Fluorescence-based Staining for Tartrate-resistant Acidic Phosphatase (TRAP) in Osteoclasts Combined with Other Fluorescent Dyes and Protocols

Abstract

Osteoclasts are the only bone-resorbing cells. In addition to other specific properties, osteoclasts are characterized by their expression of tartrate-resistant acidic phosphatase (TRAP), which is usually detected using a histochemical method for light microscopy. Using ELF97 phosphatase substrate, this study describes a new fluorescence-based method for TRAP detection. The fluorescence-based ELF97 TRAP stain not only results in a better resolution of the TRAP-positive granules, because confocal microscopy can be applied for image acquisition and analysis, but it reveals additional and more specific information about osteoclasts because it can be combined with other fluorescence-based methods.

Osteoclasts are the only bone-resorbing cells (Vaananen et al. 2000). They are derived from bone marrow stem cells and develop along the myeloid lineage. In a precursor stage they are found in the peripheral blood among the monocytic population (Salhoub et al. 2000). Through the blood circulation, osteoclast precursors are recruited to internal or external bone surfaces throughout the skeleton where bone resorption is required. Under the influence of differentiation and maturation factors, including macrophage colony-stimulating factor (MCSF) (Fujikawa et al. 2001) and receptor activator of NF-κB ligand (RANK-L) (Feige 2001), from the bone-resorbing environment, the precursors develop into mature osteoclasts with distinct characteristics. First, there are specific morphological qualities, such as multiple nuclei and the basal ruffled border, surrounded by a membrane stabilizing extent ring of actin filaments (Akisaka et al. 2001). Then, there is the specific expression of cathepsin K, which is found in the lysosomal granules. Finally, there is the expression of TRAP, which is regarded as an osteoclast marker, although there are other cells of the macrophage/dendritic cell family that express TRAP under certain conditions (Tsuboi et al. 2003; Walsh et al. 2003). TRAP is mostly characterized by its biochemical property, i.e., through its acid phosphatase function, which cannot be blocked by tartrate.

A histochemical method, commercially available as a kit by Sigma Diagnostics (St Louis, MO), is widely used and has proved to be the standard method for specific detection of osteoclasts. This histochemical TRAP staining results in a color precipitate that can be easily detected using light microscopy. Although this classical TRAP stain is easy to perform, it has the disadvantage that it is rather difficult to combine with other specific stains such as those used in immunohistological methods. Because of this problem, a fluorescence-based TRAP staining method was developed and tested that could be combined with other fluorescence-based stains, including a nuclear DNA stain, an immunofluorescence stain, and an actin stain using fluorescence-labeled phalloidin. For that purpose, ELF97 phosphatase subtrate (Molecular Probes; Eugene, OR) was tested because it has been used successfully for the detection of alkaline phosphatase activity with fluorescence microscopy (Cox and Singer 1999).

First, the ELF97 TRAP protocol was established and compared with the classical TRAP method. Human osteoclast-like cells were generated in vitro from adherent peripheral blood mononuclear cells (PBMCs). Briefly, PBMCs were isolated from buffy coats [Australian Red Cross Blood Service (ARCBS); Perth, WA, Australia] through Ficoll-gradient centrifugation. The PBMCs were cultured (37C, humidified, 5% CO₂) in 25-cm² culture flasks (Sarstedt; Nuernbrecht, Germany) in RPMI 1640 medium (GIBCO; Auckland, New Zealand), supplemented with antibiotics (GIBCO) and 10% human A serum (ARCBS). After 1 hr in culture, the non-adherent PBMCs were discarded and the adherent cells washed twice with PBS before being cultured in osteoclast differentiation medium composed of RPMI 1640, antibiotics, 5% human A serum, and cytokines (recombinant human M-CSF, 10 ng/ml; recombinant human RANK-L, 10 ng/ml; R&D, Minneapolis, MN). Freshly isolated adherent PBMCs and cells, cultured for 1–30 days were fixed with 1–3% paraformaldehyde in PBS for 30 min to 24 hr and subsequently processed for cytospins and TRAP staining. After 3–30 days, the cultured cells were stained using the classical TRAP protocol to confirm that they were TRAP-positive (Figure 1A). The kit from Sigma Diagnostics was used according to the recommended protocol and it produced positive TRAP staining in all cultures (blood samples from more than 20 individuals were tested). The results were analyzed using a Leica DM RBE microscope and recorded by a LEAF digital camera (Leica; Wetzlar, Germany). In parallel, using the same cells, the ELF97 protocol was established. The same buffer used for the classical TRAP staining was used for the ELF97 protocol (110 mM acetate buffer, pH 5.2, 1.1 mM sodium nitrite, 7.4 mM tartrate). Different dilutions and incubation times were tested, indicating that a 200-μM concentration of ELF97 and a 15-min incubation time were the most appropriate conditions. These conditions were used for all future fluorescence TRAP stains. It should be noted that, similar to the classical TRAP staining, no cell membrane permeabilization step is needed for the ELF97 TRAP staining. The fluorescence signal was analyzed and documented by conventional fluorescence microscopy or by confocal microscopy (Bio-Rad MRC 1024, Coherent Enterprise argon ion 250-mW multi-line UV, American Laser 100-mW argon ion multiline laser, Melles-Griot 0.5-mW green helium-neon laser; Bio-Rad, Hercules, CA). By omitting the ELF97 substrate (data not shown) or by staining freshly isolated TRAP-negative cells (Figure 1B), no fluorescence signal for TRAP was detected. However, when cultured over 3 days or longer in osteoclast differentiation medium, the adherent PBMCs became TRAP-positive. This indicates that the cells had differentiated towards osteoclast-like cells under these culture conditions. Comparing the resulting morphology of the classical and the ELF97 TRAP stains, ELF97 gave much better results. The fluorescent precipitate was clearly restricted to the area of the granules, whereas the classical TRAP staining resulted in a rather diffuse granular pattern in which the single granules were difficult to separate from each other. The accuracy of the ELF97 TRAP stain was enhanced and clearer when confocal microscopy was used.

