The authors state that they have no conflicts of interest.
Three years of daily alendronate treatment increases microdamage in vertebral bone but does not significantly increase it beyond levels of microdamage found after 1 yr of treatment. This suggests microdamage accumulation peaks during the early period of bisphosphonate treatment and does not continue to accumulate with longer periods of treatment.
Introduction: Clinically relevant doses of alendronate increase vertebral microdamage by 4- to 5-fold in skeletally mature beagles after 1 yr of treatment. The goal of this study was to determine whether microdamage would continue to accumulate with 3 yr of alendronate treatment in an intact beagle dog model.
Materials and Methods: One-year-old female beagles were treated with daily oral doses of vehicle (VEH, 1 ml/kg/d) or alendronate (ALN, 0.2 or 1.0 mg/kg/d) for 3 yr. These ALN doses were chosen to approximate, on a milligram per kilogram basis, those used to treat osteoporosis (ALN0.2) and Paget's disease (ALN1.0). Microdamage accumulation, static and dynamic histomorphometry, densitometry, and mechanical properties of lumbar vertebrae were assessed. Comparisons were made among the three groups treated for 3 yr and also within each treatment group to animals treated under the same conditions for 1 yr.
Results: Overall microdamage accumulation (crack surface density) was not significantly higher in animals treated for 3 yr with either dose of ALN, whereas crack density increased significantly (100%; p < 0.05) with the higher dose of ALN compared with VEH. Both ALN doses significantly suppressed the rate of bone turnover (−60% versus VEH). There was no difference among groups for any of the structural biomechanical properties-ultimate load, stiffness, or energy absorption. However, when adjusted for areal BMD, ALN-treated animals had significantly lower energy absorption (−20%) compared with VEH. Toughness, the energy absorption capacity of the bone tissue, was significantly lower than VEH for both ALN0.2 (−27%) and ALN1.0 (−33%). Compared with animals treated for 1 yr, there was no significant difference in microdamage accumulation for either ALN dose. VEH-treated animals had significantly lower bone turnover (−58%) and significantly higher levels of microdamage (+300%) compared with values in 1-yr animals. Toughness was significantly lower in animals treated for 3 yr with ALN1.0 (−18%) compared with animals treated for 1 yr, whereas there was no difference in toughness between the two treatment durations for either VEH or ALN0.2.
Conclusions: Although 3 yr of ALN treatment resulted in higher microcrack density in vertebral trabecular bone compared with control dogs, the amount of microdamage was not significantly higher than animals treated for 1 yr with similar doses. This suggests that bisphosphonate-associated increases in microdamage occur early in treatment. Because toughness continued to decline significantly over 3 yr of treatment at the higher ALN dose, decreases in toughness are probably not dependent on damage accumulation.
Microdamage accumulates with age(1,2) and may play an important role in age-associated bone fragility.(3,4) Microdamage formation occurs in response to mechanical loads,(5–7) preferentially at sites of increased tissue mineralization,(8–10) and is removed by remodeling.(6,7) The level of skeletal microdamage is determined by the balance between microdamage formation and its removal. Therefore, conditions that either increase microdamage formation, or decrease its removal, can have a significant impact on the accumulation of microdamage and bone fragility.
Bisphosphonates are efficacious for reducing fractures because of their suppression of bone remodeling.(11–13) However, because reductions in remodeling are permissive for the accumulation of microdamage, bisphosphonate treatment also increases skeletal microdamage. Numerous animal studies have noted significant increases in microdamage after bisphosphonate treatment.(14–18) This accumulation of microdamage occurs with alendronate and risedronate doses comparable to those used for the treatment of postmenopausal osteoporosis, although the accumulation is greater when higher doses are given (e.g., those approximating doses used for treatment of Paget's disease).(18)
Whether microdamage accumulation continues or plateaus with extended bisphosphonate treatment is not known, yet has significant implications as some patients now enter their second decade of treatment. Studies to date have assessed microdamage at a single time-point, most often 1 yr. Recently, Komatsubara et al.(16,17) reported that 3 yr of daily incadronate, at 2.5 or 5 times the clinical dose, significantly increased the accumulation of microdamage in both the vertebrae and rib of dogs. Because no data are available concerning microdamage levels with shorter-term incadronate treatment (<3 yr), this study was not able to address whether microdamage accumulation continues or plateaus with prolonged bisphosphonate treatment.
