Ann Rheum Dis 1996;55:816-822

Human osteoclast formation and bone resorptionby monocytes and synovial macrophages inrheumatoid arthritis

Y Fujikawa, A Sabokbar, S Neale, N A Athanasou

AbstractObjective-To determine whether syno-vial macrophages and monocytes isolatedfrom patients with rheumatoid arthritispatients are capable of differentiating intoosteoclastic bone resorbing cells; and thecellular and humoral conditions requiredfor this to occur.Methods-Macrophages isolated from thesynovium and monocytes from theperipheral blood of rheumatoid arthritispatients were cultured on bone slices andcoverslips, in the presence and absence ofUMR 106 rat osteoblast-like cells,1,25-dihydroxy vitamin D3 (1,25(0H)2D3)and macrophage colony stimulating factor(M-CSF), and assessed for cytochemicaland functional evidence of osteoclastdifferentiation.Results-Isolated calcitonin receptor(CTR), tartrate resistant acid phos-phatase (TRAP), and vitronectin receptor(VNR) negative, CDllb and CD14 positivemonocytes and macrophages differenti-ated into CTR, TRAP, and VNR positivemultinucleated cells capable of extensivelacunar bone resorption when co-culturedfor 14 d withUMR 106 cells in the presence1,25(OH)2D3 and M-CSF.Conclusions-Mononuclear phagocytes(monocytes and macrophages) fromrheumatoid arthritis patients are capableof differentiating into multinucleated cellsshowing all the cytochemical andfunctional criteria of mature osteoclasts.Synovial macrophage-osteoclast differen-tiation may represent an important cellu-lar mechanism in the bone destructionassociated with rheumatoid arthritis.(Ann Rheum Dis 1996;55:816-822)

Rheumatoid arthritis is a chronic inflammatorydisease of unknown aetiology that results inprogressive damage to joint tissues. There isextension of exuberant inflamed synovial tissue(pannus) over the articular surface, leading toits destruction.' Rheumatoid pannus alsoextends into juxta-articular bone, producingmarginal erosions and cyst-like areas of bonedestruction. In the inflamed synovium fromwhich pannus develops, there is marked thick-ening of the synovial intima, which iscomposed largely of lining cells of macrophagephenotype2-5; there is also a heavy subintimalchronic inflammatory cells infiltrate whichincludes numerous lymphocytes, plasma cells,

and macrophages.3'- Both intimal and subin-timal synovial macrophages in rheumatoidarthritis are derived from the circulation, bloodmonocytes passing through the tall endothe-lium of postcapillary venules to enter the syno-vial tissues.'0Macrophages rather than T cells have been

found to predominate at the articular marginswhere there is bone and cartilage destruc-tion." 1 A significant correlation has beenreported between the degree of joint erosionand the number of synovial macrophages." Atthe site of marginal erosions, bone resorption,however, appears to be effected largely by rec-ognisable osteoclasts.'"' Various macrophagederived cytokines are known to enhance osteo-clastic bone resorption indirectly (throughosteoblast stimulation)'5 and several of thesecytokines also promote the release ofcollagenase, elastase, plasminogen activator,and prostaglandins by resident cells of boneand joint,'617 factors which may play a role inbone destruction in rheumatoid arthritis.'8

Several recent animal studies have shownthat monocytes and macrophages (includingthose isolated from tumours and inflammatorylesions) are capable of differentiating intoosteoclastic bone resorbing cells whenco-cultured with bone derived stromal cells inthe presence of 1,25-dihydroxyvitamin D3(1,25(OH)2D3)'9-22; the mononuclear andmultinucleated cells formed in these co-cultures show all the phenotypic characteristicsof osteoclasts, including expression of tartrateresistant acid phosphatase (TRAP), calcitoninreceptors (CTR), and most importantly, thefunctional ability to produce resorptionlacunae in bone. In the context of aninflammatory disease such as rheumatoidarthritis, where there is a heavy macrophageinfiltrate in synovial tissues, osteoclast differen-tiation by monocytes and macrophages couldrepresent an important mechanism for thepathogenesis of bone destruction. In this studywe have sought to determine whether this cel-lular mechanism of osteoclast formation andpathological bone resorption operates in rheu-matoid arthritis. Peripheral blood monocytesand synovial macrophages from rheumatoidarthritis patients were co-cultured with osteo-blastic cells in the presence and absence ofhumoral factors which are known to be impor-tant in mouse osteoclast formation, in order todetermine the cellular and humoral conditionsfor mononuclear phagocyte-osteoclast differ-entiation in rheumatoid arthritis.2'25

Nuffield DepartmentofOrthopaedicSurgery, University ofOxford, Oxford,United KingdomY FujikawaA Sabokbar

Department ofOrthopaedics andTrauma, University ofAdelaide, Adelaide,AustraliaS Neale

Department ofPathology, NuffieldOrthopaedic Centre,Oxford, UnitedKingdomN A Athanasou

Correspondence to:DrN A Athanasou,Department of Pathology,Nuffield OrthopaedicCentre, Oxford OX3 7LD.

