SEARCH

SEARCH BY CITATION

Keywords:

  • osteoclasts;
  • osteoclastogenesis;
  • myeloid progenitors

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. References

Murine BM was fractionated using a series of hematopoietic markers to characterize its osteoclast progenitor populations. We found that the early osteoclastogenic activity in total BM was recapitulated by a population of cells contained within the CD11b−/low CD45RCD3CD115high fraction.

Introduction: Osteoclasts are of hematopoietic origin and they have been shown to share the same lineage as macrophages. We further characterized the phenotype of osteoclast progenitor populations in murine bone marrow (BM) by analyzing their cell surface markers.

Materials and Methods: We used fluorescence-activated cell sorting (FACS) to identify the subsets of BM cells that contained osteoclast progenitors. We fractionated BM according to several markers and cultured the sorted populations for a period of 2–6 days with macrophage-colony stimulating factor (M-CSF) and RANKL. The numbers of multinucleated osteoclast-like cells (OCLs) that formed in the cultures were counted.

Results: We found that the CD45RCD11b−/low population recapitulated the early osteoclastogenic activity of total BM. In addition, although previous experiments indicated that osteoclastogenic activity was enriched within the CD45R+ population, we found that highly purified CD45R+ BM was incapable of differentiating into osteoclasts in vitro. We also found that CD45RCD11bhigh BM cells were an inefficient source of osteoclast progenitors. However, CD11b was transiently upregulated by cells of the CD45RCD11b−/low fraction early (within 24 h) during culture with M-CSF. Finally, further fractionation of BM using CD115 and CD117 showed that, as osteoclast precursor cells matured, they downregulate CD117 but remain CD115+. Curiously, pure populations of CD117 (CD115high) cells isolated fresh from BM have low osteoclastogenic activity in vitro.

Conclusions: We provided a refined analysis of the precise subpopulations of murine BM that are capable of differentiating into OCLs in vitro when treated with M-CSF and RANKL.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. References

THE HEMATOPOIETIC ORIGIN of osteoclasts was first shown >30 years ago by DJ Walker. (1, 2) Multiple evidence has led to the hypothesis that osteoclasts and macrophages share a common hematopoietic progenitor. Both cell types share some morphological features, including their ability to fuse to form polykaryon. Also, granulocyte macrophage-colony forming cells (GM-CSF)(3, 4) and macrophage cell lines(5) are able to differentiate into osteoclasts on addition of vitamin D3. Furthermore, data from various knockout animal models support this lineage relationship. The op/op mouse, which has a spontaneous early termination mutation in the M-CSF gene, (6) is unable to form normal numbers of osteoclasts or macrophages (at least early in its development). Similarly, PU.1−/− animals display a lack of both macrophages and osteoclasts. (7) In addition, animals deficient in c-fos are osteopetrotic and display elevated levels of mature macrophages defined as CD11b+ (also known as Mac-1) F4/80+ cells, which suggests that a lineage shift between osteoclasts and macrophages has occurred. (8, 9)

Several groups have defined osteoclasts progenitor populations from bone marrow (BM) and spleen using in vitro read out assays that include co-culture with stimulated stroma or direct culture of candidate populations with RANKL and macrophage-colony stimulating factor (M-CSF), which drive cells to differentiate into osteoclasts. (10, 11) Whole bone marrow suspensions incubated with M-CSF expand a population of CD11b+ cells with high osteoclastogenic potential when cultured with RANKL. From this observation, it has been proposed that CD11b is a marker for osteoclast progenitors. More defined dissections of progenitor populations from BM have been reported. Muguruma and Lee(12) identified progenitor populations in fractions of BM negative for markers against B lymphocytes (CD45R, also known as B220), granulocytes (Gr-1), macrophages (CD11b), and erythroid progenitors (YW25.12.7). These populations were distinct to hematopoietic stem cells because of their lack of reactivity with the Sca-1 antibody. Later studies by Arai et al. (13) characterized different populations with the ability to generate osteoclasts after dissection of the CD117 (also known as c-kit) positive fraction from BM using antibodies against the M-CSF receptor, CD115 (also known as c-fms) and anti-CD11b. From those studies, a population characterized by the phenotype: CD117+CD115+CD11blow, with high osteoclast precursor frequency, was isolated. These cells were not committed solely to osteoclast lineage because they also could generate macrophages and dendritic cells. (14)

In summary, these data imply that there exists a strong relationship between osteoclasts and macrophages. They also show that multiple populations in the myeloid lineage have the capacity to differentiate into osteoclasts. Interestingly, several reports (including our previous work(15)) have shown that CD45R/B220 might identify a unique population of osteoclast progenitors in the BM, (16, 17) which is regulated by estrogen. (18, 19)

However, the sum of available information has not generated a consensus regarding the identification of progenitor populations from murine BM that are able to differentiate into osteoclasts. Therefore, we attempted to dissect the osteoclast lineage pathway by defining its cellular components. We used fluorescence-activated cell sorting (FACS) to purify various populations of cells from murine BM and assayed these for their ability to differentiate into mature osteoclast-like cells (OCLs) when treated in vitro with M-CSF and RANKL.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. References

Animals

BM cells were derived from 7- to 10-week-old male C57BL/6 mice, which were purchased from Charles River. Animals were housed in the Center for Laboratory Animal Care at the University of Connecticut Health Center. The Institutional Animal Care and Use Committee approved all animal protocols used in this study.

Antibodies

The antibodies used for sorting and flow cytometric analysis were the following: anti-mouse CD45R-FITC (clone RA3-6B2), anti-mouse CD3-FITC (clone 145–2C11), and anti-mouse CD11b-PE (clone M1/70) or CD45R-PECy7 (clone RA3-6B2), CD3-PECy7 (clone145-2C11), anti-mouse CD11b-FITC (clone M1/70-conjugated in-house), anti-mouse CD117-APC (clone 2B8), and anti-mouse CD115-biotin (AFS98 hybridoma, a kind gift of Dr Shin-Ichi Nishikawa, Kyoto University Medical School) used in combination with streptavidin-phycoerythrin (PE). Unless indicated, all antibodies and secondary step reagents were purchased from commercial sources (Pharmingen and eBioscience).

