Human tumour-associated macrophages differentiate into osteoclastic bone-resorbing cells

, . 184: 31–36 (1998)

HUMAN TUMOUR-ASSOCIATED MACROPHAGESDIFFERENTIATE INTO OSTEOCLASTIC

BONE-RESORBING CELLS

. . . 1, . ’. 1 . . 1,2*1University of Oxford, Nuffield Department of Pathology and Bacteriology, John Radcliffe Hospital, Oxford OX3 9DU, U.K.

2Department of Pathology, Nuffield Orthopaedic Centre, Oxford OX3 7LD, U.K.

SUMMARY

Macrophages are commonly found within osteolytic secondary carcinomas in bone, but the manner in which these cells contribute tomalignant bone resorption is uncertain. Macrophages isolated from primary breast carcinomas were co-cultured for up to 21 days withUMR 106 rat osteoblast-like cells on bone slices and glass coverslips in the presence and absence of 1,25-dihydroxyvitamin D3[1,25(OH)2D3] and human macrophage colony stimulating factor (M-CSF). Cell cultures were then assessed for the presence ofphenotypic markers of macrophage and osteoclast differentiation. Isolated cells were negative for osteoclast markers includingtartrate-resistant acid phosphatase (TRAP), vitronectin receptor (VNR), and the ability to carry our lacunar bone resorption, but werepositive for CD11b and CD14, macrophage markers which are not present on osteoclasts. In 21-day co-cultures of breast carcinomatumour-associated macrophages (TAMs) and UMR 106 cells, incubated in the presence of 1,25(OH)2D3 and M-CSF, numerous TRAP-and VNR-positive multinucleated cells capable of extensive lacunar resorption were formed. Contact with UMR 106 cells and thepresence of 1,25(OH)2D3 and M-CSF were absolute requirements for differentiation of human breast carcinoma TAMs into maturefunctional osteoclasts. TAM–osteoclast differentiation may represent an important cellular mechanism of osteolysis in metastaticskeletal carcinomas. ? 1998 John Wiley & Sons, Ltd.

J. Pathol. 184: 31–36, 1998.

KEY WORDS—macrophage; carcinoma; osteoclast; metastasis; bone resorption

INTRODUCTION

Malignant tumours, including several commoncarcinomas such as breast, prostate, and lung, have thecapacity to metastasize to bone, where they are associ-ated with considerable bone destruction. This tumourosteolysis causes considerable morbidity, including bonepain, pathological fracture, and hypercalcaemia. Osteo-clasts, which are highly specialized multinucleatedcells responsible for bone resorption, are present atsites of tumour osteolysis, and they are particularlyprominent in the early stages of the establishmentof a tumour metastasis in bone, when osteolysis isproceeding rapidly.1Osteoclasts and macrophages are known to have

several morphological, cytochemical, and functionalcharacteristics in common and to derive from the samecommitted population of haematopoietic precursors.2,3In vitro studies have recently shown that in both mouseand man, osteoclasts may form directly from precursorcell populations of monocytes and macrophages.4–6 Thisfinding may be of significance in terms of malignantbone resorption, as a prominent macrophage inflam-matory infiltrate is commonly present in both primaryand metastatic carcinomas.7–9 A role for tumour-

associated macrophages (TAMs) in causing malignantbone resorption by differentiation of these cells intoosteoclasts is supported by studies which have shownthat TAMs isolated from primary human lung andmouse mammary carcinomas, when co-cultured withbone-derived stromal cells in the presence of 1,25-dihydroxyvitamin D3 [1,25(OH)2D3], are capable ofdifferentiating into multinucleated cells which can effecta small amount of lacunar resorption.10,11 This resorp-tion was due to TAMs, and not carcinoma cells,as shown by studies where, following subcutaneousimplants of human breast, colon, and cervical carcinomacell lines into athymic mice, murine TAMs were shownto be the cells which differentiated into bone-resorbingosteoclasts.12As TAMs are a major component of the host

