Terminal Differentiation of Osteoblasts to Osteocytes Is Accompanied by Dramatic Changes in the Distribution of Actin-Binding Proteins

Authors

  • Hiroshi Kamioka,

    1. Department of Orthodontics and Dentofacial Orthopedics, Graduate School of Medicine and Dentistry, Okayama University, Okayama, Japan
    Search for more papers by this author
  • Yasuyo Sugawara,

    1. Department of Orthodontics and Dentofacial Orthopedics, Graduate School of Medicine and Dentistry, Okayama University, Okayama, Japan
    Search for more papers by this author
  • Tadashi Honjo,

    1. Department of Orthodontics and Dentofacial Orthopedics, Graduate School of Medicine and Dentistry, Okayama University, Okayama, Japan
    Search for more papers by this author
  • Takashi Yamashiro,

    1. Department of Orthodontics and Dentofacial Orthopedics, Graduate School of Medicine and Dentistry, Okayama University, Okayama, Japan
    Search for more papers by this author
  • Teruko Takano-Yamamoto

    Corresponding author
    1. Department of Orthodontics and Dentofacial Orthopedics, Graduate School of Medicine and Dentistry, Okayama University, Okayama, Japan
    • Address reprint requests to: Teruko Takano-Yamamoto, PhD, DDS, 2–5–1 Shikata, Okayama 700–8525, Japan
    Search for more papers by this author

  • The authors have no conflict of interest.

Abstract

Immunofluorescence staining of actin-binding proteins in osteoblasts and osteocytes was performed. α-Actinin, myosin, and tropomyosin showed similar organization in both osteoblastic stress fibers and osteocyte processes. However, fimbrin, villin, filamin, and spectrin showed dramatic differences in distribution between osteoblasts and osteocytes. This study suggested that terminal differentiation of osteoblasts to osteocytes is accompanied by highly dramatic changes in the distribution of actin-binding proteins.

Introduction: We previously reported that osteocyte shape is dependent on actin filaments. To analyze the terminal differentiation from osteoblasts to osteocytes, we investigated the actin-binding proteins, which are the control elements in the dynamic organization of the actin cytoskeleton.

Materials and Methods: We used primary chick osteocytes and osteoblasts, the phenotypes of which were confirmed by use of OB7.3, a chick osteocyte-specific monoclonal antibody and by detection of alkaline phosphatase activity, respectively. Immunofluorescence staining was performed for visualizing actin-binding proteins. Furthermore, we applied shear stress at 12 dyns/cm2 to the cells and compared the changes in fimbrin distribution.

Results: Immunofluorescence staining of fimbrin and α-actinin showed their presence in the processes of osteocytes, with especially strong signals of fimbrin at the sites of divarication of the processes. Anti-villin was reactive with the osteocyte cytoplasm but not with the processes. Interestingly, anti-villin immunoreactivity was much stronger in osteocytes than in osteoblasts. Filamin was localized along the stress fibers of osteoblasts but was seen only in those in the proximal base of osteocyte processes. Myosin and tropomyosin were found to have a similar pattern in both stress fibers of osteoblasts and osteocyte processes. The difference in the distribution of anti-spectrin staining was highly dramatic. Osteoblasts immunostained with anti-spectrin showed punctate signals on their cytoplasmic membranes, whereas anti-spectrin in osteocytes detected a filamentous organization; and the spectrin was totally colocalized with actin from the distal portion of the cytoplasmic processes to the cell center. In osteoblasts, shear stress induced recruitment of fimbrin to the end of stress fibers. However, fimbrin in the osteocyte processes did not change its localization.

Conclusion: We found that terminal differentiation of osteoblasts to osteocytes was accompanied by highly dramatic changes in the distribution of actin-binding proteins, changes of which may affect cellular function.

