Effect of anti-angiogenesis induced by chemotherapeutic monotherapy, chemotherapeutic/bisphosphonate combination therapy and anti-VEGFA mAb therapy on tooth extraction socket healing in mice (original) (raw)

NAOSITE: Nagasaki University’s Academic Output SITE

1 Effect of anti-angiogenesis induced by chemotherapeutic monotherapy,

2 chemotherapeutic/bisphosphonate combination therapy and anti-VEGFA

3 mAb therapy on tooth extraction socket healing in mice

4 Yuri Akita 1{ }^{1}, Shinichiro Kuroshima 1,2,∗{ }^{1,2,}{ }^{*}, Kazunori Nakajima 1{ }^{1}, Hiroki Hayano 1{ }^{1}, Riho Kanai 1{ }^{1}, Muneteru Sasaki 1{ }^{1}, Takashi Sawase 1{ }^{1}
1{ }^{1} Department of Applied Prosthodontics, Graduate School of Biomedical Sciences, Nagasaki
7 University, 1-7-1, Sakamoto, Nagasaki, 852-8588, Japan
2{ }^{2} Oral & Maxillofacial Implant Center, Nagasaki University Hospital, 1-7-1, Sakamoto, Nagasaki, 852-8588, Japan

10

11 Yuri Akita and Shinichiro Kuroshima contributed equally to this work.

14 Corresponding author and person to whom reprints should be addressed:
15 Shinichiro Kuroshima, D.D.S, Ph.D
16 Department of Applied Prosthodontics, Graduate School of Biomedical Sciences, Nagasaki
17 University, 1-7-1, Sakamoto, Nagasaki, 852-8588, Japan.
18 Oral & Maxillofacial Implant Center, Nagasaki University Hospital, 1-7-1, Sakamoto, Nagasaki, 852-8588, Japan.
20 Phone: +81-95-819-7685.
21 Fax: +81-95-819-7689.
22 e-mail: kuroshima@nagasaki-u.ac.jp

24 Keywords: bisphosphonate, chemotherapy, blood vessels, osteonecrosis of the jaw, angiogenesis

Abstract

Osteonecrosis of the jaw (ONJ), which is a rare but severe adverse effect, mainly occurs in oncology patients receiving chemotherapeutic agents and bisphosphonates. However, the combined impact of chemotherapy and bisphosphonates on wound healing after tooth extraction remains unknown. The aim of this study was to determine the precise etiology of ONJ induced by chemotherapy and bisphosphonate combination therapy. Mice received zoledronate (ZA) monotherapy, cyclophosphamide (CY) monotherapy, or CY/ZA combination therapy. The maxillary first molars were extracted 3 weeks after the initiation of drug treatment. Moreover, antivascular endothelial growth factor A (VEGFA) monoclonal antibody (mAb) was administered once every 2 days just after tooth extraction for 2 weeks. Soft and hard tissue wound healing were evaluated 2 and 4 weeks post-extraction using histomorphometry, microcomputed tomography, and immunohistochemistry. ZA monotherapy did not induce impaired oral wound healing and ONJ-like lesions 2 and 4 weeks post-extraction, respectively. Tooth extraction socket healing worsened with severe anti-angiogenesis by CY monotherapy and CY/ZA combination therapy 2 weeks postextraction. However, CY monotherapy rarely induced ONJ-like lesions with severe angiogenesis suppression, whereas CY/ZA combination therapy frequently induced ONJ-like lesions with severe angiogenesis inhibition 4 weeks post-extraction. Interestingly, anti-VEGFA mAb therapy delayed osseous wound healing with normal soft tissue wound healing of tooth extraction sockets, although this therapy significantly suppressed blood vessel formation. Our findings suggest that antiangiogenesis alone is not the main cause of ONJ-like lesions induced by CY/ZA combination therapy. The combination of suppressed osteoclasts and anti-angiogenesis, in addition to other risk factors such as chemotherapy, may contribute to the development of ONJ.

1

Introduction

Osteonecrosis of the jaw (ONJ), which was first reported as “avascular necrosis of the jaws” [1], is a rare but severe adverse effect observed in patients taking antiresorptive drugs such as bisphosphonates and anti-receptor activator of nuclear factor kappa-B ligand antibody (denosumab) [2,3]. Recently, medication-related ONJ has been proposed in place of antiresorptive-related ONJ since anti-angiogenic agents such as bevacizumab also induced ONJ in oncology patients, although these patients have other risk factors such as chemotherapy and steroid therapy use [2]. ONJ significantly increases the Oral Health Impact Profile, resulting in a decreased patient oral healthrelated quality of life due to difficulty in receiving optimal dental treatments [4]. Approximately 70%70 \% of ONJ patients have received tooth extractions [3]. Tooth extraction injures soft and hard tissues with open wounds. Normally, wound repair occurs with new bone formation in extraction sockets and wound closure by soft tissue regeneration. Blood vessel formation is an essential factor in recruiting immune cells including polymorphonuclear cells (PMNs) and supporting bone formation and soft tissue reconstruction during healing processes. Several vascular endothelial growth factors (VEGFs), which are produced by macrophages, dendritic cells, fibroblasts, osteoblasts, and endothelial cells, are required for angiogenesis in wounds. Studies have reported that bisphosphonates inhibit blood vessel formation [5,6], suggesting that angiogenesis suppression by bisphosphonates is a pathoetiology of ONJ. However, some reports have indicated that bisphosphonate therapy does not suppress angiogenesis in oral wounds [7-10]. Moreover, it has not been reported that denosumab inhibits angiogenesis [3], although both bisphosphonates and denosumab cause the same pathophysiology in ONJ. Therefore, it remains controversial whether angiogenesis suppression is a crucial factor in inducing ONJ following tooth extraction.