cellular response to carcinomas, and osteoclastic boneresorption both accompanies and is necessary for theestablishment of a carcinomatous skeletal metastasis,the aim of this study was to determine the essentialcellular and humoral conditions which are requiredfor differentiation of TAMs derived from primaryhuman breast carcinomas into bone-resorbing osteo-clasts. This study has focused in particular on therole of osteoblasts, 1,25(OH)2D3, and macrophage-colony stimulating factor (M-CSF), all of which havebeen shown to be important in osteoclast differen-tiation from precursor cells found in haematopoietictissue and in the circulation, in both mouse andman.6,13,14

*Correspondence to: N. A. Athanasou, Department of Pathology,Nuffield Orthopaedic Centre, Oxford OX3 7LD, U.K.

Contract grant sponsors: Cancer Research Campaign; WellcomeTrust.

CCC 0022–3417/98/010031–06 $17.50 Received 6 November 1996? 1998 John Wiley & Sons, Ltd. Accepted 13 June 1997

MATERIALS AND METHODS

Isolation of tumour-associated macrophages (TAMs)from breast carcinomas and co-culture with UMR 106cells on bone slices and coverslips

TAMS were isolated from seven cases of primarybreast ductal adenocarcinoma, none of whom hadevidence of skeletal metastases. Tumour tissue was cutinto small pieces (1 mm#1 mm#1 mm) and placed insterile containers filled with 10 ml of Hanks balancedsalt solution (HBSS) containing 1 mg/ml collagenasetype I (Sigma, U.K.).15 This mixture was gently agitatedby a magnetic stirrer for 2 h at 37)C. The supernatantwas removed and centrifuged for 10 min and the cellpellet was resuspended in 10 ml of HBSS before furthercentrifugation. The cell pellet was finally resuspendedin 10 ml of alpha minimal essential medium with10 per cent added fetal calf serum (MEM/FCS) (1#106cells per ml). 100 ìl of the cell suspension was thenadded to human cortical bone slices (10 mm2) or 6 mmdiameter coverslips, on half of which 4#104 osteoblast-like UMR 106 cells (supplied by Professor T. J. Martin,Melbourne), a rat osteosarcoma cell line,16 had pre-viously been cultured for 24 h in MEM/FCS. Cultureswere maintained in 1 ml of MEM/FCS for up to 21 days,both in the presence and in the absence of 10"7 1,25(OH)2D3 (Solvay Duphar, The Netherlands), 25 ng/ml human macrophage-colony stimulating factor(M-CSF) (R & D Systems, U.K.), and 10"8 dexa-methasone. In all cultures, the medium (with addedfactors) was entirely replaced every 3 days. As controls,cells were similarly isolated from breast tissue at themargins of the specimens studied; this tissue was un-involved by breast carcinoma and was composed ofnormal breast tissue and fat. Evidence of histochemical/immunophenotypic macrophage/osteoclast differenti-ation (lacunar resorption by cells cultured on glasscoverslips and human cortical bone slices, respectively)was assessed after 24 h and 21 days.To determine the cellular and humoral requirements

for TAM–osteoclast differentiation, cultures on boneslices were incubated both in the presence and in theabsence of UMR 106 cells, 1,25(OH)2D3 or M-CSF.As controls, UMR 106 cells alone were cultured onglass coverslips and bone slices, both in the presenceand in the absence of 1,25(OH)2D3 and M-CSF. Intwo experiments, to determine whether a specificstromal cell was required to support TAM–osteoclastdifferentiation, mouse fibroblast-like L929 cells (SirWilliam Dunn School of Pathology, Oxford, U.K.)were substituted for UMR 106 cells. To determinewhether UMR 106 cells release a soluble factor thatstimulates differentiation of TAMs into bone-resorbingcells, TAMs were also settled onto bone slices devoidof UMR 106 cells. After washing, the bone slices werethen placed into fresh wells containing a glass coverslipupon which UMR 106 cells had previously been cul-tured for 24 h in MEM/FCS; these TAM cultureson bone slices, separated from the cultured UMR 106cells on glass coverslips, were then incubated for21 days in MEM/FCS with added 1,25(OH)2D3 andM-CSF.

