Mice Lacking the Integrin β5 Subunit Have Accelerated Osteoclast Maturation and Increased Activity in the Estrogen-Deficient State


  • The authors have no conflict of interest.


Integrin αvβ5 is expressed on osteoclast precursors and is capable of recognizing the same amino acid motif as αvβ3. Three-month-old β5−/− female OVX mice had increased osteoclastogenesis ex vivo, and μCT assessment of trabecular bone volume was 53% lower than WT-OVX animals. These preliminary data suggest αvβ5 integrin's presence on osteoclast precursors may inhibit of osteoclast formation.

Introduction: Osteoclasts are unique resorptive skeletal cells, capable of degrading bone on contact to the juxtaposed matrix. Integrin αvβ5 is expressed on osteoclast precursors, structurally similar to αvβ3, and capable of recognizing the same amino acid motif. Given the structural relationship and reciprocal regulation of αvβ3 and αvβ5, the purpose of this study was to evaluate how αvβ5 might contribute to osteoclast maturation and activity.

Materials and Methods: Three-month-old wildtype (WT) and β5−/− female mice had ovariectomy (OVX) or sham operations. The osteoclastogenic capacity of marrow-derived precursors, the kinetic, the circulating, and structural parameters of bone remodeling, was determined after 6 weeks of paired feeding.

Results and Conclusions: OVX increased osteoclastogenesis ex vivo and in vivo. Osteoclast formation and prolonged pre-osteoclast survival were substantially enhanced in cultures containing β5−/− cells whether obtained from sham-operated or OVX mice. Expression of cathepsin K, β3 integrin subunit, and calcitonin receptor were accelerated in cultured β5−/−osteoclasts. β5−/− osteoclasts from OVX animals showed a 3-fold enhancement of net resorptive activity, with quantitative μCT showing trabecular bone volume loss after OVX 53% greater in β5−/− OVX compared with similarly treated WT OVX mice (p < 0.05). α5β3 seems to be an inhibitor of osteoclast formation, in contrast to αvβ3. In addition, loss of αvβ5 seems to accelerate osteoclast formation in the OVX model. Further examination of αvβ5 signaling pathways may enhance our understanding of the activation of bone resorption.


BONE IS CONTINUOUSLY removed and replaced by the coupled actions of osteoclasts and osteoblasts, resulting in renewal of the skeleton, maintenance of its structure, and mineral homeostasis.(1) Osteoclasts are derived from bone marrow precursors of the monocyte-macrophage lineage,(2) whose differentiation, survival, and function are controlled by two stromal cell/osteoblast produced cytokines, RANKL and macrophage-colony stimulating factor (M-CSF).(3) The discovery that these two agents are sufficient to generate pure populations of cells committed to the osteoclast phenotype, in culture, has yielded major insights into the mechanisms by which osteoclasts are formed. Because both events require cell-matrix recognition, plasma membrane-residing molecules that mediate osteoclast-bone attachment are central to the resorptive process.

Integrins are transmembrane heterodimers consisting of various families of associated α and β chains.(4-8) The integrin αvβ3 plays an essential role in osteoclast differentiation and its capacity to degrade bone.(9) The osteoclast recruiting and stimulating functions of αvβ3 depend on intracellular signaling through the cytoplasmic domain of its β3 subunit, particularly Ser752.(10)

αvβ5 is structurally related to αvβ3(11) and recognizes the same tripeptide, arg-gly-asp (RGD), in all of its known ligands.(12) Interestingly, αvβ3 and αvβ5 are reciprocally regulated during osteoclastogenesis. For example, αvβ3 is absent and αvβ5 is abundant in immature osteoclast precursors such as bone marrow macrophages.(13) As the cells assume the osteoclast phenotype, αvβ5 disappears and αvβ3 appears. Other cytokines that augment the osteoclastogenic process, such as TNFα, exert a similar reciprocal effect on the two matrix receptors.(14)

Given the structural relationship and reciprocal regulation of αvβ3 and αvβ5, we hypothesized that αvβ5, like αvβ3, might contribute to the generation of osteoclasts. In this study we found that αvβ5 does affect osteoclast differentiation and function, but does so in a manner opposite to that of αvβ3. Specifically, β5−/− mice generate more osteoclast precursors and in the estrogen deplete state, generate hyper-resorptive osteoclasts ex vivo and in vivo, resulting in profoundly accelerated bone loss. Thus, αvβ5, as opposed to αvβ3, is anti-osteoclastogenic and blunts the loss of bone in the estrogen-deficient state.



