Loss of skeletal weight bearing or skeletal unloading as occurs during spaceflight inhibits bone formation and stimulates bone resorption. These are associated with a decline in the osteoblast (Ob.S/BS) and an increase in the osteoclast (Oc.S/BS) bone surfaces. To determine the temporal relationship between changes in the bone cells and their marrow precursor pools during sustained unloading, and whether genetic background influences these relationships, we used the hindlimb unloading model to induce bone loss in two strains of mice known to respond to load and having significantly different cancellous bone volumes (C57BL/6 and DBA/2 male mice). Skeletal unloading caused a progressive decline in bone volume that was accompanied by strain-specific changes in Ob.S/BS and Oc.S/BS. These were associated with a sustained reduction in the osteoprogenitor population and a dramatic but transient increase in the osteoclast precursor pool size in both strains. The results reveal that bone adaptation to skeletal unloading involves similar rapid changes in the osteoblast and osteoclast progenitor populations in both strains of mice but striking differences in Oc.S/BS dynamics, BFR, and cancellous bone structure. These strain-specific differences suggest that genetics plays an important role in determining the osteoblast and osteoclast populations on the bone surface and the dynamics of bone loss in response to skeletal unloading.
- skeletal unloading
- bone loss
- bone remodeling
loss of weight bearing during spaceflight and simulated spaceflight induces osteopenia in both young growing and adult animals (3, 18, 29, 32). Skeletal unloading inhibits bone formation, decreases cancellous bone volume and cortical thickness, and promotes the accumulation of marrow fat (4, 14, 30). These changes are associated with a reduction in the number of osteoblasts on the bone surface (Ob.S/BS) (8, 11, 27). In adult human bed rest studies and during space flight, bone formation and Ob.S/BS are also reduced (17, 31).
Studies in mice and rats after 10 and 14 days of unloading suggest that the unloading-induced decline in Ob.S/BS is associated with a reduction of calcium nodule formation in bone marrow stromal cell cultures (2, 10, 16, 22, 26). This amounts to a reduction in the size of the osteoprogenitor pool.
The effects of skeletal unloading on bone resorption are less clear. In the mouse and rat, unloading has been reported to increase the number of osteoclasts on the bone surface (Oc.S/BS) and bone resorption (7, 25, 32) or to have no effect (2, 8, 11). In adult human bed rest studies and in astronauts, there is general agreement that resorption increases rapidly with unloading or spaceflight. Resorption markers increase significantly within a few days after loss of weight bearing and remain elevated for the duration of unloading (12, 17, 31, 33). Bone biopsies from subjects in these studies have shown that both cancellous and cortical bone-eroded and osteoclast surfaces increase with loss of weight bearing (31, 33). Studies in mice and rats after 10 and 14 days of unloading suggest that the reported increase in Oc.S/BS is associated with an increase in the size of the osteoclast precursor pool (2, 10, 26). The effects of unloading or loss of weight bearing on bone are rapid and similar between astronauts and mice in our unloading model. Loss of bone in other pathologies such as aging is slower, and the changes in Ob.S/BS and Oc.S/BS are smaller.
The response of bone to skeletal load in mice is strain specific and may vary in proportion to bone mass (1, 13, 15, 24, 30). Mechanical loading increases cortical bone area and periosteal bone formation in the lower bone mass DBA/2 and C57BL/6 strains but not in the high bone mass C3H strain (15, 24). Comparison of the mechanical strain response in the ulna to axial loading shows the DBA/2 mouse to be most sensitive (24). Studies to examine the effects of genetic background on the response of bone to spaceflight or unloading in humans are very limited. It appears that some astronauts lose more bone than others (19, 20), suggesting that genetic components are playing a role in how bone responds to unloading.
Collectively, the unloading-induced changes in Ob.S/BS and Oc.S/BS appear to be linked to changes in the progenitor populations in the bone marrow and may be influenced by mouse strain. To extend these studies, we examined the temporal changes in the Ob.S/BS and Oc.S/BS and their relation to the osteoprogenitor and osteoclast precursor populations following 7, 14, and 28 days of skeletal unloading. To assess the influence of genetic background on the changes in bone induced by unloading, we studied two mouse strains (C57BL/6 and DBA/2) known to respond to loading/unloading and with differing bone volumes [DBA/2, femoral metaphyseal bone volume/total volume (BV/TV) = 6% and C57BL/6, BV/TV = 15%].
