The authors have no conflict of interest.
Hyaluronan Increases RANKL Expression in Bone Marrow Stromal Cells Through CD44†
Article first published online: 18 OCT 2004
Copyright © 2005 ASBMR
Journal of Bone and Mineral Research
Volume 20, Issue 1, pages 30–40, January 2005
How to Cite
Cao, J. J., Singleton, P. A., Majumdar, S., Boudignon, B., Burghardt, A., Kurimoto, P., Wronski, T. J., Bourguignon, L. Y. and Halloran, B. P. (2005), Hyaluronan Increases RANKL Expression in Bone Marrow Stromal Cells Through CD44. J Bone Miner Res, 20: 30–40. doi: 10.1359/JBMR.041014
- Issue published online: 4 DEC 2009
- Article first published online: 18 OCT 2004
- Manuscript Accepted: 24 AUG 2004
- Manuscript Revised: 16 AUG 2004
- Manuscript Received: 30 JAN 2004
HA activates CD44 to stimulate RANKL expression in bone marrow stromal cells. HA stimulation of RANKL is blocked by anti-CD44 antibody and is absent in cells from CD44−/− mice. CD44−/− mice exhibit thicker cortical bone and a smaller medullary cavity, but indices of bone resorption are not affected.
Introduction: Hyaluronan (HA), the major nonprotein glycosaminoglycan component of the extracellular matrix in mammalian bone marrow, functions in part through its receptor, CD44, to stimulate a series of intracellular signaling events that lead to cell migration, adhesion, and activation. To determine whether HA activation of CD44 influences RANKL and osteoprotegerin (OPG) expression and whether CD44 is functionally important in bone metabolism, we studied whole bone and bone marrow stromal cells (BMSCs) from wildtype and CD44−/− mice.
Materials and Methods: BMSCs from wildtype and CD44−/− mice at 7 weeks of age were cultured and treated with either HA or anti-CD44 antibody. The levels of mRNA of RANKL, OPG, CD44, alkaline phosphatase (ALP), osteocalcin (OC), and αI collagen (COLL) were determined by quantitative real-time RT-PCR. Levels of RANKL and CD44 protein were measured by immunoblotting, and expression of CD44 in whole bone was determined by immunohistochemical staining. Double immunofluorescence staining and confocal microscopy were used to study colocalization of Cbfa1, CD44, and HA. Tibias were imaged using μCT, and cancellous and cortical parameters were measured. Osteoblast and osteoclast surface in the distal femoral metaphysis and osteoclast on the endocortical surface at the tibio-fibular junction were measured using quantitative histomorphometry. Differences were analyzed using ANOVA and the Newman-Keuls test.
Results: Addition of HA dose-dependently increased RANKL mRNA (3.6-fold) and protein (3-fold) levels in BMSCs. Stimulation of RANKL by HA could be blocked with anti-CD44 antibody. Treatment of cells with HA or anti-CD44 antibody had no significant effect on OPG mRNA levels. Both CD44 and HA localized on the plasma membrane in cells expressing Cbfa1. HA localization on the cell membrane disappeared when cells were preincubated with anti-CD44 antibody. Compared with control mice, cortical bone of CD44−/− was thicker, and medullary area was smaller at both 7 and 17 weeks, but at 7 weeks, indices of bone resorption were normal. At 17 weeks of age, tibial mass of CD44−/− mice was higher than control mice. CD44−/− animals expressed less RANKL in whole bone (−30%) and in BMSCs (−50%). Cells from CD44−/− animals failed to respond to either HA or CD44 antibody treatment.
Conclusions: HA can increase RANKL expression in BMSCs through CD44.
HYALURONAN (HA), a high molecular weight glycosaminoglycan polymer, is the major nonprotein glycosaminoglycan component of the extracellular matrix in mammalian bone marrow.(1-4) It is required for growth, development, cell motility, and cushioning of joints.(5,6) HA is often anchored to CD44, a ubiquitous, abundant, and structurally/functionally important surface receptor that displays HA binding site(s).(7,8) CD44 is widely expressed on the cell surface of a variety of cell types including leukocytes, fibroblasts, osteoblasts, osteocytes, osteoclasts, epithelial, and endothelial cells. The binding of HA to CD44 activates a series of intracellular signaling molecules, including the Src family tyrosine kinases,(9,10) p185HER2,(11) Rho-Kinase,(12) and transforming growth factor (TGF)-β receptor kinases,(13) and leads to cell migration, adhesion, and activation. In addition, HA-mediated CD44 signaling has been shown to be involved in the production of cytokines(14) and hormones(13) in many cell types.