INTRODUCTION

THERE IS A GROWING consensus that the network of interconnected osteocytes provides a cellular system that senses changes in bone loading and subsequently activates osteoclasts and osteoblasts to resorb or produce bone matrix.(1, 2) The formation of the network results from the morphological changes from the cuboidal shape of osteoblasts to the stellate shape of the osteocyte, with its multiple, slender cytoplasmic processes.(3–5) Changes in cell shape are the results of regulated changes in the assembly and disassembly kinetics of the cytoskeleton. Furthermore, it is generally agreed that the cytoskeleton is somehow involved in the mechanism by which individual cells sense mechanical signals.(6-8) We previously reported that actin filaments were crucial for the maintenance of osteocyte processes and pointed out that two actin-binding proteins, fimbrin and α-actinin, were present in the osteocyte processes.(9) In general, actin-binding proteins are the control element for the dynamic reorganization of the actin cytoskeleton. Therefore, understanding how the actin-binding proteins are organized in osteoblasts and osteocytes will provide critical information as to how the osteoblast differentiates into the osteocyte as well as about the mechanosensory capabilities of osteocytes.

The success of this study stemmed from the successful isolation and maintenance of osteocytes in culture, which allowed us to conduct extensive analysis of the osteocyte cytoskeleton. Without the problematic autofluorescence from bone matrix, we used immunofluorescence microscopy to examine the relative distribution of actin filaments and actin-binding proteins. The actin-binding proteins examined were actin-bundling proteins including fimbrin, α-actinin, and villin(10); the gel-forming protein filamin(11); and the plasma membrane-anchored cytoskeletal protein spectrin.(12, 13) Furthermore, myosin, which has motor activity on actin filaments,(14, 15) and tropomyosin, which is localized along the groove of the F-actin filament and regulates the actin-myosin interaction,(16) were also analyzed. α-Actinin, myosin, and tropomyosin showed similar organization in both osteoblastic stress fiber and osteocyte processes. However, other actin-binding proteins analyzed in this study showed dramatic differences in distribution between osteoblasts and osteocytes.

MATERIALS AND METHODS

Preparation of coated coverslips

Both human fibronectin and poly-D-lysine, purchased from Beckton Dickinson Labware, were dissolved with double-distilled water and diluted to a final concentration of 20 mg/ml. For double coating with poly-D-lysine and fibronectin, poly-D-lysine was applied to glass coverslips, and the coverslips were incubated for 1 h at 37°C, washed several times with double-distilled water, and air dried. Thereafter, the same processes were repeated for fibronectin.

Isolation of chick osteoblasts and osteocytes

Osteocytes and osteoblasts were isolated from 16-day-old embryonic chicken calvaria, with a modification of the technique described earlier.(9) After the periosteum had been stripped off, the calvaria were trimmed to remove the noncalcified and marrow-rich areas, as previously determined by histological analysis. They were dissected into small pieces and treated for 30 minutes with 1 mg/ml collagenase type I in a bone isolation buffer (25 mM HEPES [pH 7.4], 10 mM NaHCO3, 70 mM NaCl, 60 mM sorbitol, 3 mM K2HPO4, 1 mM CaCl2, 30 mM KCl, 1 mg/ml bovine serum albumin [BSA], 5 mg/ml glucose, and 300 nM α-tosyl-L-lysyl chloromethane); incubated in 5 mM EDTA in PBS containing 0.1% BSA (Sigma) for 30 minutes; and again with collagenase for 40 minutes at 37°C. Cells released after the final collagenase digestion were collected and seeded onto culture dishes. After 1 h, the cells were washed with PBS and incubated 15–18 h in fresh medium with 2% fetal bovine serum (FBS). This method yields a population of cells that contain about 90% osteocytes, as determined by subsequent immunocytochemistry with OB7.3, a chick osteocyte-specific antibody.(17) After fluorescence micrographs of the OB7.3-labeled cells had been acquired, the coverslips were rinsed; the samples were incubated for 20 minutes with a mixture of 50 mg/ml Naphtol ASMX (Sigma), 0.5% N,N-dimethylformamide, and 0.6 mg/ml of fast red violet LB salt in 0.1 M Tris-HCL, pH 8.5, at room temperature. Osteoblasts were identified as alkaline phosphatase activity-positive cells. More than 80% of the nondentritic cells were regarded as osteoblasts.

All procedure was performed in accordance with National Institutes of Health (NIH) guidelines for the use of animals in research.

Fluorescence microscopy

Pan-anti-fimbrin, a rabbit polyclonal antibody against chicken intestinal fimbrin (R163.3), was a gift from Dr P Matsudaira (Whitehead Institute, MIT, Cambridge, MA, USA). Commercially purchased primary antibodies included anti-actin mouse IgG (Chemicon), anti-vinculin mouse IgG (Sigma), anti-α-actinin mouse IgM (Sigma), anti-myosin light-chain mouse IgM (Sigma), anti-tropomyosin rabbit polyclonal (LSL), anti-filamin mouse IgG (Sigma), anti-villin mouse IgG (Chemicon), and anti-spectrin mouse IgM (Sigma).

