Osteonecrosis of the jaw (ONJ) has been temporally associated with the use of nitrogen (N)-containing bisphosphonates (N-BPs)1–5 and, more recently, with the use of receptor activator of nuclear factor kappa B (RANK) ligand (L) antibodies.6–9 However, no cause-and-effect relationship has yet been established. N-BPs are powerful antiresorptive agents that bind to bone surfaces. They are ingested by osteoclasts, after which they decrease the activity and survival of these bone resorbing cells.10, 11 RANKL antibodies inhibit the formation, function, and survival of osteoclasts by specifically binding to RANKL, preventing its binding to the RANK receptor and hence, inhibiting RANKL/RANK signaling.12–14 Because of the earlier introduction and wider use of N-BPs compared to RANKL antibodies, more research has been conducted on N-BP–associated ONJ than RANKL-associated ONJ. N-BP–associated ONJ has been defined as an area of exposed bone in the maxillofacial region that does not heal within 8 weeks after identification by a healthcare provider, in a patient who is receiving or had been exposed to an N-BP and had not had radiation therapy to the craniofacial region.1, 15, 16 ONJ was reported to occur in 6% to 10% of N-BP–treated cancer patients.1, 4, 5, 17 It is observed most frequently in patients with multiple myeloma or breast cancer treated intravenously with high doses of N-BPs, particularly zoledronic acid (ZOL) or pamidronate,3, 18, 19 for inhibition of bone metastases. ONJ has also been reported to occur in postmenopausal women undergoing treatment for osteoporosis with the N-BPs alendronate (ALN), ibandronate, and risedronate.4, 5, 19, 20 Although ONJ incidence in N-BP–treated osteoporotic patients is much less than in cancer patients, more total ONJ cases may be identified in osteoporotic patients because many more women are treated with N-BPs for osteoporosis than for cancer.
Periodontitis has been identified as one of the most important risk factors for the development of ONJ.1, 4 The rice rat has been shown to be extraordinarily susceptible to the initiation and progression of a spontaneous form of periodontitis that does not require mechanical manipulation.21–25 Periodontal lesions develop more rapidly when rats are fed a diet high in sucrose and casein (H-SC).22–24, 26–28 Importantly, periodontitis in rice rats appears somewhat similar to that observed in humans in its development, location and appearance.22, 23, 27, 29
We have recently reexamined the rice rat model for periodontitis and characterized the periodontal lesions.25 We confirmed that rice rats, particularly when consuming a pelleted H-SC diet, represent a good small animal model for the accelerated horizontal and vertical alveolar bone loss (ABL) (moderate periodontitis) observed in humans with early or late onset periodontitis. Based on these findings, we considered that rice rats would represent a useful tool to study ONJ. We made use of this animal model of periodontitis and hypothesized that ONJ in this model is a two-stage process in which: (1) risk factors, such as periodontitis, initiate pathologic processes in the oral cavity that lead to a supranormal rate of hard tissue necrosis; and (2) a reduced level of bone resorption (eg, as produced by powerful antiresorptives) decreases the rate of removal of necrotic bone sufficiently to allow its net accumulation in the jaw.
We aimed to establish the effects of osteoporosis doses of ZOL and ALN, and oncologic doses of ZOL on: (1) the progression of periodontitis; (2) the integrity of alveolar bone; (3) the status of bone resorption and formation; (4) angiogenesis; and (5) osteocyte viability. This information allowed us to determine whether N-BPs can induce ONJ-like lesions in rice rats with periodontitis.
Subjects and Methods
Animals and experimental design
Study 1: Effects of N-BPs on the jaws of rice rats with inducible periodontitis after 6, 12, and 18 weeks of treatment
Seventeen pairs of marsh rice rats (Oryzomys palustris) were provided by Dr. Kent Edmonds, Department of Biology, Indiana University Southeast, New Albany, IN, USA. Rice rats were maintained in quarantine for 12 weeks before being transferred to a conventional room. Rice rats were paired using a monogamous continuous-breeding system until the complete number of experimental animals was generated. After a lactation period of 25 to 28 days, litters were weaned, separated by gender, and randomized to nine different experimental groups, composed of 15 to 19 (9–11 males and 5–9 females) rats in each group.
