J Clin Pathol 1998;51:831-837

Study of the cell biology and biochemistry ofcherubism

John Southgate, Usha Sarma, John V Townend, Jeffrey Barron, Adrienne M Flanagan

AbstractAims-To establish whether the multinu-cleate cells in lesions of patients withcherubism are also osteoclasts and if thisis the case whether they were responsive tocalcitonin; to carry out cytogenetic studieson two members of the same familyaffected by cherubism in an attempt toidentify any major chromosomal defects;and to perform an in-depth modernbiochemical study of four children in thesame family.Subjects and methods-Four related chil-dren with cherubism were studied. Tissuetaken from one of the children at electivedecompression of an optic nerve was sub-mitted to in vitro bone resorption studies.Cytogenetic studies were done on two ofthe children and biochemical studies onall four.Results-The multinucleate cells in thecherubic lesions were shown to be osteo-clasts since they synthesised tartrate re-sistant acid phosphatase, expressed thevitronectin receptor, and resorbed bone.Bone resorption by the cultured multinu-cleate cells was significantly inhibited bycalcitonin. High resolution cytogeneticstudies failed to detect any chromosomalabnormalities in two children with cheru-bism. The biochemistry profile of all fourchildren with cherubism showed thatserum calcium, parathyroid hormone,parathyroid related hormone, calcitonin,and alkaline phosphatase were within nor-mal levels. Urine analysis of pyridiniumand deoxypyridinium cross links, hydro-xyproline, and calcium in relation to urinecreatinine were measured to assess boneresorption in these children, and thevalues were at the upper end ofthe normalrange in all four.Conclusions-Further studies are re-quired to determine whether calcitonintreatment will control this grossly de-forming disease until the time when thephysiological changes that occur at pu-berty rectify the pathology. It is notrecommended that biochemical markersofbone resorption are used in isolation tomonitor the activity of cherubism in indi-viduals because the results are based on asmall number of children and because ofreports of marked interindividual varia-tion in the levels of these markers,particularly in children.(, Clin Pathol 1998;51:831-837)

Keywords: cherubism; osteoclast; calcitonin; boneresorption; parathyroid hormone

Cherubism is a grossly deforming disease ofthe jaw and is generally inherited as anautosomal dominant trait,' 2 with incompletepenetrance and variable expressivity.' Sporadiccases have also been reported.3` At birth, thejaw of the affected individuals is apparentlynormal and the disease usually presentsbetween two and five years of age, althoughthere is a report of the diagnosis being made asearly as 18 months.6 The disease progressesuntil puberty, at which time it stabilises andregresses so that by the time the individuals arein their mid-twenties, their facial features aregenerally not recognised as having beenaffected by the disease. However, residualdeformity of the jaws has been reported.6 7

In a typical case, painless symmetric, multi-locular, radiolucent, expansile lesions developat the angle and vertical ramus of the lower jaw.In more severely affected cases, the lesionsramify through both the upper and lower jaws,with extensive bone loss. This is associatedwith interference with tooth development, andpremature tooth exfoliation may occur. Non-facial bones are generally not affected, al-though bifid ribs have been reported. A furtherunexplained feature of the condition is markedcervical lymphadenopathy.6 8The histology of this lesion is well

described.9 10 It comprises a mononuclearfibroblastic stroma in which large numbers ofmultinucleate cells are present and there maybe cyst formation. Newly formed osteoid orbone is often seen at the periphery of lesionaltissue. The histology is virtually identical tothat seen in peripheral and central giant cellgranulomata of jaw." Giant cell granuloma ofthe jaw and osteoclastoma are also histologi-cally similar but the former is usually unilat-eral, whereas cherubism is a symmetric lesion,involving both sides of the jaw. Osteoclastomararely occurs at this site. The differential diag-nosis of cherubism includes fibrous dysplasiaand hyperparathyroidism, as these lesions alsocontain large numbers of osteoclasts. However,the histological features of fibrous dysplasiadiffer from those in cherubism in that, in theclassical form, Chinese-figure-like trabeculaeof immature bone, which are not rimmed byosteoblasts, are present within the proliferatingstroma. Furthermore, polyostotic fibrous dys-plasia first presents in the second or third dec-ade (for a review, see Riefkohl et al'2).Hyperparathyroidism, as opposed to cheru-bism, rarely affects the jaw in an isolated man-ner. The histology also differs from cherubismin that it does not contain the mononuclearstromal cell population which is characteristicof cherubism. Finally, peritrabecular fibrosis is

