Bisphosphonate withdrawal: Effects on bone formation and bone resorption in maturing male mice

Abstract

Bisphosphonates are being increasingly used to treat pediatric patients with skeletal disorders. However, the effects of long-term bisphosphonate therapy and cessation of therapy during growth are unclear. Thus, studies were undertaken to determine the effects of alendronate discontinuation following treatment of C57Bl/6 mice during the period of rapid skeletal growth. Compared to vehicle-treated mice, 16 weeks of alendronate treatment starting at 18 days of age resulted in a 3.7 fold increase in trabecular bone in the setting of suppressed bone formation. Alendronate therapy for 8 weeks followed by 8 weeks of vehicle treatment resulted in a more pronounced increase in trabecular bone compared to mice treated with alendronate for 16 weeks (1.7 fold) and to vehicle-treated controls (6.5 fold). Mice that received alendronate for 8 weeks followed by 8 weeks of vehicle exhibited increased osteoblast surface (2.5 fold), mineralizing surface (5.7 fold), and bone formation rate (5.1 fold) compared to mice treated continuously with alendronate. However, these parameters were not restored to the levels observed in the vehicle-treated mice. Thus, partial resumption of bone formation upon cessation of bisphosphonate therapy leads to a greater increase in trabecular bone than that seen when bisphosphonates are administered continuously to growing mice. These data suggest that intermittent administration of bisphosphonates may optimize their beneficial effects on the growing skeleton.

Introduction

Bisphosphonates are used to prevent bone loss and fractures in disorders associated with increased bone resorption (1,2). In addition to being used for treating osteoporosis in mature individuals, bisphosphonates are increasingly being used to treat children with skeletal disorders such as osteogenesis imperfecta and fibrous dysplasia, as well as to treat or prevent bone loss in children with cerebral palsy, steroid-induced osteoporosis, Duchenne muscular dystrophy, and idiopathic hypercalciuria (3–8).

Despite the increasing use of bisphosphonates in the pediatric population, the effects of discontinuation of long-term bisphosphonate therapy are still unclear. In children with osteogenesis imperfecta who received pamidronate for 4 or more years, sustained increases in lumbar spine bone mineral content (BMC) and density (BMD) and a modest increase in serum NTx (9,10) were observed in the 18–24 month follow-up period. The paucity of histomorphometric analyses of bone following cessation of bisphosphonate therapy during growth has limited our knowledge of the consequences of bisphosphonate withdrawal on the growing skeleton.

Studies in rodent models have been performed to determine the consequences of long-term bisphosphonate therapy during growth. In mice undergoing rapid skeletal growth, bisphosphonates increase metaphyseal bone and diaphyseal cortical bone thickness without impairing long bone growth (11). However, the consequences of bisphosphonate withdrawal were not examined in these studies. In mature ovariectomized rats, continued suppression of bone formation (12,13) and bone resorption, assessed by deoxypyridinoline and Oc.S/BS (13), was observed 8 months after cessation of therapy.

To examine the consequences of bisphosphonate discontinuation, C57Bl/6 male mice were treated with alendronate for 8 weeks starting day 18 and then treatment was discontinued. The mice were euthanized 8 weeks later as were mice that had received continuous vehicle or alendronate treatment for 16 weeks. Studies were undertaken to determine the effects of alendronate withdrawal on skeletal microarchitecture and bone formation in growing mice.

Materials and Methods

Animal experiments

All experimental procedures were approved by the Massachusetts General Hospital IACUC. Mice housed 3–5/cage were maintained in a pathogen-free facility and subjected to a 12-h light, 12-h dark cycle. Male C57Bl/6 littermates (Jackson Labs, Bar Harbor, Maine) were administered either saline (VEH, n = 5) or alendronate (Sigma, 100 μg/kg subcutaneously twice weekly; the dose used for studies of rapidly growing oim/oim mice, a model of osteogenesis imperfecta (14)) starting at 18 days of age. To assess the effects of long-term continuous administration of alendronate, mice were treated for 16 weeks (ALN-Cont, n = 5). To evaluate the consequences of cessation of therapy, alendronate was discontinued after 8 weeks of treatment and mice received saline twice weekly for the following 8 weeks. Power calculations based on previous data (11) demonstrated that to detect half the difference observed in those studies (in BV/TV, Ct.Th, pMOI, Imax, and Imin) at 95% power and α = 0.05, a sample size of 3 per group was sufficient, thus three mice were examined in the ALN-On/Off group. Mice were randomly distributed and analyses were performed blinded to treatment group.

