Induction of osteoclasts from CD14-positive human peripheral blood ...

Clinical Science (2000) 99, 133–140 (Printed in Great Britain)

Induction of osteoclasts from CD14-positive human peripheral blood mononuclear cells by receptor activator of nuclear factor κB ligand (RANKL) Geoffrey C. NICHOLSON*, Mary MALAKELLIS*, Fiona M. COLLIER*, Paul U. CAMERON†, Wayne R. HOLLOWAY*, Tamara J. GOUGH*, Claudia GREGORIO-KING‡, Mark A. KIRKLAND‡ and Damian E. MYERS* *Department of Medicine, University of Melbourne, Parkville, Victoria 3010, Australia, †Department of Microbiology and Immunology, University of Melbourne, Parkville, Victoria 3010, Australia, and ‡The Douglas Hocking Medical Institute, Barwon Health, The Geelong Hospital, Box 281, Geelong, Victoria 3220, Australia

A

B

S

T

R

A

C

T

Osteoclasts are bone-resorbing cells that are derived from haemopoietic precursors, including cells present in peripheral blood. The recent identification of RANKL [receptor activator of nuclear factor (NF)-κB ligand], a new member of the tumour necrosis factor ligand superfamily that has a key role in osteoclastogenesis, has allowed the in vitro generation of osteoclasts in the absence of cells of the stromal/osteoblast lineage. Human peripheral blood mononuclear cells (PBMC) cultured in vitro with soluble RANKL and human macrophage colony-stimulating factor form osteoclasts. However, PBMC are heterogeneous, consisting of subsets of monocytes and lymphocytes as well as other blood cells. As the CD14 marker is strongly expressed on monocytes, the putative osteoclast precursor in peripheral blood, we have selected CD14+ cells from PBMC to examine their osteoclastogenic potential and their expression of novel members of the tumour necrosis factor superfamily involved in osteoclastogenesis. Highly purified CD14+ cells demonstrated mRNA expression of receptor activator of NF-κB, but no expression of RANKL or osteoprotegerin, whereas PBMC expressed mRNAs for all three factors. CD14+ (but not CD14−) cells cultured on bone slices for 21 days with human macrophage colony-stimulating factor and soluble RANKL generated osteoclasts and showed extensive bone resorption. Similar numbers of osteoclasts were generated by 105 CD14+ cells and 106 PBMC, but there was significantly less intra-assay variability with CD14+ cells, suggesting the absence of stimulatory/ inhibitory factors from these cultures. The ability of highly purified CD14+ cells to generate osteoclasts will facilitate further characterization of the phenotype of circulating osteoclast precursors and cell interactions in osteoclastogenesis.

Key words : CD14-positive, osteoclastogenesis, peripheral blood mononuclear cells, receptor activator of NF-κB-ligand, receptor activator of NF-κB. Abbreviations : (h)M-CSF, (human) macrophage colony-stimulating factor ; MNC, multinuclear cells ; Mo, mononuclear cells ; NSE, non-specific esterase ; NF-κB, nuclear factor κB ; OPG, osteoprotegerin ; PBMC, peripheral blood mononuclear cells ; RANK, receptor activator of NF-κB ; (s)RANKL, (soluble) receptor activator of NF-κB ligand ; RT, reverse transcription ; TNF, tumour necrosis factor ; TRAP, tartrate-resistant acid phosphatase ; the superscripts + and − denote that cells are positive or negative respectively for the given molecule (e.g. TRAP+, CD14−). Correspondence : Professor G. C. Nicholson (e-mail g.nicholson!medicine.unimelb.edu.au).