MATERIALS AND METHODS
Male C57BL/6 and DBA/2 mice, 6 mo of age, were obtained from the National Institute of Aging colony of aging rodents (Harlan Laboratories, Bethesda, MD). The animal protocol for these studies was approved by the Animal Care and Use Committee at the Veterans Affairs Medical Center, San Francisco, CA. Animals were maintained and processed in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. Because not all mouse strains respond to loading/unloading, suggesting that genetic factors contribute to unloading-induced bone loss, we chose to study two strains that have been shown to respond to load and that have different cancellous bone volumes.
Skeletal unloading was induced using the hindlimb elevation model (9, 32). Animals (n = 8/group) of each strain were submitted to hindlimb unloading for 1, 2, and 4 wk. Control normal ambulatory mice were housed in similar suspension cages, and their food intake was controlled by pair feeding. All mice were weighed daily. To measure trabecular bone formation rates, demeclocycline (20 mg/kg) and calcein (15 mg/kg) were injected subcutaneously at 7 and 2 days, respectively, before the animals were euthanized to label bone-mineralizing surfaces. Animals in the 2- and 4-wk suspension groups were given an additional calcein label at the time of suspension to measure periosteal bone formation at the tibular-fibular junction. Total bone area, cortical area, and medullary area were also measured.
At the time of euthanasia, the left femurs were collected and processed for microcomputed tomography (μCT) analysis, and the right femur was processed for bone histomorphometry. Marrow was collected from the tibiae for cell culture and the left tibial diaphysis was prepared for measurement of periosteal bone formation rate.
Bone marrow stromal and osteoclast cell cultures.
To assess the effects of skeletal unloading on osteoprogenitor number, bone marrow stromal cells were prepared from the tibiae and cultured as described previously (5). On day 28 the cultures were washed, fixed in 3.7% phosphate-buffered formalin, and stained with 2% Alizarin red (Sigma-Aldrich, Milwaukee, WI). Dark red-stained colonies >1 mm in diameter were counted as calcified nodules. Each nodule was taken as representing a single osteoprogenitor.
To assess the effects of unloading on the number of osteoclast precursors in the bone marrow, primary bone marrow cells were cultured as described previously (6). On day 2 of culture, the nonadherent cell fraction was removed and washed twice with phosphate-buffered saline (PBS). The cells were suspended in PBS, counted using a hemocytometer, and seeded into 24-well tissue culture plates at 1 × 106 cells/well. The cultures were carried for 6 days in the stromal cell culture medium supplemented with receptor activator for NF-κB ligand (40 ng/ml) and macrophage colony-stimulating factor (10 ng/ml). Media were changed three times/wk, and on the 6th day cells were washed twice with PBS, fixed in citrate/formaldehyde solution, and stained for tartrate-resistant acid phosphatase (TRAP; Sigma Diagnostics, St Louis, MO). Dark reddish-purple cells were counted as TRAP positive (TRAP+). TRAP+, multinucleated (>3 nuclei) cells were counted as osteoclasts. Cells containing >30 nuclei were considered mature osteoclasts.
The right distal femurs were fixed in 10% phosphate-buffered formalin, dehydrated in increasing concentrations of ethanol, and defatted in xylene. Bones were then embedded in methyl methacrylate, and 4- and 10-μm-thick sections were cut longitudinally using a Leica RM2165 microtome. The 10-μm sections were not stained and were used for fluorometric analysis to measure bone formation rate. The 4-μm sections were deplastified with 2-methoxyethyl acetate and stained according to Von Kossa method with tetrachrome counterstain (Sigma-Aldrich). Cancellous bone measurements were made in the region of interest (ROI), defined as the cancellous bone compartment beginning 0.5 mm proximal to the growth plate and extending proximally 1.5 mm. This ensures exclusion of the primary spongiosa from the analyses. Two longitudinal sections taken from the center of the bone were examined from each animal, resulting in histomorphometric measurements along 30–40 mm of cancellous bone perimeter. Sections were viewed at ×200 magnification using a Zeiss Imager M1, and pictures were taken at ×200 magnification with a Zeiss Axiocam HRC. Single- and double-labeled perimeters and interlabel widths were used to calculate mineralizing surface/bone surface (MS/BS) and mineral apposition rate (MAR) and surface-based bone formation rate (BFR/BS = MAR × MS/BS; μm3·μm2·day−1). The histomorphometric parameters and rate of bone formation were derived as recommended by the American Society for Bone Mineral Research Histomorphometry Nomenclature Committee (21). Bone structure (total bone area, cortical bone area, and medullary area) and bone formation rate were measured in histological sections at a site 1 mm proximal to the bifurcation of the tibular-fibular junction. All samples were evaluated using Bio-Quant Osteo analysis software (Bio-Quant, Nashville, TN).