Accepted for publication12 August 1996

816

group.bmj.com on February 5, 2016 - Published by http://ard.bmj.com/Downloaded from

Fujikawa, Sabokbar, Neale, Athanasou

MethodsSynovial tissue was obtained fresh from sixpatients with seropositive rheumatoid arthri-tis,26 all ofwhom were undergoing knee arthro-plasty operations. The synovial samples wereobtained from five women and one man (agerange 51 to 82 years).Monocytes were isolated from the peripheral

blood of eight patients with seropositiverheumatoid arthritis; seven women and oneman were studied (age range 46 to 74 years).

ISOLATION AND CULTURE OF MONOCYTES ANDSYNOVIAL MACROPHAGESFor isolation of monocytes, blood was diluted1:1 in Hanks 's balanced salt solution (HBSS)(Gibco, UK), layered over Ficoll Hypaque(Pharmacia), centrifuged, washed, and thenresuspended in 5 ml of a minimal essentialmedium (MEM) (Gibco, UK) with 10% fetalcatf serum (Gibco, UK) containing glutamine,benzylpenicillin, and streptomycin (MEM/FCS). The number of cells in the resultantsuspension was counted in a haemocytometer.Peripheral blood mononuclear cells (5 x 10'were then added to 7 mm diameter wells of96-well plates containing human cortical boneslices (5 mm2), prepared as previouslydescribed,27 or 6 mm diameter glass coverslips,on half of which 4 x 104 osteoblast-like UMR106 cells-a rat osteosarcoma cell line28-hadpreviously been cultured for 24 h inMEMIFCS.Specimen radiography was carried out on all

samples of rheumatoid arthritis synovium todetermine whether any bone particles werepresent in the synovial membrane; blocks werealso taken for routine histology. The remainderof the synovial specimen was then cut intosmall pieces (approximately 1 mm') andwashed thoroughly in HBSS. The synovialfragments were digested in HBSS containing 1mg ml-' collagenase type I (Sigma, UK) and0.25% trypsin for 90 min at 37°C. Thesuspension was filtered through a 70 gm cellstrainer (Falcon, UK) before being centrifugedat 680 g for 10 min. The cell pellet was resus-pended in 5 ml MEM/FCS and the cellsuspension (1 x 10' cells/well) was then added

p.

9

sbation,

to bone slices and to glass coverslips, half ofwhich had been preseeded withUMR 106 cellsas described above.The monocyte-UMR 106 and macrophage-

UMR 106 co-cultures were maintained in24-well plates, in 1 ml MEM/FCS for up to 14days, both in the presence and absence of 10-'M 1,25(OH)2D3 (Solvay Duphar, TheNetherlands), 25 ng ml-' human macrophagecolony stimulating factor (M-CSF) (R & DSystems, UK), and 10' M dexamethasone. Inall cultures, the medium (with added factors)was entirely replaced every 3 days.As controls, cultures of monocytes and

synovial macrophages were also set up withand without UMR 106 cells, and in thepresence or absence of 1,25(OH)2D3, M-CSF,and dexamethasone.

CYTOCHEMICAL AND IMMUNOCYTOCHEMICALCHARACTERISATION OF ISOLATED AND CULTUREDCELLSIsolated monocytes and macrophages whichhad been cultured on coverslips for 24 hoursand 7 days were fixed in citrate acetone andstained for tartrate resistant acid phosphatase(TRAP), an osteoclast associated enzyme,29using a kit from Sigma. For immunocytochem-istry, cell cultures on coverslips were fixed incold acetone and stained by an indirect immu-noperoxidase technique30 for the detection ofCD5 1, the vitronectin receptor (VNR), ahighly osteoclast associated antigen, usingmonoclonal antibody 23C6 (a kind gift of Pro-fessorM Horton, London)." Cultures on cov-erslips were also stained immunohistochemi-cally for the presence of CD1lb and CD14,monocyte/macrophage antigens known not tobe expressed by human osteoclasts," with themonoclonal antibodies TMG6-5 andMEM-18 respectively. These monoclonal anti-bodies were derived from the 5th InternationalWorkshop on Human Leucocyte Differentia-tion Antigens."

DETECTION OF CALCITONIN RECEPTORS ONMONOCYTE CULTURESPeripheral blood mononuclear cells isolated andcultured on glass coverslips in the presence andabsence of UMR 106 cells, 1,25(OH)2D3, andM-CSF were assessed after incubation for 24hours and 14 days for the presence of calcitoninreceptors (CTR) by autoradiography using'251-labelled salmon calcitonin ligand binding, aspreviously described.'4 Briefly, peripheral bloodmononuclear cells and UMR 106 cells wereco-cultured on glass coverslips for one hour at22°C in MEM/FCS containing 0.15% bovineserum albumin (BSA) with 0.25 gCi1251-calcitonin (Amersham International, Ayles-bury, UK). After allowing labelled calcitonin tobind for two hours, the cells were washed withcold MEM, fixed for 10 minutes in 2%glutaraldehyde-10% formalin solution, and airdried. Non-specific binding was assessed in thepresence ofan excess amount ofunlabelled calci-tonin (300 nM). The coverslips were then dippedin K-5 photographic emulsion and processed forautoradiography. Negative controls consisted ofUMR 106 cells alone; positive controls consistedof adherent cultured rabbit osteoclasts.

