The senescence accelerated mouse strain P6 (SAMP6) has emerged as a useful model of senile osteoporosis because it has many features of the disease, including low trabecular bone formation and low areal bone density. We further characterized the SAMP6 model of senile osteoporosis by comparing morphological, mechanical, and densitometric properties of femurs and tibias from SAMP6 mice to those of the control strain (SAMR1) at 4 months and 12 months of age. SAMP6 long bones had increased periosteal width and endosteal area (p < 0.05), resulting in an average increase of 30% in moments of inertia (p < 0.05), but no difference in bone area (p > 0.05) compared with control. Despite their increased moments of inertia, long bones from SAMP6 mice were relatively weak and brittle. Ultimate bending moment was reduced by 25%, and both postyield displacement and energy-to-fracture were reduced by 60% compared with SAMR1 controls (p < 0.001). Average cortical ash fraction was increased slightly from 0.74 in SAMR1 to 0.76 in SAMP6 bones (p < 0.05), indicating that increased mineralization may have contributed to the brittleness of SAMP6 bones. The relative differences we observed—increased endosteal and periosteal dimensions, reduced bending strength, increased brittleness, and increased mineralization—are analogous to changes that occur in the aging human skeleton. Moreover, these features were consistently observed in young (4-month) and old (12-month) animals. These findings extend the previous descriptions of the SAMP6 mouse and identify key mechanical features that further validate its relevance as a unique and functionally relevant model of senile osteoporosis.
Senile osteoporosis is characterized by a deficit in bone formation relative to resorption that results in decreased bone mass and decreased bone strength with advanced age. Progress toward understanding the mechanisms underlying bone loss in senile osteoporosis has been slowed in part by a lack of appropriate animal models.(1) The senescence accelerated mouse strain P6 (SAMP6) has emerged as a useful model of senile osteoporosis because it has many of the salient features of the disease. Relative to the SAMR1 strain (the most appropriate control strain(2)), bones from SAMP6 mice are reported to have reduced cortical thickness(3,4) and whole-bone areal density.(2,5,6) The differences in areal density are evident at and beyond the age of peak bone mass, that is, from 4–5 months to 18–24 months.(5) In addition, the osteogenic potential of marrow cells from SAMP6 is reduced relative to SAMR1,(2,6) similar to changes that have been reported in aging humans(7,8) and mice.(9,10) This reduction in marrow osteogenesis corresponds to decreased trabecular bone formation in SAMP6 mice.(2,6)
Although the available data support the relevance of the SAMP6 mouse as a model of senile osteoporosis, there are several important issues that have not been addressed yet. First, there have been no reports of biomechanical data that would confirm the functional relevance of the SAMP6 model. One study did note the incidence of spontaneous fractures in tibias from 2-year old SAMP6 mice,(5) suggesting that SAMP6 mice have reduced skeletal strength. Second, the previous reports indicating reduced bone density in SAMP6 mice were based on areal density obtained from radiographs of whole bones(5) or from whole DXA.(2) Thus, the findings of low bone “density” in SAMP6 may be caused by in part differences in bone dimensions rather than differences in tissue density. Volumetric mineral density and ash content of bones from SAMP6 mice have not been reported.
Our objective was to characterize further the relevance of the SAMP6 model of senile osteoporosis by comparing morphological, mechanical, and densitometric properties of long bones from SAMP6 mice to those of the control strain SAMR1. We hypothesized that SAMP6 bones would have reduced bone area, whole-bone fracture force, and diaphyseal bone mineral density (BMD) compared with SAMR1, and that these deficits would be evident in both 4-month- and 12-month-old mice.
MATERIALS AND METHODS
Eighty-three SAM mice were obtained from our colony that was started with breeders provided by the Council for SAM Research of Kyoto University, Japan. They were assigned to groups (8–12 animals/group) based on strain (SAMR1/SAMP6), age (4 months/12 months), and sex. The mice were killed by CO2 hypoxia and weighed. All procedures were approved by our institutional animal studies committee. Femurs and tibias were isolated and cleaned of soft tissue and their lengths were measured using digital calipers. Then, bones were wrapped in gauze soaked with PBS, placed in a sealed tube, and stored at −20°C.