ONJ primarily occurs in malignant patients taking not only antiresorptive drugs or antiangiogenic agents, but also one or more chemotherapeutic drugs [3]. Chemotherapy, which has been widely used for malignancies, is one of the risk factors for ONJ [2]. In particular, ONJ

1 develops most commonly in multiple myeloma patients who receive alkylating agents and 2 zoledronate combination therapy [11]. Alkylating agents, which are prescribed for multiple myeloma, bind to DNA and prevent proper DNA replication, resulting in cancer cell death [12]. Alkylating agents also induce adverse effects because they cause serious damage in both cancer cells and normal cells. However, whether alkylating agents affect soft and hard tissue wound healing following tooth extraction remains unknown. Moreover, the mechanisms by which 7 antiresorptive and chemotherapeutic combination therapy contribute to the development of ONJ remain unclear, although many clinical studies have reported that chemotherapeutic agents and 9 bisphosphonate and/or denosumab combination therapies induced ONJ following tooth extraction. Therefore, the aim of this study was to investigate the combined effect of chemotherapy and bisphosphonate therapy on soft and hard tissue wound healing following tooth extraction to elucidate the pathophysiology of ONJ.

14 Materials and methods

15 Animals, tooth extraction, and bisphosphonate therapy and chemotherapy
Male C57BL/6J mice (9-week-old; CLEA Japan Inc., Osaka, Japan) were randomly divided into three treatment groups and one control group (four groups total). Mice were subcutaneously administered zoledronate (Zometa, Novartis, Stein, Switzerland) at 0.05mg/kg0.05 \mathrm{mg} / \mathrm{kg} twice a week (designated as ZA,n=7\mathrm{ZA}, \mathrm{n}=7 ) [8]. Mice were intraperitoneally injected with 150mg/kg150 \mathrm{mg} / \mathrm{kg} of cyclophosphamide (C7397, Sigma-Aldrich, St. Louis, MO, USA) twice a week before tooth extraction and once a week post-extraction (designated as CY, n=7\mathrm{n}=7 ). Mice received combination treatment of intraperitoneal CY (150mg/kg(150 \mathrm{mg} / \mathrm{kg} twice a week before extraction and once a week postextraction) and subcutaneous ZA(0.05mg/kg\mathrm{ZA}(0.05 \mathrm{mg} / \mathrm{kg} twice a week) (designated as CY/ZA,n=7\mathrm{CY} / \mathrm{ZA}, \mathrm{n}=7 ). Saline was used as a control (vehicle control; designated as VC,n=7\mathrm{VC}, \mathrm{n}=7 ). Three weeks after the initiation of drug therapy, both maxillary first molars were extracted. Mice were euthanized 2 and 4 weeks post-

1 extraction (Fig. 1a, b). Animal care and experimental procedures were performed in accordance with the Guidelines for Animal Experimentation of Nagasaki University, with approval from the Ethics Committee for Animal Research

4 Anti-vascular endothelial growth factor A (VEGFA) monoclonal antibody (mAb) therapy
To investigate the net effect of anti-angiogenesis on tooth extraction wound healing, antiVEGFA mAb therapy was carried out. Female C57BL/6J mice (8-week-old; CLEA Japan Inc.) were used. Mice (n=4)(\mathrm{n}=4) received 100μ g100 \mu \mathrm{~g} of anti-VEGFA mAb (clone 2G11-2A05; BioLegend, San Diego, CA, USA) intraperitoneally once every 2 days for 2 weeks [13] just after extraction of both maxillary first molars. Intraperitoneal injection of an equivalent volume of rat IgG2a (BLD-400516; BioLegend, San Diego, CA, USA) was also performed as a control ( n=4\mathrm{n}=4 ). Euthanasia was performed 2 weeks post-extraction.

12 Microcomputed tomography (microCT)
Right maxillae were dissected 2 and 4 weeks after tooth extraction, fixed in 10%10 \% neutral buffered formalin for 24 hours, and visualized using microCT at 20−μm20-\mu \mathrm{m} voxel resolution and 90−kV90-\mathrm{kV} tube voltage (R_mCT2, Rigaku Co. Ltd., Tokyo, Japan) [14]. Extraction sockets were segmented and reconstructed using a semi-manual contouring method [15] with TRI/3D-Bon (Ratoc System Engineering, Tokyo, Japan). In tooth extraction sockets, bone fill of sockets (bone volume/tissue volume, BV/TV), trabecular number (Tb.N), trabecular thickness (Tb.Th), trabecular separation (Tb.Sp), and bone mineral density (BMD) were semi-automatically measured in accordance with the guidelines for assessment of bone microstructures using microCT [16].

21 Histomorphometric analysis of oral wound healing
Hematoxylin and eosin (H-E) staining was performed using a standard staining protocol to evaluate soft and hard tissues. Left maxillae were dissected 2 and 4 weeks after tooth extraction and fixed in 10%10 \% neutral formalin at euthanasia. Maxillary bones were demineralized in 10%10 \%

1 EDTA, paraffin embedded, and sectioned at 5−μm5-\mu \mathrm{m} thickness in serial sagittal sections. Tartrateresistant acid phosphatase (TRAP) staining and Masson’s trichrome staining were conducted to visualize osteoclasts and collagen fibers, respectively, with commercial kits following the manufacturer’s instructions (HT15 and 386A, respectively; Sigma-Aldrich). The bones in tooth extractions were then histomorphometrically assessed to detect the following: 1) osteoclast number per bone surface (osteoclast perimeter) (OC.N/BS, #/mm); 2) living bone area (%); 3) necrotic bone area (%); 4) number of empty lacunae ( #/mm2\# / \mathrm{mm}^{2} ); 5) osteocyte numbers ( #/mm2\# / \mathrm{mm}^{2} ); 6) collagen area
8(%)8(\%); and 7) PMN infiltration ( #/mm2\# / \mathrm{mm}^{2} ). Collagen fibers were quantified in the connective tissue [area of interest (AOI), 200μ m×500μ m200 \mu \mathrm{~m} \times 500 \mu \mathrm{~m} ] above extraction sockets. PMN infiltration was assessed by quantifying the number of inflammatory cells in the connective tissue within 100μ m100 \mu \mathrm{~m} of the bone surface (AOI, 100μ m×500μ m100 \mu \mathrm{~m} \times 500 \mu \mathrm{~m} ). Necrotic bone was defined as the portion of bone in which there was ≥10\geq 10 adjacent empty or pyknotic osteocyte lacunae, since such bone is not vital [8,17].

