Transient Disturbance in Physeal Morphology Is Associated With Long-Term Effects of Nitrogen-Containing Bisphosphonates in Growing Rabbits

Authors

  • Elisabeth J Smith,

    1. Bone and Mineral Research Program, Garvan Institute of Medical Research, St Vincent's Hospital and University of New South Wales, Sydney, Australia
    2. Orthopaedic Research Unit, Childrens Hospital Westmead, Sydney, Australia
    Search for more papers by this author
  • David G Little,

    1. Orthopaedic Research Unit, Childrens Hospital Westmead, Sydney, Australia
    Search for more papers by this author
  • Julie N Briody,

    1. Orthopaedic Research Unit, Childrens Hospital Westmead, Sydney, Australia
    Search for more papers by this author
  • Anthony McEvoy,

    1. Orthopaedic Research Unit, Childrens Hospital Westmead, Sydney, Australia
    Search for more papers by this author
  • Nicholas C Smith MSc,

    Corresponding author
    1. Orthopaedic Research Unit, Childrens Hospital Westmead, Sydney, Australia
    • Bone and Mineral Research Program Garvan Institute of Medical Research 384 Victoria Street Darlinghurst, NSW 2010 Sydney, Australia
    Search for more papers by this author
  • John A Eisman,

    1. Bone and Mineral Research Program, Garvan Institute of Medical Research, St Vincent's Hospital and University of New South Wales, Sydney, Australia
    Search for more papers by this author
    • Dr Eisman served as a consultant, received corporate appointments and research support from Aventis, Eli Lilly and Company, MSD, NPS Pharmaceuticals, Novartis, Organon, Roche, and Servier. Dr Little has a patent licensing arrangement from Novartis. All other authors have no conflict of interest

  • Edith M Gardiner

    1. Bone and Mineral Research Program, Garvan Institute of Medical Research, St Vincent's Hospital and University of New South Wales, Sydney, Australia
    Search for more papers by this author

Abstract

Bisphosphonates have clinical benefit in children with severe osteogenesis imperfecta or osteoporosis and potential benefit in children with Perthes disease or undergoing distraction osteogenesis. However, there is concern about the effects of bisphosphonates on the physis and bone length. In 44 growing rabbits, zoledronic acid caused a transient disruption of physeal morphology, retention of cartilaginous matrix in trabeculae and cortical bone of the metaphysis, and a minor decrement in tibial bone length at maturity.

Introduction: Data from growing animal models suggest that bisphosphonates cause retention of longitudinal cartilaginous septa at the chondro-osseous junction, extension of trabeculae to the metaphyseal-diaphyseal junction, and varying dose-dependent effects on longitudinal growth. However, there is a lack of data regarding effects of intermittent use of nitrogen-containing bisphosphonates on the physis and on tibial length in models reaching maturity.

Materials and Methods: Contralateral tibias of juvenile rabbits were examined after right tibial distraction osteogenesis from two previous studies. Animals were randomized to receive 0.1 mg/kg zoledronic acid (ZA) IV at 8 weeks of age (ZA1) or 8 and 10 weeks of age (ZA2) or saline. Body mass was analyzed from 5 to 44 weeks of age; tibial length and proximal physeal-metaphyseal histology and histomorphometry were analyzed at 8–52 weeks of age.

Results: Tibial length was 3% less at 14 weeks of age in the ZA2-treated versus saline group (p < 0.05) in both studies, and this difference persisted at maturity in the long-term study group (26 weeks of age, p < 0.05). Total body mass gain from 5 to 26 weeks of age was 14% less in ZA2-treated than saline animals (p < 0.05). Rate of weight gain from 8 to 10 weeks of age was 76% less in ZA2 compared with saline animals (p < 0.05). Radiographs showed radiodense lines in the metaphyses of ZA-treated bones, corresponding to the number of doses. Histologically, lines resulting from the first dose of ZA contained longitudinal cartilaginous matrix cores surrounded by bone, whereas those from the second dose contained spherical cores of matrix caused by transient disruption of physeal morphology after the first dose of ZA. Resorption of these lines at later times was radiographically and histologically evident, but remnants of cartilaginous matrix remained in the cortical bone of ZA-treated animals.

Conclusions: ZA treatment within the final 13.5% of the rabbit tibial growth period caused a transient disruption in physeal morphology and resorption associated with retention of cartilaginous matrix and coinciding with a persistent 3% decrement in tibial length. Disruption of physeal morphology and potential loss of bone length should be considered when administering nitrogen-containing bisphosphonates to children before closure of the major physes.

INTRODUCTION

BISPHOSPHONATES HAVE A strong affinity for bone mineral and inhibit bone resorption through a direct effect on osteoclast function. (1) Studies have shown their ability to prevent bone loss, increase BMD, and reduce fracture risk in osteoporotic adults. (2, 3) Bisphosphonates are widely prescribed in children with systemic conditions such as osteogenesis imperfecta (OI), juvenile rheumatoid arthritis (JRA), and juvenile idiopathic osteoporosis (JIO). (4–8). Recent animal studies suggest that bisphosphonates may benefit children undergoing distraction osteogenesis for leg lengthening and in childhood osteonecrosis. (9–14) Because these latter children usually are systemically normal with only local skeletal deficiencies, there are concerns about the impact of bisphosphonates on the open physis, chondro-osseous modeling, and consequent growth in otherwise normal long bones.