Cell sorting

Sterile BM preparations were obtained by flushing femurs, tibias, and occasionally humeri with staining medium (1× Hank's balanced salt solution {HBSS}; 10 mM HEPES; 2% newborn calf serum) using a 25-gauge needle. After washing, red blood cells were lysed using ammonium chloride and filtered. To decrease nonspecific antibody binding, cells were stained for 45 minutes on ice in the presence of rat Ig (Sigma). After washing, secondary antibody was used when necessary. Cells were washed and resuspended in staining medium containing propidium iodide (PI) (1 μg/ml), and the populations of interest were sorted by FACS using a FACS Vantage SE with FACSDiva option (BD Biosciences), excluding dead cells, highly positive for PI. When a double sort was performed, cells from the first sort were spun down and resuspended in staining medium containing PI before being sorted for a second time.

A two-step approach was taken to sort the multiple CD117/CD115 populations (i.e., populations I-VI). In an initial sort, we gated on the (CD45R, CD3, CD11b) negative (TN) population, and separated the CD117highCD115high population (IV) from the CD117neg/lowCD115high population (V + VI). In a subsequent second sort, population V was separated from VI. In addition, population IV was resorted. This approach ensured isolation of the populations to extremely high purity.

Some sorted populations were cultured in complete α-MEM without phenol red (Gibco BRL). For flow cytometry analyses or resort of these cells after culture, a nonenzymatic cell dissociation solution (Sigma) was used to recover cells from the plate/tube, followed by staining as described. The viability and purity of the sorted populations were determined by evaluating aliquots using a Calibur flow cytometer (BD Biosciences) and analyzed using CellQuest (BD Biosciences).

OCL cultures

Sorted cells were spun down and resuspended in α-MEM (Gibco BRL) supplemented with 10% heat-inactivated fetal bovine serum (HIFBS) (Hyclone, Logan, UT, USA), glutamine, penicillin, streptomycin, and M-CSF and RANKL, both at 30 ng/ml (R&D systems). The number of viable cells was determined using trypan blue. The cells were plated (density as indicated in each experiment) in triplicates in 96-well flat bottom plates, in 200 μl of complete α-MEM with 30 ng/ml M-CSF and RANKL for different periods of time as indicated. The medium was changed on days 3, 4, and 5.

To score multinucleated osteoclasts, cells were fixed with 2.5% glutaraldehyde (Polysciences, Warrington, PA, USA) in PBS and stained for TRACP using the Leukocyte Acid Phosphatase kit according to manufacturer's instructions (Sigma). TRACP+ multinucleated cells (more than three nuclei) were counted.

To evaluate the size of the OCLs over time, we counted the number of nuclei per OCL in 50 OCLs for each of the three wells (n = 150). The mean number of nuclei per OCL (MNN) and SE were calculated for each time-point, and ANOVA analysis was performed to determine statistically significant differences.

Bone resorption

The ability of sorted populations to resorb bone was assessed by culturing them on the surface of UV-sterilized devitalized bovine cortical bone slices (4.4 × 4.4 × 0.2 mm) that were placed in the wells of a 96-well plate in α-MEM with M-CSF and RANKL (30 ng/ml) for 7 or 14 days. The bone slices were fixed with 2.5% glutaraldehyde in PBS and stained for TRACP according to the manufacturer's recommendations. After TRACP+ multinucleated osteoclasts were visualized by light microscopy, bone slices were sonicated in 0.25 M NH4OH to remove osteoclasts. Bone slices were stained with 1% toluidine in 1% borax buffer to visualize resorption pits.

To assess the resorption potential of OCLs derived from different groups of progenitor populations, we determined the mean area of pits in each group at two different time-points (days 7 and 15). The areas (mm2) of 15 individual pits (minimal area/pit of ∼0.00015 mm2 to avoid pits made by mononuclear OCLs) per group and time-point were recorded using the Via-160 video Image maker-measurement system (Boeckeler Instruments), and the mean areas per pit and SEs were calculated. A two-tailed t-test was performed to determine statistically significant differences.

Calcitonin receptor binding assay

Sorted populations were plated into a 4-ml slideflask (Nalge Nunc International) in complete α-MEM with 30 ng/ml of M-CSF and RANKL. For the last 2 h of culture, cells were incubated with radiolabeled125I-salmon calcitonin (0.02 μCi, 50,000 cpm/ml; New England Nuclear/Dupont) with or without excess (100×) cold calcitonin for 2 h. Cells were washed twice in PBS, fixed with 2.5% glutaraldehyde in PBS, and stained for TRACP. Slides were dipped in LM-1 photographic emulsion (1:1 dilution with 1.7% glycerol; Amersham) and developed by autoradiography to show radiolabeled calcitonin binding to cells.

Immunofluorescence staining

Sorted cells were plated on glass chamber slides (Nalge Nunc International) and cultured in α-MEM with M-CSF and RANKL (30 ng/ml) until mature osteoclasts formed. Cells were fixed for 30 minutes at 4°C with fresh 4% paraformaldehyde in PBS, washed with cold PBS, and stained for 1 h at 4°C with CD11b-PE (clone M1/70; Pharmingen) or anti-CD4-PE (clone GK1.5; eBioscience) as an isotype control (IgG2bκ). After three washes with cold PBS, the chambers were removed, and the slides were mounted with glycerol:PBS (1:1 solution) and a cover slip. Cells were visualized using a Zeiss Axiovert 200 microscope equipped with the motorized X-Y-Z platform, a motorized fluorescent cube, and an AxioCam color digital camera. The slides were scanned under fluorescent light, briefly dipped in xylene to remove the cover slip, and TRACP stained as described previously before being scanned again by light microscopy.