Immunohistochemical analyses

Immunofluorescent staining was performed to visualize blood vessels as follows. Sections were fixed, dehydrated, subjected to antigen retrieval, and blocked with nonspecific protein. Sections were then incubated with a CD31 rabbit anti-mouse primary antibody (1:100 dilution; ab56299, Abcam, Cambridge, MA, USA) overnight at 4∘C4^{\circ} \mathrm{C}. Fluorescent-conjugated goat anti-rat Alexa Flour 594 (1:200 dilution; Invitrogen, Carlsbad, CA, USA) was used as the secondary antibody. VECTASHIELD Antifade Mounting Medium with DAPI (H-1200 Vector Laboratories, Burlingame, CA, USA) was also used. Stained sections were visualized using immunofluorescent microscopy (Axio Scope A1, Zeiss, Oberkochen, Germany). Blood vessel numbers and areas were quantitatively analyzed by semiautomatically counting the number and area of vessels in the connective tissue above extraction sockets (AOI, 200μ m×500μ m200 \mu \mathrm{~m} \times 500 \mu \mathrm{~m} ) to evaluate blood vessel formation, as previously described [7-9].

Statistics

1 CY/ZA combination therapy significantly decreased bone fills in tooth extraction sockets with 2 decreased Tb.N, decreased Tb.Th and increased Tb.Sp, as compared with VC and ZA (Fig. 3b-e). BMD in CY was significantly decreased compared with VC, while BMD in CY/ZA was the same as VC (Fig. 3f).

Osseous wound healing of tooth extraction sockets appeared to be suppressed in CY and 6 CY/ZA 4 weeks post-extraction (Fig. 3g). CY/ZA combination therapy significantly suppressed bone fill in tooth extraction sockets with decreased Tb.N, decreased Tb.Th, increased Tb.Sp, and decreased BMD as compared with all other groups (Fig. 3h-l). CY monotherapy significantly induced impaired osseous wound healing compared with VC and ZA. Osseous wound healing in CY/ZA was significantly more impaired than that in CY (Fig. 3h-l).

11 Effect of administered drugs on soft and hard tissue wound healing 2 weeks post-extraction

Localized and extended open wounds were grossly detected in CY and CA/ZA\mathrm{CA} / \mathrm{ZA}, respectively (Fig. 4a). ZA monotherapy and CY monotherapy for 5 weeks significantly suppressed osteoclast numbers, even when CY/ZA combination therapy was administered (Fig. 4a, b). ZA monotherapy, CY monotherapy, and CY/ZA combination therapy significantly decreased the living bone area compared with VC (Fig. 4c). CY monotherapy and CY/ZA combination therapy significantly increased the necrotic bone area and the number of empty lacunae compared with VC and ZA monotherapy (Fig. 4d, e). Additionally, CY monotherapy and CY/ZA combination therapy significantly decreased the number of osteocytes compared with VC (Fig. 4f). CY/ZA combination therapy significantly suppressed collagen production compared with all other groups (Fig. 4g, h). CY monotherapy and CY/ZA combination therapy significantly increased PMN infiltration, although PMN infiltration in CY/ZA was severer than that in CY (Fig. 4g, i).

23 Effect of administered drugs on soft and hard tissue wound healing 4 weeks post-extraction

Open wounds were frequently observed following CY/ZA combination therapy but not VC, ZA monotherapy, or CY monotherapy (Fig. 5a). ZA administration significantly suppressed

1 osteoclast numbers in tooth extraction sockets, irrespective of CY administration (Fig. 5a, b). However, ZA monotherapy did not change the living bone area compared with VC (Fig. 5c). CY/ZA combination therapy and CY monotherapy significantly decreased the living bone area in extraction sockets compared with VC and ZA monotherapy (Fig. 5c). CY/ZA combination therapy significantly increased the necrotic bone area compared with all other therapies (Fig. 5d). The number of empty lacunae was the same between VC and ZA monotherapy, however, CY monotherapy and CY/ZA combination therapy significantly increased the number of empty lacunae compared with VC and ZA monotherapy (Fig. 5e). Additionally, CY monotherapy and CY/ZA combination therapy significantly decreased the number of osteocytes compared with VC (Fig 5f). CY/ZA combination therapy significantly decreased collagen production and increased PMN infiltration in soft tissues of tooth extraction sockets compared with VC, ZA monotherapy, and CY monotherapy (Fig. 5g-i).

Effect of administered drugs on blood vessels in the connective tissue of tooth extraction

14
sockets 2 and 4 weeks post-extraction

Blood vessel formation was measured to compare the effect of CY monotherapy and CY/ZA combination therapy. Surprisingly, CY monotherapy and CY/ZA combination therapy significantly suppressed the number and area of blood vessels as compared with VC and ZA monotherapy 2 weeks post-extraction (Fig. 6a-c). Interestingly, these anti-angiogenic effects by CY monotherapy and CY/ZA combination therapy were significantly sustained until 4 weeks postextraction (Fig. 6d-f).

21 Effect of anti-VEGFA mAb therapy on wound healing in tooth extraction sockets

No open wounds were observed in both IgG2a and anti-VEGFA mAb 2 weeks postextraction (Fig. 7b, c). Anti-VEGFA mAb therapy significantly delayed bone formation of tooth extraction sockets compared with IgG2a therapy, although the necrotic bone area and numbers of empty lacunae and osteocytes were the same between groups (Fig. 7d-h). Conversely, collagen

production was the same between groups, although PMN infiltration was significantly increased in anti-VEGFA mAb compared with IgG2a (Fig.7i-k). Anti-VEGFA mAb therapy significantly suppressed the number of blood vessels without alteration of vessel surfaces (Fig. 7l-n).

Discussion

In the present study, we demonstrated that: 1) severe angiogenesis suppression by alkylating chemotherapeutic agent CY alone is a not the definitive cause of ONJ-like lesions, although CY monotherapy delayed osseous wound healing; 2) CY/ZA combination therapy frequently induced ONJ-like lesions, although CY/ZA combination therapy also significantly inhibited angiogenesis in the soft tissue of tooth extraction sockets; and 3) anti-VEGFA mAb therapy delayed osseous wound healing but not soft tissue healing of tooth extraction sockets.