β5 knockout mice were generated as described(15) and bred onto a 129 SvJ mice strain. Wildtype (WT) or β5−/− mice were divided into four groups and were either sham-operated or underwent bilateral ovariectomy (OVX) at 12 weeks of age and were maintained for an additional 6 weeks to assess the estrogen-deficient bone loss (Table 1). The animals were maintained at 72°F and in 12-h/12-h light/dark cycles. The animals were allowed free access to water, and the OVX mice were pair-fed to sham-controls with commercial natural diet (Teklad Rodent Laboratory Chow 8604; Harlan Teklad, Madison, WI, USA), which contained 1.46% calcium, 0.99% phosphorus, and 4.96 IU/g of vitamin D3. The average amount of food consumed by the WT-Sham and the β5−/−-Sham was about 3 g/day, and this amount was fed to all OVX mice. All animals were treated according to the USDA animal care guidelines with the approval of the UCSF Committee on Animal Research.

Table Table 1.. Experimental Groups and Body Weights During the Experiment
original image

Experimental procedures

A 24-h urine collection was performed in fasting animals 1 day before the sham or OVX surgeries and at 3 and 6 weeks after surgery in all animals. Serum samples were taken during necropsy and stored at −80°C until they were assessed for biochemical makers of bone turnover. Calcein (10 mg/kg, IP) was given 9 and 2 days before death to access bone turnover. For the surgeries and necropsies, the animals were anesthetized with isoflurane. At necropsy, the mice were exsanguinated by cardiac puncture, and successful removal of the ovaries was confirmed by failure to detect ovarian tissue and by marked atrophy of the uterine horns. At the time of death, tibias were placed in 10% phosphate-buffered formalin for 24 h and transferred to 70% ethanol for μCT analysis and bone histomorphometry preparation. Mouse femurs were separated from the skeleton, bone marrow was irrigated, and bone marrow monocytes/macrophages (BMMs) were collected for future cell culture experiments.

Ex vivo mouse osteoclastogenesis and functional assays

Osteoclast differentiation:

BMMs were collected from the mouse femurs by flushing with PBS using a 25-gauge needle.(16) Cells were collected and plated in 1 ml of α-MEM on 24-well tissue culture plates at a density of 1 × 106 cell/well (Invitrogen, Carlsbad, CA, USA) supplemented with 10 ng/ml M-CSF (R&D Biosystems, Minneapolis, MN, USA) and incubated overnight. The following day, nonadherent cells were collected, transferred to 24-well dishes, and cultured in α-MEM with 10% FBS, 1% glutamine, penicillin/streptomycin, 10 ng/ml M-CSF, and 50 ng/ml RANKL (R&D Biosystems) for an additional 6 days with addition of fresh media every 3 days. All ex vivo studies were repeated at least in triplicate.

TRACP staining:

After 7 days in culture, cells were fixed with 3.7% formaldehyde in PBS for 10 minutes. Plates were washed twice in PBS, incubated in 50% acetone/50% ethanol for 30 s, and washed with PBS. Cells were stained for TRACP using a kit and following the manufacturer's instructions (Product 435; Sigma, St Louis, MO, USA). Osteoclasts were defined by TRACP+ staining if they had more than three nuclei. The number of TRACP+ cells per well was determined by light microscopy.