Bone formation and resorption are linked through the RANKL/osteoprotegerin (OPG) axis. Expressed on osteoblasts/stromal cells, RANKL binds to RANK, a transmembrane receptor on hemopoietic osteoclast precursor cells, and initiates gene expression that results in differentiation and maturation of osteoclast precursor cells to active osteoclasts. OPG, a secreted glycoprotein, acts as a nonsignaling decoy receptor that can block activation of RANK, thus inhibiting osteoclast development.(15-21) Many factors have been found to increase RANKL expression and osteoclastogenesis, including parathyroid hormone (PTH)/PTH-related peptide (PTHrP), 1,25(OH)2D, TNFα, glucocorticoids, interleukin (IL)-1β, and IL-6(22,23), and certain physiological states such as aging(24) have been linked to increased osteoblast expression of RANKL. Furthermore, the proresorptive effects of IL-1β and IL-6 have been shown to be induced by HA.(25) Collectively, these data suggest that HA may play a role in osteoblast expression of RANKL.
To determine whether HA influences the expression of RANKL and OPG and whether it functions through CD44, we measured the expression of RANKL and OPG in whole bone and bone marrow stromal cells (BMSCs) from CD44 wildtype (CD44+/+) and CD44 knockout (CD44−/−) mice treated with HA and anti-CD44 antibody. Using μCT and bone histomorphometry, we also analyzed bone structure and cellular composition.
MATERIALS AND METHODS
Six- and 16-week-old male CD44+/+ and CD44−/− mice were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). They were provided a standard rodent diet (8640 Harlan Teklad 22/5 [W]; Harlan Teklad, Madison, WI, USA) containing 1.13% calcium and 0.94% phosphorus and were allowed to acclimate in our animal facility for 3-7 days. Animals were 7 and 17 weeks at the time of experimentation. All animals were housed in air-filtered, humidity- and temperature-controlled rooms with equal 12-h light-12-h dark cycles. The animal protocol for these studies was approved by the Animal Care and Use Committee at the Veterans Affairs Medical Center, San Francisco. Animals were maintained and processed in accordance with the NIH Guide for the Care and Use of Laboratory Animals.
Preparation of bones and cell culture
All in vitro cell culture studies were carried out using 7-week-old mice, and whole bone analyses were carried out using 7- and 17-week-old mice. Animals were killed using an overdose of a ketamine cocktail (10% ketamine [v/v]; Abbott Laboratories, North Chicago, IL, USA; 2% xylazine HCl [v/v]; Boehringer Ingelhein, St Joseph, MO, USA; 1% acepromazine [v/v]; Fermenta Animal Health, Kansas City, MO, USA). To quantify whole bone mRNA and to harvest marrow for cell culture, bones were prepared as described previously with slight modification.(24, 26, 27) Briefly, after death, both tibias and femora were quickly excised, and soft tissues were removed. For whole bone mRNA, the epiphysis of the right femur was removed with a razor blade and discarded, and the marrow was flushed out with a calcium- and magnesium-free PBS (PBS-CMF) solution. The diaphysis was flash-frozen in liquid nitrogen and stored at −70°C before pulverization with a liquid nitrogen-cooled steel mortar and pestle and RNA isolation.
BMSCs were harvested from the left femur. Bones were briefly immersed in 70% ethanol (3 s), and rinsed four times (2 minutes each) in a solution of PBS-CMF containing antibiotics (penicillin-streptomycin) and fungizone under sterile conditions. The epiphyses of each bone were removed with a razor blade, and the marrow was flushed from the diaphysis with a syringe and 26.5-gauge needle and collected in primary culture medium (αMEM containing l-glutamine, nucleosides [Mediatech, Herndon, VA, USA], supplemented with 10% FBS [Atlanta Biologicals, Nocross, GA, USA], 1% antibiotics [penicillin and streptomycin], 0.1% fungizone, and 50 μg/ml ascorbic acid). The marrow cell suspension was gently drawn through an 18-gauge needle to mechanically dissociate the mixture into a uniform single cell suspension. The cells were plated at 10 × 106 cells/10-cm tissue culture dish. On day 5, nonadherent cells were removed by aspiration. Adherent cells were replenished with secondary medium (αMEM supplemented with 10% FBS, 1% antibiotics, 0.1% fungizone, 50 μg/ml l-ascorbic acid, and 3 mM β-glycerophosphate). On day 6 of culture, cells were serum-starved overnight. On day 7 of culture, cells were treated accordingly. After the treatment, cells were washed with PBS-CMF twice and collected for RNA isolation or Western analysis.
Treatment with HA and CD44 antibody
Rooster comb hyaluronan was purchased from Sigma (H-5388; relative mass is about 2,000,000) and dissolved in PBS to form a stock solution of 2 mg/ml. Anti-CD44 antibody (217594) was purchased from Calbiochem (San Diego, CA, USA) and recognizes a common determinant of the CD44 class of glycoproteins of mouse. Cells were routinely serum-starved overnight (∼18 h; therefore deprived of serum HA) before adding HA. Cells were treated with HA for 30 minutes and/or anti-CD44 antibody for 1 h.