W,

.1

Figure 1 Rheumatoid synovial macrophagelUMR 106 co-culture after 14 d inczshowing multinucleated cels positive for TRAP (x 200).

817

group.bmj.com on February 5, 2016 - Published by http://ard.bmj.com/Downloaded from

Booij, Biewenga-Booij, Huber-Bruning, Cornelis, Jacobs, Bijlsma

Figure 2 Lacunar resorption ofhuman cortical bone slices seen in rheumatoidmonocytelUMR 106 co-cultures after 14 d incubation (black bar = 100pgm).

FUNCTIONAL ASSESSMENT OF THE ABILITY OFCULTURED CELLS TO CARRY OUT LACUNAR BONERESORPTIONRheumatoid arthritis monocytes and macro-

phages incubated on bone slices in thepresence and absence of UMR 106 cells andthe humoral factors described above were cul-tured for periods of 24 hours, 7 days, and 14days. To determine whether lacunar resorptionof bone slices had occurred, confluent cell cul-tures were removed from the surface of thebone in the following manner: each bone slicewas rinsed in phosphate buffered salinecontaining 0.2% EDTA, then placed in 0.25%trypsin solution for 10 min, after which theywere rinsed vigorously in distilled water andimmersed in 0.25 M ammonium hydroxideovernight; they were then washed in distilledwater, dehydrated in graded alcohols, and airdried. The bone slices were mounted onto alu-minium stubs using double sided Sellotape,sputtered with gold, and examined in a PhillipsSEM 505 scanning electron microscope. The

presence or absence of resorption pits on the

Figure 3 Rheumatoid monocytelUMR 106 co-culture after 14 d showing the presence ofcalcitonin receptors on multinucleated cells by autoradiography using 125I labelled calcitonin(x 400).

bone surface was noted,35 and the extent ofresorption determined semiquantitatively.

ResultsHISTOLOGY AND SPECIMEN RADIOGRAPHY OFRHEUMATOID ARTHRITIS SYNOVIAL SPECIMENSThe specimens of rheumatoid synoviumshowed the typical histological features of thiscondition with villous thickening of thesynovial membrane, intimal hyperplasia, and aheavy subintimal chronic inflammatoryinfiltrate, including numerous lymphocytes,plasma cells, and macrophages. On specimenradiography, it was noted that all rheumatoidsynovial specimens contained small radio-opaque fragments, and on histology these werefound to represent small fragments of bonederived from the eroded joint; several of thesebone fragments were surrounded by osteo-clasts.

CHARACTERISATION OF CELLS ISOLATED FROMPERIPHERAL BLOOD AND SYNOVIUM IN 24 hCULTURESAdherent cells isolated from the peripheralblood of rheumatoid arthritis patients werecharacterised as monocytes on the basis that,after being incubated alone on coverslips for 24hours, they were entirely negative for osteoclastmarkers (that is, TRAP, VNR, and CTR)29 3134;they strongly expressed CD1 lb and CD14monocyte/macrophage antigens, which areknown not to be expressed by osteoclasts." Inaddition, after 24 hours incubation on boneslices, no evidence of lacunar resorption wasfound on scanning electron microscopy.3536The above cytochemical, immunocytochemi-cal, calcitonin receptor, and functional charac-teristics were seen when these isolated cellswere cultured both in the presence or absenceofUMR 106 cells, 1,25(OH)2D3, and M-CSF.

Cells isolated from the synovium of rheuma-toid arthritis patients cultured for 24 hours onglass coverslips, both in the presence andabsence ofUMR 106 cells, 1,25(OH)2D3, andM-CSF, were also strongly CDl lb and CD14positive. Similarly, they were also entirelynegative for the osteoclast markers, TRAP andVNR, and they did not produce lacunarresorption on bone slices.

CYTOCHEMICAL AND IMMUNOPHENOTYPICCHARACTERISATION OF CELLS ISOLATED FROMPERIPHERAL BLOOD AND SYNOVIUM IN 7 dCULTURESAfter seven days incubation on glass coverslips,co-cultures of rheumatoid arthritis monocytesand UMR 106 cells, in the presence of1,25(OH)2D3 and M-CSF, contained numer-ous multinucleated cells which were positivefor TRAP and VNR. In the absence of eitherUMR 106 cells, M-CSF, or 1,25(OH)2D3,only a few TRAP positive multinucleated cellsand no VNR positive multinucleated cells wereseen in some cultures. Scattered TRAP andVNR positive mononuclear cells were seen inall co-cultures after seven days incubation.Numerous CD1 lb and CD14 positivemononuclear cells and occasional multinucle-ated cells positive for these antigens were found