Animal studies documenting increased microdamage with bisphosphonate treatment have consistently shown increases in vertebral bone strength and stiffness, leading to questions regarding the implications of increased microdamage with bisphosphonates. However, in all but one of these studies,(16) bisphosphonate treatment reduced bone toughness, the energy absorption capacity of the bone tissue.(14,15,17,18) Furthermore, when normalized for increases in BMD, energy absorption capacity at the whole bone (structural) level was significantly compromised after 1 yr of alendronate treatment.(19) If microdamage continues to accumulate with prolonged bisphosphonate treatment, it is possible that this could lead to further reductions in work to failure and toughness.
The goal of this study was to test the hypothesis that microdamage continues to accumulate throughout the duration of bisphosphonate treatment and that this continued accumulation is accompanied by a progressive decline in energy absorption and toughness. We have recently documented that clinically relevant doses of alendronate reduce vertebral bone turnover by >70%, increase microdamage by 4- to 5-fold, and nonsignificantly reduce vertebral toughness by 14–17% in skeletally mature beagles after 1 yr of treatment.(18) This study reports results from animals treated for 3 yr with the same doses of alendronate used in the 1-yr study. This allows both an across-treatment analysis at the 3-yr time-point (vehicle versus alendronate) and a within-treatment analysis across time-points (1 versus 3 yr).
MATERIALS AND METHODS
All procedures were approved before the study by the Indiana University School of Medicine Animal Care and Use Committee. Thirty-six female beagles (1–2 yr old on arrival) were purchased from LBL (Reelsville, IN, USA). On arrival, lateral X-rays of all dogs were obtained to confirm skeletal maturity (closed proximal tibia and lumbar vertebra growth plates). Animals were housed two per cage in environmentally controlled rooms at Indiana University School of Medicine's AALAC accredited facility and provided standard dog chow and water. Two dogs (both in the ALN 0.2 group) developed hernias, both in year 2, that required surgery. One of these animals developed a second hernia that progressed to the point of needing to be terminated early (month 34 of treatment); this animal was still included in all analyses. All other animals completed the 36-mo treatment without serious complications.
After 2 wk of acclimatization, animals were assigned to treatment groups (n = 12/group) by matching body weights. All dogs were treated daily for 3 yr with oral doses of vehicle (saline, 1 ml/kg/d) or alendronate sodium (0.20 or 1.00 mg/kg/d; Merck and Co.). Alendronate doses were chosen to approximate, on a milligram per kilogram basis, the doses used for treatment of postmenopausal osteoporosis and Paget's disease, respectively. Alendronate was dissolved in saline and administered to the dogs orally with a syringe. Vehicle-treated animals received 1 ml/kg/d of saline. Dosing was performed each morning after an overnight fast and at least 2 h before feeding.
Before necropsy, animals were injected with calcein (0.20 ml/kg, IV) using a 2–12–2-5 labeling schedule. Animals were killed by intravenous administration of sodium pentobarbital (0.22 mg/kg Beuthanasia-D Special). After death, lumbar vertebrae were dissected and saved for analyses. The second and third lumbar vertebrae were fixed in 10% neutral buffered formalin, whereas the fourth lumbar vertebra was wrapped in saline-soaked gauze and frozen (−20°C). All tissue preparation, processing, and analyses were similar to those used for dogs treated for 1 yr.(18)
Histology (static, dynamic, and microdamage)
Static and dynamic histomorphometric measures of trabecular bone were obtained on second lumbar vertebrae (L2). Bones were embedded undecalcified in methyl methacrylate (MMA; Aldrich). Midsagittal (4 μm) sections were cut using a Reichert-Jung 2050 microtome (Magee Scientific) and stained with McNeal's tetrachrome for static histomorphometry. Midsagittal (8 μm) sections were cut and left unstained for dynamic histomorphometry and wall thickness measures.
Third lumbar vertebrae (L3) were processed for microdamage assessment by bulk staining in basic fuchsin as previously described.(18,20) Using 1% basic fuchsin dissolved in increasing concentrations of ethanol, specimens were stained according to the following schedule: 8 h at 80% (with one change to fresh 80% after 4 h), overnight in 95% (with one change to fresh 95%), and 8 h in 100% (with one change to fresh 100% after 4 h). Bones were placed under vacuum (20 in Hg) for all stages during the day and left on the bench top overnight. After staining, bones were washed in 100% ethanol and embedded undecalcified in MMA. Midsagittal (80–100 μm) sections were cut using a diamond wire saw (Histosaw; Delaware Diamond Knives).