RNA isolation and quantitative real-time RT-PCR
Total RNA from bone and cells was extracted using RNA-STAT 60 (Tel-Test, Friendwood, TX, USA) according to the manufacturer's protocol. Concentration and purity of the RNA were determined by measuring the absorbance in TE buffer (10 mM Tris-HCl, pH 8.0, and 1 mM EDTA) at 260 and 280 nm. The levels of mRNA of RANKL, OPG, alkaline phosphatase (ALP), osteocalcin (OC), αI collagen (COLL), and CD44 were determined by quantitative real-time RT-PCR (Q-PCR). Denatured total RNA from bone and cells (2 μg) was reverse transcribed with the following components from Applied Biosystems (Foster City, CA, USA): 1× RT buffer, 5.5 mM MgCl2, 500 μM of each dNTPs, 2.5 μM random hexamers, 0.4 U/μl of RNase inhibitor, and 1.25 U/μl MultiScribe reverse transcription enzyme in a 100-μl reaction volume. The following protocol was used: hexamer incubation at 25°C for 10 minutes, reverse transcription at 37°C for 1 h, and heat inactivation of reverse transcriptase at 95°C for 5 minutes.
The oligonucleotide primers and TaqMan probes used for PCR amplification for RANKL, OPG, ALP, OC, and COLL have been described previously.(24) The primers and probe used for CD44 PCR amplification were forward primer 5′-TGAAGTTGGCCCTGAGCAA-3′, reverse primer 5′-CTCGGAATTACCACATTTCCTTCT-3′, and Taqman probe 5′-TTTGAAACATGCAGGTATGGGTTC-3′. They were designed to amplify all forms of CD44 by the Primer Express software (version 1.0; Applied Biosystems) and synthesized by Integrated DNA Technologies (IDT, Coralville, IA, USA) with HPLC purification.
After reverse transcription, the cDNA (2 μl) was amplified and quantified using a Sequence Detection System (SDS 7700) and a PCR universal protocol as follows: AmpliTaq Gold activation at 95°C for 15 s and annealing/extension at 60°C for 1 minute. The fluorescence of the accumulated double-stranded products was monitored in real time. To account for differences arising from reverse transcription efficiency and quality of the total RNA, the relative RANKL, OPG, ALP, OC, COLL, and CD44 mRNA levels were normalized to levels of GAPDH mRNA in the same sample.
Gel electrophoresis, Western blotting, and CD44 immunohistochemistry
BMSCs were harvested at day 7 of culture, washed twice with cold PBS-CMF to remove traces of media, and centrifuged at 1000g at 4°C for 5 minutes. The cell pellet was collected and solubilized in RIPA buffer (50 mM HEPES, pH 7.4, 1% deoxycholate, 1% Triton X-100, 0.1% SDS, 150 mM NaCl, 1 mM EDTA) and 1 mM 2-mercaptoethanol containing 1× protease inhibitors cocktail (1 μg/ml Leupeptin, 1 μg/ml pepstain, 20 μg/ml PMSF, 2 μg/ml aprotinin, added immediately before use; Roche Molecular Biochemicals, Indianapolis, IN, USA). The homogenate was centrifuged at 1000g for 15 minutes at 4°C. Protein content was measured by the BCA protein assay kit (Pierce, Rockford, IL, USA). Equal amounts of protein (50 μg) were resolved on a 10% SDS-PAGE (Bio-Rad, Hercules, CA, USA) and electroblotted onto Nitrocellulose membrane (Bio-Rad) using 192 mM glycine, 10 mM Tris, 0.05% SDS, and 20% methanol. The membranes were incubated with either RANKL antibody (Oncogene, Cambridge, MA, USA), CD44 antibody (Calbiochem, San Diego, CA, USA), or actin antibody (Sigma, St Louis, MO, USA) after being blocked with 5% nonfat dry milk in TBS (100 mM Tris HCl, 0.9% NaCl, pH 7.5). After washing with TBS, anti-mouse IgG horseradish peroxidase conjugate (Amersham Life Science, Piscataway, NJ, USA) or anti-rat IgG horseradish peroxidase (HRP) conjugate secondary antibody (Sigma) was applied for detection with a chemiluminescence substrate (SuperSignal West Femto Maximum Sensitivity Substrate; Pierce).
Equal protein loading and transfer were verified by actin protein expression. Band intensities were quantified using Biomax 1D image analysis software (Kodak Scientific Imaging Systems).