Department ofMaxillofacial Surgery,St Richard's Trust,Chichester, WestSussex, UKJ V Townend

Department ofChemical Pathology,Worthing Hospital,West Sussex, UKJ SouthgateJ Barron

Departmnent ofHistopathology,Imperial Coliege ofScience, Technologyand Medicine at StMary's, London W2,UKU SarmaA M Flanagan

Correspondence to:Dr Adrienne M Flanagan,Department ofHistopathology, ImperialCollege of Science,Technology and Medicine atSt Mary's, Norfolk Place,London W2 1PG, UK;email: a.flanagan(ic.ac.uk

Accepted for publication1 June 1998

831

Southgate, Sarma, Townend, et al

Child A Child Bb 1979 b 1981

Figure 1 The family tree.

also a feature of hyperparathyroidism, but notof cherubism. Serum concentrations of par-athyroid hormone and calcium also help todistinguish these lesions.

It has been shown previously that the multi-nucleate giant cells in osteoclastomas and giantcell granulomas of the jaw are osteoclasts, on

the basis that these cells are able to resorb bonein vitro.'3 '4 Therefore one of the aims of thisstudy was to establish whether the multinucle-ate cells in cherubic lesions are also osteoclastsand if this is the case whether they wereresponsive to calcitonin. Harris has previouslyreported the successful treatment of giant cellgranulomata of the jaw with calcitonin," andthis report prompted us to consider whether itmight also be possible to treat cherubism withcalcitonin. Such treatment may reduce theneed for surgery and may result in earlyregression of the disease.We also took the opportunity to carry out

cytogenetic studies on two members of thesame family affected by cherubism in an

attempt to identify any major chromosomaldefects. Finally we performed an in depthmodern biochemical study of four children inthe same family, as such a study does not atpresent exist.

Case historyThe four children (A-D) shown in the simpli-fied family tree (fig 1) inherited the diseaseknown as cherubism through their affectedfathers, who were brothers. The antecedents,many siblings, and descendants of the parentsof the index cases (not shown) are notable forthe absence of further cases. It is possible,therefore, that the disease arose as a recentmutation and only first presented clinically inthe fathers of the children, though clearly thegene must have been present in one of thegrandparents of the four children, as theaffected fathers were not twins. The fathersremember having the disease in a mild formand we have confirmed this on being shown a

photograph of them taken when they wereabout 10 years of age (not shown). Inphotographs of the fathers taken at the age of

18, they no longer had recognisable features ofthe disease, which is consistent with its normalcourse.

Child A was first recognised as havingcherubism at four years of age. The diagnosiswas made on the basis of premature exfoliationof a deciduous lower incisor, marked bilateralcervical lymphadenopathy, and the discoveryof the characteristic radiolucencies in bothjaws. Since then the disease has progressed (fig2), and has become more severe than it was infathers A and B. Subsequent radiologicalexamination of her cousins (child C at 20months, and child D at two years) showed thatthey had similar disease. Child B is minimallyaffected, with radiolucencies only in both verti-cal rami of the lower jaw and possibly in bothmaxillary tuberosities. The condition hasprogressed steadily in children A, C, and D,and all three show marked facial deformitiesand displacement or loss of teeth (fig 3). ChildA has been treated surgically for severe propto-sis of her right eye and child B has had jaw sur-gery. Children A, C, and D are currentlyundergoing medical treatment for cherubism.