Histological measurements were made using a semiautomatic analysis system (Bioquant OSTEO 7.20.10; Bioquant Image Analysis) attached to a microscope equipped with an UV light source (Optiphot 2 microscope; Nikon). A 5 × 5-mm region of interest, located 1 mm below the cranial plateau, was used for sampling. Static and dynamic variables were measured and calculated in accordance with ASBMR recommended standards.(21) Microdamage was assessed using UV fluorescence as previously described.(22) Measurements included crack length (Cr.Le, μm) and crack number (Cr.N), with calculations of crack density (Cr.Dn, #/mm2; Cr.N/bone area) and crack surface density (Cr.S.Dn, μm/mm2; Cr.N * Cr.Le/bone area).
Areal BMD (aBMD, g/cm2) of the fourth lumbar vertebra (L4), without the posterior elements or cranial/caudal endplates, was quantified using a PIXImus II densitometer (Lunar). Volumetric bone density and geometry of the L4 vertebra was quantified using a Norland Stratec XCT Research SA+ pQCT (Stratec Electronics). One slice (0.07 × 0.07 × 0.50-mm voxel size) was taken at three locations (25%, 50%, and 75% of total vertebra height). Total, trabecular, and cortical volumetric BMD (vBMD, mg/cm3) and cross-sectional area (CSA, mm2) were obtained for each slice and averaged together to obtain a single representative value for each specimen.
The biomechanical properties of L4 vertebrae were quantified using a servohydraulic testing system (MTS Bionix; MTS Corp.). Compression to failure was carried out on saline soaked specimens using displacement control mode (20 mm/min). Load versus displacement data were digitally recorded at a sampling rate of 10 Hz. Plots were analyzed for determination of ultimate force (F), stiffness (k), and work to ultimate force (w). Apparent material-level properties ultimate stress (σult), modulus (E), and toughness (U) were estimated using the following equations: σult = (F/CSA)/BV/TV; E = (k × [height/CSA])/BV/TV; U = (w/[height × CSA])/BV/TV, where CSA was from pQCT, height was measured using digital calipers, and BV/TV was from L2 histomorphometry.
All statistical tests were performed using SAS software (SAS Institute). To determine whether variables were different among treatment groups after 3 yr, data were evaluated using a one-way ANOVA with Fisher's protected least significant difference (PLSD) posthoc tests. Strength-density and energy absorption-density relationships from 3-yr treated animals were compared between VEH and ALN treatments using analyses of covariance with least square means (LSM) used to determine differences in parameters after accounting for aBMD. To determine whether changes occurred within treatment groups across time, t-tests were used to compare data from animals treated for 3 yr with results from an earlier study in our laboratory that treated animals under the same conditions for 1 yr.(18) For all tests, p ≤ 0.05 was considered statistically significant. All data are presented as mean ± SE.
At the conclusion of the study, there was no significant difference in body mass among the three groups (VEH: 12.6 ± 0.6 kg; ALN0.2: 12.4 ± 0.5 kg; ALN1.0: 11.6 ± 0.7 kg; p = 0.492).
Crack density, the number of microcracks per mm bone tissue, was significantly higher than VEH for ALN1.0 (+100%, p = 0.01) but not ALN0.2 (+50%; p = 0.12; Fig. 1A). Mean crack length was significantly smaller in both ALN-treated groups compared with VEH (−20% for both; Fig. 1B). Crack surface density, the product of crack density and crack length, was not significantly different among groups (Fig. 1C).
Activation frequency (Ac.f) was significantly lower than VEH in both ALN0.2- (−59%) and ALN1.0- (−60%) treated animals. The reduction in Ac.f resulted from significant suppression of both mineral apposition rate (MAR) and mineralizing surface (MS/BS), with no change in wall thickness. MAR was 17% lower than VEH for both doses of ALN, whereas MS/BS was −51% and −62% for ALN0.2 and ALN1.0 groups, respectively (Table 1).
