RANKL Inhibition Improves Bone Properties in a Mouse Model of Osteogenesis Imperfecta

Abstract

Recently, a new class of agents targeting the receptor activator of nuclear factor-κB ligand (RANKL) pathway has been developed for the treatment of osteoporosis and other bone diseases. In the current study, inhibition of the RANKL pathway was evaluated to assess effects on “bone quality” and fracture incidence in an animal model of osteogenesis imperfect (OI), the oim/oim mouse. Juvenile oim/oim (~6 weeks old) and wildtype (+/+) mice were treated with either a RANKL inhibitor (RANK-Fc) or saline. After treatment, bone density increased significantly in the femurs of both genotypes. Femoral length decreased with RANK-Fc in +/+ mice. Geometric measurements at mid-diaphysis in the oim/oim groups showed increases in the ML periosteal and endosteal diameters and AP cortical thickness in the treated groups. Within +/+ groups, ML cortical thickness and ML femoral periosteal diameter were significantly increased with RANK-Fc. Biomechanical testing revealed increased stiffness in oim/oim and +/+ mice. Total strain was increased with treatment in the +/+ mice. Histologically, RANKL inhibition resulted in retained growth plate cartilage in both genotypes. The average number of fractures sustained by RANK-Fc-treated oim/oim mice was not significantly decreased compared to saline treated oim/oim mice. This preclinical study demonstrated that RANKL inhibition at the current dose improved density and some geometric and biomechanical properties of oim/oim bone, but it did not decrease fracture incidence. Further studies that address commencement of therapy at earlier time points are needed to determine whether this mode of therapy will be clinically useful in OI.

INTRODUCTION

Osteogenesis imperfecta (OI) is a genetic disorder characterized by bone fragility. Clinical manifestations of this disease range from perinatal lethality, to severe disability with frequent fractures, to a relatively benign condition with few fractures over a lifetime. The disorder is most often the result of a mutation in one of two genes that encode the α chains of type I collagen, COL1A1 and COL1A2, leading to a quantitative or qualitative change in type I collagen which in turn results in abnormal bone formation and high bone turnover [1]. The treatment of OI has been extremely challenging, and until the introduction of bisphosphonate therapy, mostly ineffective. Bisphosphonates are currently the mainstay of treatment and the different treatment modalities include cyclic infusions of pamidronate and neridronate, as well as oral treatment with alendronate (ALN) [2–6]. Bisphosphonates, which inhibit osteoclast activity, have been shown to ease chronic bone pain, increase bone mineral density and strength, reduce fractures, and improve final height probably due to a decreased incidence of vertebral fractures [4, 7, 8]. However, limitations in bisphosphonate therapy remain, including occasional respiratory distress with the first infusion [9].

In addition, bisphosphonates have half-lives of up to a decade and therefore the optimal duration of treatment is unknown [10]. There are also concerns about bone healing and remodeling in bisphosphonate-treated children with OI [11, 12]. Further, these compounds are known to cross the placenta and may pose risk to a developing fetus, although the few case reports published to date do not suggest any significant ill effects for the mothers and their offspring [13].

Recently, a new class of agents directed at the receptor activator of nuclear factor-κB (RANK) RANKL/RANK/OPG pathway has been proposed for the treatment of osteoporosis and other degenerative bone diseases. This pathway consists of three key molecules that regulate the rate of bone remodeling. RANK ligand (RANKL) is an osteoblast-secreted protein that binds to RANK, a type I membrane receptor expressed on the surfaces of macrophage precursor cells. The binding of RANKL to RANK induces differentiation of these cells into mature osteoclasts. To mitigate bone loss, molecules targeting this pathway may reduce bone resorption by blocking this differentiation process. In vivo, bone homeostasis is maintained by osteoblasts, which also express a soluble decoy receptor, osteoprotegerin (OPG). Once secreted, OPG competes with RANK and inhibits RANKL activity [14]. The role of the RANKL/RANK/OPG system has been documented in a number of pathological processes ranging from Paget’s disease to osteoporosis, and several therapeutic strategies have been developed [15].