Chemotherapeutic agents, which are pivotal drugs for the treatment of malignancies, have adverse effects on normal cells. The alkylating agent CY constitutes a major class of front-line chemotherapeutic agents that inflict cytotoxic DNA damage as their mode of action. CY damages both cancer cells and normal cells by inhibiting DNA replication [18]. To our knowledge, no reports have focused on the effect of CY monotherapy on wound healing following tooth extraction. Therefore, in this study, we investigated the effect of CY monotherapy on hard and soft tissue wound healing following tooth extraction. CY monotherapy influenced osteoclasts and osteocytes with decreased bone fill and decreased BMD after CY administration. This finding was partially in accordance with that of a previous study, which demonstrated that weekly 20mg/kg20 \mathrm{mg} / \mathrm{kg} CY injections for more than 5 weeks significantly decreased osteoblast and osteoclast numbers with bone loss in rat mandibular condyles [19]. CY induces myelosuppression [20], suggesting that immune responses may be influenced. Bacteremia may occur in up to 96%96 \% of cases following bacterial invasion into extraction sockets [21,22]. Elimination of invasive bacteria in extraction sockets supports wound closure. Network development of blood vessels in wound sites plays an important

role in accumulating neutrophils, osteoclast precursors, monocyte-macrophage lineages, and PMNs
during wound healing [23]. Neutrophils, PMNs, and monocyte-macrophage lineages fight invading
bacteria within soft tissue wounds, resulting in soft tissue regeneration [24,25]. In the current study,
the influence of CY on PMN infiltration and collagen production in the soft tissue of extraction
sockets decreased in a time-dependent manner, although CY significantly suppressed angiogenesis,
irrespective of administration duration. CY has been demonstrated to inhibit angiogenesis by
preventing endothelial cell growth through suppressed basic fibroblast growth factor expression in vitro and in vivo [26]. Conversely, a recent study showed that CY metronomic chemotherapy,
which suppressed only angiogenesis without immune suppression, has been developed for use in oncology patients [27,28]. Moreover, CY enhanced the recruitment of immune cells such as PMNs and macrophages via high mobility group box 1 protein [29]. Therefore, our findings suggested that
CY monotherapy may specifically inhibit blood vessel formation, but not affect immune cells such as PMNs. As a result, CY monotherapy rarely induced ONJ-like lesions with severe angiogenesis suppression, although this therapy induced impaired osseous wound healing 2 and 4 weeks postextraction.

ONJ develops most commonly in malignant patients taking intravenous bisphosphonates following tooth extraction with incidences ranging 1.6−14.8%1.6-14.8 \% [3]. Malignant patients may take multiple chemotherapeutic agents simultaneously. Thus, we investigated the effect of ZA and CY combination therapy on tooth extraction socket healing. First, the effect of ZA monotherapy was examined. ZA is usually administered intravenously in humans. It has been demonstrated that subcutaneous injection of ZA(0.1mg/kg\mathrm{ZA}(0.1 \mathrm{mg} / \mathrm{kg} at 5 times for 20 days) improved the survival rate without reducing the number of metastases in a murine prostate cancer bone metastasis model. This therapy also significantly increased trabecular bone with the preservation of bone structures in long bones [30]. These effects of subcutaneous ZA injection on tumor-bearing mice are comparable to those of intravenous ZA injection. Hence, in this study, subcutaneous injection of ZA was applied to

1 C57B6/J mice. Interestingly, ZA monotherapy did not induce ONJ-like lesions with open wounds, although ZA monotherapy affected soft and hard tissue wound healing 2 weeks, but not 4 weeks, post-extraction. ZA did not affect the number of blood vessels, regardless of administration duration. We previously demonstrated that ZA monotherapy did not suppress angiogenesis in the soft tissue of extraction sockets during wound healing processes [8,9]. Thus, the present findings are in agreement with our previous findings. Conversely, ZA has been demonstrated to have a negative effect on cultured oral epithelial cells and fibroblasts [31]. Additionally, ZA had a negative influence on cell viability of human umbilical vascular endothelial cells [32]. Short-term ZA monotherapy (12 days) inhibited tooth extraction socket healing with inhibited angiogenesis 5 days post-extraction [33], although the most recent study has demonstrated that subcutaneous alendronate monotherapy did not suppress angiogenesis with normal wound healing 3,5,7,103,5,7,10, and 21 days post-extraction [10]. International task force on ONJ also agrees with no anti-angiogenesis with animal study using bisphosphonates [2]. However, ZA monotherapy actually suppressed angiogenesis in cancer patients [34] and ONJ mainly occurs in malignant patients taking ZA. Accordingly, ZA-induced anti-angiogenesis may not be an exact causative factor of BRONJ in human, although inhibition of blood vessel formation by ZA monotherapy negatively affects the early phase of wound healing processes following tooth extraction. Anti-angiogenesis induced by ZA monotherapy may specifically occur at malignant sites but not at wound sites, such as soft tissues of extraction sockets.

Next, we evaluated the effect of CY and ZA combination therapy. The effects of CY/ZA combination therapy on tooth extraction wound healing significantly differed from that of ZA monotherapy 2 and 4 weeks post-extraction. CY/ZA combination therapy severely impaired hard tissue healing. Surprisingly, only 20%20 \% of bone fill in tooth extraction sockets was observed following CY/ZA combination therapy even 4 weeks post-extraction. Open wounds frequently remained with severely reduced collagen production and significant PMNs infiltration in soft tissue

1 wounds. As a result, CY/ZA combination therapy severely influenced soft and hard tissues of
2 extraction sockets, resulting in frequent induction of ONJ-like lesions. Moreover, angiogenesis inhibition following CY monotherapy and CY/ZA combination therapy was almost the same. Thus, our findings demonstrated that both CY and ZA are requisite for the development of ONJ-like lesions in mice.