The left femurs were cleaned of adherent soft tissue, fixed in 10% phosphate-buffered formalin (24 h), and kept in 70% ethanol. The distal femoral metaphysis was analyzed using μCT (Scanco ViviaCT 40; Scanco Medical, Zurich, Switzerland). The ROI was the same as that used for bone histomorphometry. A threshold was determined as 22% of the maximal gray scale and applied to differentiate bone from soft tissue (lean and fat marrow). BV/TV, trabecular number (Tb.N; 1/mm), trabecular thickness (Tb.Th; μm), trabecular separation (Tb.Sp; μm), connectivity density (1/mm3), structure model index (SMI: ranges from 0–3; 0 = plate like and 3 = rod like), and bone mineral density per bone volume (BMD/BV) and tissue bone mineral density (BMD/TV; mg hydroxyapatite/cm3) were measured.
Data are presented as means ± SD and analyzed using a t-test between loaded and unloaded mice within each time point for each strain. All analyses were performed using SigmaStat 3.0 software (SPSS, Chicago, IL).
Mean body weights of unloaded mice were not significantly different from normally loaded mice at any time during the experiment (Fig. 1).
After 1 wk of skeletal unloading, bone marrow stromal cell cultures from C57BL/6 and DBA/2 mice showed significantly decreased mineralized nodule formation compared with normally loaded mice (−70 and −60%, respectively) (Fig. 2). By 2 and 4 wk, cultures from unloaded animals in both strains had one-half as many calcified nodules (osteoprogenitors) as normally loaded animals (P < 0.001). Skeletal unloading produced a rapid and sustained loss of osteoprogenitors.
The number of mature osteoclasts (>30 nuclei, TRAP+ cells) produced from nonadherent cell cultures increased three- to fourfold after 1 wk of unloading in both C57BL/6 and DBA/2 mice (Fig. 3). Osteoclast-induced erosion of mineral from dentin disks confirmed these observations (Fig. 3B). By 2 and 4 wk the number of mature osteoclasts had returned to normal (Fig. 3C). The total number of osteoclasts (>3 nuclei, TRAP+ cells) also increased transiently and then returned to normal by 2 wk (data not shown).
The Ob.S/BS in C57BL/6 was decreased after 2 wk of unloading and remained suppressed at 4 wk (Fig. 4A). In the DBA/2 mouse there was a small decrease (25%) in Ob.S/BS at 2 wk only. The Oc.S/BS in C57BL/6 mice was increased after 1 and 2 wk and then returned to normal by 4 wk (Fig. 4B). In the DBA/2 mouse, Oc.S/BS did not increase until 4 wk after unloading (Fig. 4B).
Cancellous mineralizing surface (MS/BS) and MAR in the distal femur decreased in response to unloading only in the DBA/2 mice after 2 and 4 wk unloading and were basally much higher in DBA/2 than C57BL/6 mice (37.5 ± 6.5 vs. 25.6 ± 6% and 1.5 ± 0.2 vs. 1.1 ± 0.15 μm/day, respectively). The cancellous bone formation rate (BFR/BS) was not affected by unloading in C57BL/6 mice but fell significantly in DBA/2 mice at 2 wk unloading (Fig. 5). Unloading did not produce a distinguishable change in cortical BFR measured at the tibial fibular junction, and no changes were seen in total bone area, cortical bone area, or medullary area (data not shown).