MethodsIN VITRO BONE RESORPTION STUDIESElective decompression of the right optic nervewas undertaken on the eldest of the fourchildren (child A). Tissue was taken at the timeof surgery and fixed in formalin for histologicalexamination, "dabbed" onto glass slides forimmunohistochemistry, and a small amount offresh tissue placed in phosphate buffered saline(PBS) for tissue culture studies. The latter wastransported to the laboratory on ice, where itwas minced into tiny explants and sedimentedonto wafers of devitalised bovine bone (3 x 3 x0.1 mm), as described previously,'6 in a 24 wellplate (Sterilin) containing Medium 199 withpenicillin/streptomycin and glutamine (allfrom Gibco Life Technologies). After 45minutes the bone slices were washed vigorouslyand placed in a 96 well plate in minimumessential medium with Hanks' salts (Gibco LifeTechnologies), with 0.1% bovine serum albu-min (Sigma), L-glutamine, 100 IU benzylpenicillin/ml (Gibco Life Technologies), and100 ,ug streptomycin/ml (Gibco), and incu-bated at 37°C in a humidified atmosphere of5% CO2/95% air. Salmon calcitonin (sCT)(kindly provided by Rhone-Poulenc Rorer,Pennsylvania, USA) was added to the culturesat final concentrations of 0.1 pg/ml to 10ng/ml. After 18 hours the bone slices werewashed in PBS, fixed in formalin, and stainedin 0.1% toluidine blue (Sigma) for fiveminutes. The area of bone resorption on thebone slices was quantified using reflected lightand an image analysis system (Seescan). Boneresorption was expressed as the total (mean)plan area (SEM) of resorption per bone slice.Statistical significance was determined by Stu-dent's t test.The cells were later removed from the bone

surfaces by immersing the cells in sodiumhypochlorite for five minutes. They were thensputter coated with gold and the bone slices intheir entirety were scanned for bone resorption

Antecedents unknown

Child C Child Db 1982 b 1984

832

Cell biology and biochemistry of cherubism

using a Cambridge Stereoscan S90 scanningelectron microscope (Cambridge Instruments)(with the kind permission from Professor ofTJ Chambers and technical assistance from MrR Moss in Department of Histopathology, StGeorge's Hospital Medical School, LondonSW17).

Cells were also sedimented onto bone slicesto stain them for tartrate resistant acidphosphatase. Staining for tartrate resistant acidphosphatase and the a,3, subunit of thevitronectin receptor using monoclonal anti-body 23c6,'7 18 was carried out using conven-

tional techniques. The substrate used for theformer was alkaline phosphatase and for thelatter, naphthol AS-B 1 phosphate. Tartrateresistance was assessed in the presence of 0.05M sodium tartrate. The cultures were counter-stained with toluidine blue.

CYTOGENETIC STUDIESVenous blood (10 ml) was taken from two sib-lings who are severely affected by cherubism.High resolution cytogenetic studies were car-ried out by conventional techniques on thesesamples. "

Figure 2 ChildA at 4, 8, 9, and 13 years of age (A-D). The features of cherubism are most severe at the age of 13. At 4years of age, the child would be considered normal if the photograph were examined in isolation, but when studied insequence it is clear that the early stages of cherubism are present and that there is slight puffiness of the cheeks. To preserveanonymity the eyes are obscured and the severe proptosis is not seen.

833

Southgate, Sarma, Townend, et al

--.....

Figure 3 Photograph (A) and x ray (B) of child C at 12 years of age. The deformity of thejaw seen in the photograph isthe result of marked overgrowth of the mandible and maxilla with cystformation in the bone. Note the loss of teeth andpartial obstruction of the nasal cavity.

BIOCHEMICAL ANALYSISBlood and urine samples were collected afterovernight fasting. The second void urine sam-ple was collected on rising and used for analy-sis. Routine determinations of calcium andalkaline phosphatase were made on an SMACII analyser (Technicon Instruments). Urinehydroxyproline was measured by the method ofProkofand Udenfriend.'0 Total pyridinium anddeoxypyridinium cross links were measured byhigh performance liquid chromatography,using a fluorimetric detector.2" Serum osteocal-cin concentrations were determined by radio-immunoassay using a kit supplied by IncstarCorporation. Serum parathyroid hormone(PTH) was measured using Magic Lite (CibaCorning). Parathyroid hormone related pep-tide (PTH-rP) (1-86) was measured using atwo site immunoradiometric assay from theNichols Institute (San Juan Capistrano, Cali-fornia, USA). All these assays were performedaccording to the manufacturer's instructions.Age matched reference ranges for serumcalcium and total alkaline phosphatase werederived from the data base at Worthing Hospi-tal and that for serum calcitonin from theendocrine unit at Imperial College School of

.ji...:

Medicine at Hammersmith Campus, London.We are grateful to Dr W Frazer at The RoyalLiverpool Hospital for the PTH-rP referencerange for children, as this is not available in apublished report. The reference range forserum osteocalcin was provided in the manu-facturer's instructions. Those for urine hy-droxyproline to creatinine ratios and for urine

A

;

E

I, ;.......

