Risedronate Prevents Early Radiation-Induced Osteoporosis in Mice at Multiple Skeletal Locations

Abstract

Introduction

Irradiation of normal, non-malignant bone during cancer therapy can lead to atrophy and increased risk of fracture at several skeletal sites, particularly the hip. This bone loss has been largely attributed to damaged osteoblasts. Little attention has been given to increased bone resorption as a contributor to radiation-induced osteoporosis. Our aims were to identify if radiation increases bone resorption resulting in acute bone loss, and if bone loss could be prevented by administering risedronate.

Methods

Twenty-week old female C57BL/6 mice were either: not irradiated and treated with placebo (NR+PL); whole-body irradiated with 2 Gy X-rays and treated with placebo (IR+PL); or irradiated and treated with risedronate (IR+RIS; 30μg/kg every other day). Calcein injections were administered 7 and 2 days before sacrifice. Bones were collected 1, 2, and 3 weeks after exposure. MicroCT analysis was performed at 3 sites: proximal tibial metaphysis; distal femoral metaphysis; and the body of the 5th lumbar vertebra (L5). Osteoclasts were identified from TRAP-stained histological sections. Dynamic histomorphometry of cortical and trabecular bone was performed. Circulating TRAP5b and osteocalcin concentrations were quantified.

Results

In animals receiving IR+PL, significant (P < 0.05) reduction in trabecular volume fraction relative to non-irradiated controls was observed at all three skeletal sites and time points. Likewise, radiation-induced loss of connectivity and trabecular number relative to NR+PL were observed at all skeletal sites throughout the study. Bone loss primarily occurred during the first week post-exposure. Trabecular and endocortical bone formation was not reduced until Week 2. Loss of bone volume was absent in animals receiving IR+RIS. Histology indicated greater osteoclast numbers at Week 1 within IR+PL mice. Serum TRAP5b concentration was increased in IR+PL mice only at Week 1 compared to NR+PL (P = 0.05). Risedronate treatment prevented the radiation-induced increase in osteoclast number, surface, and TRAP5b.

Conclusion

This study demonstrated a rapid loss of trabecular bone at several skeletal sites after whole-body irradiation. Changes were accompanied by an increase in osteoclast number and serum markers of bone loss. Risedronate entirely prevented bone loss, providing further evidence that an increase in bone resorption likely caused this radiation-induced bone loss.

Introduction

Improvements in cancer treatment and diagnosis have helped increase long-term cancer survivorship in patients treated for several types of tumors [1, 2]. For example, of the more than 200,000 men who are diagnosed with prostate cancer each year, the 10 year survival rate approaches 90% [3]. Cancer treatments, including radiotherapy (RT), can damage normal tissue [4, 5]. As a consequence, this growing population of survivors can develop unique, long-term medical complications. A better understanding of the etiology and progression of these conditions is therefore necessary.

Skeletal fractures are recognized to be a late-effect from radiation therapy. Rib fractures have been documented in breast cancer patients: fracture rates range from 1.8% [6] to as high as 19% [7]; where contributing variables include total dose, dose per fraction, photon energy and duration of follow-up. Similarly, patients receiving RT for pelvic tumors are at increased risk for hip fractures [8–11]. The largest study performed to date examining this phenomenon (and to the authors’ knowledge the only examination to include non-irradiated controls) retrospectively identified a 65–216% greater risk of pelvic fractures for post-menopausal women receiving RT for cervical, rectal and anal cancers [12], with hazard ratios of 1.66, 1.65 and 3.16, respectively. Fracture rates outside the treatment volume were unchanged, suggesting radiation damages bones that directly absorb dose. Hip fractures have high rates of both mortality and morbidity, therefore it is important to understand and treat these radiation-induced fractures [13].

The cause of these radiation-induced fractures is unknown. Reduced bone density and fractures following direct exposure to ionizing radiation are thought to develop as a long-term response following damage to osteoblasts and vasculature within marrow and throughout osteons [12, 14, 15]. In vitro and in vivo data suggest that radiation can impair bone formation after exposure by killing or damaging osteoblasts; specifically decreasing osteoblast proliferation and differentiation; inducing cell cycle arrest; reducing collagen production; and increasing sensitivity to apoptotic agents [14, 16–19]. The contribution of late vascular damage to skeletal complications is less clear [15]. However, late radiation-induced bone atrophy has been observed in clinical [20] and animal studies [21, 22].

The effects of ionizing radiation on osteoclast activity are poorly documented, with a preponderance of the literature indicating a decline in osteoclast numbers and bone resorption activity [23–25]. However, some studies suggest an early stimulation of active bone resorption after exposure could contribute to the etiology of radiation-induced bone damage. We have demonstrated that circulating markers of bone resorption and osteoclast numbers are elevated in the proximal tibia of mice during the first 3 days of exposure to 2 Gy X-rays [26], which has been recently confirmed [27]. From these preclinical studies, the possibility exists that an early, radiation-induced increase in osteoclast activity can contribute to late-observed loss of bone density. Clinical evidence may support this acute, rapid bone loss: declines in volumetric bone mineral content have been reported in the lumbar vertebrae of cervical cancer patients relative to a pretreatment scan 5 weeks after the bone was exposed to either a 45 or 22.5 Gy total dose of high energy photons, with no subsequent recovery one year later [28].

The purpose of this study was to document the early, functional response of trabecular and cortical bone to ionizing radiation and to determine if this bone loss can be prevented by an antiresorptive agent. Specifically, we used a 2 Gy whole-body dose of X-rays applied to skeletally mature mice and quantified changes in osteoclast number, bone microarchitecture, and bone formation during the first three weeks after exposure. Previously, this 2 Gy dose was sufficient to substantially increase osteoclast number (+44%) and surface (+213%) three days after exposure; however, no reduction in bone volume as measured by microcomputed tomography (microCT) was observed. Our intent was to characterize how any changes in osteoclast numbers correspond temporally with bone loss. Additionally, we administered an antiresorptive agent, specifically the bisphosphonate risedronate (Procter and Gamble Pharmaceuticals), in an attempt to reduce any radiation-induced stimulation of osteoclasts. Bisphosphonates are a commonly used antiresorptive drug for treatment of osteoporosis and diseases of elevated bone turnover, functioning by decreasing osteoclast activity and inducing apoptosis [29]. The efficacy of risedronate in reducing bone loss would provide more support for the hypothesis that radiation rapidly stimulates active bone resorption as an early response.

Methods

Animals and Study Design

Twenty-week-old female C57BL/6 mice (n = 115 total) were examined in this study (Taconic Farms, Inc., Hudson, NY). The animals were received at 15 weeks of age and allowed to acclimate prior to irradiation at 20 weeks of age; food and water were available ad libitum. The Institutional Animal Care and Use Committee of Clemson University approved all procedures.

Animals were grouped to ensure similar mean body masses between groups at the outset of the study (~23 g; Table 1). Two groups were determined to receive whole body irradiation (n = 70 total), with the remainder receiving no irradiation and serving as non-irradiated controls, either as a baseline group (n = 10) or exposed to a sham irradiation procedure (n = 35). The baseline group was sacrificed immediately prior to the irradiation procedure, and tissues were harvested as described below.

Table 1.