BV/TV decreased progressively in the distal femur during unloading in both C57BL/6 and DBA/2 mice (Fig. 6A). After 2 wk of unloading, BV/TV in C57BL/6 and DBA/2 mice had decreased by 26 and 17%, respectively. By 4 wk, BV/TV was 44 and 35% lower in unloaded than normally loaded mice in C57BL/6 and DBA/2 mice, respectively. Tb.Th also decreased progressively during unloading in both C57BL/6 and DBA/2 mice (Fig. 6B). By 4 wk Tb.Th had decreased by 22 and 19% in C57BL/6 and DBA/2 mice, respectively. Tb.N decreased with unloading in C57BL/6 but not DBA/2 mice (Fig. 6C). At 4 wk, Tb.N in DBA/2 unloaded mice was 11% higher (P = 0.06). Trabecular spacing (Tb.Sp) was not affected by unloading in C57BL/6 mice but decreased in DBA/2 mice (Fig. 6D). The SMI was not affected by unloading but was greater in DBA/2 than C57BL/6 mice (2.5 ± 0.2 and 2.0 ± 0.3, respectively, P < 0.05). Both C57BL/6 and DBA/2 hindlimb unloaded mice showed a decrease in BMD/BV by 4 wk of unloading (Fig. 6E). Overall, BMD/TV decreased progressively in C57BL/6, but only at 2 wk of unloading in DBA/2 mice (Fig. 6F).
Skeletal unloading produced a rapid and sustained decrease in osteoprogenitor number in both C57BL/6 and DBA/2 mice, similar to that in studies by Sakata et al. (26) in mice and Basso et al. (2) in rats. The fall in osteoprogenitor number in the C57BL/6 mice was accompanied by a fall in Ob.S/BS, but the change in ObS/BS lagged the decline in osteoprogenitor number. This suggests that, although unloading may have a rapid and sustained effect on the osteoprogenitor pool, some time is required for a change in the progenitor population to become apparent in the osteoblasts on the bone surface. Although a similar relationship was observed in DBA/2 mice, a significant reduction in ObS/BS was achieved only at 2 wk. The data suggest that Ob.S/BS decreases with unloading in both strains but that the magnitude of change is much greater in the C57BL/6 than the DBA/2. In interpreting the relationship between the osteoprogenitor population and the Ob.S/BS, it is important to recognize that the cell population being used to assess osteoprogenitor number comes from a mixture of diaphyseal and metaphyseal marrow and may not accurately reflect the cellular makeup of the marrow in the distal femoral metaphysis.
Oddly, the marked decrease (30–50%) in Ob.S/BS in C57BL/6 mice did not carry over to a decrease in bone formation rate in the femoral metaphysis, whereas in the DBA/2 mice bone formation was clearly suppressed. It may be that in normally loaded C57BL/6 mice some of the osteoblasts on the bone surface are not actively forming bone. With unloading the Ob.S/BS decreases, but a greater proportion of the cells left on the surface are forming bone. In DBA/2, unloading may have more of an effect on osteoblast activity than on Ob.S/BS. Collectively, our findings suggest that the mechanisms of bone loss induced by unloading may differ between strains of mice.
The reduction in the size of the osteoprogenitor pool may be linked to sclerostin inhibition of Wnt signaling. Hindlimb unloading has been shown to increase sclerostin expression in bone, and studies have shown that treatment of skeletally unloaded mice with antibody neutralizing sclerostin can increase osteoprogenitor number (23, 28).
The rapid rise and fall in the total number of osteoclast precursors and number of mature osteoclasts induced in vitro from the nonadherent population of bone marrow stromal cell cultures coincided with a rise and fall in Oc.S/BS in the C57BL/6 mice. Changes in Oc.S/BS appear to be related to changes in the precursor pool size, at least in the C57BL/6 mouse. A similar relationship has been reported by Sakata et al. (26). In the DBA/2 mice, the same transient increase in osteoclast precursor number occurred, but this did not translate into a coincident increase in Oc.S/BS. On the contrary, the Oc.S/BS in DBA/2 mice increased progressively. This may be a consequence of strain-specific differences in the dynamics of osteoclast recruitment and processing. The data suggest that the changes in osteoclast recruitment and residency on the bone surface induced by unloading are in part genetically determined. The mechanisms responsible for the rapid but transient increase in osteoclast precursors are not evident and need further investigation. Interestingly, the changes in the osteoprogenitor and osteoclast precursor populations induced by unloading in C57BL/6 and DBA/2 mice are nearly identical despite the relatively large inherent differences in BV/TV.