# 2000 The Biochemical Society and the Medical Research Society

133

134

G. C. Nicholson and others

INTRODUCTION The bone-resorbing cell, the osteoclast, is derived from macrophage colony-stimulating factor (M-CSF)-dependent haemopoietic precursors, including mononuclear cells present in the peripheral circulation [1–3]. In vitro generation of osteoclasts in co-cultures of adherent peripheral blood mononuclear cells (PBMC) and stromal\osteoblastic cells has been demonstrated in both murine [4,5] and human [6–9] systems. Recently, the requirement for co-culture with stromal\osteoblastic cells has been circumvented by the identification of a factor expressed on the membrane of stromal\osteoblast cells that mediates osteoclast differentiation (reviewed in [10]). This factor, a novel type II membrane protein of the tumour necrosis factor (TNF) ligand superfamily, was initially identified in dendritic cells as TNF-related activation-induced cytokine (‘ TRANCE ’) [11]. Subsequently, identical proteins were independently identified as receptor activator of nuclear factor (NF)-κB ligand (RANKL) [12], osteoclast differentiation factor [13] and osteoprotegerin (OPG) ligand [14]. As suggested [10], this factor will be referred to as RANKL in this paper. RANKL acts on osteoclast precursors via a membrane receptor, RANK (receptor activator of NF-κB) [15,16], but can also bind to a second receptor, OPG, a secreted glycoprotein [13,14,17]. OPG can thus function as a ‘ decoy ’ receptor for RANKL, in this way inhibiting osteoclastogenesis, a function which led to its identification as ‘ osteoclastogenesis inhibitory factor ’ [18]. Matsuzaki et al. [19] incubated washed adherent human PBMC with human M-CSF (hM-CSF) and soluble mouse RANKL (sRANKL ; comprising the extracellular domain of mouse RANKL) and generated large numbers of tartrate-resistant acid phosphatase-positive (TRAP+), vitronectin-receptor-positive multinucleate cells that resorbed dentine slices extensively. However, when the non-adherent cells (which are mostly lymphocytes) were not removed, dexamethasone was required for osteoclastogenesis to occur, suggesting the presence of a dexamethasone-suppressible inhibitory factor(s) in this population. Quinn et al. [20] generated bone-resorbing osteoclasts from mouse and human PBMC and mouse spleen cells treated with sRANKL and hM-CSF. In that study, 95 % of the human cells adherent to glass coverslips expressed the macrophage-associated antigens CD68, CD11b and CD14, suggesting that osteoclasts are derived from CD14-positive (CD14+) monocytes. However, phenotypic expression was not examined on cells adherent to bone slices. Nevertheless, the recent results of Massey and Flanagan [21], who found that selection of CD14+ cells from PBMC enhanced osteoclastic bone resorption in co-cultures with UMR 106 rat osteoblastlike cells, provide further evidence that the osteoclast precursor is CD14+. # 2000 The Biochemical Society and the Medical Research Society

The above results indicate that there is an osteoclast precursor within the PBMC population and, although PBMC are a heterogeneous population (monocytes, lymphocytes and smaller numbers of other blood cell types), the most likely candidate for the osteoclast precursor in this population is the CD14+ monocyte. The aim of the present study was to determine whether ‘ monocultures ’ of highly purified CD14+ cells selected from PBMC could generate osteoclasts when cultured on bone with sRANKL and hM-CSF, and also to examine the expression of RANK, RANKL and OPG mRNAs in CD14+ cells.

MATERIALS AND METHODS Materials Tissue culture media, Eagle’s minimum essential medium and Medium 199 were purchased from Sigma Chemical Co. (St. Louis, MO, U.S.A.). Fetal bovine serum was purchased from Commonwealth Serum Laboratories (Parkville, Victoria, Australia). Soluble human RANKL was purchased from PeproTech (Rocky Hill, NJ, U.S.A.). Recombinant hM-CSF was generously donated by Genetics Institute (Cambridge, MA, U.S.A.). Monoclonal antibodies to CD markers and MACS (magnetic cell sorting) magnetic separation columns were purchased from Miltenyi Biotec (Bergisch Gladbach, Germany).

Preparation of total PBMC, CD14+ PBMC and CD14-negative (CD14−) PBMC Buffy coats obtained from the blood of healthy donors were provided anonymously by the Red Cross Blood Bank, under a protocol approved by Barwon Health Human Research and Ethics Advisory Committee. To isolate PBMC, the buffy coats were diluted 1 : 1 (v\v) in PBS, layered over Ficoll-Paque solution and centrifuged (400 g for 30 min). The PBMC layer was collected and washed in 5–6 vol. of PBS, isolated by centrifugation (140 g) and resuspended in minimum essential medium containing 10 % (v\v) fetal bovine serum. Prior to CD14+ enrichment, granulocyte contamination of PBMC was reduced to 1 % by performing two Ficoll-Paque separations. A portion of 20 µl of MACS Microbeads, conjugated with a monoclonal mouse anti-(human CD14) antibody, was added to 10( PBMC in 80 µl of PBS and incubated for 15 min at 4 mC. Cells were then washed with 10–20i buffer volume and centrifuged at 300 g for 10 min. The supernatant was aspirated and the cells were resuspended in buffer at a concentration of 10) cells\ 500 µl. The cell suspension was added to a VS+ positive selection column mounted on a magnetic separator. The