Right femurs and tibias were embedded for cross-sectional geometric analysis following standard procedures for undecalcified bone (Osteo-Bed; Polysciences, Warrington, PA, USA). The embedded bones then were cut at 50% of total bone length using a diamond wafering saw (Isomet; Buehler, Lake Bluff, IL, USA). The exposed bone surfaces (proximal) were stained using 5% silver nitrate and imaged using a light microscope equipped with a CCD camera. The medial-lateral (M-L) and anterior-posterior (A-P) periosteal widths and average cortical thickness were determined using image processing software (Scion Image; Scion Corp., Frederick, MD, USA). The FORTRAN program VA-TWIST(11) was used to calculate the geometric properties of the cortical bone cross-sections: medullary area, bone area, and bending moments of inertia. The bending moments of inertia IA-P and IM-L represent the geometric resistance of the cross-section to bending about the A-P and M-L axes, respectively. The precision of the technique used to calculate cross-sectional properties was determined previously to be ≤1% for all parameters.(12)
Before mechanical testing, left femurs and tibias (wrapped in wet gauze) were scanned using peripheral quantitative computed tomography (pQCT; XCT Research M, Stratec, Germany). Slices were taken at the midpoint and ±1 mm from the midpoint (voxel size, 70 μm in plane and 500 μm out of plane). The manufacturer's software was used to determine bone mineral content (BMC) and volumetric BMD using a bone threshold of 600 mg/cm3.
Left femurs and tibias were mechanically tested to failure using four-point bending.(12) The distance between the upper (loading) points was 3.0 mm and the distance between the lower (support) points was 7.0 mm. Before testing, the bones were soaked for 1 h in PBS at room temperature (21–23°C) to ensure hydration.(13) Mechanical tests were conducted using a materials testing machine (Instron 8500R; Instron Corp., Canton, MA, USA) fitted with an 111-N force cell (model 3397; Lebow, Inc., Troy, MI, USA). Bones were loaded at a displacement rate of 0.03 mm/s in a direction that allowed for stable positioning in the loading fixtures; femurs were flexed about the M-L axis (tension on the anterior surface), while tibias were laterally bent about the A-P axis (tension on the medial surface). Force-displacement data were collected using a computerized data acquisition system (Labview 5.0; National Instruments, Austin, TX, USA). Force values were converted to moments (moment = force ∗ a/2) and displacement data were divided by the quantity (3aL − 4a2)/6, where L is the distance between the support points and a is the distance between the loading and support points (2.0 mm). The following structural properties then were determined from the moment versus normalized displacement curves (Fig. 1): rigidity (Nmm2), yield moment (Nmm), ultimate moment (Nmm), postyield displacement (mm/mm2), and energy-to-fracture (N).
After failure testing, cortical density and ash fraction were determined. Diaphyseal sections (2–3 mm long) were trimmed from the broken bones and flushed with saline to remove the marrow. Tissue volume was determined using Archimede's principle.(14) For wet weight measurements, bones were centrifuged at 14,000 rpm for 15 minutes and then weighed using an analytical balance. Dry weight was determined after drying at 90°C for 24 h. Ash weight was determined after ashing in a muffle furnace at 600°C for 24 h. Each measurement was repeated five times per bone and the average values were determined. Wet density was defined as wet weight divided by bone volume and was taken as an indirect measure of cortical porosity. Ash fraction was calculated as ash weight divided by dry weight and was taken as a measure of mineral content.(15)
Preliminary statistical analysis indicated that the effects of mouse strain and age did not differ between male and female mice, although bones from female mice were smaller and tended to be weaker than bones from male mice. For conciseness, we pooled male and female groups and used two-way ANOVA to assess the effects of mouse strain and age and Tukey tests for post hoc comparisons (StatView 5.0; Abacus Concepts, Inc., Berkeley, CA, USA). Statistical significance was considered at p < 0.05. Results are presented as mean ± SD.
Morphological data indicated that at both 4 months and 12 months, SAMP6 long bones were wider and had larger medullary cavities than SAMR1 bones but with no difference in the net amount of bone tissue (Fig. 2). SAMP6 femurs had significantly increased A-P and M-L widths, medullary area, and moments of inertia about both A-P and M-L axes compared with SAMR1 (p < 0.001; Table 1). For example, femoral medullary area was increased by an average of 40% and moments of inertia increased by 35% in SAMP6. However, cortical thickness (p = 0.60) and bone area (p = 0.093) did not differ significantly between SAMP6 and SAMR1 femurs. Similar to the differences observed in the femurs, SAMP6 tibias had increased M-L width, medullary area, and moment of inertia about the A-P axis (p < 0.001) but the same bone area (p = 0.36) compared with SAMR1. In contrast to the femurs, SAMP6 tibias had a small (9%) reduction in cortical thickness compared with SAMR1 (p = 0.004). Bone lengths differed slightly between the two strains. SAMP6 femurs were 2.5% shorter than SAMR1 femurs, and SAMP6 tibias were 1.2% longer (p < 0.01). SAMP6 mice were 15% heavier compared with SAMR1 (p < 0.001). SAMR1 mice weighed 30.5 ± 4.3 g and 35.4 ± 4.9 g at 4 months and 12 months respectively, compared with 34.8 ± 2.6 g and 41.1 ± 5.5 g for SAMP6.