Initial and final animal masses.

		Initial	Final	% Ch
Baseline		22.8 ± 0.5	na	na
Week 1	NR+PL	23.3 ± 0.2	23.4 ± 0.3	0.3
	IR+PL	23.2 ± 0.5	23.0 ± 0.5	−0.9
	IR+RIS	22.5 ± 0.2	23.0 ± 0.3	2.0c
Week 2	NR+PL	23.1 ± 0.2	23.0 ± 0.4	−0.5
	IR+PL	23.1 ± 0.3	22.6 ± 0.3	−2.5c
	IR+RIS	22.7 ± 0.4	22.7 ± 0.3	0.0
Week 3	NR+PL	22.9 ± 0.4	22.9 ± 0.5	−0.1
	IR+PL	23.1 ± 0.4	22.7 ± 0.4	−1.5c
	IR+RIS	22.8 ± 0.3	22.7 ± 0.3	−0.4

Non-irradiated mice receiving placebo injections (NR+PL); irradiated mice receiving placebo injections (IR+PL); irradiated mice receiving risedronate injections (IR+RIS) There are no differences in initial or final body mass between any of the groups.

^cindicates a significant increase or decrease in mass for a given group as determined by a paired t-test (P <0.05). All data are presented as mean ± SEM.

While under anesthesia (1.5% isoflurane), mice were irradiated in the prone position with a single field of 140 kVp X-rays to a single-fraction mid-plane dose of 2 Gy at a rate of 1.36 Gy/min (Philips Medical Systems; Bothell, WA). A 2 Gy dose is equivalent to the single daily fraction administered for treatment of many cancers, including pelvic tumors. Anaesthetized control mice were placed inside the inactive X-ray unit for the same amount of time as the irradiated animals, creating the sham procedure. The irradiation procedure served as the start of the experiment (Day 0).

One of the 2 Gy irradiated groups was selected to receive subcutaneous injections of the bisphosphonate risedronate (Procter and Gamble Pharmaceuticals; Cincinnati, OH) every other day starting immediately following the irradiation procedure at a dose of 30 μg/kg/injection (IR+RIS; n = 35). Equivalent volumes of PBS were injected as a placebo into the remaining 2 Gy irradiated (IR+PL; n = 35) and non-irradiated (NR+PL; n = 35) mice. Because of the observed rate and degree of osteoclast activation [26], 30 μg/kg injections were chosen rather than the more standard 5 μg/kg. This 6x increase in dose is equivalent to the greater risedronate dose Paget’s disease patients receive compared to post-menopausal osteoporosis (30 mg daily compared to 5 mg daily). To monitor new bone formation, a fluorescent calcein bone label (20 mg/kg) was injected subcutaneously at 7 days and 2 days prior to sacrifice for every mouse, excluding baseline.

Tissue collection

Mice from each treatment group were euthanized at 1, 2, and 3 weeks following the initial radiation exposure. Within each time point, 1/3 of the mice (n = 11–12) served as non-irradiated controls treated with placebo, 1/3 of the mice were irradiated and treated with placebo, and 1/3 of the mice irradiated and treated with risedronate. Each mouse was weighed then anesthetized using isoflurane, and blood was collected for serum analysis by cardiac puncture and exsanguination followed by cervical dislocation to ensure death. Serum was isolated by centrifugation, flash frozen in liquid nitrogen, and stored at −80°C. The left and right hind limbs as well as the vertebral column were collected for analysis. Hind limbs were removed and disarticulated. Tibiae and femora were cleaned of soft tissue and fixed in a solution of 10% formalin. Length measurements were obtained from the left femora. After 48 hours, the bones were placed in 70% ethanol. The vertebral column was cleaned and then frozen at −20°C for microcomputed tomography (microCT) analysis of the fifth lumbar vertebrae (L5).

Microcomputed tomography

The left tibiae were evaluated for trabecular microarchitecture using microCT (microCT 20, Scanco Medical AG; Bassersdorf, Switzerland), with isotropic voxels of 9 μm/side. An approximately 1 mm section of secondary spongiosa immediately adjacent to the primary spongiosa (starting 0.5 mm from the distal border of the growth plate) of the proximal growth plate was scanned. A total of 100 slices (0.9 mm) were traced and evaluated. Bone histomorphometric parameters for the proximal metaphysis of the tibia were measured as described in the report of the American Society of Bone and Mineral Research (ASBMR) Histomorphometry Nomenclature Committee [30]. Trabecular bone parameters including bone volume fraction (BV/TV), connectivity density (Conn.D), trabecular number (Tb.N), trabecular thickness (Tb.Th), trabecular separation (Tb.Sp), structure model index (SMI), and volumetric bone mineral density (vBMD) were quantified for all skeletal sites.

The L5 and distal femur were also analyzed via microCT (microCT 80, Scanco). L5 was isolated using the microCT X-ray scout view and scanned in its entirety (~3.5 mm) with a 10 μm voxel size. A section of the vertebral body measuring 0.5 mm immediately cranial to the caudal end plate was selected for analysis. This region was chosen because of its relatively high trabecular bone density and to minimize morphological differences between samples. The left femora were evaluated in 2 regions: the distal metaphysis and mid-diaphysis. A 1 mm volume of bone immediately adjacent to the primary spongiosa of the distal growth plate (starting 0.5 mm from the proximal edge of the growth plate and extending 1 mm proximally) was scanned and evaluated. A section of the mid-diaphysis measuring approximately 1.0 mm (~0.5 mm on either side of a point determined to be half-way along the length of the bone as determined from CT scans) was scanned, with the middle 0.3 mm evaluated to determine cortical porosity (Ct.Po) and polar moment of inertia (pMOI) within the diaphyseal bone.

Osteoclast and Osteoblast Identification

Following tomographic analysis, the left tibiae were decalcified using a formic acid solution (Immunocal; Decal Chemical Corp., Talman, NY) and embedded in a glycol methacrylate resin (ImmunoBed^™; Polysciences, Warrington, PA). The samples were cut into sagittal sections with a thickness of 3 μm using a microtome (Leica Microsystems, model RM2165; Witzlar, Germany). A subset of each group (n=5–6) was selected for analysis. Each slide was stained with TRAP using a commercial kit (Sigma, St. Louis, MO) to identify osteoclasts and counterstained with hematoxylin.

Histomorphometric evaluation was performed from captured micrographs (200X) throughout the metaphysis, starting approximately 0.25 mm distal from the growth plate (in order to exclude the primary spongiosa) and extending a further 0.5 mm. Analyses were performed using SigmaScan Pro Software (Systat Software, Inc., Richmond, CA). Surface measurements were quantified relative to total bone surface (BS, μm). These measurements included: osteoclast surface (Oc.S/BS; %); eroded surface with the inclusion of osteoclast surface (surface covered by Howship’s lacunae plus osteoclasts, [ES(Oc+)/BS], %); and eroded surface with the exclusion of osteoclast surface (surface covered by Howship’s lacunae, [ES(Oc−)/BS], %). The number of osteoclasts (N.Oc) within the region of interest along trabeculae of the secondary spongiosa was also determined (N.Oc/BS, mm⁻²). Osteoblast surface (Ob.S/BS; %) was also determined from these sections.