Recombinant OPG injections, anti-RANK antibodies, and RANKL vaccines are all being explored as potential treatments in clinical and preclinical studies as potential treatments of disorders of high bone turnover [16]. Two phase III studies with Denosumab, a fully human monoclonal IgG2 anti-RANKL antibody (Amgen Corp, Thousand Oaks, CA), for treatment of postmenopausal women with or at risk for osteoporosis and for metastatic bone disease in cancer patients have been published [17, 18]. These agents may be a potential therapeutic modality for children with OI; however, to date, there have been no clinical studies investigating this potential application. Denosumab does not recognize murine RANKL [19], so RANKL inhibition in the current study was achieved with a murine RANK-Fc fusion protein [20] that functions in a manner analogous to that of Denosumab.

The oim/oim mouse is an established model of moderate to severe OI, which is characterized by limb deformities, frequent fractures, osteopenia, and small size [21]. These mice have a naturally occurring mutation leading to a deficiency in proα2(I) collagen [21]. This model has been previously used to assess the effects of bisphosphonates in OI [22–25]. Heterozygous oim/+ mice have also been used to test the effect of growth hormone therapy on bone quality in OI [26]. Other animal models of OI exist including other murine models [27–30], and a model of OI in zebra fish [28, 31]. Treatment with bisphosphonates has been most extensively tested in the oim/oim model [22–25] although it has also been tried in the brtl/+ mice [32]. The bone abnormalities in the oim/oim model are more reproducible than the genetically engineered brtl/+ mice making it more suitable for evaluation of therapies.

In these mice, ALN was shown to decrease the incidence of fractures while maintaining calcified cartilage in oim/oim mice treated during a period of infancy and active growth [22, 24]. It has also been shown that ALN affects growth plate morphology [23] and increases metaphyseal bone mass [25]. A recently published paper by Delos et al. reported the effects of RANKL inhibition on fracture healing and bone strength in oim/oim mice, demonstrating that the mechanical properties of healing bone were not adversely affected [33]. We chose to extend this study to assess the effect of murine RANK-Fc (RANKL inhibition) on juvenile mice, testing the drug’s effect on fracture frequency, bone quality, bone density, and histologic parameters.

MATERIAL AND METHODS

Animals and RANK-Fc Treatment

All procedures were approved by the IACUC committee of the Hospital for Special Surgery. Homozygous oim/oim and wild-type (+/+) control mice were obtained from Jackson Laboratories (Bar Harbor, ME, USA). The animals were housed up to four mice per cage (according to genotype and sex) in a light controlled environment (12 hr light-dark cycles). They were given autoclaved water and fed whole and powdered rodent diet (Purina, St. Louis, MO, USA). Weaning was carried out at 3 weeks for +/+ and at 4 weeks for oim/oim mice. Twenty oim/oim mice were sacrificed at 6 weeks of age to provide a baseline fracture number. Starting at 6 weeks of age, mice were randomized into groups of 17–20 and subcutaneously injected with either soluble RANK-Fc (generously provided by Amgen Inc., Thousand Oaks, CA, USA) or saline two times per week for 8 weeks. Three different doses of RANK-Fc were used: low (0.15 mg/kg), medium (1.5 mg/kg), and high (15 mg/kg) to establish efficacy (10–15 mice per group). These doses of RANK-Fc were selected based on prior studies showing reduction in osteoclast numbers [20, 34].

Radiographic Analyses

Fracture Count

At sacrifice, whole body high resolution anterior-posterior radiographs were taken via Faxitron (Faxitron X-Ray, Wheeling, PA, USA) and fractures in the femora, humeri, and tibia counted by two independent investigators. A fracture was defined as present if there was evidence of a callus or obvious bone deformity.

