Osteopetrosis is a genetic disease of defective bone resorption with an estimated incidence of around 1:20,000. The severity can range from mild to severe depending on the amount of compromise to osteoclast function. Clinically, the condition can be associated with pain, increased fracture rate, osteomyelitis, and dental complications. More severe forms can develop compression of nerves leading to blindness and deafness, and eventually death.1 The only successful curative treatment for osteoclast defect-related osteopetrosis is a bone marrow transplant from a healthy donor.2 However, the complications associated with this procedure means that it is only justified for life-threatening forms of the disease, with other patients being treated symptomatically.
The incisors absent (ia/ia) rat is one of the first identified naturally occurring rodent models of osteopetrosis. It was originally published by Dr. R. Greep in 1941 as a heritable mutation causing an absence of the prominent rat incisors.3 It was later noted the ia/ia rat possessed a fundamental bone problem, and in 1950 a detailed study noted slowed tibial growth and an “osteosclerotic” phenotype that undergoes spontaneous recovery.4 Later studies showed that transplantation of wild-type spleen or bone marrow cells were able to rescue the phenotype, indicating a cellular deficiency in healthy osteoclasts was the likely cause.5–7 Notably, it is these experiments in the ia/ia rat that directly led to the current clinical management of infantile malignant osteopetrosis. Due to the biological and material properties of osteopetrotic bones, as well as their shortened length and reduced marrow space, orthopedic management of osteopetrotic fractures can present unique challenges.8 We decided to examine fracture repair in the ia/ia rat as a model of osteopetrotic fracture healing. Two studies from the 1970s reported aberrant bone healing in ia/ia rat pups (fractured at 11–17 days). Compared to normal littermates, the ia/ia pups showed greater net callus mineralization and delayed hard callus remodeling.9 This phenotype could be rescued by transplantation of normal spleen cells.10 While these studies support a role for osteoclasts in hard callus remodeling, they did not address the early events associated with soft callus formation and endochondral replacement with hard callus. Consequently, we decided to perform a comprehensive study on more skeletally mature animals paying specific attention to the early stages of soft callus remodeling.
The ia/ia osteopetrotic phenotype has been reported to recover over time and it was important that osteoclast function remain compromised during the fracture repair process. To better define the time frame of recovery, ia/ia versus ia/+ and +/+ rats were examined longitudinally in terms of body weight, femora and tibia of rats were harvested for histomorphometric parameters, bone mineral density (BMD), and serum was harvested for bone resorption markers. Based on these data, rats at 5 weeks of age were subjected to the Einhorn closed fracture model and femoral healing was studied over 3 weeks. Previous work using pharmaceutical inhibition of osteoclast function with bisphosphonates has indicated that osteoclast function may not be critical for the early phases of endochondral union, but essential for timely remodeling of the hard callus.11, 12 Consequently, we hypothesized that formation and removal of the early soft callus would proceed normally in ia/ia mutant rats, but that hard callus removal would still be impaired. This study has important implications for the orthopedic management of osteopetrotic fractures and provides further evidence to support a redundant role for osteoclasts during initial endochondral fracture repair.