Statistical analysis

All the experiments were repeated independently at least three times with similar results, and a representative experiment is shown.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. References

Ability of highly purified CD45R+ murine BM cells to differentiate into OCLs in vitro

Several reports, including our own, have indicated that osteoclasts could be derived from preparations of purified CD45R+ murine BM cells in vitro. (15, 17–19) However, because of some variability in our results from multiple experiments, we set out to determine the possibility that contamination of the purified CD45R+ preparation by a minor population of CD45R cells might influence the observed osteoclastogenic activity of purified CD45R+ murine BM cells. Therefore, we compared the in vitro osteoclastogenic activity of purified murine BM CD45R+(CD11b) cells after a single and double FACS separation. Cells were cultured for 6 days in the presence of M-CSF and RANKL (Fig. 1). Surprisingly, we found that highly purified CD45R+-sorted cells (99.87% after double sort) were unable to differentiate into osteoclasts in vitro, whereas single-sorted (99.14% purity) CD45R+ cells did so readily.

thumbnail image

Figure FIG. 1.. Double-sorted CD45R+CD11b BM cells do not differentiate into OCLs in vitro. (A) CD45R+ murine BM cells were first sorted to a purity of 99.14% (left) and then sorted a second time to a purity of 99.87% (right). (B) After each sort, 105 cells per well were plated (in triplicate) with α-MEM and 30 ng/ml M-CSF and RANKL for 4–6 days. After TRACP staining, TRACP+ multinucleated cells (more than three nuclei) were counted as OCLs. Values represent the mean of three wells ± SE.20

Download figure to PowerPoint

Ability of CD45RCD11b−/low and CD45RCD11bhigh murine BM cells to differentiate into OCLs in vitro

To further characterize the cells in the CD45R population that differentiated into OCLs in vitro, we fractionated total BM according to the expression of the antigens CD45R and CD11b. The latter marker is used to identify macrophage populations. Because macrophages and osteoclasts are believed to be closely related, we anticipated that the osteoclastogenic activity in BM cells would reside in the CD11b+ population. The CD11bhighCD45R fraction represents ∼64% of total BM cells in a C57BL/6 mouse, whereas the double negative (DN) CD11b−/lowCD45R represent about 12.5%. We double-sorted CD11bhighCD45R and DN cells to high purity (99.95% and 99.99%, respectively; Fig. 2A) and plated these populations at multiple densities in 96-well plates with M-CSF and RANKL (30 ng/ml for each). Cells were cultured for 2–6 days, and the ability of the isolated cell populations to differentiate into OCLs in vitro was assessed (Fig. 2). Interestingly, we found that CD11bhighCD45R cells were not as efficient or rapid in their ability to form OCLs, as DN populations, which formed multinucleated OCLs (more than three nuclei) within 3 days. Furthermore, OCL formation in DN cell cultures occurred at much lower plating densities than with the CD11bhighCD45R fraction (50 × 103 CD11bhighCD45R cells produced 373 ± 5 OCLs per well on day 4, whereas it only required 2.5 × 103 DN to generate the same number of OCLs, 373 ± 57 OCLs, on day 4; Fig. 2C). As a further demonstration of the efficiency of the DN population to generate OCLs, 1.25 × 103 DN cells produced 176 ± 35 OCLs in 4 days. OCLs from the DN fraction became larger as the culture progressed (Fig. 2B), as shown by the increase in the mean number of nuclei per OCLs (MNN) over time, reaching a peak in size after 4 or 5 days of culture (depending on the density at which the cells were plated). When 5 × 103 DN cells were plated per well, the mean number of OCLs per well peaked at day 4 (Fig. 2C). The mean number of nuclei per OCL peaked at day 5 (day 3-MNN = 6 ± 0.3; day 4-MNN = 16 ± 1; day 5-MNN = 32 ± 2; day 6-MNN = 11 ± 1), showing that OCLs grew bigger in size over time (the mean number of nuclei per OCL on days 3 and 4 was significantly lower than on day 5; p < 0.05). At the end of the culture period, cell death was observed in the OCLs cultures (days 5 and 6; Fig. 2B), as evidenced by the presence of TRACP+ cell remnants in the cultures (indicated by ⋄ symbol). OCLs could also be derived from CD11bhighCD45R cells, although the wells had to be seeded with a higher density of cells, and a longer culture period was required. Cell fragmentation was also observed at day 6 in higher density CD11bhighCD45R cultures.

thumbnail image

Figure FIG. 2.. Double-sorted DN and CD11bhigh CD45R BM cell populations both differentiate into OCLs in vitro but with different efficacy. (A) DN (CD11b−/lowCD45R), CD11bhighCD45R, and CD45R+CD11b populations were double sorted from the same bone marrow preparation. The purity of the samples after sorting is as indicated (%). (B) Five × 103 cells/well (DN and CD11bhigh CD45R) or 100 × 103 cells/well (CD45R+CD11b) were plated in 96-well plates (in triplicate) in α-MEM + 30 ng/ml M-CSF and RANKL for 3–6 days and stained for TRACP. (C) DN and CD11bhighCD45 cells were plated in 96-well plates at different densities: 2.5 × 103, 5 × 103, 12.5 × 103, or 50 × 103 cells per well (in triplicate) in α-MEM + 30 ng/ml M-CSF and RANKL for 2–6 days. After TRACP staining, multinucleated cells (more than three nuclei) were counted. The results are represented as the mean of 3 wells ± SE. ND, not done. ⋄ indicates remnants of fragmented OCLs.20

Download figure to PowerPoint

Individual contribution of TN and the CD11bhighCD45RCD3 fractions to the osteoclastogenic activity of total BM

To further characterize the osteoclast progenitor population in the DN fraction of total murine BM, we tested the ability of CD11b−/lowCD45RCD3+ cells to differentiate into OCLs. We found that sorted CD11b−/lowCD45RCD3+ cells plated at densities of 2.5 or 5 × 103 cells per well were unable to differentiate into OCLs (data not shown), whereas sorted “triple negative” (TN) CD11b−/lowCD45RCD3 were enriched in OCL precursors in comparison with double negative CD11b−/lowCD45R cells (data not shown).