Serum biochemical markers

Serum levels of procollagen type 1 N-terminal propeptide (P1NP, Rat/Mouse P1NP EIA, IDS, Gaithersburg, MD) and collagen type 1 C-telopeptide (RatLaps CTX-1 EIA, IDS, Gaithersburg, MD) were evaluated after 16 weeks of alendronate and/or vehicle treatment.

Static and Dynamic Histomorphometry

To assess the effects of alendronate on bone formation and resorption, static and dynamic histomorphometric analyses were performed according to the criteria established by the American Society for Bone and Mineral Research (15). Calcein was administered 10 and 3 days prior to sacrifice. Left femurs were fixed, subjected to μCT analyses, dehydrated, and embedded in poly methylmethacrylate. Five micron thick longitudinal sections were stained with Goldner’s Trichrome for evaluation of static histomorphometric parameters. Measurements were performed in the distal femoral metaphysis 500 μm proximal to the growth plate and in the femoral diaphysis 2–4 mm proximal to the growth plate using an Osteomeasure image analyzer. Images were visualized with a Nikon E800 with each field of view 400×600 μm² (20X objective, final magnification 200X). Six fields were measured in the metaphysis and 10 fields in the diaphysis. Mineralizing surface was determined by measuring the total extent of the double labeled surface plus half the extent of the single labeled surface, divided by the Bone Surface [MS = (dLS + sLS/2)/BS] (15). Mineral apposition rate was determined by dividing the distance between the two calcein labels by the time interval between the two calcein injections (7 days) and expressed as μm/day. Enzymatic detection of tartrate resistant acid phosphatase (TRAP) was performed by incubating sections in TRAP staining solution for 60 minutes at 37°C (16).

Bone mineral density and bone microarchitecture

Total body areal bone mineral density (BMD, g/cm²) was measured by dual-energy X-ray absorptiometry (PIXImus, GE Lunar, Madison, WI). μCT imaging was performed on the distal metaphysis and mid-diaphysis of the left femur using a high-resolution desktop imaging system (μCT40, Scanco Medical AG, Brüttisellen, Switzerland) in accordance with the ASBMR guidelines for the use of μCT in rodents (17). Individual femurs were stabilized in agarose in a scanning tube. After density calibration, scans were acquired using a 12 μm³ isotropic voxel size, 70 kVp and 114 mA peak x-ray tube potential and intensity, 200 ms integration time and were subjected to Gaussian filtration. Femoral lengths were determined by measuring the distance from the femoral head to the distal end of femoral condyles on μCT scans. Trabecular bone microarchitecture was evaluated in the distal metaphysis and in the diaphysis. The distal metaphyseal region analyzed began 240 μm (20 slices) above the peak of the distal growth plate and extended proximally 1.8 mm (150 slices). The diaphyseal region analyzed began 1.8 mm (150 slices) above the peak of the distal growth plate and extended proximally 3.0 mm (250 slices). Cortical bone was evaluated in the mid-diaphysis in a region that started 55% of the bone length below the femoral head and extended 600 μm (50 slices) distally. Thresholds of 486 and 733 mg HA/cm³ were used for evaluation of trabecular and cortical bone, respectively, based on adaptive-iterative thresholding (AIT) that was performed on the control group. Cancellous bone outcomes included trabecular bone volume fraction (BV/TV, %), trabecular thickness (Tb.Th, μm), trabecular separation (Tb.Sp, μm), and trabecular number (Tb.N, mm⁻¹), which was calculated using the distance transformation method. Cortical bone outcomes included cortical tissue mineral density (Ct.TMD, mg HA/mm³), marrow area (Ma.Ar, mm²), cortical area (Ct.Ar, mm²), total cross-sectional area (Tt.Ar, mm²), cortical thickness (Ct.Th, μm), cortical bone area fraction (Ct.Ar/Tt.Ar, %), cortical porosity (Ct.Po, %), polar moment of inertia (pMOI, J, mm⁴), and the maximum and minimum moments of inertia (I_max and I_min, mm⁴).

Statistics

Statistical analyses were performed using a one-way ANOVA (Prism 6.07, La Jolla, CA). Tukey HSD post-hoc comparisons of means were performed when ANOVA was significant. p-values of <0.05 indicated significance.