MATERIALS AND METHODS
ia/ia Phenotype Analysis
An ia/ia colony was established at the Westmead Department of Animal Care and experiments were performed with approval from the institutional animal ethics committee. The ia/ia phenotype was distinguished by the lack of incisors, and these animals were given normal rat chow ground to a fine grain. All rats were given food and water ad libitum. For phenotype analyses, time points at 3, 5, 7, 9, 12, and 20 weeks of age were examined, with groups of males and females analyzed separately. The earliest time point (3 weeks) was selected as the earliest feasible time to perform an orthopedic procedure on a rat. A minimum of n = 10 were used for each sex at each time point for both the ia/ia incisor absent phenotype and control phenotype (ia/+, +/+ genotypes). As shown previously, ia/+ and +/+ genotypes are indistinguishable from each other and due to the fact this rat colony could not be genotyped both these genotypes were pooled as control phenotype animals.14
At the point of cull, rats were weighed and the femora and tibiae dissected. BMD was measured by dual-energy X-ray absorptiometry (DEXA) using the GE Lunar PIXImus (Lunar Piximus Corp, Madison, WI). Samples were fixed in 4% paraformaldehyde and stored in 70% ethanol prior to processing for decalcified (paraffin) and undecalcified (resin) histology. Paraffin sections were stained for osteoclasts by tartrate-resistant acid phosphatase (TRAP) staining or Safranin O staining for cartilage to assess growth plate height. Resin sections of proximal tibia were stained with a von Kossa stain for calculations of BV/TV. Slide images were captured using a Leica DMLA CTRMC microscope (Leica Microsystems, Heerbrugg, Switzerland) and a QICAM Fast 1394 color 12 bit camera with QCapture software version 184.108.40.206 (Quantitative Imaging Corporation, Surrey, British Columbia, Canada). Images were analyzed for Osteoclast number (Oc.N), osteoclast surface (Oc.S) per unit bone surface and bone volume/total volume (BV/TV) using the BIOQUANT measure 32 Nova prime (Nashville, TX). Growth plate height was determined from an average of 15 measurements across the width of the proximal tibia. Serum samples were also taken at the experiment endpoint for measurement of Ctx peptide by ELISA as a marker of bone resorption and P1NP by ELISA as a marker of bone formation. As an in vitro assay for osteoclast formation and function, osteoclasts were generated from hematopoietic progenitors isolated from ia/ia and control rats. Whole marrow was cultured in alpha-MEM supplemented with 10% fetal calf serum (Invitrogen, Carlsbad, CA) and osteoclasts induced using RANKL (100 ng/ml) and M-CSF (20 ng/ml; R&D Systems, Minneapolis, MN). Bone resorption was measured on commercially obtained calcium phosphate-coated disks (Osteologic discs; BD Biosciences, Frederick, MD) as previously described.13
Fracture Healing in the ia/ia Rat
Fracture repair experiments were performed in 5-week-old rats based upon data from the phenotype study. Closed femoral fractures were induced in male rats using an Einhorn apparatus (drop weight and 3-point bending device with a weight drop of only 300 g reduced from 500 g in normal 9-week-old rats). In brief, rats were anesthetized by intraperitoneal injection of ketamine (75 mg/kg) and xylazine (9 mg/kg), which persisted for 15–25 min. A small incision was made at the knee and a 0.8 mm diamond tipped (required for denser ia/ia rat bones) Kirschner wire inserted into the medullary canal of the femur. A fracture was produced by 3-point bending and the position and lack of comminution confirmed by X-ray (Faxitron X-ray Corp., Wheeling, IL). Specimens were harvested at 1, 2, and 3 weeks post-fracture, which corresponded to 6–8 weeks of age. A total of n = 30 ia/ia rats (n = 10 per time point) and n = 30 control rats (n = 10 per time point) were operated on.
Radiographs of harvested samples were used to grade fractures for union. A fracture was considered united if both sides of the fracture callus was bridged by radio dense tissue. X-rays were graded by a blinded observer, although blinding was made difficult due to the obviously osteopetrotic fractures in ia/ia rat samples. Quantitative computed tomography (QCT) analysis was used to determine the mineral properties of the fracture callus using a Stratec xQCT research SA+ scanner (Norland Stratec, Durlacher, Germany). Scans were performed over a 5 mm region and total callus bone mineral content (BMC), BMD, mineralized tissue volume, and polar moment of inertia data were generated for each fracture callus. Images from all femoral QCT scans were obtained using Scion image for Windows (Scion Incorporated, North Ryde, NSW, Australia) imaging software from the center of the scanned region for observational analysis of callus remodeling and bone architecture. QCT scans were nondestructive and samples were subsequently processed for histological staining as described for phenotypic analyses. Cartilage, mesenchymal tissue, and bone in fracture sections were distinguished and visualized using an Alcian Blue, Wiegherts Hematoxylin, Picrosirius red stain. Cartilage stained blue, mesenchymal tissue and nuclei stained black and bone stained red. The area of each callus that contained avascular cartilage or mesenchymal tissue was measured as was the area of vascularized bone tissue and expressed as a percentage of the total callus area, as previously published by our group.12
For phenotype analyses, two-tailed unpaired t-tests were performed comparing the two genotypes within each of the sexes and age groups. All of the age groups within each sex graphed together. Error bars represent the standard error. For fracture analyses, differences in bridging were analyzed using a Fischer's exact test. Histomorphometric data were analyzed using two-tailed unpaired t-tests. An alpha value of 0.05 was used to determine significance.