Table Table 1.. Morphological Properties of SAM Long Bones
Despite their increased cross-sectional moments of inertia, long bones from SAMP6 mice were significantly weaker and more brittle than long bones from SAMR1 controls at both 4 months and 12 months age (Fig. 3). Femurs from SAMP6 mice had yield and ultimate moments that were approximately 20% and 35% less, respectively, than SAMR1 controls (p < 0.001; Table 2). The most striking differences between groups were in measures of femoral brittleness, in which both postyield displacement and energy to fracture were reduced by 70% in SAMP6 femurs compared with control (p < 0.001; Fig. 4). In contrast, there was no significant difference in femoral rigidity between strains (p = 0.73), indicating normal whole-bone elastic behavior in SAMP6 femurs. Tibial properties showed similar relative differences between SAMP6 and SAMR1, although the magnitude of the differences was diminished compared with the femur (Table 2). In slight contrast to the femoral data, which indicated comparable differences between SAMR1 and SAMP6 at both 4 months and 12 months, the tibial data indicated differences that were more pronounced at 12 months. For example, tibial postyield displacement was reduced by 33% in SAMP6 compared with SAMR1 at 4-months but was reduced by 66% at 12-months. The only difference in long bone bending properties that favored SAMP6 was in the tibial rigidity, which was 26% greater in SAMP6 mice compared with SAMR1 at 4 months of age.
Table Table 2.. Mechanical Properties of SAM Femurs and Tibias at 4 Months and 12 Months of Age
Several physical and pQCT measures indicated increased diaphyseal mineral density in SAMP6 bones compared with SAMR1, although these differences were small and were not consistently observed across all groups (Table 3). SAMP6 femurs had significantly increased ash fraction (p = 0.016) and cortical BMD (p = 0.012) but normal wet density (p = 0.49). SAMP6 tibias had increased ash fraction (p < 0.001) but normal wet density (p = 0.39) and BMD (p = 0.19) compared with SAMR1. BMC was greater in SAMP6 femurs compared with SAMR1 (p = 0.012), consistent with increased BMD and a nonsignificant trend for increased bone area. In contrast, tibial BMC did not differ between strains (p = 0.14), consistent with the lack of difference in both BMD and bone area.
Table Table 3.. Physical and pQCT Measures of Bone Density
With aging, bones from both SAMP6 and SAMR1 showed evidence of endosteal expansion and cortical thinning while at the same time becoming weaker. At 12 months, femurs and tibias had increased medullary area (p < 0.05) and decreased cortical thickness (p < 0.01) compared with 4 months. Bone area did not differ from 4 to 12 months (p > 0.05). At 12 months, femurs and tibias had reduced yield (p < 0.01) and ultimate (p < 0.001) moments and reduced postyield displacement (p < 0.01) and energy-to-fracture (p < 0.001) compared with bones at 4 months. Measures of bone density showed few changes with aging, although femoral BMD (p = 0.004) and tibial ash fraction (p = 0.003) were increased at 12 months compared with 4 months.
Our objective was to characterize the functional relevance of the SAMP6 murine model of senile osteoporosis by comparing mechanical, densitometric, and morphological properties of long bones from SAMP6 mice to those of the control strain SAMR1. Bones from SAMP6 mice had enlarged medullary cavities and increased periosteal widths. These differences resulted in no net change in cortical bone area but an increase in moment of inertia. Although these geometric features resulted in normal or increased elastic stiffness (rigidity), bones from SAMP6 mice had reduced whole-bone strength and were markedly more brittle than bones from SAMR1 mice. Several measures of mineral density were increased slightly in SAMP6 compared with SAMR1, indicating that increased mineralization may have contributed to their relative brittleness.
A number of our morphological results are in agreement with previous reports of the SAMP6 skeleton and support the relevance of SAMP6 as a model of accelerated skeletal senescence. Our findings of increased medullary area in SAMP6 relative to SAMR1 are consistent with previous reports describing increased endosteal dimensions(3,5) and reduced rates of trabecular bone formation.(2,3,6) Our finding of increased periosteal width is in contrast to one study(5) but consistent with another.(3) In contrast with several other studies,(3–5) we did not observe a significant reduction in cortical thickness of SAMP6 femurs, although we did observe a significant reduction in thickness in SAMP6 tibias. Taken together, our findings of increased periosteal width combined with equivalent or reduced cortical thickness are consistent with the reports of reduced areal BMD based on radiographs(5) and DXA.(2) Finally, our morphological data suggest that SAMP6 bones have “compensated” for diminished endosteal bone formation by increasing periosteal size and thus moment of inertia. These changes are comparable with age-related changes reported in C57BL/6 mice.(16) More importantly, the morphological differences between SAMP6 and SAMR1 are similar to changes that occur with aging in human long bones(17–19) and thus support the relevance of the SAMP6 mouse as a model of accelerated skeletal aging.