6 Finally, to address whether anti-angiogenesis is a critical factor in wound healing following 7 tooth extraction, we investigated the net effect of anti-angiogenesis on tooth extraction socket healing at the early phase of wound healing processes using anti-VEGFA mAb. Blood vessels are required for the recruitment of immune cells to tooth extraction sites, and bone formation and soft tissue regeneration in the tooth extraction sockets. Hence, anti-VEGFA mAb therapy delayed osseous wound healing of tooth extraction sockets. Other factors except for anti-angiogenesis may induce normal soft tissue healing of tooth extraction sockets without inhibition of collagen production, although which factors positively affect soft tissue healing remains unknown in the present study. It has been reported that ONJ occurred in cancer patients treated with an antiangiogenic agent, bevacizumab, although these patients already had many other risk factors for ONJ such as the use of chemotherapeutic drugs and abnormal immune responses [35]. This report concluded that the incidence of ONJ in metastatic breast cancer patients treated with bisphosphonates was not increased even when bevacizumab was used in combination with bisphosphonate therapies [35]. A recent study showed that suppression of blood vessel formation may not be a crucial factor in the development of ONJ [36]. Additionally, normal angiogenesis has been reported in most histological studies of bisphosphonate-related ONJ [37,38]. It has not been reported whether denosumab has anti-angiogenic effects on tooth extraction sockets. Hence, our findings and accumulating scientific data support the finding that suppressed angiogenesis alone is not a definitive cause of ONJ, although an anti-angiogenic effect may negatively influence osseous wound healing of tooth extraction. It has been reported that administration duration of anti-

1
angiogenesis agents such as bevacizumab was relatively long in ONJ patients (e.g., one report was 6 months, another report was 2 years) [39,40]. Hence, further animal studies investigating the effects of long-term anti-VEGFA mAb therapy on tooth extraction socket healing are required to clearly manifest the relationship between anti-angiogenesis and BRONJ in mice.

In summary, we demonstrated that suppressed angiogenesis did not induce ONJ-like lesions in mice. The combination of osteoclast suppression and angiogenesis inhibition, in addition to other factors such as the use of chemotherapeutic drugs and compromised immune reactions, may be essential for the development of ONJ. Moreover, ZA monotherapy did not evoke ONJ-like lesions in mice, unlike in humans, indicating that caution should be exercised during the evaluation of impaired wound healing when using experimental rodent models.

12 Acknowledgements

This work was supported by JSPS KAKENHI Grant Nos. 25870523 and 15K11258.

15 Conflict of interest

The authors declare that they have no conflict of interest.

18 References

1. Marx RE (2003) Pamidronate (Aredia) and zoledronate (Zometa) induced avascular necrosis of the jaws: a growing epidemic. J Oral Maxillofac Surg 61:1115-7

2. Khan AA, Morrison A, Hanley DA, Felsenberg D, McCauley LK et al. (2015) Diagnosis and management of osteonecrosis of the jaw: a systematic review and international consensus. J Bone Miner Res 30:3-23 doi:10.1002/jbmr. 2405

3. Ruggiero SL, Dodson TB, Fantasia J, Goodday R, Aghaloo T, Mehrotra B, O’Ryan F, Surgeons AAOSM (2014) American Association of Oral and Maxillofacial Surgeons position paper on medication-related osteonecrosis of the jaw–2014 update . J Oral Maxillofac Surg 72:1938-56 doi:10.1016/j.joms.2014.04.031

4. Miksad RA, Lai KC, Dodson TB, Woo SB, Treister NS, Akinyemi O, Bihrle M, Maytal G, August M, Gazelle GS, Swan JS (2011) Quality of life implications of bisphosphonateassociated osteonecrosis of the jaw. Oncologist 16:121-32 doi:10.1634/theoncologist.20100183

5. Bezzi M, Hasmim M, Bieler G, Dormond O, Rüegg C (2003) Zoledronate sensitizes endothelial cells to tumor necrosis factor-induced programmed cell death: evidence for the suppression of sustained activation of focal adhesion kinase and protein kinase B/Akt. J Biol Chem 278:4360314 doi:10.1074/jbc.M308114200

6. Sharma D, Hamlet SM, Petcu EB, Ivanovski S (2016) The effect of bisphosphonates on the endothelial differentiation of mesenchymal stem cells. Sci Rep 6:20580 doi:10.1038/srep20580

7. Kuroshima S, Go VA, Yamashita J (2012) Increased numbers of nonattached osteoclasts after long-term zoledronic acid therapy in mice. Endocrinology 153:17-28 doi: 10.1210/en.20111439

8. Kuroshima S, Yamashita J (2013) Chemotherapeutic and antiresorptive combination therapy suppressed lymphangiogenesis and induced osteonecrosis of the jaw-like lesions in mice. Bone 56:101-9 doi:10.1016/j.bone.2013.05.013

9. Kuroshima S, Mecano RB, Tanoue R, Koi K, Yamashita J (2014) Distinctive tooth-extraction socket healing: bisphosphonate versus parathyroid hormone therapy. J Periodontol 85:24-33 doi:10.1902/jop.2013.130094

10. Tanoue R, Koi K, Yamashita J (2015) Effect of Alendronate on Bone Formation during Tooth Extraction Wound Healing. J Dent Res 94:1251-8 doi:10.1177/0022034515592867

11. Fliefel R, Tröltzsch M, Kühnisch J, Ehrenfeld M, Otto S (2015) Treatment strategies and outcomes of bisphosphonate-related osteonecrosis of the jaw (BRONJ) with characterization of patients: a systematic review. Int J Oral Maxillofac Surg 44:568-85 doi:10.1016/j.ijom.2015.01.026

12. Fu D, Calvo JA, Samson LD (2012) Balancing repair and tolerance of DNA damage caused by alkylating agents. Nat Rev Cancer 12:104-20 doi:10.1038/nrc3185

13. Lu R, Kujawski M, Pan H, Shively JE (2012) Tumor angiogenesis mediated by myeloid cells is negatively regulated by CEACAM1. Cancer Res 72:2239-50 doi:10.1158/0008-5472.CAN-113016