Bone Geometry

Isolated femora were radiographed by faxitron in the anterior-posterior (AP) and medial-lateral (ML) planes at a resolution of 20 linear pixels/mm. Each image included an aluminum alloy step density standard for calibration. Femoral length in the AP view was determined as the distance from the tip of the femoral head to the base of the condyles. Although reliance on radiographs for density and geometric measurements are no longer state-of-the-art, previous work by our group has demonstrated significant correlations between cortical measurements performed by faxitron radiograph and by micro-CT [35]. Endosteal (d_e) and periosteal (d_p) diameters were measured at mid-diaphysis in the AP and ML radiographic views (AP view yields the ML diameter and vice versa). Cortical thickness was calculated as $\frac{1}{2}(d_{\text{p}}-d_{\text{e}})$. The femoral cross section was assumed to be elliptical, and the moment of inertia, I, a measure of the distribution of material around a given axis, was calculated as:

$$I_{\text{ellipse}}=\frac{\pi }{4}\left[\left(\frac{AP{d}_{\text{p}}}{2}\right){\left(\frac{ML{d}_{\text{p}}}{2}\right)}^3-\left(\frac{AP{d}_{\text{e}}}{2}\right){\left(\frac{ML{d}_{\text{e}}}{2}\right)}^3\right]$$

where d_p in the AP view corresponds to the major diameter of the ellipse, and d_p in the ML view corresponds to the minor diameter. Radiographic intensity in the metaphyseal region was calculated in the AP view. For this measurement, the average intensity was obtained through a 1 mm² box drawn 2 mm from the bottom of the condyle. Each individual image was calibrated used the density standard steps with intensity units ranging from 1 to 5. All measurements were done using Image J software (NIH, Bethesda, MD, USA).

Micro-CT Analyses

Qualitative micro-CT analysis was performed on a subset of femora. The femora were scanned at a resolution of 19 μm voxel resolution on an EVS micro-CT system (MS-8 in vitro specimen scanner, Enhanced Vision Systems, Ontario, CA). Images were thresholded using Gems Microview’s (GE Medical Systems, Ontario) autothresholding function.

Histology

Selected femora were fixed in 80% ethanol and decalcified in 10% EDTA (pH 7.2–7.4) for a ~2 weeks. Tissues were embedded in paraffin, sectioned at 7 μ thickness along the coronal plate from anterior to posterior, and stained with alcian blue for qualitative assessment of calcified cartilage.

Biomechanics

Three-point bend tests were performed as previously described [25]. One femur from each mouse was used for mechanical testing. Prior to testing of the oim/oim mice, a radiograph was obtained to confirm that the femur utilized for testing contained no fractures or deformities (n = 13–16). Three-point bend tests were conducted at room temperature using a closed-loop servohydraulic uniaxial MTS test frame (MTS Systems, Eden Prairie, MN, USA) with Instron electronic controls (Instron, Canton, MA, USA). The anterior surface of the femur was placed on the load supports, and the load was applied posteriorly at the middiaphysis. For this configuration, the posterior surface is in compression and the anterior surface experiences tension. Since there was less than a 3% difference in average bone length among specimens (15.25 ± 0.38 vs. 15.60 ± 0.19 for the oim/oim versus +/+ femora, respectively), a set distance of 8 mm was used for the outside loading points. The tests were performed at a rate of 0.5 mm/sec, and load and displacement data were acquired at 20 Hz with a PC equipped with a 12 bit A/D data acquisition board. Specimens were kept moist throughout testing.

Both structural (whole bone) properties and intrinsic material (tissue) properties of the bone were calculated using accepted formulas for bone testing [34, 36]. For structural properties, maximum load was defined as the maximum load the bone sustained during testing, and displacement was assumed to be equal to the amount of crosshead deflection during testing. Stiffness was calculated as the slope of the linear region of the load-displacement curve. For material properties, the stress was calculated as:

$$\sigma =\frac{Mc}{I_{\text{ellipse}}}$$

where M is the moment (equal to F*L/4), c is the distance from the neutral axis to the most anterior point on the bone cross-section (equal to ¹/₂ APd_p when approximating the cross-section to be elliptical), and I_ellipse is the moment of inertia. Strain was calculated as:

$$\epsilon =\frac{12cd}{L^2}$$

where c is the medial-lateral outside radius (¹/₂ MLd_p), d is the crosshead displacement, and L is the span between the loading supports. Elastic modulus (the slope of the linear elastic region of the curve, a measure of the intrinsic stiffness of the material) and yield strength were determined from the resulting stress-strain curves. Toughness (a measure of the amount of energy needed to fracture the bone tissue) was calculated as the area under the stress-strain curve. Materials that sustain very little postyield strain before fracture are termed “brittle.” Thus the “brittleness” of the cortical bone, a material property, was calculated by dividing the yield strain (point of maximum elastic deformation) by the ultimate strain.