Figure 1

TRAP staining of human in vitro generated osteoclast-like cells (A-D) and of rat periosteum (E,F) using the classical TRAP protocol (A,E) or the fluorescence-based ELF97 TRAP protocol (B-D,F). The light microscopy images were acquired with a digital camera, the fluorescence-based images were acquired with a confocal microscope. (A) Human adherent PBMCs were cultured in osteoclast differentiation medium for 3 weeks before staining with the classical TRAP stain (dark red). (B) Freshly isolated adherent human PBMCs were stained using the ELF97 TRAP protocol. Because there is no TRAP expression in monocytes, there is no green fluorescence signal. Nuclei were counterstained with Syto25 (red). (C) After 3-day culture in osteoclast differentiation medium, adherent human PBMCs were stained using the ELF97 TRAP protocol. Nuclei were counterstained with Syto25 (red). Note the distinct TRAP-positive green granules. (D) After 7-day culture in osteoclast differentiation medium, adherent human PBMCs were stained using the ELF97 TRAP protocol. Nuclei were counterstained with DAPI (blue) and the actin filaments with AlexaFluor546 phalloidin (red). Note the distinct green TRAP-positive granules. (E) Rat tibial periosteum was stained using the classical TRAP protocol. TRAP-positive cells corresponding to osteoclasts are stained in dark purple. (F) Rat tibial periosteum was stained using the ELF97 TRAP protocol. MHC class II molecules (red) were stained using an immunofluorescence protocol and a specific mouse monoclonal antibody. Nuclei were counterstained with DAPI (blue). Note the distinct TRAP-positive green granules of a multinucleated cell and mono- or binucleated cells corresponding to osteoclasts. Bars = 20 μm.

Second, the ELF97 TRAP method was combined with other staining protocols. Not only fluorescence detection of surface molecules but also detection of intracellular molecules were of interest because for the latter cell membrane permeabilization is usually needed. First, a fluorescence nuclear counterstain with DAPI (1 μg/ml in PBS, 4′,6-diamidine-2′-phenylindole dihydrochloride; Roche Diagnostics, Mannheim, Germany) and Syto25 (5 nM in PBS; Molecular Probes) was tested after the ELF97 TRAP staining. No membrane permeabilization was required to get a brilliant fluorescence signal for the nuclear DNA (Figures 1B-1E). After the ELF97 TRAP staining, immunofluorescence stains of surface markers were tested, including CD45 and MHC surface molecules, which worked well and gave excellent results (data not shown). Finally, fluorescence stains for cytoskeleton elements were tested, for which membrane permeabilization (60 sec, 0.1% Triton X-100 in PBS) is needed after the ELF97 stain. The basal actin ring was stained with AlexaFluor546-labeled phalloidin (0.2 U/ml in PBS; Molecular Probes) (Figure 1D). The cellular tubulin structure was stained using an immunofluorescence protocol and a mouse anti-tubulin monoclonal antibody (data not shown). In both cases of fluorescence-based cytoskeleton staining, the results were excellent, without alteration of the ELF97 TRAP fluorescence signal.

It was of interest whether the ELF97 TRAP stain works in tissue samples and in other species. Therefore, the ELF97 TRAP protocol was tested in freshly excised rat tibial periosteum. Ten rats (Lewis; Guidelines of the Research Ethics and Animal Care Office were followed) were anesthetized before sacrifice. The periosteum from the anterior surface of both tibias was stripped from the bone and fixed in 3% paraformaldehyde/PBS. The tissue was stained as a whole tissue sample either with the classical TRAP stain (Figure 1E) or using the ELF97 TRAP protocol (Figure 1F). Omitting the phosphatase substrate resulted in negative TRAP staining (data not shown). Using either protocol resulted in clear staining of osteoclasts in the cell-rich internal layer of the rat tibial periosteum. The ELF97 TRAP stain was also combined with diverse fluorescence staining protocols, including a nuclear stain with DAPI and an immunofluorescence protocol for staining of MHC class II using OX6, a monoclonal mouse anti-rat MHC class II antibody (Serotec; Oxford, UK), followed by a polyclonal biotin-labeled sheep anti-mouse antibody (Amersham Biosciences; Piscataway, NJ) and AlexaFluor546-labeled streptavidin (Molecular Probes). This combined staining resulted in clear and specific staining of polynucleated cells for ELF97 TRAP and of mononucleated cells for MHC class II in the cell-rich internal layer of the rat periosteum (Figure 1F).

Summarizing, one can conclude that the fluorescence-based ELF97 TRAP stain provides a comparable if not even better option than the classical histochemical TRAP stain. Using the ELF97 TRAP protocol, one can take advantage of confocal microscopy to obtain high-resolution images. In addition, the ELF97 TRAP stain can probably be combined with most fluorescence stains, including immunofluorescence protocols. Elf97 TRAP stain not only works with single cells and cultured cells for cytological analysis but it also gives excellent results with tissues. Furthermore, it can probably be applied for different species, including humans and rats. Finally, fluorescence-based TRAP stain with ELF97 opens up new histological possibilities in the area of bone and osteoclast research.