Bone is clearly a target of vitamin D and as expected, the vitamin D receptor (VDR) is expressed in osteoblasts. However, the presence of VDR in other cells such as osteocytes, osteoclasts, chondroclasts, and chondrocytes is uncertain. Because of difficulties in obtaining sections of undecalcified adult bone, identification of the site of VDR expression in adult bone tissue has been problematic. In addition, the antibodies to VDR used in previous studies lacked specificity, a property crucial for unambiguous conclusions. In the present study, VDR in the various cells from neonatal and adult mouse bone tissues was identified by a highly specific and sensitive immunohistochemistry method following bone decalcification with EGTA. For accurate evaluation of weak immunosignals, samples from Demay VDR knockout mice were used as negative control. Molecular markers were used to identify cell types. Our results showed that EGTA-decalcification of bone tissue had no detectable effect on the immunoreactivity of VDR. VDR was found in osteoblasts and hypertrophic chondrocytes but not in the multinucleated osteoclasts, chondroclasts, and bone marrow stromal cells. Of interest is the finding that immature osteoblasts contain large amounts of VDR, whereas the levels are low or undetectable in mature osteoblasts including bone lining cells and osteocytes. Proliferating chondrocytes appear devoid of VDR, although low levels were found in the hypertrophic chondrocytes. These data demonstrate that osteoblasts and chondrocytes are major targets of 1α,25-dihydroxyvitamin D, but osteoclasts and chondroclasts are minor targets or not at all. A high level of VDR was found in the immature osteoblasts located in the cancellous bone, indicating that they are major targets of 1α,25-dihydroxyvitamin D. Thus, the immature osteoblasts are perhaps responsible for the vitamin D hormone signaling resulting in calcium mobilization and in osteogenesis. © 2014 American Society for Bone and Mineral Research.
Vitamin D receptor (VDR) is a member of the nuclear receptor superfamily and plays a key role in the mechanism of vitamin D action. Following binding of its active ligand, 1,25-dihydroxyvitamin D3 (1,25[OH]2D3), the VDR complex binds to vitamin D responsive elements in target genes to activate or suppress transcription of those genes.[1, 2] VDR is found in specific cell types throughout the body. The tissues with the highest VDR content such as the intestine, kidney, parathyroid gland, and bone are clearly involved in the maintenance of calcium homeostasis and bone formation.[4-7]
Bone tissue contains several cell types, including osteocytes, osteoblasts, osteoclasts, chondrocytes, and chondroclasts. It is one of the first tissues shown to accumulate 3H-1,25(OH)2D3 in the nuclei of osteoblasts. A specific high-affinity binding macromolecule for 1,25(OH)2D3 was demonstrated very early in the preparations of fetal rat bone by ligand binding. VDR was also found in the mature rat bone by ELISA. Further, 1,25(OH)2D3 strongly induces gene transcription in mouse bone.[10, 11] Although it is clear that VDR is present in the bone tissue, the cell types expressing VDR have not yet been clearly delineated.
Because of difficulties in obtaining sections of undecalcified adult bone, early work on VDR detection was done using uncalcified bone tissue from 18-day-old and 20-day-old rat fetus or 2-day-old rats using autoradiography to locate 3H-1,25(OH)2D3 in the nuclei of target tissues. These studies showed that the receptor was not only present in the osteoprogenitor cells, osteoblasts, lining cells, and osteocytes, but also in chondrocytes. The levels of the receptor were low in cells of the articular and resting zone, intermediate in the proliferating zone, and highest in hypertrophic chondrocytes. By direct measurement of the receptor in normal human fetal bone at 10 and 15 weeks gestation, VDR was found in the nuclei of the perichondral mesenchymal cells (osteoblastogenic cells) and in fetal cartilage cells. Another study by electron microscopic immunocytochemistry on neonatal mouse bone showed VDR in the nuclei of osteoblasts, but not in the osteoclasts. By immunohistochemistry, Clemens and colleagues also detected VDR in osteoblasts but not in osteoclasts. Consistent with the results indicating the absence of the receptor in osteoclasts, VDR was not demonstrable by ligand binding in chicken osteoclast preparations.
