Mechanism of bisphosphonate-related osteonecrosis of the jaw (BRONJ) revealed by targeted removal of legacy bisphosphonate from jawbone using competing inerthydroxymethylene diphosphonate

  1. Hiroko Okawa
  2. Takeru Kondo
  3. Akishige Hokugo  Is a corresponding author
  4. Philip Cherian
  5. Jesus J Campagna
  6. Nicholas A Lentini
  7. Eric C Sung
  8. Samantha Chiang
  9. Yi-Ling Lin
  10. Frank H Ebetino
  11. Varghese John
  12. Shuting Sun  Is a corresponding author
  13. Charles E McKenna  Is a corresponding author
  14. Ichiro Nishimura  Is a corresponding author
  1. Weintraub Center for Reconstructive Biotechnology, Division of Regenerative & Reconstructive Sciences, University of California, Los Angeles School of Dentistry, United States
  2. Division of Molecular & Regenerative Prosthodontics, Tohoku University Graduate School of Dentistry, Japan
  3. Regenerative Bioengineering and Repair Laboratory, Division of Plastic and Reconstructive Surgery, Department of Surgery, David Geffen School of Medicine at University of California, Los Angeles, United States
  4. BioVinc, LLC, United States
  5. Department of Neurology, David Geffen School of Medicine at University of California, Los Angeles, United States
  6. Department of Chemistry, University of Southern California, United States
  7. Division of Oral & Systemic Health Sciences, University of California, Los Angeles School of Dentistry, United States
  8. Section of Oral & Maxillofacial Pathology, University of California, Los Angeles School of Dentistry, United States
7 figures, 1 table and 1 additional file

Figures

Competitive equilibrium-based dissociation of N-BP by low potency BP (lpBP).

(A) Chemical structures (shown as the tetraacids) of N-BP, zoledronate (ZOL) and hydroxymethylene diphosphonate (HMDP). (B) ZOL but not HMDP affects mouse femur trabecular bone architecture. Mice received a bolus intravenous injection of 100 µl ZOL (40 nmol), HMDP (40 nmol) or vehicle saline (0.9% NaCl) solutions. Femurs (n=4 per group) were harvested 3 weeks after the IV injection and subjected to Micro-CT imaging. (C) HMDP did not affect the femur trabecular bone micro-architecture, whereas ZOL increased bone volume over total volume (BV/TV), connectivity density (ConnD) and trabecular thickness (Tb.Th). Trabecular number (Tb.N) and trabecular separation (Tb.Sp) measurements were not affected by ZOL. (Figure 1—source data 1) (D) In vitro demonstration of competitive displacement of legacy N-BP. Synthetic apatite-coated wells preincubated with 10 µM of 5-carboxyfluorescein-conjugated zoledronate (FAM-ZOL) were washed with MilliQ-treated pure water (MQW), then treated with 10 µM HMDP once (1×) or twice (2×) (n=3 per group). The FAM fluorescent signal measurement indicated significant reduction of the FAM-ZOL amount on the synthetic apatite. (Figure 1—source data 2) (E) Fluorescent signal measurement of the wash solutions from the in vitro experiment in (D) demonstrated removal of FAM-ZOL by the HMDP treatments (arrows). (Figure 1—source data 3) (F) In vitro osteoclastic pit formation assay. Synthetic apatite coated wells were preincubated with 10 µM ZOL followed by 10 µM HMDP treatment 1× or 2× (n=3 per group). RAW264.7 cells (2.5×104 cells per well) were then inoculated to each well in culture medium supplemented by mouse recombinant receptor activator of nuclear kappa-B ligand (RANKL). The areas of resorption pits generated by osteoclasts derived from RAW 264.7 cells were measured after 6 days of incubation and cell removal. Twice repeated HMDP treatments restored normal in vitro resorption pit formation. (Figure 1—source data 4) In (D) and (F), the graphs show the mean and SD (n=3 per group), and the Turkey test was used to analyze multiple samples. The statistical significance was determined to be at p<0.05. In (E), the graphs show the mean and SD (n=3 per group), and the Turkey test was used to analyze multiple samples within each time point. The statistical significance was determined to be at p<0.05. Different letters (e.g., a, b) are used to show statistically significant differences between multiple groups.