818

group.bmj.com on February 5, 2016 - Published by http://ard.bmj.com/Downloaded from

Fujikawa, Sabokbar, Neale, Athanasou

Figure 4 (A) Lacunar resorption pits on human cortical bone slice seen in rheumatoidsynovial macrophagelUMR 106 co-culture after 14 d incubation (black bar = 100 pm).(B) Higher power ofone of the complex resorption pits (black bar = 100 pm).

in seven day cultures both in the presence andabsence ofUMR 106 cells.Rheumatoid synovial macrophages cultured

for seven days with UMR 106 cells, in thepresence of 1,25(OH)2D3 and M-CSF, alsocontained numerous TRAP positive (fig 1) andVNR positive mononuclear and multinucle-ated cells. When incubated in the absence ofUMR 106 cells, 1,25(OH)2D3 or M-CSF, allthese cultures were found to contain mainlyCD1 lb and CD14 positive mononuclear cells;however, cultures from two cases, incubated inthe absence of the above factors, containedoccasional TRAP and VNR positive multinu-cleated cells. However, scattered TRAP andVNR positive mononuclear cells were seen inall co-cultures after seven days incubation.

LACUNAR BONE RESORPTION AND CALCITONINRECEPTOR EXPRESSION IN 14 d CO-CU1LTURES OFRHEUMATOID MONOCYTES AND MACROPHAGESWITH OSTEOBIAST-LIKE CELLSAfter 14 days in culture, numerous resorptionpits were seen on bone slices upon which rheu-matoid monocytes had been co-cultured with

UMR 106 cells in the presence of1,25(OH)2D3, M-CSF, and dexamethasone.Lacunar excavation was considerable, amount-ing to over 500 resorption pits (approximately20% of the bone slice area), in some cases (fig2). In all cases, over 50 resorption pits wereseen on each bone slice when rheumatoidmonocytes were cultured under theseconditions. This extensive lacunar resorptionwas not seen when rheumatoid monocyteswere cultured for 14 days in the absence ofUMR 106 cells, or when 1,25(OH)2D3 orM-CSF was omitted from monocyte-UMR106 co-cultures. Although dexamethasone wasnot an absolute requirement for lacunar boneresorption, its addition to rheumatoidmonocyte-UMR 106 co-cultures (with added1,25(OH)2D3 and M-CSF) considerablyincreased the extent of the lacunar resorptionseen.The presence of calcitonin receptors was

also determined by autoradiography in 14 daymonocyte cultures, where it was found thatCTR positive multinucleated cells werepresent only in co-cultures of rheumatoidmonocytes andUMR 106 cells which had beenincubated in the presence of 1,25(OH)2D3 andM-CSF (fig 3). When monocytes wereincubated alone on glass coverslips, or withUMR 106 cells alone, or with UMR 106 in theabsence of 1,25(OH)2D3 or M-CSF, CTRpositive cells were not seen.

Fourteen day co-cultures of rheumatoidarthritis synovial macrophages with UMR 106cells, when incubated with 1,25(OH)2D3,M-CSF, and dexamethasone, also showed evi-dence of extensive lacunar resorption on allbone slices, over 200 resorption pits beingnoted in some cases (fig 4). In the absence ofUMR 106 cells, 14 day rheumatoid arthritismacrophage cultures on bone slices, even whenincubated in the presence of 1,25(OH)2D3 andM-CSF, showed no evidence of lacunarresorption. Lacunar resorption was also notseen in rheumatoid macrophage-UMR 106co-cultures when 1,25(OH)2D3 or M-CSF wasomitted from the culture medium.

DiscussionThis is the first report of human osteoclast for-mation and bone resorption by mononuclearphagocytes isolated from patients withrheumatoid arthritis. It indicates that mononu-clear phagocytes, either present in the inflamedsynovial membrane or recruited to thesynovium from the circulation, may directlycontribute to the osteolysis of rheumatoidarthritis by themselves differentiating intomultinucleated cells which show all thecytochemical and functional features of osteo-clasts. This study has also characterised thecellular and humoral conditions required forthis to occur, showing that osteoblastic cells,1,25(OH)2D3, and M-CSF are essential forosteoclast formation from rheumatoid mono-cytes and macrophages.

Cells isolated from the peripheral blood andsynovial membrane were characterised asmonocytes and macrophages on the basis thatthey did not express the osteoclast markers of

819

group.bmj.com on February 5, 2016 - Published by http://ard.bmj.com/Downloaded from