Quantitative Histomorphometry of Cortical Bone

Left femora were air-dried and embedded with noninfiltrating Epo-Kwick epoxy (Buehler, Lake Bluff, IL, USA). Disks that formed were sectioned using a low-speed saw (Buehler, 12.7 cm 9 0.5 mm, diamond blade) at the mid-diaphysis of the femur. Sections were polished to a flat, smooth surface using 600- and 1200-grit carbide paper followed by polishing with a cloth impregnated with 6-μm diamond paste. Micrographs of these bone cross sections were captured at 50X magnification under a UV light (400 nm). Quantitative histomorphometric analysis was performed using these images and SigmaScan Pro software. Measurements of femoral bone morphology included total tissue volume (Tt. V, mm³), marrow cavity volume (Ma.V, mm³), cortical bone volume (Ct.V; as calculated by Tt.V − Ma.V, mm³). Endocortical surface (Ec.BS, μm²) and periosteal surface (Ps.BS, μm²) were measured. Eroded surface was measured at the endorcortical surface (Ec.ES/Ec.BS, %). The green calcein labels injected 7 and 2 days prior to sacrifice allowed measurement of cortical bone formed between those times. Bone formation rate (BFR) was calculated by dividing the measured volume of new bone formed by 5 days, and is reported relative to the endocortical (Ec.BFR/BS, μm³/[μm²*day]) and periosteal (Ps.BFR/BS, μm³/[μm²*day]) bone surface. Mineralizing surface (MS), reported with the corresponding bone surface as a referent (MS/BS, %), was obtained by adding the ratio of double-labeled surface to bone surface (dLS/BS) to 50% of the ratio of single-labeled surface to bone surface (0.5* (sLS/BS)). Mineral apposition rate (MAR, μm/day) was calculated by dividing the corresponding bone formation rate per bone surface (BFR/BS) by the mineralizing surface per bone surface (MS/BS).

Endochondral Growth and Quantitative Histomorphometry of Trabecular Bone

The right tibiae were embedded in a methyl methacrylate resin (Osteobed^™) and cut into sagittal sections with a thickness of 5 μm using a tungsten carbide blade. Slides were unstained and coverslipped. Quantitative histomorphometric analysis was performed using SigmaScan Pro software on the micrographs captured at 10X magnification under a UV light (400 nm). Endochondral growth rate in the proximal tibia was determined by measuring the distance between double labels (20 samples per micrograph) found at the osseo-chondral junction at the distal portion of the growth plate. Histomorphometric evaluation was performed throughout the metaphysis, starting approximately 0.25 mm distal from the growth plate (again excluding the primary spongiosa) and extending a further 0.5 mm. The following parameters were measured or determined as was done for the cortical bone sections: BS (μm²); BFR/BS (μm³/[μm²*day]); MS/BS (%); and MAR (μm/day).

Serum Chemistry

Serum samples were analyzed for circulating markers of bone formation and resorption using ELISA kits for osteocalcin (Biomedical Technologies, Inc., Stoughton, MA) and tartrate-resistant acid phosphatase (TRAP5b) (Immunodiagnostic Systems, Inc., Fountain Hills, AZ), respectively. The analyses were performed according to protocols provided by the manufacturers.

Statistics

All data are presented as mean ± standard error of the mean. Significance was determined using SigmaStat version 3.5 (Systat Software Inc., Richmond, CA). A one-way ANOVA was used before irradiations to ensure consistent starting animal masses. A paired t-test was used to determine if mice gained or lost weight within an individual group during the period of study. All other statistical comparisons utilized a two-way ANOVA with a Holm-Sidak post-hoc test to determine significant effects of treatment (NR+PL v. IR+PL v. IR+RIS) and time (1 v. 2 v. 3 weeks). These comparisons include final body mass, microCT parameters, histomorphometric values and quantitative serum chemistry data. The threshold for significance for all tests was set at a 5% probability of a Type I error (P = 0.05).

Results

Body Mass

Body masses were similar across all treatment groups at the initiation of the study (P > 0.05; Table 1). Likewise, there were no differences in final body mass for any of the treatment groups. Relative to starting body mass, IR+RIS treated animals were significantly larger (+2%) at Week 1 compared to their starting mass, and IR+PL animals were significantly smaller at Weeks 2 (−2.5%) and 3 (−1.5% Table 1).

Radiation Associated Degradation in Trabecular Microarchitecture: Irradiated+Placebo vs Non-Irradiated

Differences in various trabecular structures of the proximal tibial metaphysis were observed between irradiated, placebo treated (IR+PL) and non-irradiated, placebo treated (NR+PL) control groups at several time points (Figure 1; Table 2). MicroCT images from representative tibia from each group sacrificed at the Week one time point are shown in Figure 2. Relative to non-irradiated controls, BV/TV was lower than IR+PL at Week 1 (−22%), Week 2 (−25%), and Week 3 (−30%; P < 0.01 each). Likewise, Conn.D. values were lower in IR+PL animals compared with NR+PL at all time points (−43%, −40%, and −53%, respectively; P < 0.05). Values for SMI were greater in IR+PL than non-irradiated control mice at Week 1 (+ 9%) and Week 3 (+12%; P < 0.05 each). Scores for IR+PL vBMD were 14% lower than non-irradiated controls at Week 1, 19% lower at Week 2, and 21% lower at Week 3 (P < 0.01 each). Significantly lower scores for Tb.N were observed at Weeks 1 (−13%) and Week 2 (−12%; P < 0.05 each) after irradiation compared with non-irradiated control, accompanied by elevated Tb.Sp values (+15% and +13%, respectively P < 0.05).

Figure 1.

Bone volume fracture (BV/TV), trabecular connectivity (Conn.D), and structure model index (SMI) as quantified using microCT within the secondary spongisa from the proximal tibia, distal femur, or fifth lumbar vertebra (L5) from mice. Animals either received a 2 Gy whole-body dose of X-rays together with placebo injections of PBS (IR+PL; red triangle symbols) or injections of risedronate (IR+RIS; green circles) throughout the study; or served as non-irradiated controls receiving PBS (NR+PL; blue squares). * Indicates significant difference compared to NR+PL within a given time point as determined by two-way ANOVA (P < 0.05). For data different from NR+PL within each time point, the percentage above each symbol represents the difference from non-irradiated controls. ^aIndicates significant difference compared to Week 1 data within a treatment group (P < 0.05); ^b Indicates significant difference compared to Week 2 data within a treatment group (P < 0.05). Differences between IR+PL and IR+RIS are not illustrated as they are significant for every comparison execpt: for Weeks 1 and 2 SMI at the proximal tibia; Weeks 1 and 2 SMI at the distal femur; and Weeks 1, 2 and 3 Conn.D at L5.

Table 2.