In contrast, Johnson and colleagues detected VDR immunosignals in the osteoblasts, multinucleated cells, and proliferating and hypertrophic chondrocytes from fetal rat bone. In addition, VDR mRNA and VDR protein were found in the osteoclasts and osteoclast precursors from bones of patients with Paget's disease by in situ reverse transcription polymerase chain reaction (RT-PCR) and immunolocalization. A more recent study by immunohistochemical staining showed the presence of VDR in all the cell types of normal adult human bone tissue, including osteoclasts and bone marrow stromal cells. Clearly, additional study is required to determine the cell type–specific expression of VDR in the bone tissue. First, the presence of VDR in chondroclasts has not been previously studied. Second, the presence of VDR in osteoclasts remains controversial. Third, the studies on the presence of VDR in chondrocytes are incomplete. Forth, the majority of the previous studies have used fetal or uncalcified bone tissue. Finally, the antibodies to VDR used in the previous studies have not been well characterized.
We previously studied available VDR antibodies including those widely used such as rat monoclonal antibody 9A7 and rabbit polyclonal antibody C-20. The sensitivity and specificity of these antibodies were characterized by multiple immunological assays including immunoblotting, immunocytochemistry, and immunohistochemistry. We used samples from Demay's VDR knockout mice devoid of VDR to evaluate nonspecific signals, demonstrating the specificity of each antibody. In addition, comparing the intensity of the VDR band, we were able to evaluate the sensitivity of each antibody to the receptor. Among all the VDR antibodies tested, the D-6 antibody is the most specific and sensitive. It does not react with the samples from Demay VDR knockout mice, whereas other antibodies, such as C-20 and 9A7, bind to proteins in the lysates from those mice. Using this highly specific and sensitive antibody (D-6), VDR was successfully identified in the proximal renal tubule, glomerular parietal epithelial cells, podocytes, and macula densa of the juxtaglomerular apparatus.
In this study, we determined the presence of VDR in the neonatal and adult mouse bone using the D-6 antibody employing the VDR knockout mice as control. We also developed an EGTA-decalcification method that had no effect on VDR immunogenicity. Our results showed that VDR is present in osteoblasts and chondrocytes, but undetectable in osteoclasts, chondroclasts, and bone marrow stromal cells.
Subjects and Methods
Mouse anti-VDR (D-6) antibody was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Polyclonal rabbit anti-COMP, Runx2, and Sp7/Osterix antibodies were purchased from Abcam (Cambridge, MA, USA). Rat anti-mouse CD44 antibody was purchased from BD Pharmingen (Sparks, MD, USA). 4′,6-Diamidino-2-phenylindole (DAPI), Alexa Fluor 594 goat anti-mouse immunoglobulin G (IgG), Alexa Fluor 488 goat anti-rat IgG, and Alexa Fluor 488 goat anti-rabbit IgG secondary antibodies were purchased from Invitrogen (Carlsbad, CA, USA).
Animals and maintenance
Neonatal (postnatal day 1) and 7-week-old to 9-week-old C57BL/6 mice were used for experiments. Demay VDR knockout mice and C57BL/6 littermates were purchased from Jackson Laboratory (West Grove, PA, USA). After arrival in our facility, animals were maintained on chow diet. Blood was collected from adult mice for measurement of serum calcium concentration. Both the wild-type and knockout mice from the same litter of identical age were used for each experiment. Experimental protocols were reviewed and approved by the Animal Care and Use Committee (University of Wisconsin–Madison, Madison, WI, USA).
Collection of organs/tissues
Animals were euthanized with CO2. Duodenum and hind leg bone (tibia and femoral bone with a joint) of the adult mice or duodenum and skull bone of the baby mice were collected and washed gently with phosphate buffered saline (PBS) to remove contaminants. The tissues were chemically fixed with cold 4% paraformaldehyde (PFA) solution on a rocker overnight. The fixed calcified bone and intestinal tissues were subsequently decalcified on a shaker at 4°C. Briefly, the tissues were incubated with 10% EGTA (pH 7.4) in distilled water. Decalcification was performed at 4°C and solutions were changed every day. The decalcification was considered completed when the bone tissue was easily cut with a razor blade. The tissues were then washed five times in PBS for 10 minutes before dehydration.
The fixed skull tissues without decalcification and decalcified tibia bone tissues were dehydrated by progressively replacing PBS with 20%, 50%, 60%, 70%, and then 100% ethanol. Dehydrated tissues were processed for embedding in paraffin. Serial sagittal sections of bone and cross-sections of intestine with 5 µm thickness were obtained and placed on adhesive-coated glass slides to avoid detachment during the following process of immunohistochemical staining.