Booij, Biewenga-Booij, Huber-Bruning, Cornelis, Jacobs, Bijisma

TRAP 29 VNR31, and CTR,3 and did not pos-sess the ability to form resorption lacunae after24 hours in culture, a functional feature whichis characteristic of osteoclasts'536 ; these cells,however, did express CD1 lb and CD14mononuclear phagocyte antigens which areknown not to be expressed by osteoclasts."Although cells expressing these antigenspersisted in all the long term monocyte/macrophage-UMR 106 co-cultures examined,it was noted that multinucleated cells whichexpressed the above osteoclast phenotypiccharacteristics, including that of lacunarresorption, formed only when rheumatoidmonocytes or macrophages were co-culturedwith UMR 106 cells in the presence of1,25(OH)2D3 and M-CSF.Previous studies have shown that mouse

monocytes and macrophages are capable ofosteoclast differentiation when co-culturedwith UMR 106 cells or other bone derivedstromal cell types in the presence of1,25(OH)2D alone.192021-2237 The presentstudy shows that the above factors are alsoessential for the in vitro formation of humanosteoclasts from rheumatoid arthritis mono-cytes and macrophages, and that M-CSF is anadditional essential cofactor. This requirementfor human M-CSF is not surprising as it haspreviously been shown that rodent M-CSF(which would be secreted by UMR 106 cells)does not bind to the human M-CSF receptor(which would be present on the mononuclearphagocytes isolated from the circulation orsynovium of rheumatoid arthritis patients).38M-CSF is known to be necessary for the prolif-eration and differentiation of haematopoieticosteoclast progenitors,2539 and an abnormalityin the gene for M-CSF has been found to bethe cause of the deficiency in osteoclast andbone macrophage numbers in osteopetroticop/op mice.42 Synovial macrophages arethemselves known to produce 1,25(OH)2D3,"'and M-CSF is produced by both bone cellsand cells found within the inflamedsynovium.24 4 117 The synovial membrane isknown to be a source of other cytokines andgrowth factors which influence osteoclastdifferentiation.4"5' One of these cytokines isinterleukin (IL)-6, receptors for which havebeen shown to be induced on osteoblastic cellsby dexamethasone21; this may explain themarked increase in bone resorption weobserved when dexamethasone was added tomonocyte/macrophage-UMR 106 co-cultures.IL-6 and soluble IL-6 receptors in the synovialfluid of rheumatoid patients have also recentlybeen shown to be associated with theformation of osteoclast-like cells.5'

Analysis of the immunophenotype ofsynovial lining cells and subintimal macro-phages in the synovial membrane has shownthat these cells strongly express monocyte/macrophage markers such as HILA-DR, Fc,and C3 receptors, as well as a wide range ofleucocyte/macrophage antigens, includingCD1 lb and CD14.' In rheumatoid arthritis,when the synovial intima becomes hyperplas-tic, over 90% of the synovial lining cells arecommonly of macrophage phenotype, and

there is an increase in macrophages in thesubintima.9 Cellular kinetic and immunophe-notypic studies have shown that synovial mac-rophages are derived from circulatingmonocytes which pass through tall endothelialcells of postcapillary venules before migratingto their tissue location in the intima orsubintima.4510 These synovial macrophagesdifferentially express certain activation markerswhich are associated with their ability to carryout a number of complex functions.52 In rheu-matoid arthritis, circulating monocytes them-selves have been reported to show enhancedmetabolic and phagocytic activity and to expresssurface markers consistent with macrophageactivation.5' Monocytes are known to bechemotactically attracted to constituents of thebone matrix,55 and cytochemical and ultrastruc-tural studies have shown that macrophage-likemononuclear cells appear at resorption surfacesbefore mononuclear cells which expressosteoclast characteristics.56

Synovial macrophages are believed to play acrucial role in joint destruction in rheumatoidarthritis. They are known to release numerouscytokines, principally IL-1 and tumournecrosis factor aj,'5 16 which act on osteoblaststo stimulate osteoclastic bone resorptionindirectly; in addition, they are known torelease or to promote the release of prostaglan-dins and tissue proteases which may play a rolein bone lysis."'6 '7 However, mature tissuemacrophages, even in the presence of thesefactors, are incapable of causing lacunarresorption of a mineralised substrate.'6 This isalso the case for rheumatoid synovialmacrophages, as confirmed in this study.57However, our findings do indicate that synovialmacrophages, like monocytes, are capable ofaltering their phenotype to that of osteoclastswhen cultured under specific cellular andhumoral conditions. In the same manner asosteoclast precursors derived from haemat-opoietic tissues, these mononuclear phagocytesacquire osteoclastic features and lose theirmacrophage phenotypic features respectivelyin the process of osteoclast differentiation.Monocyte-osteoclast and macrophage-osteoclast differentiation does not appear to beparticular to rheumatoid arthritis, as we haveshown that in both man and experimental ani-mals peripheral blood mononuclear cells andtissue macrophages are capable of osteoclastdifferentiation when cultured with osteoblast-like cells under similar culture conditions.20'759Monocytes and macrophages do not

represent homogeneous cell populations butinclude a fraction of morphologicallyindistinguishable immature cells capable offurther division.' In the context of boneresorption in rheumatoid arthritis, it is possiblethat these immature mononuclear phagocytes,either as monocytes recruited into thesynovium or as synovial macrophages, arecapable-when placed in the cellular andhumoral microenvironment of bone (that is,when they come in contact with bone lined byosteoblasts in the presence of 1,25(OH)2D3and M-CSF)-of proliferation and differentia-tion into bone resorbing osteoclasts, the tissue

820

group.bmj.com on February 5, 2016 - Published by http://ard.bmj.com/Downloaded from

Fujikawa, Sabokbar, Neale, Athanasou

specific mononuclear phagocyte of bone. Theessential cellular and humoral conditions forhuman osteoclast formation from monocytesand macrophages are present in the in vitroco-culture system described in this study. Itshould provide a useful model for the analysisof the role that synovial macrophage-osteoclastdifferentiation plays in the bone destructionassociated with rheumatoid arthritis and otherarthritic diseases.