Statistics

Statistical analyses were performed with SigmaStat software (SPSS, Chicago, IL, USA). Means and standard deviations were calculated for each measured parameter. Two-way ANOVA was performed to test for the simultaneous effects of genotype and RANK-Fc treatment on the outcome variables. Bonferroni post-hoc tests were performed for comparison of groups with values significant at p < 0.05. Values were considered significantly different at p < 0.05 and tending toward significance at p < 0.1.

RESULTS

Dosing Study

Clinically abnormally high density of the metaphyseal region in growing subjects is associated with poor bone quality [36, 37]. Therefore the optimal dose of RANK-Fc was defined as that which would result in increased density in the metaphyseal region of the oim/oim mice, without the presence of sclerotic bone. Micro-CT images (Figure 1) demonstrated insufficient increase in mineralization with the low dose and features of osteoclerosis with the high dose. Therefore the medium dose was chosen for the remainder of the study.

FIG. 1.

Micro-CT images of femora at the end of 8 weeks of treatment. Insufficient increase in density of the distal femur is seen with the low dose RANK-Fc (0.15 mg/kg), moderate increase in density with medium dose RANK-Fc (1.5 mg/kg), and sclerotic-like changes with the high dose of RANK-Fc (15 mg/kg). In both +/+ [WT] and oim/oim [oim] mice.

Body Weight

All treated animals in the medium dose group gained weight consistently. The +/+ mice weighed significantly more than oim/oim mice at the start and end of the treatment period. There were no differences in weight at sacrifice between the saline and RANK-Fc groups for either the +/+ or oim/oim mice (Figure 2).

FIG. 2.

Average weights in +/+ and oim/oim mice at the end of treatment at 15 weeks of age. There were no significant differences in adult weight of +/+ or oim/oim mice with RANK-Fc treatment.

Histology

As previously described, saline-treated oim/oim bone differed from +/+ controls in that the bone was less robust with delicate trabeculae. +/+ RANK-Fc-treated specimens demonstrated increased central cores of growth plate cartilage with persistent chondrocytes, Erlenmeyer flask deformity (lack of remodeling of the femoral dimetaphysis with abnormal cortical thinning and lack of the concave dimetaphyseal curve resulting in an Erlenmeyer flask-like appearance [38]) and decreased marrow space. Oim/oim RANK-Fc-treated bones also showed persistent central cores of growth plate cartilage in addition to thin trabecular structures and the Erlenmeyer flask deformity. The marrow space was compromised less in the oim/oim mice compared to the controls (Figure 3).

FIG. 3.

RANKL inhibition results in retained cartilage in +/+ and oim/oim animals, Erlenmeyer flask deformity, and decreased marrow space. Hematoxylin and eosin staining. 5x and 25x Magnification. A = +/+ saline, B = +/+ RANK-Fc, C = oim/oim saline, D = oim/oim RANK-Fc.

Radiographic Results

As expected, there were no fractures in either treated or untreated +/+ mice. At 6 weeks of age, oim/oim mice had already sustained an average of 1.8 fractures per mouse. At the end of the treatment period, there was no difference in fracture incidence between the saline, and RANK-Fc-treated oim/oim mice (Figure 4). There were no differences in the type and location of fractures in the mice that received RANK-Fc or saline (data not shown).

FIG. 4.

Average number of fractures per mouse (oim/oim) at the end of treatment. There were no significant differences in the number of fractures sustained by mice in the treatment versus the control group. Of note is the relatively high number of fractures at baseline.