To test the contribution of the TN fraction to the osteoclastogenic activity of total BM, we compared the number of OCLs formed over time by total BM with the number of OCLs formed by the double-sorted CD11b−/lowCD45RCD3 (TN) and CD11bhighCD45RCD3 fractions. In these experiments, cells were cultured separately or mixed together (Fig. 3). Because the TN and CD11bhighCD45RCD3 fractions represent ∼11% and 65% of total BM, respectively, we plated 5 × 103 of the TN, 30 × 103 of the CD11bhighCD45RCD3 (separately and added together), and 45.5 × 103 of total BM, because these plating densities reproduced the ratio of each population within total murine BM. We found that the osteoclastogenic activity of BM at day 3 was entirely recapitulated by the TN fraction. This fraction also largely contributed to the osteoclastogenic activity of total BM at day 4. In contrast, at days 5 and 6, the total number of OCLs in all populations except the CD11bhighCD45RCD3 fraction (i.e., total BM, TN and TN + CD11bhighCD11bhighCD45RCD3) dramatically decreased, as the number of fragmented osteoclasts increased in these wells. In contrast, the CD11bhighCD45RCD3 fraction gave rise to a significant number of OCLs at later time-points, and this response was similar to that of the CD11bhighCD45R fraction described in Fig. 2.

thumbnail image

Figure FIG. 3.. The early osteoclastogenic activity of total bone marrow is recapitulated by the TN fraction. Total bone marrow was fractionated according to CD11b, CD45R, and CD3. Double-sorted CD11bhighCD45RCD3 (30 × 103) cells, double-sorted TN (CD11b−/lowCD45RCD3; 5 × 103) cells, both populations added together, and total bone marrow (45.5 × 103) cells were plated in triplicates in 96-well plates for 3–6 days in the presence of M-CSF and RANKL. The mean number of OCLs per well over time was plotted for each population.20

Download figure to PowerPoint

Ability of OCLs derived from the CD11bhighCD45RCD3 and TN fractions to express calcitonin receptor and to form resorption pits on bovine cortical bone slices

To further show that the OCLs we produced in vitro had the characteristics of bona fide osteoclasts, we tested their ability to form resorption pits in bovine cortical bone slices and their ability to express calcitonin receptor (Fig. 4). Both sorted TN (10 × 103 cells/bone slice) and CD11bhighCD45RCD3 (100 × 103 cells/bone slice) formed resorption pits when they were cultured on bovine cortical bone slices. However, the resorption pits generated by OCLs derived from TN populations were larger (higher mean area/pit) and more numerous (Fig. 4A; middle) than these derived from CD11bhighCD45RCD3 populations. This difference in the mean area/pit was statistically significant after 7 days of culture (0.0019 ± 0.0003 mm2 for TN samples versus 0.0011 ± 0.0002 mm2 for CD11bhighCD45RCD3 samples; p < 0.05) but not after 15 days (0.0038 ± 0.0008 mm2 for TN samples versus 0.0035 ± 0.007 mm2 for CD11bhighCD45RCD3 samples). In addition, the mean area/pit of the day 15 samples (both in the TN and CD11bhighCD45RCD3 samples) was significantly larger than the correspondent day 7 samples (p < 0.05).

thumbnail image

Figure FIG. 4.. OCLs generated from TN and CD11bhighCD45RCD3 murine BM cells are able to form resorption pits on bovine cortical bone slices and express calcitonin receptor. (A) Double-sorted populations were plated onto bovine cortical bone slices and cultured in α-MEM, 10% HIFCS, and M-CSF and RANKL (30 ng/ml). Ten × 103 TN (CD11b−/lowCD45RCD3) cells per well and 100 × 103 CD11bhighCD45RCD3 per well were incubated for 7 or 15 days. After glutaraldehyde fixation, TRACP assay was performed on the slides (top). After sonication, toluidine blue staining was done (middle and bottom) to visualize the resorption pits. (B) Double-sorted populations were cultured. Sorted 50 × 103 TN (1-4) and 750 × 103 CD11bhighCD45RCD3 (5-8) cells were plated per flasks for 4 or 5 days, respectively. For each group, some of the wells were incubated with125I-labeled calcitonin (1, 2, 5, and 6) or125I-labeled calcitonin with 100 times excess unlabeled calcitonin to monitor the specificity of binding (3, 4, 7, and 8). After fixation and TRACP staining, each slide was examined under bright field and dark field optics to visualize calcitonin binding.20

Download figure to PowerPoint

As expected for cells with osteoclast phenotype, the OCLs generated by both fractions expressed calcitonin receptors, as determined by a radiolabeled125I-salmon calcitonin-binding assay (Fig. 4B).