Early studies on the effects of bisphosphonates on the growth plate in young rats and of alendronate in mice noted the persistence of calcified cartilaginous columns in the metaphysis after treatment, (15–17) and other animal studies have shown reductions in overall bone length. (9, 18–20) Studies in children receiving continuous dosing of bisphosphonates have recorded the appearance of band-like sclerosis and increased metaphyseal density on radiographs. (5, 7, 21, 22) However, there is a lack of data from histological examinations of the growth plate and cortex both immediately and in the long-term after noncontinuous dosing with potent nitrogen-containing bisphosphonates in a well-described growth model. The New Zealand white rabbit is an ideal model in which to record the effects of systemic intervention on growth. Detailed growth data concerning the percentile distributions of tibial lengths at successive time periods, growth increments, and the relation of these parameters to body weight have been well documented. (23)

This study examined the immediate and long-term effects of singular and repeat dosing of zoledronic acid on physeal morphology, tibial length, and remodeling of endochondral cartilaginous matrix at the chondro-osseous junction and in the metaphysis of the nonoperated tibia from animals of two previous studies of distraction osteogenesis, in which only the distracted callus in the operated tibia had been analyzed. (14) Animals in these studies were growing New Zealand white rabbits and were followed out to 14 weeks of age in the short-term study and 52 weeks of age in the long-term study. The aim of this study was to examine the short- and long-term effects of bolus intravenous doses of zoledronic acid on physeal morphology, metaphyseal cartilaginous matrix remodeling, long-term longitudinal growth, and final bone length at adulthood.

MATERIALS AND METHODS

Animals and drug administration

The effect of bisphosphonates was studied on the nonoperated limb of 66, male New Zealand white rabbits that had undergone right-sided tibial distraction osteogenesis at 8 weeks of age. (9, 15) Two groups of animals were studied independently. In the initial study the latest time-point was 14 weeks of age; in the longer second study, the last collection occurred at 52 weeks of age. All animal groups were statistically equivalent with respect to body weight and tibial length at the earliest time-points of both short- and long-term studies. Animals were randomized for infusion with either saline (controls), 0.1 mg/kg zoledronic acid at 8 weeks of age (ZA1), or the same dose at both 8 and 10 weeks of age (ZA2). (11) ZA was prepared in saline from the hydrated disodium salt, molecular weight 399.5 (Novartis Pharma AG, Basel, Switzerland). Buprenorphine 0.05 mg/kg was administered at the end of surgery and again 12 h postoperatively. The animals were supplied with rabbit pellet and water ad libitum. These studies were conducted with the approval of the institutional animal ethics committee (WAEC 177.08-01).

Tibial length and body mass

Longitudinal growth data were collated from length measurements of the left nonoperated tibias made from DXA scans. Data were available from the long-term study at 10, 12, 14, 20, 26, weeks of age in n = 6 control and ZA2 animals and in n = 2 animals of the same groups at 36, 42, and 52 weeks of age. From the short-term study, data were available at 8, 10, 12, and 14 weeks of age in 10 control and 12 ZA2 animals. For DXA scans, a total body dual energy X-ray densitometer (LUNAR DPX, Madison, WI, USA) was used with the tibia oriented in the antero-posterior (AP) projection, using software specifically designed for measuring small animals (LUNAR DPX, Small Animal Software, 1.0c; Lunar, Madison, WI, USA). The HiRes <0.5 kg Slow scan mode was used (fine collimation, sample size of 0.6 × 1.2 mm, and sample interval of 1/16 s). Scans were digitized at 1.66 pixels/mm resolution in the horizontal plane and tibial length was measured using the ruler function available in the software's Manual Analysis Option and determined as the distance from the tip of the tibial intercondylar eminence to the tip of the lateral malleolus in the anterior-posterior (AP) plane. Body mass (kilograms) was recorded at 6, 7, 8, 8.5, 9, 10, and 12 weeks of age in the short-term study group and at 5, 6, 8, 10, 12, 14, 20, 26, 32, 42, and 52 weeks of age in the long-term study group. Rates of change of tibial length and body mass were calculated as the difference between values as a percentage of the earlier value.

Radiography

For radiographs, the limbs were oriented in AP and lateral projections and taken with a Siemens Multix H/UPH configuration using digital luminescent cassettes with a 50-kV and 4-mA exposure. Images were imported with no magnification into a Picture Archiving and Communication System (PACS; Seimens) and viewed in MagicView 200 (Seimens) that digitized the image at a resolution of 2500 × 2500 pixels. Radiographs of the proximal half of the left tibias at 14, 26, and 52 weeks of age were examined for radiographic differences between treatment groups.