Trabecular bone properties quantified by microCT within the proximal tibial metaphysis, distal femoral metaphysis, and the body of L5 from mice receiving a whole-body, 2 Gy dose of X-rays.

			vBMD (mg HA/cm3)	Tb.N (1/mm)	Tb.Sp (μm)	Tb.Th (μm)
Proximal Tibia	Baseline		101.6 ± 2.4	2.95 ± 0.11	343 ± 41	51.3 ± 1.1
	Week 1	NR+PL	99.4 ± 4.0	2.82 ± 0.12	357 ± 17	54.5 ± 1.0
		IR+PL	85.5 ± 2.4*	2.46 ± 0.07*	411 ± 11*	55.3 ± 0.9
		IR+RIS	100.1 ± 2.6	2.84 ± 0.06	349 ± 9	52.8 ± 1.0
	Week 2	NR+PL	96.7 ± 3.3	2.77 ± 0.07	362 ± 10	54.3 ± 1.5
		IR+PL	78.1 ± 2.9*	2.44 ± 0.08*	410 ± 14*	52.5 ± 1.4
		IR+RIS	96.7 ± 2.7	2.89 ± 0.08	352 ± 11	52.4 ± 1.0
	Week 3	NR+PL	95.5 ± 2.4	2.61 ± 0.07	384 ± 11	54.4 ± 1.0
		IR+PL	75.4 ± 2.6* a	2.52 ± 0.10	400 ± 16	54.1 ± 1.5
		IR+RIS	108.4 ± 2.0* b	2.81 ± 0.07	359 ± 10	53.6 ± 1.4
	Distal Femur
Baseline		49.0 ± 4.5	3.21 ± 0.45	312 ± 4	36.9 ± 1.2
Week 1	NR+PL	41.0 ± 3.5	3.17 ± 0.06	317 ± 6	37.1 ± 1.4
	IR+PL	25.2 ± 2.5*	2.95 ± 0.05*	341 ± 6*	34.6 ± 1.0
	IR+RIS	54.2 ± 3.4*	3.28 ± 0.06	306 ± 6	38.2 ± 1.3
Week 2	NR+PL	43.7 ± 2.3	3.07 ± 0.04	326 ± 4	37.1 ± 1.4
	IR+PL	24.4 ± 2.9*	2.81 ± 0.04*	359 ± 5*	36.0 ± 1.2
	IR+RIS	47.5 ± 3.2	3.21 ± 0.05*	313 ± 5	37.1 ± 0.9
Week 3	NR+PL	39.6 ± 4.2	2.99 ± 0.05a	336 ± 6b	37.2 ± 1.3
	IR+PL	24.4 ± 2.8*	2.81 ± 0.05*	360 ± 6* b	34.9 ± 1.9
	IR+RIS	61.4 ± 3.9* b	3.29 ± 0.06*	304 ± 5*	39.1 ± 1.1
L5 Vertebra
	Baseline		292.5 ± 5.4	6.63 ± 0.07	144 ± 2	45.3 ± 0.9
	Week 1	NR+PL	268.0 ± 7.7	6.26 ± 0.08	154 ± 3	46.2 ± 0.5
		IR+PL	259.8 ± 4.4	6.25 ± 0.09	155 ± 3	45.7 ± 0.5
		IR+RIS	308.1 ± 5.0*	6.49 ± 0.12	146 ± 3	49.2 ± 0.6*
	Week 2	NR+PL	263.6 ± 5.2	6.46 ± 0.12	149 ± 3	44.6 ± 0.7
		IR+PL	238.6 ± 6.0* a	6.04 ± 0.07*	161 ± 2*	43.7 ± 0.8
		IR+RIS	296.0 ± 4.4*	6.51 ± 0.12	145 ± 3	47.4 ± 0.5*
	Week 3	NR+PL	262.0 ± 5.7	6.52 ± 0.14	150 ± 4	43.2 ± 0.5 a
		IR+PL	239.9 ± 6.5* a	5.96 ± 0.13*	164 ± 4*	44.7 ± 0.5
		IR+RIS	302.2 ± 3.8*	6.54 ± 0.10	145 ± 3	48.5 ± 0.6

Notes: Non-irradiated mice receiving placebo injections (NR+PL); irradiated mice receiving placebo injections (IR+PL); irradiated mice receiving risedronate injections (IR+RIS); volumetric bone mineral density (vBMD); trabecular number (Tb.N); trabecular separation (Tb.Sp); and trabecular thickness (Tb.Th). All values aregiven as mean ± SEM.

^*Indicates significant difference compared to NR+PL within a given time point as determined by two-way ANOVA (P < 0.05);

^aIndicates significant difference compared to Week 1 data within a treatment group (P < 0.05);

^bIndicates significant difference compared to Week 2 data within a treatment group (P < 0.05).

Figure 2.

MicroCT images of trabecular bone in the proximal tibial metaphysis (Tibia), distal femoral metaphysis (Femur), and within the vertebral body of the fifth lumbar (L5) vertebrae from mice sacrificed at Week 1 after receiving: a 2 Gy dose of X-rays with a PBS placebo injected subcutaneously (IR+PL); a 2 Gy dose and subcutaneous risedronate injections (IR+RIS); and non-irradiated animals receiving PBS (NR+PL). The scale bar represents 500 μm.

Radiation as a single treatment was associated with generally impaired trabecular microarchitectural indices relative to non-irradiated control in the distal femoral metaphysis (Figure 1, Table 2). BV/TV was lower in IR+PL compared with control at Weeks 1 (−32%), 2 (−39%), and 3 (−43%; P < 0.01 each). Differences at these time points were accompanied by a radiation-induced decline of vBMD (−39%, −44%, and −38%, respectively P < 0.01), Tb.N (−7%, −8%, −6%, respectively P < 0.05), and greater Tb.Sp (+10%, +10%, and +7%, respectively P < 0.01) compared to NR+PL. Conn.D was lower than control by Week 2 (−67%) and Week 3 (−65%; P < 0.01 each) after radiation exposure. Likewise, SMI scores were greater at Weeks 2 and Week 3 (+15% each) after irradiation compared with non-irradiated control mice.

Trabecular microarchitectural properties of L5 also exhibited differences with radiation (Figure 1, Table 2). BV/TV (−9%, −15%, and −11%) were lower at Weeks 1, 2, and 3, respectively (P < 0.05 each). SMI scores were elevated at all time points after exposure relative to control (+70%, +48%, and +34%, respectively P < 0.01). Conn.D. was lower at Week 3 after irradiation compared with control (−21%, P < 0.01). At Week 2 and Week 3 after irradiation, IR+PL animals exhibited lower vBMD (−10% and −8%, respectively P < 0.05), Tb.N (−7% and −9%, respectively P < 0.05), and higher Tb.Sp (+8% and +10%, respectively P < 0.05) than NR+PL.

Prevention of Radiation Associated Changes In Trabecular Microarchitecture: Radiation + Risedronate vs Non-Irradiated

Irradiated mice receiving risedronate generally had microarchitectural parameters similar to non-irradiated controls (Figure 1, Table 2). The few differences between these two groups generally represent improved microarchitecture rather than radiation-induced degradation. Within the proximal tibia, vBMD was greater than NR+PL at Week 3 (+13%, P < 0.01). Likewise, BV/TV (+21%, P < 0.05) and Conn.D (+36%, P < 0.05) were greater at Week 3 relative to NR+PL.