Experimental rats were fed a pelleted H-SC diet (TestDiet; Purina Feed, Richmond, IN, USA) for periods of 6, 12, and 18 weeks to induce periodontitis, as reported.23, 25, 27, 30 The composition of the H-SC diet was based on the formula of the powdered Harvard high-sucrose 700 diet and the ration 100 diet previously used.21–24, 26, 27 Simultaneously, groups of rice rats were injected either with vehicle, ALN, or a low (LD) or high dose (HD) of ZOL. The high dose of ZOL (HD-ZOL) is referred to as an “oncologic dose” throughout the text. Vehicle or ALN (15 µg/kg) were injected subcutaneously (SC) twice weekly for 6, 12, or 18 weeks. The ALN dose was based on a clinically relevant dosage for osteoporosis that was used previously.31, 32 Vehicle, LD-ZOL (8 µg/kg), or HD-ZOL (80 µg/kg) were intravenously (IV) injected in the tail vein once monthly for the same periods of time. The LD and HD of ZOL were recommended by Dr. J. Gasser (Novartis Pharma AG, Basel, Switzerland; personal communication) as the equivalent doses for osteoporosis and oncology treatments, respectively. In order to minimize the use of rats, vehicle control groups for ALN and ZOL were combined, in which half of the rats were injected SC with the vehicle used to dissolve ALN and the other half was injected IV with the vehicle used for ZOL dilution. ALN was obtained from Merck & Co. (Rahway, NJ, USA) and dissolved in a vehicle of sterile PBS. ZOL was obtained from Novartis Pharma AG and dissolved in sterile saline.
Study 2: Effects of HD of ZOL in rice rats after 24 weeks of treatment
The definition of N-BP-associated ONJ states that a non-healed area of exposed bone must persist for at least a period of 8 weeks after identification,15, 16or for 6 to 8 weeks according to Bilezikian.1 Based on results from Study 1, we conducted Study 2 to determine whether ONJ-like lesions could also be present in rats treated with HD-ZOL for 24 weeks. For this purpose, 25- to 28-day-old rice rats (10 males and 5 females/group) were fed the pelleted H-SC diet for periods of 24 weeks and simultaneously injected IV with vehicle or HD-ZOL monthly until the end of the study. The presence of ONJ-like lesions after 24 weeks of treatment would suggest persistence of lesions for more than 6 weeks.
All experimental rats were group-housed (2–5 animals per cage) in static filter top cages (area: 143 in2), with pine shavings bedding and continuous access to food and water. The housing room was maintained at 55°F to 70°F, with an average humidity between 30% and 70%, and under a 14-hour:10-hour light:dark cycle. The Animal Care Services resource at the University of Florida is an AAALAC-accredited animal care and use program. The animal protocol was approved by the University of Florida Institutional Animal Care and Use Committee (IACUC). All animal care and experimental procedures were in accordance with federal policies and guidelines of the University of Florida IACUC. Adequate measures were taken to minimize pain and discomfort in the rats.
Necropsy and tissue collection
Rats were injected SC with declomycin and calcein (Sigma, St Louis, MO, USA) at a dose of 15 mg/kg on the 7th and 2nd days prior to sacrifice, respectively, to label sites of bone formation. Rice rats were then euthanized by CO2 inhalation followed by thoracotomy. Maxillae and mandibles were stripped of musculature, leaving the periosteum intact. All right maxillae were processed for analyses of horizontal ABL. Left maxillae of males were placed in 10% buffered formalin for analyses of histology and vertical ABL. Left mandibles of males were placed in 70% ethanol to assess static and dynamic histomorphometric parameters and vertical ABL. Left mandibles of females were placed in 4% paraformaldehyde for histological, histochemical, immunohistochemical, and vertical ABL analyses. Right mandibles of females were placed in 70% ethanol for microCT analyses. Right mandibles of male rats were frozen at −80°C and processed to histochemically determine lactate dehydrogenase (LDH) activity in osteocytes as a marker of cell viability.
Determination of horizontal ABL
Right maxillae were processed as described.25 Digital images of both palatal and buccal root surfaces of all molar teeth were captured using a stereo dissecting microscope (SteReo Discovery V8; Carl Zeiss MicroImaging, Inc., Thornwood, NY, USA), and the surface perimeter of the cementoenamel junction (CEJ) and the alveolar bone crest (ABC) was calculated.25
Maxillae of male and female rice rats were processed as previously described.25, 33, 34 Fifty consecutive 10 µm sections were discarded to eventually obtain two consecutive sections more adjacent to the buccal surface (buccal surface sections). Vertical alveolar bone height (ABH) was determined by measuring the distance from a line between the CEJs of adjacent teeth to the ABC of the interproximal alveolar bone between the first and second molars (M1-M2), and second and third molars (M2-M3). These measurements were obtained in the palatal surface and buccal surface sections using the OsteoMeasure System (OsteoMetrics, Inc., Atlanta, GA, USA) at a magnification of ×100.25 Sections of mandibles from males that were stained en bloc with basic fuchsin and embedded in methyl methacrylate were used to assess vertical ABH at the lingual surface (see Histomorphometry below).