We would like to thank Mrs Margaret Pearce for typing themanuscnpt. This work was supported by the Nuffield Founda-tion (Oliver Bird Fund) and the Wellcome Trust.

1 Gardner DC. The pathology of rheumatoid arthritis. London:Edward Arnold, 1992.

2 Burmester GR, Loacher P, Winchester RJ, Dimitriu-Bona,Kalden JR, Mohr W. The tissue architecture of synovialmembranes in inflammatory and non inflammatory jointdiseases. Rheumatol Int 1983;3:173-81.

3 Hogg N, Palmer DG, Revell PA. Mononuclear phagocytesof normal and rheumatoid synovial membrane identifiedby monoclonal antibodies. Immunology 1985;56:673-81.

4 Athanasou NA, Quinn J, Heryet A, Puddle B, Woods CG,McGee JOD. The immunohistology of synovial lining cellsin normal and inflamed synovium.J7Pathol 1988;155:133-42.

5 Henderson B, Revell PA, Edwards JC. Synovial lining cellhyperplasia in rheumatoid arthritis: dogma and fact. AnnRheum Dis 1988;47:348-9.

6 Allen CA, Highton J, Palmer GD. Increased expression ofP150, 95 and CR3 leukocyte adhesion molecules bymononuclear phagocytes in rheumatoid synovial mem-brane. Comparison with osteoarthritic and normal syno-vial membranes. Arthritis Rheum 1989;32:947-54.

7 Hale LP, Martin ME, McCollum DE, Nunley JA, SpringerTA, Singer KH, et al. Immunohistologic analysis of thedistribution of cell adhesion molecules within the inflam-matory synovial microenvironment. Arthritis Rheum 1989;32:22-30.

8 Broker BM, Edwards JCW, Fanger MW, Lydyard PM, Theprevalence and the distribution of macrophage bearingFcRI, FcRII and FcRI in synovium. Scand J Rheumatol1990;19: 123-35.

9 Athanasou NA, Immunophenotypic analysis of mononu-clear phagocyte antigens in the arthritic synovium of thehip joint. Hip Int 1993;3:67-74.

10 Ziff M. Pathways of mononuclear cell infiltration inrheumatoid synovitis. Rheumatol Int 1989;9:97-103.

11 Bromley M, Woolley DE. Histopathology of the rheumatoidlesion: identification of cell types at sites of cartilageerosion. Arthritis Rheum 1984;27:857- 63.

12 Aclard SA, Bayliss MT, Maini RN. They synovium-cartilagejunction of the normal human knee: implications for jointdestruction and repair. Arthritis Rheum 1990;33:1170-9.

13 Yanni G, Whelan A, Feighery C, Bresnihan B. Synovial tis-sue macrophages and joint erosion in rheumatoid arthritis.Ann Rheum Dis 1994;53:39-44.

14 Bromley M, Woolley D. Chondroclasts and osteoclasts atsubchondral sites of erosion in the rheumatoid joint.Arthritis Rheum 1984;27:857-63.

15 Skerry T, Gowen M. Bone cells and bone remodelling inrheumatoid arthritis. In: Henderson B, Edwards JEW, Pet-tipher ER, eds. Mechanisms and models in rheumatoid arthri-tis. London: Academic Press, 1995:205-20.

16 Brennan FM, Field M, Chu CQ, Feldmann M, Maini RM.Cytokine expression in rheumatoid arthritis. BrJ7 Rheuma-tol 1991;30(suppl 1):76-80.

17 Zvaifler NJ, Firestein GS. Pannus and pannocytes: alterna-tive models of joint destruction in rheumatoid arthritis.Arthritis Rheum 1994;37:783-9.

18 Dingle JT. The role of cellular interactions in joint erosions.Clin Orthop 1984;182:24-9.

19 Udagawa N, Takahashi N, Akatsu T, Tanaka H, Sasaki T,Nishihara T, et al. Origin of osteoclasts: mature monocytesand macrophages are capable of differentiating into osteo-clasts under a suitable microenvironment prepared bybone marrow-derived stromal cells. Proc Natl Acad SciUSA 1990;87:7260-4.

20 Quinn J, McGee JO D, Athanasou NA. Cellular andhormonal factors influencing monocyte differentiation toosteoclastic bone resorbing cells. Endocrinology 1994;134:2416-23.