Histology and histomorphometry

Four left tibias in each treatment group were collected from animals killed at 10, 12, 14, 26, and 52 weeks of age, bisected longitudinally in the medio-lateral plane, and fixed in 4% paraformaldehyde (pH 7.4) for 48 h and decalcified in 14.5% EDTA/0.5% paraformaldehyde (pH 8.0) for 5 weeks at room temperature. Decalcification was confirmed by radiography, and the decalcified specimens were dehydrated and embedded in paraffin wax. Consecutive sagittal sections (5 μm) from paraffin blocks of the left proximal end of tibias were cut, oriented parallel to the longitudinal axis of the tibia, taking care to remain in the vertical axis of the physis, and were stained with H&E or TRACP for cellular histology or with Safranin O or fast green for histomorphometric analysis of cartilage and bone. Sections were analyzed using a digitizing tablet and BioQuant software (BIOQUANT; R&M Biometric, Nashville, TN, USA). Measurements were made in areas corresponding to the resorptive zone, directly adjacent (distal) to the chondro-osseous junction (ZONE 1), and the metaphysis, which was further subdivided into zones corresponding to the tissue present at the time of the second ZA dose (ZONE 2), between doses (ZONE 3), and the first ZA dose (ZONE 4; Fig. 3A). Bone volume over tissue volume (BV/TV), endochondral cartilaginous matrix volume (CgV/TV), trabecular number (Tb.N), and thickness (Tb.Th) were determined for the resorptive and metaphyseal zones. Osteoclast (TRACP+) cell number (Oc.N) and percent surface (Oc.S) were determined in the resorptive zone (ZONE 1) of the proximal physis.

Figure FIG. 3..

Bone and cartilaginous matrix histology of the left proximal tibia at 12 and 14 weeks of age. (A) Split macroscopic view of saline (left) and ZA*2 (right) at 12 weeks of age, which is 4 weeks after the first dose and 2 weeks after the second dose. Framed areas correspond to the following: zone of resorption, ZONE 1; tissue present at time of second dose, ZONE 2; tissue present at time between doses, ZONE 3; tissue present at time of first dose, ZONE 4. Higher magnification of each zone in (B, H, N, and T) saline, (C, I, O, and U) ZA*1, and (D, J, P, and V) ZA*2 at 12 weeks of age and in (E, K, Q, and W) saline, (F, L, R, and X) ZA*1, and (G, M, S, and Y) ZA*2 at 14 weeks of age. Trabeculae containing cartilaginous matrix cores retained after dosing were gradually removed with time (J and M, V and Y, U and X). Those trabeculae retained because of the first dose in ZA*1 (U and X) appeared more remodeled than those in ZA*2 animals (V and Y). Trabeculae and cartilaginous matrix existing in ZONE 4 were less remodeled in ZA*2 (S and P) than in ZA*1 animals (O and R). The longitudinal appearance of cartilaginous matrix cores in trabeculae retained after the first dose of ZA (U, V, X, and Y) was not present in tissue retained after the second dose in ZA*2 animals (J and M). The longitudinal form of cartilaginous septa in ZONE 2 of saline animals was lost in ZA*1 and ZA*2 animals at 12 weeks of age (B-D) but appeared to be regained in ZA*1, but not ZA*2, animals at 14 weeks of age (E-G). (A) Macroscopic image: H&E staining; magnification, x1.5; scale bar, 1 mm (left). (B-Y) Microscopic images: Safranin O (red) and Fast green (blue/green) staining; magnification, x10; scale bar, 100 μm (T).20

Analysis of physis

The physis contains chondrocytes that are organized into stacks of cells whose axes cluster about a vertical axis of anisotropy. (24) In a vertical stack, chondrocytes display a continuous structural variation that reflects different maturational stages of the cell differentiation cycle such that each cell in a column represents a later stage of development than the preceding. Not only do these axial cell columns represent a chronological sequence, the cells' maturation occurs synchronously for cells located at the same horizontal levels in different columns. Bone elongation is thought to occur from the proliferation of chondrocytes within these cell stacks and enlargement of the lacunae as the chondrocytes hypertrophy. (25) In this study, we analyzed the physis by measuring its total height from the top of the reserve zone to the last transverse septum of the hypertrophic zone at 100 points along the physis. Height of the reserve, proliferative, and hypertrophic zones were also measured. Mean length of chondrocyte cell stack and mean number of chondrocytes per cell stack were each recorded in the proliferative zone, and height and width were recorded in chondrocytes of the hypertrophic zone. Two sections per animal were used for histomorphometry in each treatment group at 12 and 14 weeks of age. Serial sections throughout the entire proximal tibia were observationally examined for changes in parallel alignment of the physis with the vertical axis of the tibia.

Immunohistochemistry

Decalcified sections were dewaxed in xylene and rehydrated in graded ethanol and rinsed in water. For PCNA staining, unmasking was performed with citrate buffer and microwaves for 5 minutes and endogenous peroxidase was blocked with 3% H2O2 in methanol for 10 minutes. Sections were incubated for 30 minutes in 40% horse serum followed by 1 h in a humidity chamber at room temperature with PCNA antibody (PC10 mouse anti-rat; DAKOCYTOMATION, Carpinteria, CA, USA) at a dilution of 1:1000, followed by rinsing and incubation with secondary horse anti-mouse antibody (DAKOCYTOMATION) and visualization with streptavidin-horseradish peroxidase complex and diaminobenzidine as supplied with the ABC Vectra staining kit (Vector Laboratories, Santa Cruz, CA, USA). Sections were counterstained in Harris hematoxylin. Proliferative activity was quantified by counting the number of PCNA+ cells per total chondrocytes in the proliferative zone to yield the percent PCNA+ cells. Only cells containing a nucleus were included in the analysis. Apoptotic cells were identified by TUNEL of fragmented DNA using the TACS TdT DAB Apoptosis detection kit (R&D Systems, Minneapolis, MN, USA). Sections were counterstained with methyl green. Two sections from each 12-week-old proximal tibia were analyzed. The number of apoptotic chondrocytes was counted and expressed per square millimeter of surface area of proliferative zone in the physis. Pretreatment with DNase in control sections labeled all cells.