Region of interest
Histomorphometric evaluation for bone parameters, necrotic alveolar bone, and histochemical and immunohistochemical analyses were performed in the region of interest (ROI) located at the mandibular interproximal space between M1 and M2, the site for the highest prevalence and severity of periodontal lesions in rice rats.23, 27, 29 The ROI was similar to that reported,27, 30 except for minor modifications25 (See Supplemental Fig. 1 for details).
Evaluation of bone parameters
Left mandibles from male rats (n = 9–11/group) were processed by en bloc basic fuchsin staining and methyl methacrylate embedding to assess bone static and dynamic histomorphometric parameters within the ROI as described.25, 35–37 Sections obtained from each rat that were adequately aligned and contained alveolar bone processes from the lingual surface were used for the quantitative analysis. Variables assessed in the ROI included the percent of mineralized alveolar bone present within the ROI (alveolar bone volume) and fluorochrome-based indices of bone formation, including mineralizing surface (% cancellous bone perimeter bearing double fluorochrome labels plus one-half of singly labeled surfaces), mineral apposition rate, and bone formation rate. Bone formation rate (surface referent, BFR/BS) was calculated by multiplying mineralizing surface by mineral apposition rate.38 The terminology used, when applicable, was based on recommendations by the Histomorphometry Nomenclature Committee of the American Society of Bone and Mineral Research.39
Left mandibles from female rats (n = 5–9/group) were decalcified, embedded in paraffin, sectioned at 5 µm, and stained for tartrate-resistant acid phosphatase (TRAP; Sigma).40 Bone resorption parameters, including eroded surface, osteoclast surface, and osteoclast number, were measured on alveolar bone surfaces within the ROI using the OsteoMeasure System (OsteoMetrics, Inc.).
Assessment of necrotic alveolar bone
Necrotic alveolar bone within the ROI was defined using three complementary approaches: (1) quantification of bone areas containing nonviable osteocytes using sections histochemically stained for LDH enzyme reaction; (2) evaluation of osteocyte apoptosis by terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate-biotin nick end labeling (TUNEL) staining; and (3) measurement of bone areas devoid of basic fuchsin stain.
Right mandibles from male rice rats (n = 9–11/group) were used to calculate the amount of necrotic bone present in the alveolar bone of the ROI. Necrotic bone was histologically defined as those mineralized bone areas lacking viable osteocytes. Osteocyte viability was measured using in situ histochemical analysis of LDH activity as previously described.41, 42 Briefly, three randomized 7-µm undecalcified cryosections per animal were obtained with a Leica Cryostat (Leica, Inc.) equipped with a CryoJane Frozen Sectioning Kit (Instrumedics, Inc.).43 Cryosections were incubated for 3 hours at 37°C in 40% Polypep solution containing nicotinamide adenine dinucleotide (NAD) disodium salt, lactic acid, and nitroblue tetrazolium (NBT). Following incubation, sections were rinsed in warm water to remove the reaction mixture prior to fixation in 4% paraformaldehyde, and then mounted in Farrant's mounting medium. The total number of LDH-positive and -negative osteocyte lacunae was determined within the alveolar bone area of the ROI. Results were expressed as the percentage of LDH-negative osteocyte lacunae over the total number of osteocyte lacunae (with/without osteocytes present) within the ROI using the Osteomeasure System (Osteometrics, Inc.).
Paraffin-embedded, 5-µm longitudinal sections of mandibles from female rats (n = 5/group) were deparaffinized and rehydrated through a graded series of ethanol and water. Slides were then placed in 0.1 M citrate buffer (pH 6.0) and permeabilized by exposure to 1 minute of microwave irradiation (350 W). A known positive sample and two negative control slides were included in the runs. Staining was performed using a commercially available apoptosis kit (In Situ Cell Death Detection Kit; Roche, Applied Science) following the manufacturer's instructions. TUNEL reaction mixture was incubated on the slides for 1 hour at 37°C, with negative control slides receiving labeling mixture devoid of terminal deoxynucleotidyl transferase (TdT) enzyme. After three washes in 1× PBS, slides were coverslipped using VectaShield with DAPI (Vector Labs, Burlingame, CA, USA). Detection was observed as a fluorescent green signal contributed by fluorescein isothiocyanate (FITC).