Primary cell culture
Primary bone cells were isolated from the calvariae of newly born C57B/6 J mice at postnatal day 1 (male and female) as described[22, 23] and cultured in α modified essential medium (αMEM) containing 10% fetal bovine serum (FBS). Cells were cultured for a period of 14 days with medium changes performed every 3 days. Confluent, quiescent cells were plated onto a coverslip and cultured for VDR immunocytochemistry as described.
Staining protocol was performed as described. The calvaria cells were fixed with 1% fresh PFA in PBS containing 0.1% Tween 20 (PBS-T) on ice for 30 minutes. Cells were then treated with 0.1% Triton X-100 in PBS at room temperature for 30 minutes, followed by incubation in a blocking solution containing 3% bovine serum albumin (BSA) and 10% host animal serum for the secondary antibody (30 minutes at room temperature). Cells were incubated with the VDR primary antibody for 1 hour. After washing cells thoroughly, the primary antibody signal was visualized by fluorescent dye-conjugated secondary antibodies. The coverslips incubated with only the secondary antibody were used to evaluate the secondary antibody background. DAPI staining was applied for nuclear staining. Images were captured using a Nikon Inverted Microscope (ECLIPSE TE2000-U; Nikon Instruments Inc., Melville, NJ, USA).
A standard deparaffinizing procedure was used to remove paraffin from sections. Sections were then rehydrated and allowed to sit in deionized water for at least 2 hours. Slides were subsequently incubated for 30 minutes at room temperature in 1% Triton X-100 solution. For antigen retrieval, slides were bathed in citric acid buffer (20 mM, pH 6.0) in an enclosed chamber to maintain moisture and heated in a 14-kW microwave at full power for 12 to 16 minutes. Slides were allowed to cool for at least 3 hours in the citric acid bath. Afterward, slides were washed three times with PBS-T and blocked using a blocking solution containing 3% BSA and 10% host animal serum for the secondary antibody (30 minutes at room temperature). The primary antibody with an optimal concentration as described was allowed to incubate at 37°C on a shaker in a moist chamber for 1 hour. For VDR staining, each group also contained a slide of tissue section incubated with mouse isotype IgG to assess background staining from the secondary antibody. After thoroughly washing with PBS-T, the secondary antibody solution was added to samples and allowed to incubate at 37°C on a shaker for 1 hour. DAPI staining was applied for nuclear staining if necessary. Images were captured using a fluorescence microscope (Nikon Inverted Microscope ECLIPSE TE2000-U) or a confocal fluorescence microscope (Nikon A1R high-speed confocal microscope; Nikon Instruments Inc.).
VDR is detected in the primary bone cells isolated from uncalcified skull bone tissue
Using the D-6 antibody, we first determined the presence of VDR in the primary cultured bone cells isolated from neonatal mouse calvaria (Fig. 1). The VDR expression in these cells was co-stained with antibodies to Runx2 (Fig. 1A–F) or Sp7/Osterix (data not shown). Runx2 and Sp7/Osterix are nuclear transcriptional factors and molecular markers for bone osteoblasts.[24-26] It was clear that VDR was present in the cultured osteoblasts and perhaps chondrocytes as well because Runx2 is also present in the chondrocyte progenitors.[27, 28]
VDR is present in the osteoblasts and cartilage chondrocytes from neonatal mouse calvaria tissue
We next determined if VDR was present in the osteoblasts and cartilage chondrocytes from neonatal mouse calvaria without decalcification (Figs. 2 and 3). The calvaria tissue from the VDR knockout mice was used as negative control (Fig. 2, middle panel and Fig. 3, lower panel). Sp7/Osterix was used as a marker of osteoblasts and cartilage oligomeric matrix protein (COMP/TSP-5) as a marker for cartilage chondrocytes.[29, 30] COMP is normally secreted by the chondrocytes under the condition of normal bone calcification. However, under the condition of low extracellular calcium, ie, impaired calcium metabolism in the VDR knockout mice, COMP secretion is blocked, resulting in its accumulation in the cytoplasm of chondrocytes. In the calvaria tissue, VDR immunostaining was colocalized with the Osterix staining in most osteoblasts (Fig. 2, lower panel). In contrast, VDR immunostaining was weak in the most COMP-stained chondrocytes (Fig. 3, lower panel). In addition, no specific staining was seen in the VDR knockout samples. Thus, it was clear that VDR was found in the uncalcified osteoblasts and cartilage chondrocytes.