Identification of macrophage and osteoclastcytochemical markers in isolated and cultured cells

Following incubation for 24 h and 21 days, cultureson glass coverslips were fixed and stained cytochemicallyfor the osteoclast-associated enzyme tartrate-resistantacid phosphatase (TRAP),17 and immunocytochemi-cally, using an indirect immunoperoxidase technique,18with the monoclonal antibody 23C6 for the presenceof CD51, the osteoclast-specific vitronectin receptor(VNR).19 In addition, these cultures were immunocyto-chemically stained with the monoclonal antibodiesTMG6-5 and GRS1 for the presence of the macrophage-associated antigens CD11b and CD14 respectively, bothof which are known not to be expressed by osteoclasts.20Cultures on glass coverslips were also assessed immuno-histochemically for epithelial antigens after 24 h and21 days in culture using monoclonal antibodies E29and CAM5.2 (Dakopatts), which recognize epithelialmembrane antigen (EMA) and cytokeratin intermediatefilaments respectively.

Functional characterization of isolated and cultured cells

Slices of normal femoral cortical bone obtained fromosteotomies carried out for leg length discrepancy werecut with a diamond blade using a low-speed bone saw(Buehler, Isomet, IL, U.S.A.). The slices were ultrasoni-cated, washed, and stored at room temperature prior touse. Following culture, as described above, the cellscovering the bone slice were removed by rinsing eachslice in phosphate-buffered saline containing 0·02 percent EDTA and then placing them for 5 min in HBSS-containing trypsin (0·25 per cent); they were thenwashed vigorously in distilled water and placed over-night in 0·25 ammonium hydroxide solution. Thebone slices were again washed vigorously in distilledwater, dehydrated in graded alcohols, air-dried, andsputter-coated with gold before being examined in aPhilips SEM 505 scanning electron microscope, to deter-mine the presence or absence of lacunar resorption ofthe bone surface.21 Control bone slices on which no cellswere incubated were also maintained under cultureconditions for 21 days (both in the presence and in theabsence of hormones) and processed similarly; theyshowed only the smooth bone surface and no evidenceof lacunar resorption.

RESULTS

Characterization of TAMs isolated from breastcarcinomasAfter 24 h incubation on glass coverslips, cells isolated

from all seven tumours, when cultured in the presence orabsence of UMR 106 cells, 1,25(OH)2D3 and M-CSF,were found to be entirely negative for the osteoclastmarkers TRAP and VNR.17,19 They strongly expressedCD11b and CD14, macrophage antigens which areknown not to be present on osteoclasts20 (Fig. 1). Manyfewer adherent cells, only a few of which were positivefor CD11b and CD14, were present in 24 h cultures of

32 J. M. W. QUINN ET AL.

? 1998 John Wiley & Sons, Ltd. , . 184: 31–36 (1998)

cells isolated from control normal breast tissue and noTRAP- or VNR-positive cells were seen; 24 h cultures ofUMR 106 cells alone showed no evidence of expressionof CD11b, CD14, TRAP, or VNR. SEM examination of24 h cultures of cells isolated from normal breast tissueand breast carcinoma TAMs on bone slices [again bothin the presence and in the absence of UMR 106 cells,1,25(OH)2D3, and M-CSF] showed no evidence oflacunar bone resorption. There was no staining forEMA or cytokeratin in 24 h cultures. The cells isolatedfrom the breast carcinomas thus expressed only thephenotypic characteristics of macrophages and not ofosteoclasts or carcinoma cells.