Statistical methods

Statistical analyses were performed by one-way ANOVA within groups or two-way ANOVA between groups with linear contrasts. All data are presented as mean ± SD. The level of significance was set at p < 0.05.

RESULTS

Tibial length

There was no difference in mean tibial length between control and ZA2 groups in either study from study start through 12 weeks of age (Figs. 1A and 1B). At 14 weeks of age, however, 4 weeks after the second and final dose, tibial length was 4% less in ZA2 animals of the short-term study group (saline, 103 mm; ZA2, 98 mm; p < 0.05) and 3.4% less in ZA2 animals of the long-term study group (saline, 97 mm; ZA2, 94 mm; p < 0.05; Figs. 1A and 1B). In the long-term study, tibial length increased by 13.5% (p < 0.05) during the period of 10–26 weeks of age. The difference in length persisted at 20 and 26 weeks of age (saline, 99 mm; ZA2, 95 mm at 20 weeks; saline, 101 mm; ZA2, 98 mm at 26 weeks). No significant increase in length occurred after 26 weeks of age in either saline or ZA-treated animals, consistent with the normal cessation of rabbit growth at ∼25 weeks of age. (23)

Figure FIG. 1..

Tibial length (mm) in (A) short-term study and (B) long-term study and body mass (g) in (C) short-term study and (D) long-term study. Tibial length was less in the ZA*2 group than saline group at 14 weeks of age in both short- and long-term studies and at 20 and 26 weeks of age in the long-term study group. There was no difference in mean tibial length between groups at baseline age of 8 weeks in the short-term study group, and tibial growth rate was lower in ZA*2 than saline animals from 8 to 10 and 10 to 12 weeks of age. Body mass rate of gain was less in ZA*2 animals than saline animals from 8 to 10 weeks of age in both studies. Body mass was less in ZA*2 than saline at 9 and 12 weeks of age in the short-term study and at 26 weeks of age in the long-term study. Arrows indicate ZA administration at 8 and 10 weeks of age. In the short-term study group, n = 10 in saline group and n = 12 in ZA*2. In the long-term study group, n = 6 for each group except n = 2 per group at 32, 44, and 52 weeks of age. ≠, y-axes are broken to allow magnified view. Statistically significant difference from saline value at given age:ap < 0.05; difference in rate of change, bp < 0.05.20

Body mass

Mean body mass at 6 weeks of age was similar in all groups of the short-term study (saline, 1.37 ± 0.1; ZA2, 1.44 ± 0.1) and the long-term study (saline, 1.05 ± 0.04; ZA2, 1.11 ± 0.04; Figs. 1C and 1D). Rate of weight gain from 8 to 10 weeks of age, 2 weeks immediately after the first dosing, was 76% less in ZA2-treated animals compared with controls (p < 0.05) in the long-term study, from 1.55 g at 8 weeks to 1.81 g at 10 weeks in the saline group compared with a 1.64- to 1.71-g change during that interval in the ZA2 group (Fig. 1D). Total gain in body mass from 5 to 26 weeks of age was 1.8 kg in controls and 1.4 kg in ZA2-treated animals, resulting in a body mass 14% lower for ZA2-treated than control animals at maturity, 26 weeks of age (p < 0.05). There was no change in weight of animals from either group after 26 weeks of age. Similar results were seen in the short-term study (Fig. 1C).

Radiography

Radiographs showed radiodense lines in the metaphyses of ZA-treated animals that corresponded to the number of doses administered. These lines appeared diminished at 26 weeks of age compared with earlier times and had largely disappeared by 52 weeks of age (Fig. 2).

Figure FIG. 2..

Radiographic and macroscopic histological images of the left proximal tibias in the anterior-posterior plane in (A, D, F, H, K, and M) saline-, (B and I) ZA*1-, and (C, E, G, J, L, and N) ZA*2-treated rabbits at indicated ages (weeks). Radiodense lines on radiographs and lines of trabeculae in metaphyses of histological images corresponded to number of doses administered (arrows). Radiodense or histologically visible lines in the metaphyses had diminished by 52 weeks of age compared with those evident on images from earlier times. Original magnification, x1.5; scale bar, 10 mm (H).20