21 Quinn JM, Joyner C, Triffitt JT, Athanasou NA.Polymethylmethacrylate-induced_ inflammatory macro-phages resorb bone. JBoneJoint SurgBr 1992;74B:652- 8.

22 Quinn J, Athanasou NA. Tumour infiltrating macrophagesare capable of bone resorption. J Cell Sci 1992;101:681-6.

23 Takahashi N, Yamana H, Yoshiti S, Roodman GD, MundyGR, Jones SJ, et al. Osteoclast-like cell formation and itsregulation by osteotropic hormones in mouse bonemarrow cultures. Endocrinology 1988;122:1373-82.

24 Suda T, Udagawa N, Nakamura I, Miyaura C, Takahashi N.Modulation of osteoclast formation by local factors. Bone1995;17(suppl 2):875-915.

25 Tanaka S, Takahashi N, Udagawa N, Tamura T, Akatsu T,Stanley ER, et al. Macrophage colony stimulating factor isindispensable for both proliferation and differentiation ofosteoclast progenitors. J Clin Invest 1993;91:257-63.

26 Arnett C, Edworthy SM, Bloch DA, McShane DJ, Fries JF,Cooper NS, et al. The American Rheumatism Association1987 revised criteria for the classification of rheumatoidarthritis. Arthritis Rheum 1988;31:315-24.

27 Athanasou NA, Wells CA, Quinn J, Ferguson DP, Heryet A,McGee JOD. The origin and nature of stromal osteoclast-like multinucleated giant cells in breast carcinoma: impli-cations for tumour osteolysis and macrophage biology. Br_J Cancer 1989;59:491-498.

28 Partridge N, Alcorn D, Michelangeli V, Kemp E, Ryan G,Martin TIJ. Functional properties of hormonal responsivecultured normal and malignant rat osteoblastic cells.Endocrinology 1981;108:213-9.

29 Andersson G, Ek-Rylander B, Minkin C. Acid phos-phatases. In: Rifkin BR, Gay CV, eds. Biology andphysiology of the osteoclast. Boca Raton: CRC Press,1992:55- 80.

30 Gatter KC, Falini B, Mason DY. 1984 The use ofmonoclonal antibodies in histological diagnosis. In:Anthony PP, MacSween RNM, eds. Recent advances in his-topathology, No 12. Edinburgh: Churchill Livingstone,1984:35-67.

31 Horton MA, Lewis D, McNulty K, Pringle JAS, ChambersTJ. Monoclonal antibodies to osteoclastomas (giant cellbone tumours): definition of osteoclast specific cellularantigens. Cancer Res 1985;45:5663-9.

32 Athanasou NA, Quinn J. Immunophenotypic differencesbetween osteoclasts and macrophage polykaryons: immu-nohistological distinction and implications for osteoclastontogeny and function. J Clin Pathol 1990;43:997-1004.

33 Schlossman SF, Boumsell L, Gilks W, Harlan JM,Kishimoto T, Morimoto C, et al. Leucocyte typing V: whitecell differentiation antigens. Oxford: Oxford UniversityPress, 1995.

34 Nicholson CG, Moseley JM, Sexton PM, Mendelsohn FAO.Abundant calcitonin receptors in isolated rat osteoclasts. J7Clin Invest 1986;78:355-60.

35 Chambers TJ, Revell PA, Fuller K, Athanasou NA. Resorp-tion of bone by isolated rabbit osteoclasts. J7 Cell Sci 1984;66:383-99.

36 Chambers TJ, Horton MA. Failure of cells of the mononu-clear phagocyte series to resorb bone. Calcif Tissue Int1984;36:556-68.

37 Quinn JMW, Sabokbar A, Athanasou NA. Cells of the mono-nuclear phagocyte series differentiate into osteoclasticlacunar bone-resorbing cells.JPathol 1996;179:106-1 1.

38 Roussel MF, Downing JR, Rettenmeier CW, Sherr CJ. Apoint mutation in the extra cellular domain of the humanCSF-1 receptor (c-fms product) activates its transformingpotential. Cell 1988;55:979-88.

39 Kodama H, Nose M, Niida S, Yamasah A. Essential role ofmacrophage-colony stimulating factor in the osteoclastdifferentiation supported by stromal cells. _J Exp Med1991;173:1292-4.

40 Yoshida H, Hayashi S-I, Kunisada T, Ogawa M, NishikawaS, Okamura H, et al. The murine mutation osteopetrosis isin the coding region of the macrophage colony stimulatingfactor gene. Nature 1990;345:442-4.

41 Felix R, Cecchini MG, Fleisch H. Macrophage colonystimulating factor restores in vivo bone resorption in theop/op osteopetrotic mouse. Endocrinology 1990;127:2592-4.

42 Felix R, Cecchini MG, Hofstetter W, Elford PR, Stutzer A,Fleisch H. Impairment of macrophage colony stimulatingfactor production and lack of resident bone marrow mac-rophages in the osteopetrotic op/op mouse. J7 Bone MinerRes 1990;5:781-9.