Expression of CD11b by osteoclast progenitors

CD11b is an integrin chain (αM) that is expressed predominantly in mature myeloid cells. (20, 21) We determined whether osteoclast precursors expressed CD11b at any stage of their differentiation because expression of this antigen might explain the osteoclastogenic activity that we observed in the CD11bhighCD45RCD3 fraction. We therefore sorted TN (representing ∼12% of the total BM) and plated them overnight in the presence of M-CSF and RANKL. After overnight incubation, we found that CD11b was upregulated in ∼50% of the cells (Fig. 5B, bottom left), and virtually 100% of these CD11bhigh cells (day1-CD11bhigh) also expressed high levels of CD115 (the receptor for M-CSF). In contrast, only 14% of the starting population expressed CD115. We also found that there were no CD115high cells within the CD11b−/low population (day1-CD11b−/low) after 1 day of culture with M-CSF and RANKL. However, ∼63% of the day1-CD11b−/low cells were CD115int. We sorted the day1-CD11bhigh and day1-CD11b−/low populations, plated them with M-CSF and RANKL at a density of 5 × 103 cells per well, and determined the number of OCLs per well after an extra 2, 3, or 4 days of incubation. Results from these cultures were compared with the numbers of OCLs that were generated by the starting population (Fig. 5A). We found that the osteoclastogenic activity of TN cells at day 3, which corresponds to the osteoclastogenic activity of total BM, was totally recapitulated by the population that upregulated CD11b after 24 h of culture in M-CSF and RANKL. The number of OCLs in the cultures decreased at days 4 and 5 as OCLs started to fragment. Interestingly, cells from the day1-CD11b−/low population (after overnight culture) were also able to differentiate into OCLs but not before day 5. They also eventually upregulate the expression of CD11b and CD115 (data not shown).

thumbnail image

Figure FIG. 5.. TNs transiently upregulate CD11b. (A) Sorted TN (CD11b−/lowCD45RCD3) cells were cultured overnight with M-CSF and RANKL. Fifteen hours later, CD11bhigh and CD11b−/low cells were sorted(according to the gates in B), and 5 × 103 cells per well were plated with M-CSF and RANKL for an extra 2–4 days. TN cells were plated on day 0 as controls (5 × 103/well). The OCLs counts for each of the three populations on days 3–5 were plotted. Results are expressed as the mean of 2–3 wells/time-point ± SE. (B) Flow cytometry analysis of the CD115 and CD11b expression (%) of the starting population (TN) and the two resorted populations (day 1-CD11b−/low and CD11bhigh) after overnight incubation with M-CSF and RANKL. (C) Sorted TN were plated with M-CSF and RANKL for 6 days, fixed and stained for CD11b expression using an anti-mouse CD11b Ab (bottom left) or anti-mouse CD4-PE Ab of the same isotype (bottom right). The cells were stained for TRACP (top).20

Download figure to PowerPoint

When TN cells were cultured overnight only with M-CSF, the same results were observed as when they were cultured with M-CSF and RANKL (data not shown). In addition, cells cultured overnight only with M-CSF upregulated CD11b in the same proportion as cells cultured with M-CSF and RANKL. Sorted day 1-CD11bhigh cells in these cultures were also responsible for the osteoclastogenic activity at days 3 and 4. Sorted day 1-CD11b−/low cells from cultures that were incubated overnight with only M-CSF were also able to differentiate into OCLs by day 5. These results confirm that RANKL is not necessary for the upregulation of CD11b expression in TN cell cultures.

We also used immunofluorescence to determine if mature OCLs expressed CD11b (Fig. 5C). Takahashi et al. (22) previously reported that mature osteoclasts do not express this antigen, and our immunofluorescence data corroborated this finding. Although mature multinucleated OCLs were negative for surface expression of CD11b, the wells did contained mononucleated cells that were strongly positive. These cells likely correspond to TN cells that have upregulated CD11b as described previously (most of these CD11b+ mononuclear cells are not present in the TRACP pictures as the treatment of the slides before the TRACP staining dislodged them).

Further fractionation of the TN progenitor population with CD115 and CD117

Because most of the osteoclastogenic activity of total BM resided in the TN fraction, we used two additional surface markers to further dissect the population: CD115 and CD117 (Fig. 6A). As previously shown, CD115, the receptor for M-CSF, can be used as a marker for osteoclast progenitors, and we found that the early osteoclastogenic activity of the TN fraction was contained in the CD115high population (data not shown). CD117, also known as c-kit, is the receptor for stem cell factor, and it is a consistent marker of early hematopoietic lineage, (23) including early osteoclast progenitors. We therefore sorted the three different populations from the TN CD115high fraction according to their CD117 expression (populations IV, V, and VI) as seen in Fig. 6B, and cultured them in the presence of M-CSF and RANKL for 2 or 4 days before examining their CD11b, CD115, and CD117 expression (Fig. 6C and data not shown). In all three populations, we found that cells proliferated and that a proportion of these cells increased in size (Fig. 6C). We also found that all three populations upregulated CD11b, although with slightly different kinetics. By day 2, 96% of population VI cells were CD11bhigh, whereas 82% and 91% of populations IV and V cells, respectively, had high level expression of CD11b (Fig. 6C). However, by day 4, >95% of cells in all three populations were CD11bhigh (data not shown). In addition, by day 2, most of the cells in all three populations were CD117 (Fig. 6C). By day 4, >94% of the CD115high cells were CD117, whereas a population of CD115low cells was further expanded in all three populations (data not shown).

thumbnail image

Figure FIG. 6.. CD115 and CD117 identify several OCL progenitor populations. (A) The TN fraction (CD11b−/lowCD45RCD3) was arbitrarily divided into six populations (I-VI) according to their CD117 and CD115 expression. (B) Sort reanalysis of TN and populations IV, V, and VI. Sorted populations were plated with M-CSF and RANKL. (C) Two days later, their forward (FSC) and side (SSC) scatters, CD11b, CD45R, CD3, CD115, and CD117 surface expressions were determined by flow cytometry. The numbers on the plots represent the percent of gated events. (D) Sorted populations were also tested for their ability to differentiate into OCLs on days 3–6. The ratio of each population within the TN control fraction was conserved. The results are represented as the mean of three wells ± SE.20

Download figure to PowerPoint

We also sorted populations I, III, IV, V, and VI to determine their individual contribution to the osteoclastogenic activity of the total TN fraction over time. Sorted cells were plated for 3–6 days with M-CSF and RANKL at densities that were proportional to their ratio within the TN fraction. Populations I and III failed to generate OCLs in culture over a period of 6 or 5 days, respectively (data not shown). Furthermore, most of the early osteoclastogenic activity of the TN fraction was included in the TN CD115high fraction and more specifically within the TN CD115highCD117high population (population IV) as shown in Fig. 6D. The number of OCLs decreased by day 5 in population IV, TN, and IV + V + VI wells as OCLs became bigger and then died as shown by an increasing number of TRACP+ cell fragments in the culture. Populations V and VI gave rise to OCLs in vitro, but their contribution to the overall osteoclastogenic activity contained in the TN fraction was limited compared with that of population IV.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. References

In this paper, we found that murine BM contains distinct populations of osteoclast progenitors that can be separated by flow cytometry. We also showed that we could sequentially sort these populations from total BM, culture them in vitro for varying periods of time in the presence of M-CSF and RANKL, and determine which populations contained osteoclastogenic activity.