Histology

The radiodense lines seen in the metaphyseal region on radiographs of ZA-treated animals (Fig. 2) contained unresorbed trabeculae consisting of cartilaginous matrix cores in areas of tissue that were generated at the time of dosing (Fig. 3). In saline animals, longitudinal cartilaginous septa in the resorptive zone, directly below the physis (Fig. 3B, ZONE 1), were surrounded by osteoid, and through the process of endochondral ossification remodeled to form trabeculae that were subsequently removed from the metaphysis to form the medulla (Fig. 3, ZONES 2–4). However, in ZA-treated animals, these longitudinal cartilaginous septa were retained within trabeculae that were transiently preserved in the metaphysis and, albeit diminished with time, persisted in the cortical bone long term (Figs. 3A, 3U, 3V, 4A, 4U, 4W, and 4Y). The removal of such trabeculae containing remnants of cartilaginous matrix after dosing appeared more advanced in ZA1 than ZA2 animals (Figs. 3U, 3V, 3X, and 3Y). Furthermore, there was more tissue transiently retained during the time between doses in ZA2 than was retained in the corresponding region of animals not receiving a second dose (Figs. 3O, 3P, 3R, and 3S). The longitudinal appearance of cartilaginous matrix cores in trabeculae retained after the first dose of ZA was not present in the tissue retained after the second dose of ZA2 animals (Figs. 3J, 3V, 3Y and 3M). However, loss of this longitudinal form of cartilaginous septa appeared transient because ZA1, but not ZA2, animals showed a return of longitudinal cartilaginous septa in the resorptive zone at 14 weeks of age (Figs. 3C, 3D, 3F, and 3G). Bone in younger 10-, 12-, and 14-week-old animals of all groups stained blue-purple after safranin O and fast green staining (Figs. 3H-3Y); however, whereas cortical bone in older 52-week-old saline animals stained blue-green, bone of ZA2 animals at this age still stained with the blue-purple color of bone seen in younger animals (Figs. 4F, 4L, 4R, 4X, 4G, 4M, 4S, and 4Y).

Figure FIG. 4..

Bone and cartilage histology of the left proximal tibia at 26 and 52 weeks of age. (A) Split macroscopic view at 26 weeks of age with framed areas corresponding to the following: zone of resorption, ZONE 1; tissue present at time of second dose, ZONE 2; tissue present at time between doses, ZONE 3; and tissue present at time of first dose, ZONE 4. Each zone is shown at higher magnification in (B, H, N, T, D, J, P, and V) saline and (C, I, O, U, E, K, O, and W) ZA*2 samples in the medulla region at 26 and 52 weeks of age. Zones of cortical bone shown at 52 weeks of age in (F, L, R, and X) saline and in (G, M, S, and Y) ZA*2 samples. Cartilaginous matrix retained after ZA administration was removed from the medulla with time (I, U, K, and W; arrows) but persisted in cortical bone at 52 weeks of age (M and Y; arrows). (A) Macroscopic images: H&E staining; magnification, x1.5; scale bar, 1 mm (left). (B-Y) Microscopic images: Safranin O (red) and Fast green (blue/green) staining; magnification, x10; scale bar, 100 μm (T).20

Physeal morphology, proliferation, and apoptosis

The physes of saline animals had distinct zones of proliferating and hypertrophying chondrocytes that were organized into cell stacks orientated parallel to the vertical axis of the tibia. This allowed visualization within the plane of a 5-μm section of a complete cell stack beginning in the reserve zone and extending to the chondro-osseous junction (Fig. 5A). However, in 12-week-old ZA-treated animals, there were obvious morphological disturbances in cell stack organization and shape. Each individual chondrocyte cell stack could not be seen extending the full length of the physis within the plane of a 5-μm section, suggesting a loss of orientation parallel with the vertical axis, failure of cells to align beneath one another, or possibly shorter cell stacks because of reduced proliferation and/or apoptosis (Figs. 5C and 5D). These morphological disturbances were not as marked in ZA-treated animals at 14 weeks of age, suggesting a transient effect (Table 1). Total physeal height was decreased in ZA1 and ZA2 animals compared with saline animals at 12 weeks of age (12% and 24%, respectively; p < 0.05), and each reserve, proliferative, and hypertrophic zone within the physis was also shorter in ZA-treated animals, with the most marked height reduction occurring in the hypertrophic zone (ZA1, 36%; ZA2, 44% less than saline; p < 0.01). The proliferative zone in the physis of ZA-treated animals contained 61% fewer PCNA+ cells per total number of nucleus-containing chondrocytes compared with saline animals (p < 0.01; Fig. 5; Table 1). Apoptotic cells appeared almost exclusively among proliferative chondrocytes within the zone of proliferation of the physis with 35% fewer apoptotic chondrocytes in the ZA-treated group over controls at 12 weeks of age (p < 0.05). Within the hypertrophic zone of 12-week-old ZA1 and ZA2 animals, both height and width of chondrocyte lacunae were less than in saline animals (Figs. 5A-5D; Table 1). No apoptotic cells were detected among hypertrophic chondrocytes. There were virtually no differences in these parameters at 14 weeks of age. The physes of all groups appeared closed at 26 weeks of age (Figs. 4A-4E).

Table Table 1.. Physeal Morphology of Proximal Tibia
original image
Figure FIG. 5..