Bone volume fell progressively with unloading, and this appears to reflect, for the most part, a decrease in trabecular thickness. Interestingly, unloading decreased Tb.N in the C57BL/6 but tended to increase Tb.N in DBA/2 mice at 2 and 4 wk (P = 0.06). This produced an unexpected decrease in Tb.Sp in DBA/2 mice. It appears that the deficit in bone associated with unloading is achieved in the C57BL/6 mouse by eroding trabecular surfaces and decreasing Tb.Th. As a consequence, the number of trabeculae (Tb.N) decreases. In the DBA/2 mice, trabeculae become thin, but the trabecular meshwork is maintained or enhanced.
The changes in BV/TV and BMD/TV induced by unloading are consistent with the changes in Ob.S/BS and Oc.S/BS. In the C57BL/6 mouse, the initial decrease in bone volume appears to be linked to a decrease in the Ob.S/BS and an increase in the Oc.S/BS. BFR was not affected in C57BL/6. Thus, the decrease in bone volume appears to be driven primarily by bone resorption, as judged by Oc.S/BS. In the DBA/2 mouse, the initial decrease in bone volume appears to be more linked to decreases in osteoblastic cell activity and decreased BFR than to changes in Oc.S/BS. After 2 wk, the continued loss of bone in the DBA/2 mouse appears to be related more to Oc.S/BS despite normalization of the osteoclast precursor population. The differences in how C57BL/6 and DBA/2 mice adjust their bone mass in response to unloading are striking. In C57BL/6 mice unloading-induced bone loss appears to be sustained by suppressed Ob.S/BS and increased Oc.S/BS, whereas in DBA/2 mice continued loss appears to be sustained primarily by decreased bone formation. Collectively, these findings suggest that the mechanisms responsible for loss of bone during unloading are influenced by genetic background and basal bone mass. These results are consistent with previous studies showing strain-specific differences in bone mass and structure and response to mechanical strain (24, 30).
Trabecular bone mineral density (BMD/BV) was decreased in both strains of mice after 4 wk of unloading. This suggests that both primary and secondary phases of mineralization decrease when the skeleton experiences unloading. Given that over the 4 wk of unloading only a small fraction of the skeleton is remodeled in the adult mouse and that mineral density is reflective of the entire trabeculae of both the old and newly formed bone, the actual density of the newly formed bone in the unloaded skeleton may be even much smaller.
The bone-specific genetic differences between C57BL/6 and DBA/2 mice are noteworthy. DBA/2 mice have roughly one-half the bone volume of C57BL/6 mice, similar Ob.S/BS, roughly 50% greater Oc.S/BS, and a bone formation rate twofold greater. The dynamics of maintaining bone volume differ between strains. That SMI is greater (trabecular structure is more rod like) in DBA/2 mice may reflect a different accommodation to mechanical load/unload in the two strains of mice.
Body weight was not affected by unloading, suggesting that the changes in bone induced by unloading are not a consequence of weight loss or altered food intake.
In summary, our data define the temporal changes in the Ob.S/BS and Oc.S/BS in relation to their marrow progenitor cells as induced by skeletal unloading in two mouse strains and suggest that these changes may be linked to changes in the osteoprogenitors and osteoclast precursor populations and their appearance on bone surfaces. The strain-specific differences in the relationship between progenitors and bone surface cell populations suggest that genetic factors play an important role in determining how bone is lost in response to skeletal unloading.
This work was supported by NASA Grant NNA04CK55G, NIH Grant RO1-AR-055924, and a Veterans Affairs Merit Review Program.
No conflicts of interest, financial or otherwise, are declared by the authors.
M.S. analyzed the data; M.S., D.D.B., and B.P.H. interpreted the results of the experiments; M.S. prepared the figures; M.S., P.K., B.M.B., B.E.O., and B.P.H. drafted the manuscript; M.S., P.K., B.M.B., B.E.O., D.D.B., and B.P.H. edited and revised the manuscript; M.S., P.K., B.M.B., B.E.O., D.D.B., and B.P.H. approved the final version of the manuscript; P.K., B.M.B., and B.E.O. performed the experiments; B.P.H. did the conception and design of the research.