Manufacturing and trans-oral mucosal penetration evaluation of hydroxymethylene diphosphonate (HMDP)-DNV and AF647-ZOL-DNV.

(A) Diagram of deformable nanoscale vesicles (DNV), a liposome derivative. (B) Flow diagram of micro-fluidics based DNV synthesis. (C) Confocal laser scanning microscopy of AF647-ZOL-DNV. Approximately 100–200 nm DNV particles exhibited an AF647 signal. (D) Characterization of DNV formulations. (E) Protocol for intra-oral topical application to mouse palatal tissue. The reconstituted DNV solution in MQW (3 µl) was topically applied to the palatal gingiva between maxillary molar teeth and covered by a custom-made oral appliance fabricated with auto-polymerizing dental resin. After 1 hr, the oral appliance was removed. (F) After topical application of AF647-ZOL-DNV, mouse maxillary bones were harvested and AF647 fluorescence was measured. The AF647 fluorescent signal from the maxillary bone region of interest (ROI) increased with up to 75 µM AF7647-ZOL in DNV applied and then reached a plateau. (Figure 2—source data 1) (G) AF647-ZOL-DNV and AF647-ZOL in non-deformable formulation (AF647-ZOL-nDNV) were reconstituted in either MQW or 20% polyethylene glycol (PEG). AF647-ZOL-DNV in MQW most efficiently delivered the drug to the maxillary bone through trans-oral mucosal route. (Figure 2—source data 2) In (F) and (G), the graphs show the mean and SD (n=6 per group), and the Turkey test was used to analyze multiple samples. The statistical significance was determined to be at p<0.05. Different letters (e.g., a, b) are used to show statistically significant differences between multiple groups.

Figure 3 with 2 supplements
Disease phenotypes of mouse and human BRONJ lesion.

(A) Experimental protocol for inducing a bisphosphonate-related osteonecrosis of the jaw (BRONJ) lesion in mice. Mice received a bolus IV injection of zoledronate (ZOL) (40 nmol, 500 µg/kg) or vehicle saline solution were subjected to maxillary left first molar extraction. (B) Intra-oral photographs depicted that mouse pretreated with ZOL IV injection exhibited delayed wound healing with sustained open wound (black arrows) and gingival swelling (white arrows). (C) Human BRONJ lesions with open tooth extraction wound (black arrows) and gingival swelling (white arrows). (D) Micro-CT images of mouse maxilla depicted that delayed bone regeneration in the mesial, buccal and palatal root extraction sockets (white arrows) in ZOL-pretreated mice, a sign of BRONJ symptoms. (E) Radiographic demonstration of unhealed tooth extraction of human BRONJ. (F) Histological evaluation of mouse BRONJ lesion with osteonecrosis (dotted line; Nec), gingival inflammation (Inf) and epithelial hyperplasia reaching to the necrotic bone (black arrows). (G) A biopsy specimen of human BRONJ lesion with osteonecrosis (Nec) and gingival epithelial hyperplasia (black arrows). (H) A time course experimental diagram of hydroxymethylene diphosphonate (HMDP)-deformable nanoscale vesicles (DNV) application. All mice received ZOL IV injection. Empty-DNV, HMDP in MQW (HMDP alone) (5 nmol, 3 μl of 1.67 mM) or HMDP-DNV in MQW ( 5 nmol, 3 μl of 1.67 mM) was topically applied to the palatal gingiva prior to the maxillary left first molar extraction. (I) Micro-CT evaluation of tooth extraction socket, which remained empty (white arrows) in the groups treated with Empty-DNV or HMDP alone. Two topical applications of HMDP-DNV prior to the tooth extraction significantly increased tooth extraction socket bone regeneration compared to Empty-DNV and HMDP alone. (Figure 3—source data 1) (J) Histological evaluation. Two topical applications of HMDP-DNV prior to the tooth extraction reduced the development of osteonecrosis compared to the treatment of Empty-DNV and HMDP alone, which remained to exhibit BRONJ phenotype. (Figure 3—source data 2) In (I) and (J), the graphs show the mean and SD (n=3 per untreated control group and n=6 per experimental group), and the Turkey test was used to analyze multiple samples. The statistical significance was determined to be at p<0.05. Different letters (e.g., a, b) are used to show statistically significant differences between multiple groups.