All secondary antibodies were commercially purchased and included anti-mouse IgG-Alexa 594 conjugate (Molecular Probes), anti-mouse IgM-fluorescein (FITC) conjugate (Kirkegaard & Perry Laboratories), and anti-rabbit IgG-Alexa 488 conjugate (Molecular Probes).

After 15–18 h in culture, isolated osteocytes were rinsed with PHEM (60 mM piperazine-N,N′-bis[2-ethanesulfonic acid], 10 mM ethylene glycol-bis [2-aminoethylether]-N,N,N′,N′-tetraacetic acid, 2 mM magnesium chloride, pH 6.9), and permeabilized for 5 minutes in 1% paraformaldehyde in PHEM containing 0.15% Triton X-100 at 37°C. These specimens were fixed with 3% paraformaldehyde in PHEM for 10 minutes. To prevent nonspecific interactions, we immersed the specimens in a blocking solution of 1% BSA in PBS for 2 h. Thereafter they were incubated with a 1:200 dilution of primary antibody for 1 h at 37°C and rinsed several times with PBS. After reaction with the proper secondary antibody for 30 minutes at 37°C, the specimens were again rinsed and stained with a 1:200 dilution of Alexa 594-Phalloidin (Molecular Probes) in PHEM for 10 minutes. For double immunolabeling with anti-actin and anti-spectrin, the above protocol was repeated. After final rinsing with PBS, the specimens were mounted with fluorescence-mounting medium containing 1 mg/ml p-phenylenediamine dihydrochloride and viewed immediately.

Fluorescent images

Images of isolated osteoblasts and osteocytes were obtained with an inverted Olympus IX 70 microscope, using a 100× (NA1.4) objective (Olympus). Images of the cells were recorded with a 1024 × 1024 Hamamatsu CCD camera, controlled by Aquacosmos software (Hamamatsu photonics). Images were digitally processed by using Adobe Photoshop 5.5 (Adobe Systems Inc.).

Induction of shear stress in flow chamber

Cells were subjected to fluid flow in flow chamber connected to flow loop apparatus as described by Frangos.(18) Cells were exposed for 30 minutes to a defined shear stress of 12 dynes/cm2, as monitored by a flowmeter (SWF-5; Zepeda Co., Seattle, WA, USA). This level of shear is in the predicted physiological range,(19, 20) and total cell perfusate volume used in all experiments was approximately 45 ml. Control cells were maintained in same chamber and given fresh media in a manner identical to the flow-treated cells.

RESULTS

Immunofluorescence staining with anti-fimbrin, anti-α-actinin, anti-villin, anti-filamin, anti-spectrin, anti-myosin (light chain 20K), and anti-tropomyosin showed a unique distribution of actin-binding proteins in both the osteocyte cell body and its processes compared with that for stress fibers in osteoblasts.

Fimbrin is abundant in osteocyte processes, especially at the site of divarication

Double-label fluorescence staining using Alexa 594-phalloidin for actin filament staining and anti-pan-fimbrin showed a diffuse distribution of fimbrin in the osteoblastic cytoplasm (Figs. 1A and 1B), whereas both actin filaments and fimbrin were especially prominent in the processes of osteocytes (Figs. 1C and 1D), as reported previously.(9) In this report, we focused on the distribution of fimbrin in the osteocyte processes. Interestingly, strong signals of fimbrin were more evident at the site of divarication of the processes (arrows in Fig. 1E). Because fimbrin is present in adhesion complexes,(21) we used antibody against vinculin, which is a prominent protein in focal adhesions,(22) to confirm whether the expression of fimbrin was relate to the site of adhesion complexes. Comparison of the distribution of vinculin and fimbrin in osteocyte processes revealed a different appearance. Anti-vinculin showed short linear fragments along the processes, and the expression was not restricted to the sites of divarication (Fig. 1F). In contrast, strong signals indicating fimbrin were mostly found at the sites of divarication (Fig. 1G).

Figure FIG. 1..