Basic fuchsin stain using the en bloc staining method
The same bulk-stained mandible sections from male rats (n = 9–11) used for histomorphometry were used to determine necrotic bone. Necrotic bone was defined as those regions of the bone matrix devoid of basic fuchsin stain and was assessed in the ROI at a magnification of ×200. Necrotic bone was expressed as a percentage of the total bone area within the ROI using the Osteomeasure System (Osteometrics, Inc.). This method has been used to identify areas of necrotic bone in the mandibles of dogs treated with N-BPs.35–37
Immunohistochemical evaluation of angiogenesis/vascularity
Left mandibles from female rats (n = 5/group) were fixed in 4% paraformaldehyde and decalcified in 5% formic acid. Five micrometers (5 µm) decalcified paraffin-embedded sections were cut in the mesiodistal plane and processed for the morphometric evaluation of angiogenesis/vascularity by immunocytochemistry. Briefly, sections were deparaffinized and rehydrated through a graded series of ethanol and water. Slides were then stained with mouse anti-CD31 (DaKoCytomation, Carpinteria, CA, USA). Slides were blocked for endogenous peroxidase activity and then unmasked in Target Retrieval Solution (DaKoCytomation). Antibody was applied at 1:50 overnight at 4°C prior to identification with an anti-mouse detection kit using DAB (Elite Kit; Vector Labs). An isotype- and concentration-matched negative control section was included. The specificity of immunostaining was demonstrated by the absence of signal in sections incubated with isotype- and concentration-matched control mouse IgGs (Vector Labs) instead of the primary antibody. Slides were counterstained with hematoxylin and examined by light microscopy. Blood vessel number, expressed as total number per square millimeter (mm2) of the ROI area minus bone area, was quantified in a blinded manner at a magnification of ×200.
Five-micrometer (5 µm) decalcified paraffin-embedded sections of maxillae from males (n = 9–11/group) and mandibles from females (n = 5–9/group) rats were cut in a mesiodistal plane, stained with hematoxylin and eosin (H&E), and examined in a blind-coded fashion. As described for the vertical ABL assessment, sectioning was performed from the palatal toward the buccal surface. Blocks were oriented so as to visualize the mesial and distal roots of all molars as an indicator of proper alignment. Ten consecutive sections obtained from each rat that were adequately aligned and contained alveolar bone processes from the palatal or lingual surface were obtained to perform microscopic analyses. An inflammation scoring system was used, ranging from 0 to 5, where 0 corresponded to no histological lesions, 1 to slight, 2 to mild, 3 to moderate, 4 to severe periodontitis, and 5 to ONJ-like lesions (see Table 1 for details).
Table 1. Scoring System Used to Characterize Periodontal Lesions Observed in Maxillae and Mandibles of Rice Rats
GE = gingival epithelium; LP = lamina propria; PDL = periodontal ligament; ABC = alveolar bone crest; ONJ = osteonecrosis of the jaw.
Gingivitis: slight hyperplasia of GE, intraepithelial inflammatory cell infiltration. Accumulation of bacterial plaque. No substantial changes in the LP, PDL, or ABC.
Gingival hyperplasia, inflammatory cell infiltration of the GE and LP, accumulation of bacterial plaque. No substantial changes in the PDL or ABC.
Erosion/ulceration and hyperplasia of the GE and accumulation of bacterial plaque. Moderate inflammatory cell infiltration of the LP, disruption of the PDL, migration of the junctional epithelium, and ABC resorption.
Erosion/ulceration of GE and hyperplasia of the GE. Severe inflammatory cell infiltration of the LP, disruption of the PDL, migration of the junctional epithelium, and ABC resorption.
Extensive ulceration of GE. Exposed necrotic bone and/or osteolysis. Severe destruction of the periodontium accompanied with inflammatory cell infiltration of the LP and a pseudoepitheliomatous-like hyperplasia of GE. Bacterial colonies adhered to alveolar bone or deep in the LP.
MicroCT assessment of mandibular alveolar bone was performed in right mandibles of females (n = 5/group) as described.44 Vertical ABH was measured in the mandibular interproximal space M1-M2 in the customary fashion (distance from a line connecting the CEJs of adjacent teeth to the ABC) in up to 10 8-µm slices that subtended the buccal-to-lingual dimension. Furthermore, a three-dimensional reconstruction of the microCT scan of the whole rat mandible was performed to display horizontal alveolar bone on the buccal and lingual aspects of the first and second molars.45 In addition, bone mineral density (BMD) and bone volume (BV/TV) of alveolar bone were also assessed. All microCT measurements were performed in a blinded and randomized fashion.
For the pQCT analysis, left femurs from male (n = 9–11/group) and female (n = 5–9/group) rice rats were scanned using a Stratec XCT Research M instrument (Norland Medical Systems; Fort Atkinson, WI, USA) with software version 5.40. Scans were performed at a distance of 5 mm proximal to the distal end of the femur. This site is at the level of the secondary spongiosa of the distal femoral metaphysis. Volumetric content, bone density, and bone area were determined for total bone (trabecular and cortical bone) as described.46
Data are expressed as mean ± SD for each group. For Study 1, data were evaluated with ANOVA followed by the Holm-Sidak test for multiple comparisons. When ANOVA assumptions regarding normality of data were not met, the nonparametric Kruskal-Wallis test was used. For Study 2, differences in variables between rice rats treated with vehicle or HD-ZOL for 24 weeks were evaluated using the unpaired Student's t test. Regardless of the test employed, p values less than 0.05 were considered to be statistically significant.