TAM–osteoclast differentiation in 21 day co-cultures

After 21 days’ incubation, numerous multinucleatedcells expressing the osteoclast-associated markers TRAPand VNR were found to be present in all TAM–UMR106 co-cultures in the presence of 1,25(OH)2D3 andM-CSF (Fig. 2). Mononuclear and multinucleated cellspositive for CD11b and CD14 were also noted in thesecultures, but no staining for EMA or cytokeratin wasobserved. In the absence of either UMR 106 cells,1,25(OH)2D3, or M-CSF, VNR-positive multinucleatedcells were not seen in 21-day TAM co-cultures. No

CD11b, CD14, VNR, or TRAP staining was seen in21-day cultures of UMR 106 cells alone on glass cover-slips in the presence or absence of 1,25(OH)2D3 andM-CSF. Scattered TRAP-positive mononuclear andmultinucleated cells were, however, noted in all 21-dayTAM cultures (i.e., both in the presence and in theabsence of UMR 106 cells); this reflects the lack ofspecificity of TRAP for the positive identification ofosteoclasts in vitro.22 A few TRAP-positive mono-nuclear cells were also seen in 21-day cultures of cellsisolated from control normal breast tissue, regardless ofwhether these cells were incubated in the presence orabsence of UMR 106 cells, 1,25(OH)2D3, and M-CSF.VNR-positive multinucleated cells, however, were notformed in co-cultures of these cells with UMR 106 cellsin the presence of these factors.All 21-day co-cultures of UMR 106 cells and TAMs

on bone slices, in the presence of 1,25(OH)2D3 andM-CSF, showed functional evidence of osteoclast differ-entiation, with the formation of numerous areas oflacunar resorption on all bone slices (Fig. 3). Extensiveareas of osteolysis (up to 25 per cent of the bone slicearea, or greater than 500 resorption areas, most of whichwere composed of multiple compound areas of lacunar

Fig. 1—Indirect immunoperoxidase staining of 24 h co-culture ofTAMs and UMR 106 cells with anti-CD11b monoclonal antibodyTMG6-5, showing scattered, positively stained TAMs and unstainedUMR 106 cells (#100)

Fig. 2—Twenty-one day co-culture [in the presence of 1,25(OH)2D3 and M-CSF] of TAMs and UMR 106 cells, the latter partly detached,revealing underlying multinucleated cells that are positive for (a) TRAP (#100) and (b) VNR (CD51) (#250)

Fig. 3—SEM photomicrograph of a bone slice on which TAMs hadbeen co-cultured with UMR 106 cells for 21 days in the presence of1,25(OH)2D3 and M-CSF, showing extensive lacunar resorption of thebone surface. (Black bar=100 ìm)

33TAM–OSTEOCLAST DIFFERENTIATION


resorption on some bone slices) were seen in 21-daycultures. The mean surface area resorbed on each boneslice was 86 636 ìm2 (standard error of mean 6247 ìm2).When either UMR 106 cells, M-CSF, or 1,25(OH)2D3was omitted from TAM cultures on bone slices, lacunarresorption was not seen. Twenty-one day co-culturesof cells isolated from control normal breast tissueand UMR 106 cells, incubated in the presence of1,25(OH)2D3 and M-CSF, showed no evidence ofresorption on bone slices. In addition, 21-day culturesof UMR 106 cells alone on bone slices showed noevidence of resorption.The addition of conditioned medium from UMR 106

cells to TAMs cultured with UMR 106 cells on boneslices (both in the presence and in the absence of1,25(OH)2D3 and M-CSF) did not result in lacunarresorption after 21 days’ incubation. In addition, whenTAMs were cultured for 21 days on bone slices in thesame well but not in contact with UMR 106 cells, orwith L929 cells [in the presence of 1,25(OH)2D3 andM-CSF], lacunar resorption did not result.