Results

C57Bl/6 male mice were randomly assigned at 18 days of age to receive saline for 16 weeks, alendronate for 16 weeks, or alendronate for 8 weeks followed by saline for 8 weeks. Mice were sacrificed after 16 weeks of treatment. Neither alendronate treatment regimen altered long bone growth, based on microCT evaluation of femoral length (Table 1). Both alendronate treatment regimens increased total body areal BMD compared to vehicle treatment. Alendronate treatment suppressed serum levels of the bone resorption marker, CTX-I, and this suppression persisted 8 weeks after discontinuation of alendronate therapy. Neither alendronate treatment regimen altered serum P1NP levels. While neither alendronate treatment regimen altered body weight relative to vehicle treated mice, the weight of mice in which alendronate was discontinued was greater than that of the group that received continuous alendronate treatment (Figure 1A).

Table 1.

Femoral length, total body aBMD, and serum biochemical markers.

	VEH (n=5)	ALN-Cont (n=5)	ALN-On/Off (n=3)
Femur Length [mm]	15 ± 0.3	15 ± 0.4	15 ± 0.3
BMD [mg/cm2]	48.5 ± 1.3	52.2 ± 0.1a	57.0 ± 1.8a,b
CTX-I [ng/mL]	10.2 ± 2	4.8 ± 1a	4.1 ± 2a
P1NP [ng/mL]	44 ± 9	44 ± 5	37 ± 3

Data were analyzed on samples obtained at the time of euthanasia. Data represent Mean ± SEM of 3–5 mice per group. VEH, saline treated mice; ALN-Cont, mice treated with alendronate for 16 weeks (d18-d130); ALN-On/Off, mice treated with alendronate for 8 weeks (d18-d74), followed by saline for 8 weeks (d75-d130).

^aP < 0.05 compared to VEH;

^bP < 0.04 compared to ALN-Cont.

Figure 1.

Effects of alendronate treatment and withdrawal on weight and skeletal microarchitecture in growing mice. (A) Body weight of mice from d18 to d130. Dashed line indicates the alendronate withdrawal time point (8-week) for the ALN-On/Off group. Data represent Mean ± SEM. a = P < 0.05 ALN-Cont vs. ALN-On/Off at 16-weeks. (B) Representative microCT images of the distal femur. (C) BV/TV = bone volume/tissue volume; Tb.Th = trabecular thickness; Tb.Sp = trabecular separation; Tb.N = trabecular number. Data represent Mean ± SEM. a = P < 0.05 compared to VEH; b = P < 0.05 compared to ALN-Cont by Tukey HSD post-hoc comparisons of means following ANOVA. P < 0.05.

To further characterize the increase in areal BMD, skeletal microarchitecture was evaluated by μCT. The trabecular BV/TV was 3.7 fold higher in the 16-week continuous alendronate treatment group compared to vehicle treated mice (Figure 1B, C). When alendronate was discontinued after 8 weeks, a more pronounced increase in BV/TV was observed (6.5 fold compared to vehicle treated mice and 1.7 fold compared to mice treated continuously with alendronate). Neither alendronate treatment regimen altered trabecular thickness compared to vehicle treated mice. Both alendronate treatment regimens led to a lower trabecular separation. While continuous alendronate treatment resulted in a 2.5 fold increase in trabecular number when compared to mice treated with vehicle, discontinuation of alendronate led to a further increase in trabecular number compared to mice treated continuously with alendronate (1.3 fold) or vehicle (3.2 fold). These data suggest that an increase in bone formation following discontinuation of alendronate results a more significant increase in trabecular bone than that seen with continuous alendronate therapy.

To investigate alterations in cortical bone microarchitecture, μCT analyses of the femoral diaphyses were performed. While neither alendronate treatment regimen altered cortical thickness, cortical tissue mineral density, or cortical area fraction on μCT (Table 2), alendronate withdrawal led to an increase in marrow area and cortical area compared to vehicle treated mice. Total cross-sectional area was greater in the alendronate withdrawal group compared to both the vehicle and continuous alendronate treatment groups. A modest increase in cortical porosity was observed 8 weeks after discontinuation of alendronate relative to that seen in the vehicle or continuous alendronate treatment groups.

Table 2.