Figure 3—figure supplement 1
Mouse BRONJ model.

(A) Time course protocol to develop bisphosphonate-related osteonecrosis of the jaw (BRONJ) in the mouse. Mice received a bolus IV injection of zoledronate (ZOL) (500 µg/kg) or vehicle saline solution and were subjected to maxillary left first molar extraction. The maxillary tissue was harvested 1, 2, and 4 weeks after the tooth extraction, photographed and scanned by micro-CT. (B) Three-dimensional reconstruction of micro-CT images showed mesial, buccal and palatal root extraction sockets. (C) Quantitative BV/TV measurement of extraction sockets of mesial, buccal, and palatal roots of micro-CT images demonstrated a significant delay of bone regeneration in ZOL-treated mice, a signifier of BRONJ. The statistical significance was determined to be at p<0.05. Different letters (e.g., a, b) are used to show statistically significant differences between multiple groups.

Figure 3—figure supplement 2
Single intra-oral topical application of hydroxymethylene diphosphonate (HMDP)-deformable nanoscale vesicles (DNV) to zoledronate (ZOL)-treated mice prior to tooth extraction did not prevent development of a BRONJ lesion.

(A) Time course experimental diagram. All mice received a ZOL IV injection. HMDP-DNV (5 nmol, 3 μl of 1.67 mM), Empty-DNV or HMDP in saline solution (HMDP alone, 5 nmol, 3 μl of 1.67 mM) was topically applied to the palatal gingiva prior to the maxillary left first molar extraction. Two weeks after tooth extraction, the maxillary tissue was harvested and subjected to micro-CT imaging and demineralized paraffine embedded histological section preparation. (B) A single application of HMDP-DNV, Empty-DNV or HMDP alone did not produce any improvement of bone regeneration as measured by BV/TV of micro-CT images. (C) Histological evaluation also indicated that the single application of HMDP-DNV, Empty-DNV and HMDP alone did not show reduction of histological osteonecrosis area.

Hydroxymethylene diphosphonate (HMDP)-deformable nanoscale vesicles (DNV) topical application to the established BRONJ lesion in mice accelerated the disease resolution.

(A) The time course experimental protocol. After zoledronate (ZOL) IV injection and tooth extraction, a BRONJ lesion developed, which was treated by two topical applications of HMDP-DNV in MQW at a high dose (15 nmol per application) for the 2-week tissue collection experiment and a lower dose (3.8 nmol per application) for the 4-week tissue collection experiment. (B) Intra-oral photographs depicted the unhealed tooth extraction wound (black arrows) with gingival swelling (dotted line) in significant percentage of animals in untreated mice group (‘No Tx’ group) from 2 weeks to 4 weeks after tooth extraction. By contrast, all HMDP-DNV-treated mice exhibited a closed wound in week-2, which was well healed in 4 weeks after tooth extraction. (C) The wound opening of untreated mice remained 4 weeks after tooth extraction, while HMDP-DNV topical treatment minimized the wound opening 2 weeks after tooth extraction. (Figure 4—source data 1) (D) The gingival swelling area was also decreased by HMDP-DNV topical treatment. (Figure 4—source data 2) (E) Micro-CT analysis showed the delayed bone regeneration in the extraction sockets (white arrows) of untreated ZOL-injected mice. However, HMDP-DNV-treated ZOL-injected mice accelerated the extraction socket bone regeneration, which was further remodeled to generate bone marrow trabecular structure (white arrow heads). The bone volume/total volume of the extraction socket showed the early bone regeneration in the group of HMDP-DNV treatment. (Figure 4—source data 3) (F) The histological osteonecrosis area progressively increased in the BRONJ lesion. HMDP-DNV topical treatment halted the osteonecrosis area increase at 2 weeks and decreased at 4 weeks after tooth extraction. (Figure 4—source data 4) (G) The topical application of HMDP-DNV did not affect distant skeletal tissue in femurs. (Figure 4—source data 5) In (C), (D), (E), (F) and (G), the graphs show the mean and SD (n=5–6 per group), and the Turkey test was used to analyze multiple samples. The statistical significance was determined to be at p<0.05. Different letters (e.g., a, b, et al.) are used to show statistically significant differences between multiple groups.