In our hands, highly purified CD45R+ murine BM cells were incapable of differentiating into OCLs in vitro, which indicates that the potential of CD45R+ populations to generate OCLs, as previously reported, requires contaminating populations. Most previous reports used CD45R+ population with a purity of 97–99%. In this report, we showed that a contaminating population of <1% is required to produce the observed osteoclastogenic activity in CD45R+ murine BM cells when cultured with M-CSF and RANKL (Fig. 1). Therefore, it seems that stringent procedures such as double sorting are required to fully define the osteoclastogenic potential of murine BM fractions in vitro.

Cells from the CD11bhighCD45R and the DN fractions were both shown to contaminate the CD45R+ fraction after a single sort. However, when CD45R+ cells were sorted in the context of CD45R and CD11b (versus CD45R alone), the DN fraction was the predominant contaminant cell population (Fig. 1 and data not shown). The presence of a small number of CD45R BM cells within the CD45+ seems necessary for the osteoclastogenic activity of CD45R+ populations that are purified by a single round of FACS or immunomagnetic beads (Fig. 2 and data not shown).

Interestingly, when we set out to purify double positive CD45R+CD11b+ cells from murine BM, on reanalysis of the sorted population, only single positive CD45R+ or CD11bhigh were recovered at a one to one ratio. These results indicate that CD45R+ cells are able to associate with CD11bhighCD45R cells and form doublets (data not shown). Such interactions might partly explain the presence of CD45R cells in the CD45R+ fraction. An additional possibility is that the CD11bhighCD45R cells that form doublets in the single-sorted CD45R+ fraction are enriched for osteoclast precursor cells.

Our experimental system tested the ability of sorted populations to form osteoclasts in vitro in the presence of M-CSF and RANKL and preferentially accounted for the osteoclastogenic activity of M-CSF-responsive populations. Such cells are expected to express the M-CSF receptor, CD115. M-CSF has been shown to be sufficient for stimulating the proliferation and inducing differentiation of early osteoclast progenitors and probably represents the main physiological mediator. However, it is not essential as other factors can replace its action as shown by the spontaneous recovery of the osteoclast defect in the M-CSF-deficient op/op mouse with time. (24–26) For this reason, any osteoclastogenic activity contained in the CD115 cell fractions might not be accounted for in our assays, because these populations could require other signals to differentiate into an OCL. These could include cells contained in the CD45R+ fraction. It should be noted that all CD11bhigh cells are positive for CD115 (expression varies from low to high; data not shown). In addition, ∼14% of CD45R+ (CD11b) cells express low levels of CD115 (data not shown).

We found that the TN fraction was responsible for most of the osteoclastogenic activity of total murine BM (Fig. 3). However, both CD11bhighCD45RCD3 and TN populations were capable of differentiating into functional mature osteoclasts as assessed by calcitonin receptor expression and the generation of resorption pits on cortical bone slices (Fig. 4). Both populations were indeed able to form resorption pits on bovine cortical bone slices after 7 days of culture; however, the pits observed in the TN samples were bigger and more numerous than the pits from the CD11bhighCD45RCD3 samples. Such difference could be explained if TN cells are far more efficient than CD11bhighCD45RCD3 at differentiating into OCLs in vitro at early time-points (Figs. 2 and 3). It is also possible that the OCLs derived from the TN fraction are more efficient at forming pits than the OCLs derived from the CD11bhighCD45RCD3 fraction. It is also important to remember that the sorted populations likely behave differently on bovine cortical bone slices than on treated plastic ware. For example, the half-life of TN-derived OCLs might be increased on bovine cortical slices. Such a result could explain the statistically significant increase in the mean pit area for each group between days 7 and 15. In addition, as shown in Fig. 5, a “second wave” of OCLs generated from the TN fraction might also contribute to the pits being formed in the TN samples at a later time-point.

We expected the CD11bhigh (CD45RCD3) fraction to contain the majority of the osteoclastogenic activity in murine BM because osteoclasts and macrophages share a common lineage. However, our data shows that this is not the case. Instead, CD11bhighCD45RCD3 cells were far less efficient than TN cells at differentiating into OCLs in vitro (Figs. 2 and 3). This result could be explained by the finding that this fraction contains large numbers of terminally differentiated myeloid cells such as neutrophils and macrophages. Nevertheless, late cultures of CD11bhighCD45RCD3 were able to give rise to significant numbers of OCLs (Fig. 3). However, we cannot rule out the possibility that contamination of CD11bhighCD45RCD3 cells by TN cells was responsible for this result. Our finding that osteoclast progenitors upregulate CD11b quickly on culture in M-CSF (Fig. 5) leads us to hypothesize that the osteoclastogenic activity seen in the CD11bhighCD45RCD3 fraction is caused by a small population of intermediate day 1-CD11bhigh, which are present in steady-state conditions. The limitations of the in vitro assay that we used might explain why the osteoclastogenic activity of this population within the CD11bhigh DN fraction is only apparent in late cultures. Indeed, expansion of osteoclast progenitors and fusion must first occur, and the low density of these cells could be responsible for the delayed response observed. Such a “dilution” effect was observed when plating equivalent numbers of cells in 48-well versus 96-well plates (data not shown).