RANKL induces osteoclasts from CD14+ human mononuclear cells

Table 1

Summary of PCR primers

Abbreviation : GAPDH, glyceraldehyde-3-phosphate dehydrogenase. Gene

Forward primer

Reverse primer

Product size (bp)

GAPDH RANK OPG RANKL

5h CAGTCAGCCGCATCTTCTTTTG 3h 5h TTAAGCCAGTGCTTCACGGG 3h 5h GTACGTCAAGCAGGAGTGCAATC 3h 5h CAGCACATCAGAGCAGAGAAAGC 3h

5h TGGTTCACACCCATGACGAAC 3h 5h ACGTAGACCACGATGATGTCGC 3h 5h TTCTTGTGAGCTGTGTTGCCG 3h 5h CCCCAAAGTATGTTGCATCCTG 3h

464 497 472 517

CD14− cell population was rinsed through the column with 3i3 ml of PBS and collected for parallel culture. The column was removed from the magnetic separator and the CD14+ cell fraction flushed through. Samples of 10& cells from the PBMC, CD14+ and CD14− fractions were treated with anti-CD14 conjugated to fluorescein isothiocyanate and anti-CD45 conjugated to phycoerythrin, and analysed by fluorescence-activated cytometry using a FACSCalibur (Becton Dickinson).

Gene expression by comparative reverse transcription (RT)-PCR PBMC (2i10() or CD14+ cells (2i10') were settled on to 25 cm# culture flasks and incubated at 37 mC in 5 % CO with hM-CSF (25 ng\ml). After 24 h the cells were # lysed in RNAzol B solution and total RNA was extracted according to the manufacturer’s instructions. For RT and PCR reactions, a Perkin Elmer\Cetus DNA Thermal Cycler was used. RT was performed in the presence of 5 mM MgCl , 1 mM deoxynucleotide mixture, 3.2 µg of # random primers, 50 units of RNase inhibitor and 20 units of AMV (avian myeloblastosis virus) reverse transcriptase. The final mixture was allowed to react at 25 mC for 10 min, 42 mC for 60 min and 95 mC for 5 min to denature the enzyme. Sense and antisense primers were designed using the MacVector program and synthesized by Gibco BRL (Gaithersburg, MD, U.S.A.). Sequences and sizes are shown in Table 1. PCR products were confirmed either by restriction enzyme digestion or by sequencing, and primer pairs spanned intron–exon splice sites, allowing for the detection of mRNA only. PCR amplification was performed with cycles of denaturation at 95 mC for 1 min, annealing at 55 mC for 2 min and extension at 72 mC for 1 min. The reaction mixture contained 40 pmol of each primer, 200 mM dNTPs, 2 µl of 10i reaction buffer, optimized concentration of MgCl (1.0 mM), 1 unit of Taq DNA polymerase and # sterile distilled water to a volume of 20 µl. The mixture was then overlaid with paraffin oil. For comparative RTPCR analysis, the optimal number of cycles for each gene was determined as follows : glyceraldehyde-3-phosphate dehydrogenase, 20 cycles ; RANK and RANKL, 34 cycles ; OPG, 38 cycles. PCR products were resolved on a 1.2 % (w\v) agarose gel and visualized using ethidium

bromide. The size of the bands was confirmed by comparison with a 100 bp DNA ladder (Gibco BRL). cDNA from a sample of human giant cell tumour of bone was used as a positive control, as it was found to express all the genes studied.