For microarchitectural parameters quantified within the distal femur, BV/TV was greater for IR+RIS relative with NR+PL at Week 3 (+50%, P < 0.05). Conn.D. was greater in the IR+RIS groups relative with non-irradiated controls at Week 1 and 3 after irradiation (+87% and +134%, respectively P < 0.01 for each). SMI scores were lower than control at Week 3 (−10%, P < 0.01). The vBMD was higher than control at Week 1 (+10%) and Week 3 (+55%, P < 0.05 each). Tb.N was higher than control at Week 3 (+10%, P < 0.001). Tb.Sp. was lower than control at Week 3 (−9%, P < 0.001).

Within L5, BV/TV was greater than control at all time points (+15%, +14%, + 25%, respectively, P < 0.001 for each). Conn.D. was lower than control at Week 3 (−17%, P < 0.05). SMI scores were lower than control at Week 1 (−47%), Week 2 (−48%), and Week 3 (−74%, P < 0.01 each). The vBMD was higher at all time points after irradiation versus control (+15%, +12%, and +15%, respectively P < 0.001). Tb.Th was higher than control at Week 1 (+7%), Week 2 (+6%), and Week 3 (+12%, P < 0.01 each).

Differences in Trabecular Microarchitecture Across Time within NR+PL, IR+PL and IR+RIS Groups

There were some within group differences in microarchitectural parameters for the NR+PL, IR+PL and IR+RIS animals across time. For NR+PL, Tb.N within the femur at Week 3 was less than Week 1 (P < 0.05). Additionally, for NR+PL animals the SMI values in L5 were greater at Week 2 and 3 relative to Week 1 (P < 0.05). Conn D., values in L5 were higher at Week 3 were greater than at Week 1 (P < 0.05) and Tb.Th in L5 was lower at Week 3 than at Week 1 (P < 0.05).

Within IR+PL animals, the vBMD scores in the proximal tibia at Week 3 were lower than Week 1 (P < 0.05); vBMD of L5 at Weeks 2 and 3 were lower than Week 1 (P < 0.05); and Tb.Sp in the distal femur was higher at Week 3 relative to Week 1 (P < 0.05).

In the IR+RIS groups there were no trabecular microarchitectureal differences across time for L5. However, BV/TV and vBMD scores within the proximal tibia were greater at Week 3 than Week 2; and SMI Week 3 values were less than Week 2. There were several time-associated differences within the distal femur. Scores for BV/TV and vBMD were greater at Week 3 than Week 2. SMI scores for Week 3 was less than Week 1, and Conn.D for Weeks 1 and 2 were less than Week 3.

Rate of Change in Trabecular Microarchitecture Across Time between NR+PL and IR+PL Groups

A majority of the bone loss from irradiation appears to occur within the first week after exposure. To formally quantify this observation, linear regression was used to determine if the rate of change (slope) in trabecular parameters between the non-irradiated placebo treated and irradiated placebo treated mice were similar between Week 1 and Week 3 (Figure 3). Of the three parameters examined (BV/TV, Conn.D., and SMI) at all three skeletal sites (9 parameters), only the rate of change in Conn.D. within L5 was dissimilar (P < 0.05). Though many microarchitectural indices are significantly different between NR+PL and IR+PL groups across times and skeletal locations, the rate of change (slope) between these treatment groups were not significantly different from Week 1 to Week 3.

Figure 3.

Bone volume fracture (BV/TV), trabecular connectivity (Conn.D), and structure model index (SMI) as quantified using microCT within the secondary spongisa from the proximal tibia, distal femur, or fifth lumbar vertebra (L5) from mice. Animals either received a 2 Gy whole-body dose of X-rays together with placebo injections of PBS (IR+PL; red triangle symbols) throughout the study or served as non-irradiated controls receiving PBS (NR+PL; blue squares). The equation of the regression line from Week 1 to Week 3 is provided. * P < 0.05 Indicates a difference between the slopes of the lines as determined by linear regression.

Histology

At Weeks 1 and 3, IR+PL animals exhibited significantly higher ES(OC+)/BS (+76%, +73%, respectively P < 0.01 each, Table 3) compared to non-irradiated controls. Additionally, Oc.S/BS (+113%), and Oc.N/BS (+218%) were also higher at Week 1 post-irradiation relative to NR+PL, P < 0.01 each (Figure 4). At Weeks 1, 2 and 3 IR+RIS animals displayed lower ES(Oc+)/BS and Oc.S/BS values compared to IR+PL (−41%, −34%, −43%; and −43%, −45%, 49%, respectively P < 0.05 each, Table 3). Oc.N/BS was lower at Weeks 1 and 2 for IR+RIS compared to IR+PL (−44%, −48%; P < 0.05, Figure 4).

Table 3.

Histological parameters quantified within the proximal tibial metaphysis from mice irradiated with 2 Gy X-rays.

		BS (um2)	Ob.S/BS (%)	ES(Oc+)/BS (%)	ES(Oc−)/BS (%)	Oc.S/BS (%)
Baseline		4455 ± 752	31.5 ± 5.3	18.0 ± 4.0	7.0 ± 4.0	11.0 ± 5.0
Week 1	NR+PL	3729 ± 236	32.3 ± 4.4	14.4 ± 4.0	6.0 ± 2.1	8.4 ± 2.1
	IR+PL	2591 ± 209*	29.9 ± 3.9	25.3 ± 2.5*	7.5 ± 1.2	17.8 ± 2.2*
	IR+RIS	4909 ± 247* #	18.0 ± 6.9	15.0 ± 1.4#	4.8 ± 1.8	10.2 ± 1.6#
Week 2	NR+PL	4446 ± 281	22.1 ± 1.6	18.5 ± 2.9	6.3 ± 1.2	12.2 ± 2.2
	IR+PL	2992 ± 494*	22.5 ± 2.5	26.3 ± 2.2	10.4 ± 1.3	15.9 ± 1.6
	IR+RIS	3291 ± 433a	23.7 ± 4.7	17.3 ± 1.6#	8.6 ± 2.7	8.8 ± 1.4#
Week 3	NR+PL	3849 ± 502	20.0 ± 1.0	14.9 ± 2.8	5.2 ± 1.5	9.7 ± 2.0
	IR+PL	3390 ± 276	24.3 ± 4.5	25.7 ± 4.4*	5.1 ± 2.7	16.4 ± 2.9
	IR+RIS	3762 ± 265a	25.9 ± 3.6	14.6 ± 2.5#	6.3 ± 2.3	8.3 ± 2.0#

Notes: Non-irradiated mice receiving placebo injections (NR+PL); irradiated mice receiving placebo injections (IR+PL); irradiated mice receiving risedronate injections (IR+RIS); bone surface (BS); eroded surface with the inclusion of osteoclast surface [ES(Oc+)/BS]; eroded surface with the exclusion of osteoclast surface [ES(Oc−)/BS]; osteoclast surface (Oc.S), and osteoblast surface (Ob.S). All data are given as mean ± SEM.

^*Indicates significant difference compared to NR+PL within a given time point as determined by two-way ANOVA (P < 0.05).

^#Indicates a difference between IR+RIS and IR+PL within a given time point as determined by two-way ANOVA (P < 0.05).

Figure 4.