DISCUSSION

This study describes the essential cellular and humoralrequirements for the differentiation of human TAMsinto mature osteoclasts capable of extensive lacunarbone resorption. Animal studies employing haemato-poietic and mononuclear phagocyte precursors havepreviously shown the importance of bone-derivedstromal cells, 1,25(OH)2D3, and M-CSF in osteoclastdifferentiation.4,7,13,23 1,25(OH)2D3 and human M-CSF,as well as contact with osteoblast-like cells, were alsofound to be necessary for the in vitro differentiation ofTAMs derived from primary breast carcinomas intofunctional osteoclasts.Osteoclasts are known to have a highly specific cyto-

chemical, antigenic, and ultrastructural phenotype anduniquely, in contrast to monocytes, macrophages, andmacrophage polykaryons, are capable of lacunar boneresorption in short-term culture.2,24 Our results showthat mononuclear osteoclast precursors can be found inthe macrophage fraction of extraskeletal carcinomas,and that these cells express the phenotypic characteris-tics of macrophages and not osteoclasts; these osteo-clasts may have formed from precursor cells present inthe blood vasculature, or from cells in the macrophageinfiltrate of the breast carcinomas. Fujikawa et al.6 haverecently shown that the human osteoclast precursorcirculates in the monocyte fraction of peripheral bloodand that it expresses a monocyte/macrophage and not anosteoclast phenotype. Monocytes are known to beattracted to constituents of the bone matrix,25 andcytochemical and ultrastructural studies have shownthat macrophage-like cells appear at resorption sitesbefore mononuclear cells that express osteoclast charac-teristics.26 Osteoclast precursors have also been shownto lose and acquire macrophage and osteoclast markersrespectively, in the process of osteoclast differen-tiation.27 This is mirrored in the TAM–UMR 106co-culture system which we have employed, where

TAMs expressing only macrophage and not osteoclastcharacteristics after 24 h in culture became TRAP- andVNR-positive and acquired the ability to carry outlacunar bone resorption after long-term co-culture withUMR 106 cells.We found that co-cultures of UMR 106 cells

and mononuclear cells similarly isolated from controlnormal breast tissue did not result in the generation ofmultinucleated bone-resorbing cells under similar con-ditions. This may reflect the fact that many fewermacrophages were isolated from normal breast tissue,macrophages being known to form a prominent fractionof the host cellular response to primary breast carcino-mas.7,8 It should not be inferred from this result, how-ever, that tissue macrophage cell populations do notcontain cells capable of osteoclast differentiation, as wehave recently shown that both synovial macrophagesisolated from the synovium of rheumatoid arthritispatients28 and peritoneal macrophages isolated frompatients undergoing peritoneal dialysis29 are capable ofosteoclast differentiation in vitro under similar cellularand humoral conditions of culture. In the mouse, it hasalso been shown that macrophages isolated from a widerange of extraskeletal tissues are similarly capable ofosteoclast differentiation.5 It is known that a phase ofM-CSF-dependent proliferation appears to occur duringosteoclast differentiation,13 and that immature mono-nuclear phagocytes capable of proliferation constitute asmall proportion of the macrophage population in mostorgans.30 Such relatively immature mononuclear phago-cytes may provide the mononuclear precursor popula-tion from which osteoclastic cells formed in ourco-culture system. In carcinomas of the breast, wherethere is generally a heavy macrophage infiltrate, more ofthese immature cells capable of site-specific differentia-tion to the specialized cell of the mononuclear phagocytesystem appropriate to bone (i.e., the osteoclast) wouldappear to be present than in normal breast tissue fromwhich fewer macrophages were isolated.TAMs are derived from circulating blood mono-

cytes,31,32 and various tumour cells have been shownto secrete chemoattractant factors for monocytes.33–35TAMs thus comprise a significant component of thehost cellular response to an infiltrating carcinoma.7–9Previous studies have shown that monocytes,36 macro-phages,37 and their fused products macrophagepolykaryons,38 as well as breast carcinoma cells them-selves,39 are capable of degrading both the mineral andthe organic components of bone, but direct evidenceof lacunar bone resorption by the above mono-nuclear phagocytes or carcinoma cells cultured aloneon bone slices has not been reported.4,40 However, ithas been shown in mice that tissue macrophages,4,5including TAMs,11 but not tumour cells,12 are capableof differentiation into lacunar bone-resorbing cells.Our results show that a number of cellular and