We observed that a longer culture period was required for equivalent numbers of OCLs to be obtained from the same original number of cells when cultured in 96-well versus 48-well plates. Such a phenomenon could also explain why none of the osteoclast progenitor populations tested in vitro were capable of generating OCLs before day 3 of culture (data not shown).

In contrast to the osteoclastogenic response of CD11bhighCD45RCD3 cells, the TN fraction was very efficient at differentiating into OCLs. However, the numbers of OCLs per well in the culture of these cells dropped rapidly after they formed (Figs. 2 and 3). We believe this is a reflection of two phenomena; the first is that, as they mature, OCLs fuse to form large multinucleated cells (Fig. 2B). However, our assay does not take into consideration the surface area of OCLs; hence, the number of OCLs in the cultures may be decreasing with time without a decrease in the cell “mass” of OCLs. When 5 × 103 DN cells were plated per well, the mean number of nuclei per OCL increased over time and peaked on day 5. However, the mean number of OCLs per well peaked at day 4 and was very much decreased by day 5 (Fig. 2C). The drop in the mean number of OCLs per well at day 5 can be partly explained by the increase in size of the OCLs at day 5. Furthermore, the surface area of the well could be a limiting factor in our assays. Indeed, we found that, at higher densities, OCLs fragment and cannot be visualized. This phenomenon may cause the decrease in OCLs numbers, which we observed at late time-points (fragmentation was reduced when 48-well plates were used instead of 96-well plates; data not shown). Therefore, it was crucial to evaluate the osteoclastogenic activity of our sorted populations from a kinetic prospective.

In such studies, we showed that TN cells, which have high level surface expression of CD11b and CD115 after a 15-h incubation with M-CSF, were much more efficient than cells that did not express these surface proteins. However, day 1-CD11b−/low cells were able to efficiently differentiate into OCLs later (day 5, after they upregulated CD11b and CD115). The TN fraction is a heterogeneous population of cells. We believe that its “second wave” of OCL formation, contained within the day 1-CD11b−/low, is caused by the maturation of earlier progenitors, which require more time to differentiate into osteoclasts. Our results showed that osteoclast progenitors upregulate CD11b transiently on culture with either M-CSF or M-CSF and RANKL. Similarly, we found that most of the rapid osteoclastogenic activity of the TN fraction resided in the CD115high population. The CD115−/low (CD117high) cells were also able to differentiate into OCLs efficiently by day 5 (data not shown). This result suggests that an osteoclast progenitor pool of cells from population III increased expression of CD115 on culture with M-CSF, which is a finding that corroborates data previously published by Arai et al. (13)

Finally, we found that population IV was the most efficient fraction purified from BM for differentiating into OCLs in vitro (Fig. 6). Flow cytometric analysis of this fraction after in vitro culture showed that most of its cells upregulated CD11b and downregulated CD117 (Figs. 6B and 6C) as proposed by Arai et al. (13) Interestingly, low osteoclastogenic activities were recovered from populations V and VI, which also upregulated CD11b and downregulated CD117 in vitro.

Such results might be explained by the ability of M-CSF to act as a multipotent growth factor and regulate the differentiation and maintenance of multiple cells within the myeloid lineage. It is possible that, in the sorted populations V and VI, M-CSF treatment promoted the proliferation of a population of myeloid cells without osteoclastogenic potential, which had enhanced expression of CD11b and decreased expression of CD117. As discussed above, a low frequency of osteoclast progenitors is not easily detected in our in vitro assay. Consistent with this hypothesis is the fact that when fraction VI cells were plated at higher densities, they were more efficient at forming OCLs in vitro (data not shown).

In summary, our results show that FACS can be used to isolate different osteoclast progenitor populations from fresh preparations of murine BM. Using a set of five markers (CD11b, CD45R, CD3, CD115, and CD117), we showed that murine BM cells with the phenotype CD11b−/lowCD45RCD3CD115highCD117+ (population IV), which represents ∼2% of fresh BM preparations, contains the highest in vitro osteoclastogenic activity. We also found that this population upregulates CD11b and downregulates CD117 when it was cultured with M-CSF and RANKL to form mature OCLs.

The ability to isolate highly purified primary murine BM populations provides us with a powerful tool to determine the degree of osteoclastogenic lineage development that a population of osteoclast precursor cells has advanced and to evaluate factors that regulate their ability to form osteoclasts both in vitro and in vivo.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. References

We thank Dr David Rowe for access to his microscopy facility and Dr Ivana Debeljak-Boban for expert assistance. Funding: This work was supported by NIH/NIAID Grant RO1-AI46708 (to HLA) and NIH/NIAMS Grant RO1-AR4871401 (to JAL). This work also used the Flow Cytometry Facility component of the Institutional Core Center for Musculoskeletal Research Grant NIH/NIAMS 5P30-AR046026.

