Osteonecrosis of the Jaws (ONJ), a significant side effect of anti-resorptive or antiangiogenic medications, is described as exposed bone in the oral cavity that often presents with pain, infection, and often significantly decreases quality of life of affected patients. Largely seen in patients taking anti-resorptives for osteoporosis, bone malignancies, and bone metastases, the pathophysiology of ONJ has remained largely elusive. Additionally, treatment options for the disease have remained empirical, often focused on managing symptomology. Others have recommended surgical resection of the affected region, often leaving large defects that can be devastating. Here, we develop an ONJ animal model in rats, harness its power to investigate the effects of anti-resorptive discontinuation. Additionally, we explore whether other medications with possible anti-resorptive properties can lead to development of ONJ. Finally, we evaluate clinical data to better understand ONJ pathophysiology, as well as ONJ treatment.

Following clinical observations, where tooth extraction occurs only in diseased teeth, such as those with periapical or periodontal disease, we hypothesized that significant periapical disease prior to tooth extraction will lead to ONJ. Thus, we treated rats with high-dose bisphosphonates, and extracted teeth with experimental periapical disease (EPD). All BP-naïve animals healed uneventfully. Interestingly, only animals treated with BPs with extraction of teeth with EPD developed clinical, radiographic, and histologic signs of ONJ. Our findings here show that pre-existing periapical disease is necessary to induce ONJ in rats treated with BPs.

Using the animal model described above, we investigated with discontinuation of BPs or the RANK-L inhibitor, OPG-Fc (a denosumab surrogate for rodents), is time or drug dependent. First, we observed that discontinuation of BPs prior to tooth extraction did not ameliorate ONJ burden, while discontinuation of OPG-Fc prior to tooth extraction ameliorated ONJ burden. In contrast, discontinuation of BPs or OPG-Fc in rats with established ONJ did not lead to resolution. These findings have direct clinical correlation and can be used for the development of clinical guidelines.

With our understanding of ONJ and its relationship to systemic anti-resorptive use, we investigated whether treatment with romosozumab, a monoclonal antibody against sclerostin used for the treatment of osteoporosis, would lead to ONJ. In clinical studies, romosozumab led to decreased bone resorption markers;​ thus, while not a purely catabolic agent, it may have the potential to cause ONJ. Using an established model of experimental periodontitis (EP) in ovariectomized rats, we found that treatment with romosozumab did not lead to ONJ.

Driven by these translational findings, we evaluated a cohort of patients with osteopetrosis and pycnodysostosis, inherited genetic diseases that lead to an anti-resorptive like phenotype. In these patients, we observed areas of exposed, non-healing bone that were phenotypically identically to patients with ONJ from anti-resorptive use. Radiographic and histologic findings confirmed findings;​ in-vitro resorption pit assays revealed an absence of osteoclastic resorptive activity in these patients.

Finally, we retrospectively evaluated a cohort of ONJ patients that had been treated with local wound care. In these patients, we found that over 75% of patients healed with bone sequestration, leaving healed mucosa on the underlying surface. We found that healing time correlated with their wound care abilities, with patients with better wound care exfoliating the necrotic bone prior to those with poor wound care.