCES
BANCROFT, J. D. & STEVENS, A. (1977). Alkaline phosphatase by the naphthol A5-B1 method. In
Theory and Practice of Histological Techniques (ed. J. D. Bancroft & A. Stevens), pp. 291.
Edinburgh, London, New York: Churchill Livingstone.
BOYDE, A. & HOBDELL, M. H. (1969). Scanning electron microscopy of lamellar bone. Z.
Zellforsch. mikrosk. Anat. 93, 213-231.
CHAMBERS, T . J. (1980). The cellular basis of bone resorption. Clin. Orthop.rel.Res. 151, 283-293.
CHAMBERS, T . J. (1982). Osteoblasts release osteoclasts from calcitonin-induced quiescence. J . Cell
Sci. 57, 247-260.
CHAMBERS, T . J. & DUNN, C. J. (1983). Pharmacological control of osteoclastic motility. Calc.
Tiss. Int. 35, 566-579.
CHAMBERS, T . J., REVELL, P. A., FULLER, K. & ATHANASOU, N. A. (1984a). Resorption of bone
by isolated rabbit osteoclasts. J. Cell Sci. 66, 383-399.
CHAMBERS, T . J., THOMPSON, B. M. & FULLER, K. (19846). The effect of substrate composition
on bone resorption by rabbit osteoclasts. J . Cell Sci. 70, 61-71.
FORNASIER, V. L. (1977). Osteoid: an ultrastructural study. Hum. Path. 8, 243-254.
FORNASIER, V. L. (1980). Transmission electron microscopy studies of osteoid maturation. Metab.
Bone Disease rel. Res. 25, 103-108.
GLOWACKI, J., ALTOBELLI, D. & MULLIKEN, J. B. (1981). Fate of mineralised and demineralised
osseous implants in cranial defects. Calc. Tiss. Int. 33, 71-76.
HAMILTON, J. A., LINGELBACK, S. R., PARTRIDGE, N. C. & MARTIN, T . J. (1984). Stimulation
of plasminogen activator in osteoblast-like cells by bone-resorbing hormones. Biochem. Biophys.
Res. Commun. 122, 230-236.
HEATH, J. K., ATKINSON, S. J., MEIKLE, M. C. & REYNOLDS, J. J. (1984). Mouse osteoblasts
synthesise collagenase in response to bone resorbing agents. Biochim. Biophys. Acta 802,
151-154.
LOUTIT, J. F. & NISBET, N. W. (1982). The origin of the osteoclast. Immunology 161, 193-203.
MARKS, S. C. (1983). The origin of osteoclasts. J. oral Path. 12, 226-256.
O'GRADY, R. L. & CAMERON, D. A. (1971). Demonstration of binding sites of parathyroid hormone in bone cells. In Endocrinology 1971: Proc. 3rd Int. Symp. pp. 376-379. London:
Heinemann.
PARFITT, A. M. (1976). The actions of parathyroid hornione bone. Part I: Morphological and
kinetic approaches to bone tumour. Metabolism 25, 810-844.
RAINA, V. (1972). Normal osteoid tissue. J . clin. Path. 25, 229-232.
RAO, L. G., MURRAY, T . M. & HEERSCHE, J. N. M. (1983). Immunohistological demonstration
of parathyroid hormone binding to specific cell types in fixed rat bone tissue. Endocrinology 113,
805-810.
RODAN, G. A. & MARTIN, T . J. (1981). The role of osteoblasts in hormonal control of bone
resorption. Calc. Tiss. Int. 33, 349-351.
SAKAMOTO, S. & SAKAMOTO, M. (1982). Biochemical and immunohistochemical studies on
collagenase in resorbing bone in tissue culture..?, periodont. Res. 17, 523-526.
SILVE, C. M., HRADEK, G. T., JONES, A. L. & ARNAUD, C. D. (1982). Parthyroid hormone
receptor in intact embryonic chicken bones: characterisation and cellular localisation.^. CellBiol.
93, 379-386.
SODEK, J. & HEERSCHE, J. N. M. (1981). Uptake of collagenolytic enzymes by bone cells during
isolation from embryonic rat calvaria. Calc. Tiss. Int. 33, 255-260.
TRAN VAN, P., VIGNERY, A. & BARON, R. (1982). Cellular kinetics of the bone remodelling
sequence in the rat. Anat. Rec. 202, 445-451.
VANDERWIEL, C. J. (1980). An ultrastructural study of the components which make up the resting
surface of bone. Metab. Bone Disease rel. Res. 25, 109-116.
{Received 13 November 1984-Accepted 11 January 1985)