VDR is present in adult mouse bone tissue
We next addressed if VDR is present in the adult mouse bone tissue. Recently, the effects of decalcification on the immunoreactivity of certain antigen proteins have been shown.[31, 32] These effects are largely varied with the types of decalcification solutions, incubation time, and even target proteins. Decalcification using calcium chelates, such as EDTA or ethylene glycol tetraacetic acid (EGTA), can preserve antigens. We first determined if the decalcification process could alter the immunoreactivity of the VDR antigen. Immunostaining of the samples from the EGTA-treated or untreated mouse duodenal tissues showed that VDR was highly expressed in the nuclei of gut epithelial cells (data not shown). The immunosignals were strong and there was no difference in the patterns of VDR immunostaining between the EGTA-treated and untreated tissues.
Immunohistochemical examination of the tissue distribution of VDR in normal adult mouse tibia bone tissue showed that the strong VDR immunostaining was visualized in the nuclei of bone cells, located at the region of cancellous bone (Fig. 4B, D, F, H). These cells were positively stained by either Runx2 or Sp7/Osterix. Sp7/Osterix and Runx2 are selectively expressed in the progenitors of osteoblasts and immature osteoblasts. In addition, there was no significant staining either in the VDR knockout samples or in the cellular nucleus from the wild-type bone samples stained with mouse IgG isotype (data not shown). Therefore, VDR was highly expressed in the immature osteoblasts from cancellous bone (Fig. 4A–F). There were no obvious correlations between VDR and Runx2 staining or between VDR and Osterix staining.
Immunostaining also showed that VDR was detectable in some osteocytes (Fig. 5A–D). The majority of bone lining cells, which were located at the limit of cortical bone tissue, were not stained by the VDR antibody, but a few cells were. Therefore, the mature osteoblasts in the cortical bone region seemed to have low or undetectable levels of VDR. A small population of osteocytes and lining cells was also stained by Osterix (Fig. 5B, D, F) and Runx2 (data not shown), perhaps indicating their incomplete developmental stages.
In contrast, the background staining in the mature multinucleated osteoclasts from the VDR knockout samples was identical to that of the wild-type samples. Therefore, there was no VDR in these cells (Fig. 6A–D). Similarly, no VDR-specific immunostaining was seen in the nuclei of bone marrow stromal cells (Fig. 6A–D). There was consistent staining in the outside of the cells from both the wild-type and Demay VDR knockout mice. This staining was probably endogenous IgG because it also appeared in the mouse IgG controls.
VDR is present in adult mouse cartilage tissue
Strong COMP staining was seen in the hyaline cartilage matrix (Fig. 7A), indicating this tissue highly expressed COMP. Proliferating chondrocytes were COMP-positive whereas the hypertrophic ones were surrounded by the COMP-positive matrix (Fig. 7B, D, F). In contrast to the hypertrophic phenotype, the proliferating chondrocytes had a special disc-like morphology. Clearly, VDR was low but detectable in the hypertrophic chondrocytes (Fig. 7C, D). No detectable staining was observed in the proliferating chondrocytes. As expected, the VDR-specific staining was not seen in the samples from the VDR knockout mice (Fig. 7E, F).
Chondroclasts were found in the cancellous bone region and were stained positively for CD44, a molecular marker for chondroclasts. The chondroclasts highly expressed CD44 on the surface of chondroclasts. The background staining in the chondroclasts from the VDR knockout samples was identical to that from the wild-type ones, suggesting no detectable levels of VDR in the chondroclasts (Fig. 7G, H).
There is no doubt that bone is a target of 1,25(OH)2D, but which cells and at what stages of differentiation is either unknown or controversial.
Because of difficulties in obtaining sections of undecalcified adult bone tissue, previous studies on VDR detection in the bone tissue were performed using the samples either from fetal or neonatal animals. A single study was performed on human adult bone tissue. However, this study showed the presence of VDR in all types of bone cells, which is not consistent with the results from other studies.[6, 13, 14] It is unclear if decalcification[31, 32] or the VDR antibodies used[19, 34] could elicit errors in the immunostaining results. In our study, we developed an EGTA decalcification method that has no noticeable effect on the immunoreactivity of VDR. Most importantly, the D-6 antibody is highly specific and sensitive. In addition, we used the samples from the VDR knockout mice as negative control. This unique experimental setup is critical to accurately evaluate the immunostaining results. Our study also included a panel of cell type–specific molecular markers to aid the identification of bone cell types.