humoral factors are required for the differentiationof TAMs into osteoclasts. Direct contact of TAMswith osteoblastic cells appears to be essential, as TAM–osteoclast differentiation was not supported by therelease of a soluble factor by UMR 106 cells. Theinvolvement of a particular type of stromal cell is also



required, as osteoclast differentiation did not occurwhen TAMs were co-cultured with L929 fibroblast-likecells. This may reflect the need for stromal cells support-ing osteoclast differentiation to be of bone (althoughnot necessarily of osteoblast) origin.7 Osteoclast differ-entiation was only found when 1,25(OH)2D3 andM-CSF were added to TAM–UMR 106 co-cultures.Both of these factors have been shown to be necessaryfor the proliferation and/or maturation of osteoclastprecursors.7,13 Murine mammary tumours have pre-viously been shown to produce factors that have osteo-clast colony-stimulating activity41 and tumour cells areknown to express vitamin D receptors.42 M-CSF is alsoknown to be a secretory product of a number of humantumours including adenocarcinomas of the breast.43 Inaddition, M-CSF and 1,25(OH)2D3 are known productsof macrophages themselves.44 All the above cellular andhumoral elements which we have identified as essentialfor TAM–osteoclast differentiation can thus be found inthe bone micro-environment of a skeletal carcinomatousmetastasis. TAM–osteoclast differentiation could there-fore represent a means whereby lacunar resorptionresults in these osteolytic tumours.The majority of TAMs are found at the host–tumour

interface45 and their numbers within metastatic lesionsare generally higher in earlier small metastases, decreas-ing later as the lesions enlarge.9 Osteoclasts in osteolyticbone secondaries are also known to predominate in theearly phase when a tumour metastasis is becomingestablished in bone and bone resorption is proceedingrapidly; later, when the bone metastasis appears toenlarge more slowly, osteoclasts are correspondinglyless prominent.1 The number of TAMs, and thus thecapacity for osteoclast differentiation by this relativelyabundant fraction of tumour-infiltrating leucocytes,may thus be a determining factor in defining thesetwo more or less distinct phases of osteolytic bonedestruction.

ACKNOWLEDGEMENTS

We thank Mrs M. Pearce for typing the manuscript.This study was supported by the Cancer ResearchCampaign and the Wellcome Trust.

REFERENCES

1. Galasko CBS. Mechanisms of bone destruction in the development ofskeletal metastasis. Nature 1976; 276: 726–728.

2. Suda T, Takahashi N, Martin TJ. Modulation of osteoclast differentiation.Endocrine Rev 1992; 13: 66–80.

3. Kurihara N, Chenu C, Miller M, Civin C, Roodman GD. Identification ofcommitted mononuclear precursors for osteoclast-like cells formed in longterm human marrow cultures. Endocrinology 1990; 126: 2733–2741.

4. Udagawa N, Takahashi N, Akatsu T, et al. Origin of osteoclasts: maturemonocytes and macrophages are capable of differentiating into osteoclastsunder a suitable microenvironment prepared by bone marrow-derivedstromal cells. Proc Natl Acad Sci USA 1990; 87: 7260–7264.

5. Quinn JMW, Sabokbar AS, Athanasou NA. Cells of the mononuclearphagocyte series differentiate into osteoclastic lacunar bone-resorbing cells.J Pathol 1996; 179: 106–111.

6. Fujikawa Y, Quinn JMW, Sabokbar AS, McGee J’OD, AthanasouNA. The human osteoclast precursor circulates in the monocyte fraction.Endocrinology 1996; 137: 4058–4061.

7. Wood GW, Gollahon KA. Detection and quantification of macrophageinfiltration into primary human tumors with the use of cell-surface markers.J Natl Cancer Inst 1977; 59: 1081–1087.