Figure 4 Haematoxylin and eosin stained tissue sectionfrom thejaw of childA at the age of 13 years. There are

many multinucleate cells set in a mononuclear cellpopulation ofstromal cells.

Figure S (A) Multinucleate cellsfrom thejaw ofchildA,dabbed onto a glass slide, are immunoreactive formonoclonal antibody 23c6 (this identifies the vitronectinreceptor), which is expressed on the surface membrane ofosteoclasts (x200). (B) A scanning electronphotomicrograph of a surface excavation on devitalisedbovine bone which was resorbed by an osteoclastfrom thejaw tumour of child A.

834

Cell biology and biochemistry of cherubism

;-C,

3000x

E

aL).2

ux 2000

a)c

0

a)

o 1000

0

0.

0

a)

o 0o

_-

*

0 0.001 0.01 0.1

*

sCT (ng/ml)Figure 6 Plan area of bone resorbed per bondhours of incubation of cellsfrom a cherubism IA, in the presence and absence ofsalmon calci(n = 6). *p < 0. 001 v control (no sCT addea

pyridinium and deoxypyridiniumwere previously reported by WhartcRauch et al, respectively.22 23

ResultsHISTOLOGY AND IN VITRO STUDIESHistological sections of the lesionspindle cell stroma in which therenumbers of multinucleate giant ce

The multinucleate cells stained strmonoclonal antibody 23c6 (fig 5) aresistant acid phosphatase (not si

also found excavations on the surnbone slices when cells from thelesions were cultured on them (fig 5was reduced significantly in a dosomanner in the presence of sCT (fig

CYTOGENETIC STUDIESHigh resolution cytogenetic studiedetect any chromosomal abnormalit

BIOCHEMISTRY

Serum calcium was found to be withmal range in the four children with (

We also measured intact serum PTcondition for the first time, and repunequivocally normal in all fou

children (table 1). Serum alkaline phosphataseand osteocalcin were also found to be normalfor ages, and electrophoresis showed mainlybone subfractions, which is the normal isoen-zyme pattern for the children's ages. PTH-rPwas certainly not raised in the three mostaffected children (table 1), but this measure-

* ment was made after they had been treatedT* with calcitonin for a month as the assay was not

available during the early part of the study andold samples were not suitable for analysis. Thefinding of a normal PTH-rP concentration inthe least affected but untreated child B istherefore important to note (table 1).

1 1Urine from all four children showed high

fasting calcium to creatinine ratios comparedwith published reference ranges.24 25 However,

e slice after 18 urine salt excretion, which is now recognised toesion in child be associated with hypercalciuria, was not

ln ).........measured at this time. Urine analysis of

hydroxyproline to creatinine ratios gave valueslinks at the upper end of reported normal range

.(table 1). Likewise, the urine concentrations

of pyridinium and deoxypyridinium cross linksappeared to be at the upper end of the normalrange (table 1).2

showed a Discussion

were large We found that the multinucleate cells in cheru-

flls (fig 4). bic lesions had phenotypic characteristics of

ongly with osteoclasts: they labelled with antibody 23c6,nd tartrate which identifies the atj3 subunit of the

iown). We vitronectin receptor, and with tartrate resistant

face of the acid phosphatase. Furthermore cells from the

cherubic cherubic lesions resorbed bone in vitro and this

i); this area function was reduced in the presence of sCT,

e-response implying that the multinucleate cells were

6). responsible for the bone resorption. Our

findings raise the possibility that treatment ofcherubic individuals with calcitonin may cur-

s failed to tail the disease and reduce the need for surgery.

ties. It is not possible to determine from these data

whether bone resorption was reduced bydownregulating osteoclast activity or whether

lin the nor- osteoclast numbers were reduced in the

cherubism. presence of calcitonin, since the osteoclasts

rH in this were not counted. However, it has previouslyort it to be been shown that inhibition of calcitonin

r affected induced bone resorption is associated with

Table 1 Biochemical results in four children with cherubism

ReferenceVariable ranget ChildA Child B Child C Child D

Age (years) 14 12 10 8

Serum valuesTotal calcium (mmol/l) 2.1-2.7 2.56 2.44 2.61 2.58Alkaline phosphatase (IU/1) 81 (50-150) 283 (100-400) 353 (100-400) 398 (100-300)PTH (pmol/l) 1.3-7.6 2.9 1.1 2.3 3.0PTH-RP (pmolIA) <1.4 0.4* 0.2 0.3* 0.4*Osteocalcin (ng/ml) <11 8.58 (-) 9.86 10.0Calcitonin (ng/ml) <0.08 <0.04 <0.04 <0.04 <0.04Serum electrophoresis Normal in all four children