The ia/ia Rat Exhibit a Persistent Osteopetrotic Phenotype
A statistically significant decrease in femur length was found in ia/ia rats from the earliest time point examined (3 weeks) in both males and females (Fig. 1A,B). While no difference in weight between ia/ia and control rats was seen at 3 weeks, the ia/ia rats were significantly smaller from 5 weeks onwards (Fig. 1C,D). The BV/TV within the distal femur in a region one field of view (2.5× magnification) below the growth plate was significantly higher in the ia/ia rats at all time points, consistent with its known osteopetrotic phenotype. The BV/TV values peaked between 7 and 9 weeks, indicating some functional recovery of osteoclasts occurred after this age (Fig. 1E,F). The overall femoral BMD as measured by DEXA continued to increase over time in both ia/ia and control rats, likely due to an increased amount of metaphyseal bone (Fig. 1G,H). Reduced height at the growth plate is another indicator of osteoclast dysfunction as osteoclasts are critical for normal growth. This was also found to recover in the ia/ia rats between 7 and 9 weeks (Fig. 1I,J). These data indicate a strong and persistent osteopetrotic phenotype in the ia/ia rats that nevertheless regains some a limited functional recovery between 7 and 9 weeks. Consistent with previous findings, Oc.N was elevated in ia/ia rats at all time points and were independent of recovery of the phenotype (Fig. 1K,L). Oc.S values showed high variation and were rarely statistically different between ia/ia and control rats (data not shown).
Further functional studies were performed to look at resorption in the ia/ia rats as well as the capacity of osteoclastic progenitors to form functional osteoclasts. Serum CTX was measured by ELISA (Fig. 2A), and this showed a marked decrease in young ia/ia rats. However, this normalized over time, consistent with the spontaneous recovery of the phenotype, again with the major difference being observed between 7 and 9 weeks. In addition, we assessed serum P1NP levels and showed no increases in bone formation at any stage, confirming the increase in bone mass was primarily due to the lack of resorption. In fact, at the 9-week time point we actually saw a 34% decrease in P1NP in the ia/ia rats compared to controls, suggesting reduced bone formation at this stage (p < 0.01, Fig. 2B). Next, an in vitro system was employed where primary osteoclasts are generated using RANKL/M-CSF induction from ia/ia and control rat bone marrow. While no difference was observed in Oc.N at all time points except 5 weeks (Fig. 2C), the function of osteoclasts generated from young rats was significantly inferior to that of control rats at 5, 9, and 12 weeks (Fig. 2D). Again, this was observed to recover over time, with some resorptive function in cells isolated from 9-week-old ia/ia rats, and the majority of function restored in cells from 12-week-old ia/ia rats and complete recovery by 20 weeks.
Early Phases of Fracture Repair Occur Normally in the ia/ia Rat
Closed femoral fractures were generated in 5-week-old ia/ia and control (ia/+, +/+) rats using an Einhorn apparatus as discussed in the Materials and Methods Section.