Study 1: Effects of N-BPs on the jaws of rice rats with inducible periodontitis after 6, 12, and 18 weeks of treatment
Rice rats fed the H-SC diet and simultaneously treated with vehicle, ALN, or ZOL (LD or HD) consistently grew and increased body weight from weaning (4 weeks of age) to the end of the study (22 weeks of age) (Supplemental Fig. 2). Male rice rats became progressively larger in size and heavier than females after 7 weeks of age. However, there were no intragender differences in body weight (BW) between N-BP–treated groups compared to age-matched, vehicle-treated groups.
Gross examination of the jaws revealed no lesions of any type in rice rats treated for 6 weeks in any of the treatment groups. In contrast, serious gross lesions were observed more frequently in mandibles from rice rats treated with oncologic doses of ZOL for 18 weeks (36% in male and 40% in female rats) compared to age-matched controls (0% in male and 11% in female rats) (data not shown). These serious lesions appeared to resemble ONJ-like lesions described in humans. They were usually bilateral and located on the lingual surface of the mandible, particularly at the interproximal area between M1 and M2. ONJ-like lesions were grossly characterized by erosion or ulceration of the gingiva, exposed alveolar bone, molar furcation, and loss of alveolar bone support for the nearby teeth (Fig. 1). Consistent with the gross findings, histologic examination revealed minimal lesions in maxillae and mandibles of rice rats treated for 6 weeks in any of the groups. Except for 1 male rat treated with LD-ZOL and another with HD-ZOL for 18 weeks, which developed histologic inflammatory lesions with a score of 5, there were no significant differences in the degree of periodontal lesions in maxillae of rats treated with N-BPs compared to age-matched, vehicle-treated controls (Fig. 2C).
In contrast, a significant increase in the degree of periodontal lesions with a score of 5 was observed in mandibles of female rice rats treated with HD-ZOL for 18 weeks compared to vehicle controls (Fig. 2D). Periodontal lesions with a score of 5 were defined as ONJ-like lesions and were characterized by extensive ulceration of the gingival epithelium (GE), exposed necrotic bone and/or osteolysis, severe destruction of the periodontium accompanied by inflammatory cell infiltration of the lamina propria (LP), pseudoepitheliomatous-like hyperplasia of the GE, and the presence of bacterial colonies adhered to the surface, within the necrotic alveolar bone, or deep in the LP (Fig. 2).
As expected, we observed a potent antiresorptive effect of ALN and ZOL (LD and HD) at the distal femoral metaphyses in rice rats treated for 6, 12, and 18 weeks compared to age-matched, vehicle-treated rats (Supplemental Fig. 3). This effect was inferred by the significant increase in BMC and BMD detected by pQCT at this appendicular location. Remarkably, the effects of ALN and ZOL were different in the jaw bones than in the long bones. In maxillae, ALN and HD-ZOL prevented horizontal ABL only after 18 weeks of treatment (Fig. 3). In mandibles, HD-ZOL significantly increased BMD at 6 and 12 weeks but not at 18 weeks (Fig. 4A). Furthermore, HD-ZOL significantly increased bone volume and reduced vertical ABL at the lingual and buccal surfaces at 12 weeks, but not at 18 weeks (Fig. 4B–D). Consistent with this, the histometric analysis showed that HD-ZOL ameliorated maxillary and mandibular vertical ABL in male rice rats at 12 weeks but not at 18 weeks (data not shown). Importantly, severe mandibular ABL and bone osteolysis were apparent in mandibles of rice rats treated with HD-ZOL for 18 weeks by microCT analysis (Fig. 4E, F). Furthermore, increased porosity and a honeycomb-like appearance were frequently observed in the ONJ-like lesions in the alveolar bone of rats treated with HD-ZOL for 18 weeks (Fig. 4G, H).
The histomorphometric analysis performed in mandibles showed that both LD-ZOL and HD-ZOL dramatically reduced bone formation at 6, 12 and 18 weeks, whereas ALN only decreased it after 18 weeks of treatment (Fig. 5). Osteoclast number was reduced by ALN and ZOL at 6 weeks, but not at later time points (Fig. 5D). In addition, no significant differences were observed in blood vessel number between N-BP–treated rice rats compared to vehicle-treated control rats. However, a trend toward a reduction in blood vessel number was observed between HD-ZOL and vehicle-treated groups at 12 and 18 weeks (Fig. 5E).