Urine (fasting samples)Calcium/creatinine (mmol/mol) <700 1097 839 733 702Hydroxyproline/creatinine 118 (22-94) 216 (45-123) 182 (39-104) 132 (36-107)

(mmol/mol)Pyridinoline/creatinine (nmol/mmol) 198 (117+69) 302 (186±80) 244 (186+80) 337 (159+52)Deoxypyridinoline/creatinine 55 (29±23) 84 (50+23) 86 (50±23) 124 (43±13)

(nmol/mmol)

Pyridinoline and deoxypyridinoline ratios to creatinine are raised compared with age related reference ranges in the boy child D at8 years.23 The other children show cross links at the higher end of the age related reference range. tThe figures in brackets showthe normal age related ranges.23 *Denotes samples taken after one month's treatment with nasal calcitonin.

835

Southgate, Sarma, Townend, et al

reduced cell motility and a dramatic shapechange, and that these morphological changesare reversed when the calcitonin is removedfrom the medium.26 This suggests that calci-tonin inhibits bone resorption by inducing a

quiescent state in osteoclasts and not by reduc-ing osteoclast numbers.

It is thought that osteoblasts play a centralrole in osteoclast recruitment, formation, andresorption,27 28 and it has been proposed thatosteoclast-rich lesions-including the osteo-clastoma, central and peripheral granulomas ofthe jaw, and osteoclast-rich malignant fibroushistiocytoma-contain a cell of the osteoblastlineage which is responsible for the recruitmentof the osteoclast or its precursor into theselesions.'3 14 29 It is likely that a similar mech-anism accounts for the presence of osteoclastsin cherubic jaws, particularly since there is a

conspicuous stromal cell population in thesetumours, similar to that described in othergiant cell lesions of the jaw. Since calcitoninmediates its effect directly on the osteoclastand not on the osteoblast-type cell, it would bepredicted that if the disease were to remit as a

result of calcitonin treatment, it should relapseif the treatment was stopped. Interestingly, thisis what happened to one of the children fromHarris's group. 5 However, when treatment wasrestarted and given for a further six months, thedisease went into remission and there was no

sign of a recurrence five years later. Harris'sother two children were identified as beingclinically and radiologically healed after a

single course of calcitonin. The fourth patientin that group was an adult, who responded wellto treatment initially but required surgicalexcision for the lesion to be cured. It is not clearwhy these children were cured with calcitonintreatment, but it is possible that their diseasewas curtailed by the treatment and then wentinto spontaneous remission at puberty. Thispossibility is proposed because cherubism,which has marked similarities to giant cellgranulomas of jaw, is known to remit spontane-ously when children reach puberty. 12 Thereason why this occurs is not clear but sinceosteoclast formation is reduced by sex

steroids,30 and since plasma concentrations ofoestradiol and testosterone are increased atpuberty, it suggests that the genetic defectresponsible for the localised increase in osteo-clasts in cherubism is overridden and normal-ised by the physiologically increased synthesisof sex steroids.The most common use for calcitonin, as an

antiresorptive agent, is in postmenopausalwomen for the prevention of osteoporosis.' Inthis condition it reduces bone loss transiently,but over a period of time the effect of the drugdeclines, an effect known as "escape."32 Recentexperiments have shown that an importantcomponent of escape from inhibition of boneresorption by osteoclasts is resistance of themature osteoclasts to calcitonin, as a result ofdownregulation of the expression of thecalcitonin receptor.33 If this escape phenom-enon were to occur to children, calcitonintreatment would be of little value as children

with cherubism would require treatment forseveral years.Documentation of the biochemistry in the