Hydroxymethylene diphosphonate (HMDP)-deformable nanoscale vesicles (DNV) topical application normalized tooth extraction wound healing of zoledronate (ZOL)-pretreated mice.

(A) Histological evaluation depicted BRONJ lesion at the tooth extraction site of the untreated ZOL-injected mice (No Tx) exhibiting that the abnormal epithelial hyperplasia (E. Hyp) extending to the necrotic alveolar bone (Nec) appeared to facilitate the sustained open wound, the uneven bone regeneration in the tooth extraction sockets (Soc) and localized infiltration of inflammatory cells (Inf) on the alveolar bone surface. The abnormal wound healing pattern was also observed at 4 weeks after tooth extraction with fistula formation by epithelial hyperplasia reaching to the necrotic bone and a localized pustule lesion (Pus). The unwounded side of ZOL-injected mice did not show any abnormality of remaining tooth (Tooth), alveolar bone (Bone) and overlining gingival tissue (Gingiva). After the HMDP-DNV topical treatment, gingival wound was found closed with more diffused inflammation (Inf) and a sign of bone resorption (white arrowheads) was depicted on the surface of alveola bone. The tooth extraction socket (Soc) showed bone regeneration at week-2, which was further remodeled and matured at week-4. (B) High magnification histology of 2 weeks after tooth extraction demonstrated the dense inflammatory cell infiltration (Inf) in the gingival connective tissue and osteonecrosis area (Nec) in untreated ZOL-injected mice (No Tx). The HMDP-DNV treatment attenuated the inflammatory cell infiltration and increased signs of osteoclastic bone resorption (white arrowheads). At 4 weeks after tooth extraction, the untreated mice continued to show dense inflammatory cell infiltration and osteonecrosis. However, mice with the HMDP-DNV treatment showed subsided inflammation and the minimized osteonecrosis area likely due to bone resorption, which appeared to result in alveolar bone loss (dotted white line and arrow) as compared to the regenerated bone in the tooth extraction socket (Soc). (C) Cathepsin K (Ctsk) immune-stained osteoclasts on the alveolar bone surface of the untreated ZOL-injected mice were observed not only at the proximal area (Zone A) of the tooth extraction socket (Soc) but also under the palatine neuro-vascular complex (Zone B) with inflammation (Inf). Characteristically, these Ctsk-positive osteoclasts were small and flattened (black arrowheads). In some specimens, Ctsk-positive cells were observed away from the bone surface (small arrowheads). The HMDP-DNV-treated mice showed large Ctsk-positive osteoclasts in deep bone lacunae adjacent to the tooth extraction socket (Zone A). (D) Total number of osteoclasts defined as Ctsk-positive multi-nuclear cells on the alveolar bone surface. (Figure 5—source data 1) (E) Osteoclasts in Zone A and Zone B were separately counted. (Figure 5—source data 2) In (D) and (E), the graphs show the mean and SD (n=5 per group), and the Turkey test was used to analyze multiple samples. The statistical significance was determined to be at p<0.05. Different letters (e.g., a, b) are used to show statistically significant differences between multiple groups.