14. Sasaki M, Kuroshima S, Aoki Y, Inaba N, Sawase T (2015) Ultrastructural alterations of osteocyte morphology via loaded implants in rabbit tibiae. J Biomech 48:4130-41 doi:10.1016/j.jbiomech.2015.10.025

15. Kuroshima S, Kovacic BL, Kozloff KM, McCauley LK, Yamashita J (2013) Intra-oral PTH

2 administration promotes tooth extraction socket healing. J Dent Res 92:553-9 doi:10.1177/0022034513487558
16. Bouxsein ML, Boyd SK, Christiansen BA, Guldberg RE, Jepsen KJ, Müller R (2010) Guidelines for assessment of bone microstructure in rodents using micro-computed tomography. J Bone Miner Res 25:1468-86 doi:10.1002/jbmr. 141
17. Yamashita J, Koi K, Yang DY, McCauley LK (2011) Effect of zoledronate on oral wound healing in rats. Clin Cancer Res 17:1405-14 doi: 10.1158/1078-0432.CCR-10-1614
18. Cheung-Ong K, Giaever G, Nislow C (2013) DNA-damaging agents in cancer chemotherapy: serendipity and chemical biology. Chem Biol 20:648-59 doi:10.1016/j.chembiol.2013.04.007
19. Wang TM, Shih C (1986) Study of histomorphometric changes of the mandibular condyles in neonatal and juvenile rats after administration of cyclophosphamide. Acta Anat (Basel) 127:939
20. Vadhan-Raj S, Broxmeyer HE, Hittelman WN, Papadopoulos NE, Chawla SP, Fenoglio C, Cooper S, Buescher ES, Frenck RW, Holian A (1992) Abrogating chemotherapy-induced myelosuppression by recombinant granulocyte-macrophage colony-stimulating factor in patients with sarcoma: protection at the progenitor cell level. J Clin Oncol 10:1266-77
21. Okabe K, Nakagawa K, Yamamoto E (1995) Factors affecting the occurrence of bacteremia associated with tooth extraction. Int J Oral Maxillofac Surg 24:239-42
22. Tomás I, Alvarez M, Limeres J, Potel C, Medina J, Diz P (2007) Prevalence, duration and aetiology of bacteraemia following dental extractions. Oral Dis 13:56-62 doi:10.1111/j.16010825.2006.01247.x
23. Singer AJ, Clark RA (1999) Cutaneous wound healing. N Engl J Med 341:738-46 doi:10.1056/NEJM199909023411006
24. Fournier BM, Parkos CA (2012) The role of neutrophils during intestinal inflammation. Mucosal Immunol 5:354-66 doi:10.1038/mi. 2012.24
25. Leoni G, Neumann PA, Sumagin R, Denning TL, Nusrat A (2015) Wound repair: role of immune-epithelial interactions. Mucosal Immunol 8:959-68 doi:10.1038/mi. 2015.63
26. Browder T, Butterfield CE, Kräling BM, Shi B, Marshall B, O’Reilly MS, Folkman J (2000) Antiangiogenic scheduling of chemotherapy improves efficacy against experimental drugresistant cancer. Cancer Res 60:1878-86

1. Hanahan D, Bergers G, Bergsland E (2000) Less is more, regularly: metronomic dosing of cytotoxic drugs can target tumor angiogenesis in mice. J Clin Invest 105:1045-7 doi:10.1172/JCI9872

2. Kerbel RS, Kamen BA (2004) The anti-angiogenic basis of metronomic chemotherapy. Nat Rev Cancer 4:423-36 doi:10.1038/nrc1369

3. Guerriero JL, Ditsworth D, Catanzaro JM, Sabino G, Furie MB, Kew RR, Crawford HC, Zong WX (2011) DNA alkylating therapy induces tumor regression through an HMGB1-mediated activation of innate immunity. J Immunol 186:3517-26 doi:10.4049/jimmunol. 1003267

4. Ji Hung TT, Chan J, Russell PJ, Power CA (2011) Zoledronic acid preserves bone structure and increases survival but does not limit tumour incidence in a prostate cancer bone metastasis model. PLoS One 6:e19389 doi:10.1371/journal.pone. 0019389

5. Ravosa MJ, Ning J, Liu Y, Stack MS (2011) Bisphosphonate effects on the behaviour of oral epithelial cells and oral fibroblasts. Arch Oral Biol 56:491-8 doi:10.1016/j.archoralbio.2010.11.003

6. Walter C, Klein MO, Pabst A, Al-Nawas B, Duschner H, Ziebart T (2010) Influence of bisphosphonates on endothelial cells, fibroblasts, and osteogenic cells. Clin Oral Investig 14:3541 doi:10.1007/s00784-009-0266-4

7. Kobayashi Y, Hiraga T, Ueda A, Wang L, Matsumoto-Nakano M, Hata K, Yatani H, Yoneda T (2010) Zoledronic acid delays wound healing of the tooth extraction socket, inhibits oral epithelial cell migration, and promotes proliferation and adhesion to hydroxyapatite of oral bacteria, without causing osteonecrosis of the jaw, in mice. J Bone Miner Metab 28:165-75 doi:10.1007/s00774-009-0128-9

8. Santini D, Vincenzi B, Dicuonzo G, Avvisati G, Massacesi C, Battistoni F, Gavasci M, Rocci L, Tirindelli MC, Altomare V, Tocchini M, Bonsignori M, Tonini G (2003) Zoledronic acid induces significant and long-lasting modifications of circulating angiogenic factors in cancer patients. Clin Cancer Res 9:2893-7

9. Guarneri V, Miles D, Robert N, Diéras V, Glaspy J, Smith I, Thomssen C, Biganzoli L, Taran T, Conte P (2010) Bevacizumab and osteonecrosis of the jaw: incidence and association with bisphosphonate therapy in three large prospective trials in advanced breast cancer. Breast Cancer Res Treat 122:181-8 doi:10.1007/s10549-010-0866-3

10. Francini F, Pascucci A, Francini E, Miano ST, Bargagli G, Ruggiero G, Petrioli R (2011) Osteonecrosis of the jaw in patients with cancer who received zoledronic acid and bevacizumab. J Am Dent Assoc 142:506-13