The left tibia was fixed in 4% paraformaldehyde in 0.1 M phosphate buffer for 24 h at 4°C. The tibias were demineralized in 10% EDTA (pH 7.3). Demineralized bones were bisected longitudinally and embedded in paraffin, and 5-μm-thick sections from the proximal tibia were placed on positively charged glass slides (Fisher Scientific, Pittsburgh, PA, USA). Anti-CD44 antibody for immunohistochemistry was purchased from Calbiochem. Immunohistochemical stains were performed using the Vectastain ABC kit (Vector Laboratories, Burlingame, CA, USA). Tissue sections were deparaffinized in xylene, followed by two washes in ethanol and a brief wash in PBS containing 0.1% Tween 20, pH 7.4. Sections were immersed in 0.3% hydrogen peroxide in absolute methanol for 15 minutes to quench endogenous peroxidase activity. Before the addition of the primary antibody, nonspecific tissue binding was blocked by incubating the tissue section for 30 minutes with 4% bovine serum in PBS buffer containing 0.05% Tween 20 and 0.5% teleostean skin gelatin. The primary antibodies were applied and incubated overnight at 4°C in a humid environment. The sections were washed three times in 0.1% Tween 20. The secondary biotinylated antibody and the streptavidin-HRP conjugate complex were applied in a humidified chamber for 60 and 30 minutes, respectively. After washing in buffer, the chromogen diaminobenzidine was applied for 5 minutes followed by a counterstain with Mayer's hematoxylin. Negative controls included substituting the primary antisera with preimmune sera from the same species and omitting the primary antibody. All controls revealed the expected negative results.
BMSCs were grown on cover slides for 9 days and switched to serum-free medium overnight. Cells were treated with or without anti-CD44 antibody for 1 h and incubated in staining medium either with or without 100 μg/ml of BODIPY FL-HA (H-23379; Molecular Probes, Eugene, OR, USA) for 2 h. Cells were washed with PBS buffer and fixed in 3% paraformaldehyde for 20 minutes. After blocking in 3% goat serum in PBS for 1 h, cells were either stained with R-phycoerythrin (R-PE) conjugated rat anti-mouse CD44 antibody (Caltag Laboratories, Burlingame, CA, USA) for 1 h at room temperature and then stained with rabbit anti-mouse Cbfa1 (Core binding factor a1, a transcription factor specific to osteoblasts; Molecular Probes) after being rendered permeable with 0.3% Triton X-100 for 10 minutes or stained with rabbit anti-mouse Cbfa1 directly after permeabilization. Cells were washed with PBS three times and then incubated with appropriate fluorescent-labeled second antibody (either FITC-conjugated or rhodamine-red conjugated goat anti-rabbit IgG) for 1 h to detect Cbfa1 localization. To detect nonspecific antibody binding, cells were incubated with FITC-conjugated or rhodamine-red conjugated normal goat IgG after blocking in 3% goat serum in PBS for 1 h. No labeling was observed in control samples. The fluorescein, R-PE, and rhodamine red-labeled samples were examined with a confocal laser scanning microscope.
Determination of CFU-f, CFU-AP+, and calcium nodule formation
BMSCs were grown in 10-cm plates for up to 28 days. Colony forming units-fibroblastic (CFU-f), CFU-f ALP positive (CFU-ALP+), and calcium nodule number were determined as previously described.(22, 26, 27) Briefly, total CFU-f was visualized by staining with a solution of 0.2% crystal violet (CV) in 2% ethanol for 30 minutes at room temperature. The staining solution was aspirated, and unbound stain was removed by rinsing the cultures four times with distilled water, and the CV stain was eluted with 0.2% Triton X-100 (Sigma) and measured in a spectrophotometer at 590 nm. To visualize CFU-ALP+ colonies, cells were stained with a commercially available kit (diagnostic kit #86; Sigma) after fixation for 1 h with 10% formalin. To quantify ALP activity, the same destained cultures from CV staining were incubated for 15 minutes at 37°C with 10 ml of a solution containing equal parts p-nitrophenol phosphate (104 phosphatase substrate; Sigma) and alkaline buffer solution (221; Sigma), and absorbance was measured at 410 nm. Calcium nodule number at day 28 of culture was determined by staining with 2% Alizarin red (AR; Sigma) for 10 minutes. The stain was aspirated, and the cultures were rinsed five times with distilled water to remove loosely bound stain. The specifically bound stain was eluted with a solution of 0.5 N HCl/5% SDS and measured in a spectrophotometer at 415 nm.
μCT measurements and histomorphometry
μCT measurements and bone histomorphometry were carried out using 7- and 17-week-old mice. The right tibia of the mouse was defatted by sequential extraction overnight in ethanol and diethyl ether using a Soxhlet apparatus and dried overnight at 95°C. The fat-free weight and tibial length were obtained. Tibias were imaged using a μCT (μCT-20; Scanco Medical AG, Bassersdorf, Switzerland) as described in detail previously.(28) From the resulting images, bone volume fraction (BV/TV), trabecular thickness (Tb.Th), trabecular number (Tb.N), and trabecular separation (Tb.Sp) in the proximal tibia and cortical thickness (Ct.Th), cortical area (Ct.Ar), medullary area (Me.Ar), and total bone area at the tibia-fibular junction were calculated.