Osteoclast generation PBMC (10'\well), CD14+ cells (10&\well) or CD14− cells (10&\well) were added to 6 mm-diameter culture wells containing slices of sterilized bovine cortical bone of dimensions 4 mmi4 mmi0.5 mm (six slices\treatment). For the PBMC cultures, the wells were aspirated at 2 h to remove non-adherent cells and fresh medium was added. The cultures were maintained in 200 µl of minimum essential medium containing 10 % (v\v) fetal bovine serum and 100 units\ml benzyl penicillin for 21 days in the presence of hM-CSF (25 ng\ml) and the presence or absence of sRANKL (40 ng\ml). All media and added factors were replaced twice weekly. In the time-course assay, cell counts at 24 h (when TRAP reactions were not evident) were obtained by staining with Methylene Blue. At later time points TRAP staining allowed the quantification of TRAP+ mononuclear cells (Mo) and TRAP+ multinuclear ( 2 nuclei) cells (MNC). At various times, the cells were then removed from the bone slices by ultrasonication and the slices were processed for scanning electron microscopy. The percentage of bone surface resorbed was quantified at 100–150i magnification.

Cytochemistry of non-specific esterase (NSE) Following 18 h of culture, the bone slices were rinsed to remove non-adherent cells and the adherent cells were fixed in ice-cold acetone for 10 min, then air-dried. The fixed cells were tested for α-naphthyl acetate esterase (NSE) in the absence and presence of 18 mM NaF according to the manufacturer’s instructions (Sigma).

CD marker immunocytochemistry Bone slices containing PBMC or CD14+ cells were removed from culture at various times for analysis of CD # 2000 The Biochemical Society and the Medical Research Society

135

136


marker expression on adherent cells. They were placed in 100 µl of culture medium at 4 mC for immunostaining with fluorescent conjugates of monoclonal antibodies against CD markers. The cells were fixed in ice-cold acetone or 4 % paraformaldehyde and then stained with TOTO-3 (Molecular Probes) in the presence of RNase. Monoclonal antibodies against combinations CD14\ CD13 and CD19\CD3 were used to determine the proportion of cells on the slices expressing the different markers. To determine the proportion of CD14+ cells to the total leucocyte number, the antibody combination CD14\CD45 was used. Fluorescently labelled IgG2a and IgG1 isotype immunoglobulins were used as controls for non-specific binding. Analyses were performed on an inverted Leica DMR-1B microscope with a Bio-Rad 1024 confocal laser scanning microscope.

RESULTS Fluorescence-activated cytometric analysis Fluorescence-activated cytometric analysis of the PBMC samples showed that 13.8p2.8 % (meanpS.E.M. ; n l 6) of the cells were CD14+. Following CD14 enrichment using the MACS system, 89p4 % (n l 4) were CD14+, whereas 0.03 % of the CD14− fraction were CD14+ (results not shown).

Figure 2 Osteoclast generation from peripheral blood CD14+ cells

Figure 1 Comparative expression of glyceraldehyde-3phosphate dehydrogenase (GAPDH), RANK, OPG and RANKL mRNAs in unfractionated PBMC and in purified CD14+ cells, as demonstrated by RT-PCR

Total RNA was extracted from cells and reverse-transcribed in the presence of reverse transcriptase. cDNA samples are : lane A, PBMC ; lane B, CD14+ sample 1 ; lane C, CD14+ sample 2 ; lane D, human giant cell tumour of bone (positive control) ; lane E, negative control (no cDNA). # 2000 The Biochemical Society and the Medical Research Society

CD14+ cells were cultured on bone slices (105/slice) with hM-CSF (25 ng/ml) and sRANKL (40 ng/ml). (A) Culture at 2 h. Cells were labelled with anti-CD14 antibodies conjugated to fluorescein isothiocyanate (yellow–green colour) and viewed using a confocal laser scanning microscope. When compared with the total leucocyte count on this slice, quantified with anti-CD45 antibodies conjugated to phycoerythrin, 86 % of the cells were CD14+. (B) Culture at 21 days. Cells were tested for TRAP and viewed by light microscopy. Numerous TRAP+ MNC (dark red colour) are shown. Nuclei are unstained and appear pale red or white. Areas of a faint TRAP reaction on the bone surface, consistent with resorption lacunae, are indicated by arrows. (C) Culture at 21 days : bone slice viewed by scanning electron microscopy, demonstrating multiple resorption lacunae. Bars l 100 µm.