Distribution of actin filaments and fimbrin in (A and B) an osteoblast and (C and D) an osteocyte was examined by double-label fluorescence staining with (A and C) Alexa 594-phalloidin and (B and D) anti-pan-fimbrin. Anti-pan-fimbrin reactivity in the osteoblast appears diffuse in the cytoplasm, whereas fimbrin in the osteocyte is prominent in the processes. Especially, a strong signal of anti-fimbrin is evident at the site of divarication (E; merged image of inset in C and D). Comparison of the distribution of vinculin and fimbrin was performed by double-labeling with anti-vinculin (green, F) and Alexa 594-phalloidin (red, F) and with anti-pan-fimbrin (green, G) and Alexa 594-phalloidin (red, G). Fimbrin is abundant at the site of divarication, but vinculin is not. Bars: 10 μm.

α-Actinin is in osteocyte processes: villin is abundant only in the cell body

The distribution of other actin-bundling proteins such as α-actinin and villin in osteoblasts and osteocytes was analyzed by immunofluorescence imaging with their antibodies. In the osteoblasts, linear and thick stress fibers were evident (Fig. 2A). Anti-α-actinin showed a pattern of linear fragments and dots, indicating the localization of the protein along and at the end of the stress fibers (Fig. 2B). In the osteocytes, α-actinin-positive sites appeared as serial dots (Fig. 2D). The distance between the dots in osteocyte was less than that in osteoblasts (inset in Figs. 2B and 2D).

Figure FIG. 2..

Distribution of actin filament and α-actinin in (A and B) an osteoblast and (C and D) an osteocyte was examined by the double-label fluorescence staining with (A and C) Alexa 594-phalloidin and (B and D) anti-α-actinin. In the osteoblast, anti-α-actinin showed periodic dots along the stress fibers and at their ends (inset in B). In the osteocyte, the anti-α-actinin appeared as periodic dots in the processes (inset in D). Bars, 10 μm.

In the osteoblasts, anti-villin staining showed punctate dots in the cytoplasm, which were seen especially in the perinuclear region (Fig. 3B). The merged image showed that the distribution of villin was not identical to that of the stress fiber (Fig. 3C). Anti-villin staining of osteocytes resulted in strong signals in the cell body (Fig. 3E). Comparison of the distribution of villin in osteocytes and an adjacent osteoblast (arrow in Fig. 3E) revealed a higher expression of villin in osteocytes. Despite high expression of villin in the cell body, villin was not seen in the osteocyte processes (Fig. 3F).

Figure FIG. 3..

Distribution of actin filament and villin in (A-C) an osteoblast and (D-F) an osteocyte was examined by the double-label fluorescence staining with (A and D) Alexa 594-phalloidin and (B and E) anti-villin. (C and F) Merged images of double-label fluorescence. In the osteoblast, anti-villin appears as punctate dots in the cytoplasm. In the osteocyte, strong signals of anti-villin are seen in the cytoplasm, but no signal is evident in the processes (E). Comparison of the villin signal in osteocytes and an adjacent osteoblast (arrow in E) reveals an abundance of villin in osteocytes but not in osteoblasts (F). Bars, 10 μm.

Filamin shows a unique distribution at the terminal web of the processes but is not detected along the length of processes

Filamin was diffusely distributed in the osteoblastic cytoplasm, and strong signals were observed in the stress fibers (Figs. 4A and 4B). The pattern of the distribution was similar to that of anti-α-actinin and showed periodic dots (Fig. 4C). However, the accumulation at the end of stress fibers, which was seen with anti-α-actinin, was not observed. In the osteocytes, anti-filamin showed a unique localization of the protein. Anti-filamin staining appeared in the cell body but not in the processes (Fig. 4E). Strong linear signals were observed only in the proximal base (terminal web) of the osteocyte processes (Fig. 4F).

Figure FIG. 4..

Distribution of actin filament and filamin in (A-C) an osteoblast and (D-F) an osteocyte was examined by the double-label fluorescence of (A and D) Alexa 594-phalloidin and (B and E) anti-filamin. (C) Filamin appears as dots along the stress fibers. In the osteocyte, filamin is observed only at the proximal base of the processes (F: merged image of inset in D and E). Bars, 10 μm.