References

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. References
  • 1
    Walker DG 1972 Congenital osteopetrosis in mice cured by parabiotic union with normal siblings. Endocrinology 91:916920.
  • 2
    Walker DG 1975 Control of bone resorption by hematopoietic tissue. The induction and reversal of congenital osteopetrosis in mice through use of bone marrow and splenic transplants. J Exp Med 142:651663.
  • 3
    Kurihara N, Chenu C, Miller M, Civin C, Roodman GD 1990 Identification of committed mononuclear precursors for osteoclast-like cells formed in long term human marrow cultures. Endocrinology 126:27332741.
  • 4
    Udagawa N, Takahashi N, Akatsu T, Tanaka H, Sasaki T, Nishihara T, Koga T, Martin TJ, Suda T 1990 Origin of osteoclasts: Mature monocytes and macrophages are capable of differentiating into osteoclasts under a suitable microenvironment prepared by bone marrow-derived stromal cells. Proc Natl Acad Sci USA 87:72607264.
  • 5
    Bar-Shavit Z, Teitelbaum SL, Reitsma P, Hall A, Pegg LE, Trial J, Kahn AJ 1983 Induction of monocytic differentiation and bone resorption by 1,25-dihydroxyvitamin D3. Proc Natl Acad Sci USA 80:59075911.
  • 6
    Wiktor-Jedrzejczak W, Bartocci A, Ferrante AW Jr, Ahmed-Ansari A, Sell KW, Pollard JW, Stanley ER 1990 Total absence of colony-stimulating factor 1 in the macrophage-deficient osteopetrotic (op/op) mouse. Proc Natl Acad Sci USA 87:48284832.
  • 7
    Tondravi MM, McKercher SR, Anderson K, Erdmann JM, Quiroz M, Maki R, Teitelbaum SL 1997 Osteopetrosis in mice lacking haematopoietic transcription factor PU.1. Nature 386:8184.
  • 8
    Wang ZQ, Ovitt C, Grigoriadis AE, Mohle-Steinlein U, Ruther U, Wagner EF 1992 Bone and haematopoietic defects in mice lacking c-fos. Nature 360:741745.
  • 9
    Grigoriadis AE, Wang ZQ, Cecchini MG, Hofstetter W, Felix R, Fleisch HA, Wagner EF 1994 c-Fos: A key regulator of osteoclast-macrophage lineage determination and bone remodeling. Science 266:443448.
  • 10
    Udagawa N, Takahashi N, Jimi E, Matsuzaki K, Tsurukai T, Itoh K, Nakagawa N, Yasuda H, Goto M, Tsuda E, Higashio K, Gillespie MT, Martin TJ, Suda T 1999 Osteoblasts/stromal cells stimulate osteoclast activation through expression of osteoclast differentiation factor/RANKL but not macrophage colony-stimulating factor: Receptor activator of NF-kappa B ligand. Bone 25:517523.
  • 11
    Suda T, Takahashi N, Udagawa N, Jimi E, Gillespie MT, Martin TJ 1999 Modulation of osteoclast differentiation and function by the new members of the tumor necrosis factor receptor and ligand families. Endocr Rev 20:345357.
  • 12
    Muguruma Y, Lee MY 1998 Isolation and characterization of murine clonogenic osteoclast progenitors by cell surface phenotype analysis. Blood 91:12721279.
  • 13
    Arai F, Miyamoto T, Ohneda O, Inada T, Sudo T, Brasel K, Miyata T, Anderson DM, Suda T 1999 Commitment and differentiation of osteoclast precursor cells by the sequential expression of c-Fms and receptor activator of nuclear factor kappaB (RANK) receptors. J Exp Med 190:17411754.
  • 14
    Miyamoto T, Ohneda O, Arai F, Iwamoto K, Okada S, Takagi K, Anderson DM, Suda T 2001 Bifurcation of osteoclasts and dendritic cells from common progenitors. Blood 98:25442554.
  • 15
    Katavic V, Grcevic D, Lee SK, Kalinowski J, Jastrzebski S, Dougall W, Anderson D, Puddington L, Aguila HL, Lorenzo JA 2003 The surface antigen CD45R identifies a population of estrogen-regulated murine marrow cells that contain osteoclast precursors. Bone 32:581590.
  • 16
    Blin-Wakkach C, Wakkach A, Sexton PM, Rochet N, Carle GF 2004 Hematological defects in the oc/oc mouse, a model of infantile malignant osteopetrosis. Leukemia 18:15051511.
  • 17
    Blin-Wakkach C, Wakkach A, Rochet N, Carle GF 2004 Characterization of a novel bipotent hematopoietic progenitor population in normal and osteopetrotic mice. J Bone Miner Res 19:11371143.
  • 18
    Manabe N, Kawaguchi H, Chikuda H, Miyaura C, Inada M, Nagai R, Nabeshima Y, Nakamura K, Sinclair AM, Scheuermann RH, Kuro-o M 2001 Connection between B lymphocyte and osteoclast differentiation pathways. J Immunol 167:26252631.
  • 19
    Sato T, Shibata T, Ikeda K, Watanabe K 2001 Generation of bone-resorbing osteoclasts from B220+ cells: Its role in accelerated osteoclastogenesis due to estrogen deficiency. J Bone Miner Res 16:22152221.
  • 20
    Springer T, Galfre G, Secher DS, Milstein C 1979 Mac-1: A macrophage differentiation antigen identified by monoclonal antibody. Eur J Immunol 9:301306.
  • 21
    Sanchez-Madrid F, Simon P, Thompson S, Springer TA 1983 Mapping of antigenic and functional epitopes on the alpha- and beta-subunits of two related mouse glycoproteins involved in cell interactions, LFA-1 and Mac-1. J Exp Med 158:586602.
  • 22
    Takahashi N, Udagawa N, Tanaka S, Murakami H, Owan I, Tamura T, Suda T 1994 Postmitotic osteoclast precursors are mononuclear cells which express macrophage-associated phenotypes. Dev Biol 163:212221.
  • 23
    Ikuta K, Weissman IL 1992 Evidence that hematopoietic stem cells express mouse c-kit but do not depend on steel factor for their generation. Proc Natl Acad Sci USA 89:15021506.
  • 24
    Niida S, Kaku M, Amano H, Yoshida H, Kataoka H, Nishikawa S, Tanne K, Maeda N, Kodama H 1999 Vascular endothelial growth factor can substitute for macrophage colony-stimulating factor in the support of osteoclastic bone resorption. J Exp Med 190:293298.
  • 25
    Begg SK, Radley JM, Pollard JW, Chisholm OT, Stanley ER, Bertoncello I 1993 Delayed hematopoietic development in osteopetrotic (op/op) mice. J Exp Med 177:237242.
  • 26
    Kodama I, Niida S, Sanada M, Yoshiko Y, Tsuda M, Maeda N, Ohama K 2004 Estrogen regulates the production of VEGF for osteoclast formation and activity in op/op mice. J Bone Miner Res 19:200206.