Actin rings:

Osteoclasts were fixed as above for TRACP staining. Plates were incubated with 0.5 ml of 0.1% Triton X-100/PBS for 20 minutes to permeabilize the cells. Twenty microliters per well of Alexa Fluor phalloidin (Molecular Probes, Eugene, OR, USA) was added, and the plates were incubated in the dark at room temperature for 30 minutes. Plates were washed with PBS and analyzed by fluorescent microscopy.(16)

Resorption assays:

BMMs, harvested as described above, were plated in 96-well tissue culture plates at a density of 1 × 105 cell/well in 200 μl of α-MEM with 10 ng/ml M-CSF and incubated overnight. The following day, nonadherent cells were collected and cultured on dentine slices (ALPCO Diagnostics, Windham, NH, USA) in α-MEM with 10% FBS, 10 ng/ml M-CSF, and 50 ng/ml RANKL for 10 days. The media were changed every 3 days. Dentine discs were fixed with 3.7% formaldehyde for 30 minutes, and cells were gently removed with 0.25% ammonium hydroxide and mechanical agitation. The dentine slices were stained with 1% toluidine blue in 0.5% sodium tetraborate solution. Resorption pits were analyzed by light microscope. Total resorption area in each dentine slice was evaluated with a light microscope using OsteoMeasure software (Osteometrics, Atlanta, GA, USA).

RNA isolation and time series RT-PCR:

Total RNA was extracted from unstimulated nonadherent cells (day 0) or from the adherent cell populations on days 1-7(16) with TRIzol (Invitrogen, Carlsbad, CA, USA) following the manufacturer's instructions. The isolated RNA pellet was air-dried and resuspended in 1 mM EDTA. RT reactions used 1 μg oligo dT (Amersham Biosciences, Piscataway, NJ, USA), 5 μg total RNA, and 20 μl H2O. The samples were subjected to a denaturation step for 5 minutes at 95°C and placed at 42–55°C for 20 minutes to anneal primer. The samples were filled to 30 μl with DEPC water containing 1 μl of Ribonuclease Inhibitor Transcriptase (Rnasin; Roche Dianostics), 2 μl 10 mM dNTP, 1 μl avian myeloblastosis virus RT, and 10 μl 5× RT buffer and incubated for 1 h at 42°C. Subsequently, DNA was denatured with 0.5 μl 0.5 M EDTA, 2 μl 1 M NaOH, boiled for 3 minutes, and precipitated by addition of 2 μl 1 M HCl, 5.5 μl 3 M NaOAc, and 145 μl ETOH followed by placement at −70°C for 15 minutes. Precipitated cDNA was washed with 70% ethanol, air-died, and resuspended in 50 μl TE. Two-microliter aliquots of cDNA or RT(−) controls were used per 30-μl PCR amplifications. Each 30-μl PCR reaction used 1 μl 10 mM dNTPs, 1 μM each primer, 0.5 μl DNA Taq polymerase (Roche, Indianapolis, IN, USA), and 3 μl 10× buffer. A Perkin Elmer GeneAmp Cycler was used with the following protocol: an initial step of 10 minutes at 94°C, denaturation at 94°C for 30 s, annealing at 50°C for 30 s, and extension at 72°C for 30 s for 30 cycles except for GAPDH (26 cycles). Aliquots of PCR products were supplemented with a loading buffer (5% glycerol, 10 mM EDTA, 0.025% bromphenol blue dyes) and fractionated on 2% agarose gel. The following forward and reverse primers were used: integrin β3 (β3), 5′-GCTCATTGGCCTTGCTACTC-3′ and 5′-TGTCCCACTTAGCTCTGGCT-3′; calcitonin receptor (CTR), 5′-ACCGACGAGCAACGCCTACGC-3′ and 5′-GCCTTCACAGCCTTCAGGTAC-3′; cathepsin K, 5′-ACGGAGGCATCGACTCTGAA-3′ and 5′-GATGCCAAGCTTGCGTCGAT-3′; and GAPDH, 5′-ACGACAGTCCATGCCATCAC-3′ and 5′-TCCACCACCCTGTTGCTGTA-3′.