Gene expression In PBMC, expression of RANK, RANKL and OPG mRNAs was evident when assessed by RT-PCR. However, in purified CD14+ cells only expression of RANK


Table 2 Comparison of TRAP+ MNC generation and percentage bone resorption in assays employing unfractionated PBMC versus purified CD14+ cells

PBMC (106/well) or CD14+ cells (105/well) were incubated on bone slices for 21 days with hM-CSF (25 ng/ml) and sRANKL (40 ng/ml). Values are meanspS.E.M. CV, coefficient of variation. Significance of differences : *P 0.05 compared with PBMC. TRAP+ MNC

Bone resorption

Cells

No. (cells/slice)

CV (%)

Amount (% of slice)

CV (%)

PBMC (n l 6 experiments) CD14+ cells (n l 4 experiments)

220p12 165p47

109 63*

22p6 25p6

129 69*

mRNA was evident (Figure 1). The faint bands seen in lanes B and C of Figure 1 are larger than the OPG mRNA, and when sequenced were non-specific. In addition, no OPG mRNA was detected in CD14+ preparations using TaqMan (real time) PCR (results not shown).

Characterization of the bone-adherent population TRAP and NSE enzyme cytochemistry In cultures of PBMC, CD14+ and CD14−, the cells adherent to the bone slices at 18 h were uniformly mononuclear and negative for TRAP. In the CD14+ cultures at least 95 % of the cells showed moderate to strong reactions for fluoride-sensitive NSE, a characteristic of monocytes (results not shown). In the PBMC cultures the proportion of fluoride-sensitive NSE+ cells was variable (40–80 %) in different regions of the bone slices. The CD14− cultures were not tested for NSE.

CD markers When the CD14+ cells were settled on to bone slices for 2 h and then washed, 88p1.2 % (meanpS.E.M. ; n l 4) showed CD14+ expression, as assessed by direct immunocytochemistry using a confocal laser scanning microscope (Figure 2A). Other markers were rarely seen (CD3, 1 % ; CD13, 2 % ; CD19, 1 %). After 24 and 84 h of culture with hM-CSF and sRANKL, the percentage of cells showing CD14 decreased rapidly, to 32p4.5 % and 7.6p6.0 % respectively.

Osteoclast generation Comparison of assays employing unfractionated PBMC, CD14+ cells and CD14− cells Assays employing either PBMC (10'\well) or CD14+ cells (10&\well) incubated on bone slices in the presence of hM-CSF (25 ng\ml) and sRANKL (40 ng\ml) for 3

weeks consistently generated large numbers of TRAP+ MNC and extensive pit formation on the bone slices. Representative examples of CD14+ assays showing TRAP+ MNC generation and bone resorption are shown in Figures 2(B) and 2(C) respectively. A summary of the results obtained with six assays employing PBMC and four assays employing CD14+ is shown in Table 2. The efficiency of osteoclast generation and the degree of bone resorption obtained with 10& CD14+ cells were similar to those achieved with 10' PBMC, which is likely to reflect the prevalence of CD14+ cells in PBMC. However, when comparing PBMC and CD14+ cell osteoclastogenesis assays, we observed a greater intra-assay variability (P 0.05) in the number of TRAP+ MNC per bone slice and in the percentage bone resorption when PBMC were used than when CD14+ cells were used (Table 2). When CD14− cells (10&\well) were settled on the bone slices, few were adherent at 84 h (88p26\slice), and fewer (47p7\slice) TRAP+ Mo were present after 3 weeks of culture. No TRAP+ MNC or bone resorption were evident after 3 weeks of culture (results not shown).

Time course of TRAP+ Mo and TRAP+ MNC generation and bone resorption in assays employing CD14+ cells When CD14+ cells were incubated on bone slices in the presence of hM-CSF alone, large numbers of TRAP+ Mo were evident by 1 week (Figure 3A). The number of TRAP+ Mo increased in a linear manner, resulting in an approximate doubling during the 21-day incubation period. Small numbers of TRAP+ MNC were present at 14 days, which had increased by 21 days, but no resorption was detected by scanning electron microscopy, indicating that these cells were not osteoclasts. In the presence of both hM-CSF and sRANKL there was a 30 % increase in Mo from 24 h to 7 days, but throughout the 21 days of the culture there was no overall trend to increase. Substantial numbers of TRAP+ MNC were present at 14 days, with a marked increase by 21 days (Figure 3B). A small amount of bone resorption was seen at 7 days (when no TRAP+ MNC were present) and this increased progressively, with an average of 20–25 % # 2000 The Biochemical Society and the Medical Research Society