With these unique techniques, we confirmed the presence of VDR in the osteoblasts, consistent with the results from previous studies.[5, 8, 12, 13] Unexpectedly, we found that the expression of VDR in the osteoblasts was developmentally regulated. VDR is highly expressed in the maturing osteoblasts, but low or undetectable in the matured osteoblasts including bone lining cells and osteocytes. The immature osteoblasts are mainly located in cancellous bone whereas mature osteocytes are found in cortical bone. Our finding of different levels of VDR in these sites appears consistent with the previous observation of different responses of trabecular and cortical bone to 1,25(OH)2D3 infusion.[35, 36] Calcium mobilization mediated by 1,25(OH)2D and perhaps others is spatially restricted to the cancellous bone.[37, 38] In addition, bone loss related to estrogen or calcium deficiencies, particularly in cancellous bone, can be confined to discrete skeletal sites or even discrete trabeculae.[39, 40] Our results suggest that the various levels of VDR expression may be a determinant for 1,25-(OH)2D-mediated bone calcium mobilization.
Chondrocytes are the only cells found in cartilage. They produce and maintain the cartilaginous matrix, which consists mainly of collagen and proteoglycans. Several studies by autoradiography or immunohistochemical staining[5, 15] showed the presence of VDR in cartilage chondrocytes. We clearly demonstrate low or undetectable level of VDR in the hypertrophic chondrocytes but not in the proliferating chondrocytes. Lack of VDR in the proliferating chondrocytes is consistent with the previous finding that the chondrocyte-specific inactivation of VDR did not alter chondrocyte development in the growth plate.
The results on the presence of VDR in osteoclasts are contradictory. Several studies showed there was no VDR in osteoclasts,[6, 13, 14] whereas others reported the detection of VDR in osteoclasts by immunohistochemical staining[15, 17, 18] or by in situ RT-PCR. The main possible reason that causes the contradiction is the specificity of the methods used in each study. Using a highly specific and sensitive method, we did not detect VDR in the osteoclasts. It is likely that a low number of copies of VDR mRNA may be found in osteoclasts; however, these are either insignificant to elicit a measureable amount of VDR protein or some unknown regulation prevents translation into protein.
Chondroclasts refer to the cells against the nonmineralized transversal septae, whereas osteoclasts refer to cells along the mineralized longitudinal trabeculae. Chondroclasts are multinucleated cells that express less tartrate-resistant acid phosphatase (TRAP) activity and have less developed ruffled-border membranes than osteoclasts in vivo.[43-45] Chondroclasts are involved in cartilage resorption and also appear in calcified cartilage during osteoneogenesis.[33, 46] VDR in this type of cell has not been investigated. Here, we show that VDR is undetectable in these cells.
Two previous studies showed the presence of VDR in bone marrow–derived stromal cells.[18, 47] However, the specificity of the methods used in each study has not been addressed. We find that the receptor is undetectable in this tissue.
In this study, we demonstrate that the osteoblasts and chondrocytes are major targets of vitamin D whereas the osteoclasts and chondroclasts are minor targets or not at all. The high level of VDR in the immature osteoblasts indicates that these cells are major targets of vitamin D, perhaps responsible for 1,25(OH)2D-mediated calcium mobilization from cancellous bone. It also could explain the bone anabolic action of 1,25(OH)2D3 and some of its analogs. The absence of VDR in osteoclasts supports the finding of an indirect action of 1,25(OH)2D3 on the osteoclastic activity, perhaps through osteoblasts and receptor activator of NF-κB ligand (RANKL).[48-50]
All authors state that they have no conflict of interest.
This project was funded by the Wisconsin Alumni Research Foundation (135-A003). We sincerely thank Jean Prahl for providing primary mouse calvaria bone cells and neonatal mice. We also thank Laboratory Manager Lance Rodenkirch in the W.M. Keck Laboratory for Biological Imaging at UW-Madison for his assistance in taking confocal images.
Authors' roles: YW and HFD contributed to the study conception and design. YW collected data. The data were analyzed and interpreted by YW, JZ, and HFD. YW drafted the manuscript. All authors critically revised the content of the manuscript and approved its final version. HFD takes responsibility for the integrity of the data analysis.