The distal end of the right femur and the tibio-fibular junction (TFJ) were prepared for quantitative histomorphometry. Osteoblast and osteoclast surface in the distal femoral metaphysis and endocortical osteoclast surface at the TFJ were quantified.(29) Bone resorption on the endocortical surface was measured as described by Turner.(30) Briefly, calcein was administered subcutaneously (10 mg/kg) to two groups of animals: one group was killed 24 h after the calcein injection and the other group was killed 7 days later. Maximum labeled surface occurred at 24 h. Over the next 7 days, label is gradually lost as a consequence of endocortical resorption. The percent of labeled surface remaining at day 7 gives a relative value of resorptive activity.
All data are reported as means ± SD. Students t-test, ANOVA, and the Newman-Keuls test were used to analyze differences among treatments. All experiments were performed at least twice with similar results.
HA stimulates expression of RANKL but not OPG
To determine the effect of HA on expression of RANKL and OPG, we treated BMSCs at day 7 of culture for 30 minutes with different concentrations of HA. Addition of HA increased progressively RANKL mRNA levels as measured by Q-PCR (Fig. 1A). A dose of 25 nM produced a peak response (3.6-fold above controls, p < 0.05). Although, HA addition tended to decrease OPG expression the change did not reach significance.
To characterize the time course, we measured expression of RANKL and OPG at 0, 0.5, 1, 3, and 8 h after addition of 25 nM HA (Fig. 1B). Treatment with HA caused a rapid increase in RANKL mRNA in BMSCs (p < 0.05), with a peak response between 0.5 (4.33-fold) and 1 h (4.15-fold). Expression was still elevated at 8 h (p < 0.05). Expression of OPG gradually decreased with treatment time, but the change in expression did not reach significance.
Stimulation of RANKL by HA is mediated through CD44
To determine whether HA regulates RANKL and OPG expression through the CD44 receptor, we studied the effects of HA treatment on RANKL and OPG mRNA levels after preincubation of BMSCs with or without anti-CD44 antibody (1 μg/ml) for 1 h (Fig. 2A). In the absence of antibody, HA treatment increased RANKL mRNA levels in BMSCs (3.9-fold over control, p < 0.05). When cells were preincubated with CD44 antibody alone or treated with HA after CD44 antibody incubation, RANKL mRNA levels were similar to that in control cells. Neither pretreatment of cells with anti-CD44 antibody nor treatment of cells with HA after anti-CD44 antibody incubation affected OPG mRNA levels.
The effect of anti-CD44 antibody on HA-mediated RANKL expression was dose-dependent (Fig. 2B). Treatment of cells with 25 nM HA significantly increased RANKL mRNA levels (3.7-fold over control, p < 0.05), whereas preincubation of cells with CD44 antibody decreased RANKL expression stimulated by HA. Basal levels of RANKL expression were unaffected by anti-CD44 antibody. Contrary to the effect of anti-CD44 antibody on HA-stimulated RANKL expression, increasing concentrations of anti-CD44 antibody preincubation had no significant effect on OPG expression in BMSCs (Fig. 2B).
Hyaluronan also increased RANKL protein levels (Fig. 3). RANKL protein levels in BMSCs increased roughly 3-fold after 25 nM HA treatment for 1 h. HA stimulated RANKL protein levels were completely blocked by 1-h pretreatment with anti-CD44 antibody. Neither CD44 nor actin protein expression were affected by HA or CD44 antibody treatment.
CD44 and HA co-localize in cells expressing Cbfa1
Using confocal microscopy, fluorescently labeled HA, and antibodies against CD44 and Cbfa1, we studied the cellular distribution of CD44, HA, and Cbfa1 in BMSCs. CD44 and HA co-localized on the plasma membrane, whereas Cbfa1 was found predominantly in the nucleus but also at lower concentrations throughout the cytoplasm (Fig. 4). Cells staining positive for CD44 and HA (60% of total) also expressed Cbfa1. Preincubation with anti-CD44 antibody eliminated HA binding (data not shown).
Ablation of CD44 reduces RANKL expression
We used CD44−/− as a model to further determine whether the stimulation of RANKL expression occurs through CD44. No CD44 was expressed in mouse bone or bone marrow cells. As seen in Fig. 5A, CD44 mRNA levels in CD44+/+ animals were 240-fold higher than in mice lacking the CD44 gene as measured by Q-PCR. No CD44 protein was detected in CD44−/− mice as seen in Western blot, whereas control animals expressed predominantly the standard form of CD44 (Fig. 5B). Immunostaining of mouse distal femur showed that CD44 protein was expressed in bone marrow cells, osteoblasts, osteoclasts, and osteocytes (Figs. 6A–6D). Expression is highest in osteocytes and osteoclasts, and although the majority of osteoblasts appeared CD44 negative, we did find some stained for CD44 (10%). Examination of the growth plate of the femur revealed that CD44 is present in the ossification zone but not in other regions of the growth plate (Fig. 6C). CD44−/− animals showed no CD44 expression anywhere in the distal femur (Fig. 6B) or in the growth plate area (Fig. 6D). Compared with CD44+/+, CFU-f, CFU-ALP+, and calcium nodule number were all reduced in BMSCs from CD44−/− mice (Fig. 7). When quantified, total protein (crystal violet staining), ALP activity and calcium were also less in CD44−/− than in CD44+/+ mice. When corrected for protein, ALP activity was the same in both groups of mice. Decreased CFU-f, CFU-ALP+, and calcium nodule number was also seen in cells from wildtype mice treated with 1 μg/ml anti-CD44 antibody (data not shown).