Single-cell RNA sequencing of gingival cells of untreated and hydroxymethylene diphosphonate (HMDP)-deformable nanoscale vesicles (DNV)-treated zoledronate (ZOL)-injected mice.

(A) Two weeks after tooth extraction, gingival tissue adjacent to the tooth extraction site was harvested for cell dissociation followed by single cell RNA-sequencing. (B) Using signature gene expression, myeloid cells, T cells, and B cells were identified. (C) T cell-related gene expression indicated the presence of Cd8a+cytototoxic, Cd27+matured T cells derived from mouse bisphosphonate-related osteonecrosis of the jaw (BRONJ) gingiva. Il17f expression phenotype was decreased by HMDP-DNV treatment, which increased Foxp3 Treg phenotype. (D) Macrophage-related genes demonstrated an increase in M1 macrophages in untreated BRONJ gingiva, which was decreased by HMDP-DNV treatment. (E) Human BRONJ biopsy samples showing that necrotic bones (Nec) were associated with a large cluster of neutrophils (white arrows).

Characterization of myeloid immune cells.

(A) The myeloid cell fraction of scRNA-seq was further divided into neutrophils and macrophages. (B) Myeloid immune cells identified by Trem1 demonstrated a significant decrease of neutrophils by hydroxymethylene diphosphonate (HMDP)-deformable nanoscale vesicles (DNV) treatment. (C) Proinflammatory cytokines Il1a, Il1b, and Tnf were highly expressed in macrophages and neutrophils from bisphosphonate-related osteonecrosis of the jaw (BRONJ) gingiva. (D) Gingival macrophage and neutrophil myeloid cells after HMDP-DNV treatment expressed the multiple signature genes of the myeloid-derived suppressor cell (MDSC): Arg1, Arg2, CD84, Wfdc17, Ifitm1, Ifitm2, Clec4d, Clec4e, Ctsd, Cstdc4 (Alshetaiwi et al., 2020) or M2 macrophage phenotypes Arg1, Arg2, as well as anti-inflammatory cytokines Il1f9 and Il10. (E) A hypothetical model of BRONJ.

Tables

Table 1
Estimated MRONJ case numbers in the US based on MRONJ incidence for the major underlying diseases*.
Underlying diseasesNew cases in the US (Year)MRONJ incidenceEstimated MRONJ cases (Year)*
Multiple myeloma34,9205.16% (Rugani et al., 2016)1,802
Breast cancer330,8402.09% (Rugani et al., 2016)6,915
Prostate cancer248,5303.80% (Rugani et al., 2016)9,444
Osteoporosis/Low bone mass53,500,000§0.01% (Khan et al., 2017)5,350
Estimated annual incidents of MRONJ23,511
  1. MRONJ: Medication-related osteonecrosis in the jawbone.

  2. *

    Based on an assumption that all these patients were treated by antiresorptive medications.

  3. American Cancer Society. Cancer Facts & Figures 2021. Atlanta, Ga: American Cancer Society; 2021.

  4. Invasive breast cancer and ductal carcinoma: American Cancer Society. How Common Is Breast Cancer? Jan. 2020. Available at: https://www.cancer.org/cancer/breast-cancer/about/how-common-is-breast-cancer.html.

  5. §

    Center for Disease Control and Prevention at: https://www.cdc.gov/nchs/products/databriefs/db405htm.

Additional files

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Hiroko Okawa
  2. Takeru Kondo
  3. Akishige Hokugo
  4. Philip Cherian
  5. Jesus J Campagna
  6. Nicholas A Lentini
  7. Eric C Sung
  8. Samantha Chiang
  9. Yi-Ling Lin
  10. Frank H Ebetino
  11. Varghese John
  12. Shuting Sun
  13. Charles E McKenna
  14. Ichiro Nishimura
(2022)
Mechanism of bisphosphonate-related osteonecrosis of the jaw (BRONJ) revealed by targeted removal of legacy bisphosphonate from jawbone using competing inerthydroxymethylene diphosphonate
eLife 11:e76207.
https://doi.org/10.7554/eLife.76207