137

138


DISCUSSION

Figure 3 Time course of generation of human osteoclasts from CD14+ cells

CD14+ cells were cultured on bone slices (105/slice ; six slices/treatment) with hMCSF (25 ng/ml) in the presence or absence of sRANKL (40 ng/ml). After culture for 1, 7, 14 and 21 days, the number of TRAP+ cells were counted and bone resorption was quantified by scanning electron microscopy. (A) TRAP+ Mo ; (B) TRAP+ MNC ; (C) percentage of bone slice resorbed. Note that, in the absence of sRANKL, no bone resorption was seen. Columns that do not share the same letter (a, b, c, d) are significantly different (P 0.05 ; ANOVA Fischer’s multiple comparison). * At 1 day CD14+ cells were not TRAP+, and therefore the Mo number at this time was quantified by Methylene Blue staining.

of the bone slice surface resorbed at 21 days (Figure 3C). Approx. 5i10$ bone-adherent CD14+ cells generated 165p47 TRAP+ MNC (containing " 10$ nuclei) and approx. 5i10$ TRAP+ Mo, indicating that about 20 % of the nuclei are eventually contained within osteoclasts. This calculation does not account for cell proliferation and apoptosis that is likely to occur in these cultures. # 2000 The Biochemical Society and the Medical Research Society

In the present study we have shown that highly enriched CD14+ cells derived from human peripheral blood express mRNA for RANK and differentiate into boneresorbing osteoclasts in the presence of sRANKL and hM-CSF. This was associated with a rapid loss of the CD14 marker over the first few days of culture. Bone resorption was evident by day 7 of culture, preceding the appearance of TRAP+ MNC, indicating that a proportion of the CD14+ cells differentiate rapidly to form mononuclear bone-resorbing osteoclasts. In contrast, a significant proportion of the cells remained mononuclear after 21 days. These observations suggest that heterogeneity exists both in the maturity of the CD14+ osteoclast-precursor phenotype and in the potential of the cells to become osteoclasts. However, as we did not extend the cultures beyond 21 days, the possibility exists that these mononuclear cells may differentiate into multinuclear osteoclasts with longer incubation times. The ability of 10& CD14+ cells and 10' PBMC to produce similar numbers of TRAP+ MNC and similar percentage bone resorption is consistent with approx. 14 % of PBMC being CD14+ monocytes. However, the variability of osteoclast generation from PBMC was consistently greater, with some replicates showing virtually no osteoclastogenesis despite the presence of large numbers of TRAP+ Mo. Interestingly, we found that unfractionated PBMC expressed OPG mRNA, derived from CD8−\CD4− lymphocytes (results not shown). This raises the possibility that increased production of OPG, or other factors, by lymphocytes within some PBMC preparations may result in reduced osteoclast generation. In addition to the strongly CD14+ monocyte, there exists a subset comprising about 10 % of all blood monocytes which is only weakly CD14+ but strongly CD16+ [22], as well as other minor subsets including CD14−\CD16++, CD14+\CD33++ and CD14++\CD56+ [23]. The role of these subsets in osteoclastogenesis is unknown. In vivo, osteoclastogenesis is a complex process involving not only osteoclast haemopoietic precursors and mesenchymal stromal\osteoblasts but also other haemopoietic cells that are not osteoclast precursors but may regulate the process. An example of this was seen in the recent experiments of Horwood and colleagues [24], who showed that inhibition of osteoclastogenesis by interleukin-18 was an indirect effect, mediated by T lymphocyte granulocyte\macrophage colony-stimulating factor production. Thus, although the availability of recombinant sRANKL has allowed the development of osteoclastogenesis models which do not require co-culture with stromal\osteoblasts, these models employing PBMC essentially remain ‘ co-cultures ’ of osteoclast precursors and accessory haemopoietic cells.