Table Table 1.. Dynamic Histomorphometry of the Second Lumbar Vertebrae
Structural biomechanical properties-ultimate load, stiffness, and energy to ultimate load-were not significantly different among the treatment groups (Table 2). When normalized for aBMD, there was no difference in the strength-density relationship between VEH- and ALN-treated animals (Fig. 2A). The slope of the energy absorption-density relationship was similar between treatments, yet at a given aBMD, the energy absorption capacity was significantly lower in vertebrae from ALN-treated animals (−20%, p = 0.01) compared with VEH (Fig. 2B). For both the strength-density and energy absorption-density relationships, the two doses of ALN were pooled because the results were similar when doses were assessed separately.
Table Table 2.. Compressive Biomechanical Properties of the Fourth Lumbar Vertebrae
Toughness, the energy absorption capacity of the bone tissue, was significantly lower in both ALN0.2 (−26%) and ALN1.0 (−33%) groups compared with VEH (Table 2). There was no difference among groups for the other two material-level properties, ultimate stress and modulus.
Vertebral aBMD was not significantly different among groups, whereas vBMD tended to be higher (p = 0.056) in both ALN0.2 and ALN1.0 groups (both +7%) versus VEH (Table 3). Trabecular vBMD, cortical vBMD, and CSA were not different among the three treatment groups. Trabecular bone volume, assessed by histology, was significantly greater in both ALN0.2 (+23%) and ALN1.0 (+31%) treatment groups compared with VEH (Table 3).
Table Table 3.. Lumbar Vertebrae Bone Mineral Density, Geometry, and Bone Volume
After 3 yr of treatment Ac.f. was significantly lower in ALN0.2 (−40%, p = 0.01), but not ALN1.0 (−30%, p = 0.30), compared with similar treatment groups at 1 yr (Fig. 3A). The level of microdamage (both Cr.Dn and Cr.S.Dn) was not significantly different at 3 yr compared with 1 yr for either ALN group (Fig. 3B). Compared with 1 yr of treatment, ALN0.2 had higher ultimate load (+21%), stiffness (+55%), and modulus (+65%) at 3 yr, whereas ALN1.0 had significantly higher stiffness (+42%) and modulus (+30%) and lower toughness (−18%; Table 4; Fig. 3C). VEH-treated animals had significantly lower Ac.f. (−58%), higher microdamage accumulation (+301%), and higher structural- and material-level strength and stiffness at 3 yr compared with VEH-treated animals after 1 yr (Table 4; Fig. 3).
Table Table 4.. Percent Difference Between Animals Treated for 1 and 3 Years Within Treatment
Animal studies have consistently documented higher levels of microdamage in bisphosphonate-treated animals,(14–18) yet it has remained unclear whether microdamage accumulation continues or plateaus with extended bisphosphonate treatment. Recently, we documented that clinically relevant doses of alendronate increase microdamage by 4- to 5-fold in skeletally mature beagles after 1 yr of treatment.(18) We now present data to show that the level of microdamage in vertebral trabecular bone does not significantly increase with an additional 2 yr of alendronate treatment (3-yr total treatment duration) at doses approximating those used to treat postmenopausal osteoporosis or Paget's disease.
Because remodeling is necessary to remove microdamage,(6,7) bisphosphonate treatment would be expected to allow accumulation of damage caused by turnover suppression. Whereas the degree of turnover suppression is correlated to the degree of microdamage accumulation,(15,17,18) even mild suppression of turnover (∼40%) with bisphosphonate treatment is sufficient to allow significant increases in microdamage.(18) This study shows that the initial suppression of turnover with bisphosphonate treatment has the greatest influence on microdamage accumulation. After 1 yr of ALN treatment, vertebral bone turnover is suppressed by ∼70%, associated with a 4- to 5-fold increase in microdamage.(18) With an additional 2 yr of treatment, and a continued decline in turnover (−30% to −40% compared with values in 1-yr animals), microdamage was not significantly increased (1.3- and 1.6-fold higher than VEH). The most plausible explanations for this finding are (1) microdamage can be controlled at a new equilibrium level even with only 30% of normal bone turnover and/or (2) there is a reduced formation of microdamage. The latter could result from the lowering of trabecular strains caused by the 20–30% increase in bone volume (Table 3).
Consistent with the relationship between turnover suppression and microdamage accumulation, animals treated for 3 yr with vehicle had significantly lower turnover (−58%) and significantly higher levels of microdamage (+300%) compared with those treated for 1 yr. These data highlight that microdamage accumulation is not caused by bisphosphonates, per se, but rather the reduction in turnover brought about by bisphosphonate treatment.