The number of osteoclasts (Oc.N) relative to the bone surface (BS) within the proximal tibial metaphysis of mice at Week 1, 2, and 3 after initiating the experiment where mice either received a 2 Gy whole-body dose of X rays then daily injections of PBS (IR+PL; open triangles); 2 Gy X-rays with daily injections of risedronate (IR+RIS; open circles); or served as non-irradiated controls (NR+PL; open squares). * Indicates significant difference compared to NR+PL within a given time point; ^# Indicates a difference between IR+RIS and IR+PL within a given time point as determined by two-way ANOVA (P < 0.05).

Endochondral Growth Rate

Endochondral growth rate was significantly lower in IR+PL (2.25 ± 0.21 μm/day) and IR+RIS (1.86 ± 0.21 μm/day) relative to NR+PL (3.16 ± 0.23 μm/day) at Week 2 only (P < 0.01); however endochondral growth rate was slow and in all groups and timepoints was below 3.5 μm/day (Data not shown).

Trabecular Bone Formation

Bone formation between groups was similar along the trabeculae of the proximal tibial metaphysis at Week 1 after irradiation (Table 4). By the end of Week 2, BFR had declined in both irradiated groups, though much more pronounced in IR+RIS (−94%) than IR+PL (−37%) relative to NR+PL, P < 0.001 (Table 4). NR+PL remained elevated at Week 3 relative to IR+PL and IR+RIS (−38% and −50%, respectively P < 0.01 each). A similar pattern was observed for MAR. Interestingly, MAR scores were actually significantly greater for IR+PL than both IR+RIS as well as for NR+PL by Week 1 (Table 4; P < 0.05), as BFR/BS was slightly higher in the irradiated group than for control or risedronate-treated animals.

Table 4.

Quantitative measures of new bone formation and mineralization from trabecular bone within theproximal tibial metaphysis from mice irradiated with 2 Gy X-Rays.

		BFR/BS (μm3/[μm2 day])	MS/BS (%)	MAR (μm/day)
Week 1	NR+PL	0.71 ± 0.07	46.8 ± 3.5	1.51 ± 0.11
	IR+PL	0.88 ± 0.10	43.3 ± 1.9	2.05 ± 0.28*
	IR+RIS	0.65 ± 0.05	41.3 ± 2.0	1.57 ± 0.08#
Week 2	NR+PL	0.65 ± 0.11	46.8 ± 6.2	1.36 ± 0.12
	IR+PL	0.41 ± 0.06* a	39.7 ± 3.4	1.00 ± 0.08a
	IR+RIS	0.04 ± 0.02*#a	20.6 ± 3.0*#a	0.16 ± 0.09*#a
Week 3	NR+PL	0.74 ± 0.06	48.3 ± 1.6	1.53 ± 0.11
	IR+PL	0.46 ± 0.07* a	35.3 ± 1.0*	1.28 ± 0.16a
	IR+RIS	0.33 ± 0.08*ab	35.2 ± 5.5* b	0.80 ± 0.17*#ab

Notes: Non-irradiated mice receiving placebo injections (NR+PL); irradiated mice receiving placebo injections (IR+PL); irradiated mice receiving risedronate injections (IR+RIS); bone formation rate per bone surface (BFR/BS); mineralizing surface per bone surface (MS/BS);and mineral apposition rate (MAR).

^*Indicates significant difference compared to NR+PL within a given time point as determined by two-way ANOVA (P < 0.05).

^#Indicates a difference between IR+RIS and IR+PL within a given time point as determined by two-way ANOVA (P < 0.05).

^aIndicates significant difference compared to Week 1 data within a treatment group (P < 0.05);

^bIndicates significant difference compared to Week 2 data within a treatment group (P < 0.05).

Cortical Histomorphometry and Bone Formation

No differences were observed for endocortical or periosteal BS, Ma.V, Ct.V, Ct.Po or pMOI within each time point (Table 5). However, the resorption at the endocortical surface was different between groups by Week 3: Ec.ES/Ec.BS for IR+PL was greater than NR+PL (+68%, P < 0.05) and IR+RIS (+155%, P = 0.01).

Table 5.

Cortical bone properties calculated at femur mid-diaphysis based on quantitative histomorphometry of embedded cross sections as well as microCT measurements.

		Ec.BS (mm2)	Ps.BS (mm2)	Ma.V (mm3)	Ct.V (mm3)	Ec.ES/Ec.BS (%)	Ct.Po (%)	pMOI (mm4) * 103
Week 1	NR+PL	3.32 ± 0.06	4.80 ± 0.04	0.798 ± 0.028	0.852 ± 0.008	12.2 ± 5.4	5.55 ± 0.22	333 ± 7.0
	IR+PL	3.29 ± 0.10	4.69 ± 0.05	0.769 ± 0.045	0.820 ± 0.018	21.9 ± 5.4	5.58 ± 0.14	334 ± 11.9
	IR+RIS	3.41 ± 0.05	4.81 ± 0.03	0.840 ± 0.024	0.828 ± 0.017	11.9 ± 7.1	5.86 ± 0.23	330 ± 8.1
Week 2	NR+PL	3.37 ± 0.06	4.88 ± 0.06	0.811 ± 0.023	0.860 ± 0.023	19.2 ± 2.5	5.95 ± 0.24	333 ± 6.0
	IR+PL	3.40 ± 0.06	4.83 ± 0.08	0.833 ± 0.026	0.819 ± 0.008	18.4 ± 1.8	6.39 ± 0.28	323 ± 8.1
	IR+RIS	3.37 ± 0.08	4.88 ± 0.06	0.806 ± 0.027	0.850 ± 0.009	10.1 ± 2.5	6.60 ± 0.20	325 ± 8.7
Week 3	NR+PL	3.36 ± 0.03	4.89 ± 0.05	0.816 ± 0.013	0.865 ± 0.016	14.1 ± 2.4	6.61 ± 0.32 a	331 ± 10.9
	IR+PL	3.36 ± 0.03	4.82 ± 0.04	0.813 ± 0.013	0.833 ± 0.015	23.7 ± 5.0 *	6.31 ± 0.27	325 ± 8.5
	IR+RIS	3.35 ± 0.05	4.84 ± 0.05	0.812 ± 0.022	0.865 ± 0.015	9.3 ± 1.7 #	6.58 ± 0.24	332 ± 9.5

Notes: Non-irradiated mice receiving placebo injections (NR+PL); irradiated mice receiving placebo injections (IR+PL); irradiated mice receiving risedronate injections (IR+RIS); endocortical (Ec) and periosteal (Ps.) Bone surface (BS); marrow volume (Ma.V); cortical volume (Ct.V); eroded surface along the length of the endocortical surface (ES/Ec.BS); cortical porosity (Ct.Po) and polar moment of inertia (pMOI). All data are given as mean ± SEM.

^*Indicates significant difference compared to NR+PL within a given time point as determined by two-way ANOVA (P < 0.05).

^#Indicates a difference between IR+RIS and IR+PL within a given time point as determined by two-way ANOVA (P < 0.05).

^aIndicates significant difference compared to Week 1 data within a treatment group (P < 0.05).