Small numbers of osteoclasts have been generated from haemopoietic precursors, including granulocyte\ macrophage colony-stimulating factor-mobilized PBMC [25], and CD34+ cells [26] in the absence of stromal\ osteoblasts or the addition of exogenous sRANKL. Faust et al. [27] cultured dense populations of unfractionated PBMC (1.5i10' cells\cm#) for 22 days in the presence of serum alone. A number of calcitonin-receptor-positive mono- and multi-nuclear cells and a minor degree of bone resorption were found, indicating that some genuine osteoclasts were generated. Interestingly, RANKL mRNA was not detectable at day 1, but was present at day 5, declining progressively thereafter to undetectable levels by day 18. In this model, it seems likely that expression of RANKL and M-CSF by subpopulation(s) of cells within the PBMC facilitated the differentiation of small numbers of osteoclasts. Taken together, these studies suggest that, under some circumstances, all the essential elements for osteoclast generation [osteoclast precursors, M-CSF and RANKL, or a factor(s) with a similar effect] can be found in the blood. There is also evidence that non-adherent PBMC (which are predominantly lymphocytes) can generate an inhibitor(s) of osteoclastogenesis that is suppressed by dexamethasone [19]. Therefore, in order to determine ‘ direct ’ versus ‘ indirect ’ effects on osteoclastogenesis, homogeneous populations of osteoclast precursors must be prepared. We believe that the use of highly purified CD14+ cells, as demonstrated here, constitutes the first step in this endeavour. It will now be important to define the phenotype of the subpopulation(s) within the CD14+ population with the highest potential for differentiation into osteoclasts, and also to characterize the effects of other haemopoietic cells on these osteoclast precursors.

ACKNOWLEDGMENTS This work was supported by the National Health and Medical Research Council of Australia (Project Grant 980746). D.E.M. holds a Research Fellowship of the Faculty of Medicine, Dentistry and Health Sciences, The University of Melbourne.

REFERENCES 1 Chambers, T. (1992) Regulation of osteoclast development and function. In Biology and Physiology of the Osteoclast (Rifkin, B. R. and Gay, C. V., eds.), pp. 105–128, CRC Press, Boca Raton, FL 2 Suda, T., Takahashi, N. and Martin, T. J. (1992) Modulation of osteoclast differentiation. Endocr. Rev. 13, 66–80 3 Roodman, G. D. (1999) Cell biology of the osteoclast. Exp. Hematol. 27, 1229–1241