Physeal morphology at 12 weeks of age. Normal physis showing (a) reserve, (b) proliferative, and (c) hypertrophic zones in (A and C) saline-treated animals and (B and D; arrowheads) ZA*2 showing disruption of parallel alignment to the vertical axis in proliferating chondrocyte stacks, (B and D; arrows) lack of enlarged lacunae in the hypertrophic zone compared with those in saline (A and C; arrows), and (D, inset, white arrow) failure of cells to align beneath one another. PCNA+ cells in (E) saline and (F) ZA*2 physes. Tissue in ZONE 4 along the periosteal surface present at the time of the first dose in (G) saline and (H) ZA*2. (A-D, G, and H) Safranin O (red) and Fast green (blue/green) staining. (E and F) PCNA staining with hematoxylin counter staining. Magnification, x10 (A and B), x20 (C-F), x40 (C and D, insets), and x5 (G and H); scale bars 100 μm.20

Quantification of cartilaginous matrix and bone and osteoclasts in resorptive zone and metaphysis

Cartilaginous matrix volume (CgV/TV) within the resorptive zone (ZONE 1) of saline animals averaged 6% at 10, 12, and 14 weeks of age and was reduced to zero at 26 and 52 weeks of age. Treatment with ZA, however, was associated with a 1.5-fold increase (p < 0.05) in the amount of CgV/TV present in this zone compared with saline animals at 10 weeks of age, which persisted at 14 weeks of age (Figs. 3 and 6). CgV/TV in ZA2 animals was eventually reduced to 0.5% of total tissue volume in ZONE 1 at 26 weeks of age, with trace remnants of cartilaginous matrix persisting in cortical bone at 52 weeks of age (Figs. 4G and 6A).

Figure FIG. 6..

Histomorphometry of bone and cartilage of the proximal tibia. Endochondral cartilage matrix volume over tissue volume CgV/TV (%) (A-C) and bone volume over tissue volume, BV/TV (%) (D-F) in ZONES 1, 2, and 4 as described in Figs. 3A and 4A in saline (white bars), ZA*1 (striped bars), and ZA*2 (black bars) at indicated ages (weeks). n = 4 for each group except n = 2 per group at 52 weeks of age. Significant differences vs. saline:ap < 0.05.20

In the metaphysis, ZONES 4 and 2 (Fig. 3A) correspond to the tissue that was generated at 8 and 10 weeks of age, when the first and second doses of ZA were administered, respectively. In saline animals, CgV/TV was <3% in these zones of 12-week-old animals and was reduced to zero in 26-week-old animals (Figs. 3A and 4A). However in ZA-treated animals, more cartilaginous matrix was retained within the tissue generated at the times of dosing. In ZONE 4 of ZA1 animals, CgV/TV was 4-, 10-, and 50-fold greater than saline animals at 10, 12, and 14 weeks of age, respectively (p < 0.01; Fig. 6C). Results were similar for cartilaginous matrix retained in ZONE 4 of ZA2 animals at 12 weeks of age; however, removal of cartilaginous matrix from this zone in ZA2 animals was slower, presumably because of effects of the second dose of ZA. At 14 weeks of age, CgV/TV in ZONE 4 of ZA2 was 115-fold greater than saline and 2-fold greater than ZA1 (p < 0.05). As a result of the second dose administered to ZA2 animals, CgV/TV in ZONE 2 was 6- and 20-fold greater than in saline animals and 3- and 10-fold greater than ZA1 animals at 12 and 14 weeks of age, respectively (p < 0.05). Although eventually reduced to 4% of total tissue volume at 26 weeks of age, cartilaginous matrix persisted in metaphyseal bone at 52 weeks of age in ZA2, but not saline animals. ZA2 CgV/TV was therefore 50- and 250-fold greater than saline in ZONES 4 and 2, respectively (p < 0.01; Fig. 6B).

Bone volume (BV/TV) within the resorptive zone of the proximal tibia was comparable in saline and both ZA-treated groups at 10, 12, 14, 26, and 52 weeks of age (Fig. 6). In ZONE 4 of the metaphyseal region, BV/TV was greater by 2.4-fold in ZA1 and by 2.6-fold in ZA2 compared with saline animals at 12 weeks of age because of retention of trabeculae in the medulla of ZA-treated animals (p < 0.01). In ZONE 2, BV/TV in ZA2 was 1.5-fold greater than that of saline animals at 12 weeks of age (p < 0.01). No differences in BV/TV were seen between groups at later times (Figs. 3 and 6).

TRACP+ cell number but not percent surface was decreased in ZA2 animals compared with saline animals by 58% at 12 weeks and 34% at 14 weeks (p < 0.05). A similar trend was seen in ZA1 animals, but differences did not reach statistical significance (Table 1). No apoptotic osteoclasts were detected in any group at 12 weeks (data not shown).

DISCUSSION

The use of nitrogen-containing bisphosphonates in children with systemic bone diseases or local skeletal defects but otherwise normal skeletal growth raises concerns about their impact on the growth plate and subsequent bone growth. In this study, which analyzed the nonoperated tibias in juvenile rabbits from previous studies analyzing distracted callus, (14) administration of ZA at 0.1 mg/kg at 8 and/or 8 and 10 weeks of age caused a growth disturbance associated with the disruption of organizational cell morphology of the physis and delayed removal of cartilaginous matrix, leading to a 3% reduction in bone length. This loss of bone length occurred at a time when tibial length was already achieved 86.5% of its final value, and along with retained endochondral cartilage matrix within cortical bone, persisted long term.