Within +/+ groups, femoral length decreased with RANKL inhibition, while it was not significantly decreased for the oim/oim mice (Table 1). Within both the +/+ group and the oim/oim group, the radiographic metaphyseal density was significantly higher in the treated mice. Geometric measurements at middiaphysis in the oim/oim groups showed significant increases in the ML view of AP periosteal and endosteal diameters, as well as AP view ML cortical thickness, and a trend toward increased ML view AP cortical thickness and AP view ML femoral periosteal diameter in the RANK-Fc-treated group compared to the saline (Table 2). Within +/+ groups, ML view AP cortical thickness and ML view AP femoral periosteal diameter were significantly increased with RANK-Fc while there was a trend to increased AP view ML femoral endosteal diameter (Table 2). Within saline, and RANK-Fc-treated groups, oim/oim mice had significantly lower metaphyseal density, smaller AP and ML femoral periosteal diameters, as well as AP and ML cortical thickness (Table 2). Within RANK-Fc-treated mice AP view ML femoral endosteal diameter was lower for oim/oim mice (Table 2).

TABLE 1.

Average femur length after 8 weeks of treatment

Genotype	Treatment	Femur Length (mm)	No. of Animals
+/+	Saline	15.54 ± 0.27	20
+/+	RANK-Fc	15.19 ± 0.35*	19
oim/oim	Saline	14.48 ± 0.28◇	18
oim/oim	RANK-Fc	14.42 ± 0.56◇	17

^*Significantly different compared to same genotype saline (p < 0.05).

^◇Significantly different compared to same treatment +/+ (p < 0.05).

TABLE 2.

Geometrical properties of bone after 8 weeks of treatment (mean ± SD)

	+/+ Saline	+/+ RANK-Fc	oim/oim Saline	oim/oim RANK-Fc
Metaphyseal density	1.44 ± 0.15	2.68 ± 0.40*	1.08 ± 0.05•	1.87 ± 0.29*,•
ML femoral endosteal diameter, AP view(mm)	1.05 ± 0.12	1.11 ± 0.11◇	0.99 ± 0.11	1.02 ± 0.1•
ML femoral periosteal diameter, AP view(mm)	1.69 ± 0.16	1.73 ± 0.14	1.41 ± 0.11•	1.49 ± 0.15◇
ML cortical thickness AP view(mm)	0.32 ± 0.04	0.31 ± 0.04	0.21 ± 0.02•	0.24 ± 0.05*,•
AP femoral endosteal diameter, ML view(mm)	0.65 ± 0.07	0.66 ± 0.08	0.61 ± 0.07	0.65 ± 0.05*
AP femoral periosteal diameter, ML view(mm)	1.22 ± 0.09	1.27 ± 0.09*	1.03 ± 0.06•	1.13 ± 0.07*,•
AP cortical thickness, ML view(mm)	0.28 ± 0.03	0.31 ± 0.03*	0.21 ± 0.03•	0.24 ± 0.04◇

^*Significantly different compared to same genotype saline (p < 0.05).

^◇Trending to significance compared to same genotype saline (p = 0.05–0.1).

^•Significantly different compared to same treatment +/+ (p < 0.05).

Biomechanical Results

Extrinsic (Structural) Properties

Within the oim/oim groups, no treatment-related difference was observed for yield displacement, total displacement, or moment of inertia. However, stiffness was significantly greater for the RANK-Fc-treated compared to saline-treated mice and maximum load tended to increase. For the +/+ groups, stiffness was also greater for the RANK-Fc-treated mice, and total displacement tended to increase while yield displacement, maximum load, and moment of inertia were unchanged (Table 3).

TABLE 3.

Biomechanical data: structural and material properties of femora after 8 week of treatment