aggressive form of cherubism is limited. Serumcalcium has generally been reported as normaland we confirmed this in the current study.This is not unexpected in view of the localisedsite of the disease and it is consistent with thereport from Davis et al, who found that serumcalcium and PTH concentrations were withinthe normal range in a patient with multifocalosteoclastic lesions.34 However, in contrast toour findings in cherubism, this individual hadhigh levels of PTH-rP, alkaline phosphatase(not separate components), and serum calci-tonin. Urine analysis was not performed in thisstudy.The urine hydroxyproline to creatinine ratio

is largely a reflection of total collagen in thefasting state: it is not specific for bone collagenand furthermore it is known to be significantlyhigher in growing children than in adults.22 Wealso measured bone turnover using the fluores-cent assay for total pyridinium and deoxypyrid-inium cross links (which are newer and morespecific markers of bone turnover, the latterbeing more specific for bone collagen35). Theconcentrations of these markers were repeat-edly at the upper end of the reference data ofRausch et al,23 and it is interesting to speculatethat they may reflect the activity of the localisedbone disease in these children. However, it iswell documented that there is marked day today interindividual variation, especially in chil-dren, in the levels of urinary pyridinium anddeoxypyridinium cross links. Hence we do notrecommend that this assay is used in isolationas a means of monitoring children with cheru-bism, particularly since our data are derivedfrom a small number of patients.

Identification of the genetic mutation whichis responsible for cherubism would have wide-spread implications in terms of understandingthe biology of the osteoclast. The osteoclast is arare cell in the human body and its poor avail-ability makes study of its biology difficult.Studies which have had the greatest impact onour understanding of the biology of the osteo-clast have come from recent molecular investi-gations of natural and genetically engineeredmutations in mice.3$35

Cherubism, as one of the very few geneticallydetermined osteoclastic lesions in the humanbody, is an ideal disease for the study of osteo-clast biology and pathology.

We are grateful to Beverley Holland, Cytogenetic Unit, StMary's Hospital, London, for performing the cytogenetic stud-ies.

1 Peters WJN. Cherubism: a study of twenty cases from onefamily. Oral Surg Oral Med Oral Pathol 1979;47:307-11.

2 Jones WA. Familial multilocular cystic disease of the jaws.Am JCancer 1933;17:946-50.

3 Topazian RG, Cotish ER. Familial fibrous dysplasia of thejaws (cherubism): report of a case. Oral Surg 1965;23:559-66.

4 Grunebaum M, Tiqva P. Non familial cherubism: report oftwo cases. J Oral Surg 1973;31:632-5.

5 Kaugars GE, Svirsky JA. Cherubism: diagnosis, treatmentand comparison with central giant cell granulomas andgiant cell tumour. Oral Surg Oral Med Oral Pathol 1992;73:369-74.

6 Killey HC, Seward GR, Kay LW. An outline of oral surgery,part II. Bristol: Wright and sons, 1978:92-4.

836

Cell biology and biochemistry of cherubism

7 Wayman JB. Cherubism: a report on three cases. Br J OralSurg 1978;16:47-56.

8 Jones WA. Cherubism. Oral Surg Oral Med Oral Pathol1965;20:648-53.

9 Pindborg JJ, Kramer IRH, Torlini H. Histological typing ofodontogenic tumours, jaw cysts and allied lesions. Geneva:World Health Organisation, 1976.

10 Regezi JA, Sciubba JJ. Oral pathology: clinical-pathologic corre-lations. Philadelphia: WB Saunders, 1989:379-82, 437-9.

11 Jaffe HL. Giant cell reparative granuloma, traumatic bonecyst and fibrous (fibro-osseous) dysplasia of the jaw bones.Oral Surg Oral Med Oral Pathol 1953;6:159-75.

12 Riefkohl R, Georgiade GS, Georgiade NG. Cherubism. AnnPlastic Surg 1985;14:85-9.

13 Chambers TJ, Fuller K, McSheehy PMJ, et al. The effect ofcalcium-regulating hormones on bone resorption byisolated osteoclastoma cells. _f Pathol 1985;145:297-305.

14 Flanagan AM, Tinkler SMB, Horton MA, et al. The multi-nucleate cells in giant cell granulomas of the jaw are osteo-clasts. Cancer 1988;62:1139-45.

15 Harris M. Central giant cell granulomas of the jaws regresswith calcitonin therapy. Brf_ Oral Maxillofac Surg 1 993;31:89-94.