11. Lesclous P, Abi Najm S, Carrel JP, Baroukh B, Lombardi T, Willi JP, Rizzoli R, Saffar JL, Samson J (2009) Bisphosphonate-associated osteonecrosis of the jaw: a key role of inflammation?. Bone 45:843-52 doi:10.1016/j.bone.2009.07.011

12. Hansen T, Kunkel M, Weber A, James Kirkpatrick C (2006) Osteonecrosis of the jaws in patients treated with bisphosphonates - histomorphologic analysis in comparison with infected osteoradionecrosis. J Oral Pathol Med 35:155-60 doi:10.1111/j.1600-0714.2006.00391.x

13. Hopp RN, Pucci J, Santos-Silva AR, Jorge J (2012) Osteonecrosis after administration of intravitreous bevacizumab. J Oral Maxillofac Surg 70:632-5 doi:10.1016/j.joms.2011.02.104

14. Santos-Silva AR, Belizário Rosa GA, Castro Júnior G, Dias RB, Prado Ribeiro AC, Brandão TB (2013) Osteonecrosis of the mandible associated with bevacizumab therapy. Oral Surg Oral Med Oral Pathol Oral Radiol 115:e32-6 doi:10.1016/j.oooo.2013.02.001

14 Figure legends

15 Fig. 1 Experimental design. (a) Time schedules for ZA monotherapy, CY monotherapy, and CY/ZA combination therapy. Tooth extraction was performed 3 weeks after the onset of drug treatments. Euthanasia was conducted 2 and 4 weeks post-extraction. VC: vehicle control, CY: cyclophosphamide, ZA: zoledronate. (b) Both maxillary first molars (M1) were selected as tooth extraction sites. M2 and M3 indicate second and third molars, respectively.

Fig. 2 Effect of administered drugs on gross wound healing in tooth extraction sockets. (a) Representative gross wound healing 2 weeks post-extraction. Black dotted lines indicate second (M2) and third molars (M3). Red dotted lines indicate the outlines of wounds. (b) No open wounds were observed in VC and ZA. Open wounds were observed in 64.2%64.2 \% of CY and 100%100 \% of CY/ZA. © Increased wound areas were observed in CY and CY/ZA compared with VC and ZA, although the wound area in CY/ZA was significantly larger than that in CY. (d) Representative gross wound healing 4 weeks post-extraction. Black dotted lines indicate second (M2) and third molars (M3). Red dotted lines indicate the outlines of wounds. (e) No open wounds were observed in VC and ZA.

1 Open wounds were noted in only 7.14%7.14 \% of CY and 100%100 \% of CY/ZA. (f) The wound area in CY/ZA
2 was significantly larger than that in all other groups. n=7/\mathrm{n}=7 / group; ∗,p<0.05,∗∗∗,p<0.001{ }^{*}, p<0.05,{ }^{* * *}, p<0.001, compared with VC. †,p<0.05,††,p<0.01,†††,p<0.001{ }^{\dagger}, p<0.05,{ }^{\dagger \dagger}, p<0.01,{ }^{\dagger \dagger \dagger}, p<0.001, compared with other groups except for 4 VC.

5 Fig. 3 Effect of administered drugs on osseous wound healing 2 and 4 weeks post-extraction. (a) Representative cross-sectional microcomputed tomography (microCT) images of tooth extraction 7 sockets 2 weeks post-extraction. M1 and M2 indicate first molars and second molars, respectively. 8 (b) Bone fill of tooth extraction sockets in CY and CY/ZA was significantly impaired compared 9 with VC and ZA. © Trabecular number (Tb.N) in CY and CY/ZA was significantly decreased compared with VC and ZA. (d) Trabecular thickness (Tb.Th) was significantly smaller in CY compared with VC and ZA. Tb.Th was smaller in CY/ZA compared with VC without significant difference. (e) Trabecular separation (Tb.Sp) was significantly greater in CY and CY/ZA compared 13 with VC and ZA. (1) Increased bone mineral density (BMD) was observed in ZA compared with other groups. Decreased BMD was noted in CY and CY/ZA compared with VC. (g) Representative cross-sectional microCT images of tooth extraction sockets 4 weeks post-extraction. Yellow dotted lines indicate the mesial and distal roots of tooth extraction sockets. (h) Bone fill of tooth extraction 17 sockets in CY and CY/ZA was significantly impaired compared with VC. Significantly greater bone fill was noted in ZA compared with VC. (i) Tb.N in ZA, CY, and CY/ZA was significantly decreased compared with VC. (j) Tb.Th was significantly smaller in CY/ZA compared with VC, whereas Tb.Th was greater in ZA compared with VC. (k) Tb.Sp was significantly greater in CY and CY/ZA compared with VC and ZA. (1) Increased BMD was observed in ZA compared with VC. Decreased BMD was noted in CY and CY/ZA compared with VC. n=7/\mathrm{n}=7 / group; ∗,p<0.05,∗∗,p<{ }^{*}, p<0.05,{ }^{* *}, p< 0.01,∗∗∗,p<0.0010.01,{ }^{* * *}, p<0.001, compared with VC. †,p<0.05,††,p<0.01,†††,p<0.001{ }^{\dagger}, p<0.05,{ }^{\dagger \dagger}, p<0.01,{ }^{\dagger \dagger \dagger}, p<0.001, compared with 24 other groups except for VC.