Histologic examination revealed that ONJ-like lesions contained areas of exposed alveolar bone with empty osteocyte lacunae and/or osteolytic lesions. We observed a significant decrease in LDH reactivity in osteocyte lacunae in the mandibular alveolar bone of male rice rats treated with HD-ZOL for 18 weeks (Fig. 6A, D, E). Furthermore, we observed an increased percentage of osteocyte apoptosis in the mandibular alveolar bone of female rice rats treated with HD-ZOL for 18 weeks (Fig. 6B, H–K). Consistent with the decreased LDH negative osteocyte lacunae observed in males, we found extensive areas of DAPI-negative osteocyte lacunae in the alveolar bone of female rice rats treated under the same regimen (Fig. 6J). In addition, the percentage of areas devoid of basic fuchsin was significantly higher in mandibular alveolar bone of rats treated with HD-ZOL for 18 weeks compared to age-matched, vehicle-treated controls (Fig. 6F, G).
Study 2: Effects of HD of ZOL in rice rats after 24 weeks of treatment
HD-ZOL administered for 24 weeks induced mandibular gross lesions that resemble ONJ lesions (Fig. 7A, B) (33% in male and 67% in female rats) compared to age-matched controls (0% in male and 0% in female rats) (data not shown). These lesions were grossly characterized by extensive ulceration of the gingiva, exposed alveolar bone, molar furcation, and tooth loss. We found that several vehicle-treated rats developed moderate periodontitis after 24 weeks (Fig. 7C). Importantly, only rice rats treated with HD-ZOL developed ONJ-like lesions (Fig. 7G). These lesions were similar to or with an even more destructive pattern compared to those observed after 18 weeks of treatment. Furthermore, bacterial biofilms morphologically resembling Actinomyces sp. were found attached to or within the necrotic alveolar bone, and also on the surface of the ulcerated gingival epithelium (Fig. 7C–F). As observed in rats treated with HD-ZOL for 18 weeks, this potent N-BP reduced mineralizing surface and BFR/BS, but not osteoclast number (Fig. 7H–K). Finally, blood vessel number was significantly decreased in rice rats treated with HD-ZOL compared to vehicle controls (Fig. 7L).
Thirty years ago, Gotcher and Jee30 used the rice rat model of periodontitis to determine whether dichloromethylene diphosphonate (Cl2MDP), a non–nitrogen-containing BP, could retard or prevent the bone loss associated with periodontal disease. They observed that Cl2MDP indeed prevented alveolar bone loss. However, and perhaps relevant to the pathophysiology of ONJ, they found an abnormal amount, location, and morphologic pattern of the alveolar bone in mandibles of rats that were treated for 18 weeks. Gotcher and Jee30 referred to these lesions as “areas of alveolar bone protruded and exposed in the oral cavity with portions that were devoid of bone cells and devitalized.” This description attracted our attention since it has similarities to the ONJ lesions described in N-BP–treated patients. Nevertheless, the study30 did not provide conclusive evidence to support the existence of ONJ, leaving open the possibility to infer that a different bone disorder occurred. Inspired by these observations, we reexamined the periodontitis model in rice rats, previously described,22–24, 27, 29 with methodologies now commonly used in periodontal research, including morphometry, histomorphometry, immunocytochemistry, histochemistry, and microCT techniques.25, 47 We next performed the present study to determine the effects of two N-BPs now approved for the treatment of osteoporosis and, in the latter case, for metastatic cancer, ALN and ZOL, on the progression of periodontitis, the integrity of alveolar bone, bone resorption and formation, vascularity, and osteocyte viability.
To our knowledge, very few studies30, 48 have considered periodontal disease as a comorbid factor in the development of an animal model for ONJ. Taking this risk factor into account improves the potential relevance of an animal model of ONJ, since clinically and radiographically evident periodontitis has been identified as one of the most important risk factors for the development of ONJ in humans.1, 4, 49
Our study shows that about 40% of the rice rats (males and females) treated with oncologic doses of ZOL for 18 weeks developed gross mandibular lesions that appear to resemble those observed in ONJ patients. In contrast, only 20% of these animals developed ONJ-like lesions in maxillae. This observation represents an interesting epizootiologic aspect of the model, since it is well established that the incidence of ONJ in humans is two-fold greater in the mandible than in the maxilla.1, 5, 17 Similar incidence of ONJ-like lesions, including osteonecrosis (47%) and exposed bone (21%), were found and elegantly illustrated by Aghaloo and colleagues48, 49 in maxillae of ZOL-treated rats with ligature-induced periodontitis. Our study also shows that none of the rats treated with ALN developed ONJ-like lesions. Our findings are in agreement with epidemiological data in humans showing a high incidence (6%–10%) of ONJ lesions in N-BP-treated cancer patients3, 18, 19 and a very low incidence (∼1/10,000) in postmenopausal women undergoing treatment for osteoporosis.4, 5, 19, 20 This study did not investigate the effects of N-BPs on rice rats that were fed a standard diet instead of the H-SC diet. Data obtained from such an experiment, in which rice rats would have little or no periodontitis, could contribute to determining whether predisposing factors, in this particular case, moderate/severe periodontitis, are sine qua non conditions for the development of ONJ. This study also did not investigate the effect of introducing N-BPs into rice rats with periodontitis that had been established by 18 weeks' administration of the H-SC diet. Data obtained from such an experiment would more closely mimic the situation in adult humans with established periodontitis who begin to take N-BPs as oncologic or osteoporosis therapy.