8. Van Ravenswaay Claasen HH, Kluin PM, Fleuren G. Tumor infiltratingcells in human cancer: on the possible role of CD16+ macrophages inantitumor cytotoxicity. Lab Invest 1992; 67: 166.

9. Bugelski PJ, Corwin SP, North SM, Kirsh RL, Nicolson GL, Poste G.Macrophage content of spontaneous metastases at different stages ofgrowth. Cancer Res 1987; 47: 4141–4145.

10. Athanasou NA, Quinn J. Human tumour-associated macrophages arecapable of bone resorption. Br J Cancer 1992; 65: 523–526.

11. Quinn J, Athanasou NA. Tumour infiltrating macrophages are capable ofbone resorption. J Cell Sci 1992; 101: 681–686.

12. Quinn J, Matsumura Y, Tarin D, McGee JO’D, Athanasou NA. Cellularand hormonal mechanisms associated with malignant bone resorption. LabInvest 1994; 71: 465–471.

13. Tanaka S, Takahashi N, Udagawa N, et al. Macrophage colony stimulatingfactor is indispensable for both proliferation and differentiation of osteo-clast progenitors. J Clin Invest 1993; 91: 257–263.

14. Sarma U, Flanagan AM. Macrophage colony-stimulating factor inducessubstantial osteoclast generation and bone resorption in human bonemarrow cultures. Blood 1996; 88: 2531–2540.

15. Haskill S. Collection of macrophages from tumours. In: Herscowitz NB,Holden HT, Bellani JA, Ghaffer A, eds. Manual of MacrophageMethodology: Collection, Characterization and Function. New York:Marcel Dekker, 1981; 43–50.

16. Partridge NC, Alcorn D, Michelangeli VP, Ryan G, Martin TJ. Morpho-logical and biochemical characterisation of four clonal osteogenic sarcomacell lines of rat origin. Cancer Res 1983; 43: 4308–4312.

17. Minkin C. Bone acid phosphatase-tartrate-resistant acid phosphatase as amarker of osteoclast function. Calcif Tissue Int 1982; 34: 285–290.

18. Gatter KC, Falini B, Mason DY. The use of monoclonal antibodies inhistological diagnosis. In: Anthony PP, MacSween RNM, eds. RecentAdvances in Histopathology, No. 12. Edinburgh: Churchill Livingstone,1984; 35–67.

19. Horton MA, Lewis D, McNulty K, Pringle JAS, Chambers TJ. Monoclonalantibodies to osteoclastomas (giant cell bone tumours): definition of osteo-clast specific cellular antigens. Cancer Res 1985; 45: 5663–5669.

20. Athanasou NA, Quinn J. Immunophenotypic differences between osteo-clasts and macrophage polykaryons: immunohistological distinction andimplications for osteoclast ontogeny and function. J Clin Pathol 1990; 43:997–1004.

21. Chambers TJ, Revell PA, Fuller K, Athanasou NA. Resorption of bone byisolated rabbit osteoclasts. J Cell Sci 1984; 66: 383–399.

22. Hattersley G, Chambers TJ. Generation of osteoclastic function inmouse bone marrow cultures: multinuclearity and tartrate-resistant acidphosphatase are unreliable markers for osteoclastic differentiation.Endocrinology 1989; 124: 1689–1695.

23. Takahashi N, Akatsu T, Udagawa N, et al. Osteoblastic cells are involved inosteoclast formation. Endocrinology 1988; 123: 2600–2602.

24. Chambers TJ, Horton MA. Failure of cells of the mononuclear phagocyteseries to resorb bone. Calcif Tissue Int 1984; 36: 556–558.

25. Mundy GR, Varani J, Orr W, Gonde MD, Ward PA. Resorbing bone ischemotactic for monocytes. Nature 1978; 275: 132–135.

26. Baron R, Tran Van P, Nefussi JR, Vignery A. Kinetic and cytochemicalidentification of osteoclast precursors and their differentiation into multi-nucleated osteoclasts. Am J Pathol 1986; 128: 2324–2335.