Strong perinuclear concentration of spectrin in osteocyte is evident, with some filamentous bundles continuing into the cellular processes

In the osteoblasts, anti-spectrin decorated one side of the perinuclear region and also bound to cytoplasmic membranes (inset in Fig. 5B). Co-localization of stress fibers and spectrin was not observed (inset in Fig. 5C). In the osteocytes, a strong perinuclear concentration of spectrin was also seen (inset in Fig. 5E). However, spectrin showed a dense distribution along filamentous structures, which was not seen in osteoblasts. Some filamentous bundles of spectrin continued throughout the entire length of the processes (Fig. 5F). To confirm the interaction of spectrin with perinuclear actin filaments in osteocytes, we used anti-actin antibody. Double immunostaining with anti-actin and anti-spectrin confirmed that spectrin was colocalized with actin filaments in both the perinuclear region and osteocyte processes (compare Figs. 5G and 5H).

Figure FIG. 5..

Distribution of actin filament and spectrin in (A-C) an osteoblast and (D-F) an osteocyte was examined by the double-label fluorescence of (A and D) Alexa 594-phalloidin and (B and E) anti-spectrin. (C and F) Merged images of double-label fluorescence. Anti-spectrin in osteoblast was apparent in one side of perinuclear region and appeared as punctate dots (inset in B). Co-localization of stress fibers and spectrin was not seen (inset in C). In the osteocyte, filamentous bundles of spectrin continued from perinuclear region throughout the entire length of the processes (inset in F). Double-label immunofluorescence obtained with (G) anti-actin and (H) anti-spectrin shows spectrin is colocalized with actin filaments in perinuclear region and in their processes (I). Bars, 10 μm.

Myosin and tropomyosin are present in the osteocyte processes

Myosin is the major contractile component of the thick filaments of all muscle cells. Myosin consists of two heavy chains and two pairs of light chains. We used anti-myosin (20-K light chain) for staining myosin. In the osteoblasts, the myosin appeared as short, linear fragments with a periodic distribution (Fig. 6B), consistent with their location in stress fibers (Figs. 6A and 6B). In the osteocyte, the periodic distribution of anti-myosin was evident in both the cell body and the processes (Fig. 6D). Signals of the anti-myosin were clearly seen in the processes, indicating periodic short fragments (inset in Fig. 6D). Anti-tropomyosin staining in an osteoblasts revealed immunoreactive sites along the stress fibers (Figs. 7A and 7B), and the density of the serial dots was greater than that seen with anti-α-actinin or anti-myosin light chain. Tropomyosin in osteocytes appeared as immunoreactive sites throughout the cytoplasm (Fig. 7D). In some osteocytic processes, signals of anti-tropomyosin were observed and appeared as linear fragments (Fig. 7D, inset).

Figure FIG. 6..

Distribution of actin filaments and myosin in (A and B) an osteoblast and (C and D) an osteocyte examined by the double-label fluorescence staining with (A and C) Alexa 594-phalloidin and (B and D) anti-myosin (20-K light chain). In the osteoblast, myosin appears as periodic short linear fragments along the stress fibers (inset in B). In the osteocyte, myosin appears as periodic short linear fragments in the processes (inset in D). Bars, 10 μm.

Figure FIG. 7..

Distribution of actin filaments and tropomyosin in (A and B) an osteoblast and (C and D) an osteocyte revealed by double-label fluorescence staining with (A and C) Alexa 594-phalloidin and (B and D) anti-tropomyosin. In the osteoblast, anti-tropomyosin shows a high density of dots along the stress fibers (inset in B). In the osteocyte, anti-tropomyosin appears as linear fragments in the processes (inset in D). Bars, 10 μm.

Shear stress induces recruitment of fimbrin at the end of stress fibers in osteoblasts: fimbrin in the osteocyte processes did not change its localization

In the control osteoblasts (no flow), the slender actin bundles were randomly arranged (red staining in Fig. 8A). Additionally, double-labeling of fimbrin in the same sample showed the fimbrin to be diffusely distributed in the cytoplasm (green staining in Fig. 8A). During the course of fluid shear, osteoblasts slowly retracted, as observed by phase-contrast microscopy (data not shown). After 30 minutes of flow, linear and thick stress fibers were evident (Fig. 8B). Interestingly, double-labeling of fimbrin in the same sample showed that fimbrin was dramatically reorganized at the end of the stress fibers (Fig. 8B). Some of fimbrin localized along the thicker stress fibers (inset in Fig. 8B). However, compared with osteoblasts, osteocytes showed little change in their shape during 30 minutes of flow as observed by phase-contrast microscopy (data not shown), nor did fimbrin change its localization after subjection to shear stress in the osteocyte processes (compare Figs. 8C and 8D).