Tibial length, as opposed to femoral length, is less variable in the rabbit model because of its geometry and ease of positioning for measurement. (23) In the rabbit, tibial growth begins to slow appreciably at sexual maturation between 10 and 12 weeks of age but continues until maturity at 25 weeks of age. (23) This study recorded tibial length from 10 to 52 weeks of age and, consistent with previous reports, a final 13.5% increment of tibial length was gained between 10 and 26 weeks of age. The physes were fused by 26 weeks of age and tibial length did not change thereafter. Transient growth disturbance after ZA administration at 8 and 10 weeks of age was associated with a 3% decrement in final tibial length. This decrement occurred in both short- and long-term study groups presented here and is consistent with previous short-term animal data. (11, 17–20, 26) Notably, in this study the growth disturbance associated with the reduced final bone length occurred when only 13.5% of less rapid tibial growth remained. The final effect on bone length after administration of a nitrogen-containing bisphosphonate may be different if dosing were to occur during earlier, more rapid growth periods. Importantly, in some other rodent studies non-nitrogen-containing bisphosphonates have had no effect on longitudinal growth. (16, 27) The effect of bisphosphonates on bone length is likely to be dependent on type, dose, frequency, age, and clinical background of the subject. In this study, the repeat dose of 0.1 mg/kg of ZA given 2 weeks apart in a juvenile rabbit distraction osteogenesis model would be equivalent to a repeat dosing of 5 mg given 4–6 weeks apart in a 50-kg adult human, which is at the upper end of the clinical range.

A small decrease in final long bone length may be clinically acceptable in certain circumstances. In the context of the treatment of leg length discrepancy, for example, the most common current treatment regimens include shortening the long side by up to 4 cm through epiphysiodesis. (28) A 3% reduction in height as noted in this study would translate to a 2.0- to 2.5-cm reduction of height in adolescents; however, the possibility of a height decrement would need careful consideration when planning treatment in children with localized skeletal defects. Although clinical studies in children with osteogenesis imperfecta, osteoporosis, and fibrous dysplasia show no detectable growth disturbance, (6, 29, 30) close monitoring of longitudinal growth should be performed in all children on bisphosphonates to provide further definitive clinical information on this topic. Anecdotal reports have suggested that increased mobility in these patients caused by the positive effects of bisphosphonates may act as a stimulus for growth through increased loading. These reports also suggest that, in children, the growth spurt at puberty may overcome a transient growth disruption or arrest in early childhood. (6) This study suggests that this may not occur if treatment were administered after puberty but before closure of the major physes.

Body weight gain, an important growth parameter, increases between 4 and 16 weeks of age in rabbits and ceases at 28 weeks of age. (23) In both the short- and long-term study groups reported here, the rate of weight gain in the 2 weeks immediately after dosing at 8 weeks of age was reduced in ZA-treated animals compared with controls, which led to ZA-treated animals being 14% lighter than controls at 9 weeks of age in the short-term study group and persisting at 26 weeks of age in the long-term study group. The reduced weight gain in ZA-treated animals in this study may be associated with a bisphosphonate-related disturbance in growth or to nonspecific effects on the well being of the animals. Initial doses of bisphosphonates are known to be strongly associated with a flu-like illness with fever, malaise, and even vomiting in children. (31) Preliminary data on ZA in children document this syndrome occurring in 85% of children, with 68% having a fever and 41% experiencing nausea or vomiting. (32) It is possible this common effect was subclinical in the rabbits but translated to decreased weight gain in these animals immediately after dosing.

Transverse sclerotic lines in the metaphysis seen on radiographs of tibias in this study have been seen previously in radiographs of long bones in both animals and humans receiving bisphosphonates. (5, 7, 21, 22) Some animal studies have shown that histologically these lines contain bone and cartilaginous matrix, (15–17) and recently this observation was reported in an osteogenesis imperfecta patient receiving cyclical intravenous pamidronate. (22) The detailed histological observations and histomorphometric analysis in this study provide new insights into the nature of these lines, with the 2-week spacing between doses clarifying the reduced removal of the longitudinal cartilaginous septa from the resorptive zone and of cartilaginous matrix and trabeculae from the metaphyseal zone in the tibia of animals receiving zoledronic acid. The resulting remnants of endochondral cartilaginous matrix within trabeculae remaining in the medulla appeared diminished with time; however, similar remnants persisted in the cortical bone at adulthood. These findings complement our previous observation of the long-term retention of cartilaginous matrix in the callus tissue of the operated tibias of these animals. (14)

In this study, the repeat dose of 0.1 mg/kg of ZA acid not only preserved more cartilaginous matrix than usual within the tissue being generated at the time, but also further delayed the removal of cartilaginous matrix remnants preserved at the initial dose and existing in tissue generated between doses. A longer period between doses may ameliorate this by allowing time for recovery from the negative effects of the first dose. In the clinical situation, determining a frequency of subsequent dosing that would avoid delaying the removal of cartilaginous matrix remnants that were preserved after an initial dose may prevent retention of such remnants in cortical bone at adulthood, which could otherwise jeopardize bone integrity.