At the endocortical surface bone formation was similar at the first week after irradiation between all groups (Table 6). By Week 2, Ec.BFR/BS in IR+PL had declined by 54% relative to NR+PL and by 75% in IR+RIS (P < 0.001 each; Table 6). By Week 3, Ec.BFR/BS remained reduced in the IR+RIS group relative to control (−45%, P < 0.01). No changes were observed in BFR at the periosteal surface between any group or time point after exposure. Ec.MS/BS was likewise lower than NR+PL at both Week 2 and 3 for IR+PL (−46% and −36%, respectively P < 0.001) as well as IR+RIS (−69% and −41%, respectively P < 0.01). IR+RIS was lower at Week 2 than at Week 1 (P < 0.01). IR+RIS was lower at Week 2 than at Week 1 (P < 0.01), and NR+PL was lower at Week 2 and 3 than at Week 1 (P < 0.01 each).

Table 6.

Quantitative measures of new bone formation and mineralization from cortical bone at the femoral mid-diaphysis from mice irradiated with 2 Gy X-Rays.

		BFR/BS (μm3/[μm2 day])		MS/BS (%)		MAR (μm/day)
		Endocortical	Periosteal	Endocortical	Periosteal	Endocortical	Periosteal
Week 1	NR+PL	0.80 ± 0.07	0.64 ± 0.04	42.2 ± 3.2	26.5 ± 2.0	1.91 ± 0.11	2.48 ± 0.15
	IR+PL	0.49 ± 0.11	0.57 ± 0.13	28.8 ± 5.5	25.7 ± 5.3	1.69 ± 0.15	2.09 ± 0.18
	IR+RIS	0.57 ± 0.09	0.60 ± 0.10	29.4 ± 3.4	25.6 ± 3.3	1.87 ± 0.13	2.32 ± 0.25
Week 2	NR+PL	0.82 ± 0.07	0.66 ± 0.04	44.5 ± 1.4	33.6 ± 1.1	1.85 ± 0.13	1.95 ± 0.10 a
	IR+PL	0.37 ± 0.09 *	0.58 ± 0.06	24.0 ± 6.0 *	28.7 ± 2.6	1.57 ± 0.06	2.00 ± 0.10
	IR+RIS	0.21 ± 0.03 * a	0.43 ± 0.09	13.6 ± 1.6 * a	24.9 ± 4.8	1.49 ± 0.06	1.70 ± 0.08 a
Week 3	NR+PL	0.71 ± 0.09	0.67 ± 0.04	43.6 ± 4.4	33.1 ± 1.4	1.61 ± 0.06	2.01 ± 0.09 a
	IR+PL	0.49 ± 0.07	0.63 ± 0.05	27.9 ± 2.7 *	32.3 ± 1.9	1.72 ± 0.12	1.95 ± 0.09
	IR+RIS	0.39 ± 0.06 *	0.62 ± 0.06	25.6 ± 3.9 *	32.0 ± 3.1	1.55 ± 0.09	1.92 ± 0.06

Notes: Non-irradiated mice receiving placebo injections (NR+PL); irradiated mice receiving placebo injections (IR+PL); irradiated mice receiving risedronate injections (IR+RIS); bone formation rate per bone surface (BFR/BS); mineralizing surface per bone surface (MS/BS); and mineral apposition rate (MAR).

^*Indicates significant difference compared to NR+PL within a given time point as determined by two-way ANOVA (P < 0.05).

^aIndicates significant difference compared to Week 1 data within a treatment group (P < 0.05).

Serum Chemistry

Elevated TRAP5b concentrations were observed within IR+PL individuals at Week 1 (+21%) versus NR+PL (Table 7). No subsequent differences were observed between irradiated and non-irradiated animals at Week 2 or Week 3. Risedronate-treated animals had significantly lower values of circulating TRAP5b at Week 1 (−36%), Week 2 (−35%), and Week 3 (−37%, P < 0.01 each) relative to NR+PL (Table 7). Serum osteocalcin levels for IR+RIS at Week 1 were lower than IR+PL (−51%, P < 0.05), with no other differences across time and treatment (Table 7).

Table 7.

Circulating serum concentration of TRAP5b and osteocalcin as quantified from mice receiving 2 Gy of X-Rays.

		TRAP5b (U/L)	Osteocalcin (ng/ml)
Baseline		6.60 ± 0.28	81 ± 12
Week 1	NR+PL	8.88 ± 0.53	116 ± 21
	IR+PL	10.71 ± 0.56*	128 ± 19
	IR+RIS	5.66 ± 0.42* #	63 ± 11#
Week 2	NR+PL	8.37 ± 0.88	111 ± 22
	IR+PL	9.25 ± 0.57	128 ± 24
	IR+RIS	5.43 ± 0.52* #	80 ± 15
Week 3	NR+PL	8.06 ± 0.58	102 ± 19
	IR+PL	8.96 ± 0.87	112 ± 24
	IR+RIS	5.07 ± 0.33*#	73 ± 11

Notes: Non-irradiated mice receiving placebo injections (NR+PL); irradiated mice receiving placebo injections (IR+PL); irradiated mice receiving risedronate injections (IR+RIS); TRAP5b is a serum marker for osteoclast activity; Osteocalcin is a serum marker for bone formation;All data are given as mean ± SEM.

^*Indicates significant difference compared to NR+PL within a given time point as determined by two-way ANOVA (P < 0.05).

^#Indicates a difference between IR+RIS and IR+PL within a given time point as determined by two-way ANOVA (P < 0.05).

Discussion

In women with pelvic tumors, bones that are directly exposed to ionizing radiation are more prone to fracture after RT than non-irradiated skeletal elements [12]. Reduced bone mineral content has been observed 5 weeks after exposure in cervical cancer patients within the field of radiation [31]. Using an animal model exposed to a whole body dose of X-rays, we show that radiation results in an increase in osteoclast number and bone resorption causing a decline in trabecular bone volume fraction by the first week after exposure, in the absence of reduced bone formation. Substantial loss of bone and declines in the microarchitectural parameters of trabecular bone is evident at multiple skeletal locations, and this deterioration remains present for weeks. Treatment with risedronate prevented loss of bone and impaired microarchitecture, and suppressed the post-exposure elevation in osteoclast numbers lining the tibial metaphyses.

Ionizing radiation appears to stimulate bone loss as an acute response within the boundaries of this animal model and from limited clinical observations. Early trabecular bone loss likely occurred as a result of an increase in active bone resorption during the first week. This increase in bone resorption was followed by a reduction in bone formation for the remainder of the study along the trabeculae. Atrophy of bone is typically detected at late-time points after exposure: trabecular bone loss in mice in the proximal tibia has been observed at 4 months following 2 Gy doses of γ-rays and high LET carbon and iron ions [22], with bone loss at 4 months after a 1 Gy dose of protons [21]. An early, active increase in bone resorption could contribute to the reduced bone density and increased fracture rates among the directly irradiated bones of cancer RT patients and irradiated animal models, perhaps in combination with the subsequent reduction in bone formation. Likewise, commercially available therapeutic agents that suppress bone turnover might be effective in reducing these fractures, should these observations translate into the clinic.