4 Udagawa, N., Takahashi, N., Akatsu, T. et al. (1990) Origin of osteoclasts : mature monocytes and macrophages are capable of differentiating into osteoclasts under a suitable microenvironment prepared by bone marrowderived stromal cells. Proc. Natl. Acad. Sci. U.S.A. 87, 7260–7264 5 Quinn, J. M., McGee, J. O. and Athanasou, N. A. (1994) Cellular and hormonal factors influencing monocyte differentiation to osteoclastic bone-resorbing cells. Endocrinology 134, 2416–2423 6 Fujikawa, Y., Quinn, J. M., Sabokbar, A., McGee, J. O. and Athanasou, N. A. (1996) The human osteoclast precursor circulates in the monocyte fraction. Endocrinology 137, 4058–4060 7 Fuller, K. and Chambers, T. J. (1998) Parathyroid hormone induces bone resorption in human peripheral blood mononuclear cells. Int. J. Exp. Pathol. 79, 223–233 8 Quinn, J. M., Neale, S., Fujikawa, Y., McGee, J. O. and Athanasou, N. A. (1998) Human osteoclast formation from blood monocytes, peritoneal macrophages, and bone marrow cells. Calcif. Tissue Int. 62, 527–531 9 Matsuzaki, K., Katayama, K., Takahashi, Y. et al. (1999) Human osteoclast-like cells are formed from peripheral blood mononuclear cells in a coculture with SaOS-2 cells transfected with the parathyroid hormone (PTH)\PTHrelated protein receptor gene. Endocrinology 140, 925–932 10 Suda, T., Takahashi, N., Udagawa, N., Jimi, E., Gillespie, M. T. and Martin, T. J. (1999) Modulation of osteoclast differentiation and function by the new members of the tumor necrosis factor receptor and ligand families. Endocr. Rev. 20, 345–357 11 Wong, B. R., Josien, R., Lee, S. Y. et al. (1997) TRANCE (tumor necrosis factor [TNF]-related activation-induced cytokine), a new TNF family member predominantly expressed in T cells, is a dendritic cell-specific survival factor. J. Exp. Med. 186, 2075–2080 12 Anderson, D. M., Maraskovsky, E., Billingsley, W. L. et al. (1997) A homologue of the TNF receptor and its ligand enhance T-cell growth and dendritic-cell function. Nature (London) 390, 175–179 13 Yasuda, H., Shima, N., Nakagawa, N. et al. (1998) Osteoclast differentiation factor is a ligand for osteoprotegerin\osteoclastogenesis-inhibitory factor and is identical to TRANCE\RANKL. Proc. Natl. Acad. Sci. U.S.A. 95, 3597–3602 14 Lacey, D. L., Timms, E., Tan, H. L. et al. (1998) Osteoprotegerin ligand is a cytokine that regulates osteoclast differentiation and activation. Cell 93, 165–176 15 Nakagawa, N., Kinosaki, M., Yamaguchi, K. et al. (1998) RANK is the essential signaling receptor for osteoclast differentiation factor in osteoclastogenesis. Biochem. Biophys. Res. Commun. 253, 395–400 16 Hsu, H., Lacey, D. L., Dunstan, C. R. et al. (1999) Tumor necrosis factor receptor family member RANK mediates osteoclast differentiation and activation induced by osteoprotegerin ligand. Proc. Natl. Acad. Sci. U.S.A. 96, 3540–3545 17 Simonet, W. S., Lacey, D. L., Dunstan, C. R. et al. (1997) Osteoprotegerin : a novel secreted protein involved in the regulation of bone density. Cell 89, 309–319 18 Tsuda, E., Goto, M., Mochizuki, S. et al. (1997) Isolation of a novel cytokine from human fibroblasts that specifically inhibits osteoclastogenesis. Biochem. Biophys. Res. Commun. 234, 137–142 19 Matsuzaki, K., Udagawa, N., Takahashi, N. et al. (1998) Osteoclast differentiation factor (ODF) induces osteoclastlike cell formation in human peripheral blood mononuclear cell cultures. Biochem. Biophys. Res. Commun. 246, 199–204 20 Quinn, J. M., Elliott, J., Gillespie, M. T. and Martin, T. J. (1998) A combination of osteoclast differentiation factor and macrophage-colony stimulating factor is sufficient for both human and mouse osteoclast formation in vitro. Endocrinology 139, 4424–4427 21 Massey, H. M. and Flanagan, A. M. (1999) Human osteoclasts derive from CD14-positive monocytes. Br. J. Haematol. 106, 167–170


139

140


22 Ziegler-Heitbrock, H. M. (1996) Heterogeneity of human blood monocytes : the CD14jCD16j subpopulation. Immunol. Today 17, 424–428 23 Rothe, G., Gabriel, H., Kovacs, E. et al. (1996) Peripheral blood mononuclear phagocyte subpopulations as cellular markers in hypercholesterolemia. Arterioscler. Thromb. Vasc. Biol. 16, 1437–1447 24 Horwood, N. J., Udagawa, N., Elliott, J. et al. (1998) Interleukin 18 inhibits osteoclast formation via T cell production of granulocyte macrophage colony-stimulating factor. J. Clin. Invest. 101, 595–603 25 Purton, L. E., Lee, M. Y. and Torok-Storb, B. (1996)

Normal human peripheral blood mononuclear cells mobilized with granulocyte colony-stimulating factor have increased osteoclastogenic potential compared to nonmobilized blood. Blood 87, 1802–1808 26 Matayoshi, A., Brown, C., DiPersio, J. F. et al. (1996) Human blood-mobilized hematopoietic precursors differentiate into osteoclasts in the absence of stromal cells. Proc. Natl. Acad. Sci. U.S.A. 93, 10785–10790 27 Faust, J., Lacey, D. L., Hunt, P. et al. (1999) Osteoclast markers accumulate on cells developing from human peripheral blood mononuclear precursors. J. Cell. Biochem. 72, 67–80 Received 25 November 1999/7 March 2000; accepted 5 April 2000