	+/+ Saline	+/+ RANK-Fc	oim/oim Saline	oim/oim RANK-Fc
Number of samples	20	19	16	13
Yield strength (kPa)	172.17 ± 25.95	168.91 ± 30.64	121.86 ± 26.35•	124.32 ± 27.66•
Yield strain	0.05 ± 0.01	0.05 ± 0.01	0.03 ± 0.01•	0.02 ± 0.01•
Yield displacement (mm)	0.24 ± 0.03	0.24 ± 0.06	0.16 ± 0.05•	0.15 ± 0.04•
Total displacement (mm)	0.5 ± 0.12	0.62 ± 0.28◇	0.18 ± 0.08•	0.15 ± 0.04•
Total strain	0.09 ± 0.02	0.12 ± 0.05*	0.03 ± 0.01•	0.03 ± 0.01•
Max load (N)	27.01 ± 6.18	29.83 ± 7.26	11.60 ± 2.33•	15.50 ± 3.21◇
Moment of inertia (mm4)	0.14 ± 0.05	0.16 ± 0.05	0.07 ± 0.01•	0.09 ± 0.03•
Young’s modulus (GPa)	6678 ± 1446	6849 ± 1620	6257 ± 1786	6773 ± 1430
Brittleness (%elasticity)	54.27 ± 16.96	47.85 ± 26.37	96.86 ± 10.14•	99.76 ± 0.87•
Toughness (N* mm)	11.59 ± 3.22	13.57 ± 5.44◇	1.94 ± 0.88•	1.91 ± 0.77•
Stiffness (N/mm)	182.22 ± 41.61	220.87 ± 43.54*	85.88 ± 19.55•	130.15 ± 23.63*,•

^*Significantly different compared to same genotype saline (p < 0.05).

^◇Trending to significance compared to same genotype saline (p = 0.05–0.1).

^•Significantly different compared to same treatment +/+ (p < 0.05).

Intrinsic (Material) Properties

RANKL inhibition did not affect yield strength, yield strain, brittleness, or Young’s modulus for either +/+ or oim/oim mice. Additionally, there were no significant differences in total strain or toughness for the oim/oim mice. Within the +/+ mice, total strain increased with treatment and toughness tended to increase (Table 3).

All extrinsic and intrinsic properties measured except for Young’s modulus were significantly different between +/+ and oim/oim mice for both saline and RANK-Fc (Table 3).

DISCUSSION

Antiresorptives are the most frequently used therapeutics for OI patients, in particular for children with OI [7]. In the current preclinical study, we examined the safety and efficacy of a RANKL inhibitor, RANK-Fc, in a growing mouse model of OI.

RANKL inhibition did not adversely affect weight gain or long bone growth over the 8-week treatment period in the oim/oim mice but it did lead to shortened femora in the +/+ mice. A similar study with ALN performed in 6-week-old oim/oim mice led to decreased femoral length, but in a follow-up study of oim/oim mice treated starting at 2 weeks of age, this adverse affect was not found [22, 24]. In the +/+ mice the femoral periosteal diameter in the ML view increased significantly in the current study as it did in the +/+ mice treated with ALN starting at 6 weeks of age, and as it tended to increase in the mice who were treated with ALN starting at 2 weeks of age [22, 24]. In addition, +/+ mice had significantly increased ML cortical thickness, similar to ALN mice in the earlier study [22]. Other similarities to ALN treatment were found in the increased metaphyseal density and increased stiffness in oim/oim bones [24].

We had expected a decrease in the endosteal diameter because of osteoclast inhibition with decreased resorption, similarly to what happens with ALN [24], but this was not observed. We also did not observe the oim/oim bone becoming less brittle and the +/+ bone more brittle with treatment as has been seen in studies with ALN [24]. In contrast to the ALN-treated mice which showed a reduction in fracture number, no significant differences were seen in the number of fractures in the oim/oim mice when treated with RANK-Fc. Thus, whereas RANKL inhibition improved some biomechanical and geometric properties of bone in the oim/oim mice in the current study, it did not decrease the fracture incidence.

Interestingly, a study by Delos et al. [33] that evaluated RANKL inhibition in fracture healing found that work to failure increased in the intact femora of treated +/+ mice compared to saline controls. We did not find comparable results in the current study, but the same type of testing was not performed. Delos et al. [33] performed 4-point bend tests, while 3-point bend tests were performed in the current study. The 4-point bend test in the Delos et al. study resulted in a completely elastic fracture of the bones, with no region of plasticity evident. However, since this is not the case for the 3-point bend tests, the results of the two cannot be considered comparable. In the Delos study, no biomechanical differences were found in the oim/oim mice with treatment [33], and in the current study only one biomechanical property was affected by treatment. Nonetheless, both studies show that RANKL inhibition did not have any detrimental effects with respect to mechanical strength of oim/oim bones, which is important for OI patients who may suffer many fractures even with the best treatments.