43 Hayes ME, Denton J, Freemont AJ, Mawer EB. Synthesis ofthe active metabolite of vitamin D, 1,25(OH)2D3, bysynovial fluid macrophages in arthritic disease. Ann RheumDis 1989;48:723-9.

44 Shern CJ. Colony-stimulating factor-i receptor. Blood1990;75:1-12.

45 Weir EC, Horowitz MC, Baron R, Centrolla M, KacinskiBM, Insogna KL. Macrophage-colony stimulating factorrelease and receptor expression in bone cells. J7 Bone MinerRes 1993;8:1507-18.

46 Smith JB, Borchieri MH, Smith JB, Sherbin-Allen L,Abruzzo JL. Colony stimulating factor occurs in bothinflammatory and non inflammatory synovial fluids. Rheu-matolInt 1990;10:131-4.

47 Seitz M, Loestscher P, Fey MF, Tobler A. ConstitutivemRNA and protein production of macrophage-colonystimulating factor but not of other cytokines by synovialfibroblasts from rheumatoid arthritis and osteoarthritispatients. Br3 Rheumatol 1994;33:613-9.

48 Chin J, Rupp E, Cameron PM, MacNaul KL, Lotke PA,Tocci MJ, et al. Identification of a high-affinity receptor forinterleukin 1 and interleukin 1 on cultured humanrheumatoid synovial cells. _J Clin Invest 1 988;82:420-6.

49 Xu WD, Firestein GS, Taetle R, Kaushansky K, Zvaifler NJ.Cytokines in chronic inflammatory arthritis II: granulocyte-macrophage colony stimulating factor in rheumatoid synovialeffusions. JClin Invest 1989;83:876-82.

50 Guerne P-A, Zuraw BL, Vaughan JH, Carson DA, Lotz M.Synovium as a source of interleukin-6 in vitro. J Clin Invest1989;83:585-92.

51 Kotake S, Sato K, Kim KJ, Takahashi N, Udagawa N,Nakamura I, et al. Interleukin-6 and soluble interleulin-6receptors in the synovial fluids from rheumatoid arthritis

821

group.bmj.com on February 5, 2016 - Published by http://ard.bmj.com/Downloaded from

Booij, Biewenga-Booij, Huber-Bruning, Cornelis, Jtacobs, Bijlsma

patients are responsible for osteoclast-like cell formation.Bone Miner Res 1996:11:88-95.

52 Adams DO, Hamilton TA. The cell biology of macrophageactivation. Annu Rev Immunol 1984;2:283-318.

53 Fujii I, Shinju M, Noburaga M. Monocyte activation inearly onset rheumatoid arthritis. Ann Rheum Dis 1990;49:497-503.

54 McCarthy D, Taylor MJ, Bernhagen J, Perry JD, HamblinAS. Leucocyte integrin and CRI expression on peripheralblood leucocytes of patients with rheumatoid arthritis.Ann Rheum Dis 1992;51:307-12.

55 Mundy GR, Varani J, Orr W, Gondek MD, Ward PA.Resorbing bone is chemotactic for monocytes. Natuire1978;275:132-5.

56 Baron R, Tran Van P, Nefussi JR, Vignery A. Kinetic andcytochemical identification of osteoclast precursors and

their differentiation into multinucleated osteoclasts. AmPathol 1986:122:363-78.

57 Chang JS, Quinn JMW, Demaziere A, Buistrode CJ, FrancisMJ, Duthie RB, et al. Bone resorption by cells isolatedfrom the rheumatoid synovium. Ann Rheum Dis 1992;51:1223-9.

58 Takahashi N, Udagawa N, Tanaka S, Murakanni H, Owan I,

Tamura T, et al. Postmitotic osteoclast precursors are

mononuclear cells which express macrophage-associatedphenotypes. Dev Biol 1994;163:212-21.

59 Fujikawa Y, Quinn JMW, Sabokbar AS, McGee JO D, Atha-nasou NA. The human mononuclear osteoclast precursor

circulates in the monocyte fraction. Endocrinology (inpress).

60 van Furth R. Origin and turnover of monocytes and macro-

phages. Curr Top Pathol 1989;79:125-50.

822

group.bmj.com on February 5, 2016 - Published by http://ard.bmj.com/Downloaded from

macrophages in rheumatoid arthritis.resorption by monocytes and synovial Human osteoclast formation and bone

Y Fujikawa, A Sabokbar, S Neale and N A Athanasou

doi: 10.1136/ard.55.11.8161996 55: 816-822 Ann Rheum Dis 

http://ard.bmj.com/content/55/11/816Updated information and services can be found at:

These include:

serviceEmail alerting

box at the top right corner of the online article. Receive free email alerts when new articles cite this article. Sign up in the

Notes

http://group.bmj.com/group/rights-licensing/permissionsTo request permissions go to:

http://journals.bmj.com/cgi/reprintformTo order reprints go to:

http://group.bmj.com/subscribe/To subscribe to BMJ go to:

group.bmj.com on February 5, 2016 - Published by http://ard.bmj.com/Downloaded from