1 Fig. 4 Effect of administered drugs on tooth extraction socket healing 2 weeks post-extraction. (a) Representative sagittal H-E- and TRAP-stained images of tooth extraction sockets. Bar: 200μ m200 \mu \mathrm{~m} and 50μ m50 \mu \mathrm{~m}, respectively. In H-E-stained images, localized and extended open wounds were grossly observed in CY and CY/ZA, respectively. In TRAP-stained images, black arrowheads indicate osteoclasts on bone surfaces. (b) Osteoclast numbers were significantly decreased in ZA, CY, and CY/ZA compared with VC. © The living bone area was significantly decreased in CY and CY/ZA compared with VC and ZA. (d) The necrotic bone area was significantly increased in CY and CY/ZA compared with VC and ZA. (e) The number of empty lacunae was significantly increased in CY and CY/ZA compared with VC and ZA. (f) The number of osteocytes was significantly decreased in CY and CY/ZA compared with VC and ZA. (g) Representative trichrome-stained images of tooth extraction sockets. Bar: 200μ m200 \mu \mathrm{~m}. Yellow dotted lines indicate the border between alveolar bone and connective tissues. (h) The collagen area was significantly decreased in CY/ZA compared with all other groups. (i) PMN infiltration in CY and CY/ZA was significantly decreased compared with that in VC and ZA. Significantly more PMN infiltration was observed in CY/ZA compared with CY. n=7/\mathrm{n}=7 / group; ∗,p<0.05,∗∗,p<0.01,∗∗∗,p<0.001{ }^{*}, p<0.05,{ }^{* *}, p<0.01,{ }^{* * *}, p<0.001, compared with VC. †,p<{ }^{\dagger}, p< 0.05,††,p<0.01,†††,p<0.0010.05,{ }^{\dagger \dagger}, p<0.01,{ }^{\dagger \dagger \dagger}, p<0.001, compared with other groups except for VC.

17 Fig. 5 Effect of administered drugs on tooth extraction socket healing 4 weeks post-extraction. (a) Representative sagittal H-E- and TRAP-stained images of tooth extraction sockets. Bar: 200μ m200 \mu \mathrm{~m} and 50μ m50 \mu \mathrm{~m}, respectively. In H-E-stained images, extended open wounds were grossly observed in CY/ZA. In TRAP-stained images, black arrowheads indicate osteoclasts on bone surfaces. (b) Osteoclast numbers were significantly decreased in ZA and CY/ZA compared with VC and CY. © The living bone area was significantly decreased in CY and CY/ZA compared with VC and ZA. (d) The necrotic bone area was significantly increased in CY/ZA compared with all other groups. (e) The number of empty lacunae was significantly increased in CY and CY/ZA compared with VC. (f)

1 The number of osteocytes was significantly decreased in CY and CY/ZA compared with VC. (g) Representative trichrome-stained images of tooth extraction sockets. Bar: 200μ m200 \mu \mathrm{~m}. Yellow dotted 3 lines indicate the border between alveolar bone and connective tissues. (h) Collagen area was significantly decreased in CY/ZA compared with all other groups. (i) A significantly greater number of PMNs were observed in CY/ZA compared with all other groups. n=7/\mathrm{n}=7 / group; ∗,p<0.05{ }^{*}, p<0.05, ∗∗∗,p<0.001{ }^{* * *}, p<0.001, compared with VC. ††,p<0.01,†††,p<0.001{ }^{\dagger \dagger}, p<0.01,{ }^{\dagger \dagger \dagger}, p<0.001, compared with other groups except 7 for VC.

8 Fig. 6 Effect of administered drugs on the formation of blood vessels in soft tissue wounds. (a) Representative immunostaining using anti-CD31 antibody 2 weeks post-extraction. Nuclei were stained with DAPI. White arrowheads indicate CD31-positive vessels. Ep: epithelium, Ct: connective tissue, Ab: alveolar bone, Bar: 100μ m100 \mu \mathrm{~m}. (b) The number of blood vessels was significantly decreased in CY and CY/ZA compared with VC and ZA. The number of blood vessels was the same between CY and CY/ZA. © Vessel surface was significantly decreased in ZA, CY, and CY/ZA compared with VC. (d) Representative immunostaining using anti-CD31 antibody 4 weeks post-extraction. Nuclei were stained with DAPI. White arrowheads indicate CD31-positive vessels. Bar: 100μ m100 \mu \mathrm{~m}. (e) The number of blood vessels was the same between VC and ZA but significantly decreased in CY and CY/ZA compared with VC and ZA. (f) Vessel area was the same between VC and ZA but significantly smaller in CY and CY/ZA compared with VC and ZA. n=\mathrm{n}= 7/group; *, p<0.05,∗∗,p<0.01,∗∗∗,p<0.001p<0.05,{ }^{* *}, p<0.01,{ }^{* * *}, p<0.001, compared with VC. †,p<0.05,†††,p<0.001{ }^{\dagger}, p<0.05,{ }^{\dagger \dagger \dagger}, p<0.001, compared with other groups except for VC.

21 Fig. 7 Effect of anti-VEGFA mAb on tooth extraction socket healing. (a) Time schedule for IgG2a and anti-VEGFA mAb therapy. IgG2a and mAb therapies were performed once every 2 days just after tooth extraction for 2 weeks. (b) Representative gross wound healing 2 weeks post-extraction. M2 and M3 indicate the second and third molars, respectively. © No open wounds were observed

1 following both IgG2a and mAb therapies. (d) Representative H-E-stained images of tooth extraction 2 sockets. Open wounds and necrotic bone were not grossly observed. Bar: 200μ m200 \mu \mathrm{~m}. (e) The living 3 bone area was significantly decreased in mAb compared with IgG2a. (f-h) The necrotic bone area, 4 number of empty lacunae, and osteocyte number were the same between IgG2a and mAb . (i) Representative trichrome-stained images of tooth extraction sockets. Yellow dotted lines indicate 6 the border between alveolar bone and connective tissues. Bar: 200μ m200 \mu \mathrm{~m}. (j) Collagen production was 7 the same between IgG2a and mAb . (k) Significant PMN infiltration was observed in mAb 8 compared with IgG2a. (1) Representative immunostaining for CD31. Nuclei were stained with 9 DAPI. White arrowheads indicate CD31-positive vessels. Bar: 100μ m100 \mu \mathrm{~m}. (m) The number of blood vessels was significantly decreased in mAb compared with IgG2a. (n) Vessel surface was similar between IgG2a and mAb. n=4/\mathrm{n}=4 / group; ∗∗,p<0.01{ }^{* *}, p<0.01.

FIG. 1.
a

b
Anterior

Posterior

FIG. 2.
a

b
c

d

e
f

FIG. 3.
a

b

g

FIG. 4.
a

FIG. 5.
a

FIG. 6.
a

b

d

e



f

FIG. 7.
a

d