Figure FIG. 8..

Distribution of actin filaments and fimbrin in (A and B) osteoblasts and (C and D) osteocytes before and after shear stress. Double-label fluorescence staining with Alexa 594-phalloidin (red) and anti-fimbrin (green) showed that fimbrin in a control osteoblast was diffusely distributed in cytoplasm, whereas that in a control osteocyte was localized in processes. After 30 minutes of flow at 12 dynes/cm2, flow-treated osteoblasts showed accumulation of fimbrin in the end of stress fiber. However, fimbrin in osteocyte processes did not change its localization. Bars, 10 μm.

DISCUSSION

Our previous study, using immunofluorescence and SEM, indicated that the actin filament bundles of osteocytes in culture nearly filled the entire length of the cell processes as well as the perinuclear region as a dense meshwork of filaments. Furthermore, depolymerization of actin filaments by use of latrunclin B or cytochalasin D demonstrated actin filaments to be a necessary component for a stable framework to maintain cell shape of osteocyte.(9) To analyze the reorganization of actin filaments with the morphological changes from osteoblasts to osteocytes, we examined the expression and localization of various actin-binding proteins. Table 1 summarizes our results.

Table Table 1.. Comparison of the Distribution of Actin-Binding Proteins in Osteoblasts and Osteocytes
original image

Actin filament cross-linking protein: fimbrin, α-actinin, villin, and filamin

Fimbrin and α-actinin are widely distributed bundling proteins.(23, 24) Fimbrin is abundant in the parallel filament bundles at the leading edge of cells, particularly in microspikes of filopodia, and in focal adhesions.(21, 25, 26) α-Actinin is thought to be responsible for the relatively loose cross-linking of actin filaments in these contractile bundles and also to help to form focal adhesions.(27, 28) We reconfirmed these actin-bundling proteins to be present along the entire length of osteocyte processes. Moreover, we observed strong signals of fimbrin at sites of divarication but not at those of focal adhesion. This result suggests that high expression of fimbrin might be related to branch formation at the site of divarication rather than for attachment of osteocytes to their substrate. In neurons, focal accumulation of F-actin occurs at axon branch points, and furthermore, F-actin-microtubule interactions are essential for axon branch formation.(29) Therefore, fimbrin at a site of divarication may serve to bundle the actin filaments for branching. However, we previously observed that microtubules did not fill the entire length of osteocyte processes and were absent at the distal portion of the processes.(9) These results suggest that branch formation in osteocyte process may proceed by a mechanism different from involved in axon branch formation.

Villin serves a central function in the assembly of the actin bundles in brush border of intestinal microvilli.(30) Although villin helps to bundle actin filaments into such cytoplasmic projections, we did not detect the villin molecule in the osteocyte processes. Thus, the osteocyte processes may not be dependent on villin for actin filament bundling. Moreover, osteocyte processes are longer than most of the cytoplasmic projection; unlike microvilli, they have the ability to branch. Thus, our findings strongly suggest that differences in localization of actin-bundling proteins in cytoplasmic projections are related to the morphology of the cytoplasmic process formed. Interestingly, the expression of villin in osteocytes was much stronger than that in osteoblasts. It is reported that microinjection of villin into cells caused disruption of stress fibers and induced long-lasting changes in cel morphology.(31) Thus, the highly expression of villin in osteocytes might be related the regulation of morphological changes of osteoblasts and osteocytes. In addition, villin, like fimbrin, may be potentially useful as an intracellular marker for distinguishing osteocytes from osteoblasts as well as play a role in osteocyte differentiation.

Filamin is a homodimer that organizes filamentous actin into both gel-forming networks and actin bundles.(12) Interestingly, the type of actin filament organization depends on the filamin-to-actin ratio.(11) The formation of the actin filament bundle is promoted when the molar ratio of filamin to actin is high, whereas the formation of orthogonal actin networks is induced when this molar ratio is low. In our study, we observed filamin in osteoblasts to be distributed periodically along entire length of the stress fibers, indicating an abundance of the protein in the stress fiber. However, filamin in osteocytes was observed only at the base of the their processes. Filamin at the base of the processes might function to concentrate the cytoplasmic actin web into the processes. These differences also affirm the unique organization of actin filaments in the osteocyte processes and distinguish them from stress fibers in osteoblasts.