As a primary outcome measure, radiographs were graded as either bridged or not bridged based on the presence of new radio opaque bone tissue across the fracture gap. No difference in the rate of bridging was seen between the ia/ia and the control genotypes at any time point (Fig. 3A).
QCT Analysis of Fracture Callus and the Contralateral Limb
Analysis of hard callus volume using QCT demonstrated significant changes in the BMC, volume, BMD, periosteal circumference (circumference), and polar moment of inertia (moment of inertia) of the fracture callus in the mutant ia/ia rats compared to the ia/+, +/+ controls (Table 1). A 5 mm area of the bone was analyzed centered on the fracture callus.
As early as 1-week post-fracture, ia/ia rats showed increases over controls in terms of callus BMC (23%, p < 0.01), BMD (8%, p < 0.01), and mineralized callus volume (13%, p < 0.05). At this stage, there was no difference in periosteal circumference of the callus or moment of inertia. The differences in hard callus were more pronounced at 2 weeks with the ia/ia rats showing increased callus BMC (58%, p < 0.01) and mineralized callus volume (41%, p < 0.05). The moment of inertia was higher but did not reach statistical significance (63% increase, p < 0.06). Finally, by 3 weeks union was achieved in all samples. The ia/ia rats showed further increases in callus BMC (132%, p < 0.01) and mineralized callus volume (124%, p < 0.01). Callus circumference was also significantly increased (38%, p < 0.01), leading to a 251% increase in polar moment of inertia (p < 0.01). BMD was no longer different in ia/ia fracture calluses at 2 and 3 weeks.
Visual inspection of the cross-sectional QCT scanned images also revealed obvious differences in callus architecture between the ia/ia and control groups (Fig. 3B). This was most apparent at the 3-week time point, where the control fractures had commenced remodeling and exhibited a clearly defined neo-cortex. In contrast, ia/ia calluses were larger and lacked any signs of a neo-cortex, suggestive of a delay in hard callus remodeling.
Femora from the nonoperated leg were also analyzed to compare the systemic changes in bone in the ia/ia and control rats. Comparable regions to the fracture callus were analyzed over a 5 mm area (Table 1). The fracture time points of 1–3 weeks corresponded to rat ages of 6–8 weeks. While a difference in cortical BMC was initially noted at week 6 (20%; p < 0.01), this was reduced at week 7 (8%, p < 0.01), and no difference was seen by week 8. However, cortical BV remained higher in ia/ia bones throughout weeks 6–8 (p < 0.01). This meant that BMD of the contralateral leg steadily decreased in the ia/ia rats compared to wild types over the course of the study.
Histological Analysis of Cartilage Removal
Although delays in hard callus remodeling were apparent from QCT data, assessment of any associated delays in soft callus remodeling and cartilage removal could not be fully assessed radiologically. Specimens were therefore processed for histomorphometric analyses.
Quantitative analysis revealed a difference between the ia/ia and control rats in the total bone tissue area at 3 weeks post-fracture comparable to the mineralized BV differences observed by QCT. The ia/ia rats showed a 61% larger total callus area compared to controls (p < 0.05; Fig. 4A). The area of each callus that contained avascular cartilage and mesenchymal tissue was not different between genotypes throughout the experimental period (Fig. 4B). In both ia/ia rats and controls, the calluses contained 85–88% vascularized bone tissue (red) in week 1, 96–97% in week 2, and were completely (100%) ossified by week 3. This was also illustrated in histological images (Fig. 4C).
In this study we have performed a detailed analysis of the ia/ia rat phenotype and examined closed fracture healing in this animal model of osteopetrosis. The ia/ia rats exhibit high BMC and BV together with stunted growth with reduced long bone length. These changes were seen in both males and females out to final time point of 140 days. The increases in bone mass were associated with extensive reductions in bone resorption, as measured histologically and through serum Ctx. Importantly, serum P1NP was not increased in ia/ia rats compared to controls confirming no enhancement of bone formation in ia/ia rats, confirming the ia/ia rat increased bone mass phenotype was purely a result of reduced bone resorption. The retardation in bone growth is initially associated with an increased growth plate height in ia/ia animals, but the difference was found to disappear between 7 and 9 weeks of age.