CD44−/− mice expressed less RANKL (−30%, p < 0.05) in whole bone than control animals (Fig. 8A). Similarly, RANKL mRNA levels in cultured BMSCs from knockout mice were less than in cells from control animals (−50%, p < 0.05; Fig. 8B). OPG mRNA levels in bone and BMSCs from CD44−/− mice were not different from control (Figs. 8A and 8B). To determine whether CD44 deficiency had any effect on biochemical markers of bone formation, we measured the expression of ALP, OC, and COLL. No significant differences were observed between CD44−/− and control mice in either whole bone (Fig. 8A) or BMSCs (data not shown).
Cells from CD44−/− animals failed to respond to either HA or anti-CD44 antibody treatment as seen in wildtype mice (Fig. 9). There were no clear trends in OPG expression in BMSCs from either control or CD44−/− mice after treatment with HA and anti-CD44 antibody (data not shown).
Bone structure and histomorphometry
To determine the effects of CD44 deficiency on whole bone, we used μCT and quantitative histomorphometry. Tibial fat-free weight was not significantly affected by CD44 deficiency, but the tibias in CD44−/− mice were shorter than wildtype animals (p < 0.05; Table 1). BV/TV, Tb.Th, Tb.N, and Tb.Sp in the proximal tibia of CD44−/− mice did not differ from wildtype animals (Table 1). However, compared with wildtype mice, Ct.Th was greater (p = 0.002), Ct.Ar tended to be greater (p = 0.058), and Me.Ar was smaller (p < 0.05) in CD44−/− mice. Total bone area was the same in knockout and wildtype animals. Osteoblast and osteoclast surfaces in the cancellous bone of the distal femoral metaphysis are reported in Table 2. A marginal difference (p = 0.053) in osteoclast surface (week 17) was observed. At the TFJ, endocortical osteoclast number and surface were the same in CD44−/− and CD44+/+ mice. Loss of calcein label, a measure of bone resorption, on the endocortical surface of CD44−/− mice (15%, n = 5) was not different from CD44+/+ mice (15%, n = 4).
In this study, we presented evidence that HA regulates expression of RANKL in BMSCs. Preincubating cells with CD44 antibody blocked induced expression of RANKL, and HA had no effect on RANKL expression in cells from CD44−/− mice. Co-localization of CD44 and exogenously administered HA on the plasma membrane of cells expressing Cbfa1, a marker of the osteoblast linage (Fig. 4), and the characteristic maturation of these cells into colonies that form calcified nodules strongly suggested that the increase in RANKL expression induced by HA is associated with the osteoblast.
BMSC cultures are not homogeneous. Accordingly, it is possible that cells other than the osteoblast may contribute to CD44-mediated HA stimulation of RANKL expression. Macrophages are likely the most significant single contaminant in mouse BMSC cultures (20-30%; our unpublished data). However, mouse primary macrophages do not express RANKL either basally or after HA treatment (unpublished data); thus, these cells are not likely to directly influence our conclusions. However, the presence of macrophages in the culture may indirectly affect RANKL expression induced by HA. Khaldoyanidi et al.(25) reported that HA induces IL-1β and IL-6 secretion in bone marrow macrophages. These cytokines have been shown to increase RANKL expression in osteoblasts. Thus, it is possible that the HA-stimulated expression of RANKL occurs indirectly through the macrophage. It is also possible that other contaminating cells in the culture are also affecting our results. The predominance of evidence, however, supports the contention that HA stimulates directly RANKL expression in the osteoblast linage. Whether HA activation of CD44 directly increases RANKL expression in the osteoblast or whether upregulation of RANKL expression is mediated through other signaling pathways that are modulated by CD44 is not clear.
CD44 is expressed in many cell types including leukocytes, fibroblasts, epithelial cells, keratinocytes, and some endothelial cells.(31) Our histochemical staining (Fig. 6) indicates that, in the femoral growth plate, CD44 is present on the surfaces of cells in the zones of erosion and ossification, but not in other regions.(32) In the femoral metaphysis, osteoclasts, osteocytes, and some osteoblasts are CD44+, findings consistent with those of Nakamura et al.(32) and others.(33,34) Whereas many CD44 isoforms are expressed in other cell types,(35) the standard form (85 kDa) of CD44 is the only form expressed in osteoblasts (our data)(36) and in osteoclasts.(37) These results support the contention that osteoblasts express CD44 both in vitro and in vivo.