There have been no previous reports of mechanical properties from the SAMP6 skeleton and few reports describing age-related changes in bone strength in mice. The only previous study of bone strength in SAM mice reported an ultimate bending moment of ∼26 Nmm in SAMR1 femurs at 11 months of age,(20) which is comparable with our value of 34 Nmm at 12 months of age. Relative to SAMR1 controls, we measured markedly decreased ultimate moment and energy to fracture in SAMP6 bones at both 4 months and 12 months of age. We also determined that from 4 to 12 months the ultimate moment decreased by ∼20% and energy to fracture decreased by 40% in both strains. Taken together, our mechanical results are consistent with the observation reported by Takeda and coworkers of spontaneous tibial fractures in 14% of SAMP6 mice that survived to 22 months of age.(4,5) The age-related decrease we observed in diaphyseal bending strength is in contrast to results reported for C57BL/6 mice(16) and C57BL/6 × C3H/He hybrid mice,(21) in which there was no evidence of a decline in strength up to 24 months and 18 months of age, respectively. However, the age-related changes in mechanical properties that occurred in SAM long bones are comparable with those reported for both human whole bones(17,22) and human cortical bone tissue,(23,24) further highlighting the clinical relevance of the SAMP6 model.
The relative weakness and brittleness of SAMP6 long bones was not suggested by previous descriptions of the skeletal phenotype. Moreover, the increased periosteal width and bending moments of inertia in SAMP6 long bones we observed, taken alone, would have suggested increased whole-bone strength compared with the smaller SAMR1 bones. The only measure that would have indicated a brittle phenotype was the increase in ash fraction, although the changes in ash fraction were so slight (∼2%) that the magnitude of the reduction in fracture energy or postyield displacement would have been difficult to predict. The apparent discrepancy between increased bone size and decreased bone strength in SAMP6 highlights the critical role of mechanical testing in assessing the functional relevance of murine models.(25) To put the magnitude of the increased brittleness into perspective, the reductions in postyield displacement and energy to failure (∼70%) were comparable with the values reported in murine models of osteogenesis imperfecta. In bones from Mov13 mice, postyield displacement was reduced by 61% compared with control,(26) and in bones from oim/oim mice fracture energy was reduced by 62%.(21) These comparisons suggest that, in addition to the increased mineralization of SAMP6 long bones, a defect in collagen structure also might contribute to their brittle behavior. Additional investigations are needed to assess the relative contributions of mineral and matrix to the brittle behavior of the SAMP6 skeleton.
SAMP6 mice were ∼15% heavier than SAMR1 controls. Therefore, it is possible that the increased bone widths and moments of inertia in SAMP6 bones resulted in part from adaptation to increased skeletal loading. To account for differences in body weight, we compared morphological parameters normalized by body weight. Many (but not all) of the normalized morphological parameters were not significantly different between strains (data not shown), suggesting that SAMP6 bones may be appropriately sized relative to their increased body weight. For example, normalized M-L width and M-L moment of inertia were not different between SAMP6 and SAMR1 (p > 0.05 for both tibia and femur). Regardless of whether or not SAMP6 bones are appropriately sized for their body weight, increased body weight in SAMP6 mice does not result in stronger bones. Either in absolute terms or normalized to their body weight (data not shown), SAMP6 long bones are less strong and more brittle.
One limitation of our study is that we assessed mechanical properties at sites of cortical rather than trabecular bone. Thus, we cannot address the relevance of the SAMP6 mouse for studies on age-related changes in trabecular bone. Nevertheless, the mechanical “defect” in SAMP6 bones that we observed (i.e., decreased cortical bone strength and increased brittleness) does correspond to changes that occur with aging in human cortical bone.(23,24) We believe that testing long bones has provided a valuable initial assessment of the skeletal phenotype of the SAMP6 mouse and that future studies therefore are warranted to assess whether or not these findings extend to regions containing trabecular bone.
In summary, long bones from SAMP6 mice have increased endosteal and periosteal dimensions, reduced bending strength and increased brittleness, and increased mineralization, all of which are analogous to changes that occur in the aging human skeleton. Moreover, these features are consistently observed in young (4 month) and old (12 month) animals. These findings extend the previous descriptions of the SAMP6 mouse and identify key mechanical features that further validate its relevance as a unique and functionally relevant model of senile osteoporosis.
We thank Daniel Touhey for help in embedding and imaging bone samples. This work was funded in part by the National Institutes of Arthritis and Musculoskeletal and Skin Diseases (R01 AR47867). The breeding pairs were generously provided by the Council for SAM Research, Kyoto University, Japan.