Cortical bone microCT

	VEH	ALN-Cont	ALN-On/Off
Cortical Thickness [μm]	157 ± 6.9	158 ± 6.1	170 ± 4.7
Cortical TMD [mg HA/cm3]	1150 ± 20	1140 ± 10	1140 ± 10
Marrow Area [mm2]	1.12 ± 0.07	1.30 ± 0.03	1.56 ± 0.11#
Cortical Area [mm2]	0.71 ± 0.04	0.75 ± 0.04	0.91 ± 0.02#
Total Cross-sectional Area [mm2]	1.83 ± 0.11	2.06 ± 0.06	2.47 ± 0.12*
Cortical Area Fraction [%]	38.7 ± 0.90	36.5 ± 0.91	36.9 ± 1.30
Cortical Porosity [%]	0.29 ± 0.08	0.30 ± 0.02	0.57 ± 0.09*
pMOI [mm4]	0.36 ± 0.06	0.41 ± 0.03	0.60 ± 0.03*
Imax [mm4]	0.25 ± 0.04	0.26 ± 0.02	0.39 ± 0.02*
Imin [mm4]	0.11 ± 0.02	0.15 ± 0.01	0.21 ± 0.01*

Data represent Mean ± SEM of 3–5 mice per group. VEH, saline treated mice; ALN-Cont, mice treated with alendronate for 16 weeks (d18-130); ALN-On/Off, mice treated with alendronate for 8 weeks (d18-d74), followed by saline for 8 weeks (d75-d130).

^*P < 0.05 compared to VEH and ALN-Cont;

^#P < 0.05 compared to VEH

The effects of alendronate treatment on inferred biomechanical parameters were evaluated by μCT analyses. Continuous alendronate treatment did not alter femoral diaphyseal morphology. However, a significant increase in pMOI, I_max, and I_min was observed 8 weeks after discontinuation of alendronate relative to that seen in the vehicle or continuous alendronate treatment groups.

To evaluate the effects of alendronate withdrawal on metaphyseal bone formation and resorption, dynamic histomorphometry was performed on the distal femoral metaphysis of treated and control mice. Continuous alendronate treatment led to a decrease in trabecular Ob.S/BS compared to vehicle treatment (Figure 2 A and B). Although the Ob.S/BS remained lower than that of vehicle treated mice, cessation of alendronate therapy led to an increase in Ob.S/BS by 2.5 fold compared to mice treated continuously with alendronate. While both alendronate treatment regimens led to a decrease in MS/BS and BFR/BS compared to vehicle treated mice, the MS/BS and BFR/BS observed 8 weeks after alendronate discontinuation was 5.7 and 4.5 fold higher than that of mice receiving continuous alendronate treatment. Thus, these studies suggest that resumption of bone formation after cessation of alendronate underlies the increase in trabecular bone observed.

Figure 2.

Effects of alendronate treatment and withdrawal on trabecular and cortical bone histomorphometry in growing mice. (A) Endosteal cortical bone calcein double labels (inserts) and trabecular osteoblasts (yellow triangle). Trabecular (B) and cortical endosteal (C) and periosteal (D) bone histomorphometry parameters. Ob.S/BS = osteoblast surface/bone surface; MS/BS = mineralizing surface/bone surface; BFR/BS = bone formation rate/bone surface; MAR = mineral apposition rate. Data represent Mean ± SEM. a = P < 0.05 compared to VEH; b = P < 0.05 compared to ALN-Cont by Tukey HSD post-hoc comparisons of means following ANOVA. P < 0.05.

To evaluate the effects of alendronate withdrawal on cortical bone formation, dynamic histomorphometry was performed on the femoral mid-diaphysis of treated and control mice. Both alendronate treatment regimens led to a decrease in endosteal MAR compared to vehicle treated mice (Figure 2 A and C). However, cessation of alendronate therapy led to a 2.5 fold increase in endosteal MS/BS and a 2.9 fold increase in endosteal BFR/BS compared to mice treated continuously with alendronate. Neither alendronate treatment regimen altered periosteal MAR, MS/BS, or BFR/BS (Figure 2D).

Both alendronate treatment regimens led to a >6 fold increase in osteoclast number and a >3 fold increase in Oc.S/BS relative to vehicle treated mice. The size of the osteoclasts was also significantly larger in the alendronate treated mice (Figure 3).

Figure 3.