Flow cytometry

BMMs were collected as described above. One million cells from each group were aliquoted into tubes and incubated with 2.4 G2 blocking antibodies on ice for 30 minutes. Cells were incubated with FITC-labeled rat antimouse monoclonal CD11b antibody (Miltenyi Biotec, Auburn, CA, USA) for 10 minutes in the dark at 4°C. Cells were washed, resuspended in 300 μl PBS, and placed on ice until they were subjected to FACS analysis in a Becton Dickinson FACSan with Cell Quest software without gating.(17)

BMM apoptosis assay

BMMs were cultured with 100 ng/ml M-CSF for 3 days. Pre-osteoclasts were lifted with 1× Trypsin/EDTA (Invitrogen) for 10 minutes. The reaction was stopped by 1× Trypsin Inhibitor Solution (Sigma). Cells (5 × 105) were replated in 6-well plates coated with type I collagen (Cohesion Technology, Palo Alto, CA, USA) or 2 μg/ml vitronectin (Sigma) in serum-free α-MEM containing 1 ml of M-CSF for 16 h and suspended in a Teflon beaker (Nalgene). The apoptosis rate was measured by the Cell Death Detection ELISA Plus Kit (Roche Diagnostics).(18)

Biochemical markers of bone turnover

Urinary levels of deoxypyridinoline cross-links and creatinine (DPD/Cr) were measured using ELISA kits from Quidel (Mountain View, CA, USA).(19,20) Serum levels of osteocalcin were measured using a mouse sandwich ELISA kit from Biomedical Technologies (Stroughton, MA, USA). The manufacturers protocols were followed, and all samples were assayed in duplicate. A standard curve was generated for each kit, and the absolute concentrations were extrapolated from the standard curve. The interassay and intra-assay CVs in our laboratory for the ELISA assays for DPD/Cr were <10% and for osteocalcin were 4-8%.(19,20) These results are similar to those reported by the manufacturers and other laboratories.


The right proximal tibial metaphyses were scanned by μCT (μCT-20; Scanco Medical, Bassersdorf, Switzerland). For image acquisition, the tibias was placed in a 17-mm holder and scanned. The image consisted of 200 slices with a voxel size of 9 μm in all three axes. A evaluation of trabecular bone structural parameters was done in a region that consisted of about 120 slices starting at about 0.1 mm from the lowest point of the growth plate and moving distally. Cancellous bone was separated from the cortical shaft by semiautomatically drawn contours. Images were segmented using a low-pass filter to remove noise, and a threshold was generated to edit the regions of cancellous bone from cortical bone.(21,22) This same threshold setting was used for all samples. The following 3D parameters were measured: bone volume, bone surface, and surface-to-volume ratio. A 3D cubical voxel model of bone was built, and the following calculations were made: relative bone volume over total bone volume (BV/TV), trabecular number (Tb.N), thickness (Tb.Th), separation (Tb.Sp), and connectivity density (Conn. Dens.). The diameter of spheres filling the structure was taken as Tb.Th, the thickness of the marrow spaces as Tb.Sp, and the inverse of the mean distances of the skeletal structure was calculated as Tb.N. By displacing the surface of the structure into infinitesimal amounts (dr), the structure model index (SMI) was calculated as SMI = 6 × (BV × dBS/dr)/BS2. The SMI quantifies the plate versus rod characteristics of trabecular bone, in which an SMI of 0 pertains to a purely plate-shaped bone, an SMI of 3 designates a purely rod-like bone, and values between stand for mixtures of plates and rods.(21-24) These methods have been used and previously published by our research group.(19,20)

Bone histomorphometry

The right proximal tibial metaphyses were dehydrated in ethanol, embedded undecalcified in methylmethacrylate, and sectioned longitudinally with a Leica/Jung 2065 microtome at 4- and 8-μm-thick sections. Bone histomorphometry was performed using a semiautomatic image analysis OsteoMeasure system (OsteoMetrics) linked to a microscope equipped with transmitted and fluorescence light.(20,25)

Osteoclast numbers (Oc.N) and osteoclast surface (Oc.S) were measured from TRACP-stained decalcified slides. With epifluorescent illumination, single- (sL.Pm) and double-labeled perimeter (dL.Pm) and interlabel width (Ir.L.Wi) were measured from the unstained undecalcified slides. These indices were used to calculate mineral apposition rate (MAR) and mineralizing surface (MS/BS). Bone volume-based bone formation rate (BFR/BV) was calculated by multiplying mineralizing surface (single labeled surface/2 + double labeled surface) according to Parfitt et al.(26,27) We have reported similar methodology in other experiments in our laboratory.(20)

Statistical analysis

Data on animal weights, biochemical markers of bone turnover, trabecular architecture by μCT and bone histomorphometry, TRACP+ cells, and resorption pit area are expressed as group means ± SD. Differences between the experimental groups at the same time-point were analyzed using one-way ANOVA. Differences within and between groups at different time-points (body weights and DPD/Cr cross-link excretion) were analyzed by repeated measurement of ANOVA with a grouping factor. The repeated measure was time, and the experimental group was the grouping factor. Posthoc analysis was performed using Tukey's method (SPSS, Chicago, IL, USA). In all of the analyses, p < 0.05 was considered statistically significant.