CD44-deficient mice are viable without obvious developmental defects and show no overt abnormalities, although CD44-deficient lymphocytes exhibit impaired entry into the adult thymus.(38) Our μCT data indicate that the changes induced in bone by CD44 deficiency are subtle but significant (Table 1). Cancellous bone volume in the metaphysis is normal. Osteoblast (7 and 17 weeks) and osteoclast (17 weeks) surfaces tend to be less in CD44−/− than in CD+/+ mice, but the difference does not reach significance (Table 2). That there is a trend to have fewer osteoblasts in CD44−/− mice is consistent with the observation that there are fewer CFU-Fs and less protein in cultures of BMSCs from CD44−/− mice. Overall, bone mass is normal (7 weeks) or increased (17 weeks), and tibias are shorter in CD44−/− mice. Cortical thickness is increased and medullary area is decreased. These changes are consistent with diminished endocortical resorption. Assuming that HA functions in part to maintain chronic stimulation of RANKL synthesis through CD44, loss of the CD44 receptor would be expected to decrease bone resorption and reduce marrow cavity expansion. This could account for the decreased medullary area and increased cortical thickness. However, osteoclast number per millimeter of endocortical surface and the percent of endocortical surface covered by osteoclasts are normal at 7 weeks of age (Table 2). This inconsistency may be a consequence of deficient resorption at an earlier age that produces a sustained decrease in medullary area.
Although previous studies have not directly linked HA with RANKL expression, HA and bone metabolism are known to be closely interrelated. Siczkowski et al.(4) reported that HA regulates human bone marrow-derived stromal cells to support hematopoiesis. PTH, which is known to stimulate RANKL expression,(22) can also stimulate HA synthesis in an osteoblast-like cell line.(39) Thus, the stimulatory effect of PTH on RANKL expression may be mediated in part by increasing HA synthesis.
That bone is nearly normal in CD44−/− mice is remarkable. Given the potent effect of CD44-mediated stimulation of RANKL expression by HA and the importance of CD44 in cell migration, adhesion, and activity, we expected to see a reduction in resorptive activity. Our results suggest that resorption is, at best, marginally reduced. That resorptive activity is not more affected by CD44 deficiency is puzzling. It may be that compensatory mechanisms come into play to offset the effects of CD44 deficiency on RANKL expression in vivo, or it may be that osteoclast deficiency of CD44 in the CD44−/− mouse offsets the normal effects of RANKL on osteoclasts (e.g., CD44 deficiency may increase osteoclast activity).
In summary, our data indicate that HA can stimulate RANKL expression in BMSCs and that this process is mediated, at least in part, by CD44. The subtle changes in bone phenotype induced by CD44 deficiency suggest that the CD44/HA axis may play a significant role in bone metabolism.
The authors thank Sandra Chang (Department of Dermatology at University of California, San Francisco, CA) and Rosemarie Arzaga (University of Florida) for technical assistance. This work was supported by an NIH fellowship (DK07418-21) to JC, the Veteran's Affairs Merit Review Program (BH and LB), NIH Grants CA66163 and CA78633 (LB), and Department of Defense Grant DAMD17-99-1-9291 (LB).
- 11968 The glycosaminoglycans of dog compact bone and epiphyseal cartilage in the normal state and in experimental hyperparathyroidism. Calcif Tissue Res 2: 314–333.,
- 21987 Compartmentalization of a haematopoietic growth factor (GM-CSF) by glycosaminoglycans in the bone marrow microenvironment. Nature 326: 403–405., , ,
- 31990 Regulation of hemopoiesis by bone marrow stromal cells and their products. Annu Rev Immunol 8: 111–137.
- 41993 Hyaluronic acid regulates the function and distribution of sulfated glycosaminoglycans in bone marrow stromal cultures. Exp Hematol 21: 126–130., , ,
- 52000 Hyaluronan: A multifunctional, megaDalton, stealth molecule. Curr Opin Cell Biol 12: 581–586.,
- 62002 Hyaluronan and its catabolic products in tissue injury and repair. Matrix Biol 21: 25–29.