Spectrin: a membrane-associated, actin-binding protein

Spectrin molecules are tetramers made up of two α-and two β-spectrin chains, and they form a dense meshwork of proteins in close association with the plasma membrane.(32) Spectrin binds to short actin filaments (37 nm) on the one hand and is linked to the membrane through interactions with proteins such as ankyrin or band 4.1 protein on the other.(33, 34) Punctate anti-spectrin-stained sites were found in all cytoplasmic membranes of osteoblasts, as expected. Such positive sites were also seen in the perinuclear region. It is reported that spectrin isoforms also become localized in the Golgi apparatus, where spectrin arrays may form vesicular coats and mediate vesicle sorting and trafficking.(35, 36) Thus, the signal seen in the perinuclear region of osteoblasts might be related to the Golgi apparatus. However, anti-spectrin staining in osteocytes revealed a filamentous organization of spectrin that was totally colocalized with actin and extended from the cell center to the distal portion of the cytoplasmic processes. We previously proposed that the actin bundles spanning the cell body and osteocyte processes could be an important part of the osteocyte's mechanosensory capability, because it would allow for the direct transmission of physical stimuli from the periphery to the nuclear region, possibly influencing activities within the nucleus. So the reorganization of spectrin in the osteocyte might to be related to the process of acquiring mechanosensory capability.

Myosin and tropomyosin

A major role of myosin is motor activity on actin filaments in contractile bundles.(37, 38) Tropomyosin is localized along each of the two grooves of the actin filament(39) and is known to propagate actin interaction with myosin. Thus, we analyzed the localization of both myosin and tropomyosin in stress fibers and osteocyte processes. Both myosin and tropomyosin were present along the entire length of the osteoblastic stress fibers and osteocyte processes. The distribution pattern of myosin in both locations, that is, periodic short, linear fragments, was similar. Furthermore, we confirmed the presence of α-actinin, which form the looser packing of actin bundles and allow myosin molecules to enter, in the osteocyte processes. In general, the presence of actin and myosin is related to motile functions in stress fibers. Furthermore, myosin is an enzyme (ATPase) and its activity is the immediate source of the free energy that drives muscle contraction.(40) So far, the motility of osteocyte processes in canaliculi is unknown. However, this result implies to us the dynamic motility of the osteocyte process in the canaliculi.

Fimbrin in osteoblasts and osteocytes shows different behavior in response to fluid shear stress

Changes in cell shape are the results of regulated changes in the assembly and disassembly kinetics of the cytoskeleton. Numerous previous observations have been made about cytoskeletal changes in response to mechanical stimulation.(20, 41) Therefore, to clarify what the differences in cytoskeletal organization mean with respect to function of osteoblasts and osteocytes, we applied fluid shear stress to the cells and compared the shape changes elicited by such stress. Phase-contrast microscopic observation revealed prominent morphological changes, such as contraction and loss of contact with neighboring cells, in the osteoblasts during the stress treatment with fluid flow. Similar shape changes were reported for endothelial cells and cells of the osteoblastic cell line MC3T3-E1.(20, 42) On the other hand, osteocytes showed little change in shape; especially the shape of osteocyte processes remained almost unaltered during the 30 minutes of flow. Therefore, we consider that unlike most cells that are put into culture, osteocytes are unique, and maintain their shape. Furthermore, this stable structure of processes would serve for the maintenance of the osteocyte network as a mechanosensory syncytium.

Interestingly, as far as we know, ours are the first observations of highly dramatic reorganization of fimbrin after shear stress in osteoblasts, but not in osteocyte processes. It is well known that many responses, including alterations of hormonal receptor, regulation of signal pathways, and expression of specific genes, are some of the changes affecting cytoskeletal organization that occur in response to mechanical loading.(43, 44) Therefore, the difference in behavior of actin-binding protein seen between osteoblasts and osteocytes might be responsible for the critical difference in biological response between these cells.

Acknowledgements

This study was supported in part by the Japan Society for the Promotion of Science in the form of Grants-in-Aid for Scientific Research (14704052, 14207092, 15390635, 15659491).

Ancillary