All mice tolerated the surgery without complications and appeared healthy. The animals in all four treatment groups gained 2-4 g during the experimental period (Table 1).

Because other investigators have reported previously that ex vivo osteoclastogenesis of marrow derived from β3−/− mice is arrested,(9) we first determined if the same occurs in cells from β5−/− mice. Osteoclast precursors, in the form of BMMs, were cultured in the presence of M-CSF and RANKL for 7 days, after which they were fixed and stained for TRACP activity. OVX increased osteoclastogenesis ex vivo in both WT and β5−/− genotypes (Fig. 1A; Table 2). However, osteoclast formation was substantially enhanced in cultures containing β5−/− cells, whether obtained from sham or OVX animals (Table 1). Interestingly, the osteoclast in cultures from β5−/−-Sham BMMs were similar to those from WT-OVX cells. Whereas all osteoclasts were spread and contained well-developed actin rings, β5−/− polykaryons seemed somewhat larger than their WT counterparts (Fig. 1B).

Figure FIG. 1.

(A) TRACP staining. In vitro generation of multinucleated TRACP+-stained cells. The β5−/− animals formed significantly more TRACP+ cells than the WT-Sham group (p < 0.05). Equal numbers of TRACP+ cells were formed in the β5−/− Sham and WT-OVX groups. However, β5−/−-OVX formed significantly more TRACP+ cells than the other three experimental groups (p < 0.05). In addition, the mature osteoclasts from β5−/−-OVX group were larger. Table 2 reports the mean ± SD for the TRACP+ cells for all experimental groups. These data are representative of five experiments. (B) Actin ring formation. Osteoclasts from all experimental groups exhibited actin ring formation. However, the osteoclasts from the β5−/−-OVX group formed both larger and more actin rings than the three other experimental groups.20

Table Table 2.. TRACP+ Multinucleated Cells/Well Cultured From Bone Marrow Monocytes/Macrophages After 6 Weeks of Sham-operation or OVX
original image

The increase in the number and size of osteoclasts derived from β5−/− mice, as well as their capacity to spread and form actin rings, suggests that net resorptive activity in cultures of these cells might also be enhanced. To examine this possibility, osteoclasts were maintained on dentine slices for 10 days. At termination of the culture, dentine resorption pits were photographed, and the magnitude of degradation was quantified histomorphometrically. In all circumstances, osteoclasts formed well-demarcated resorption lacunae (Fig. 2A). Mirroring their abundance and size, the net resorptive activity of osteoclasts generated from OVX β5−/− mice was ∼2-fold greater than similarly treated WT-OVX cells (p < 0.05) and 4-fold greater than the β5−/−-Sham group (p < 0.05; Fig. 2B). There were no significant differences in the resorptive capacity of osteoclasts generated from Sham-operated animals of either genotype.

Figure FIG. 2.

Dentine resorption assays. Resorption pits (black arrows) on the dentine surface show that the osteoclasts are functional. The β5−/−-OVX group seemed to have more dentine resorption pits compared with the three other groups. These data are representative of three experiments. (B) Quantitation of the osteoclast resorption pit area on the dentine slices was evaluated using OsteoMeasure software and expressed as the percentage of resorption pit area/total dentine slice. The resorbed pit area was significantly greater in both the WT-OVX and the β5−/−-OVX animals compared with WT-Sham and β5−/−-Sham groups.ap < 0.05 from WT-Sham group;bp < 0.05 from WT-Sham, WT-OVX, and β5−/−-Sham groups. These data are representative of five experiments.20