Toughness, the energy absorption capacity of the material, is consistently reduced in bisphosphonate-treated animals.(14,15,17,18) This change has often been attributed to microdamage accumulation, although a cause and effect has yet to be established. These current results provide two pieces of evidence to suggest microdamage accumulation is not causing reduced toughness in bisphosphonate-treated bone. First, despite higher levels of microdamage in VEH-treated animals after 3 yr (compared with 1 yr), there was no change in bone toughness. Second, despite no significant difference in microdamage accumulation between animals treated for 1 and 3 yr with either dose of ALN, animals treated with ALN1.0 had significantly lower bone toughness at 3 yr compared with 1 yr. Whereas these data do not disprove a cause/effect relationship, they strongly suggest that bisphosphonate-associated reductions in bone toughness extend beyond simply the accumulation of microdamage.
Structural biomechanical properties-ultimate load, stiffness, and energy absorption-were not significantly different than VEH after 3 yr of ALN treatment. These results differ from those at 1 yr, where both doses of ALN significantly increased vertebral stiffness(18) and the higher dose of ALN increased strength.(15) The absence of difference among these groups treated for 3 yr is likely the result of significant increases in VEH-treated animals, which had significantly higher ultimate load (+24%) and stiffness (+68%) compared with values in 1-yr treated animals. These higher structural-level mechanical properties in 3-yr VEH-treated animals compared with 1-yr VEH-treated animals likely result from age-associated periosteal expansion. Vertebral CSA, which plays a significant role in determining structural parameters and results from continued periosteal expansion, was significantly higher (+17%) in the 3-yr VEH group compared with the 1-yr group. Material-level properties-ultimate stress and modulus-were also higher in VEH-treated animals at 3 yr compared with 1 yr. We have recently documented increases in collagen cross-linking and collagen maturity of vertebrae that are attributable to turnover suppression.(23) Because the organic matrix is known to affect material properties, we hypothesize that the reduction in turnover between 1 and 3 yr in vehicle-treated animals (−58%) results in an increase in collagen cross-linking and maturity that, in conjunction with other parameters such as mineralization and microdamage, determine material-level biomechanical properties.(24)
An alternative approach to study the effects of bisphosphonate treatment on biomechanical properties is to compare the relationships between BMD and biomechanical properties. Proposed by Hernandez and Keaveny,(25) these relationships allow the determination of changes in bone strength or energy to fracture that are not accounted for by a change in bone mass (aBMD). ALN-treated animals had 20% lower energy absorption capacity at a given aBMD, indicating that an increase in BMD is necessary with alendronate treatment to maintain energy absorption capacity at a level comparable to nontreated bone. This result is consistent with the 22% lower energy absorption at a given aBMD after 1 yr of treatment with doses of ALN approximating those used to treat osteoporosis.(19)
Given the invasive nature of both microdamage and biomechanical property measures, it proves difficult to determine whether the changes noted in this study extend to humans treated with bisphosphonates. Higher levels of microdamage exist in bisphosphonate-treated women,(26) although there is no data to support whether there exists a similar treatment duration accumulation pattern as noted in this study. Bisphosphonates have clear antifracture efficacy suggestive of improved biomechanical properties.(11–13) However, given the multifactorial nature of fractures, it remains possible that reduced toughness or lower energy absorption at a given aBMD could exist even in light of an overall population-wide reduction in fracture risk with bisphosphonates. Indeed, both toughness and energy absorption are compensated for by the increased BMD that routinely occurs with bisphosphonate treatment, but the material properties of the tissue are nevertheless compromised.
In conclusion, 3 yr of alendronate treatment resulted in higher microcrack density in vertebral trabecular bone of intact beagle dogs, yet the amount of microdamage was not significantly higher than in animals treated with equivalent doses for 1 yr. This suggests that increased skeletal microdamage associated with turnover suppression occurs early in treatment and does not progress with longer treatment duration. Because toughness continued to decline significantly over 3 yr of treatment at the higher ALN dose, decreases in toughness are probably not dependent on damage accumulation.
The authors thank Keith Condon, Diana Jacob, and Lauren Waugh for histological preparation and Andrew Koivuniemi and Mark Koivuniemi for assistance with densitometry and mechanical testing. This work was supported by NIH Grants AR047838 and AR007581 and used an animal facility constructed with support from Research Facilities Improvement Program Grant C06 RR10601-01 from the National Center for Research Resources, National Institutes of Health.