Remarkably, we observed that rice rats that had mandibular histologic lesions with an inflammatory score of 5 (ONJ-like lesions), were those that displayed evident serious gross lesions including erosion or ulceration of the gingiva, exposed alveolar bone, molar furcation, and loss of alveolar bone supporting the teeth. This correspondence suggests that in our model, gross examination of mandibles can be used similar to a clinical examination, to study the progression of ONJ or to test different treatment modalities for ONJ in vivo before euthanizing the animals for ex vivo tissue analysis.
We further observed that ONJ-like lesions were not only present in rats that received HD-ZOL for 18 weeks but also in those treated for 24 weeks. This is important, because it suggests that ONJ-like lesions in this model were not transient, since they likely persisted for at least 6 weeks after their identification at 18 weeks of treatment. Although this time period does not fully comply with the most recent definitions of N-BP-associated ONJ,15, 16 it could represent a comparable persistence for a rodent animal model.
The definition of ONJ only specifies “an area of exposed bone in the maxillofacial region.” However, numerous clinical reports in humans have described additional bone lesions present in ONJ patients, including osteolysis, osteosclerosis, presence of woven bone, honeycomb-like appearance of affected areas of jaw bones, and pathologic fractures.4, 50–57 Importantly, we observed some of these lesions in our model. Rice rats treated with oncologic doses of ZOL for 18 to 24 weeks displayed, concomitantly with or without exposed bone, extensive areas of osteolysis, and alveolar bone areas with a honeycomb-like appearance as evidenced by microCT. The histologic and histomorphometric analyses of these honeycomb-like areas showed a correspondence to alveolar bone areas that were replaced by woven bone, periodontal ligament, and/or fibrous tissue, and were sometimes mineralized. Taken together, these data suggest that our model can reproduce not only the presence of exposed necrotic bone in the oral cavity, but also other pathological bone lesions frequently observed in patients with ONJ.
The clinical staging system updated on the 2009 AAOMS expanded abbreviation guidelines16 has served to more accurately categorize patients with ONJ. We attempted to use this system for ONJ-like lesions found in rice rats. We observed that HD-ZOL administered for 18 weeks induced the development of ONJ-like lesions that could be comparable to a stage 1 or 2, since the rats did not lose weight (as an indicator of well-being and lack of pain or distress), but had periodontal lesions frequently accompanied by infection. On the other hand, ONJ-like lesions present in rice rats treated with HD-ZOL for 24 weeks could be comparable to those in stage 2 or 3, because exposed necrotic bone was frequently accompanied by infection, osteolytic lesions, and the loss of a significant amount of body weight (data not shown). The development of a staging system for the rice rat model would be highly desirable in order to advance the development of early diagnostic protocols and treatments, particularly in the early stages of the disease (phase 0), when no clinical evidence of necrotic bone but nonspecific clinical findings and symptoms appear to be present in humans.
In contrast to the consistent and potent antiresorptive effect of ALN and ZOL observed at the distal femoral metaphyses, different responses were seen in jaw bones. Responses were associated with the degree of histological lesions found at these locations. When periodontal lesions were absent or present to a low degree, as found in maxillae at 18 weeks and in mandibles at 12 weeks, HD-ZOL exerted a potent antiresorptive effect. In contrast, the antiresorptive effect of HD-ZOL was lost when a high degree of damage, destruction, and commensal bacterial (Actinomyces-like) growth were present. It is well established that oral bacterial infection, particularly with Actinomyces sp., is a common finding in patients with ONJ.52, 58–62 When this takes place, microorganisms can stimulate bone resorption by inducing cytokines and/or RANKL.63–65 Our findings are in agreement with what Reid and Cornish52 defined as the “ONJ paradox,” in which aggressive local osteolysis could be present in patients treated with antiresorptives as a result of infection-induced bone resorption.