RANKL inhibition resulted in increased calcified cartilage under the growth plate, similar to the effect seen with ALN treatment [22]. This is likely the effect of suppression of osteoclasts that results in the primary spongiosa not being resorbed, and could in fact be beneficial for bone strength in patients with OI [39], although some recent studies suggest that this may not be the case [40]. Similar findings were observed in iliac crest biopsies from children with OI treated with ALN. These areas of unresorbed cartilage seem to correspond to the white lines commonly seen in radiographs of long bones in patients who have received bisphosphonate treatment [39].

In contrast to the previous ALN studies in the oim/oim mouse model, the current study did not find reduced fracture incidence in oim/oim mice treated with RANK-Fc, despite increased metaphyseal density and bone strength. This could be due to inadequate power to determine a fracture difference given the large standard deviations. For a power of 0.8 to detect a 50 % difference in fracture incidence 21 mice per group were required, and to detect a 25 % difference in fracture incidence, 79 mice per group were required. It is well known from a number of preclinical and clinical studies that bone mineral density (BMD) changes are not predictive of fracture risk. A clinical study of two different bisphosphonates, ALN and risedronate, showed that although the former produced a greater increase in BMD, the two therapies were similar in reducing fraction risk in a population of over 1000 postmenopausal women [41, 42]. A meta-analysis of trials of antiresorptives in osteoporosis showed that the magnitude of reductions in nonvertebral fracture risk was not associated with the magnitude of increased in BMD after 1 year of therapy [43].

There are a few recent studies that indicate that in addition to osteoclast inhibition, bisphosphonates may have other effects beneficial to bone such as prevention of osteocyte and osteoblast apoptosis [44, 45]. Although we did not investigate cell function in the current study, the absence of an effect on fracture reduction by RANKL blockade possibly could be explained by a lack of effect on osteoblast and osteocyte apoptosis. That being said, one of the few studies directly comparing the effects of a RANKL inhibitor versus a bisphosphonate on bone strength reported improvements in femur bending strength with OPG-Fc but not with ALN [46]. There are no published data suggesting significant differences in bone material properties with RANKL inhibitors versus bisphosphonates.

An alternative explanation of the lack of therapeutic efficacy with respect to fracture incidence concerns the number of fractures prior to the start of treatment. The mice in the current study accumulated most of their fractures prior to the start of the treatment period, which is clearly a limitation of the study. The increase in fracture number over the 8 week period among both treated and untreated mice was small compared to the number of fractures they already sustained by 6 weeks (1.8 fractures per mouse at baseline and 2.5 fractures per mouse in the RANK-Fc and saline-treated mice after 8 weeks of treatment). It is known that patients with OI, especially those with more severe disease, fracture more frequently as children than they do in later life [47]. This may be due to naturally increasing bone density with age, and especially with puberty. This phenomenon has also been confirmed in the Brittle IV mouse model of OI [29]. We did not treat our mice during the period of earliest infancy when they may be most susceptible to fractures because of uncertainty regarding how they would tolerate RANK-Fc at that early age.

Another important aspect with respect to fracture incidence concerns the fact that RANKL inhibition delays callus resorption [48]. Therefore, calluses in RANK-Fc-treated mice may still be apparent on a faxitron, whereas those fractures in their saline-treated counterparts may have healed with full callus resolution by the time of sacrifice. Consequently, the data presented here do not give us the definitive answer as to the potential of RANKL inhibitors to reduce fractures. Further studies should be performed where RANKL inhibitors are administered to younger animals, e.g., 2 weeks of age, similar to the McCarthy et al. ALN study [24]. Nonetheless, this initial study has demonstrated no obvious adverse effects of RANK-Fc in OI mice at the administered dose. Increments in BMD with RANK-Fc were numerically similar to increases found previously with ALN treatment using a similar study design. RANK-Fc improved certain femoral geometry parameters and improved femur bending strength and certain bone material properties. These factors together would support the further investigation of RANKL inhibitors as therapeutic agents for OI patients.