In this study, ZA disrupted physeal morphology in a transient and dose-dependent manner. Although not described in detail, a disruption in growth plate organizational cell morphology was previously noted in animals receiving multiple weekly doses of alendronate. (20) The physis disruption in this double-dosed model appears to have distinct characteristics, because the chondrocyte cell stacks within the physis appeared to lose their orientation parallel to the vertical axis of the entire bone and individual chondrocytes failed to consistently stack beneath one another after mitotic division. These disruptions led to a change in the form of cartilaginous matrix left after removal of hypertrophic chondrocytes from the irregular clusters of lacunae. Normally, removal of cellular material from continuous, vertical hypertrophic cell stacks flanked by columns of cartilaginous matrix in the physis leaves longitudinal cartilaginous septa that are thinned by resorption before bone deposition on them. However, if the cell stacks within the physis are not continuous or longitudinally organized, once removed, the cartilaginous septa surrounding these irregular cellular clusters do not have a longitudinal form. This is reflected here in the form of cartilaginous matrix septa preserved immediately after the first dose of ZA, which had a longitudinal appearance, and after the second dose, which appeared spherical and adjoined. This occurs presumably because of the disruptive effects of the first dose on physeal morphology, such that removal of the noncontinuous and disorganized cell stacks left the surrounding irregular cores of cartilaginous matrix that were preserved immediately after the second dose. Importantly, regularity of chondrocyte cell stacks in the physis in animals given only one dose appeared to return with time, but the irregularity was more pronounced in animals receiving a second dose. Total physis height, as well as height of the reserve, proliferative, and hypertrophic zones within, were all transiently reduced in ZA double-treated animals. Furthermore, there was a dose-dependent reduction in the size of chondrocytes of the hypertrophic zone, and a 60% reduction in number of proliferating chondrocytes within the proliferative zone, both of which may have temporarily reduced bone elongation. (24, 25, 33, 34) The disorganization of the chondrocyte stacks may relate to the reduced chondrocyte proliferation and/or apoptosis in the proliferative zone.

The mechanism of endochondral bone elongation relies on clonal expansion and subsequent hypertrophy of chondrocytes, along with vascular invasion and resorption of the calcified cartilaginous septa at the chondro-osseous junction and subsequent influx of cells required for the formation and resorption of primary spongiosa. (35, 26) Furthermore, continued remodeling of the primary spongiosa is essential for metaphyseal integrity. (20) Preservation of cartilaginous matrix in the metaphysis after bisphosphonate therapy is presumed to occur through an inhibitory action of bisphosphonates on osteo(chondro)clastic resorbing cell function, (20, 37, 38) consistent with the reduction in TRACP+ cell number observed in this study. It is not known, however, how resorption of mineralized cartilaginous matrix in the resorptive zone is associated with bone elongation. It is possible that retention of this matrix may disrupt cellular signals required for normal physis function and subsequent bone elongation. However, elongation of tail vertebrae can continue despite clodronate ablated osteo(chondro)clastic resorption, (27) suggesting that bone elongation is independent of resorption of mineralized cartilaginous matrix; thus, the decrement in bone length seen in studies using nitrogen-containing bisphosphonates may be caused by a direct effect on the physis rather than through antiresorptive properties. Nitrogen-containing bisphosphonates have been reported to have anti-apoptotic effects on chondrocytes(39) and ZA-treated animals had a reduced number of apoptotic chondrocytes in the proliferative zone of the physis. ZA has also been shown to suppress MMP-9 expression and metalloprotease activity in osteolytic bone metastases(40) and to reduce matrix metalloproteinase (MMP)-9 expression in macrophages and subsequent association of vascular endothelial growth factor (VEGF) with its receptor on angiogenic endothelial cells preventing endothelial cell apoptosis. (41) These cellular and molecular targets of nitrogen-containing bisphosphonates are possible modes by which they could disrupt the function of the physis.

In conclusion, this study presents evidence from a growing animal model that nitrogen-containing bisphosphonates can cause transient effects on physis cell morphology and retention of cartilaginous matrix coinciding with a growth disturbance. This leads to a reduction in final long bone length at maturity. It emphasizes the need for an understanding of both the positive and negative effects of third-generation bisphosphonates in growing animal models to determine an effective dose that can be given at a frequency and age that minimizes detrimental effects to achieve maximal therapeutic potential in children requiring bisphosphonate therapy.

Acknowledgements

The authors thank Rachael Bugler, Madeleine Thompson, and Max Walford for technical assistance, Kathy Mikulec and Hao-Xu Lu for ethical care of the rabbits, Dr Christopher Cowell for encouragement and stimulating discussions, and Associate Professor Daniel Cass, Dr Amanda Sainsbury-Salis, and Dr Paul Baldock for comments on the manuscript. This study was funded in part by grants from the Children's Hospital Westmead Orthopaedic Research Fund and Novartis Pharmaceuticals Pty Ltd. AM was funded by contributions from Smith & Nephew Surgical and Ingham Enterprises. ES was funded by the NH&MRC Dora Lush BioMedical Scholarship and the J Scougall Orthopaedic Scholarship. EG was funded by a project grant from the Australian National Health & Medical Research Council.

Ancillary