27. Takahashi N, Udagawa N, Tanaka S, et al. Postmitotic osteoclast pre-cursors are mononuclear cells which express macrophage-associatedphenotypes. Dev Biol 1994; 163: 212–221.

28. Fujikawa Y, Sabokbar A, Neale S, Athanasou NA. Human osteoclastformation and bone resorption by monocytes and synovial macrophages inrheumatoid arthritis. Ann Rheum Dis 1996; 55: 1–7.

29. Quinn JM, Neale S, Fujikawa Y, McGee JO’D, Athanasou NA. Humanmonocytes, peritoneal macrophages and bone marrow cells are capable ofosteoclast differentiation. (Submitted).

30. Van Furth R. Origin and turnover of monocytes and macrophages. CurrTopics Pathol 1989; 79: 125–150.

31. McBride WH. Phenotype and functions of intratumoral macrophages.Biochim Biophys Acta 1986; 865: 27–41.

32. Mantovani A, Bottazzi B, Colotta F, Sozzani S, Ruco L. The originand function of tumor-associated macrophages. Immunol Today 1992; 13:265–276.

33. Bottazzi B, Erba E, Nobili N, Fazioli F, Roubaldi A, Mantovani A. Aparacrine circuit in the regulation of the proliferation of macrophagesinfiltrating murine sarcomas. J Immunol 1990; 144: 2409–2412.

34. Martinet N, Beck G, Bernard V, et al. Mechanism for the recruitment ofmacrophages to cancer sites: in vivo concentration gradient of monocytechemotactic activity. Cancer 1992; 70: 854–860.

35. Yamashiro S, Takeya M, Nishi T, et al. Tumor-derived monocyte chemo-attractant protein-1 induces intratumoral infiltration of monocyte-derivedmacrophage subpopulation in transplanted rat tumors. Am J Pathol 1994;145: 856–867.

36. Mundy GR, Altman AJ, Gondek MD, Bandelin JG. Direct resorption ofbone by human monocytes. Science 1977; 196: 1109–1111.

37. Teitelbaum SL, Stewart CC, Kahn AJ. Rodent peritoneal macrophages asbone resorbing cells. Calcif Tissue Int 1979; 27: 255–261.

35TAM–OSTEOCLAST DIFFERENTIATION


38. Fallon MD, Teitelbaum SL, Stewart CC, Kahn AJ. Multinucleationenhances macrophage-mediated bone resorption. Lab Invest 1983; 49:159–164.

39. Eilon G, Mundy GR. Direct resorption of bone by human breast cancercells in vitro. Nature 1978; 276: 726–728.

40. Boyde A, Maconnachie E, Reid SA. Scanning electron microscopy in bonepathology: review of methods, potential and application. Scanning ElectronMicrosc 1986; 4: 1537–1554.

41. Lee MY, Eyre DR, Osborne WRA. Isolation of a murine osteo-clast colony-stimulating factor. Proc Natl Acad Sci USA 1991; 88:8500–8504.

42. Suda T, Miyaura C, Abe E, Kuroki T. Modulation of cell differentiation,immune responses and tumor promotion by vitamin D components. In:Peck W, ed. Bone and Mineral Research, Vol. 4. Amsterdam: Elsevier, 1986;1–47.

43. Kacinski BM. CSF-1 and its receptor in ovarian, endometrial and breastcancer. Ann Med 1995; 27: 79–85.

44. Nathan CF. Secretory products of macrophages. J Clin Invest 1987; 79:319–326.

45. Svennevig JL, Savaar H. Content and distribution of macrophages andlymphocytes in solid malignant human tumors. Int J Cancer 1979; 24:754–758.



Documents

Human tumour-associated macrophages differentiate into osteoclastic bone-resorbing cells