Alendronate increases osteoclast number, surface, and size. (A) Representative Trichrome and TRAP stain. Arrows indicate osteoclasts. (B) Oc.N = osteoclast number; Oc.S/B.S = osteoclast surface/bone surface; Oc.Ar = osteoclast area. Data represent Mean ± SEM. a = P < 0.05 compared to VEH by Tukey HSD post-hoc comparisons of means following ANOVA. P < 0.05.

Discussion

These studies demonstrate novel consequences of discontinuation of alendronate in rapidly growing mice. Notably, a marked increase in trabecular bone was observed in mice that received alendronate for 8 weeks and saline for the following 8 weeks, compared to mice treated continuously with alendronate or vehicle. This occurs in the setting of increased osteoblast surface, mineralizing surface and bone formation rate, and continued suppression of bone resorption based on serum CTX-I. Interestingly, the cortical parameters of mice in which alendronate was discontinued also differ significantly from those of mice that received continuous vehicle or alendronate treatment. While the modest increase in weight observed at 16 weeks in the ALN-On/Off mice vs ALN-Cont mice could contribute, the weight of the ALN-On/Off mice did not differ from that of the vehicle treated mice at any time point.

The effects of stopping alendronate after rapid skeletal growth are not well studied. In mature ovariectomized rats (4–6 months old), alendronate-mediated suppression of bone formation and resorption persists for 8 months after cessation of therapy, as assessed by dynamic histomorphometry (12,13). However, osteoblast surface was not evaluated in these studies. These data contrast with our study, which demonstrates a partial recovery of bone formation and osteoblast surface 2 months after discontinuation of alendronate, leading to a more pronounced increase in trabecular bone following cessation of treatment than is seen with continuous treatment with alendronate. The discrepancy between these studies may reflect the consequences of ovariectomy or the ability of the growing skeleton to resume bone formation after alendronate withdrawal, while bone resorption remains suppressed. Although the current studies cannot distinguish between the effects of alendronate withdrawal and the total cumulative dose of alendronate, our investigations were designed to reflect clinical practice where the administered dose is constant but the duration of therapy differs (18).

Despite suppressed serum CTX-I levels, both alendronate treatment regimens led to a significant increase in osteoclast surface and size. This contrasts with investigation in rodents which demonstrate that alendronate therapy decreases osteoclast number and surface post-ovariectomy (12,13). However, increases in osteoclast number and size are observed in postmenopausal women treated long-term with alendronate (19). Unlike clodronate or etidronate that inhibit bone resorption by promoting osteoclast apoptosis, alendronate is thought to impair osteoclast activity primarily through disruption of GTPases that regulate the formation of the ruffled border (20). It has been postulated that the increased osteoclast surface and size observed with prolonged alendronate treatment is due to an increase in osteoclast lifespan associated with a prolonged duration of the DNA-fragmentation phase of apoptosis (19).

The major limitations of these investigations are the absence of biomechanical end points and the concern that studies in growing animals may not precisely predict consequences of interventions in children, since unlike humans, mice do not exhibit intracortical remodeling. However, these studies do provide important insight into the basis for the effects of therapeutic interventions. The current investigations have clinical implications for the use of long-term bisphosphonate therapy and suggest that introduction of “drug holidays” may maximize the increase in bone in bisphosphonate-treated children, while decreasing the risks that accompany longterm bisphosphonate therapy. In adults, the benefit of intermittent discontinuation of bisphosphonates is less clear, since while bone formation, mineralizing surface, and biochemical markers of bone turnover increase after cessation of therapy, bone mineral density decreases (19,18). In addition, an increase in clinical vertebral fractures is observed 5 years after alendronate withdrawal compared to that observed in those treated continuously with alendronate (18). However, 2 years after stopping bisphosphonates, an increase in bone mineral density and bone mineral content associated with increased bone resorption markers is observed in children (9). Recent case reports (21–24) of atypical femoral fractures in pediatric patients receiving long-term bisphosphonate therapy highlight the risks of chronic bisphosphonate therapy in growing children. The pre-clinical data presented in the current studies suggest that intermittent “drug holidays” may present an opportunity for resumption of bone formation in the setting of continued suppression of bone resorption, thereby enhancing the increase in bone seen with bisphosphonates while minimizing the long term risks of chronic bisphosphonate therapy on the growing skeleton.