Historically, the deleterious effects of ionizing radiation on bone has been largely attributed to reduced bone formation resulting from arrested proliferation of osteoblast precursors, osteoid production, and cell death [14, 16–19]. A reduction in BFR/BS was observed at Weeks 2 and 3 within the tibial metaphysis of irradiated mice, and at Week 2 along the endocortical surface. Bone formation has been shown to be suppressed within rabbit tibiae exposed to a 50 Gy electron beam from the first examination at 4 weeks until 24 weeks [32]. In primate models, osteoblasts and mesenchymal progenitors within the periosteum and marrow of mandibles together with new osteoid production were largely absent by 6 months after a 45 Gy total dose applied in fractions over 12 days [33]. In vitro studies have shown that irradiation can reduce alkaline phosphatase activity and expression levels at 2 Gy of X–rays in C2C12 mouse myoblast cells [18]. Studies using MC3T3-E1 osteoblast-like cells under different conditions have shown radiation can induce terminal differentiation after a 5 Gy X-rays [34], reduce collagen production and proliferation following irradiation of 4 Gy [17], or induce cell cycle arrest and increase sensitivity to apoptotic agents after exposure to doses of 15 and 30 Gy of γ-rays, respectively [19].

An increase in active bone resorption, as opposed to decreased formation, appears to have primarily contributed to the observed loss of bone in the present study, which used a substantially lower dose of radiation than many in vivo investigations noting reduction in bone formation. The main deleterious effect of radiation on bone appears to have occurred during the first week after exposure. While irradiation resulted in substantially less bone by the first week relative with non-irradiated controls, the rate of change in bone volume and microarchitecture between control and irradiated groups was similar from Week 1 to Week 3 (Figure 3). BV/TV was lower at all skeletal sites and all time points early after exposure. Irradiated trabeculae were generally more disconnected and rod-like, as represented by increases in SMI. Such rapid changes (primarily during the first week) would suggest an increase in bone resorption via osteoclast activity, especially without a noted reduction in osteoblast surface, bone formation rate as determined from dynamic histomorphometry, and osteocalcin concentration. The number of TRAP+ osteoclasts was greater in IR+PL mice than non-irradiated controls at Week 1 only. In mature C57BL/6 mice receiving 2 Gy X-rays, a very early (on Day 3) increase in circulating TRAP5b as well as osteoclast numbers within the proximal tibial metaphysis has been shown to occur after exposure [26]. Indeed, the bone loss in this study occurred in the presence of elevated osteoclast numbers, eroded surface, as well as TRAP5b concentration, all of which indicate an increase in active resorption after exposure. These observations are consistent with early reductions in trabecular bone functional parameters observed in rodents after whole-body exposure to radiation, though from the proximal tibiae of weanling rats exposed to a 6 Gy whole body dose of X-rays and within the tibia of a high bone density mouse model after a 2 Gy exposure [27, 35]. It seems likely that this very early increase in osteoclast number and activity accounts for the severe loss of bone and deteriorated microarchitecture observed by Week 1.

Reduction in bone volume and declines in many bone mircoarchitectural indices (e.g., Conn.D) was virtually absent in animals receiving both radiation and risedronate injections. Baseline scores were largely preserved across parameters and at all sites. Within the metaphysis, bone formation in IR+RIS was unchanged relative to NR+PL at Week 1, followed by a drastic reduction in formation at Week 2. However, bone resorption in IR+RIS (as indicated by circulating TRAP5b levels) was substantially reduced by Week 1 relative to NR+PL, and remained lower throughout the study. This reduction in resorption occurred despite having elevated osteoclast numbers by Week 1 relative to control (79%), though these differences were not significant with two-way ANOVA comparison (P = 0.01 using one-way ANOVA with same Holm Sidak followup; data not shown). Thus, it is presumed that suppression of bone resorption in the IR+RIS animals via action of the bisphosphonate mitigated the bone loss that was observed in IR+PL animals. These observations again support a radiation-induced increase in active bone resorption occurring early after exposure that contributed greatly to the observed deficits in bone.

The effect of radiation on cortical bone structure and overall bone strength is poorly understood. We report no change in overall morphology of the femoral mid-diaphysis with radiation or risedronate treatment throughout the study, though Ec.ES/Ec.BS is elevated within a month after exposure. Thus, at the endocortical surface, resorption of bone is elevated early after exposure, without a corresponding increase in porosity, but with a decline in bone formation at Week 2. Prior studies have shown significant increases in cortical bone porosity and reductions in strength following high-dose irradiation [32, 36]. After exposure to a single 50 Gy dose, rabbit tibiae showed increased intracortical porosity between 12 and 52 weeks [32]. Sections of cortical bone from these irradiated proximal metaphyses showed reduced ultimate strength by 12 weeks, with slight recovery by 52 weeks, coincident with the increase in porosity. Fractionation seems to have a profound effect on changes in fracture strength of the mid-diaphysis of rat tibiae [36]. High doses of irradiation (40 – 60 Gy) administered in 1.25 Gy twice-daily fractions resulted in no changes in maximal load to fracture, while single dose application, or higher fractionated doses (2.5 Gy daily), reduced maximum load to fracture. Recent studies have found no changes in cortical parameters following exposure to doses of 2 Gy or less of several types of radiation in the femur of mice (e.g., protons, iron and carbon ions, and photons) at 4 months after exposure [21, 22]. However, an ~ 0.5 Gy dose of mixed types of ion radiation, including iron, helium, protons, and other nuclear fragments resulted in increased cortical porosity within the humeral diaphysis of mice at 9 weeks after exposure [37]. Thus, while many studies suggest lower doses or fractionation can result in a limited effect on long-term cortical bone morphology and strength, the true nature of the relationship remains undefined at this time.

One of the many unresolved issues regarding the mechanism behind radiation-induced bone loss is whether the cellular and structural changes result from systemic factors, direct effects, or a combination of both. For this study the entire animal was exposed to X-rays, which can alter the function of several organ systems important in bone homeostasis, such as the hypothalamic-pituitary axis. Radiation-induced reduction in growth hormone and subsequent IGF-1 production has been shown to be associated with long-term reduced BMD in childhood cancer survivors [38]. Additionally, we have recently identified that heavy ion (Fe²⁶⁺) radiation is associated with long term loss of trabecular and cortical bone at a distant, non-irradiated site, though correlated with overall loss of body mass [39]. Thus, we do not discount that exposure could have induced a systemic response that resulted in elevated osteoclast activity. However, in clinical conditions, reduced bone volume or density appears to be isolated within the irradiated bone treatment volume [20, 31]. Likewise, the risk of fracture after treatment (e.g., the femoral neck) seems constrained to the bones that absorb radiation and not distant sites (e.g., the wrist) [12]. Therefore, while future examinations of radiation-induced osteoporosis must target skeletal elements and avoid exposing organs such as the gonads, we hypothesize that the bone loss occurred largely as a direct effect of radiation on osteoclasts and cells within the irradiated bones.

In conclusion, a whole body, 2 Gy dose of X-rays resulted in an early increase in osteoclast number and bone loss at several skeletal locations in female C57BL/6 mice. The deterioration of trabecular bone occurred primarily during the first week, when bone formation was unaltered and without a radiation-induced reduction of body mass. Loss of bone was prevented by the administration of an antiresorptive agent, the bisphosphonate risedronate. The mechanisms behind how radiation could induce an increase in osteoclast number early after exposure, leading to acute bone loss, is unknown. Radiation represents a substantial challenge to bone health, and thus mechanisms inducing these changes must be more thoroughly studied in order to effectively prevent bone loss and fractures in patients receiving radiotherapy.