We also observed that rice rats treated with oncologic doses of ZOL for 18 weeks have areas of exposed alveolar bone that contain empty osteocyte lacunae and/or osteolytic lesions. In line with these findings, we also observed decreased reactivity to LDH in osteocyte lacunae in males and an increased percentage of osteocyte apoptosis in female rice rats at this time point. Taken together, these data suggest that indeed, HD-ZOL is associated with increased levels of apoptosis/necrosis in alveolar bone osteocytes of rats with induced periodontitis. There are some differences in the assessments of these tests. The assessment of LDH-negative osteocyte lacunae considers not only the presence of unreactive osteocytes (necrotic), but also empty osteocyte lacunae. Therefore, the presence of cellular remnants and absence of osteocytes in lacunae are also expressions of necrotic cells. In contrast, the percentage of TUNEL-positive osteocytes, accounted only for those cells that have fragmented DNA, a characteristic of both apoptotic as well as necrotic cells, from the total of DAPI-positive osteocytes in the ROI. Therefore, empty osteocyte lacunae were excluded from this calculation. The presence of empty osteocyte lacunae is a characteristic of the alveolar bone described in ONJ patients.19, 59, 62 Taken together, these findings suggest that the LDH assay would be a better indicator of the index of osteocyte necrosis in the alveolar bone compared to TUNEL staining, which underestimates the real number of necrotic/apoptotic osteocytes. Whether apoptosis/necrosis of osteocytes results from side effects of the inflammatory response, bacterial endotoxins, and/or end-products, a direct toxic effect of HD-ZOL, or other mechanisms, remains to be defined.
Finally, we observed a tendency toward reduced blood vessel number in the periodontium of rice rats treated with HD-ZOL, but only after 24 weeks of treatment. We have previously described an antiangiogenic effect of another N-BP, ALN, in a tooth extraction rat model.32 However, in the latter study, the effect appeared early and was transient.
ZOL has been reported to inhibit oral keratinocytes proliferation,66 induce oral epithelial cell apoptosis,67 and exert a potent antiangiogenic effect.68–70 It is well established that the integrity of the gingival epithelium is a physical barrier for the oral microbiota to invade, colonize, and induce inflammation and disease. Taken together, our data suggest that high doses of ZOL could enhance or promote the persistence of periodontitis and tissue destruction by inhibiting re-epithelization of damaged gingival epithelium, allowing commensal bacterial growth and periodontal infection. Bacterial infection can then stimulate production of proinflammatory cytokines and increase the expression of RANKL by different cell types in the periodontal tissue, resulting in more tissue destruction, decreased angiogenesis, and the paradoxical local elimination of the antiresorptive effect of this potent N-BP.
Most ONJ cases reported in the literature of the past 5 years have been associated with the use of oncologic doses of N-BPs.1–5 The more recent observation of ONJ in N-BP–naïve patients receiving RANKL antibodies6–9 suggests that it may be the physiologic effect of strongly inhibiting bone resorption by any means, rather than through the specific use of an N-BP, that drives much of ONJ's pathophysiology. In this report, we show that rice rats with moderate/severe periodontitis develop ONJ-like lesions when coadministered oncologic doses of ZOL. A similar experiment in which RANKL antibodies or osteoprotegerin were administered to rice rats with moderate/severe periodontitis would test this hypothesis.
In summary, our data show that rice rats treated with oncologic doses of ZOL appear to be a promising animal model for studies of the pathogenesis and treatment of ONJ. Our study also showed that osteoporosis doses of ALN and ZOL produced no such lesions in a similar timeframe. Our study suggests that HD-ZOL exacerbates the inflammatory response and periodontal tissue damage in rice rats, inducing lesions that resemble ONJ. Our data show that HD-ZOL inhibits bone resorption in the jaw of rice rats when the integrity of the periodontium is preserved. However, the antiresorptive effect is lost simultaneously with extensive ulceration of gingival epithelium, exposed necrotic bone and/or osteolysis, severe inflammatory cell infiltration of the lamina propria, and presence of bacterial biofilm adhered to alveolar bone or deep in the lamina propria.
DBK is a consultant to Xradia, Amgen, Lexicon, and Bayer; and owns Merck stock and options. All other authors state that they have no conflicts of interest.
This research was supported by NIH grant R03DE018924-01A1 from the National Institute of Dental and Craniofacial Research (NIDCR). We thank M Rivera, X Xia, Katherine Neuville, and A Lepper for technical support. The authors are grateful to Novartis Pharma (Basel, Switzerland) for kindly providing the zoledronic acid used in these studies. We thank Dr. Kent Edmonds, Department of Biology, Indiana University Southeast, New Albany, IN, USA, for providing the original rice rat breeders.
Author's roles: Study design: JIA, DBK, and TJW. Study conduct: JIA, JEP, AW, and MJ. Data collection: JIA, JEP, and MPA. Data analysis: JIA and MPA. Data interpretation: JIA, DBK, TJW, MPA, MJ, and LK. Drafting manuscript: JIA. Revising manuscript content: JIA, DBK, TJW, MPA, MJ, and LK. Approving final version of manuscript: JIA, MPA, DBK, JEP, AW, MJ, LK, and TJW. JIA takes responsibility for the integrity of the data analysis.