The data presented thus far raised the possibility that the absence of the β5 integrin subunit might accelerate osteoclast differentiation. To directly examine this possibility, BMMs were harvested 42 days after OVX and maintained in M-CSF and RANKL. The cultures were terminated daily for 7 days, RNA was extracted, and mRNA markers of commitment to the osteoclast phenotype was measured by RT-PCR. The appearance of mRNA coding for the osteoclast-specific proteins cathepsin K, the β3 integrin subunit, and the calcitonin receptor were accelerated in both the β5−/-Sham and β5−/−-OVX cultures (Fig. 3). However, there seemed to be a lower amount of cathepsin K and β3 integrin subunit in the β5−/−-Sham cultures throughout the 7-day experiment.

Figure FIG. 3.

Molecular characterization of osteoclast maturation by serial RT-PCR. Cathepsin K (Cath K), calcitonin receptor (CTR), integrin β3 (β3), and GAPDH (internal control) were all present by day 7 in all experimental groups. However, earlier maturation of osteoclasts in the β5−/−-OVX group was found, with β3, cath K, and CTR present by day 3, which was at least 2 days earlier than in the WT-OVX group. The β5−/−-Sham group had evidence for early osteoclast maturation with the presence of β3 developing 1-2 days before the WT-Sham group, but the expression of both β3 and the other markers were very weak. These data were representative of more than five experiments.20

We established that the increased number of osteoclasts in β5 integrin subunit reflects, at least in part, stimulated precursor differentiation. We further investigated whether stimulated macrophage proliferation or prolonged lifespan of the differentiated osteoclasts may also be contributing factors. Thus, equal numbers of BMMs from all four experimental groups were subjected to FACS analysis using an antiCD11b mAb, which identifies cells of the macrophage lineage. There was an increase in the number of the bone marrow macrophages in the OVX animals lacking β5 (p < 0.05; Fig. 4A) compared with all other experimental groups. Moreover, the presence of histone-associated DNA fragments, a marker of apoptosis, was less in osteoclasts generated in vitro from both the β5−/−-OVX and Sham-operated animals compared with WT groups (p < 0.05; Fig. 4B). Therefore, the greater numbers of osteoclasts present in β5−/− mice most likely reflects a combination of an increased number of osteoclast precursors and or prolonged osteoclast lifespan.

Figure FIG. 4.

(A) Flow cytometric analysis of isolated marrow cells showed that CD11b+ cells were about 20% higher in the β5−/−-OVX group compared with both WT groups and the β5−/−-Sham group.bp < 0.05 from WT-Sham, WT-OVX, and β5−/−-Sham groups. (B) Osteoclast apoptosis assay. The apoptotic rate was about 25% lower in the β5−/− osteoclasts compared with WT cells.ap < 0.05 from WT-Sham group;bp < 0.05 from WT-OVX group.20

Because deletion of the β5 integrin subunit enhanced osteoclastogenesis ex vivo, we examined if the same is true in vivo. We prepared undecalcified histological sections of proximal tibia and quantitated the abundance of osteoclasts (Table 3). As expected, ovariectomy increases osteoclast number. Moreover, like the ex vivo observations, deletion of the β5 integrin enhanced osteoclast number in the sham and OVX mice (p < 0.05) compared with the WT-Sham. In addition, total osteoclast resorptive activity assessed by urinary DPD/Cr cross-link excretion was accelerated in both WT-OVX and β5−/− OVX mice and was significantly different from baseline values within the OVX groups and from both sham-operated groups (p < 0.05) by 21 days after OVX (Table 4).

Table Table 3.. Histomorphometric Assessment of Trabecular Bone Turnover of the Proximal Tibial Metaphyses*
original image
Table Table 4.. Changes in Biochemical Markers of Bone Turnover: Deoxypyridinoline/Creatinine Crosslinks
original image

The rates of bone formation at either the cellular or tissue levels are similar in the two groups of sham mice (Table 3). The coupling of osteoclast and osteoblast function in the remodeling process of bone formation, like bone resorption, was accelerated in OVX animals. This finding was reflected by bone histomorphometry assessment of BFR/BV and serum osteocalcin measurements. Interestingly, osteocalcin levels, a marker of bone formation, were significantly higher in β5−/−-OVX mice than in WT-OVX mice (Table 5).