16 Chambers TJ, Revell PA, Fuller K, et al. Resorption of boneby isolated rabbit osteoclasts. if Cell Sci 1984;66:383-99.

17 Horton MA, Lewis D, McNulty K, et al. Monoclonal anti-bodies to osteoclastomas (giant cell bone tumours): defini-tion of osteoclast-specific cellular antigens. Cancer Res1985;45:5663-9.

18 Davies J, Warwick J, Tutty N, et al. The osteoclast functionalantigen, implicated in the regulation of bone resorption, isbiochemically related to the vitronectin receptor. if Cell Biol1989;109:1817-26.

19 Gosden CM, Davidson C, Robertson M. In: Rooney DE,Czepulkowski BH, eds. Human cytogenetics: a practicalapproach. New York: Oxford University Press, 1992:31-54.

20 Prokof DJ, Udenfriend S. A specific method for the analysisof hydroxyproline in tissues and urine. Anal Biochem 1960;1:228.

21 Ibrahim S, Mojiminiyi S, Barron J. High performance liquidchromatographic determination of pyridinium crosslinks inserum, urine and dialysate of patients in chronic renal fail-ure. Ann Clin Biochem 1996;33:31-5.

22 Wharton BA, Gough G, Williams A, et al. Urinary totalhydroxyproline:creatinine ratio. Range of normal, andclinical applications in British children. Arch Dis Child1972;47:74-9.

23 Rauch F, Schonau E, Woitge H, et al. Urinary excretion ofhydroxy-pyridinium cross-links of collagen reflects skeletal

growth velocity in normal children. Exp Clin Endocrinol1994;102:94-7.

24 Matos V, van Melle G, Boulat 0, et al. Urinaryphosphate/creatinine, calcium/creatinine and magnesium/creatinine ratios in a healthy pediatric population. if Pediatr1997;131:252-7.

25 Sweid HA, Bagga A, Vaswani M, et al. Urinary excretions ofminerals, oxalate and uric acid in north Indian children.PediatrNephrol 1997;11:189-92.

26 Chambers TJ, Magnus CJ. Calcitonin alters behaviour ofisolated osteoclasts. if Pathol 1982;136:27-39.

27 Chambers TJ. The cellular basis of bone resorption. ClinOrthop Rel Res 1980;151:283-93.

28 Rodan GA, Martin TJ. The role of osteoblasts in hormonalcontrol of bone resorption. Calcif Tiss Int 1981;33:349-51.

29 Flanagan AM, Chambers TJ. Osteoclasts are present in thegiant cell variant of malignant fibrous histiocytoma. ifPathol 1989;159:53-7.

30 Sarma U, Edwards M, Motoyoshi K, et al. 17P-Estradiolinhibits human osteoclast formation in vitro. if Cell Physiol1998;175:99-108.

31 Overgaard K, Riis BJ, Christiansen C, et al. Effect ofcalcitonin given intranasally on early postmenopausal boneloss. BMJ 1989; 305:477-9.

32 Silva OL, Becker KL. Salmon calcitonin in the treatment ofhypercalcaemia. Arch Intern Med 1973;132:337-9.

33 Wada S, Udagawa N, Nagata N, et al. Calcitonin receptordown-regulation relates to calcitonin resistance in maturemouse osteoclasts. Endocrinology 1996;137: 1042-8.

34 Davis JP, Archer DJ, Fisher C, et al. Multiple recurrent giantcell lesions associated with high circulating levels ofparathyroid hormone-related peptide in a young adult. Bri7Oral Maxillofac Surg 199 1;29:102-5.

35 James IT, Walne AJ, Perrett D. The measurement of pyrid-inium crosslinks: a methodological overview. Ann Clin Bio-chem 1996;33:397-420.

36 Yoshida H, Hayashi S-I, Kunisada T, et al. The murinemutation osteopetrosis is in the coding region of the mac-rophage colony stimulating factor gene. Nature 1990;345:442-4.

37 Grigoriadis AE, Wang Z-Q, Cecchini MG, et al. c-Fos: a keyregulator of osteoclast-macrophage lineage determinationand bone remodeling. Science 1994;266:443-7.

38 Soriano P, Montgomery C, Geske R, et al. Targeted disrup-tion of the c-src proto-oncogene leads to osteopetrosis inmice. Cell 1991;64:693-702.

837