- 71990 CD44 is the principal cell surface receptor for hyaluronate. Cell 61: 1303–1313., , , ,
- 81993 CD44 and its interaction with extracellular matrix. Adv Immunol 54: 271–335., ,
- 91998 The ankyrin-binding domain of CD44s is involved in regulating hyaluronic acid-mediated functions and prostate tumor cell transformation. Cell Motil Cytoskeleton 39: 209–222.,
- 102001 CD44 interaction with c-Src kinase promotes cortactin-mediated cytoskeleton function and hyaluronic acid-dependent ovarian tumor cell migration. J Biol Chem 276: 7327–7336., , ,
- 111997 Interaction between the adhesion receptor, CD44, and the oncogene product, p185HER2, promotes human ovarian tumor cell activation. J Biol Chem 272: 27913–27918., , , , ,
- 121999 Rho-kinase (ROK) promotes CD44v(3,8–10)-ankyrin interaction and tumor cell migration in metastatic breast cancer cells. Cell Motil Cytoskeleton 43: 269–287., , , ,
- 132002 Hyaluronan promotes signaling interaction between CD44 and the transforming growth factor beta receptor I in metastatic breast tumor cells. J Biol Chem 277: 39703–39712., , ,
- 142001 CD44-mediated oncogenic signaling and cytoskeleton activation during mammary tumor progression. J Mammary Gland Biol Neoplasia 6: 287–297.
- 151998 Osteoprotegerin ligand is a cytokine that regulates osteoclast differentiation and activation. Cell 93: 165–176., , , , , , , , , , , , , , , , , , , , , , ,
- 161998 Identity of osteoclastogenesis inhibitory factor (OCIF) and osteoprotegerin (OPG): A mechanism by which OPG/OCIF inhibits osteoclastogenesis in vitro. Endocrinology 139: 1329–1337., , , , , , , , , , , , , ,
- 171998 Osteoclast differentiation factor is a ligand for osteoprotegerin/osteoclastogenesis-inhibitory factor and is identical to TRANCE/RANKL. Proc Natl Acad Sci USA 95: 3597–3602., , , , , , , , , , , , , , ,
- 181997 Osteoprotegerin: A novel secreted protein involved in the regulation of bone density. Cell 89: 309–319., , , , , , , , , , , , , , , , , , , , , , , ,
- 191999 The ligand for osteoprotegerin (OPGL) directly activates mature osteoclasts. J Cell Biol 145: 527–538., , , , , , , , , , ,
- 201997 Characterization of a novel TNF-like ligand and recently described TNF ligand and TNF receptor superfamily genes and their constitutive and inducible expression in hematopoietic and non-hematopoietic cells. Gene 204: 35–46., , , , , , ,
- 211999 OPGL is a key regulator of osteoclastogenesis, lymphocyte development and lymph-node organogenesis. Nature 397: 315–323., , , , , , , , , , , , , , , ,
- 222001 Osteoprotegerin: A physiological and pharmacological inhibitor of bone resorption. Curr Pharm Des 7: 613–635.,
- 232003 Osteoclast differentiation and activation. Nature 423: 337–342., ,
- 242003 Expression of RANKL and OPG correlates with age-related bone loss in male C57BL/6 mice. J Bone Miner Res 18: 270–277., , ,
- 251999 Hyaluronate-enhanced hematopoiesis: Two different receptors trigger the release of interleukin-1beta and interleukin-6 from bone marrow macrophages. Blood 94: 940–949., , , ,
- 261999 Skeletal unloading causes resistance of osteoprogenitor cells to parathyroid hormone and to insulin-like growth factor-I. J Bone Miner Res 14: 21–31., , , , ,
- 272001 Osteoblasts respond to pulsatile fluid flow with short-term increases in PGE(2) but no change in mineralization. J Appl Physiol 90: 1849–1854., , , ,
- 282002 Changes in bone structure and mass with advancing age in the male C57BL/6J mouse. J Bone Miner Res 17: 1044–1050., , , , ,
- 291988 Temporal relationship between bone loss and increased bone turnover in ovariectomized rats. Calcif Tissue Int 43: 179–183., ,
- 301994 Cancellous bone turnover in growing rats: Time-depemdemt changes in association between calcein label and ostoeblasts. J Bone Miner Res 9: 1419–1424.
- 312002 The hyaluronan receptor, CD44. Int J Biochem Cell Biol 34: 718–721.,
- 321995 Localization of CD44, the hyaluronate receptor, on the plasma membrane of osteocytes and osteoclasts in rat tibiae. Cell Tissue Res 280: 225–233., , ,
- 331994 CD44 expression in human bone: A novel marker of osteocyte differentiation. J Bone Miner Res 9: 39–44., ,
- 341996 Immunolocalization of CD44 and the ERM in bone cells of mouse tibiae. J Bone Miner Res 11: 1715–1722.,
- 352001 CD44-mediated uptake and degradation of hyaluronan. Matrix Biol 21: 15–23., ,
- 361996 CD44 expression in fetal rat bone: In vivo and in vitro analysis. Exp Cell Res 223: 467–477.,
- 372003 Rho-dependent Rho kinase activation increases CD44 surface expression and bone resorption in osteoclasts. J Biol Chem 278: 29086–29097., , , ,
- 381999 CD44-deficient mice develop normally with changes in subpopulations and recirculation of lymphocyte subsets. J Immunol 163: 4917–4923., , ,
- 391994 Parathyroid hormone stimulates hyaluronan synthesis in an osteoblast-like cell line. J Biol Chem 269: 13200–13206., ,