Table Table 5.. Changes in Biochemical Markers of Osteocalcin
original image

To determine whether the increased bone resorption observed ex vivo in β5−/−-OVX mice altered trabecular bone architecture in vivo, the proximal tibial metaphyses from all animals underwent quantitative μCT (Fig. 5A). In WT-OVX animals, compared with the WT-sham group, total trabecular volume (BV/TV) decreased by 41%, and trabecular number decreased by 24% (all p < 0.05) after 42 days after OVX (Fig. 5B). However, trabecular and cortical thickness were not different between the WT-OVX and WT-Sham groups at day 42 after OVX. The SMI, which reflects the pattern of trabeculae (combination of rods and plates), was 18% higher in WT-OVX compared with WT-Sham mice (p < 0.05; data on file).

Figure FIG. 5.

(A) μCT images of trabecular bone of the proximal tibial metaphyses. Whereas both WT-OVX and β5−/−-OVX mice had significantly less bone volume than the WT and β5−/−-Sham groups, β5−/−-OVX had significantly less trabecular bone volume, number, and connectivity density than WT-OVX mice. (B) Quantification of trabecular bone structure of the right proximal tibial metaphyses obtained with μCT. Trabecular bone volume (BV/TV), trabecular number (Tb.N.), and trabecular connectivity (Tb. Conn) were significantly reduced in both WT-OVX and β5−/−-OVX compared with WT-Sham and β5−/−-Sham mice. However, the BV/TV in β5−/−-OVX group was 70% lower than β5−/−-Sham and significantly lower than the all of the other experimental groups.ap < 0.05 from WT-Sham group;bp < 0.05 from WT-Sham, WT-OVX, and β5−/−-Sham groups.20

Compared with the β5−/−-Sham group, the β5−/−-OVX animals had 70% less total trabecular bone volume (BV/TV), 84% less trabecular bone connectivity density, and 35% for trabecular number (all p < 0.05; Figs. 5A and 5B). The SMI increased by 36% in the β5−/−-OVX group compared with the β5−/−-Sham group but was not significant (data on file).


Osteoclasts or their precursors express three major β integrin subunits, namely β3, β1, and β5. Whereas the collagen receptor, α2β1, is posited to be a significant participant in the resorptive process,(28,29) data suggest αvβ3 as the principal attachment molecule regulating osteoclast function. Specifically, the β3 integrin subunit knockout mouse is osteosclerotic because of dysfunctional osteoclasts.(9) Thus, attention has turned to αvβ3 blockade as an anti-osteoporosis strategy, and this strategy is currently being evaluated in clinical trials.(7) The success of αvβ3 blockade in preventing bone loss underscores the importance of identifying and targeting other matrix receptors residing on osteoclasts and/or their precursors.

The β5 subunit, which associates only with αv, is 56% homologous to β3 and establishes the pair as the two most closely related integrin β subunits.(12) Because of this homology and the fact that both αvβ3 and αvβ5 are expressed on osteoclasts (albeit at different stages of development), we asked whether, like αvβ3, αvβ5 contributes to osteoclast differentiation and function. This issue has important therapeutic implications given that both integrins recognize the same bone-residing ligands such as osteopontin and bone sialoprotein(13, 14, 30) and thus may be similarly impacted by blocking agents mimicking the RGD motif. αvβ5 is present on immature osteoclast precursors, but it disappears as the cells differentiate into resorptive polykaryons in the presence of RANKL and is thus a negative marker of osteoclast commitment.(13) On the other hand, the integrin persists in the presence of M-CSF alone and is a component of non-osteoclastogenic macrophage differentiation.(13) Its functional importance in macrophages is suggested by the fact that it, and not αvβ3, regulates their spreading on bone matrix.(12,14)