One intriguing aspect of the ia/ia phenotype is its apparent reversion over time. A prior published histological study similarly found that some resorption begins to resume between 30 and 50 days, but that the osteopetrotic phenotype persists for >100 days.4 However, another report indicated that PTH-stimulated bone resorption was decreased by 22% in ia/ia rats at 16 days, but increased by 18% at 23 days compared to controls.14 Thus, the timeline for spontaneous recovery reported thus far remains ambiguous. Impaired bone resorption in ia/ia rats has been attributed to deficient ruffle border formation in osteoclasts.7, 15 Prior studies looking at in vivo osteoclasts have observed no improvement before day 50, which does not necessarily correspond to the previously noted functional improvements.7 We have isolated hematopoietic progenitors and induced osteoclast formation from ia/ia and control rats at different ages. While we did not specifically address ruffled border formation, functional in vitro assays of mineral resorption showed decreased osteoclast function that recovered with time. This was also consistent with a normalization of growth plate height. These data suggest that there is an intrinsic deficiency with osteoclast progenitors that recovers with time, rather than changes in the environmental cues that promote osteoclast formation. More recently, the genetic cause of the ia/ia phenotype was linked to a mutation in the Plekhm1 gene, and the human ortholog PLEKHM has been found to be associated with an intermediate form of osteopetrosis.16
In this study, we have used the ia/ia rat system to model osteopetrotic fracture repair. Patients with osteopetrosis not only develop more frequent fractures, but present with additional challenges for orthopedic fracture management.8 Long bone fractures generally heal by endochondral processes where a cartilaginous template is replaced with ossified bone, which is subsequently remodeled. Current models of fracture repair distinguish the early “nonspecific” anabolic and catabolic responses associated with normal wound repair and late “bone-specific” responses mediated by osteoblasts and osteoclasts.17 Although fracture repair has been examined in ia/ia rats previously, these authors failed to examine closely the initial nonspecific anabolic responses and cartilage tissue removal and focused mainly on hard callus remodeling.9, 18 Our data support the concept that osteoclast function is not essential for nonmineralized cartilage removal during the early stages of endochondral fracture healing. In contrast, the osteoclast is critical for subsequent hard callus remodeling with ia/ia rats showing extensive increases in hard callus BMC and volume, due to hindered resorption and not through enhanced bone formation. It should be noted that some mineralized cartilage struts trapped within woven bone in the primary spongiosa do remain, as these are hard tissue.
In our osteopetrotic fracture healing experiments, the key orthopedic outcome of fracture union was unaffected by phenotype/genotype. Thus, osteoclast dysfunction does not appear to adversely affect the achievement of or time to union. Polar moment of inertia increased in the ia/ia fractures, indicative of an increase in mechanical strength at week 3. These data suggest that the challenges that face the orthopedic care of osteoporotic fractures are primarily anatomical and focus around fixation rather than an intrinsic biological deficiency.
The overly dense bones of osteopetrotic patients can be challenging to stabilize in a clinical setting. We sought to address whether a biological deficiency in osteoclasts may translate to additional defects in fracture healing. We utilized the osteopetrotic ia/ia rat strain that, while maintaining a high bone mass phenotype, begins to show signs of recovery of osteoclast function between weeks 7 and 9. In a closed femoral fracture model in 5-week old ia/ia rats, there was no delay in soft callus remodeling leading to initial union, but a significant delay in hard callus remodeling was observed. From an orthopedic standpoint in osteopetrotic patients, this suggests once stable fixation is achieved, fracture healing should proceed normally, but delays in hard callus remodeling can be anticipated.
We would like to thank the Westmead animal facility staff for assistance with breeding and maintaining the ia/ia colony for the duration of the project.