Unlike mice lacking β3, ovary-intact β5−/− mice seem to have a relatively normal bone phenotype. Interestingly, when placed in culture, BMMs from both β5−/−-Sham and -OVX mice generate more osteoclasts than do their WT counterparts, and as evidenced by expression of markers of osteoclastogenesis, do so more rapidly. In addition, the β5−/−-Sham-derived precursors generate more osteoclasts in vitro and animals exhibit more osteoclasts in vivo than the WT-Sham animals in both TRACP staining and histomorphometric assessment of osteoclasts on the trabecular bone surface. However, these osteoclasts from the β5−/−-Sham animals did not actively resorb bone. Most likely, additional factors such as cytokines released after OVX are required to activate these osteoclasts in β5−/−-Sham animals to resorb bone.(31)

Postulating that a state of accelerated bone remodeling may uncover changes in osteoclast function obscured in basal conditions, we subjected β5−/− mice to OVX. We discovered that αvβ5 actually prevents osteoclast maturation. Osteoclast precursors, regardless of genotype, pre-exposed to an estrogen-deplete environment, in vivo, differentiated more rapidly into increased numbers of mature resorptive polykaryons. Genes characteristic of mature osteoclasts such as the calcitonin receptor, the β3 integrin subunit, and cathepsin K were expressed much more rapidly in OVX β5−/− marrow cells after exposure to M-CSF and RANKL.

β5 integrin deletion enhanced the capacity of individual osteoclasts to degrade matrix, ex vivo and in vivo, but did so only in the context of estrogen deprivation. Whereas the number of osteoclasts, in vivo, was somewhat greater in OVX β5−/− compared with similarly treated WT mice, bone degradation was accelerated as evidenced by urinary excretion of DPD/Cr. Also, 6 weeks after OVX, there was significantly greater trabecular bone loss and acceleration of erosion of trabecular plates compared with estrogen-depleted WT mice. Interestingly, TNFα, which is produced by T-lymphocytes in OVX mice and is essential for their accelerated loss of bone,(18) also inhibits αvβ5 expression by osteoclast precursors.(14) This observation, taken with the osteoclast-inductive effect of β5 deletion, raises the possibility that the osteoporosis caused by estrogen deprivation reflects, at least in part, suppression of the αvβ5 integrin by TNFα. Hence, analogous to αvβ3 and M-CSF, there seems to be interplay between αvβ5 and cytokines expressed in abundance with OVX.

Increased numbers of osteoclasts may reflect enhanced precursor proliferation or reduced apoptosis. The abundance of osteoclasts in β5−/−-Sham and -OVX mice seems to reflect these two events. Compared with their WT counterparts and the β5−/−-Sham animals, macrophage numbers were higher in β5−/−-OVX mice. Because BMMs cease to express αvβ5 on exposure to RANKL,(32) we performed our apoptotic assay without RANKL so as to maintain αvβ5 in WT osteoclast precursors. Our results showed that cell death was reduced in β5−/− mice, suggesting that absence of β5 in an estrogen-depleted state may be associated with enhanced β3 expression and enhanced survival of pre-osteoclasts.(18)

The capacity of osteoclasts to degrade bone is reflective of a unique process of cytoskeletal organization. As the osteoclast contacts bone, it forms a seal that isolates the acidic resorptive microenvironment from the general extracellular space. This circumferential sealing zone, which consists predominantly of fibrillar actin, is visualized in the normal matrix-opposed osteoclast and is absent in a number of resorption defective mutants such as those lacking αvβ3.(9) Similarly, β3−/− osteoclasts fail to adequately form a ruffled membrane, another cytoskeletal structure that is the cell's resorptive organelle.(9) Similar to other hyper-resorptive osteoclasts, such as those derived from mice lacking SHIP1(33) or patients with Paget's disease of the bone, those derived from β5−/− mice are enlarged and have well-developed cytoskeletons.

This report documents that the absence of the β5−/− integrin subunit directly accelerates differentiation of osteoclast precursors. Moreover, mice lacking the β5 integrin, when made estrogen deficient, have a hyper-resorptive phenotype, ex vivo and in vivo. Thus, additional studies that elucidate the mechanisms surrounding the β5 integrin's role in osteoclast precursor maturation will further our understanding of osteoclastogenesis.


This work was supported by NIH Grants 1R01AR43052 and AR07304, an HHMI UCSF Institutional grant to DS, the Rosalind Russell Arthritis Research Center, and NIH Grants AR32788, AR46523, and AR48853 to SLT and AR46852 and AR48812 to FPR.